200b: x32 LPDDR4 SDRAM
Features
LPDDR4 SDRAM
RS384M32LZ4D2ANP, RS768M32LZ4D4ANQ
Features
Options
Marking
• VDD1/VDD2: 1.8V/1.1V
• Array configuration
– 384 Meg x 32 (2 channels x16 I/O)
– 768 Meg x 32 (4 channels x16 I/O)
• Device configuration
– 384M16 x 2 die in package
– 384M16 x 4 die in package
• FBGA “green” package
– 200-ball WFBGA (10mm x 14.5mm x
0.8mm, Ø0.28 SMD )
– 200-ball VFBGA (10mm x 14.5mm x
0.95mm, Ø0.28 SMD )
– 200-ball WFBGA (10mm x 14.5mm x
0.8mm, Ø0.35 SMD)
– 200-ball VFBGA (10mm x 14.5mm x
0.95mm, Ø0.35 SMD)
– Speed grade, cycle time
– 755ps @ RL = 24/28
– 625ps @ RL = 28/32
– 535ps @ RL = 32/36
• Ultra-low-voltage core and I/O power supplies
– VDD1 = 1.70–1.95V; 1.8V nominal
– VDD2/VDDQ = 1.06–1.17V; 1.1V nominal
• Frequency range
– 1866–10 MHz (data rate range: 3733–20 Mb/s/
pin)
• 16n prefetch DDR architecture
• 1-channel partitioned architecture for low RD/WR
energy and low average latency
• 8 internal banks per channel for concurrent operation
• Single-data-rate CMD/ADR entry
• Bidirectional/differential data strobe per byte lane
• Programmable READ and WRITE latencies (RL/WL)
• Programmable and on-the-fly burst lengths (BL =
16, 32)
• Directed per-bank refresh for concurrent bank operation and ease of command scheduling
• Up to 7.5 GB/s per die
• On-chip temperature sensor to control self refresh
rate
• Partial-array self refresh (PASR)
• Selectable output drive strength (DS)
• Clock-stop capability
• RoHS-compliant, “green” packaging
• Programmable V SS (ODT) termination
A
384M32
768M32
D2
D4
NP
NQ
DS1
DT1
-75
-62
-53
• Operating temperature range
– –25°C to +85°C
• Revision
Table 1: Key Timing Parameters
WRITE Latency
READ Latency
Speed
Grade
Clock Rate
(MHz)
Data Rate
(Mb/s/pin)
Set A
Set B
-75
1333
2667
12
22
24
28
-62
1600
3200
14
26
28
32
-53
1866
3733
16
30
32
36
1
(DBI Disabled) (DBI Enabled)
200b: x32 LPDDR4 SDRAM
Features
SDRAM Addressing
The table below shows the addressing for the 6Gb die density. All bank, row, and column addresses are shown perdie.
Table 2: Device Addressing
Configuration
384M32 (12Gb)
768M32 (24Gb)
2
4
Device density (per die)
6Gb
6Gb
Device density (per channel)
6Gb
6Gb
48Mb x 16 DQ x 8 banks
x 2 channels
96Mb x 16 DQ x 8 banks
x 2 channels
Number of channels (per die)
1
1
Number of ranks per channel
1
2
Number of banks (per channel)
8
8
256
256
49,152
49,152
Die per package
Configuration
Array prefetch (bits) (per channel)
Number of rows (per bank)
Number of columns (fetch boundaries)
64
64
2048
2048
Channel density (bits per channel)
6,442,450,944
6,442,450,944
Total density (bits per die)
6,442,450,944
6,442,450,944
Page size (bytes)
Bank address
x16
BA[2:0]
BA[2:0]
Row addresses
R[15:0]
R[15:0]
Column addresses
C[9:0]
C[9:0]
64-bit
64-bit
Burst starting address boundary
Notes:
1. The lower two column addresses (C0–C1) are assumed to be zero and are not transmitted on the CA bus.
2. Row and column address values on the CA bus that are not used for a particular density are "Don't Care."
3. For non-binary memory densities, only a quarter of the row address space is invalid. When the MSB address
bit is HIGH, then the MSB - 1 address bit must be LOW.
2
200b: x32 LPDDR4 SDRAM
Features
Part Number Ordering Information
Figure 1: Part Number Chart
3
200b: x32 LPDDR4 SDRAM
Features
Contents
General Description ....................................................................................................................................... 16
General Notes ............................................................................................................................................ 16
Package Block Diagrams ................................................................................................................................. 17
Ball Assignments and Descriptions ................................................................................................................. 19
Package Dimensions ....................................................................................................................................... 22
MR0, MR[6:4], MR8, MR13 Readout ................................................................................................................. 26
IDD Parameters ............................................................................................................................................... 27
Functional Description ................................................................................................................................... 31
Monolithic Device Addressing ......................................................................................................................... 32
Simplified Bus Interface State Diagram ............................................................................................................ 36
Power-Up and Initialization ............................................................................................................................ 37
Voltage Ramp ............................................................................................................................................. 38
Reset Initialization with Stable Power .......................................................................................................... 40
Power-Off Sequence ....................................................................................................................................... 41
Controlled Power-Off .................................................................................................................................. 41
Uncontrolled Power-Off .............................................................................................................................. 41
Mode Registers ............................................................................................................................................... 42
Mode Register Assignments and Definitions ................................................................................................ 42
Commands and Timing .................................................................................................................................. 67
Truth Tables ................................................................................................................................................... 67
ACTIVATE Command ..................................................................................................................................... 70
Read and Write Access Modes ......................................................................................................................... 71
Preamble and Postamble ................................................................................................................................ 72
Burst READ Operation .................................................................................................................................... 75
Read Timing ............................................................................................................................................... 77
tLZ(DQS), tLZ(DQ), tHZ(DQS), tHZ(DQ) Calculation ..................................................................................... 77
tLZ(DQS) and tHZ(DQS) Calculation for ATE (Automatic Test Equipment) .................................................... 78
tLZ(DQ) and tHZ(DQ) Calculation for ATE (Automatic Test Equipment) ........................................................ 79
Burst WRITE Operation .................................................................................................................................. 81
Write Timing .............................................................................................................................................. 84
tWPRE Calculation for ATE (Automatic Test Equipment) .............................................................................. 85
tWPST Calculation for ATE (Automatic Test Equipment) ............................................................................... 85
MASK WRITE Operation ................................................................................................................................. 86
Mask Write Timing Constraints for BL16 ...................................................................................................... 88
Data Mask and Data Bus Inversion (DBI [DC]) Function ................................................................................... 90
Preamble and Postamble Behavior .................................................................................................................. 94
Preamble, Postamble Behavior in READ-to-READ Operations ...................................................................... 94
READ-to-READ Operations – Seamless ........................................................................................................ 94
READ-to-READ Operations – Consecutive ................................................................................................... 95
WRITE-to-WRITE Operations – Seamless ................................................................................................... 102
WRITE-to-WRITE Operations – Consecutive ............................................................................................... 105
PRECHARGE Operation ................................................................................................................................. 109
Burst READ Operation Followed by Precharge ............................................................................................ 109
Burst WRITE Followed by Precharge ........................................................................................................... 110
Auto Precharge .............................................................................................................................................. 111
Burst READ With Auto Precharge ............................................................................................................... 111
Burst WRITE With Auto Precharge .............................................................................................................. 112
RAS Lock Function .................................................................................................................................... 116
Delay Time From WRITE-to-READ with Auto Precharge .............................................................................. 117
REFRESH Command ..................................................................................................................................... 118
4
200b: x32 LPDDR4 SDRAM
Features
Refresh Requirement ..................................................................................................................................... 123
SELF REFRESH Operation .............................................................................................................................. 124
Self Refresh Entry and Exit ......................................................................................................................... 124
Power-Down Entry and Exit During Self Refresh ......................................................................................... 125
Command Input Timing After Power-Down Exit ......................................................................................... 126
Self Refresh Abort ...................................................................................................................................... 127
MRR, MRW, MPC Commands During tXSR, tRFC ........................................................................................ 127
Power-Down Mode ........................................................................................................................................ 130
Power-Down Entry and Exit ....................................................................................................................... 130
Input Clock Stop and Frequency Change ........................................................................................................ 140
Clock Frequency Change – CKE LOW ......................................................................................................... 140
Clock Stop – CKE LOW ............................................................................................................................... 140
Clock Frequency Change – CKE HIGH ........................................................................................................ 140
Clock Stop – CKE HIGH ............................................................................................................................. 141
MODE REGISTER READ Operation ................................................................................................................ 142
MRR After a READ and WRITE Command .................................................................................................. 143
MRR After Power-Down Exit ...................................................................................................................... 145
MODE REGISTER WRITE ............................................................................................................................... 146
Mode Register Write States ......................................................................................................................... 147
VREF Current Generator (VRCG) ..................................................................................................................... 148
VREF Training ................................................................................................................................................. 150
VREF(CA) Training ........................................................................................................................................ 150
VREF(DQ) Training ....................................................................................................................................... 155
Command Bus Training ................................................................................................................................. 160
Command Bus Training Mode .................................................................................................................... 160
Training Sequence for Single-Rank Systems ................................................................................................ 161
Training Sequence for Multiple-Rank Systems ............................................................................................ 162
Relation Between CA Input Pin and DQ Output Pin ..................................................................................... 163
Write Leveling ............................................................................................................................................... 167
Mode Register Write-WR Leveling Mode ..................................................................................................... 167
Write Leveling Procedure ........................................................................................................................... 167
Input Clock Frequency Stop and Change .................................................................................................... 168
MULTIPURPOSE Operation ........................................................................................................................... 171
Read DQ Calibration Training ........................................................................................................................ 176
Read DQ Calibration Training Procedure .................................................................................................... 176
Read DQ Calibration Training Example ...................................................................................................... 178
MPC[READ DQ CALIBRATION] After Power-Down Exit ............................................................................... 179
Write Training ............................................................................................................................................... 179
Internal Interval Timer .............................................................................................................................. 185
DQS Interval Oscillator Matching Error ...................................................................................................... 187
OSC Count Readout Time .......................................................................................................................... 188
Thermal Offset .............................................................................................................................................. 190
Temperature Sensor ...................................................................................................................................... 190
ZQ Calibration ............................................................................................................................................... 191
ZQCAL Reset ............................................................................................................................................. 192
Multichannel Considerations ..................................................................................................................... 193
ZQ External Resistor, Tolerance, and Capacitive Loading ............................................................................. 194
Frequency Set Points ..................................................................................................................................... 195
Frequency Set Point Update Timing ........................................................................................................... 196
Pull-Up and Pull-Down Characteristics and Calibration .................................................................................. 200
On-Die Termination for the Command/Address Bus ....................................................................................... 201
ODT Mode Register and ODT State Table .................................................................................................... 201
5
200b: x32 LPDDR4 SDRAM
Features
ODT Mode Register and ODT Characteristics ............................................................................................. 202
ODT for CA Update Time ........................................................................................................................... 204
DQ On-Die Termination ................................................................................................................................ 205
Output Driver and Termination Register Temperature and Voltage Sensitivity .............................................. 207
ODT Mode Register ................................................................................................................................... 208
Asynchronous ODT ................................................................................................................................... 208
DQ ODT During Power-Down and Self Refresh Modes ................................................................................ 210
ODT During Write Leveling Mode .............................................................................................................. 210
Target Row Refresh Mode ............................................................................................................................... 211
TRR Mode Operation ................................................................................................................................. 211
Post-Package Repair ...................................................................................................................................... 213
Failed Row Address Repair ......................................................................................................................... 213
Read Preamble Training ................................................................................................................................. 215
Electrical Specifications ................................................................................................................................. 216
Absolute Maximum Ratings ....................................................................................................................... 216
AC and DC Operating Conditions ................................................................................................................... 216
AC and DC Input Measurement Levels ........................................................................................................... 218
Input Levels for CKE .................................................................................................................................. 218
Input Levels for RESET_n and ODT_CA ...................................................................................................... 218
Differential Input Voltage for CK ................................................................................................................ 218
Peak Voltage Calculation Method ............................................................................................................... 219
Single-Ended Input Voltage for Clock ......................................................................................................... 220
Differential Input Slew Rate Definition for Clock ......................................................................................... 221
Differential Input Cross-Point Voltage ........................................................................................................ 222
Differential Input Voltage for DQS .............................................................................................................. 223
Peak Voltage Calculation Method ............................................................................................................... 223
Single-Ended Input Voltage for DQS ........................................................................................................... 224
Differential Input Slew Rate Definition for DQS .......................................................................................... 225
Differential Input Cross-Point Voltage ........................................................................................................ 226
Input Levels for ODT ................................................................................................................................. 227
Output Slew Rate and Overshoot/Undershoot specifications ........................................................................... 227
Single-Ended Output Slew Rate .................................................................................................................. 227
Differential Output Slew Rate ..................................................................................................................... 228
Overshoot and Undershoot Specifications .................................................................................................. 229
Driver Output Timing Reference Load ............................................................................................................ 229
LVSTL I/O System .......................................................................................................................................... 230
Input/Output Capacitance ............................................................................................................................. 231
IDD Specification Parameters and Test Conditions ........................................................................................... 232
IDD Specifications ...................................................................................................................................... 238
AC Timing ..................................................................................................................................................... 240
CA Rx Voltage and Timing .............................................................................................................................. 249
DQ Tx Voltage and Timing ............................................................................................................................. 252
DRAM Data Timing ................................................................................................................................... 252
DQ Rx Voltage and Timing ............................................................................................................................. 253
Clock Specification ........................................................................................................................................ 256
tCK(abs), tCH(abs), and tCL(abs) ................................................................................................................ 257
Clock Period Jitter .......................................................................................................................................... 257
Clock Period Jitter Effects on Core Timing Parameters ................................................................................. 257
Cycle Time Derating for Core Timing Parameters ........................................................................................ 258
Clock Cycle Derating for Core Timing Parameters ....................................................................................... 258
Clock Jitter Effects on Command/Address Timing Parameters ..................................................................... 258
Clock Jitter Effects on READ Timing Parameters .......................................................................................... 258
6
200b: x32 LPDDR4 SDRAM
Features
Clock Jitter Effects on WRITE Timing Parameters ........................................................................................ 259
Revision History ............................................................................................................................................ 260
Rev. D – 3/17 ............................................................................................................................................. 260
Rev. C – 11/16 ............................................................................................................................................ 260
Rev. B – 8/16 .............................................................................................................................................. 260
Rev. A – 5/16 .............................................................................................................................................. 260
7
200b: x32 LPDDR4 SDRAM
Features
List of Figures
Figure 1: Part Number Chart ............................................................................................................................ 3
Figure 2: Dual-Die, Dual-Channel, Single-Rank Package Block Diagram .......................................................... 17
Figure 3: Quad-Die, Dual-Channel, Dual-Rank Package Block Diagram ........................................................... 18
Figure 4: 200-Ball Dual-Channel, Single-Rank Discrete FBGA .......................................................................... 19
Figure 5: 200-Ball Dual-Channel, Dual-Rank Discrete FBGA ............................................................................ 20
Figure 6: 200-Ball WFBGA – 10mm × 14.5mm (Package Code: NP) ................................................................... 22
Figure 7: 200-Ball VFBGA – 10mm × 14.5mm (Package Code: NQ) ................................................................... 23
Figure 8: 200-Ball WFBGA – 10mm × 14.5mm (Package Code: DS) ................................................................... 24
Figure 9: 200-Ball WFBGA – 10mm × 14.5mm (Package Code: DT) ................................................................... 25
Figure 10: Functional Block Diagram ............................................................................................................. 32
Figure 11: Simplified State Diagram ............................................................................................................... 36
Figure 12: Simplified State Diagram ............................................................................................................... 37
Figure 13: Voltage Ramp and Initialization Sequence ...................................................................................... 39
Figure 14: ACTIVATE Command .................................................................................................................... 71
Figure 15: tFAW Timing .................................................................................................................................. 71
Figure 16: DQS Read Preamble and Postamble – Toggling Preamble and 0.5nCK Postamble ............................. 72
Figure 17: DQS Read Preamble and Postamble – Static Preamble and 1.5nCK Postamble .................................. 73
Figure 18: DQS Write Preamble and Postamble – 0.5nCK Postamble ................................................................ 73
Figure 19: DQS Write Preamble and Postamble – 1.5nCK Postamble ................................................................ 74
Figure 20: Burst Read Timing ......................................................................................................................... 75
Figure 21: Burst Read Followed by Burst Write or Burst Mask Write .................................................................. 76
Figure 22: Seamless Burst Read ...................................................................................................................... 76
Figure 23: Read Timing .................................................................................................................................. 77
Figure 24: tLZ(DQS) Method for Calculating Transitions and Endpoint ............................................................ 78
Figure 25: tHZ(DQS) Method for Calculating Transitions and Endpoint ........................................................... 78
Figure 26: tLZ(DQ) Method for Calculating Transitions and Endpoint .............................................................. 79
Figure 27: tHZ(DQ) Method for Calculating Transitions and Endpoint ............................................................. 80
Figure 28: Burst WRITE Operation ................................................................................................................. 82
Figure 29: Burst Write Followed by Burst Read ................................................................................................ 83
Figure 30: Write Timing ................................................................................................................................. 84
Figure 31: Method for Calculating tWPRE Transitions and Endpoints ............................................................... 85
Figure 32: Method for Calculating tWPST Transitions and Endpoints ............................................................... 85
Figure 33: MASK WRITE Command – Same Bank ........................................................................................... 86
Figure 34: MASK WRITE Command – Different Bank ...................................................................................... 87
Figure 35: MASKED WRITE Command with Write DBI Enabled; DM Enabled .................................................. 92
Figure 36: WRITE Command with Write DBI Enabled; DM Disabled ................................................................ 93
Figure 37: READ Operations: tCCD = MIN, Preamble = Toggle, 1.5nCK Postamble ............................................ 94
Figure 38: Seamless READ: tCCD = MIN + 1, Preamble = Toggle, 1.5nCK Postamble .......................................... 95
Figure 39: Consecutive READ: tCCD = MIN + 1, Preamble = Toggle, 0.5nCK Postamble ..................................... 95
Figure 40: Consecutive READ: tCCD = MIN + 1, Preamble = Static, 1.5nCK Postamble ...................................... 96
Figure 41: Consecutive READ: tCCD = MIN + 1, Preamble = Static, 0.5nCK Postamble ...................................... 96
Figure 42: Consecutive READ: tCCD = MIN + 2, Preamble = Toggle, 1.5nCK Postamble ..................................... 97
Figure 43: Consecutive READ: tCCD = MIN + 2, Preamble = Toggle, 0.5nCK Postamble ..................................... 98
Figure 44: Consecutive READ: tCCD = MIN + 2, Preamble = Static, 1.5nCK Postamble ...................................... 98
Figure 45: Consecutive READ: tCCD = MIN + 2, Preamble = Static, 0.5nCK Postamble ...................................... 99
Figure 46: Consecutive READ: tCCD = MIN + 3, Preamble = Toggle, 1.5nCK Postamble .................................... 100
Figure 47: Consecutive READ: tCCD = MIN + 3, Preamble = Toggle, 0.5nCK Postamble .................................... 100
Figure 48: Consecutive READ: tCCD = MIN + 3, Preamble = Static, 1.5nCK Postamble ..................................... 101
Figure 49: Consecutive READ: tCCD = MIN + 3, Preamble = Static, 0.5nCK Postamble ..................................... 101
Figure 50: Seamless WRITE: tCCD = MIN, 0.5nCK Postamble ......................................................................... 102
8
200b: x32 LPDDR4 SDRAM
Features
Figure 51: Seamless WRITE: tCCD = MIN, 1.5nCK Postamble, 533 MHz < Clock Frequency 800 MHz, ODT
Worst Timing Case ..................................................................................................................................... 103
Figure 52: Seamless WRITE: tCCD = MIN, 1.5nCK Postamble ......................................................................... 104
Figure 53: Consecutive WRITE: tCCD = MIN + 1, 0.5nCK Postamble ................................................................ 105
Figure 54: Consecutive WRITE: tCCD = MIN + 1, 1.5nCK Postamble ................................................................ 105
Figure 55: Consecutive WRITE: tCCD = MIN + 2, 0.5nCK Postamble ................................................................ 106
Figure 56: Consecutive WRITE: tCCD = MIN + 2, 1.5nCK Postamble ................................................................ 106
Figure 57: Consecutive WRITE: tCCD = MIN + 3, 0.5nCK Postamble ................................................................ 107
Figure 58: Consecutive WRITE: tCCD = MIN + 3, 1.5nCK Postamble ................................................................ 108
Figure 59: Consecutive WRITE: tCCD = MIN + 4, 1.5nCK Postamble ................................................................ 108
Figure 60: Burst READ Followed by Precharge – BL16, Toggling Preamble, 0.5nCK Postamble .......................... 110
Figure 61: Burst READ Followed by Precharge – BL32, 2tCK, 0.5nCK Postamble ............................................... 110
Figure 62: Burst WRITE Followed by PRECHARGE – BL16, 2nCK Preamble, 0.5nCK Postamble ........................ 111
Figure 63: Burst READ With Auto Precharge – BL16, Non-Toggling Preamble, 0.5nCK Postamble ..................... 112
Figure 64: Burst READ With Auto Precharge – BL32, Toggling Preamble, 1.5nCK Postamble ............................. 112
Figure 65: Burst WRITE With Auto Precharge – BL16, 2 nCK Preamble, 0.5nCK Postamble ................................ 113
Figure 66: Command Input Timing with RAS Lock ......................................................................................... 117
Figure 67: Delay Time From WRITE-to-READ with Auto Precharge ................................................................. 117
Figure 68: All-Bank REFRESH Operation ....................................................................................................... 120
Figure 69: Per Bank REFRESH Operation ....................................................................................................... 121
Figure 70: Postponing REFRESH Commands (Example) ................................................................................. 123
Figure 71: Pulling In REFRESH Commands (Example) ................................................................................... 123
Figure 72: Self Refresh Entry/Exit Timing ...................................................................................................... 125
Figure 73: Self Refresh Entry/Exit Timing with Power-Down Entry/Exit .......................................................... 126
Figure 74: Command Input Timings after Power-Down Exit During Self Refresh ............................................. 127
Figure 75: MRR, MRW, and MPC Commands Issuing Timing During tXSR ....................................................... 128
Figure 76: MRR, MRW, and MPC Commands Issuing Timing During tRFC ...................................................... 129
Figure 77: Basic Power-Down Entry and Exit Timing ...................................................................................... 131
Figure 78: Read and Read with Auto Precharge to Power-Down Entry ............................................................. 132
Figure 79: Write and Mask Write to Power-Down Entry .................................................................................. 133
Figure 80: Write With Auto Precharge and Mask Write With Auto Precharge to Power-Down Entry ................... 134
Figure 81: Refresh Entry to Power-Down Entry .............................................................................................. 135
Figure 82: ACTIVATE Command to Power-Down Entry .................................................................................. 135
Figure 83: PRECHARGE Command to Power-Down Entry .............................................................................. 136
Figure 84: Mode Register Read to Power-Down Entry ..................................................................................... 137
Figure 85: Mode Register Write to Power-Down Entry .................................................................................... 138
Figure 86: MULTI PURPOSE Command for ZQCAL Start to Power-Down Entry ............................................... 139
Figure 87: MODE REGISTER READ Operation ............................................................................................... 143
Figure 88: READ-to-MRR Timing .................................................................................................................. 144
Figure 89: WRITE-to-MRR Timing ................................................................................................................. 145
Figure 90: MRR Following Power-Down ......................................................................................................... 146
Figure 91: MODE REGISTER WRITE Timing .................................................................................................. 146
Figure 92: VRCG Enable Timing .................................................................................................................... 149
Figure 93: VRCG Disable Timing ................................................................................................................... 149
Figure 94: V REF Operating Range (VREF,max, V REF,min) ....................................................................................... 150
Figure 95: V REF Set-Point Tolerance and Step Size .......................................................................................... 151
Figure 96: tVref for Short, Middle, and Long Timing Diagram .......................................................................... 152
Figure 97: V REF(CA) Single-Step Increment ...................................................................................................... 152
Figure 98: V REF(CA) Single-Step Decrement ..................................................................................................... 153
Figure 99: V REF(CA) Full Step from V REF,min to V REF,max ...................................................................................... 153
Figure 100: V REF(CA) Full Step from V REF,max to V REF,min .................................................................................... 153
Figure 101: V REF Operating Range (VREF,max, V REF,min) ..................................................................................... 155
9
200b: x32 LPDDR4 SDRAM
Features
Figure 102:
Figure 103:
Figure 104:
Figure 105:
Figure 106:
Figure 107:
Figure 108:
Figure 109:
Figure 110:
Figure 111:
Figure 112:
Figure 113:
Figure 114:
Figure 115:
Figure 116:
Figure 117:
Figure 118:
Figure 119:
Figure 120:
Figure 121:
Figure 122:
Figure 123:
Figure 124:
Figure 125:
Figure 126:
Figure 127:
Figure 128:
Figure 129:
Figure 130:
Figure 131:
Figure 132:
Figure 133:
Figure 134:
Figure 135:
Figure 136:
Figure 137:
Figure 138:
Figure 139:
Figure 140:
Figure 141:
Figure 142:
Figure 143:
Figure 144:
Figure 145:
Figure 146:
Figure 147:
Figure 148:
Figure 149:
Figure 150:
Figure 151:
Figure 152:
Figure 153:
V REF Set Tolerance and Step Size .................................................................................................. 156
V REF(DQ) Transition Time for Short, Middle, or Long Changes ........................................................ 157
V REF(DQ) Single-Step Size Increment ............................................................................................. 157
V REF(DQ) Single-Step Size Decrement ............................................................................................ 158
V REF(DQ) Full Step from V REF,min to V REF,max ................................................................................... 158
V REF(DQ) Full Step from V REF,max to V REF,min ................................................................................... 158
Command Bus Training Mode Entry – CA Training Pattern I/O with V REF(CA) Value Update ............ 163
Consecutive V REF(CA) Value Update .............................................................................................. 164
Command Bus Training Mode Exit with Valid Command .............................................................. 165
Command Bus Training Mode Exit with Power-Down Entry .......................................................... 166
Write Leveling Timing – tDQSL (MAX) .......................................................................................... 168
Write Leveling Timing – tDQSL (MIN) .......................................................................................... 168
Clock Stop and Timing During Write Leveling .............................................................................. 169
DQS_t/DQS_c to CK_t/CK_c Timings at the Pins Referenced from the Internal Latch .................... 170
WRITE-FIFO – tWPRE = 2nCK, tWPST = 0.5nCK ............................................................................ 172
READ-FIFO – tWPRE = 2nCK, tWPST = 0.5nCK, tRPRE = Toggling, tRPST = 1.5nCK ......................... 173
READ-FIFO – tRPRE = Toggling, tRPST = 1.5nCK ........................................................................... 174
Read DQ Calibration Training Timing: Read-to-Read DQ Calibration ............................................ 177
Read DQ Calibration Training Timing: Read DQ Calibration to Read DQ Calibration/Read ............ 177
MPC[READ DQ CALIBRATION] Following Power-Down State ....................................................... 179
WRITE-to-MPC[WRITE-FIFO] Operation Timing ......................................................................... 181
MPC[WRITE-FIFO]-to-MPC[READ-FIFO] Timing ........................................................................ 182
MPC[READ-FIFO] to Read Timing ............................................................................................... 183
MPC[WRITE-FIFO] with DQ ODT Timing .................................................................................... 184
Power-Down Exit to MPC[WRITE-FIFO] Timing ........................................................................... 185
Interval Oscillator Offset – OSCoffset ............................................................................................. 187
In Case of DQS Interval Oscillator is Stopped by MPC Command .................................................. 188
In Case of DQS Interval Oscillator is Stopped by DQS Interval Timer ............................................. 189
Temperature Sensor Timing ........................................................................................................ 191
ZQCAL Timing ............................................................................................................................ 193
Frequency Set Point Switching Timing ......................................................................................... 197
Training for Two Frequency Set Points ......................................................................................... 199
Example of Switching Between Two Trained Frequency Set Points ................................................ 199
Example of Switching to a Third Trained Frequency Set Point ....................................................... 200
ODT for CA ................................................................................................................................. 201
ODT for CA Setting Update Timing in 4-Clock Cycle Command .................................................... 204
Functional Representation of DQ ODT ........................................................................................ 205
Asynchronous ODTon/ODToff Timing ......................................................................................... 209
Target Row Refresh Mode ............................................................................................................ 212
Post-Package Repair Timing ........................................................................................................ 214
Read Preamble Training .............................................................................................................. 215
Input Timing Definition for CKE .................................................................................................. 218
Input Timing Definition for RESET_n and ODT_CA ...................................................................... 218
CK Differential Input Voltage ....................................................................................................... 219
Definition of Differential Clock Peak Voltage ................................................................................ 220
Clock Single-Ended Input Voltage ................................................................................................ 220
Differential Input Slew Rate Definition for CK_t, CK_c .................................................................. 221
V ix Definition (Clock) .................................................................................................................. 222
DQS Differential Input Voltage .................................................................................................... 223
Definition of Differential DQS Peak Voltage .................................................................................. 224
DQS Single-Ended Input Voltage ................................................................................................. 224
Differential Input Slew Rate Definition for DQS_t, DQS_c ............................................................. 225
10
200b: x32 LPDDR4 SDRAM
Features
Figure 154:
Figure 155:
Figure 156:
Figure 157:
Figure 158:
Figure 159:
Figure 160:
Figure 161:
Figure 162:
Figure 163:
Figure 164:
Figure 165:
Figure 166:
Figure 167:
Figure 168:
Figure 169:
Figure 170:
Figure 171:
Figure 172:
Figure 173:
V ix Definition (DQS) .................................................................................................................... 226
Single-Ended Output Slew Rate Definition ................................................................................... 228
Differential Output Slew Rate Definition ...................................................................................... 228
Overshoot and Undershoot Definition ......................................................................................... 229
Driver Output Timing Reference Load ......................................................................................... 230
LVSTL I/O Cell ............................................................................................................................ 230
Pull-Up Calibration ..................................................................................................................... 231
tCMDCKE Timing ....................................................................................................................... 244
tESCKE Timing ........................................................................................................................... 246
CA Receiver (Rx) Mask ................................................................................................................ 249
Across Pin V REF (CA) Voltage Variation ........................................................................................... 249
CA Timings at the DRAM Pins ..................................................................................................... 250
CA tcIPW and SRIN_cIVW Definition (for Each Input Pulse) .......................................................... 250
CA V IHL_AC Definition (for Each Input Pulse) ................................................................................ 250
Read Data Timing Definitions – tQH and tDQSQ Across DQ Signals per DQS Group ....................... 252
DQ Receiver (Rx) Mask ................................................................................................................ 253
Across Pin V REF DQ Voltage Variation ........................................................................................... 253
DQ-to-DQS tDQS2DQ and tDQDQ .............................................................................................. 254
DQ tDIPW and SRIN_dIVW Definition for Each Input Pulse .......................................................... 255
DQ V IHL(AC) Definition (for Each Input Pulse) ............................................................................... 255
11
200b: x32 LPDDR4 SDRAM
Features
List of Tables
Table 1: Key Timing Parameters ....................................................................................................................... 1
Table 2: Device Addressing .............................................................................................................................. 2
Table 3: Ball/Pad Descriptions ....................................................................................................................... 21
Table 4: Mode Register Contents .................................................................................................................... 26
Table 5: IDD Parameters – Single-Die .............................................................................................................. 27
Table 6: IDD6 Full-Array Self Refresh Current ................................................................................................... 28
Table 7: IDD Parameters – Single-Die .............................................................................................................. 28
Table 8: IDD6 Full-Array Self Refresh Current ................................................................................................... 30
Table 9: Monolithic Device Addressing – 2 Channels per Die ........................................................................... 33
Table 10: Monolithic Device Addressing – 1 Channel per Die ........................................................................... 34
Table 11: Mode Register Default Settings ........................................................................................................ 38
Table 12: Voltage Ramp Conditions ................................................................................................................ 38
Table 13: Initialization Timing Parameters ...................................................................................................... 40
Table 14: Reset Timing Parameter .................................................................................................................. 41
Table 15: Power Supply Conditions ................................................................................................................ 41
Table 16: Power-Off Timing ............................................................................................................................ 42
Table 17: Mode Register Assignments ............................................................................................................. 42
Table 18: MR0 Device Feature 0 (MA[7:0] = 00h) .............................................................................................. 43
Table 19: MR0 Op-Code Bit Definitions .......................................................................................................... 43
Table 20: MR1 Device Feature 1 (MA[7:0] = 01h) .............................................................................................. 44
Table 21: MR1 Op-Code Bit Definitions .......................................................................................................... 44
Table 22: Burst Sequence for Read .................................................................................................................. 46
Table 23: Burst Sequence for Write ................................................................................................................. 46
Table 24: MR2 Device Feature 2 (MA[7:0] = 02h) .............................................................................................. 47
Table 25: MR2 Op-Code Bit Definitions .......................................................................................................... 47
Table 26: Frequency Ranges for RL, WL, nWR, and nRTP Settings .................................................................... 49
Table 27: MR3 I/O Configuration 1 (MA[7:0] = 03h) ......................................................................................... 49
Table 28: MR3 Op-Code Bit Definitions .......................................................................................................... 50
Table 29: MR4 Device Temperature (MA[7:0] = 04h) ........................................................................................ 51
Table 30: MR4 Op-Code Bit Definitions .......................................................................................................... 51
Table 31: MR5 Basic Configuration 1 (MA[7:0] = 05h) ...................................................................................... 52
Table 32: MR5 Op-Code Bit Definitions .......................................................................................................... 52
Table 33: MR6 Basic Configuration 2 (MA[7:0] = 06h) ...................................................................................... 52
Table 34: MR6 Op-Code Bit Definitions .......................................................................................................... 52
Table 35: MR7 Basic Configuration 3 (MA[7:0] = 07h) ...................................................................................... 52
Table 36: MR7 Op-Code Bit Definitions .......................................................................................................... 52
Table 37: MR8 Basic Configuration 4 (MA[7:0] = 08h) ...................................................................................... 53
Table 38: MR8 Op-Code Bit Definitions .......................................................................................................... 53
Table 39: MR9 Test Mode (MA[7:0] = 09h) ....................................................................................................... 53
Table 40: MR9 Op-Code Definitions ............................................................................................................... 53
Table 41: MR10 Calibration (MA[7:0] = 0Ah) ................................................................................................... 53
Table 42: MR10 Op-Code Bit Definitions ........................................................................................................ 54
Table 43: MR11 ODT Control (MA[7:0] = 0Bh) ................................................................................................. 54
Table 44: MR11 Op-Code Bit Definitions ........................................................................................................ 54
Table 45: MR12 Register Information (MA[7:0] = 0Ch) ..................................................................................... 55
Table 46: MR12 Op-Code Bit Definitions ........................................................................................................ 55
Table 47: MR13 Register Control (MA[7:0] = 0Dh) ............................................................................................ 55
Table 48: MR13 Op-Code Bit Definition .......................................................................................................... 56
Table 49: Mode Register14 (MA[7:0] = 0Eh) ..................................................................................................... 57
Table 50: MR14 Op-Code Bit Definition .......................................................................................................... 57
12
200b: x32 LPDDR4 SDRAM
Features
Table 51: V REF Setting for Range[0] and Range[1] ............................................................................................. 58
Table 52: MR15 Register Information (MA[7:0] = 0Fh) ..................................................................................... 59
Table 53: MR15 Op-code Bit Definition .......................................................................................................... 59
Table 54: MR15 Invert Register Pin Mapping ................................................................................................... 59
Table 55: MR16 PASR Bank Mask (MA[7:0] = 010h) .......................................................................................... 59
Table 56: MR16 Op-Code Bit Definitions ........................................................................................................ 59
Table 57: MR17 PASR Segment Mask (MA[7:0] = 11h) ...................................................................................... 60
Table 58: MR17 PASR Segment Mask Definitions ............................................................................................ 60
Table 59: MR17 PASR Segment Mask .............................................................................................................. 60
Table 60: MR18 Register Information (MA[7:0]=12h) ....................................................................................... 61
Table 61: MR18 LSB DQS Oscillator Count ...................................................................................................... 61
Table 62: MR19 Register Information (MA[7:0] = 13h) ..................................................................................... 61
Table 63: MR19 DQS Oscillator Count ............................................................................................................ 61
Table 64: MR20 Register Information (MA[7:0] = 14h) ..................................................................................... 62
Table 65: MR20 Register Information ............................................................................................................. 62
Table 66: MR20 Invert Register Pin Mapping ................................................................................................... 62
Table 67: MR21 Register Information (MA[7:0] = 15h) ..................................................................................... 62
Table 68: MR22 Register Information (MA[7:0] = 16h) ..................................................................................... 62
Table 69: MR22 Register Information ............................................................................................................. 63
Table 70: MR23 Register Information (MA[7:0] = 17h) ..................................................................................... 64
Table 71: MR23 Register Information ............................................................................................................. 64
Table 72: MR24 Register Information (MA[7:0] = 18h) ..................................................................................... 64
Table 73: MR24 Register Information ............................................................................................................. 65
Table 74: MR25 Register Information (MA[7:0] = 19h) ..................................................................................... 65
Table 75: MR25 Register Information ............................................................................................................. 65
Table 76: MR26:31 Register Information (MA[7:0] = 1Ah–1Fh) ......................................................................... 66
Table 77: MR32 Register Information (MA[7:0] = 20h) ..................................................................................... 66
Table 78: MR32 Register Information ............................................................................................................. 66
Table 79: MR33:39 Register Information (MA[7:0] = 21h–27h) .......................................................................... 66
Table 80: MR40 Register Information (MA[7:0] = 28h) ..................................................................................... 66
Table 81: MR40 Register Information ............................................................................................................. 67
Table 82: MR41:47 Register Information (MA[7:0] = 29h–2Fh) .......................................................................... 67
Table 83: MR48:63 Register Information (MA[7:0] = 30h–3Fh) .......................................................................... 67
Table 84: Command Truth Table .................................................................................................................... 68
Table 85: Reference Voltage for tLZ(DQS), tHZ(DQS) Timing Measurements ..................................................... 79
Table 86: Reference Voltage for tLZ(DQ), tHZ(DQ) Timing Measurements ........................................................ 80
Table 87: Method for Calculating tWPRE Transitions and Endpoints ................................................................ 85
Table 88: Reference Voltage for tWPST Timing Measurements ......................................................................... 86
Table 89: Same Bank (ODT Disabled) ............................................................................................................. 88
Table 90: Different Bank (ODT Disabled) ........................................................................................................ 88
Table 91: Same Bank (ODT Enabled) .............................................................................................................. 89
Table 92: Different Bank (ODT Enabled) ......................................................................................................... 89
Table 93: Function Behavior of DMI Signal During WRITE, MASKED WRITE, and READ Operations ................. 90
Table 94: Precharge Bank Selection ............................................................................................................... 109
Table 95: Timing Between Commands (PRECHARGE and AUTO PRECHARGE): DQ ODT is Disable ................. 113
Table 96: Timing Between Commands (PRECHARGE and AUTO PRECHARGE): DQ ODT is Enable ................. 116
Table 97: Bank and Refresh Counter Increment Behavior ............................................................................... 118
Table 98: REFRESH Command Timing Constraints ........................................................................................ 120
Table 99: Legacy REFRESH Command Timing Constraints ............................................................................. 122
Table 100: Modified REFRESH Command Timing Constraints ........................................................................ 122
Table 101: Refresh Requirement Parameters .................................................................................................. 123
Table 102: MRR ............................................................................................................................................ 142
13
200b: x32 LPDDR4 SDRAM
Features
Table 103:
Table 104:
Table 105:
Table 106:
Table 107:
Table 108:
Table 109:
Table 110:
Table 111:
Table 112:
Table 113:
Table 114:
Table 115:
Table 116:
Table 117:
Table 118:
Table 119:
Table 120:
Table 121:
Table 122:
Table 123:
Table 124:
Table 125:
Table 126:
Table 127:
Table 128:
Table 129:
Table 130:
Table 131:
Table 132:
Table 133:
Table 134:
Table 135:
Table 136:
Table 137:
Table 138:
Table 139:
Table 140:
Table 141:
Table 142:
Table 143:
Table 144:
Table 145:
Table 146:
Table 147:
Table 148:
Table 149:
Table 150:
Table 151:
Table 152:
Table 153:
Table 154:
Truth Table for MRR and MRW ..................................................................................................... 147
MRR/MRW Timing Constraints: DQ ODT is Disable ...................................................................... 147
MRR/MRW Timing Constraints: DQ ODT is Enable ....................................................................... 148
VRCG Enable/Disable Timing ....................................................................................................... 149
Internal V REF(CA) Specifications ..................................................................................................... 154
Internal V REF(DQ) Specifications .................................................................................................... 159
Mapping MR12 Op Code and DQ Numbers ................................................................................... 161
Mapping of CA Input Pin and DQ Output Pin ................................................................................ 163
Write Leveling Timing Parameters ................................................................................................. 169
Write Leveling Setup and Hold Timing .......................................................................................... 169
MPC Command Definition ........................................................................................................... 171
MPC Commands .......................................................................................................................... 172
Timing Constraints for Training Commands .................................................................................. 174
Invert Mask Assignments .............................................................................................................. 176
Read DQ Calibration Bit Ordering and Inversion Example .............................................................. 178
MR Setting vs. DMI Status ............................................................................................................. 179
MPC[WRITE-FIFO] AC Timing ...................................................................................................... 185
DQS Oscillator Matching Error Specification ................................................................................. 187
DQS Interval Oscillator AC Timing ................................................................................................ 189
Temperature Sensor ..................................................................................................................... 191
ZQ Calibration Parameters ........................................................................................................... 192
Mode Register Function with Two Physical Registers ...................................................................... 195
Relation Between MR Setting and DRAM Operation ...................................................................... 196
Frequency Set Point AC Timing ..................................................................................................... 197
tFC Value Mapping ....................................................................................................................... 197
tFC Value Mapping: Example ........................................................................................................ 198
Pull-Down Driver Characteristics – ZQ Calibration ........................................................................ 200
Pull-Up Characteristics – ZQ Calibration ....................................................................................... 200
Valid Calibration Points ................................................................................................................ 200
Command Bus ODT State ............................................................................................................. 202
ODT DC Electrical Characteristics – up to 3200 Mbps .................................................................... 202
ODT DC Electrical Characteristics – Beyond 3200 Mbps ................................................................. 203
ODT DC Electrical Characteristics – up to 3200 Mbps .................................................................... 205
ODT DC Electrical Characteristics – Beyond 3200 Mbps ................................................................. 206
Output Driver and Termination Register Sensitivity Definition ....................................................... 207
Output Driver and Termination Register Temperature and Voltage Sensitivity ................................. 208
ODTLON and ODTLOFF Latency Values .......................................................................................... 209
Termination State in Write Leveling Mode ..................................................................................... 210
Post-Package Repair Timing Parameters ........................................................................................ 214
Absolute Maximum DC Ratings .................................................................................................... 216
Recommended DC Operating Conditions ..................................................................................... 216
Input Leakage Current .................................................................................................................. 217
Input/Output Leakage Current ..................................................................................................... 217
Operating Temperature Range ...................................................................................................... 217
Input Levels ................................................................................................................................. 218
Input Levels ................................................................................................................................. 218
CK Differential Input Voltage ........................................................................................................ 219
Clock Single-Ended Input Voltage ................................................................................................. 221
Differential Input Slew Rate Definition for CK_t, CK_c ................................................................... 221
Differential Input Level for CK_t, CK_c .......................................................................................... 222
Differential Input Slew Rate for CK_t, CK_c .................................................................................... 222
Cross-Point Voltage for Differential Input Signals (Clock) ............................................................... 223
14
200b: x32 LPDDR4 SDRAM
Features
Table 155:
Table 156:
Table 157:
Table 158:
Table 159:
Table 160:
Table 161:
Table 162:
Table 163:
Table 164:
Table 165:
Table 166:
Table 167:
Table 168:
Table 169:
Table 170:
Table 171:
Table 172:
Table 173:
Table 174:
Table 175:
Table 176:
Table 177:
Table 178:
Table 179:
Table 180:
Table 181:
Table 182:
Table 183:
Table 184:
Table 185:
Table 186:
Table 187:
Table 188:
Table 189:
Table 190:
DQS Differential Input Voltage ...................................................................................................... 223
DQS Single-Ended Input Voltage ................................................................................................... 225
Differential Input Slew Rate Definition for DQS_t, DQS_c .............................................................. 225
Differential Input Level for DQS_t, DQS_c ..................................................................................... 226
Differential Input Slew Rate for DQS_t, DQS_c ............................................................................... 226
Cross-Point Voltage for Differential Input Signals (DQS) ................................................................ 227
Input Levels ................................................................................................................................. 227
Single-Ended Output Slew Rate .................................................................................................... 227
Differential Output Slew Rate ....................................................................................................... 228
AC Overshoot/Undershoot Specifications ..................................................................................... 229
Overshoot/Undershoot Specification for CKE and RESET .............................................................. 229
Input/Output Capacitance ........................................................................................................... 231
IDD Measurement Conditions ....................................................................................................... 232
CA Pattern for IDD4R ...................................................................................................................... 232
CA Pattern for IDD4W ..................................................................................................................... 233
Data Pattern for IDD4W (DBI Off) ................................................................................................... 233
Data Pattern for IDD4R (DBI Off) .................................................................................................... 234
Data Pattern for IDD4W (DBI On) .................................................................................................... 236
Data Pattern for IDD4R (DBI On) .................................................................................................... 237
IDD Specification Parameters and Operating Conditions ................................................................ 238
Clock Timing ............................................................................................................................... 240
Read Output Timing ..................................................................................................................... 240
Write Voltage and Timing ............................................................................................................. 242
CKE Input Timing ........................................................................................................................ 243
Command Address Input Timing .................................................................................................. 244
Boot Timing Parameters (10–55 MHz) ........................................................................................... 244
Mode Register Timing Parameters ................................................................................................. 245
Core Timing Parameters ............................................................................................................... 245
CA Bus ODT Timing ..................................................................................................................... 247
CA Bus Training Parameters .......................................................................................................... 247
Asynchronous ODT Turn On and Turn Off Timing ......................................................................... 248
Temperature Derating Parameters ................................................................................................ 248
DRAM CMD/ADR, CS ................................................................................................................... 251
DQs In Receive Mode ................................................................................................................... 255
Definitions and Calculations ........................................................................................................ 256
tCK(abs), tCH(abs), and tCL(abs) Definitions ................................................................................. 257
15
200b: x32 LPDDR4 SDRAM
General Description
General Description
The 6Gb Mobile Low-Power DDR4 SDRAM (LPDDR4) is a high-speed CMOS, dynamic
random-access memory. The device is internally configured with x16 I/O, 8-banks.
Each of the x16’s 805,306,368-bit banks is organized as 49,152 rows by 1024 columns by
16 bits.
General Notes
Throughout the data sheet, figures and text refer to DQs as “DQ.” DQ should be interpreted as any or all DQ collectively, unless specifically stated otherwise.
“DQS” and “CK” should be interpreted as DQS_t, DQS_c and CK_t, CK_c respectively,
unless specifically stated otherwise. “CA” includes all CA pins used for a given density.
In timing diagrams, “CMD” is used as an indicator only. Actual signals occur on CA[5:0].
VREF indicates V REF(CA) and V REF(DQ).
Complete functionality may be described throughout the entire document. Any page or
diagram may have been simplified to convey a topic and may not be inclusive of all requirements.
Any specific requirement takes precedence over a general statement.
Any functionality not specifically stated herein is considered undefined, illegal, is not
supported, and will result in unknown operation.
16
200b: x32 LPDDR4 SDRAM
Package Block Diagrams
Package Block Diagrams
Figure 2: Dual-Die, Dual-Channel, Single-Rank Package Block Diagram
VDD1 VDD2 VSS VDDQ
VDDQ
RZQ
Die
ZQ0
RESET_n
CS0_A
CKE0_A
LPDDR4
Channel A
CK_t_A
CK_c_A
CA[5:0]_A
ODT_CA
DMI[1:0]_A
DQ[15:0]_A
DQS[1:0]_t_A
DQS[1:0]_c_A
ODT_CA_A
Die
CS0_B
CKE0_B
CK_t_B
CK_c_B
CA[5:0]_B
LPDDR4
Channel B
ODT_CA
17
DMI[1:0]_B
DQ[15:0]_B
DQS[1:0]_t_B
DQS[1:0]_c_B
ODT_CA_B
200b: x32 LPDDR4 SDRAM
Package Block Diagrams
Figure 3: Quad-Die, Dual-Channel, Dual-Rank Package Block Diagram
VDD1 VDD2 VSS VDDQ
VDDQ
RZQ
Die
ZQ0
RESET_n
CS0_A
CKE0_A
DMI[1:0]_A
DQ[15:0]_A
DQS[1:0]_t_A
DQS[1:0]_c_A
LPDDR4
CK_t_A
CK_c_A
CA[5:0]_A
ODT_CA
ODT_CA_A
Die
CS0_B
DMI[1:0]_B
DQ[15:0]_B
DQS[1:0]_t_B
DQS[1:0]_c_B
LPDDR4
CKE0_B
CK_t_B
CK_c_B
CA[5:0]_B
ODT_CA
ODT_CA_B
VDDQ
Die
RZQ
ZQ1
CS1_A
CKE1_A
LPDDR4
ODT_CA
VSS
Die
CS1_B
CKE1_B
LPDDR4
ODT_CA
VSS
Note:
1. ODT(ca) for Rank 0 of each channel is wired to the respective ODT ball. ODT(ca) for
Rank 1 of each channel is wired to VSS in the package.
18
200b: x32 LPDDR4 SDRAM
Ball Assignments and Descriptions
Ball Assignments and Descriptions
Figure 4: 200-Ball Dual-Channel, Single-Rank Discrete FBGA
1
2
3
4
5
A
DNU
DNU
VSS
VDD2
B
DNU
DQ0_A
VDDQ
C
VSS
DQ1_A
D
VDDQ
VSS
E
VSS
F
VDD1
DQ3_A
VDDQ
G
VSS
ODT_CA_A
H
VDD2
J
K
8
9
10
11
12
ZQ0
NC
VDD2
VSS
DNU
DNU
DQ7_A
VDDQ
VDDQ
DQ15_A
VDDQ
DQ8_A
DNU
DMI0_A
DQ6_A
VSS
VSS
DQ14_A
DMI1_A
DQ9_A
VSS
DQS0_t_A
VSS
VDDQ
VDDQ
VSS
DQS1_t_A
VSS
VDDQ
VSS
VSS
DQ4_A
VDD2
VDD2
VSS
VDD1
VSS
CA0_A
NC
CS0_A
VSS
CA1_A
VSS
VDD2
VSS
N
VDD2
P
DQ2_A DQS0_c_A DQ5_A
6
7
DQ13_A DQS1_c_A DQ10_A
VSS
DQ12_A
VDDQ
DQ11_A
VDD1
VSS
VDD1
VSS
NC
VSS
VDD2
VDD2
CA2_A
CA3_A
CA4_A
VDD2
CKE0_A
NC
CK_t_A
CK_c_A
VSS
CA5_A
VSS
VDD2
VSS
NC
NC
VSS
VDD2
VSS
VDD2
VSS
VDD2
VSS
NC
NC
VSS
VDD2
VSS
VDD2
VSS
CA1_B
VSS
CKE0_B
NC
CK_t_B
CK_c_B
VSS
CA5_B
VSS
R
VDD2
CA0_B
NC
CS0_B
VDD2
VDD2
CA2_B
CA3_B
CA4_B
VDD2
T
VSS
ODT_CA_B
VSS
VDD1
VSS
VSS
VDD1
VSS
RESET_n
VSS
U
VDD1
DQ3_B
VDDQ
DQ4_B
VDD2
VDD2
DQ12_B
VDDQ
DQ11_B
VDD1
V
VSS
VSS
VSS
W
VDDQ
VSS
DQS0_t_B
VSS
VDDQ
VDDQ
VSS
DQS1_t_B
VSS
VDDQ
Y
VSS
DQ1_B
DMI0_B
DQ6_B
VSS
VSS
DQ14_B
DMI1_B
DQ9_B
VSS
AA
DNU
DQ0_B
VDDQ
DQ7_B
VDDQ
VDDQ
DQ15_B
VDDQ
DQ8_B
DNU
AB
DNU
DNU
VSS
VDD2
VSS
VSS
VDD2
VSS
DNU
DNU
1
2
3
4
5
8
9
10
11
12
L
M
DQ2_B DQS0_c_B DQ5_B
6
7
DQ13_B DQS1_c_B DQ10_B
VSS
Top View (ball down)
DDR4_A (Channel A)
DDR4_B (Channel B)
19
ZQ, ODT_CA, RESET
Supply
Ground
200b: x32 LPDDR4 SDRAM
Ball Assignments and Descriptions
Figure 5: 200-Ball Dual-Channel, Dual-Rank Discrete FBGA
1
2
3
4
5
A
DNU
DNU
VSS
VDD2
B
DNU
DQ0_A
VDDQ
C
VSS
DQ1_A
D
VDDQ
VSS
E
VSS
F
VDD1
DQ3_A
VDDQ
G
VSS
ODT_CA_A
H
VDD2
J
K
8
9
10
11
12
ZQ0
ZQ1
VDD2
VSS
DNU
DNU
DQ7_A
VDDQ
VDDQ
DQ15_A
VDDQ
DQ8_A
DNU
DMI0_A
DQ6_A
VSS
VSS
DQ14_A
DMI1_A
DQ9_A
VSS
DQS0_t_A
VSS
VDDQ
VDDQ
VSS
DQS1_t_A
VSS
VDDQ
VSS
VSS
DQ4_A
VDD2
VDD2
VSS
VDD1
VSS
CA0_A
CS1_A
CS0_A
VSS
CA1_A
VSS
VDD2
VSS
N
VDD2
P
DQ2_A DQS0_c_A DQ5_A
6
7
DQ13_A DQS1_c_A DQ10_A
VSS
DQ12_A
VDDQ
DQ11_A
VDD1
VSS
VDD1
VSS
NC
VSS
VDD2
VDD2
CA2_A
CA3_A
CA4_A
VDD2
CKE0_A
CKE1_A
CK_t_A
CK_c_A
VSS
CA5_A
VSS
VDD2
VSS
NC
NC
VSS
VDD2
VSS
VDD2
VSS
VDD2
VSS
NC
NC
VSS
VDD2
VSS
VDD2
VSS
CA1_B
VSS
CKE0_B
CKE1_B
CK_t_B
CK_c_B
VSS
CA5_B
VSS
R
VDD2
CA0_B
CS1_B
CS0_B
VDD2
VDD2
CA2_B
CA3_B
CA4_B
VDD2
T
VSS
ODT_CA_B
VSS
VDD1
VSS
VSS
VDD1
VSS
RESET_n
VSS
U
VDD1
DQ3_B
VDDQ
DQ4_B
VDD2
VDD2
DQ12_B
VDDQ
DQ11_B
VDD1
V
VSS
VSS
VSS
W
VDDQ
VSS
DQS0_t_B
VSS
VDDQ
VDDQ
VSS
DQS1_t_B
VSS
VDDQ
Y
VSS
DQ1_B
DMI0_B
DQ6_B
VSS
VSS
DQ14_B
DMI1_B
DQ9_B
VSS
AA
DNU
DQ0_B
VDDQ
DQ7_B
VDDQ
VDDQ
DQ15_B
VDDQ
DQ8_B
DNU
AB
DNU
DNU
VSS
VDD2
VSS
VSS
VDD2
VSS
DNU
DNU
1
2
3
4
5
8
9
10
11
12
L
M
DQ2_B DQS0_c_B DQ5_B
6
7
DQ13_B DQS1_c_B DQ10_B
VSS
Top View (ball down)
DDR4_A (Channel A)
DDR4_B (Channel B)
20
ZQ, ODT_CA, RESET
Supply
Ground
200b: x32 LPDDR4 SDRAM
Ball Assignments and Descriptions
Table 3: Ball/Pad Descriptions
Symbol
Type
Description
CK_t_A, CK_c_A,
CK_t_B, CK_c_B
Input
Clock: CK_t and CK_c are differential clock inputs. All address, command and control
input signals are sampled on positive edge of CK_t and the negative edge of CK_c. AC
timings for CA parameters are referenced to clock. Each channel (A, B) has its own
clock pair.
CKE0_A, CKE1_A,
CKE0_B, CKE1_B
Input
Clock enable: CKE HIGH activates and CKE LOW deactivates the internal clock signals,
input buffers, and output drivers. Power-saving modes are entered and exited via CKE
transitions. CKE is sampled at the rising edge of CK.
CS0_A, CS1_A, CS0_B,
CS1_B
Input
Chip select: Each channel (A, B) has its own CS signals.
CA[5:0]_A, CA[5:0]_B
Input
Command/address inputs: Provide the command and address inputs according to
the command truth table. Each channel (A, B) has its own CA signals.
ODT_CA_A,
ODT_CA_B
Input
CA ODT Control: The ODT_CA pin is used in conjunction with the mode register to
turn on/off the on-die termination for CA pins. It is bonded to VDD2 within the package, or at the package ball, for the terminating rank, and the non-terminating ranks
are bonded to VSS (or left floating with a weak pull-down on the DRAM die). The terminating rank is the DRAM that terminates the CA bus for all die on the same channel.
DQ[15:0]_A,
DQ[15:0]_B
I/O
Data input/output: Bidirectional data bus.
DQS[1:0]_t_A,
DQS[1:0]_c_A,
DQS[1:0]_t_B,
DQS[1:0]_c_B
I/O
Data strobe: DQS_t and DQS_c are bi-directional differential output clock signals
used to strobe data during a READ or WRITE. The data strobe is generated by the
DRAM for a READ and is edge-aligned with data. The data strobe is generated by the
SoC memory controller for a WRITE and is trained to precede data. Each byte of data
has a data strobe signal pair. Each channel (A, B) has its own DQS_t and DQS_c strobes.
DMI[1:0]_A,
DMI[1:0]_B
I/O
Data Mask/Data Bus Inversion: DMI is a dual use bi-directional signal used to indicate data to be masked, and data which is inverted on the bus. For data bus inversion
(DBI), the DMI signal is driven HIGH when the data on the data bus is inverted, or driven LOW when the data is in its normal state. DBI can be disabled via a mode register
setting. For data mask, the DMI signal is used in combination with the data lines to
indicate data to be masked in a MASK WRITE command (see the Data Mask (DM) and
Data Bus Inversion (DBI) sections for details). The data mask function can be disabled
via a mode register setting. Each byte of data has a DMI signal. Each channel has its
own DMI signals.
ZQ0, ZQ1
Reference ZQ Calibration Reference: Used to calibrate the output drive strength and the termination resistance. There is one ZQ pin per die. The ZQ pin shall be connected to
VDDQ through a 240Ω ±1% resistor.
VDDQ, VDD1, VDD2
Supply
Power supplies: Isolated on the die for improved noise immunity.
VSS
Supply
Ground Reference: Power supply ground reference.
RESET_n
Input
DNU
–
Do not use: Must be grounded or left floating.
NC
–
No connect: Not internally connected.
RESET: When asserted LOW, the RESET pin resets all channels of the die.
21
200b: x32 LPDDR4 SDRAM
Package Dimensions
Package Dimensions
Figure 6: 200-Ball WFBGA – 10mm × 14.5mm (Package Code: NP)
Seating plane
A
200X Ø0.312
Dimensions apply
to solder balls postreflow on Ø0.28 SMD
ball pads.
0.08 A
Ball A1 ID
(covered by SR)
12 11 10 9 8
5 4 3 2
Ball A1 ID
1
A
B
C
D
E
F
G
H
J
K
L
M
N
P
R
T
U
V
W
Y
AA
AB
14.5 ±0.1
13.65 CTR
0.65 TYP
0.7 ±0.1
0.8 TYP
8.8 CTR
10 ±0.1
Notes:
0.225 ±0.05
1. All dimensions are in millimeters.
2. Solder ball composition: SAC302 with NiAu pads (96.8% Sn, 3Ag, 0.2% Cu).
3. The package height does not include room temperature warpage.
22
200b: x32 LPDDR4 SDRAM
Package Dimensions
Figure 7: 200-Ball VFBGA – 10mm × 14.5mm (Package Code: NQ)
Seating plane
A
200X Ø0.312
Dimensions apply
to solder balls postreflow on Ø0.28 SMD
ball pads.
0.08 A
Ball A1 ID
(covered by SR)
12 11 10 9 8
5 4 3 2
Ball A1 ID
1
A
B
C
D
E
F
G
H
J
K
L
M
N
P
R
T
U
V
W
Y
AA
AB
14.5 ±0.1
13.65 CTR
0.65 TYP
0.85 ±0.1
0.8 TYP
8.8 CTR
10 ±0.1
Notes:
0.225 ±0.05
1. All dimensions are in millimeters.
2. Solder ball composition: SAC302 with NiAu pads (96.8% Sn, 3Ag, 0.2% Cu).
3. The package height does not include room temperature warpage.
23
200b: x32 LPDDR4 SDRAM
Package Dimensions
Figure 8: 200-Ball WFBGA – 10mm × 14.5mm (Package Code: DS)
Seating plane
A
200X Ø0.363
Dimensions
apply to solder
balls post-reflow
on Ø0.35 SMD
ball pads.
0.08 A
Ball A1 ID
12 11 10 9 8
5 4 3 2
Ball A1 ID
1
A
B
C
D
E
F
G
H
J
K
L
M
N
P
R
T
U
V
W
Y
AA
AB
14.5 ±0.1
13.65 CTR
0.65 TYP
0.7 ±0.1
0.8 TYP
0.227 ±0.05
8.8 CTR
10 ±0.1
Notes:
1. All dimensions are in millimeters.
2. Solder ball composition: SAC302 with NiAu pads (96.8% Sn, 3Ag, 0.2% Cu).
3. The package height does not include room temperature warpage.
24
200b: x32 LPDDR4 SDRAM
Package Dimensions
Figure 9: 200-Ball WFBGA – 10mm × 14.5mm (Package Code: DT)
Seating plane
A
200X Ø0.363
Dimensions
apply to solder
balls post-reflow
on Ø0.35 SMD
ball pads.
0.08 A
Ball A1 ID
12 11 10 9 8
5 4 3 2
Ball A1 ID
1
A
B
C
D
E
F
G
H
J
K
L
M
N
P
R
T
U
V
W
Y
AA
AB
14.5 ±0.1
13.65 CTR
0.65 TYP
0.85 ±0.1
0.8 TYP
0.227 ±0.05
8.8 CTR
10 ±0.1
Notes:
1. All dimensions are in millimeters.
2. Solder ball composition: SAC302 with NiAu pads (96.8% Sn, 3Ag, 0.2% Cu).
3. The package height does not include room temperature warpage.
25
200b: x32 LPDDR4 SDRAM
MR0, MR[6:4], MR8, MR13 Readout
MR0, MR[6:4], MR8, MR13 Readout
Table 4: Mode Register Contents
Mode Register
OP7
OP6
OP5
OP4
OP3
OP2
OP1
OP0
REF
MR0
0b: Both legacy and modified refresh mode supported
SR Abort
MR4
0b: Disable (default)
1b: Enable
Manufacturer ID
MR5
1111 1111b : Micron
Revision ID1
MR6
0000 0001b
Density
MR8
OP[5:2] = 0011b: 6Gb per die (6Gb per channel)
VRO
MR13
0b: Normal operation (default)
1b: Output the VREF(CA) value on DQ7 and VREF(DQ) value on DQ6
Notes:
1. The contents of MR0, MR[6:4], MR8, and MR13 will reflect information specific to each
in these packages.
2. Other bits not defined above and other mode registers are referred to in Mode Register
Assignments and Definitions section.
26
200b: x32 LPDDR4 SDRAM
IDD Parameters
IDD Parameters
Refer to IDD Specification Parameters and Test Conditions section for detailed conditions.
Table 5: IDD Parameters – Single-Die
VDD2, VDDQ = 1.06–1.17V; VDD1 = 1.70–1.95V; TC = -25°C to +85°C
Speed Grade
Parameter
IDD01
Supply
3200 Mbps
3733 Mbps
Unit
VDD1
3.5
3.5
mA
IDD02
VDD2
45
45
IDD0Q
VDDQ
0.75
0.75
IDD2P1
VDD1
1.2
1.2
IDD2P2
VDD2
3
3
IDD2PQ
VDDQ
0.75
0.75
IDD2PS1
VDD1
1.2
1.2
IDD2PS2
VDD2
3
3
IDD2PSQ
VDDQ
0.75
0.75
IDD2N1
VDD1
2
2
IDD2N2
VDD2
27
27
IDD2NQ
VDDQ
0.75
0.75
IDD2NS1
VDD1
2
2
IDD2NS2
VDD2
23
23
IDD2NSQ
VDDQ
0.75
0.75
IDD3P1
VDD1
1.2
1.2
IDD3P2
VDD2
9
9
IDD3PQ
VDDQ
0.75
0.75
IDD3PS1
VDD1
1.2
1.2
IDD3PS2
VDD2
9
9
IDD3PSQ
VDDQ
0.75
0.75
IDD3N1
VDD1
2.25
2.25
IDD3N2
VDD2
30
30
IDD3NQ
VDDQ
0.75
0.75
IDD3NS1
VDD1
2.25
2.25
IDD3NS2
VDD2
26
26
IDD3NSQ
VDDQ
0.75
0.75
IDD4R1
VDD1
2.25
2.25
IDD4R2
VDD2
275
315
IDD4RQ
VDDQ
150
160
IDD4W1
VDD1
2.25
2.25
IDD4W2
VDD2
210
240
IDD4WQ
VDDQ
55
65
27
Note
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
3
200b: x32 LPDDR4 SDRAM
IDD Parameters
Table 5: IDD Parameters – Single-Die (Continued)
VDD2, VDDQ = 1.06–1.17V; VDD1 = 1.70–1.95V; TC = -25°C to +85°C
Speed Grade
Parameter
IDD51
Supply
3200 Mbps
3733 Mbps
Unit
VDD1
10
10
mA
IDD52
VDD2
90
90
IDD5Q
VDDQ
0.75
0.75
IDD5AB1
VDD1
2.5
2.5
IDD5AB2
VDD2
30
30
IDD5ABQ
VDDQ
0.75
0.75
IDD5PB1
VDD1
2.5
2.5
IDD5PB2
VDD2
30
30
IDD5PBQ
VDDQ
0.75
0.75
Notes:
Note
mA
mA
1. Published IDD values except IDD4RQ are the maximum of the distribution of the arithmetic
mean. Refer to the following note for IDD4RQ; refer to IDD6 Full-Array Self Refresh Current table for IDD6.
2. IDD4RQ value is reference only. Typical value. DBI disabled, VOH = VDDQ/3, TC = 25°C.
Table 6: IDD6 Full-Array Self Refresh Current
VDD2, VDDQ = 1.06–1.17V; VDD1 = 1.70–1.95V
Temperature
Supply
25°C
85°C
Note:
Full-Array Self Refresh Current
Unit
VDD1
0.3
mA
VDD2
0.5
VDDQ
0.1
VDD1
2.0
VDD2
5.0
VDDQ
0.75
1. IDD6 25°C is the typical, and IDD6 85°C is the maximum of the distribution of the arithmetic mean.
Table 7: IDD Parameters – Single-Die
VDD2, VDDQ = 1.06–1.17V; VDD1 = 1.70–1.95V; TC = 0°C to +105°C
Speed Grade
Parameter
Supply
3200 Mbps
3733 Mbps
Unit
IDD01
VDD1
3.5
3.5
mA
IDD02
VDD2
45
45
IDD0Q
VDDQ
0.75
0.75
IDD2P1
VDD1
1.2
1.2
IDD2P2
VDD2
3
3
IDD2PQ
VDDQ
0.75
0.75
28
mA
Note
200b: x32 LPDDR4 SDRAM
IDD Parameters
Table 7: IDD Parameters – Single-Die (Continued)
VDD2, VDDQ = 1.06–1.17V; VDD1 = 1.70–1.95V; TC = -25°C to +85°C
Speed Grade
Parameter
IDD2PS1
Supply
3200 Mbps
3733 Mbps
Unit
VDD1
1.2
1.2
mA
IDD2PS2
VDD2
3
3
IDD2PSQ
VDDQ
0.75
0.75
IDD2N1
VDD1
2
2
IDD2N2
VDD2
30
30
IDD2NQ
VDDQ
0.75
0.75
IDD2NS1
VDD1
2
2
IDD2NS2
VDD2
23
23
IDD2NSQ
VDDQ
0.75
0.75
IDD3P1
VDD1
1.2
1.2
IDD3P2
VDD2
9
9
IDD3PQ
VDDQ
0.75
0.75
IDD3PS1
VDD1
1.2
1.2
IDD3PS2
VDD2
9
9
IDD3PSQ
VDDQ
0.75
0.75
IDD3N1
VDD1
2.25
2.25
IDD3N2
VDD2
36
36
IDD3NQ
VDDQ
0.75
0.75
IDD3NS1
VDD1
2.25
2.25
IDD3NS2
VDD2
30
30
IDD3NSQ
VDDQ
0.75
0.75
IDD4R1
VDD1
2.25
2.25
IDD4R2
VDD2
310
350
IDD4RQ
VDDQ
165
170
IDD4W1
VDD1
2.25
2.25
IDD4W2
VDD2
235
265
IDD4WQ
VDDQ
60
70
IDD51
VDD1
10
10
IDD52
VDD2
90
90
IDD5Q
VDDQ
0.75
0.75
IDD5AB1
VDD1
2.5
2.5
IDD5AB2
VDD2
30
30
IDD5ABQ
VDDQ
0.75
0.75
29
Note
mA
mA
mA
mA
mA
mA
mA
mA
mA
mA
3
200b: x32 LPDDR4 SDRAM
IDD Parameters
Table 7: IDD Parameters – Single-Die (Continued)
VDD2, VDDQ = 1.06–1.17V; VDD1 = 1.70–1.95V; TC = -25°C to +105°C
Speed Grade
Parameter
IDD5PB1
Supply
3200 Mbps
3733 Mbps
Unit
VDD1
2.5
2.5
mA
IDD5PB2
VDD2
30
30
IDD5PBQ
VDDQ
0.75
0.75
Notes:
Note
1. Published IDD values except IDD4RQ are the maximum of the distribution of the arithmetic
mean. Refer to another note for IDD4RQ. And refer to another table for IDD6.
2. IDD4RQ value is reference only. Typical value. DBI disabled, VOH = VDDQ/3, TC = 25°C
Table 8: IDD6 Full-Array Self Refresh Current
VDD2, VDDQ = 1.06–1.17V; VDD1 = 1.70–1.95V
Temperature
Supply
105°C
Note:
Full-Array Self Refresh Current
Unit
VDD1
2
mA
VDD2
8
VDDQ
1
1. IDD6 105°C is the maximum of the distribution of the arithmetic mean.
30
200b: x32 LPDDR4 SDRAM
Functional Description
Functional Description
The Mobile Low-Power DDR4 SDRAM (LPDDR4) is a high-speed CMOS, dynamic random-access memory internally configured with either 1 or 2 channels. Each channel is
comprised of 16 DQs and 8 banks.
LPDDR4 uses a 2-tick, single-data-rate (SDR) protocol on the CA bus to reduce the
number of input signals in the system. The term "2-tick" means that the command/
address is decoded across two transactions, such that half of the command/address is
captured with each of two consecutive rising edges of CK. The 6-bit CA bus contains
command, address, and bank information. Some commands such as READ, WRITE,
MASKED WRITE, and ACTIVATE require two consecutive 2-tick SDR commands to
complete the instruction.
LPDDR4 uses a double-data-rate (DDR) protocol on the DQ bus to achieve high-speed
operation. The DDR interface transfers two data bits to each DQ lane in one clock cycle
and is matched to a 16n-prefetch DRAM architecture. A write/read access consists of a
single 16n-bit-wide data transfer to/from the DRAM core and 16 corresponding n-bitwide data transfers at the I/O pins.
Read and write accesses to the device are burst-oriented. Accesses start at a selected
column address and continue for a programmed number of columns in a programmed
sequence.
Accesses begin with the registration of an ACTIVATE command to open a row in the
memory core, followed by a WRITE or READ command to access column data within
the open row. The address and bank address (BA) bits registered by the ACTIVATE command are used to select the bank and row to be opened. The address and BA bits registered with the WRITE or READ command are used to select the bank and the starting
column address for the burst access.
Prior to normal operation, the LPDDR4 SDRAM must be initialized. Following sections
provide detailed information about device initialization, register definition, command
descriptions and device operations.
31
200b: x32 LPDDR4 SDRAM
Monolithic Device Addressing
Figure 10: Functional Block Diagram
VDDQ
RZQ
ZQ
To CLK, CS, CA ODT calibration
To DQS, DQ, DMI ODT calibration
ZQ Cal
RESET
CKE
CK_t, CK_c
CS_n
CA[5:0]
ODT_CA
DQ ODT control
RCVRS
Command/Address
Multiplex and
Decode
Control
logic
Mode
registers
x
Refresh
counter
CA ODT control
Rowaddress
MUX
Bank 7
Bank 7
Bank 6
Bank 6
Bank 5
Bank 5
Bank 4
Bank 4
Bank 3
Bank 3
Bank 2
Bank 2
Bank 1
Bank 1
Bank 0
Bank 0
rowMemory array
address
latch
and
decoder
COL[3:0]
16n
Read 16n
MUX
latch
n
DATA
DRVRS
Read data path
DQS
generator DQS_t,
DQS_c
(1...n)
VSS
Sense amplifier
RTT,nom
(1...n)
16n
VSS
RTT,nom
I/O gating
DM mask logic
0–7
SW
0–7
Bank
control
logic
Columnaddress
counter/
latch
y-4
Column
decoder
n
WRITE
FIFO
Mask
16n and
drivers
CK out
CK_t,
16n
CK in
CK_c
Data
SW
n/16
Input
registers
n
RCVRS
Write data path
DQ[n-1:0]
DQS_t, DQS_c
DMI
COL[3:0]
4
Monolithic Device Addressing
The table below includes all monolithic device addressing options defined by JEDEC.
Under the SDRAM Addressing heading near the beginning of this data sheet are addressing details for this product data sheet.
32
Table 9: Monolithic Device Addressing – 2 Channels per Die
Memory density
(per die)
6Gb
8Gb
12Gb
16Gb
2Gb
3Gb
4Gb
6Gb
8Gb
Configuration
16Mb x 16DQ x
8 banks x 2 channels
24Mb x 16DQ x
8 banks x 2 channels
32Mb x 16DQ x
8 banks x 2 channels
48Mb x 16DQ x
8 banks x 2 channels
64Mb x 16DQ x
8 banks x 2 channels
Number of
channels (per die)
2
2
2
2
2
Number of banks
(per channel)
8
8
8
8
8
Array prefetch
(bits, per channel)
256
256
256
256
256
Number of rows
(per channel)
16,384
24,576
32,768
49,152
65,536
Number of columns
(fetch boundaries)
64
64
64
64
64
Page size (bytes)
2048
2048
2048
2048
2048
Channel density
(bits per channel)
2,147,483,648
3,221,225,472
4,294,967,296
6,442,450,944
8,589,934,592
Total density
(bits per die)
4,294,967,296
6,442,450,944
8,589,934,592
12,884,901,888
17,179,869,184
Bank address
BA[2:0]
BA[2:0]
BA[2:0]
BA[2:0]
BA[2:0]
x16
Row
add
R[13:0]
R[14:0]
(R13 = 0 when R14 = 1)
R[14:0]
R[15:0]
(R14 = 0 when R15 = 1)
R[15:0]
Col.
add
C[9:0]
C[9:0]
C[9:0]
C[9:0]
C[9:0]
64 bit
64 bit
64 bit
64 bit
64 bit
Memory density
(per channel)
33
Burst starting
address boundary
200b: x32 LPDDR4 SDRAM
Monolithic Device Addressing
4Gb
Table 10: Monolithic Device Addressing – 1 Channel per Die
Memory density (per die)
4Gb
6Gb
8Gb
Memory density (per channel)
4Gb
6Gb
8Gb
Configuration
32Mb x 16 DQ x 8 banks
48Mb x 16 DQ x 8 banks
64Mb x 16 DQ x 8 banks
Number of channels (per die)
1
1
1
Number of banks (per channel)
8
8
8
Array prefetch (bits, per channel)
Number of rows (per channel)
Number of columns (fetch boundaries)
256
256
256
32,768
49,152
65,536
64
64
64
2048
2048
2048
Channel density (bits per channel)
4,294,967,296
6,442,450,944
8,589,934,592
Total density (bits per die)
4,294,967,296
6,442,450,944
8,589,934,592
BA[2:0]
BA[2:0]
BA[2:0]
R[14:0]
R[15:0]
(R14 = 0 when R15 = 1)
R[15:0]
Page size (bytes)
Bank address
x16
Row add
Column add
34
Burst starting address boundary
C[9:0]
C[9:0]
C[9:0]
64 bit
64 bit
64 bit
200b: x32 LPDDR4 SDRAM
Monolithic Device Addressing
200b: x32 LPDDR4 SDRAM
Monolithic Device Addressing
Notes:
1. The lower two column addresses (C[1:0]) are assumed to be zero and are not transmitted on the CA bus.
2. Row and column address values on the CA bus that are not used for a particular density
should be at valid logic levels.
3. For non-binary memory densities, only a quarter of the row address space is invalid.
When the MSB address bit is HIGH, then the MSB - 1 address bit must be LOW.
35
200b: x32 LPDDR4 SDRAM
Simplified Bus Interface State Diagram
Simplified Bus Interface State Diagram
The state diagram provides a simplified illustration of the bus interface, supported state
transitions, and the commands that control them. For a complete description of device
behavior, use the information provided in the state diagram with the truth tables and
timing specifications. The truth tables describe device behavior and applicable restrictions when considering the actual state of all banks. For command descriptions, see the
Commands and Timing section.
Figure 11: Simplified State Diagram
Automatic sequence
Command sequence
MPCbased
training
Power-on
MR
write
MPCbased
training
L
E
M
=
CK
E
MPC
H
n
T_
SE
RE = H
=
RR
M
M
REF
MPC
All bank
refresh
Idle
SRX
RW
REF
SRE
Self
refresh
MPC
MR
write
Per bank
refresh
RW
CK
MPCbased
training
MRW
MRW
RW
MR
write
Reset
Command
bus
training
MR
read
MPC
M
n
T_
SE
RE = L
SR
powerdown
MR
R
M
W
MR
RR
MR
W
=L
CKE
Command
bus
training
MR
read
R
MRW
MR
CKE
=H
MR read
MPCbased
training
ACT
Idle
powerdown
MR write
MR read
Activating
Active
powerdown
MR
write
M
=
CK
E
RW
CK
L
E
R
MR
=
H
MPCbased
training
R
or
Per bank
refresh
RD
MW
MPC
MPCbased
training
M
RD
RR
or
Read
RA
W
A
WRA or
MWRA
RDA
PRE or PREA
A
RE
PR
E
Write or mask
write with
auto
precharge
or
E
PR
PR
EA
Precharging
Notes:
MR
read
RD
Write or
mask write
M
W
RA
WR
W
MR
REF
Bank
active
WR or MWR
MR
write
MR read
P
or
Read
with auto
precharge
PRE(A) = PRECHARGE (ALL)
ACT = ACTIVATE
WR(A) = WRITE (with auto precharge)
MWR(A) = Mask WRITE
(with auto precharge)
RD(A) = READ (with auto precharge)
MRW = MODE REGISTER WRITE
MRR = MODE REGISTER READ
"CKE = L" = Enter power-down
"CKE = H" = Exit power-down
SRE = Enter self refresh
SRX = Exit self refresh
REF = REFRESH
MPC = Mult-purpose command (with NOP)
1. From the self refresh state, the device can enter power-down, MRR, MRW, or any of the
training modes initiated with the MPC command. See the Self Refresh section.
36
200b: x32 LPDDR4 SDRAM
Power-Up and Initialization
2. All banks are precharged in the idle state.
3. In the case of using an MRW command to enter a training mode, the state machine will
not automatically return to the idle state at the conclusion of training. See the applicable training section for more information.
4. In the case of an MPC command to enter a training mode, the state machine may not
automatically return to the idle state at the conclusion of training. See the applicable
training section for more information.
5. This diagram is intended to provide an overview of the possible state transitions and
commands to control them; however, it does not contain the details necessary to operate the device. In particular, situations involving more than one bank are not captured
in complete detail.
6. States that have an "automatic return" and can be accessed from more than one prior
state (that is, MRW from either idle or active states) will return to the state where they
were initiated (that is, MRW from idle will return to idle).
7. The RESET pin can be asserted from any state and will cause the device to enter the reset state. The diagram shows RESET applied from the power-on and idle states as an example, but this should not be construed as a restriction on RESET.
8. MRW commands from the active state cannot change operating parameters of the device that affect timing. Mode register fields which may be changed via MRW from the
active state include: MR1-OP[3:0], MR1-OP[7], MR3-OP[7:6], MR10-OP[7:0], MR11OP[7:0], MR13-OP[5], MR15-OP[7:0], MR16-OP[7:0], MR17-OP[7:0], MR20-OP[7:0], and
MR22-OP[4:0].
Figure 12: Simplified State Diagram
a) FIFO-Based Write/Read Timing
MPC
Automatic sequence
Command sequence
MPC
MPCbased
training
MPC
b) Read DQ Calibration
MPC
Write
-FIFO
MPC
Read
-FIFO
MPC
MPC
DQ
Calibration
=
WRW
MPC
MRW
WRW
c) ZQCAL Start
MPC
d) ZQCAL Latch
ZQ
Calibration
Start
MPC
ZQ
Calibration
Latch
Power-Up and Initialization
To ensure proper functionality for power-up and reset initialization, default values for
the MR settings are provided in the table below.
37
200b: x32 LPDDR4 SDRAM
Power-Up and Initialization
Table 11: Mode Register Default Settings
Item
Mode Register Setting
Default Setting
FSP-OP/WR
MR13 OP[7:6]
00b
FSP-OP/WR[0] are enabled
Description
WLS
MR2 OP[6]
0b
WRITE latency set A is selected
WL
MR2 OP[5:3]
000b
WL = 4
RL
MR2 OP[2:0]
000b
RL = 6, nRTP = 8
nWR
MR1 OP[6:4]
000b
nWR = 6
DBI-WR/RD
MR3 OP[7:6]
00b
Write and read DBI are disabled
CA ODT
MR11 OP[6:4]
000b
CA ODT is disabled
DQ ODT
MR11 OP[2:0]
000b
VREF(CA) setting
MR12 OP[6]
1b
VREF(CA) value
MR12 OP[5:0]
001101b
VREF(DQ) setting
MR14 OP[6]
1b
VREF(DQ) value
MR14 OP[5:0]
001101b
DQ ODT is disabled
VREF(CA) range[1] is enabled
Range1: 27.2% of VDD2
VREF(DQ) range[1] enabled
Range1: 27.2% of VDDQ
The following sequence must be used to power up the device. Unless specified otherwise, this procedure is mandatory. The power-up sequence of all channels must proceed simultaneously.
Voltage Ramp
1. While applying power (after Ta), RESET_n should be held LOW (0.2 × V DD2), and all
other inputs must be between V IL,min and V IH,max. The device outputs remain at High-Z
while RESET_n is held LOW. Power supply voltage ramp requirements are provided in
the table below. V DD1 must ramp at the same time or earlier than V DD2. V DD2 must ramp
at the same time or earlier than V DDQ.
Table 12: Voltage Ramp Conditions
After...
Applicable Conditions
Ta is reached
VDD1 must be greater than VDD2
VDD2 must be greater than VDDQ - 200mV
Notes:
1. Ta is the point when any power supply first reaches 300mV.
2. Voltage ramp conditions in above table apply between Ta and power-off (controlled or
uncontrolled).
3. Tb is the point at which all supply and reference voltages are within their defined operating ranges.
4. Power ramp duration tINIT0 (Tb–Ta) must not exceed 20ms.
5. The voltage difference between any VSS and VSSQ must not exceed 100mV.
2. Following completion of the of the voltage ramp (Tb), RESET_n must be held LOW for
tINIT1. DQ, DMI, DQS_t, and DQS_c voltage levels must be between V
SSQ and V DDQ
during voltage ramp to avoid latch-up. CK_t and CK_c, CS, and CA input levels must be
between V SS and V DD2 during voltage ramp to avoid latch-up. Voltage ramp power supply requirements are provided in the table below.
38
200b: x32 LPDDR4 SDRAM
Power-Up and Initialization
3. Beginning at Tb, RESET_n must remain LOW for at least tINIT1(Tc), after which RESET_n can be de-asserted to HIGH(Tc). At least 10ns before CKE de-assertion, CKE is
required to be set LOW. All other input signals are "Don't Care."
Figure 13: Voltage Ramp and Initialization Sequence
Ta
Tb
Power Ramp
Tc
Reset
Td
Te
Tf
Tg
Th
Initialization
Ti
Tj
Tk
Training
tINIT4=5tCK(MIN)
CK_c
CK_t
tINIT0=20ms(MAX)
tINIT1=200μs(MIN)
Supplies
RESET_n
tINIT2=10ns(MIN)
tINIT3=2ms(MIN)
CKE
tINIT5=2μs(MIN)
CA[5:0]
CS
Exit PD
DES
MRW
MRR
tZQCAL=1μs(MIN)
DES
DQs
ZQ Cal
Start
DES
ZQ Cal
Latch
tZQLAT=MAX(30ns, 8 t CK)(MIN)
DES
CA BUS
Training
Valid
DES
Write
Leveling
Valid
DES
DQ
Training
DES
Valid
Valid
Don’t Care
Note:
1. Training is optional and may be done at the system designer's discretion. The order of
training may be different than what is shown here.
4. After RESET_n is de-asserted(Tc), wait at least tINIT3 before activating CKE. CK_t,
CK_c must be started and stabilized for tINIT4 before CKE goes active(Td). CS must remain LOW when the controller activates CKE.
5. After CKE is set to HIGH, wait a minimum of tINIT5 to issue any MRR or MRW commands(Te). For MRR and MRW commands, the clock frequency must be within the
range defined for tCKb. Some AC parameters (for example, tDQSCK) could have relaxed
timings (such as tDQSCKb) before the system is appropriately configured.
6. After completing all MRW commands to set the pull-up, pull-down, and Rx termination values, the controller can issue the ZQCAL START command to the memory(Tf).
This command is used to calibrate the V OH level and the output impedance over process, voltage, and temperature. In systems where more than one device share one external ZQ resistor, the controller must not overlap the ZQ calibration sequence of each device. The ZQ calibration sequence is completed after tZQCAL (Tg). The ZQCAL LATCH
command must be issued to update the DQ drivers and DQ + CA ODT to the calibrated
values.
7. After tZQLAT is satisfied (Th), the command bus (internal V REF(CA), CS, and CA)
should be trained for high-speed operation by issuing an MRW command (command
bus training mode). This command is used to calibrate the device's internal V REF and
align CS/CA with CK for high-speed operation. The device will power-up with receivers
configured for low-speed operations and with V REF(CA) set to a default factory setting.
Normal device operation at clock speeds higher than tCKb may not be possible until
command bus training is complete. The command bus training MRW command uses
the CA bus as inputs for the calibration data stream, and it outputs the results asynchro-
39
200b: x32 LPDDR4 SDRAM
Power-Up and Initialization
nously on the DQ bus. See command bus training in the MRW section for information
on how to enter/exit the training mode.
8. After command bus training, the controller must perform write leveling. Write leveling mode is enabled when MR2 OP[7] is HIGH(Ti). See the Write Leveling section for a
detailed description of the write leveling entry and exit sequence. In write leveling
mode, the controller adjusts write DQS timing to the point where the device recognizes
the start of write DQ data burst with desired WRITE latency.
9. After write leveling, the DQ bus (internal V REF(DQ), DQS, and DQ) should be trained
for high-speed operation using the MPC TRAINING commands and by issuing MRW
commands to adjust V REF(DQ). The device will power-up with receivers configured for
low-speed operations and with V REF(DQ) set to a default factory setting. Normal device
operation at clock speeds higher than tCKb should not be attempted until DQ bus training is complete. The MPC[READ DQ CALIBRATION] command is used together with
MPC[READ-FIFO] or MPC[WRITE-FIFO] commands to train the DQ bus without disturbing the memory array contents. See the DQ Bus Training section for more information on the DQ bus training sequence.
10. At Tk, the device is ready for normal operation and is ready to accept any valid command. Any mode registers that have not previously been configured for normal operation should be written at this time.
Table 13: Initialization Timing Parameters
Parameter
Min
Max
Unit
tINIT0
–
20
ms
Maximum voltage ramp time
tINIT1
200
–
μs
Minimum RESET_n LOW time after completion of voltage
ramp
tINIT2
10
–
ns
Minimum CKE LOW time before RESET_n goes HIGH
tINIT3
2
–
ms
Minimum CKE LOW time after RESET_n goes HIGH
tINIT4
5
–
tCK
Minimum stable clock before first CKE HIGH
tINIT5
2
tCKb
Note
–
1, 2
Notes:
Note
1, 2
Comment
μs
Minimum idle time before first MRW/MRR command
ns
Clock cycle time during boot
1. Minimum tCKb guaranteed by DRAM test is 18ns.
2. The system may boot at a higher frequency than dictated by minimum tCKb. The higher
boot frequency is system dependent.
Reset Initialization with Stable Power
The following sequence is required for RESET at no power interruption initialization.
1. Assert RESET_n below 0.2 × V DD2 anytime when reset is needed. RESET_n needs
to be maintained for minimum tPW_RESET. CKE must be pulled LOW at least
10ns before de-asserting RESET_n.
2. Repeat steps 4–10 in Voltage Ramp section.
40
200b: x32 LPDDR4 SDRAM
Power-Off Sequence
Table 14: Reset Timing Parameter
Value
Parameter
Min
Max
Unit
tPW_RESET
100
–
ns
Comment
Minimum RESET_n LOW time for reset initialization
with stable power
Power-Off Sequence
Controlled Power-Off
While powering off, CKE must be held LOW (0.2 × V DD2); all other inputs must be between V IL,min and V IH,max. The device outputs remain at High-Z while CKE is held LOW.
DQ, DMI, DQS_t, and DQS_c voltage levels must be between V SSQ and V DDQ during the
power-off sequence to avoid latch-up. CK_t, CK_c, CS, and CA input levels must be between V SS and V DD2 during the power-off sequence to avoid latch-up.
Tx is the point where any power supply drops below the minimum value specified in
the minimum DC Operating Condition.
Tz is the point where all power supplies are below 300mV. After Tz, the device is powered off.
Table 15: Power Supply Conditions
The voltage difference between VSS and VSSQ must not exceed 100mV
Between...
Applicable Conditions
VDD1 must be greater than VDD2
Tx and Tz
VDD2 must be greater than VDDQ - 200mV
Uncontrolled Power-Off
When an uncontrolled power-off occurs, the following conditions must be met.
• At Tx, when the power supply drops below the minimum values specified in the Recommended DC Operating Conditions table, all power supplies must be turned off and
all power supply current capacity must be at zero, except for any static charge remaining in the system.
• After Tz (the point at which all power supplies first reach 300mV), the device must
power off. During this period, the relative voltage between power supplies is uncontrolled. V DD1 and V DD2 must decrease with a slope lower than 0.5 V/μs between Tx
and Tz.
An uncontrolled power-off sequence can occur a maximum of 400 times over the life of
the device.
41
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 16: Power-Off Timing
Parameter
Power-off ramp time
Symbol
Min
Max
Unit
tPOFF
–
2
sec
Mode Registers
Mode Register Assignments and Definitions
Mode register definitions are provided in the Mode Register Assignments table. In the
access column of the table, R indicates read-only; W indicates write-only; R/W indicates
read- or write-capable or enabled. The MRR command is used to read from a register.
The MRW command is used to write to a register.
Table 17: Mode Register Assignments
Notes 1–5 apply to entire table
MR# MA[7:0]
Function
Access
OP7
OP6
OP5
RFU
RFU
OP4
OP3
OP2
RZQI
OP1
OP0
RFU
REF
Link
Go to MR0
0
00h
Device info
R
CATR
1
01h
Device feature 1
W
RDPST
2
02h
Device feature 2
W
WR
Lev
WLS
WL
3
03h
I/O config-1
W
DBIWR
DBI-RD
PDDS
4
04h
Refresh and
training
R /W
TUF
Thermal offset
PPRE
5
05h
Basic config-1
R
Manufacturer ID
Go to MR5
6
06h
Basic config-2
R
Revision ID1
Go to MR6
7
07h
Basic config-3
R
Revision ID2
Go to MR7
8
08h
Basic config-4
R
9
09h
Test mode
W
10
0Ah
I/O calibration
W
11
0Bh
ODT
W
nWR (for AP)
I/O width
RDPRE
WRPRE
BL
Go to MR1
RL
PPRP
SR
abort
Go to MR2
WRPST
PUCAL
Refresh rate
Density
Type
Vendor-specific test mode
Go to MR4
Go to MR8
Go to MR9
RFU
RFU
Go to MR3
ZQ RST Go to MR10
CA ODT
RFU
DQ ODT
Go to MR11
R/W
RFU
VRCA
W
FSP-OP
FSPWR
VREF(DQ)
R/W
RFU
VRDQ
0Fh
DQI-LB
W
16
10h
PASR_Bank
W
PASR bank mask
Go to MR16
17
11h
PASR_Seg
W
PASR segment mask
Go to MR17
18
12h
IT-LSB
R
DQS oscillator count – LSB
Go to MR18
19
13h
IT-MSB
R
DQS oscillator count – MSB
Go to MR19
20
14h
DQI-UB
W
Upper-byte invert register for DQ calibration
Go to MR20
21
15h
Vendor use
W
RFU
Go to MR21
12
0Ch
VREF(CA)
13
0Dh
Register control
14
0Eh
15
VREF(CA)
DMD
RRO
VRCG
VRO
Go to MR12
RPT
VREF(DQ)
Lower-byte invert register for DQ calibration
42
CBT
Go to MR13
Go to MR14
Go to MR15
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 17: Mode Register Assignments (Continued)
Notes 1–5 apply to entire table
MR# MA[7:0]
Function
Access
22
16h
ODT feature 2
W
23
17h
DQS oscillator
stop
W
24
18h
TRR control
25
19h
PPR resources
26–31 1Ah~1F
h
32
20h
–
DQ calibration
pattern A
33–39 21h≈27h Do not use
OP7
OP6
RFU
OP5
OP4
OP3
OP2
ODTD- ODTE- ODTECA
CS
CK
R/W
TRR
mode
R
B7
TRR mode BAn
B6
B5
B4
Unltd
MAC
B3
–
Reserved for future use
W
See DQ calibration section
Do not use
See DQ calibration section
41–47 29h≈2Fh Do not use
–
Do not use
48–63 30h≈3Fh Reserved
–
Reserved for future use
DQ calibration
pattern B
Notes:
Go to MR23
MAC value
B2
Link
Go to MR22
DQS oscillator run-time setting
–
28h
OP0
SoC ODT
W
40
OP1
B1
Go to MR24
B0
Go to MR25
Go to MR32
Go to MR40
1. RFU bits must be set to 0 during MRW commands.
2. RFU bits are read as 0 during MRR commands.
3. All mode registers that are specified as RFU or write-only shall return undefined data
when read via an MRR command.
4. RFU mode registers must not be written.
5. Writes to read-only registers will not affect the functionality of the device.
Table 18: MR0 Device Feature 0 (MA[7:0] = 00h)
OP7
OP6
CATR
OP5
OP4
RFU
OP3
OP2
RZQI
OP1
RFU
OP0
REF
Table 19: MR0 Op-Code Bit Definitions
Register Information
Tag
Type
OP
Refresh mode
REF
Read
only
OP[0]
Built-in self-test for RZQ
information
RZQI
Read
only
OP[4:3]
Definition
Notes
0b: Both legacy and modified refresh mode supported
1b: Only modified refresh mode supported
00b: RZQ self-test not supported
01b: ZQ may connect to VSSQ or float
10b: ZQ may short to VDDQ
11b: ZQ pin self-test completed, no error condition
detected (ZQ may not connect to VSSQ, float, or
short to VDDQ)
43
1–4
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 19: MR0 Op-Code Bit Definitions (Continued)
Register Information
CA terminating rank
Notes:
Tag
Type
OP
CATR
Read
only
OP[7]
Definition
Notes
0b: CA for this rank is not terminated
1b: CA for this rank is terminated
1. RZQI, if supported, will be set upon completion of MPC[ZQCAL START] command.
(tZQCAL after MPC[ZQCAL START] command.) RZQI value will be lost after reset.
2. If ZQ is connected to VSSQ to set default calibration, OP[4:3] must be set to 01b. If ZQ is
not connected to VSSQ, either OP[4:3] = 01b or OP[4:3] = 10b might indicate a ZQ pin assembly error. It is recommended that the assembly error be corrected.
3. In the case of possible assembly error, the device will default to factory trim settings for
RON, and will ignore ZQ CALIBRATION commands. In either case, the device may not
function as intended.
4. If the ZQ pin self-test returns OP[4:3] = 11b, the device has detected a resistor connected
to the ZQ pin. However, this result cannot be used to validate the ZQ resistor value or
that the ZQ resistor meets the specified limits (that is, 240Ω
Table 20: MR1 Device Feature 1 (MA[7:0] = 01h)
OP7
OP6
RD-PST
OP5
OP4
nWR (for AP)
OP3
OP2
RD-PRE
WR-PRE
OP1
OP0
BL
Table 21: MR1 Op-Code Bit Definitions
Feature
Type
OP
BL
Burst length
Write
only
OP[1:0]
Definition
Notes
00b: BL = 16 sequential (default)
1, 5, 6
01b: BL= 32 sequential
10b: BL = 16 or 32 sequential (on-the-fly)
11b: Reserved
WR-PRE
Write preamble length
Write
only
OP[2]
RD-PRE
Read preamble type
Write
only
OP[3]
nWR
Write-recovery for autoprecharge command
Write
only
OP[6:4]
0b: Reserved
1b: WR preamble = 2 ×
5, 6
tCK
0b: RD preamble = Static (default)
3, 5, 6
1b: RD preamble = Toggle
000b: nWR = 6 (default)
001b: nWR = 10
010b: nWR = 16
011b: nWR = 20
100b: nWR = 24
101b: nWR = 30
110b: nWR = 34
111b: nWR = 40
44
2, 5, 6
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 21: MR1 Op-Code Bit Definitions (Continued)
Feature
Type
OP
RD-PST
Read postamble length
Write
only
OP[7]
Notes:
Definition
Notes
0b: RD postamble = 0.5 × tCK (default)
1b: RD postamble = 1.5 ×
4, 5, 6
tCK
1. Burst length on-the-fly can be set to either BL = 16 or BL = 32 by setting the BL bit in the
command operands. See the Command Truth Table.
2. The programmed value of nWR is the number of clock cycles the device uses to determine the starting point of an internal precharge after a write burst with auto precharge
(AP) enabled. See Frequency Ranges for RL, WL, and nWR Settings table.
3. For READ operations, this bit must be set to select between a toggling preamble and a
non-toggling preamble. (See the Preamble section.)
4. OP[7] provides an optional READ postamble with an additional rising and falling edge
of DQS_t. The optional postamble cycle is provided for the benefit of certain memory
controllers.
5. There are two physical registers assigned to each bit of this MR parameter, designated
set point 0 and set point 1. Only the registers for the set point determined by the state
of the FSP‐WR bit (MR13 OP[6]) will be written to with an MRW command to this MR
address.
6. There are two physical registers assigned to each bit of this MR parameter, designated
set point 0 and set point 1. The device will operate only according to the values stored
in the registers for the active set point, that is, the set point determined by the state of
the FSP‐OP bit (MR13 OP[7]). The values in the registers for the inactive set point will be
ignored by the device and may be changed without affecting device operation.
45
Table 22: Burst Sequence for Read
C4 C3 C2 C1 C0 1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
16-Bit READ Operation
V
0
0
0
0
0
1
2
3
4
5
6
7
8
9
A
B
C
D
E
F
V
0
1
0
0
4
5
6
7
8
9
A
B
C
D
E
F
0
1
2
3
V
1
0
0
0
8
9
A
B
C
D
E
F
0
1
2
3
4
5
6
7
V
1
1
0
0
C
D
E
F
0
1
2
3
4
5
6
7
8
9
A
B
32-Bit READ Operation
0
0
0
0
0
0
1
2
3
4
5
6
7
8
9
A
B
C
D
E
F
10 11 12 13 14 15 16 17 18 19 1A 1B 1C 1D 1E 1F
0
0
1
0
0
4
5
6
7
8
9
A
B
C
D
E
F
0
1
2
3
14 15 16 17 18 19 1A 1B 1C 1D 1E 1F 10 11 12 13
0
1
0
0
0
8
9
A
B
C
D
E
F
0
1
2
3
4
5
6
7
18 19 1A 1B 1C 1D 1E 1F 10 11 12 13 14 15 16 17
D
E
F
0
1
2
3
4
5
6
7
8
9
A
B 1C 1D 1E 1F 10 11 12 13 14 15 16 17 18 19 1A 1B
46
0
1
1
0
0
C
1
0
0
0
0
10 11 12 13 14 15 16 17 18 19 1A 1B 1C 1D 1E 1F
0
1
2
3
4
5
6
7
8
9
A
B
C
D
E
F
1
0
1
0
0
14 15 16 17 18 19 1A 1B 1C 1D 1E 1F 10 11 12 13
4
5
6
7
8
9
A
B
C
D
E
F
0
1
2
3
1
1
0
0
0
18 19 1A 1B 1C 1D 1E 1F 10 11 12 13 14 15 16 17
8
9
A
B
C
D
E
F
0
1
2
3
4
5
6
7
1
1
1
0
0
1C 1D 1E 1F 10 11 12 13 14 15 16 17 18 19 1A 1B C
D
E
F
0
1
2
3
4
5
6
7
8
9
A
B
Notes:
1. C[1:0] are not present on the CA bus; they are implied to be zero.
2. The starting burst address is on 64-bit (4n) boundaries.
Table 23: Burst Sequence for Write
C4 C3 C2 C1 C0 1
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
1
2
3
4
5
6
7
8
9
A
B
C
D
E
F
1
2
3
4
5
6
7
8
9
A
B
C
D
E
F
16-Bit WRITE Operation
V
0
0
0
0
0
32-Bit WRITE Operation
0
0
0
0
0
0
Notes:
1.
2.
3.
4.
10 11 12 13 14 15 16 17 18 19 1A 1B 1C 1D 1E 1F
C[1:0] are not present on the CA bus; they are implied to be zero.
The starting burst address is on 256-bit (16n) boundaries for burst length 16.
The starting burst address is on 512-bit (32n) boundaries for burst length 32.
C[3:2] must be set to 0 for all WRITE operations.
200b: x32 LPDDR4 SDRAM
Mode Registers
2
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 24: MR2 Device Feature 2 (MA[7:0] = 02h)
OP7
OP6
WR Lev
WLS
OP5
OP4
OP3
OP2
OP1
WL
OP0
RL
Table 25: MR2 Op-Code Bit Definitions
Feature
Type
OP
RL
READ latency
Writeonly
OP[2:0]
Definition
Notes
RL and nRTP for DBI-RD disabled (MR3 OP[6] = 0b)
1, 3, 4
000b: RL = 6, nRTP = 8 (default)
001b: RL = 10, nRTP = 8
010b: RL = 14, nRTP = 8
011b: RL = 20, nRTP = 8
100b: RL = 24, nRTP = 10
101b: RL = 28, nRTP = 12
110b: RL = 32, nRTP = 14
111b: RL = 36, nRTP = 16
RL and nRTP for DBI-RD enabled (MR3 OP[6] = 1b)
000b: RL = 6, nRTP = 8
001b: RL = 12,nRTP = 8
010b: RL = 16, nRTP = 8
011b: RL = 22, nRTP = 8
100b: RL = 28, nRTP = 10
101b: RL = 32, nRTP = 12
110b: RL = 36, nRTP = 14
111b: RL = 40, nRTP = 16
47
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 25: MR2 Op-Code Bit Definitions (Continued)
Feature
Type
OP
WL
WRITE latency
Writeonly
OP[5:3]
Definition
Notes
WL set A (MR2 OP[6] = 0b)
1, 3, 4
000b: WL = 4 (default)
001b: WL = 6
010b: WL = 8
011b: WL = 10
100b: WL = 12
101b: WL = 14
110b: WL = 16
111b: WL = 18
WL set B (MR2 OP[6] = 1b)
000b: WL = 4
001b: WL = 8
010b: WL = 12
011b: WL = 18
100b: WL = 22
101b: WL = 26
110b: WL = 30
111b: WL = 34
WLS
WRITE latency
set
Writeonly
OP[6]
WR Lev
Write leveling
Writeonly
OP[7]
Notes:
0b: Use WL set A (default)
1, 3, 4
1b: Use WL set B
0b: Disable write leveling (default)
2
1b: Enable write leveling
1. See Latency Code Frequency Table for allowable frequency ranges for RL/WL/nWR.
2. After an MRW command to set the write-leveling enable bit (OP[7] = 1b), the device remains in the MRW state until another MRW command clears the bit (OP[7] = 0b). No
other commands are allowed until the write-leveling enable bit is cleared.
3. There are two physical registers assigned to each bit of this MR parameter, designated
set point 0 and set point 1. Only the registers for the set point determined by the state
of the FSP‐WR bit (MR13 OP[6]) will be written to with an MRW command this MR address, or read from with an MRR command to this address.
4. There are two physical registers assigned to each bit of this MR parameter, designated
set point 0 and set point 1. The device will operate only according to the values stored
in the registers for the active set point, that is, the set point determined by the state of
the FSP‐OP bit (MR13 OP[7]). The values in the registers for the inactive set point will be
ignored by the device and may be changed without affecting device operation.
5. nRTP is valid for BL16 only. For BL32, the SDRAM will add 8 clocks to the nRTP value before starting a precharge.
48
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 26: Frequency Ranges for RL, WL, nWR, and nRTP Settings
READ Latency
WRITE Latency
No DBI
w/DBI
Set A
Set B
nWR
nRTP
Lower
Frequency
Limit (>)
6
6
4
4
6
8
10
266
10
12
6
8
10
8
266
533
14
16
8
12
16
8
533
800
20
22
10
18
20
8
800
1066
24
28
12
22
24
10
1066
1333
28
32
14
26
30
12
1333
1600
32
36
16
30
34
14
1600
1866
36
40
18
34
40
16
1866
2133
Notes:
Upper
Frequency
Limit(≤
≤)
Units
Notes
MHz
1–6
1. The device should not be operated at a frequency above the upper frequency limit or
below the lower frequency limit shown for each RL, WL, or nWR value.
2. DBI for READ operations is enabled in MR3 OPO[6]. When MR3 OP[6] = 0, then the "No
DBI" column should be used for READ latency. When MR3 OP[6] = 1, then the "w/DBI"
column should be used for READ latency.
3. WRITE latency set A and set B are determined by MR2 OP[6]. When MR2 OP[6] = 0, then
WRITE latency set A should be used. When MR2 OP[6] = 1, then WRITE latency set B
should be used.
4. The programmed value for nRTP is the number of clock cycles the device uses to determine the starting point of an internal PRECHARGE operation after a READ burst with AP
(auto precharge) enabled . It is determined by RU(tRTP/tCK).
5. The programmed value of nWR is the number of clock cycles the device uses to determine the starting point of an internal PRECHARGE operation after a WRITE burst with
AP (auto precharge) enabled. It is determined by RU(tWR/tCK).
6. nRTP shown in this table is valid for BL16 only. For BL32, the device will add 8 clocks to
the nRTP value before starting a precharge.
Table 27: MR3 I/O Configuration 1 (MA[7:0] = 03h)
OP7
OP6
DBI-WR
DBI-RD
OP5
OP4
OP3
PDDS
49
OP2
OP1
OP0
PPRP
WR-PST
PU-CAL
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 28: MR3 Op-Code Bit Definitions
Feature
OP
Definition
Notes
PU-CAL
(Pull-up calibration point)
Type
OP[0]
0b: VDDQ/2.5
1–4
WR-PST (WR postamble length)
OP[1]
1b: VDDQ/3 (default)
0b: WR post-amble=0.5 × tCK (default)
2, 3, 5
1b: WR post-amble=1.5 × tCK
PPRP (Post-package repair protection)
OP[2]
0b: PPR protection disabled (default)
6
1b: PPR protection enabled
PDDS
(Pull-down drive strength)
000b: RFU
1, 2, 3
001b: RZQ/1
010b: RZQ/2
Write-only
011b: RZQ/3
OP[5:3]
100b: RZQ/4
101b: RZQ/5
110b:RZQ/6 (default)
111b: Reserved
DBI-RD
(DBI-read enable)
OP[6]
DBI-WR
(DBI-write enable)
OP[7]
0b: Disabled (default)
2, 3
1b: Enabled
0b: Disabled (default)
2, 3
1b: Enabled
Notes:
1. All values are typical. The actual value after calibration will be within the specified tolerance for a given voltage and temperature. Recalibration may be required as voltage and
temperature vary.
2. There are two physical registers assigned to each bit of this MR parameter, designated
set point 0 and set point 1. Only the registers for the set point determined by the state
of the FSP‐WR bit (MR13 OP[6]) will be written to with an MRW command to this MR
address, or read from with an MRR command to this address.
3. There are two physical registers assigned to each bit of this MR parameter, designated
set point 0 and set point 1.The device will operate only according to the values stored in
the registers for the active set point, i.e., the set point determined by the state of the
FSP‐OP bit (MR13 OP[7]). The values in the registers for the inactive set point will be determined by the state of the FSP‐OP bit (MR13 OP[7]). The values in the registers for the
inactive set point will be ignored by the device, and may be changed without affecting
device operation.
4. For dual channel device, PU‐CAL (MR3‐OP[0]) must be set the same for both channels on
a die. The SDRAM will read the value of only one register (Ch.A or Ch.B), vendor-specific, so both channels must be set the same.
5. Refer to the supplier data sheet for vender-specific function. 1.5 × tCK apply ≥ 1.6GHz
clock.
6. If MR3 OP[2] is set to 1b, PPR protection mode is enabled. The PPR protection bit is a
sticky bit and can only be set to 0b by a power on reset. MR4 OP[4] controls entry to PPR
mode. If PPR protection is enabled then the DRAM will not allow writing of 1b to MR4
OP[4].
50
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 29: MR4 Device Temperature (MA[7:0] = 04h)
OP7
OP6
TUF
OP5
Thermal offset
OP4
OP3
PPRE
SR abort
OP2
OP1
OP0
Refresh rate
Table 30: MR4 Op-Code Bit Definitions
Feature
Refresh rate
Type
OP
Read-only
OP[2:0]
Definition
Notes
000b: SDRAM low temperature operating limit exceeded
001b: 4x refresh
1–4,
7–9
010b: 2x refresh
011b: 1x refresh (default)
100b: 0.5x refresh
101b: 0.25x refresh, no derating
110b: 0.25x refresh, with derating
111b: SDRAM high temperature operating limit exceeded
SR abort (Self Refresh
Abort)
Write
OP[3]
PPRE
(Post-package repair
entry/exit)
Write
OP[4]
Thermal Offset-Controller offset to TCSR (Vendor-specific function)
Write
0b: Disable (default)
9
1b: Device dependent
0b: Exit PPR mode (default)
5, 9,
1b: Enter PPR mode (Reference MR25 OP[7:0] for available PPR resources)
OP[6:5]
00b: No offset, 0~5°C gradient (default)
9
01b: 5°C offset, 5~10°C gradient
10b: 10°C offset, 10~15°C gradient
11b: Reserved
TUF
(Temperature update
flag)
Read-only
OP7
0b: OP[2:0] No change in OP[2:0] since last MR4 read (default)
6–8
1b: Change in OP[2:0] since last MR4 read
Notes:
1. The refresh rate for each MR4 OP[2:0] setting applies to tREFI, tREFIpb, and tREFW. MR4
OP[2:0] = 011b corresponds to a device temperature of 85°C. Other values require either
a longer (2x, 4x) refresh interval at lower temperatures or a shorter (0.5x, 0.25x) refresh
interval at higher temperatures. If MR4 OP[2] = 1b, the device temperature is greater
than 85°C.
2. At higher temperatures (>85°C), AC timing derating may be required. If derating is required the device will set MR4 OP[2:0] = 110b. See derating timing requirements in the
AC Timing section.
3. DRAM vendors may or may not report all of the possible settings over the operating
temperature range of the device. Each vendor guarantees that their device will work at
any temperature within the range using the refresh interval requested by their device.
4. The device may not operate properly when MR4 OP[2:0 ] = 000b or 111b.
5. Post‐package repair can be entered or exited by writing to MR4 OP[4].
6. When MR4 OP[7] = 1b, the refresh rate reported in MR4 OP[2:0] has changed since the
last MR4 read. A mode register read from MR4 will reset MR4 OP[7] to 0b.
7. MR4 OP[7] = 0b at power‐up. MR4 OP[2:0] bits are undefined at power‐up.
51
200b: x32 LPDDR4 SDRAM
Mode Registers
8. See the Temperature Sensor section for information on the recommended frequency of
reading MR4.
9. MR4 OP[6:3] can be written in this register. All other bits will be ignored by the device
during an MRW command to this register.
Table 31: MR5 Basic Configuration 1 (MA[7:0] = 05h)
OP7
OP6
OP5
OP4
OP3
OP2
OP1
OP0
OP1
OP0
OP1
OP0
Manufacturer ID
Table 32: MR5 Op-Code Bit Definitions
Feature
Manufacturer ID
Type
OP
Definition
Read-only
OP[7:0]
1111 1111b : Micron
All others: Reserved
Table 33: MR6 Basic Configuration 2 (MA[7:0] = 06h)
OP7
OP6
OP5
OP4
OP3
OP2
Revision ID1
Note:
1. MR6 is vendor-specific.
Table 34: MR6 Op-Code Bit Definitions
Feature
Revision ID1
Note:
Type
OP
Definition
Read-only
OP[7:0]
xxxx xxxxb: Revision ID1
1. MR6 is vendor-specific.
Table 35: MR7 Basic Configuration 3 (MA[7:0] = 07h)
OP7
OP6
OP5
OP4
OP3
OP2
Revision ID2
Table 36: MR7 Op-Code Bit Definitions
Feature
Revision ID2
Note:
Type
OP
Definition
Read-only
OP[7:0]
1. MR7 is vendor-specific.
52
xxxx xxxxb: Revision ID2
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 37: MR8 Basic Configuration 4 (MA[7:0] = 08h)
OP7
OP6
OP5
OP4
I/O width
OP3
OP2
OP1
Density
OP0
Type
Table 38: MR8 Op-Code Bit Definitions
Feature
Type
Type
OP
Read-only
OP[1:0]
Definition
00b: S16 SDRAM (16n prefetch)
All others: Reserved
Density
Read-only
OP[5:2]
0000b: 4Gb per die (2Gb per channel)
0001b: 6Gb per die (3Gb per channel)
0010b: 8Gb per die (4Gb per channel) or 4Gb per die (4Gb per channel)
0011b: 12Gb per die (6Gb per channel) or 6Gb per die (6Gb per channel)
0100b: 16Gb per die (8Gb per channel) or 8Gb per die (8Gb per channel)
0101b: 24Gb per die (12Gb per channel)
0110b: 32Gb per die (16Gb per channel)
All others: Reserved
I/O width
Read-only
OP[7:6]
00b: x16/channel
All others: Reserved
Table 39: MR9 Test Mode (MA[7:0] = 09h)
OP7
OP6
OP5
OP4
OP3
OP2
OP1
OP0
Vendor-specific test mode
Table 40: MR9 Op-Code Definitions
Feature
Test mode
Type
OP
Write-only
OP[7:0]
Definition
0000000b; Vendor-specific test mode disabled (default)
Table 41: MR10 Calibration (MA[7:0] = 0Ah)
OP7
OP6
OP5
OP4
OP3
RFU
OP2
OP1
OP0
ZQ RESET
53
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 42: MR10 Op-Code Bit Definitions
Feature
Type
OP
ZQ reset
Write-only
OP[0]
Definition
0b: Normal operation (default)
1b: ZQ reset
Notes:
1. See AC Timing table for calibration latency and timing.
2. If ZQ is connected to VDDQ through RZQ, either the ZQ CALIBRATION function or default
calibration (via ZQ reset) is supported. If ZQ is connected to VSS, the device operates
with default calibration and ZQ CALIBRATION commands are ignored. In both cases, the
ZQ connection must not change after power is supplied to the device.
Table 43: MR11 ODT Control (MA[7:0] = 0Bh)
OP7
OP6
RFU
OP5
OP4
OP3
CA ODT
OP2
RFU
OP1
OP0
DQ ODT
Table 44: MR11 Op-Code Bit Definitions
Notes 1-3 apply to entire table
Feature
Type
DQ ODT
DQ bus receiver on-dietermination
Write-only
OP
OP[2:0]
Definition
000b: Disable (default)
001b: RZQ/1
010b: RZQ/2
011b: RZQ/3
100b: RZQ/4
101b: RZQ/5
110b: RZQ/6
111b: RFU
CA ODT
CA bus receiver on-dietermination
Write-only
OP[6:4]
000b: Disable (default)
001b: RZQ/1
010b: RZQ/2
011b: RZQ/3
100b: RZQ/4
101b: RZQ/5
110b: RZQ/6
111b: RFU
Notes:
1. All values are typical. The actual value after calibration will be within the specified tolerance for a given voltage and temperature. Re‐calibration may be required as voltage
and temperature vary.
2. There are two physical registers assigned to each bit of this MR parameter, designated
set point 0 and set point 1. Only the registers for the set point determined by the state
of the FSP‐WR bit (MR13 OP[6]) will be written to with an MRW command to this MR
address, or read from with an MRR command to this address.
54
200b: x32 LPDDR4 SDRAM
Mode Registers
3. There are two physical registers assigned to each bit of this MR parameter, designated
set point 0 and set point 1. The device will operate only according to the values stored
in the registers for the active set point, i.e., the set point determined by the state of the
FSP‐OP bit (MR13 OP[7]). The values in the registers for the inactive set point will be ignored by the device and may be changed without affecting device operation.
Table 45: MR12 Register Information (MA[7:0] = 0Ch)
OP7
OP6
RFU
VRCA
OP5
OP4
OP3
OP2
OP1
OP0
VREF(CA)
Table 46: MR12 Op-Code Bit Definitions
Feature
Type
OP
VREF(CA)
VREF(CA) settings
Read/
Write
OP[5:0]
VRCA
VREF(CA) range
Read/
Write
OP[6]
Notes:
Data
Notes
000000b–110010b: See VREF Settings Table
1–3, 5, 6
All others: Reserved
0b: VREF(CA) range[0] enabled
1, 2, 4, 5,
6
1b: VREF(CA) range[1] enabled (default)
1. This register controls the VREF(CA) levels for frequency set point[1:0]. Values from either
VR(ca)[0] or VR(ca)[1] may be selected by setting MR12 OP[6] appropriately.
2. A read to MR12 places the contents of OP[7:0] on DQ[7:0]. Any RFU bits and unused DQ
will be set to 0. See the MRR Operation section.
3. A write to MR12 OP[5:0] sets the internal VREF(CA) level for FSP[0] when MR13 OP[6] = 0b
or sets the internal VREF(CA) level for FSP[1] when MR13 OP[6] = 1b. The time required for
VREF(CA) to reach the set level depends on the step size from the current level to the new
level. See the VREF(CA) training section.
4. A write to MR12 OP[6] switches the device between two internal VREF(CA) ranges. The
range (range[0] or range[1]) must be selected when setting the VREF(CA) register. The value, once set, will be retained until overwritten or until the next power‐on or reset
event.
5. There are two physical registers assigned to each bit of this MR parameter, designated
set point 0 and set point 1. Only the registers for the set point determined by the state
of the FSP‐WR bit (MR13 OP[6]) will be written to with an MRW command to this MR
address, or read from with an MRR command to this address.
6. There are two physical registers assigned to each bit of this MR parameter, designated
set point 0 and set point 1. The device will operate only according to the values stored
in the registers for the active set point, i.e., the set point determined by the state of the
FSP‐OP bit (MR13 OP[7]). The values in the registers for the inactive set point will be ignored by the device, and may be changed without affecting device operation.
Table 47: MR13 Register Control (MA[7:0] = 0Dh)
OP[7]
OP[6]
OP[5]
OP[4]
OP[3]
OP[2]
OP[1]
OP[0]
FSP-OP
FSP-WR
DMD
RRO
VRCG
VRO
RPT
CBT
55
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 48: MR13 Op-Code Bit Definition
Feature
CBT
Command bus training
Type
OP
Write-only
OP[0]
Definition
0b: Normal operation (default)
Notes
1
1b: Command bus training mode enabled
RPT
Read preamble training
OP[1]
0b: Disabled (default)
VRO
VREF output
OP[2]
0b: Normal operation (default)
VRCG
VREF current generator
OP[3]
0b: Normal operation (default)
RRO
Refresh rate option
OP[4]
DMD
Data mask disable
OP[5]
FSP-WR
Frequency set point write /
read
OP[6]
FSP-OP
Frequency set point operation mode
OP[7]
1b: Read preamble training mode enabled
2
1b: Output the VREF(CA) and VREF(DQ) values on DQ bits
3
1b: Fast response (high current) mode
0b: Disable MR4 OP[2:0] (default)
4, 5
1b: Enable MR4 OP[2:0]
0b: DATA MASK operation enabled (default)
6
1b: DATA MASK operation disabled
Notes:
0b: Frequency set point[0] (default)
7
1b: Frequency set point[1]
0b: Frequency set point[0] (default)
8
1b: Frequency set point[1]
1. A write to set OP[0]=1 causes the LPDDR4‐SDRAM to enter the Command Bus Training
mode. When OP[0]=1 and CKE goes LOW, commands are ignored and the contents of
CA[5:0] are mapped to the DQ bus. CKE must be brought HIGH before doing a MRW to
clear this bit (OP[0]=0) and return to normal operation. See the Command Bus Training
section for more information.
2. When set, the device will output the VREF(CA) and VREF(DQ) voltage on DQ pins. Only the
"active" frequency set point, as defined by MR13 OP[7], will be output on the DQ pins.
This function allows an external test system to measure the internal VREF levels. The DQ
pins used for VREF output are vendor-specific.
3. When OP[3] = 1, the VREF circuit uses a high current mode to improve VREF settling time.
4. MR13 OP[4] RRO bit is valid only when MR0 OP[0] = 1. For LPDDR4‐SDRAM with MR0
OP[0] = 0, MR4 OP[2:0] bits are not dependent on MR13 OP[4].
5. When OP[4] = 0, only 001b and 010b in MR4 OP[2:0] are disabled. LPDDR4‐SDRAM must
report 011b instead of 001b or 010b in this case. Controller should follow the refresh
mode reported by MR4 OP[2:0], regardless of RRO setting. TCSR function does not depend on RRO setting.
6. When enabled (OP[5] = 0b) data masking is enabled for the device. When disabled
(OP[5] = 1b), the device will ignore any mask patterns issued during a MASKED WRITE
command. See the Data Mask section for more information.
7. FSP‐WR determines which frequency set point registers are accessed with MRW and
MRR commands for the following functions such as VREF(CA) setting, VREF(CA) range,
VREF(DQ) setting, VREF(DQ) range. For more information, refer to Frequency Set Point section.
8. FSP‐OP determines which frequency set point register values are currently used to specify device operation for the following functions such as VREF(CA) setting, VREF(CA) range,
VREF(DQ) setting, VREF(DQ) range. For more information, refer to Frequency Set Point section.
56
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 49: Mode Register14 (MA[7:0] = 0Eh)
OP[7]
OP[6]
RFU
VRDQ
OP[5]
OP[4]
OP[3]
OP[2]
OP[1]
OP[0]
VREF(DQ)
Table 50: MR14 Op-Code Bit Definition
Feature
Type
OP
VREF(DQ)
VREF(DQ) setting
Read/
Write
OP[5:0]
VRDQ
VREF(DQ) range
Definition
Notes
000000b–110010b: See VREF Settings Table
1–3, 5, 6
All others: Reserved
OP[6]
0b: VREF(DQ) range[0] enabled
1, 2, 4–6
1b: VREF(DQ) range[1] enabled (default)
Notes:
1. This register controls the VREF(DQ) levels for frequency set point[1:0]. Values from either
VRDQ[vendor defined] or VRDQ[vendor defined] may be selected by setting OP[6] appropriately.
2. A read (MRR) to this register places the contents of OP[7:0] on DQ[7:0]. Any RFU bits and
unused DQ will be set to 0. See the MRR Operation section.
3. A write to OP[5:0] sets the internal VREF(DQ) level for FSP[0] when MR13 OP[6] = 0b, or
sets FSP[1] when MR13 OP[6] = 1b. The time required for VREF(DQ) to reach the set level
depends on the step size from the current level to the new level. See the VREF(DQ) training section.
4. A write to OP[6] switches the device between two internal VREF(DQ) ranges. The range
(range[0] or range[1]) must be selected when setting the VREF(DQ) register. The value,
once set, will be retained until overwritten, or until the next power‐on or reset event.
5. There are two physical registers assigned to each bit of this MR parameter, designated
set point 0 and set point 1. Only the registers for the set point determined by the state
of the FSP‐WR bit (MR13 OP[6]) will be written to with an MRW command to this MR
address, or read from with an MRR command to this address.
6. There are two physical registers assigned to each bit of this MR parameter, designated
set point 0 and set point 1. The device will operate only according to the values stored
in the registers for the active set point, i.e., the set point determined by the state of the
FSP‐OP bit (MR13 OP[7]). The values in the registers for the inactive set point will be ignored by the device, and may be changed without affecting device operation.
57
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 51: VREF Setting for Range[0] and Range[1]
Notes 1-3 apply to entire table
Range[0] Values
Range[1] Values
VREF(CA) (% of VDD2 )
Function
VREF Setting
for MR12
and MR14
OP
VREF(CA) (% of VDD2 )
VREF(DQ) (% of VDDQ )
VREF(DQ) (% of VDDQ )
OP[5:0] 000000b: 10.0%
011010b: 20.4%
000000b: 22.0%
011010b: 32.4%
000001b: 10.4%
011011b: 20.8%
000001b: 22.4%
011011b: 32.8%
000010b: 10.8%
011100b: 21.2%
000010b: 22.8%
011100b: 33.2%
000011b: 11.2%
011101b: 21.6%
000011b: 23.2%
011101b: 33.6%
000100b: 11.6%
011110b: 22.0%
000100b: 23.6%
011110b: 34.0%
000101b: 12.0%
011111b: 22.4%
000101b: 24.0%
011111b: 34.4%
000110b: 12.4%
100000b: 22.8%
000110b: 24.4%
100000b: 34.8%
000111b: 12.8%
100001b: 23.2%
000111b: 24.8%
100001b: 35.2%
001000b: 13.2%
100010b: 23.6%
001000b: 25.2%
100010b: 35.6%
001001b: 13.6%
100011b: 24.0%
001001b: 25.6%
100011b: 36.0%
001010b: 14.0%
100100b: 24.4%
001010b: 26.0%
100100b: 36.4%
001011b: 14.4%
100101b: 24.8%
001011b: 26.4%
100101b: 36.8%
001100b: 14.8%
100110b: 25.2%
001100b: 26.8%
100110b: 37.2%
001101b: 15.2%
100111b: 25.6%
001101b: 27.2% default
100111b: 37.6%
001110b: 15.6%
101000b: 26.0%
001110b: 27.6%
101000b: 38.0%
001111b: 16.0%
101001b: 26.4%
001111b: 28.0%
101001b: 38.4%
010000b: 16.4%
101010b: 26.8%
010000b: 28.4%
101010b: 38.8%
010001b: 16.8%
101011b: 27.2%
010001b: 28.8%
101011b: 39.2%
010010b: 17.2%
101100b: 27.6%
010010b: 29.2%
101100b: 39.6%
010011b: 17.6%
101101b: 28.0%
010011b: 29.6%
101101b: 40.0%
010100b: 18.0%
101110b: 28.4%
010100b: 30.0%
101110b: 40.4%
010101b: 18.4%
101111b: 28.8%
010101b: 30.4%
101111b: 40.8%
010110b: 18.8%
110000b: 29.2%
010110b: 30.8%
110000b: 41.2%
010111b: 19.2%
110001b: 29.6%
010111b: 31.2%
110001b: 41.6%
011000b: 19.6%
110010b: 30.0%
011000b: 31.6%
110010b: 42.0%
011001b: 20.0%
All Others: Reserved
011001b: 32.0%
All Others: Reserved
Notes:
1. These values may be used for MR14 OP[5:0] and MR12 OP[5:0] to set the VREF(CA) or
VREF(DQ) levels in the device.
2. The range may be selected in each of the MR14 or MR12 registers by setting OP[6] appropriately.
3. Each of the MR14 or MR12 registers represents either FSP[0] or FSP[1]. Two frequency set
points each for CA and DQ are provided to allow for faster switching between terminated and un‐terminated operation or between different high‐frequency settings, which
may use different terminations values.
58
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 52: MR15 Register Information (MA[7:0] = 0Fh)
OP[7]
OP[6]
OP[5]
OP[4]
OP[3]
OP[2]
OP[1]
OP[0]
Lower-byte invert register for DQ calibration
Table 53: MR15 Op-code Bit Definition
Feature
Lower-byte invert for
DQ calibration
Type
OP
Definition
Write-Only
OP[7:0]
Notes
The following values may be written for any operand
OP[7:0] and will be applied to the corresponding DQ locations DQ[7:0] within a byte lane
1–3
0b: Do not invert
1b: Invert the DQ calibration patterns in MR32 and
MR40
Default value for OP[7:0] = 55h
Notes:
1. This register will invert the DQ Calibration pattern found in MR32 and MR40 for any single DQ or any combination of DQ. Example: If MR15 OP[7:0] = 00010101b, then the DQ
calibration patterns transmitted on DQ[7, 6, 5, 3, 1] will not be inverted, but the DQ calibration patterns transmitted on DQ[4, 2, 0] will be inverted.
2. DM[0] is not inverted and always transmits the "true" data contained in MR32 and
MR40.
3. No data bus inversion (DBI) function is enacted during DQ read calibration, even if DBI is
enabled in MR3-OP[6].
Table 54: MR15 Invert Register Pin Mapping
PIN
DQ0
DQ1
DQ2
DQ3
DMIO
DQ4
DQ5
DQ6
DQ7
MR15
OP0
OP1
OP2
OP3
No invert
OP4
OP5
OP6
OP7
Table 55: MR16 PASR Bank Mask (MA[7:0] = 010h)
OP7
OP6
OP5
OP4
OP3
OP2
OP1
PASR bank mask
Table 56: MR16 Op-Code Bit Definitions
Feature
Bank[7:0] mask
Type
OP
Write-only
OP[7:0]
Definition
0b: Bank refresh enabled (default)
1b: Bank refresh disabled
OP[n]
Bank Mask
8-Bank SDRAM
0
xxxxxxx1
Bank 0
1
xxxxxx1x
Bank 1
59
OP0
200b: x32 LPDDR4 SDRAM
Mode Registers
OP[n]
Bank Mask
8-Bank SDRAM
2
xxxxx1xx
Bank 2
3
xxxx1xxx
Bank 3
4
xxx1xxxx
Bank 4
5
xx1xxxxx
Bank 5
6
x1xxxxxx
Bank 6
7
1xxxxxxx
Bank 7
Notes:
1. When a mask bit is asserted (OP[n] = 1), refresh to that bank is disabled.
2. PASR bank masking is on a per-channel basis; the two channels on the die may have different bank masking.
Table 57: MR17 PASR Segment Mask (MA[7:0] = 11h)
OP7
OP6
OP5
OP4
OP3
OP2
OP1
OP0
PASR segment mask
Table 58: MR17 PASR Segment Mask Definitions
Feature
Segment[7:0] mask
Type
OP
Write-only
OP[7:0]
Definition
0b: Segment refresh enabled (default)
1b: Segment refresh disabled
Table 59: MR17 PASR Segment Mask
Density (per channel)
2Gb
3Gb
4Gb
6Gb
8Gb
12Gb
16Gb
R[13:11]
R[14:12]
R[14:12]
R[15:13]
R[15:13]
TBD
TBD
110b
Not
allowed
110b
Segment
OP
Segment
Mask
0
0
XXXXXXX1
000b
1
1
XXXXXX1X
001b
2
2
XXXXX1XX
010b
3
3
XXXX1XXX
011b
4
4
XXX1XXXX
100b
5
5
XX1XXXXX
6
6
X1XXXXXX
110b
7
7
1XXXXXXX
111b
Notes:
101b
Not
allowed
110b
111b
Not
allowed
111b
111b
1. This table indicates the range of row addresses in each masked segment. "X" is “Don’t
Care” for a particular segment.
2. PASR segment-masking is on a per-channel basis. The two channels on the die may have
different segment masking.
3. For 3Gb, 6Gb, and 12Gb density per channel, OP[7:6] must always be LOW (= 00b).
60
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 60: MR18 Register Information (MA[7:0]=12h)
OP7
OP6
OP5
OP4
OP3
OP2
OP1
OP0
DQS oscillator count - LSB
Table 61: MR18 LSB DQS Oscillator Count
Notes 1–3 apply to entire table
Function
Type
DQS oscillator
count (WR training
DQS oscillator)
Read-only
Notes:
OP
OP[7:0]
Definition
0h–FFh LSB DRAM DQS oscillator count
1. MR18 reports the LSB bits of the DRAM DQS oscillator count. The DRAM DQS oscillator
count value is used to train DQS to the DQ data valid window. The value reported by
the DRAM in this mode register can be used by the memory controller to periodically
adjust the phase of DQS relative to DQ.
2. Both MR18 and MR19 must be read (MRR) and combined to get the value of the DQS
oscillator count.
3. The value in this register is reset each time an MPC command is issued to start in the
DQS oscillator counter.
Table 62: MR19 Register Information (MA[7:0] = 13h)
OP7
OP6
OP5
OP4
OP3
OP2
OP1
OP0
DQS oscillator count – MSB
Table 63: MR19 DQS Oscillator Count
Notes 1–3 apply to the entire table
Function
Type
DQS oscillator
count – MSB (WR
training DQS oscillator)
Read-only
Notes:
OP
OP[7:0]
Definition
0h–FFh MSB DRAM DQS oscillator count
1. MR19 reports the MSB bits of the DRAM DQS oscillator count. The DRAM DQS oscillator
count value is used to train DQS to the DQ data valid window. The value reported by
the DRAM in this mode register can be used by the memory controller to periodically
adjust the phase of DQS relative to DQ.
2. Both MR18 and MR19 must be read (MRR) and combined to get the value of the DQS
oscillator count.
3. A new MPC[START DQS OSCILLATOR]] should be issued to reset the contents of MR18/
MR19.
61
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 64: MR20 Register Information (MA[7:0] = 14h)
OP7
OP6
OP5
OP4
OP3
OP2
OP1
OP0
Upper-byte invert register for DQ calibration
Table 65: MR20 Register Information
Notes 1–3 apply to entire table
Function
Type
Upper-byte invert
for DQ calibration
OP
Write-only
OP[7:0]
Definition
The following values may be written for any operand OP[7:0] and will
be applied to the corresponding DQ locations DQ[15:8] within a byte
lane
0b: Do not invert
1b: Invert the DQ calibration patterns in MR32 and MR40
Default value for OP[7:0] = 55h
Notes:
1. This register will invert the DQ calibration pattern found in MR32 and MR40 for any single DQ or any combination of DQ. For example, if MR20 OP[7:0] = 00010101b, the DQ
calibration patterns transmitted on DQ[15, 14, 13, 11, 9] will not be inverted, but the DQ
calibration patterns transmitted on DQ[12, 10, 8] will be inverted.
2. DM[1] is not inverted and always transmits the true data contained in MR32 and MR40.
3. No data bus inversion (DBI) function is enacted during DQ read calibration, even if DBI is
enabled in MR3 OP[6].
Table 66: MR20 Invert Register Pin Mapping
Pin
DQ8
DQ9
DQ10
DQ11
DMI1
DQ12
DQ13
DQ14
DQ15
MR20
OP0
OP1
OP2
OP3
No invert
OP4
OP5
OP6
OP7
Table 67: MR21 Register Information (MA[7:0] = 15h)
OP7
OP6
OP5
OP4
OP3
OP2
OP1
OP0
OP2
OP1
OP0
RFU
Table 68: MR22 Register Information (MA[7:0] = 16h)
OP7
OP6
RFU
OP5
OP4
OP3
ODTD-CA
ODTE-CS
ODTE-CK
62
SOC ODT
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 69: MR22 Register Information
Function
Type
OP
Data
Notes
SOC ODT (controller ODT val- Write-only OP[2:0] 000b: Disable (default)
ue for VOH calibration)
001b: RZQ/1
1, 2, 3
010b: RZQ/2
011b: RZQ/3
100b: RZQ/4
101b: RZQ/5
110b: RZQ/6
111b: RFU
ODTE-CK (CK ODT enabled
for non-terminating rank)
Write-only
OP[3]
2, 3, 4, 6, 8
1b: ODT-CK override enabled
ODTE-CS (CS ODT enabled for Write-only
non-terminating rank)
OP[4]
ODTD-CA (CA ODT termination disable)
OP[5]
Notes:
0b: ODT-CK override disabled (default)
Write-only
0b: ODT-CS override disabled (default)
2, 3, 5, 6, 8
1b: ODT-CS override enabled
0b: CA ODT obeys ODT_CA bond pad (default)
2, 3, 6, 7, 8
1b: CA ODT disabled
1. All values are typical.
2. There are two physical registers assigned to each bit of this MR parameter: designated
set point 0 and set point 1. Only the registers for the set point determined by the state
of the FSP-WR bit (MR13 OP[6]) will be written to with an MRW command or read from
with an MRR command to this address.
3. There are two physical registers assigned to each bit of this MR parameter: designated
set point 0 and set point 1. The device will operate only according to the values stored
in the registers for the active set point determined by the state of the FSP-OP bit (MR13
OP[7]). The values in the registers for the inactive set point will be ignored by the device
and may be changed without affecting device operation.
4. When OP[3] = 1 the CK signals will be terminated to the value set by MR11 OP[6:4] regardless of the state of the ODT_CA bond pad. This overrides the ODT_CA bond pad for
configurations where CA is shared by two or more devices but CK is not, enabling CK to
terminate on all devices.
5. When OP[4] = 1 the CS signal will be terminated to the value set by MR11 OP[6:4] regardless of the state of the ODT_CA bond pad. This overrides the ODT_CA bond pad for
configurations where CA is shared by two or more devices but CS is not, enabling CS to
terminate on all devices.
6. For system configurations where the CK, CS, and CA signals are shared between packages, the package design should provide for the ODT_CA ball to be bonded on the system board outside of the memory package. This provides the necessary control of the
ODT function for all die with shared command bus signals.
7. When OP[5] = 0, CA[5:0] will terminate when the ODT_CA bond pad is HIGH and MR11
OP[6:4] is valid and disable termination when ODT_CA is LOW or MR11 OP[6:4] is disabled. When OP[5] = 1, termination for CA[5:0] is disabled regardless of the state of the
ODT_CA bond pad or MR11 OP[6:4].
8. To ensure proper operation in a multi-rank configuration, when CA, CK or CS ODT is enabled via MR11 OP[6:4] and also via MR22 or ODT_CA pad setting, the rank providing
ODT will continue to terminate the command bus in all DRAM states including Active,
Self-refresh, Self-refresh Power-down, Active Power-down and Precharge Power-down.
63
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 70: MR23 Register Information (MA[7:0] = 17h)
OP7
OP6
OP5
OP4
OP3
OP2
OP1
OP0
DQS interval timer run-time setting
Table 71: MR23 Register Information
Notes 1–2 apply to entire table
Function
Type
DQS interval timer runtime
OP
Write-only
OP[7:0]
Data
00000000b: Disabled (default)
00000001b: DQS timer stops automatically at the 16th clock after
timer start
00000010b: DQS timer stops automatically at the 32nd clock after
timer start
00000011b: DQS timer stops automatically at the 48th clock after
timer start
00000100b: DQS timer stops automatically at the 64th clock after
timer start
--------- Through --------00111111b: DQS timer stops automatically at the (63 × 16)th clock
after timer start
01XXXXXXb: DQS timer stops automatically at the 2048th clock after
timer start
10XXXXXXb: DQS timer stops automatically at the 4096th clock after
timer start
11XXXXXXb: DQS timer stops automatically at the 8192nd clock after timer start
Notes:
1. MPC command with OP[6:0] = 1001101b (stop DQS Interval Oscillator) stops the DQS interval timer in the case of MR23 OP[7:0] = 00000000b.
2. MPC command with OP[6:0] = 1001101b (stop DQS Interval Oscillator) is illegal with valid nonzero values in MR23 OP[7:0].
Table 72: MR24 Register Information (MA[7:0] = 18h)
OP7
TRR mode
OP6
OP5
OP4
OP3
TRR mode BAn
Unlimited
MAC
64
OP2
OP1
MAC value
OP0
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 73: MR24 Register Information
Function
Type
MAC value
Read
OP
Data
Notes
OP[2:0] 000b: Unknown (OP[3] = 0) or Unlimited
(OP[3]=1)
1,2
001b: 700K
010b: 600K
011b: 500K
100b: 400K
101b: 300K
110b: 200K
111b: Reserved
Unlimited MAC
Read
OP[3]
0b: OP[2:0] defines the MAC value
1b: Unlimited MAC value
TRR mode BAn
Write
2, 3
OP[6:4] 000b: Bank 0
001b: Bank 1
010b: Bank 2
011b: Bank 3
100b: Bank 4
101b: Bank 5
110b: Bank 6
111b: Bank 7
TRR mode
Write
OP[7]
0b: Disabled (default)
1b: Enabled
Notes:
1. Unknown means that the device is not tested for tMAC and pass/fail values are unknown. Unlimited means that there is no restriction on the number of activates between refresh windows.
2. There is no restriction to the number of activates.
3. MR24 OP[2:0] set to 000b.
Table 74: MR25 Register Information (MA[7:0] = 19h)
OP7
OP6
OP5
OP4
OP3
OP2
OP1
OP0
Bank 7
Bank 6
Bank 5
Bank 4
Bank 3
Bank 2
Bank 1
Bank 0
Table 75: MR25 Register Information
Function
PPR resources
Type
OP
Data
Read-only
OP[7:0]
0b: PPR resource is not available
1b: PPR resource is available
Note:
1. When OP[n] = 0, there is no PPR resource available for that bank. When OP[n] = 1, there
is a PPR resource available for that bank, and PPR can be initiated by the controller.
65
200b: x32 LPDDR4 SDRAM
Mode Registers
Table 76: MR26:31 Register Information (MA[7:0] = 1Ah–1Fh)
OP7
OP6
OP5
OP4
OP3
OP2
OP1
OP0
OP2
OP1
OP0
Reserved for future use
Table 77: MR32 Register Information (MA[7:0] = 20h)
OP7
OP6
OP5
OP4
OP3
DQ calibration pattern A (default = 5Ah)
Table 78: MR32 Register Information
Function
Type
OP
Data
Notes
Return DQ calibration pattern Write-only OP[7:0] Xb: An MPC command issued with OP[6:0] =
MR32 + MR40
1000011b causes the device to return the DQ calibration pattern contained in this register and
(followed by) the contents of MR40. A default
pattern 5Ah is loaded at power-up or reset, or
the pattern may be overwritten with a MRW to
this register. The contents of MR15 and MR20
will invert the MR32/MR40 data pattern for a given DQ (see MR15/MR20 for more information).
Notes:
1, 2, 3
1. The patterns contained in MR32 and MR40 are transmitted on DQ[15:0] and DMI[1:0]
when DQ read calibration is initiated via an MPC command. The pattern is transmitted
serially on each data lane and organized little endian such that the low-order bit in a
byte is transmitted first. If the data pattern is 27H, the first bit transmitted is a 1 followed by 1, 1, 0, 0, 1, 0, and 0. The bit stream will be 00100111.
2. MR15 and MR20 may be used to invert the MR32/MR40 data pattern on the DQ pins.
See MR15 and MR20 for more information. Data is never inverted on the DMI[1:0] pins.
3. The data pattern is not transmitted on the DMI[1:0] pins if DBI-RD is disabled via MR3
OP[6].
4. No data bus inversion (DBI) function is enacted during DQ read calibration, even if DBI is
enabled in MR3 OP[6].
Table 79: MR33:39 Register Information (MA[7:0] = 21h–27h)
OP7
OP6
OP5
OP4
OP3
OP2
OP1
OP0
OP3
OP2
OP1
OP0
Do not use
Table 80: MR40 Register Information (MA[7:0] = 28h)
OP7
OP6
OP5
OP4
DQ calibration pattern B (default = 3Ch)
66
200b: x32 LPDDR4 SDRAM
Commands and Timing
Table 81: MR40 Register Information
Function
Type
OP
Data
Notes
Return DQ calibration pattern Write-only OP[7:0] Xb: A default pattern 3Ch is loaded at power-up
MR32 + MR40
or reset, or the pattern may be overwritten with
a MRW to this register. See MR32 for more information.
Notes:
1, 2, 3
1. The pattern contained in MR40 is concatenated to the end of MR32 and transmitted on
DQ[15:0] and DMI[1:0] when DQ read calibration is initiated via an MPC command. The
pattern is transmitted serially on each data lane and organized little endian such that
the low-order bit in a byte is transmitted first. If the data pattern in MR40 is 27H, the
first bit transmitted will be a 1, followed by 1, 1, 0, 0, 1, 0, and 0. The bit stream will be
00100111.
2. MR15 and MR20 may be used to invert the MR32/MR40 data patterns on the DQ pins.
See MR15 and MR20 for more information. Data is never inverted on the DMI[1:0] pins.
3. The data pattern is not transmitted on the DMI[1:0] pins if DBI-RD is disabled via MR3
OP[6].
4. No data bus inversion (DBI) function is enacted during DQ read calibration, even if DBI is
enabled in MR3 OP[6].
Table 82: MR41:47 Register Information (MA[7:0] = 29h–2Fh)
OP7
OP6
OP5
OP4
OP3
OP2
OP1
OP0
OP2
OP1
OP0
Do not use
Table 83: MR48:63 Register Information (MA[7:0] = 30h–3Fh)
OP7
OP6
OP5
OP4
OP3
Reserved for future use
Commands and Timing
Commands transmitted on the CA bus are encoded into two parts and are latched on
two consecutive rising edges of the clock. This is called 2-tick CA capture because each
command requires two clock edges to latch and decode the entire command.
Truth Tables
Truth tables provide complementary information to the state diagram. They also clarify
device behavior and applicable restrictions when considering the actual state of the
banks.
Unspecified operations and timings are illegal. To ensure proper operation after an illegal event, the device must be either reset by asserting the RESET_n command or powered down and then restarted using the specified initialization sequence before normal
operation can continue.
CKE signal has to be held HIGH when the commands listed in the command truth table
input.
67
200b: x32 LPDDR4 SDRAM
Truth Tables
Table 84: Command Truth Table
Commands are transmitted to the device across a six-lane interface and use CK, CKE, and CS to control the capture of
transmitted data
SDR CA Pins
Command
CS
CA0
CA1
CA2
CA3
CA4
CA5
MRW-1
H
L
H
H
L
L
OP7
1
L
MA0
MA1
MA2
MA3
MA4
MA5
2
H
L
H
H
L
H
OP6
1
L
OP0
OP1
OP2
OP3
OP4
OP5
2
H
L
H
H
H
L
V
1
L
MA0
MA1
MA2
MA3
MA4
MA5
2
REFRESH
(all/per bank)
H
L
L
L
H
L
AB
1
L
BA0
BA1
BA2
V
V
V
2
ENTER SELF REFRESH
H
L
L
L
H
H
V
1
ACTIVATE-1
H
H
L
R12
R13
R14
R15
1
L
BA0
BA1
BA2
R16
R10
R11
2
H
H
H
R6
R7
R8
R9
1
L
R0
R1
R2
R3
R4
R5
2
H
L
L
H
L
L
BL
1
L
BA0
BA1
BA2
V
C9
AP
2
EXIT SELF REFRESH
H
L
L
H
L
H
V
1
MASK WRITE-1
H
L
L
H
H
L
BL
1
L
BA0
BA1
BA2
V
C9
AP
2
H
L
L
H
H
H
V
1
MRW-2
MRR-1
ACTIVATE-2
WRITE-1
RFU
L
V
L
RFU
H
H
L
L
RFU
H
H
L
READ-1
CAS-2
(WRITE-2,
MASKED
WRITE-2,
READ-2, MRR-2,
MPC (except
NOP)
H
L
1, 2, 12
1, 2, 3, 4
1, 2
1, 2, 3, 11
1, 11
1, 2, 3, 6,
7, 9
1, 2
1, 2, 3, 5,
6, 7, 9
1, 2
L
V
2
1, 2
2
H
H
V
V
2
1, 2
2
H
L
H
L
L
L
BL
1
L
BA0
BA1
BA2
V
C9
AP
2
H
L
H
L
L
H
C8
1
L
C2
C3
C4
C5
C6
C7
2
68
1, 11
2
V
L
1, 11
2
V
L
Notes
2
V
L
CK Edge
1, 2, 3, 6,
7, 9
1, 8, 9
200b: x32 LPDDR4 SDRAM
Truth Tables
Table 84: Command Truth Table (Continued)
Commands are transmitted to the device across a six-lane interface and use CK, CKE, and CS to control the capture of
transmitted data
SDR CA Pins
Command
CS
CA0
CA1
CA2
CA3
CA4
CA5
PRECHARGE
(all/per bank)
H
L
L
L
L
H
AB
1
L
BA0
BA1
BA2
V
V
V
2
MPC
(TRAIN, NOP)
H
L
L
L
L
L
OP6
1
L
OP0
OP1
OP2
OP3
OP4
OP5
2
DESELECT
L
Notes:
X
CK Edge
1
Notes
1, 2, 3, 4
1, 2, 13
1, 2
1. All commands except for DESELECT are two clock cycles and are defined by the current
state of CS and CA[5:0] at the rising edge of the clock. DESELECT command is one clock
cycle and is not latched by the device.
2. V = H or L (a defined logic level); X = "Don't Care," in which case CS, CK_t, CK_c, and
CA[5:0] can be floated.
3. Bank addresses BA[2:0] determine which bank is to be operated upon.
4. AB HIGH during PRECHARGE or REFRESH commands indicate the command must be applied to all banks, and the bank addresses are "Don't Care."
5. MASK WRITE-1 command only supports BL16. For MASK WRITE-1 commands, CA5 must
be driven LOW on the first rising clock cycle (R1).
6. AP HIGH during a WRITE-1, MASK WRITE-1, or READ-1 command indicates that an auto
precharge will occur to the bank the command is operating on. AP LOW indicates that
no auto precharge will occur and the bank will remain open upon completion of the
command.
7. When enabled in the mode register, BL HIGH during a WRITE-1, MASK-WRITE-1, or
READ-1 command indicates the burst length should be set on-the-fly to BL = 32; BL LOW
during one of these commands indicates the burst length should be set on-the-fly to BL
= 16. If on-the-fly burst length is not enabled in the mode register, this bit should be
driven to a valid level and is ignored by the device.
8. For CAS-2 commands (WRITE-2, MASK WRITE-2, READ-2, MRR-2, or MPC (only WRITEFIFO, READ-FIFO, and READ DQ CALIBRATION)), C[1:0] are not transmitted on the CA
[5:0] bus and are assumed to be zero. Note that for CAS-2 WRITE-2 or CAS-2 MASK
WRITE-2 command, C[3:2] must be driven LOW.
9. WRITE-1, MASK-WRITE-1, READ-1, MODE REGISTER READ-1, or MPC (only WRITE-FIFO,
READ-FIFO, and READ DQ CALIBRATION) command must be immediately followed by
CAS-2 command consecutively without any other command in between. WRITE-1, MASK
WRITE-1, READ-1, MRR-1, or MPC (only WRITE-FIFO, READ-FIFO, and READ DQ CALIBRATION) command must be issued first before issuing CAS-2 command. MPC (only START
and STOP DQS OSCILLATOR, ZQCAL START and LATCH) commands do not require CAS-2
command; they require two additional DES or NOP commands consecutively before issuing any other commands.
10. The ACTIVATE-1 command must be followed by the ACTIVATE-2 command consecutively
without any other command between them. The ACTIVATE-1 command must be issued
prior to the ACTIVATE-2 command. When the ACTIVATE-1 command is issued, the ACTIVATE-2 command must be issued before issuing another ACTIVATE-1 command.
11. The MRW-1 command must be followed by the MRW-2 command consecutively without
any other command between them. The MRW-1 command must be issued prior to the
MRW-2 command.
69
200b: x32 LPDDR4 SDRAM
ACTIVATE Command
12. The MRR-1 command must be followed by the CAS-2 command consecutively without
any other commands between them. The MRR-1 command must be issued prior to the
CAS-2 command.
13. The MPC command for READ or WRITE TRAINING operations must be followed by the
CAS-2 command consecutively without any other commands between them. The MPC
command must be issued prior to the CAS-2 command.
ACTIVATE Command
The ACTIVATE command must be executed before a READ or WRITE command can be
issued. The ACTIVATE command is issued in two parts: The bank and upper-row addresses are entered with activate-1 and the lower-row addresses are entered with ACTIVATE-2. ACTIVATE-1 and ACTIVATE-2 are executed by strobing CS HIGH while setting
CA[5:0] at valid levels (see Command table) at the rising edge of CK.
The bank addresses (BA[2:0]) are used to select the desired bank. The row addresses
(R[15:0]) are used to determine which row to activate in the selected bank. The ACTIVATE-2 command must be applied before any READ or WRITE operation can be executed. The device can accept a READ or WRITE command at time tRCD after the ACTIVATE-2 command is sent. After a bank has been activated, it must be precharged to
close the active row before another ACTIVATE-2 command can be applied to the same
bank. The bank active and precharge times are defined as tRAS and tRP, respectively.
The minimum time interval between successive ACTIVATE-2 commands to the same
bank is determined by the row cycle time of the device (tRC). The minimum time interval between ACTIVATE-2 commands to different banks is tRRD.
Certain restrictions must be observed for bank ACTIVATE and REFpb operations.
• Four-activate window (tFAW): No more than 4 banks may be activated (or refreshed,
in the case of REFpb) per channel in a rolling tFAW window. Convert to clocks by dividing tFAW[ns] by tCK[ns] and rounding up to the next integer value. As an example
of the rolling window, if RU[(tFAW/tCK)] is 64 clocks, and an ACTIVATE command is
issued on clock N, no more than three additional ACTIVATE commands may be issued between clock N + 1 and N + 63. REFpb also counts as bank activation for the
purposes of tFAW.
• 8-bank per channel, precharge all banks (AB) allowance: tRP for a PRECHARGE ALL
BANKS command for an 8-bank device must equal tRPab, which is greater than
tRPpb.
70
200b: x32 LPDDR4 SDRAM
Read and Write Access Modes
Figure 14: ACTIVATE Command
T0
T1
T2
T3
Ta0
RA
RA
BA0
RA
RA
RA
Ta1
Ta2
Ta3
RA
RA
Tb0
Tb1
Tb2
Tb3
Valid
BA0
CA
CA
Tc0
Tc1
Td0
Td1
Td2
Td3
RA
RA
Td4
Td5
CK_c
CK_t
CKE
CS
CA
RA
BA1
RA
BA0
Valid
RA
BA0
tRP
tRRD
tRCD
Command
ACTIVATE-1
ACTIVATE-2
DES
ACTIVATE-1
ACTIVATE-2
CAS2
READ1
DES
DES
PRECHARGE
per bank
DES
DES
ACTIVATE-2
ACTIVATE-1
DES
tRAS
tRC
Don’t Care
1. A PRECHARGE command uses tRPab timing for all-bank precharge and tRPpb timing for
single-bank precharge. In this figure, tRP is used to denote either all-bank precharge or
a single-bank precharge. tCCD = MIN, 1.5nCK postamble, 533 MHz < clock frequency ≤
800 MHz, ODT worst timing case.
Note:
Figure 15: tFAW Timing
T0
T1
T2
T3
Ta0
RA
RA
BA0
RA
RA
RA
Ta1
Ta2
Ta3
RA
RA
Tb0
Tb1
Tb2
Tb3
RA
BA2
RA
RA
Tc0
Tc1
Tc2
Tc3
RA
RA
Tc4
Td0
Td1
Td2
Td3
Td4
RA
RA
BA4
RA
RA
CK_c
CK_t
CKE
CS
CA
Command
ACTIVATE-1
ACTIVATE-2
DES
RA
BA1
ACTIVATE-1
ACTIVATE-2
RA
DES
tRRD
ACTIVATE-1
ACTIVATE-2
RA
DES
RA
BA3
ACTIVATE-1
ACTIVATE-2
DES
DES
ACTIVATE-1
ACTIVATE-2
tRRD
tRRD
t FAW
Don’t Care
Note:
1. REFpb may be substituted for one of the ACTIVATE commands for the purposes of tFAW.
Read and Write Access Modes
After a bank has been activated, a READ or WRITE command can be executed. This is
accomplished by asserting CKE asynchronously, with CS and CA[5:0] set to the proper
state (see Command Truth Table) on the rising edge of CK.
The device provides a fast column access operation. A single READ or WRITE command
will initiate a burst READ or WRITE operation, where data is transferred to/from the device on successive clock cycles. Burst interrupts are not allowed; however, the optimal
burst length may be set on-the-fly (see Command Truth Table).
71
200b: x32 LPDDR4 SDRAM
Preamble and Postamble
Preamble and Postamble
The DQS strobe for the device requires a preamble prior to the first latching edge (the
rising edge of DQS_t with data valid), and it requires a postamble after the last latching
edge. The preamble and postamble options are set via MODE REGISTER WRITE commands.
The read preamble is two tCK in length and is either static or has one clock toggle before
the first latching edge. The read preamble option is enabled via MRW to MR1 OP[3] (0 =
Static; 1 = Toggle).
The read postamble has a programmable option to extend the postamble by 1nCK
(tRPSTE). The extended postamble option is enabled via MRW to MR1 OP[7] (0 =
0.5nCK; 1 = 1.5nCK).
Figure 16: DQS Read Preamble and Postamble – Toggling Preamble and 0.5nCK Postamble
T0
T1
T2
T3
T4
Ta0
Ta1
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Tb6
Tc0
Tc1
Tc2
Tc3
Tc4
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
Command
RD-1
CAS-2
RL
tDQSCK
tRPRE
DQS_c
DQS_t
tDQSQ
DQ
DMI
tRPST
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0
n1
n2
n3
n4
n5
n10 n11 n12 n13 n14 n15
Notes:
1. BL = 16, Preamble = Toggling, Postamble = 0.5nCK.
2. DQS and DQ terminated VSSQ.
3. DQS_t/DQS_c is "Don’t Care" prior to the start of tRPRE. No transition of DQS is implied,
as DQS_t/DQS_c can be HIGH, LOW, or High-Z prior to tRPRE.
72
200b: x32 LPDDR4 SDRAM
Preamble and Postamble
Figure 17: DQS Read Preamble and Postamble – Static Preamble and 1.5nCK Postamble
T0
T1
T2
T3
T4
Ta0
Ta1
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Tb6
Tc0
Tc1
Tc2
Tc3
Tc4
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
Command
CAS-2
RD-1
tDQSCK
RL
tRPRE
DQS_c
DQS_t
tRPSTE
tDQSQ
DQ
DMI
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0
n1
n2
n3
n4
n5
n10 n11 n12 n13 n14 n15
Notes:
1. BL = 16, Preamble = Static, Postamble = 1.5nCK (extended).
2. DQS and DQ terminated VSSQ.
3. DQS_t/DQS_c is "Don’t Care" prior to the start of tRPRE. No transition of DQS is implied,
as DQS_t/DQS_c can be HIGH, LOW, or High-Z prior to tRPRE.
Figure 18: DQS Write Preamble and Postamble – 0.5nCK Postamble
T0
T1
T2
T3
Valid
Valid
Valid
Valid
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Tb6
Tb7
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CKE
CS
CA
Command
WRITE-1
CAS-2
WL
tDQSS
t
WPRE
t WPST
DQS_c
DQS_t
BL/2
tDQS2DQ
DQ
DMI
DIN
n0
DIN
n1
DIN
n2
DIN
n3
DIN
n8
DIN
n9
DIN
n10
DIN
n11
DIN
n12
DIN
n13
DIN
n14
DIN
n15
Don’t Care
Notes:
1. BL = 16, Postamble = 0.5nCK.
2. DQS and DQ terminated VSSQ.
73
200b: x32 LPDDR4 SDRAM
Preamble and Postamble
3. DQS_t/DQS_c is "Don’t Care" prior to the start of tWPRE. No transition of DQS is implied,
as DQS_t/DQS_c can be HIGH, LOW, or High-Z prior to tWPRE.
Figure 19: DQS Write Preamble and Postamble – 1.5nCK Postamble
T0
T1
T2
T3
Valid
Valid
Valid
Valid
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Tb6
Tb7
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CKE
CS
CA
Command
WRITE-1
CAS-2
WL
tDQSS
t
t WPST
WPRE
DQS_c
DQS_t
BL/2
tDQS2DQ
DQ
DMI
DIN
n0
DIN
n1
DIN
n2
DIN
n3
DIN
n8
DIN
n9
DIN
n10
DIN
n11
DIN
n12
DIN
n13
DIN
n14
DIN
n15
Don’t Care
Notes:
1. BL = 16, Postamble = 1.5nCK.
2. DQS and DQ terminated VSSQ.
3. DQS_t/DQS_c is "Don’t Care" prior to the start of tWPRE. No transition of DQS is implied,
as DQS_t/DQS_c can be HIGH, LOW, or High-Z prior to tWPRE.
74
200b: x32 LPDDR4 SDRAM
Burst READ Operation
Burst READ Operation
A burst READ command is initiated with CKE, CS, and CA[5:0] asserted to the proper
state on the rising edge of CK, as defined by the Command Truth Table. The command
address bus inputs determine the starting column address for the burst. The two loworder address bits are not transmitted on the CA bus and are implied to be 0; therefore,
the starting burst address is always a multiple of four (that is, 0x0, 0x4, 0x8, 0xC).
The READ latency (RL) is defined from the last rising edge of the clock that completes a
READ command (for example, the second rising edge of the CAS-2 command) to the
rising edge of the clock from which the tDQSCK delay is measured. The first valid data is
available RL × tCK + tDQSCK + tDQSQ after the rising edge of clock that completes a
READ command.
The data strobe output is driven tRPRE before the first valid rising strobe edge. The first
data bit of the burst is synchronized with the first valid (post-preamble) rising edge of
the data strobe. Each subsequent data-out appears on each DQ pin, edge-aligned with
the data strobe. At the end of a burst, the DQS signals are driven for another half cycle
postamble, or for a 1.5-cycle postamble if the programmable postamble bit is set in the
mode register. The RL is programmed in the mode registers. Pin timings for the data
strobe are measured relative to the cross-point of DQS_t and DQS_c.
Figure 20: Burst Read Timing
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
T5
T6
T7
T15
T16
T17
T18
T19
BL
BA0,
CA, AP
CAm
CAm
T20
T21
T22
T23
DES
DES
DES
T33
T34
T35
T36
T41
T42
T43
T44
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
DES
DES
DES
t
DES
DES
READ-1
CAS-2
DES
CCD = 16
RL = 14
RL = 14
t DQSCK
t DQSCK
BL/2 = 8
BL/2 = 16
t
t RPST
RPRE
DQS_c
DQS_t
tDQSQ
tDQSQ
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n7 n26 n27 n28 n29 n30 n31 m0 m1 m10 m11 m12 m13 m14 m15
n1 n2
n6
n0
n3
n4
n5
DQ
DMI
Don’t Care
Notes:
1. BL = 32 for column n, BL = 16 for column m, RL = 14, Preamble = Toggle, Postamble =
0.5nCK, DQ/DQS: VSSQ termination.
2. DOUT n/m = data-out from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
75
200b: x32 LPDDR4 SDRAM
Burst READ Operation
Figure 21: Burst Read Followed by Burst Write or Burst Mask Write
T0
T1
T2
T3
BL
BA0,
CA, AP
CA
CA
T4
T5
Ta0
Ta1
Ta2
BL
BA0,
CA, AP
Ta3
Ta4
CA
CA
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
DES
DES
DES
DES
DES
DES
Tb6
Tb7
Tc0
Tc1
DES
DES
DES
DES
Tc2
Tc3
Tc4
Tc5
Tc6
Tc7
DES
DES
CK_c
CK_t
CS
CA
READ-1
Command
CAS-2
DES
WR-1/MWR-1
DES
DES
CAS-2
RL + RU( tDQSCK(MAX)/ tCK) + BL/2
+ RD( tRPST) - WL + tWPRE
WL
t DQSCK
RL
t
DES
DES
DES
DES
t DQSS
BL/2 = 8
t
RPRE
WPRE
DQS_c
DQS_t
tDQSQ
DQ
DMI
tDQS2DQ
tRPST
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0
n9 n10 n11 n12 n13 n14 n15
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n9 n10 n11 n12 n13 n14 n15
n0
Don’t Care
1. BL=16, Read preamble = Toggle, Read postamble = 0.5nCK, Write preamble = 2nCK,
Write postamble = 0.5nCK, DQ/DQS: VSSQ termination.
2. DOUT n = data-out from column n and DIN n = data-in to column n.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Notes:
Figure 22: Seamless Burst Read
T0
T1
T2
T3
Ta0
Ta1
BL
BA0,
CA, AP
CAn
CAn
BL
BA0,
CA, AP
Ta2
Ta3
CAm
CAm
Tb0
Tb1
Tb2
Tb3
Tb4
Tc0
Tc1
DES
DES
DES
Tc2
Tc3
Td0
Td1
Td2
Td3
Te0
Te1
DES
DES
DES
DES
DES
Te2
Te3
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
DES
CAS-2
READ-1
BL
DES
BA1,
CA, AP
CAn
READ-1
CAn
CAS-2
DES
DES
DES
DES
t DQSCK
RL
t DQSCK
RL
RL
DES
t DQSCK
t
RPRE
DQS_c
DQS_t
tDQSQ
DQ
DMI
tDQSQ
tDQSQ
tRPST
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n15
n10 n11 n12 n13 n14
n0
n1 n10 n11 n12 n13 n14 n15 m0 m1 m10 m11 m12 m13 m14 m15 n0 n1
Bank 0
Bank 1
Don’t Care
Notes:
1. BL = 16, tCCD = 8, Preamble = Toggle, Postamble = 0.5nCK, DQ/DQS: VSSQ termination.
2. DOUT n/m = data-out from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
76
200b: x32 LPDDR4 SDRAM
Burst READ Operation
Read Timing
Figure 23: Read Timing
T0
T1
T2
T3
T4
Ta0
Ta1
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Tb6
Tb7
Tc0
Tc1
Tc2
Tc3
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
Command
RD-1
CAS-2
t
HZ(DQS)
t
RL
t
DQSCK
LZ(DQS)
t
RPRE
DQS_c
DQS_t
t
DQSQ
t
RPST
t
HZ(DQ)
t
LZ(DQ)
DQ
DMI
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0
n1
n2
n3
n4
n5
n10 n11 n12 n13 n14 n15
Notes:
1.
2.
3.
4.
BL = 16, Preamble = Toggling, Postamble = 0.5nCK.
DQS, DQ, and DMI terminated VSSQ.
Output driver does not turn on before an endpoint of tLZ(DQS) and tLZ(DQ).
Output driver does not turn off before an endpoint of tHZ(DQS) and tHZ(DQ).
tLZ(DQS), tLZ(DQ), tHZ(DQS), tHZ(DQ)
Calculation
tHZ
and tLZ transitions occur in the same time window as valid data transitions. These
parameters are referenced to a specific voltage level that specifies when the device output is no longer driving tHZ(DQS) and tHZ(DQ), or begins driving tLZ(DQS) and
tLZ(DQ). This section shows a method to calculate the point when the device is no longer driving tHZ(DQS) and tHZ(DQ), or begins driving tLZ(DQS) and tLZ(DQ), by measuring the signal at two different voltages. The actual voltage measurement points are
not critical as long as the calculation is consistent. The parameters tLZ(DQS), tLZ(DQ),
tHZ(DQS), and tHZ(DQ) are defined as single ended.
77
200b: x32 LPDDR4 SDRAM
Burst READ Operation
tLZ(DQS)
and tHZ(DQS) Calculation for ATE (Automatic Test Equipment)
Figure 24: tLZ(DQS) Method for Calculating Transitions and Endpoint
CK_t – CK_c crossing at the second CAS-2 of READ command
CK_t
CK_c
tLZ(DQS)
DQS_c
VOH
0.5 x VOH
VSW2
VSW1
End point: Extrapolated point
0V
Notes:
1. Conditions for calibration: Pull down driver RON = 40 ohm, VOH = VDDQ/3.
2. Termination condition for DQS_t and DQS_C = 50 ohm to VSSQ.
3. The VOH level depends on MR22 OP[2:0] and MR3 OP[0] settings as well as device tolerances. Use the actual VOH value for tHZ and tLZ measurements.
Figure 25: tHZ(DQS) Method for Calculating Transitions and Endpoint
CK_t – CK_c crossing at the second CAS-2 of READ command
CK_t
CK_c
tHZ(DQS)
End point: Extrapolated point
VOH
VSW2
0.5 x VOH
VSW1
DQS_c
0V
Notes:
1. Conditions for calibration: Pull down driver RON = 40 ohm, VOH = VDDQ/3.
2. Termination condition for DQS_t and DQS_C = 50 ohm to VSSQ.
78
200b: x32 LPDDR4 SDRAM
Burst READ Operation
3. The VOH level depends on MR22 OP[2:0] and MR3 OP[0] settings as well as device tolerances. Use the actual VOH value for tHZ and tLZ measurements.
Table 85: Reference Voltage for tLZ(DQS), tHZ(DQS) Timing Measurements
Measured Parameter
Symbol
Vsw1
Vsw2
Unit
DQS_c Low-Z time
from CK_t, CK_c
tLZ(DQS)
0.4 × VOH
0.6 × VOH
V
DQS_c High-Z time
from CK_t, CK_c
tHZ(DQS)
0.4 × VOH
0.6 × VOH
Measured Parameter
tLZ(DQ)
and tHZ(DQ) Calculation for ATE (Automatic Test Equipment)
Figure 26: tLZ(DQ) Method for Calculating Transitions and Endpoint
CK_t – CK_c crossing at the second CAS-2 of READ command
CK_t
CK_c
t LZ(DQ)
DQs
VOH
0.5 x VOH
VSW2
VSW1
End point: Extrapolated point
0V
Notes:
1. Conditions for calibration: Pull down driver RON = 40 ohm, VOH = VDDQ/3.
2. Termination condition for DQ and DMI = 50 ohm to VSSQ.
3. The VOH level depends on MR22 OP[2:0] and MR3 OP[0] settings as well as device tolerances. Use the actual VOH value for tHZ and tLZ measurements.
79
200b: x32 LPDDR4 SDRAM
Burst READ Operation
Figure 27: tHZ(DQ) Method for Calculating Transitions and Endpoint
CK_t – CK_c crossing at the second CAS-2 of READ command
CK_t
CK_c
tHZ(DQ)
End point: Extrapolated point
VOH
VSW2
0.5 x VOH
VSW1
DQs
0V
Notes:
1. Conditions for calibration: Pull down driver RON = 40 ohm, VOH = VDDQ/3.
2. Termination condition for DQ and DMI = 50 ohm to VSSQ.
3. The VOH level depends on MR22 OP[2:0] and MR3 OP[0] settings as well as device tolerances. Use the actual VOH value for tHZ and tLZ measurements.
Table 86: Reference Voltage for tLZ(DQ), tHZ(DQ) Timing Measurements
Measured Parameter
Symbol
Vsw1
Vsw2
Unit
DQ Low-Z time
from CK_t, CK_c
tLZ(DQ)
0.4 × VOH
0.6 × VOH
V
DQ High-Z time
from CK_t, CK_c
tHZ(DQ)
0.4 × VOH
0.6 × VOH
Measured Parameter
80
200b: x32 LPDDR4 SDRAM
Burst WRITE Operation
Burst WRITE Operation
A burst WRITE command is initiated with CKE, CS, and CA[5:0] asserted to the proper
state at the rising edge of CK, as defined by the Command Truth Table. Column addresses C[3:2] should be driven LOW for burst WRITE commands, and column addresses
C[1:0] are not transmitted on the CA bus and are assumed to be zero so that the starting
column burst address is always aligned with a 32-byte boundary. The WRITE latency
(WL) is defined from the last rising edge of the clock that completes a WRITE command
(for example, the second rising edge of the CAS-2 command) to the rising edge of the
clock from which tDQSS is measured. The first valid latching edge of DQS must be driven WL × t CK + tDQSS after the rising edge of clock that completes a WRITE command.
The device uses an unmatched DQS DQ path for lower power, so the DQS strobe must
arrive at the SDRAM ball prior to the DQ signal by tDQS2DQ. The DQS strobe output
must be driven tWPRE before the first valid rising strobe edge. The tWPRE preamble is
required to be 2 × tCK at any speed ranges. The DQS strobe must be trained to arrive at
the DQ pad latch center-aligned with the DQ data. The DQ data must be held for
TdiVW, and the DQS must be periodically trained to stay roughly centered in the TdiVW.
Burst data is captured by the SDRAM on successive edges of DQS until the 16- or 32-bit
data burst is complete. The DQS strobe must remain active (toggling) for tWPST (write
postamble) after the completion of the burst WRITE. After a burst WRITE operation,
tWR must be satisfied before a PRECHARGE command to the same bank can be issued.
Signal input timings are measured relative to the cross point of DQS_t and DQS_c.
81
200b: x32 LPDDR4 SDRAM
Burst WRITE Operation
Figure 28: Burst WRITE Operation
T0
T1
T2
T3
BL
BA0,
CA, AP
CA
CA
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Ta6
Tb0
Tb1
Tb2
Tc0
Tc1
Tc2
Valid
BA0
Tc3
Tc4
Td0
Td1
Td2
Td3
Tb4
Td5
RA
BA0,
RA
RA
RA
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
DES
DES
DES
DES
DES
WL
DES
DES
DES
DES
DES
DES
DES
tWPRE
PRECHARGE
tWR
BL/2 + 1 Clock
tDQSS
DES
DES
DES
DES
DES
ACT-1
ACT-2
tRP
(MIN)
tDSH
DQS_c
tDSS
tWPST
DQS_t
tDQS2DQ
DIN DIN DIN DIN DIN DIN DIN DIN DIN DIN
n0 n1 n2 n3 n4 n5 n12 n13 n14 n15
DQ
tDQSS
(Nominal)
tWPRE
DQS_c
DQS_t
tDQS2DQ
DIN DIN DIN DIN DIN DIN DIN DIN DIN DIN
n0 n1 n2 n3 n4 n11 n12 n13 n14 n15
tDQSS (MAX)
DQ
tWPRE
DQS_c
DQS_t
tDQS2DQ
DIN DIN DIN DIN DIN DIN DIN DIN DIN DIN
n0 n1 n2 n3 n4 n11 n12 n13 n14 n15
DQ
Don’t Care
Notes:
1.
2.
3.
4.
BL = 16, Write postamble = 0.5nCK, DQ/DQS: VSSQ termination.
DIN n = data-in to column n.
tWR starts at the rising edge of CK after the last latching edge of DQS.
DES commands are shown for ease of illustration; other commands may be valid at
these times.
82
200b: x32 LPDDR4 SDRAM
Burst WRITE Operation
Figure 29: Burst Write Followed by Burst Read
T0
T1
T2
T3
BL
BA0,
CA, AP
CA
CA
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Ta6
Tb0
Tb1
Tb2
Tc0
Tc1
Tc2
Tc3
Tc4
Tc5
Tc6
BL
BA0,
CA, AP
CA
CA
Tc7
Tc8
Tc9
Tc10
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
DES
DES
DES
DES
DES
WL
DES
DES
DES
DES
DES
DES
DES
tWPRE
DES
DES
READ-1
tWTR
BL/2 + 1 Clock
tDQSS
DES
CAS-2
RL
(MIN)
tDSH
tDSS
tWPST
DQS_c
DQS_t
tDQS2DQ
DIN DIN DIN DIN DIN DIN DIN DIN DIN DIN
n0 n1 n2 n3 n4 n5 n12 n13 n14 n15
DQ
Don’t Care
Notes:
1. BL = 16, Write postamble = 0.5nCK, DQ/DQS: VSSQ termination.
2. DIN n = data-in to column n.
3. The minimum number of clock cycles from the burst WRITE command to the burst READ
command for any bank is [WL + 1 + BL/2 + RU(tWTR/tCK)].
t
4. WTR starts at the rising edge of CK after the last latching edge of DQS.
5. DES commands are shown for ease of illustration; other commands may be valid at
these times.
83
200b: x32 LPDDR4 SDRAM
Burst WRITE Operation
Write Timing
Figure 30: Write Timing
T0
T1
T2
T3
BL
BA0,
CA, AP
CA
CA
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Ta6
DES
DES
DES
DES
DES
DES
DES
DES
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
CAS-2
WRITE-1
DES
WL
tDQSS
tDQSS
tWPRE
(MIN)
(MIN)
tDSH
tDSS
tWPST
DQS_c
DQS_t
tDQS2DQ
DIN DIN DIN DIN DIN DIN DIN DIN DIN DIN
n0 n1 n2 n3 n4 n5 n12 n13 n14 n15
DQ
tDQSS
tDQSS
(Nominal)
(Nominal)
tDSH
tWPRE
tDSS
DQS_c
DQS_t
tDQS2DQ
DIN DIN DIN DIN DIN DIN DIN DIN DIN DIN
n0 n1 n2 n3 n4 n11 n12 n13 n14 n15
DQ
tDQSS
tDQSS
(MAX)
(MAX)
tWPRE
tDSH
tDSS
DQS_c
DQS_t
tDQS2DQ
tDQSH
tDQSL
DIN DIN DIN DIN DIN DIN DIN DIN DIN DIN
n0 n1 n2 n3 n4 n11 n12 n13 n14 n15
DQ
Don’t Care
Notes:
1. BL = 16, Write postamble = 0.5nCK.
2. DIN n = data-in to column n.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
84
200b: x32 LPDDR4 SDRAM
Burst WRITE Operation
tWPRE
Calculation for ATE (Automatic Test Equipment)
Figure 31: Method for Calculating tWPRE Transitions and Endpoints
CK_t
Vref(CA)
CK_c
Resulting differential signal
relevant for tWPRE specification
Vsw2
Vsw1
DQS_t - DQS_c
0V
Begin point:
Extrapolated point
tWPRE
Note:
1. Termination condition for DQS_t, DQS_c, DQ, and DMI = 50 ohm to VSSQ.
Table 87: Method for Calculating tWPRE Transitions and Endpoints
Measured Parameter
Measured Parameter
Symbol
Vsw1
Vsw2
Unit
tWPRE
VIHL_AC × 0.3
VIHL_AC × 0.7
V
DQS_t, DQS_c
differential write preamble
tWPST
Calculation for ATE (Automatic Test Equipment)
Figure 32: Method for Calculating tWPST Transitions and Endpoints
CK_t
Vref(CA)
CK_c
Resulting differential signal
relevant for tWPST specification
0V
Vsw2
Vsw1
DQS_t - DQS_c
tWPST
Notes:
End point:
Extrapolated point
1. Termination condition for DQS_t, DQS_c, DQ, and DMI = 50 ohm to VSSQ.
2. Write postamble: 0.5tCK
3. The method for calculating differential pulse widths for 1.5tCK postamble is same as
0.5tCK postamble.
85
200b: x32 LPDDR4 SDRAM
MASK WRITE Operation
Table 88: Reference Voltage for tWPST Timing Measurements
Measured Parameter
Measured Parameter
Symbol
Vsw1
Vsw2
Unit
tWPST
–(VIHL_AC × 0.7)
–(VIHL_AC × 0.3)
V
DQS_t, DQS_c
differential write postamble
MASK WRITE Operation
The device requires that WRITE operations that include a byte mask anywhere in the
burst sequence must use the MASK WRITE command. This allows the device to implement efficient data protection schemes based on larger data blocks. The MASK
WRITE-1 command is used to begin the operation, followed by a CAS-2 command. A
MASKED WRITE command to the same bank cannot be issued until tCCDMW later, to
allow the device to finish the internal READ-MODIFY-WRITE operation. One datamask-invert (DMI) pin is provided per byte lane, and the data-mask-invert timings
match data bit (DQ) timing. See Data Mask Invert for more information on the use of
the DMI signal.
Figure 33: MASK WRITE Command – Same Bank
T0
T1
T2
T3
BL
BA0,
CA, AP
CA
CA
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Ta6
Tb0
Tb1
Tb2
Tc0
Tc1
Tc2
Tc3
Tc4
BL
BA0,
CA, AP
Tc5
Tc6
CA
CA
Tc7
Tc8
Tc9
Tc10
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
MASK WRITE-1
CAS-2
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
tCCDMW
WL
DES
DES
DES
MASK WRITE-1
CAS-2
WL
tDQSS(MIN)
t
tWPRE
WPST
DQS_c
DQS_t
tDQS2DQ
DIN
n0
DQ
DMI
DIN DIN
n1 n2
DIN DIN
n3 n4
DIN DIN DIN DIN
n5 n12 n13 n14
DIN
n15
Don’t Care
Notes:
1. BL = 16, Write postamble = 0.5nCK, DQ/DQS: VSSQ termination.
2. DIN n = data-in to column n.
3. Mask-write supports only BL16 operations. For BL32 configuration, the system needs to
insert only 16-bit wide data for MASKED WRITE operation.
4. DES commands are shown for ease of illustration; other commands may be valid at
these time.
86
200b: x32 LPDDR4 SDRAM
MASK WRITE Operation
Figure 34: MASK WRITE Command – Different Bank
T0
T1
T2
T3
T8
T9
BL
BA0,
CA, AP
CA
CA
BL
BA0,
CA, AP
T10
T11
CA
CA
T16
T17
T18
T19
T24
T25
T26
T27
T32
T33
T34
T35
BL
BA0,
CA, AP
CA
CA
BL
BA0,
CA, AP
CA
CA
BL
BA0,
CA, AP
CA
CA
T36
T37
T38
DES
DES
DES
CK_c
CK_t
CS
CA
Command MASK WRITE-1
CAS-2
MASK WRITE-1
DES
CAS-2
DES
MASK WRITE-1
tCCD
DES
CAS-2
MASK WRITE-1
tCCD
DES
CAS-2
tCCD
MASK WRITE-1
CAS-2
tCCD
tCCDMW
WL
tDQSS
tWPRE
DQS_c
DQS_t
tDQS2DQ
DQ
DMI
DIN
n0
DIN
n1
DIN
n2
DIN DIN
n3 n10
DIN DIN DIN DIN DIN
n11 n12 n13 n14 n15
DIN
n0
DIN DIN
n1 n2
DIN DIN
n3 n10
DIN DIN
n11 n12
DIN DIN DIN
n13 n14 n15
DIN
n0
DIN
n1
DIN DIN
n2 n10
DIN DIN
n11 n12
DIN DIN DIN
n13 n14 n15
DIN
n0
DIN
n1
DIN
n2
DIN DIN
n3 n4
DIN
n5
DIN
n6
DIN DIN
n8
n7
Don’t Care
Notes:
1. BL = 16, DQ/DQS/DMI: VSSQ termination.
2. DIN n = data-in to column n.
3. Mask-write supports only BL16 operations. For BL32 configuration, the system needs to
insert only 16-bit wide data for MASKED WRITE operation.
4. DES commands are shown for ease of illustration; other commands may be valid at
these time.
87
200b: x32 LPDDR4 SDRAM
MASK WRITE Operation
Mask Write Timing Constraints for BL16
Table 89: Same Bank (ODT Disabled)
Next CMD
Current CMD
ACTIVE
READ
(BL = 16 or 32)
WRITE
(BL = 16 or 32)
MASK WRITE
PRECHARGE
Illegal
RU(tRCD/tCK)
RU(tRCD/tCK)
RU(tRCD/tCK)
RU(tRAS/tCK)
READ
(with BL = 16)
Illegal
81
RL +
RL +
RU(tDQSCK(MAX)/ RU(tDQSCK(MAX)/
tCK) + BL/2 - WL +
tCK) + BL/2 - WL +
tWPRE + RD( tRPST) tWPRE + RD( tRPST)
BL/2 +
MAX{(8,RU(tRTP/
tCK)} - 8
READ
(with BL = 32)
Illegal
162
RL +
RL +
RU(tDQSCK(MAX)/ RU(tDQSCK(MAX)/
tCK) + BL/2 - WL +
tCK) + BL/2 - WL +
tWPRE + RD( tRPST) tWPRE + RD( tRPST)
BL/2 +
MAX{(8,RU(tRTP/
tCK)} - 8
WRITE
(with BL = 16)
Illegal
WL + 1+ BL/2 +
RU(tWTR/tCK)
81
WRITE
(with BL = 32)
Illegal
WL + 1 + BL/2 +
RU(tWTR/tCK)
162
MASK WRITE
Illegal
WL + 1 + BL/2 +
RU(tWTR/tCK)
tCCD
tCCDMW3
WL + 1 + BL/2 +
RU(tWR/tCK)
PRECHARGE
RU(tRP/tCK),
RU(tRPab/tCK)
Illegal
Illegal
Illegal
4
ACTIVE
Notes:
1.
2.
3.
4.
tCCDMW3
tCCDMW
+ 84
WL + 1 + BL/2 +
RU(tWR/tCK)
WL + 1 + BL/2 +
RU(tWR/tCK)
In the case of BL = 16, tCCD is 8 × tCK.
In the case of BL = 32, tCCD is 16 × tCK.
tCCDMW = 32 × tCK (4 × tCCD at BL = 16).
WRITE with BL = 32 operation is 8 × tCK longer than BL = 16.
Table 90: Different Bank (ODT Disabled)
Next CMD
Current CMD
ACTIVE
READ
(with BL = 16)
ACTIVE
READ
(BL = 16 or 32)
WRITE
(BL = 16 or 32)
MASK WRITE
PRECHARGE
RU(tRRD/tCK)
4
4
4
22
4
81
RL +
RL +
22
RU(tDQSCK(MAX)/
RU(tDQSCK(MAX)/
tCK)
+ BL/2 - WL + tCK) + BL/2 - WL +
+ RD( tRPST) tWPRE + RD( tRPST)
tWPRE
READ
(with BL = 32)
4
162
RL +
RL +
RU(tDQSCK(MAX)/
RU(tDQSCK(MAX)/
tCK)
22
+ BL/2 - WL + tCK) + BL/2 - WL +
+ RD( tRPST) tWPRE + RD( tRPST)
tWPRE
WRITE
(with BL = 16)
4
WL + 1+ BL/2 +
RU(tWTR/tCK)
81
81
22
WRITE
(with BL = 32)
4
WL + 1 + BL/2 +
RU(tWTR/tCK)
162
162
22
MASK WRITE
4
WL + 1 + BL/2 +
RU(tWTR/tCK)
81
81
22
88
200b: x32 LPDDR4 SDRAM
MASK WRITE Operation
Table 90: Different Bank (ODT Disabled) (Continued)
Next CMD
Current CMD
ACTIVE
READ
(BL = 16 or 32)
WRITE
(BL = 16 or 32)
MASK WRITE
PRECHARGE
4
4
4
4
4
PRECHARGE
Notes:
1. In the case of BL = 16, tCCD is 8 × tCK
2. In the case of BL = 32, tCCD is 16 × tCK
Table 91: Same Bank (ODT Enabled)
Next CMD
Current CMD
ACTIVE
READ
(BL = 16 or 32)
WRITE
(BL = 16 or 32)
MASK WRITE
PRECHARGE
ACTIVE
Illegal
RU(tRCD/tCK)
RU(tRCD/tCK)
RU(tRCD/tCK)
RU(tRAS/tCK)
READ
(with BL = 16)
Illegal
81
RL + RU(
RL + RU(
BL/2 +
MAX{(8,RU(tRTP/
tCK)} - 8
tDQSCK(MAX)/ tCK) tDQSCK(MAX)/ tCK)
+ BL/2 + RD( tRPST) + BL/2 + RD( tRPST)
- ODTLon - RD(
- ODTLon - RD(
tODTon(MIN)/ tCK) tODTon(MIN)/ tCK)
READ
(with BL = 32)
162
Illegal
RL + RU(
RL + RU(
tDQSCK(MAX)/ tCK) tDQSCK(MAX)/ tCK)
+ BL/2 + RD( tRPST) + BL/2 + RD( tRPST)
- ODTLon - RD(
- ODTLon - RD(
tODTon(MIN)/ tCK) tODTon(MIN)/ tCK)
WRITE
(with BL = 16)
Illegal
WL + 1+ BL/2 +
RU(tWTR/tCK)
81
WRITE
(with BL = 32)
Illegal
WL + 1 + BL/2 +
RU(tWTR/tCK)
162
MASK WRITE
Illegal
WL + 1 + BL/2 +
RU(tWTR/tCK)
tCCD
tCCDMW3
WL + 1 + BL/2 +
RU(tWR/tCK)
PRECHARGE
RU(tRP/tCK),
RU(tRPab/tCK)
Illegal
Illegal
Illegal
4
Notes:
1.
2.
3.
4.
tCCDMW3
BL/2 +
MAX{(8,RU(tRTP/
tCK)} - 8
tCCDMW
+ 84
WL + 1 + BL/2 +
RU(tWR/tCK)
WL + 1 + BL/2 +
RU(tWR/tCK)
In the case of BL = 16, tCCD is 8 × tCK.
In the case of BL = 32, tCCD is 16 × tCK.
tCCDMW = 32 × tCK (4 × tCCD at BL = 16).
WRITE with BL = 32 operation is 8 × tCK longer than BL = 16.
Table 92: Different Bank (ODT Enabled)
Next CMD
Current CMD
ACTIVE
READ
(with BL = 16)
ACTIVE
READ
(BL = 16 or 32)
WRITE
(BL = 16 or 32)
MASK WRITE
PRECHARGE
RU(tRRD/tCK)
4
4
4
22
4
81
RL + RU(
RL + RU(
22
tDQSCK(MAX)/ tCK) tDQSCK(MAX)/ tCK)
+ BL/2 + RD( tRPST) + BL/2 + RD( tRPST)
- ODTLon - RD(
- ODTLon - RD(
tODTon(MIN)/ tCK) tODTon(MIN)/ tCK)
89
200b: x32 LPDDR4 SDRAM
Data Mask and Data Bus Inversion (DBI [DC]) Function
Table 92: Different Bank (ODT Enabled) (Continued)
Next CMD
Current CMD
ACTIVE
READ
(BL = 16 or 32)
READ
(with BL = 32)
4
162
WRITE
(with BL = 16)
4
WL + 1+ BL/2 +
RU(tWTR/tCK)
81
81
22
WRITE
(with BL = 32)
4
WL + 1 + BL/2 +
RU(tWTR/tCK)
162
162
22
MASK WRITE
4
WL + 1 + BL/2 +
RU(tWTR/tCK)
81
81
22
PRECHARGE
4
4
4
4
4
Notes:
WRITE
(BL = 16 or 32)
MASK WRITE
RL + RU(
RL + RU(
tDQSCK(MAX)/ tCK) tDQSCK(MAX)/ tCK)
+ BL/2 + RD( tRPST) + BL/2 + RD( tRPST)
- ODTLon - RD(
- ODTLon - RD(
tODTon(MIN)/ tCK) tODTon(MIN)/ tCK)
PRECHARGE
22
1. In the case of BL = 16, tCCD is 8 × tCK.
2. In the case of BL = 32, tCCD is 16 × tCK.
Data Mask and Data Bus Inversion (DBI [DC]) Function
Data mask (DM) is supported for WRITE operations and the data bus inversion DBI
(DC) is supported for READ, WRITE, MASK WRITE, MRR, and MRW operations. DM
and DBI (DC) functions are supported with byte granularity. DBI (DC) for READ operations (READ, MRR) can be enabled or disabled via MR3 OP[6]. DBI (DC) for WRITE operations (WRITE, MASK WRITE, MRW) can be enabled or disabled via MR3 OP[7]. DM
for MASK WRITE operations can be enabled or disabled via MR13 OP[5]. The device has
one data mask inversion (DMI) pin per byte and a total of two DMI pins per channel.
The DMI signal is a bidirectional DDR signal, is sampled with the DQ signals, and is
electrically identical to a DQ signal.
There are eight possible states for the device with the DM and DBI (DC) functions.
Table 93: Function Behavior of DMI Signal During WRITE, MASKED WRITE, and READ Operations
DMI Signal
DM
Function
Write DBI
(DC)
Read DBI
(DC)
During
WRITE
During
MASKED
WRITE
During
READ
During
During
During
MPC[WRIT MPC[READ- MPC[READ
E-FIFO]
FIFO]
DQ CAL]
Disabled
Disabled
Disabled
Don't Care1
Illegal1, 3
High-Z2
Don't Care1
High-Z2
High-Z2
Illegal3
High-Z2
Train9
Train10
Train11
Disabled
Enabled
Disabled
Disabled
Disabled
Enabled
Don't Care1
Illegal3
DBI (DC)5
Train9
Train10
Train11
Disabled
Enabled
Enabled
DBI (DC)4
Illegal3
DBI (DC)5
Train9
Train10
Train11
Enabled
Disabled
Disabled
Don't Care6
DM7
High-Z2
Train9
Train10
Train11
High-Z2
Train9
Train10
Train11
DBI (DC)5
Train9
Train10
Train11
Enabled
Enabled
Disabled
Enabled
Disabled
Enabled
DBI
(DC)4
DBI
(DC)4
Don't Care6
DBI
(DC)8
DM7
90
200b: x32 LPDDR4 SDRAM
Data Mask and Data Bus Inversion (DBI [DC]) Function
Table 93: Function Behavior of DMI Signal During WRITE, MASKED WRITE, and READ Operations
(Continued)
DMI Signal
DM
Function
Write DBI
(DC)
Read DBI
(DC)
During
WRITE
During
MASKED
WRITE
Enabled
Enabled
Enabled
DBI (DC)4
DBI (DC)8
Notes:
During
READ
DBI (DC)5
During
During
During
MPC[WRIT MPC[READ- MPC[READ
E-FIFO]
FIFO]
DQ CAL]
Train9
Train10
Train11
1. The DMI input signal is "Don’t Care." DMI input receivers are turned off.
2. DMI output drivers are turned off.
3. The MASK WRITE command is not allowed and is considered an illegal command when
the DM function is disabled.
4. The DMI signal is treated as DBI and indicates whether the device needs to invert the
write data received on DQ within a byte. The device inverts write data received on the
DQ inputs if DMI is sampled HIGH and leaves the write data non-inverted if DMI is sampled LOW.
5. The device inverts read data on its DQ outputs associated within a byte and drives the
DMI signal HIGH when more than four data bits = 1 within a given byte lane; otherwise,
the device does not invert the read data and drives DMI signal LOW.
6. The device does not perform a MASK operation when it receives a WRITE (or MRW)
command. During the WRITE burst, the DMI signal must be driven LOW.
7. The device requires an explicit MASKED WRITE command for all MASKED WRITE operations. The DMI signal is treated as a data mask (DM) and indicates which bytes within a
burst will be masked. When the DMI signal is sampled HIGH, the device masks that beat
of the burst for the given byte lane. All DQ input signals within a byte are "Don't Care"
(either HIGH or LOW) when DMI is HIGH. When the DMI signal is sampled LOW, the device does not perform a MASK operation and data received on the DQ inputs is written
to the array.
8. The device requires an explicit MASKED WRITE command for all MASKED WRITE operations. The device masks the write data received on the DQ inputs if five or more data
bits = 1 on DQ[2:7] or DQ[10:15] (for lower byte or upper byte respectively) and the DMI
signal is LOW. Otherwise, the device does not perform the MASK operation and treats it
as a legal DBI pattern. The DMI signal is treated as a DBI signal, and data received on
the DQ input is written to the array.
9. The DMI signal is treated as a training pattern. The device does not perform any MASK
operation and does not invert write data received on the DQ inputs.
10. The DMI signal is treated as a training pattern. The device returns the data pattern written to the WRITE-FIFO.
11. The DMI signal is treated as a training pattern. For more information, see the Read DQ
Calibration Training section.
91
200b: x32 LPDDR4 SDRAM
Data Mask and Data Bus Inversion (DBI [DC]) Function
Figure 35: MASKED WRITE Command with Write DBI Enabled; DM Enabled
T0
T1
T2
T3
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Ta6
Tb0
DES
DES
DES
DES
DES
DES
DES
DES DES
CK_c
CK_t
CKE
CS
CA Valid Valid Valid Valid
Command
MASK WRITE-1
CAS-2
t
WL
DQSS
DQS_c
DQS_t
t
t
DQS2DQ
WPRE
DQ[7:0]
Valid Valid
Valid Valid Valid Valid Valid Valid Valid Valid
N1 I 2 I
M3 N
I
N M N N
DMI[0]
Don’t Care
Notes:
1. N: Input data is written to DRAM cell.
2. I: Input data is inverted, then written to DRAM cell.
3. M: Input data is masked. The total count of 1 data bits on DQ[7:2] is equal to or greater
than five.
4. Data mask (DM) is enable: MR13 OP [5] = 0, Data bus inversion (DBI) write is enable:
MR3 OP[7] = 1.
92
200b: x32 LPDDR4 SDRAM
Data Mask and Data Bus Inversion (DBI [DC]) Function
Figure 36: WRITE Command with Write DBI Enabled; DM Disabled
T0
T1
T2
T3
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Ta6
Tb0
DES
DES
DES
DES
DES
DES
DES
DES DES
CK_c
CK_t
CKE
CS
CA Valid Valid Valid Valid
Command
WRITE-1
CAS-2
t
WL
DQSS
DQS_c
DQS_t
t
DQS2DQ
t
WPRE
DQ[7:0]
Valid Valid Valid
N1 N
Valid Valid Valid Valid Valid Valid Valid
I2 I
N N
I
N N
N
DMI[0]
Don’t Care
Notes:
1. N: Input data is written to DRAM cell.
2. I: Input data is inverted, then written to DRAM cell.
3. Data mask (DM) is disable: MR13 OP [5] = 1, Data bus inversion (DBI) write is enable:
MR3 OP[7] = 1.
93
200b: x32 LPDDR4 SDRAM
Preamble and Postamble Behavior
Preamble and Postamble Behavior
Preamble, Postamble Behavior in READ-to-READ Operations
The following illustrations show the behavior of the device's read DQS_t and DQS_c
pins during cases where the preamble, postamble, and/or data clocking overlap.
DQS will be driven with the following priority
1. Data clocking edges will always be driven
2. Postamble
3. Preamble
Essentially the data clocking, preamble, and postamble will be ordered such that all
edges will be driven.
Additional examples of seamless and borderline non-overlapping cases have been included for clarity.
READ-to-READ Operations – Seamless
Figure 37: READ Operations: tCCD = MIN, Preamble = Toggle, 1.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
T7
T8
T9
T10
T11
CAm
CAm
T12
T13
T14
T15
T16
T17
T18
T19
T20
T26
T27
T28
T29
T30
T31
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
BL
DES
DES
t
BA0,
CA, AP
READ-1
CAS-2
CCD = 8
RL = 6
t
RL = 6
t
DQSCK
DQSCK
t
t
RPST
RPRE
DQS_c
DQS_t
High-Z
High-Z
t
DQ
DMI
High-Z
t
DQSQ
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0 n1 n2 n3 n4 n5 n6 n7 n8 n9 n10 n11 n12 n13 n14 n15 m0 m1 m12 m13 m14 m15
BL/2 = 8
High-Z
BL/2 = 8
Don’t Care
Notes:
1. BL = 16 for column n and column m; RL = 6; Preamble = Toggle; Postamble = 1.5nCK.
2. DOUT n/m = data-out from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
94
200b: x32 LPDDR4 SDRAM
Preamble and Postamble Behavior
READ-to-READ Operations – Consecutive
Figure 38: Seamless READ: tCCD = MIN + 1, Preamble = Toggle, 1.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
T7
T8
T9
T10
T11
T12
BL
BA0,
CA, AP
CAm
CAm
T13
T14
T15
T16
T17
T18
T19
T20
T21
T26
T27
T28
T29
T30
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
DES
DES
t
DES
READ-1
CAS-2
CCD = 9
t
RL = 6
t
RL = 6
DQSCK
DQSCK
t
t
RPST
RPRE
t
RPST
DQS_c
High-Z
DQS_t
High-Z
t
DQ
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0 n1 n2 n3 n4 n5 n6 n7 n8 n9 n10 n11 n12 n13 n14 n15
High-Z
DMI
t
DQSQ
High-Z
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT
m0 m1 m12 m13 m14 m15
BL/2 = 8
High-Z
BL/2 = 8
Don’t Care
1. BL = 16 for column n and column m; RL = 6; Preamble = Toggle; Postamble = 1.5nCK.
2. DOUT n/m = data-out from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Notes:
Figure 39: Consecutive READ: tCCD = MIN + 1, Preamble = Toggle, 0.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
T7
T8
T9
T10
T11
T12
CAm
CAm
T13
T14
T15
T16
T17
T18
T19
T20
T21
T26
T27
T28
T29
T30
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
BL
DES
DES
t
DES
BA0,
CA, AP
READ-1
CAS-2
CCD = 9
RL = 6
t
RL = 6
t
DQSCK
DQSCK
t
t
RPRE
t
RPST
RPRE
t
RPST
DQS_c
DQS_t
High-Z
High-Z
t
DQ
DMI
High-Z
t
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0 n1 n2 n3 n4 n5 n6 n7 n8 n9 n10 n11 n12 n13 n14 n15
BL/2 = 8
High-Z
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT
m0 m1 m12 m13 m14 m15
High-Z
BL/2 = 8
Don’t Care
Notes:
1. BL = 16 for column n and column m; RL = 6; Preamble = Toggle; Postamble = 0.5nCK.
2. DOUT n/m = data-out from column n and column m.
95
200b: x32 LPDDR4 SDRAM
Preamble and Postamble Behavior
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Figure 40: Consecutive READ: tCCD = MIN + 1, Preamble = Static, 1.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
T7
T8
T9
T10
T11
T12
BL
BA0,
CA, AP
CAm
CAm
T13
T14
T15
T16
T17
T18
T19
T20
T21
T26
T27
T28
T29
T30
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
DES
DES
t
DES
READ-1
CAS-2
CCD = 9
t
RL = 6
t
RL = 6
DQSCK
DQSCK
t
t
RPRE
t
RPST
RPST
DQS_c
High-Z
DQS_t
High-Z
t
DQ
DMI
t
DQSQ
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0 n1 n2 n3 n4 n5 n6 n7 n8 n9 n10 n11 n12 n13 n14 n15
High-Z
High-Z
DOUT DOUT DOUT DOUT DOUT DOUT
m0 m1 m12 m13 m14 m15
BL/2 = 8
High-Z
BL/2 = 8
Don’t Care
1. BL = 16 for column n and column m; RL = 6; Preamble = Static; Postamble = 1.5nCK.
2. DOUT n/m = data-out from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Notes:
Figure 41: Consecutive READ: tCCD = MIN + 1, Preamble = Static, 0.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
T7
T8
T9
T10
T11
T12
CAm
CAm
T13
T14
T15
T16
T17
T18
DES
DES
DES
DES
DES
DES
T19
T20
T21
T26
T27
T28
T29
T30
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
BL
DES
DES
t
DES
BA0,
CA, AP
READ-1
CAS-2
CCD = 9
RL = 6
RL = 6
t
t
DQSCK
DQSCK
t
t
RPRE
t
RPRE
RPST
DQS_c
DQS_t
High-Z
High-Z
t
DQ
DMI
High-Z
t
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0 n1 n2 n3 n4 n5 n6 n7 n8 n9 n10 n11 n12 n13 n14 n15
BL/2 = 8
High-Z
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT
m0 m1 m12 m13 m14 m15
High-Z
BL/2 = 8
Don’t Care
Notes:
1. BL = 16 for column n and column m; RL = 6; Preamble = Static; Postamble = 0.5nCK.
96
200b: x32 LPDDR4 SDRAM
Preamble and Postamble Behavior
2. DOUT n/m = data-out from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Figure 42: Consecutive READ: tCCD = MIN + 2, Preamble = Toggle, 1.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
T7
T8
T9
T10
T11
T12
T13
BL
BA0,
CA, AP
CAm
CAm
T14
T15
T16
T17
T18
T19
T20
T21
T22
T28
T29
T30
T31
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
DES
DES
t
DES
DES
READ-1
CAS-2
CCD = 10
RL = 6
t
RL = 6
t
DQSCK
DQSCK
t
t
RPRE
RPST
t
t
RPRE
RPST
DQS_c
DQS_t
High-Z
High-Z
t
DQ
DMI
High-Z
t
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0 n1 n2 n3 n4 n5 n6 n7 n8 n9 n10 n11 n12 n13 n14 n15
BL/2 = 8
High-Z
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT
m0 m1 m12 m13 m14 m15
High-Z
BL/2 = 8
Don’t Care
Notes:
1. BL = 16 for column n and column m; RL = 6; Preamble = Toggle; Postamble = 1.5nCK.
2. DOUT n/m = data-out from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
97
200b: x32 LPDDR4 SDRAM
Preamble and Postamble Behavior
Figure 43: Consecutive READ: tCCD = MIN + 2, Preamble = Toggle, 0.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
T7
T8
T9
T10
T11
T12
T13
BL
BA0,
CA, AP
CAm
CAm
T14
T15
T16
T17
T18
T19
T20
T21
T22
T28
T29
T30
T31
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
DES
DES
t
DES
DES
READ-1
CAS-2
CCD = 10
t
RL = 6
t
RL = 6
DQSCK
DQSCK
t
RPRE
t
t
RPST
RPRE
t
DQS_c
RPST
High-Z
DQS_t
High-Z
t
DQ
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0 n1 n2 n3 n4 n5 n6 n7 n8 n9 n10 n11 n12 n13 n14 n15
High-Z
DMI
t
High-Z
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT
m0 m1 m12 m13 m14 m15
High-Z
BL/2 = 8
BL/2 = 8
Don’t Care
1. BL = 16 for column n and column m; RL = 6; Preamble = Toggle; Postamble = 0.5nCK.
2. DOUT n/m = data-out from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Notes:
Figure 44: Consecutive READ: tCCD = MIN + 2, Preamble = Static, 1.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
T7
T8
T9
T10
T11
T12
T13
BL
BA0,
CA, AP
CAm
CAm
T14
T15
T16
T17
T18
T19
T20
T21
T22
T28
T29
T30
T31
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
DES
DES
t
DES
DES
READ-1
CAS-2
CCD = 10
RL = 6
t
RL = 6
t
DQSCK
DQSCK
t
t
RPRE
RPST
t
t
RPRE
RPST
DQS_c
DQS_t
High-Z
High-Z
t
DQ
DMI
High-Z
t
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0 n1 n2 n3 n4 n5 n6 n7 n8 n9 n10 n11 n12 n13 n14 n15
BL/2 = 8
High-Z
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT
m0 m1 m12 m13 m14 m15
High-Z
BL/2 = 8
Don’t Care
Notes:
1. BL = 16 for column n and column m; RL = 6; Preamble = Static; Postamble = 1.5nCK.
2. DOUT n/m = data-out from column n and column m.
98
200b: x32 LPDDR4 SDRAM
Preamble and Postamble Behavior
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Figure 45: Consecutive READ: tCCD = MIN + 2, Preamble = Static, 0.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
T7
T8
T9
T10
T11
T12
T13
BL
BA0,
CA, AP
CAm
CAm
T14
T15
T16
T17
T18
T19
T20
T21
T22
T28
T29
T30
T31
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
DES
DES
t
DES
DES
READ-1
CAS-2
CCD = 10
RL = 6
t
RL = 6
t
DQSCK
DQSCK
t
RPRE
t
RPST
t
RPRE
t
RPST
DQS_c
DQS_t
High-Z
High-Z
t
DQ
DMI
High-Z
t
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0 n1 n2 n3 n4 n5 n6 n7 n8 n9 n10 n11 n12 n13 n14 n15
BL/2 = 8
High-Z
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT
m0 m1 m12 m13 m14 m15
High-Z
BL/2 = 8
Don’t Care
Notes:
1. BL = 16 for column n and column m; RL = 6; Preamble = Static; Postamble = 0.5nCK.
2. DOUT n/m = data-out from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
99
200b: x32 LPDDR4 SDRAM
Preamble and Postamble Behavior
Figure 46: Consecutive READ: tCCD = MIN + 3, Preamble = Toggle, 1.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
T7
T8
T9
T10
T11
T12
T13
T14
BL
BA0,
CA, AP
CAm
CAm
T15
T16
T17
T18
T19
T20
T21
T22
T23
T29
T30
T31
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
DES
DES
t
DES
DES
DES
READ-1
CAS-2
CCD = 11
t
RL = 6
t
RL = 6
DES
DQSCK
DQSCK
t
t
RPRE
t
RPST
t
RPRE
RPST
DQS_c
High-Z
DQS_t
High-Z
t
DQ
DMI
t
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0 n1 n2 n3 n4 n5 n6 n7 n8 n9 n10 n11 n12 n13 n14 n15
High-Z
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT
m0 m1 m12 m13 m14 m15
High-Z
BL/2 = 8
High-Z
BL/2 = 8
Don’t Care
1. BL = 16 for column n and column m; RL = 6; Preamble = Toggle; Postamble = 1.5nCK.
2. DOUT n/m = data-out from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Notes:
Figure 47: Consecutive READ: tCCD = MIN + 3, Preamble = Toggle, 0.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
T7
T8
T9
T10
T11
T12
T13
T14
BL
BA0,
CA, AP
CAm
CAm
T15
T16
T17
T18
T19
T20
T21
T22
T23
T29
T30
T31
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
DES
DES
t
DES
DES
DES
READ-1
CAS-2
CCD = 11
RL = 6
t
RL = 6
t
DQSCK
DQSCK
t
t
RPRE
t
RPST
RPRE
t
DQS_c
DQS_t
High-Z
DMI
High-Z
RPST
High-Z
t
DQ
DES
High-Z
t
DQSQ
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0 n1 n2 n3 n4 n5 n6 n7 n8 n9 n10 n11 n12 n13 n14 n15
BL/2 = 8
High-Z
DOUT DOUT DOUT DOUT DOUT DOUT
m0 m1 m12 m13 m14 m15
High-Z
BL/2 = 8
Don’t Care
Notes:
1. BL = 16 for column n and column m; RL = 6; Preamble = Toggle; Postamble = 0.5nCK.
2. DOUT n/m = data-out from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
100
200b: x32 LPDDR4 SDRAM
Preamble and Postamble Behavior
Figure 48: Consecutive READ: tCCD = MIN + 3, Preamble = Static, 1.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
T7
T8
T9
T10
T11
T12
T13
T14
BL
BA0,
CA, AP
CAm
CAm
T15
T16
T17
T18
T19
T20
T21
T22
T23
T29
T30
T31
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
DES
DES
t
DES
DES
DES
READ-1
CAS-2
CCD = 11
t
RL = 6
t
RL = 6
DES
DQSCK
DQSCK
t
t
RPRE
t
RPST
t
RPRE
RPST
DQS_c
High-Z
DQS_t
High-Z
t
DQ
DQSQ
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0 n1 n2 n3 n4 n5 n6 n7 n8 n9 n10 n11 n12 n13 n14 n15
High-Z
DMI
t
DOUT DOUT DOUT DOUT DOUT DOUT
m0 m1 m12 m13 m14 m15
High-Z
BL/2 = 8
High-Z
BL/2 = 8
Don’t Care
1. BL = 16 for column n and column m; RL = 6; Preamble = Static; Postamble = 1.5nCK.
2. DOUT n/m = data-out from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Notes:
Figure 49: Consecutive READ: tCCD = MIN + 3, Preamble = Static, 0.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
T7
T8
T9
T10
T11
T12
T13
T14
BL
BA0,
CA, AP
CAm
CAm
T15
T16
T17
T18
T19
T20
DES
DES
DES
DES
DES
DES
T21
T22
T23
T29
T30
T31
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
DES
DES
t
DES
DES
DES
READ-1
CAS-2
CCD = 11
RL = 6
t
RL = 6
t
DES
DQSCK
DQSCK
t
t
RPRE
t
RPST
RPRE
t
RPST
DQS_c
DQS_t
High-Z
High-Z
t
DQ
DMI
High-Z
High-Z
t
DQSQ
DQSQ
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0 n1 n2 n3 n4 n5 n6 n7 n8 n9 n10 n11 n12 n13 n14 n15
BL/2 = 8
High-Z
DOUT DOUT DOUT DOUT DOUT DOUT
m0 m1 m12 m13 m14 m15
High-Z
BL/2 = 8
Don’t Care
Notes:
1. BL = 16 for column n and column m; RL = 6, Preamble = Static; Postamble = 0.5nCK
2. DOUT n/m = data-out from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
101
200b: x32 LPDDR4 SDRAM
Preamble and Postamble Behavior
WRITE-to-WRITE Operations – Seamless
Figure 50: Seamless WRITE: tCCD = MIN, 0.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA
CAn
CAn
T4
T5
T6
T7
T8
T9
T10
T11
BL
BA0,
CA
CAm
CAm
T12
T13
T14
T15
T16
T17
T18
T23
T24
T25
T26
T27
T28
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
DES
DES
t
DES
DES
WRITE-1
CAS-2
CCD = 8
WL = 4
WL = 4
t
t
t
DQSS
DQSS
WPRE
t
WPST
DQS_c
DQS_t
t
DQ
DIN
n0
DMI
t
DQS2DQ
DIN DIN DIN
n1 n2 n3
DIN
n4
DIN DIN
n5 n6
DIN
n7
DIN
n8
BL/2 = 8
DIN
n9
DIN DIN DIN DIN DIN DIN
n10 n11 n12 n13 n14 n15
DIN
m0
DQS2DQ
DIN
m1
DIN
m2
DIN
m3
DIN
m12
DIN DIN DIN
m13 m14 m15
BL/2 = 8
Don’t Care
Notes:
1. BL = 16, Write postamble = 0.5nCK.
2. DIN n/m = data-in from column n and column m.
3. The minimum number of clock cycles from the burst WRITE command to the burst
WRITE command for any bank is BL/2.
4. DES commands are shown for ease of illustration; other commands may be valid at
these times.
102
200b: x32 LPDDR4 SDRAM
Preamble and Postamble Behavior
Figure 51: Seamless WRITE: tCCD = MIN, 1.5nCK Postamble, 533 MHz < Clock Frequency ≤ 800 MHz,
ODT Worst Timing Case
T0
T1
T2
T3
BL
BA0,
CA
CAn
CAn
T4
T7
T8
T9
T10
T11
BL
BA0,
CA
CAm
CAm
T12
T13
T14
T15
T16
T17
T23
T24
T25
DES
DES
DES
DES
DES
DES
DES
DES
DES
T31
T32
T33
T34
T35
T36
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
DES
DES
t
WRITE-1
CAS-2
CCD = 8
WL = 12
WL = 12
t
t
t
DES
DQSS
DQSS
t
WPRE
WPST
DQS_c
DQS_t
t
DQ
DIN
n0
DMI
ODTLon = 6
DRAM RTT
t
ODTon (MAX)
t
DQS2DQ
DIN DIN DIN DIN DIN
n1 n2 n13 n14 n15
BL/2 = 8
DIN
m0
DQS2DQ
DIN
m1
DIN
m2
DIN DIN DIN
m13 m14 m15
BL/2 = 8
ODT on
ODT High-Z
ODTLoff = 22
ODT High-Z
t
ODToff (MIN)
Don’t Care
Notes:
Clock frequency = 800 MHz, tCK(AVG) = 1.25ns.
BL = 16, Write postamble = 1.5nCK.
DIN n/m = data-in from column n and column m.
The minimum number of clock cycles from the burst WRITE command to the burst
WRITE command for any bank is BL/2.
5. DES commands are shown for ease of illustration; other commands may be valid at
these times.
1.
2.
3.
4.
103
200b: x32 LPDDR4 SDRAM
Preamble and Postamble Behavior
Figure 52: Seamless WRITE: tCCD = MIN, 1.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA
CAn
CAn
T4
T7
T8
T9
T10
T11
BL
BA0,
CA
CAm
CAm
T12
T15
T16
T17
T18
T19
T25
T26
T27
DES
DES
DES
DES
DES
DES
DES
DES
DES
T33
T34
T35
T36
T37
T38
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
DES
DES
t
CCD = 8
WRITE-1
CAS-2
WL = 14
WL = 14
t
t
t
DES
DQSS
DQSS
t
WPRE
WPST
DQS_c
DQS_t
t
DQ
t
DQS2DQ
DIN
n0
DMI
DIN DIN DIN DIN DIN
n1 n2 n13 n14 n15
BL/2 = 8
DIN
m0
DQS2DQ
DIN
m1
DIN
m2
DIN DIN DIN
m13 m14 m15
BL/2 = 8
Don’t Care
Notes:
1. BL = 16, Write postamble = 1.5nCK.
2. DIN n/m = data-in from column n and column m.
3. The minimum number of clock cycles from the burst WRITE command to the burst
WRITE command for any bank is BL/2.
4. DES commands are shown for ease of illustration; other commands may be valid at
these times.
104
200b: x32 LPDDR4 SDRAM
Preamble and Postamble Behavior
WRITE-to-WRITE Operations – Consecutive
Figure 53: Consecutive WRITE: tCCD = MIN + 1, 0.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA
CAn
CAn
T4
T8
T9
T10
T11
T12
BL
BA0,
CA
CAm
CAm
T13
T14
T15
T16
T17
T23
T24
T25
T26
T32
DES
DES
DES
DES
DES
DES
DES
DES
DES
T33
T34
T35
T36
T37
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
DES
WRITE-1
DES
t
CAS-2
CCD = 9
WL = 12
WL = 12
t
t
t
DES
DQSS
DQSS
WPRE
t
WPST
DQS_c
DQS_t
t
DQ
DMI
t
DQS2DQ
DQS2DQ
DIN
m0
DIN DIN DIN DIN DIN
n1 n2 n13 n14 n15
DIN
n0
DIN
m2
DIN
m1
DIN DIN DIN
m13 m14 m15
BL/2 = 8
BL/2 = 8
Don’t Care
Notes:
1. BL = 16, Write postamble = 0.5nCK.
2. DIN n/m = data-in from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Figure 54: Consecutive WRITE: tCCD = MIN + 1, 1.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA
CAn
CAn
T4
T8
T9
T10
T11
T12
BL
BA0,
CA
CAm
CAm
T13
T14
T15
T16
T17
T23
T24
T25
T26
T32
DES
DES
DES
DES
DES
DES
DES
DES
DES
T33
T34
T35
T36
T37
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
DES
WRITE-1
DES
t
CAS-2
CCD = 9
WL = 12
WL = 12
t
t
t
DES
DQSS
DQSS
t
WPRE
WPST
DQS_c
DQS_t
t
DQ
DIN
n0
DMI
DQS2DQ
DIN DIN DIN DIN DIN
n1 n2 n13 n14 n15
BL/2 = 8
t
DQS2DQ
DIN
m0
DIN
m1
DIN
m2
DIN DIN DIN
m13 m14 m15
BL/2 = 8
Don’t Care
Notes:
1. BL = 16, Write postamble = 1.5nCK.
2. DIN n/m = data-in from column n and column m.
105
200b: x32 LPDDR4 SDRAM
Preamble and Postamble Behavior
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Figure 55: Consecutive WRITE: tCCD = MIN + 2, 0.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA
CAn
CAn
T4
T9
T10
T11
T12
T13
BL
BA0,
CA
CAm
CAm
T14
T15
T16
T17
T23
T24
T25
T26
T27
T33
T34
T35
T36
T37
T38
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
DES
WRITE-1
DES
t
CAS-2
CCD = 10
WL = 12
WL = 12
t
t
t
DQSS
DQSS
t
WPRE
WPRE
t
WPST
DQS_c
DQS_t
t
DQ
DMI
t
DQS2DQ
DQS2DQ
DIN
m0
DIN DIN DIN DIN DIN
n1 n2 n13 n14 n15
DIN
n0
DIN
m2
DIN
m1
DIN DIN DIN
m13 m14 m15
BL/2 = 8
BL/2 = 8
Don’t Care
Notes:
1. BL = 16, Write postamble = 0.5nCK.
2. DIN n/m = data-in from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Figure 56: Consecutive WRITE: tCCD = MIN + 2, 1.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA
CAn
CAn
T4
T9
T10
T11
T12
T13
BL
BA0,
CA
CAm
CAm
T14
T15
T16
T17
T23
T24
T25
T26
T27
T33
T34
T35
T36
T37
T38
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
DES
WRITE-1
DES
t
CAS-2
CCD = 10
WL = 12
WL = 12
t
t
t
DQSS
DQSS
t
WPRE
t
WPRE
WPST
DQS_c
DQS_t
t
DQ
DIN
n0
DMI
DQS2DQ
DIN DIN DIN DIN DIN
n1 n2 n13 n14 n15
BL/2 = 8
t
DQS2DQ
DIN
m0
DIN
m1
DIN
m2
DIN DIN DIN
m13 m14 m15
BL/2 = 8
Don’t Care
Notes:
1. BL = 16, Write postamble = 1.5nCK.
106
200b: x32 LPDDR4 SDRAM
Preamble and Postamble Behavior
2. DIN n/m = data-in from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Figure 57: Consecutive WRITE: tCCD = MIN + 3, 0.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA
CAn
CAn
T4
T9
T10
T11
T12
T13
T14
BL
BA0,
CA
CAm
CAm
T15
T16
T17
T23
T24
T25
T26
T27
T28
T34
T35
T36
T37
T38
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
DES
DES
DES
t
WRITE-1
CAS-2
CCD = 11
WL = 12
WL = 12
t
t
t
DQSS
DQSS
WPRE
t
WPST
t
WPRE
t
WPST
DQS_c
DQS_t
t
DQ
DIN
n0
DMI
DQS2DQ
DIN DIN DIN DIN DIN
n1 n2 n13 n14 n15
BL/2 = 8
t
DQS2DQ
DIN
m0
DIN
m1
DIN
m2
Din DIN DIN
m13 m14 m15
BL/2 = 8
Don’t Care
Notes:
1. BL = 16, Write postamble = 0.5nCK.
2. DIN n/m = data-in from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
107
200b: x32 LPDDR4 SDRAM
Preamble and Postamble Behavior
Figure 58: Consecutive WRITE: tCCD = MIN + 3, 1.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA
CAn
CAn
T4
T9
T10
T11
T12
T13
T14
BL
BA0,
CA
CAm
CAm
T15
T16
T17
T23
T24
T25
T26
T27
T28
T34
T35
T36
T37
T38
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
DES
WRITE-1
DES
DES
t
CAS-2
CCD = 11
WL = 12
WL = 12
t
t
t
DQSS
DQSS
t
WPRE
t
WPST
t
WPRE
WPST
DQS_c
DQS_t
t
DQ
DMI
t
DQS2DQ
DQS2DQ
DIN
m0
DIN DIN DIN DIN DIN
n1 n2 n13 n14 n15
DIN
n0
DIN
m2
DIN
m1
DIN DIN DIN
m13 m14 m15
BL/2 = 8
BL/2 = 8
Don’t Care
Notes:
1. BL = 16, Write postamble = 1.5nCK.
2. DIN n/m = data-in from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Figure 59: Consecutive WRITE: tCCD = MIN + 4, 1.5nCK Postamble
T0
T1
T2
T3
BL
BA0,
CA
CAn
CAn
T4
T9
T10
T11
T12
T13
T14
BL
BA0,
CA
CAm
CAm
T15
T16
T17
T23
T24
T25
T26
T27
T28
T29
T35
T36
T37
T38
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
DES
DES
WRITE-1
DES
t
CAS-2
CCD = 12
WL = 12
WL = 12
t
t
t
DQSS
DQSS
t
WPRE
WPST
t
t
WPRE
WPST
DQS_c
DQS_t
t
DQ
DIN
n0
DMI
DQS2DQ
DIN DIN DIN DIN DIN
n1 n2 n13 n14 n15
BL/2 = 8
t
DIN
m0
DQS2DQ
DIN
m1
DIN
m2
DIN DIN DIN
m13 m14 m15
BL/2 = 8
Don’t Care
Notes:
1. BL = 16, Write postamble = 1.5nCK.
2. DIN n/m = data-in from column n and column m.
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
108
200b: x32 LPDDR4 SDRAM
PRECHARGE Operation
PRECHARGE Operation
The PRECHARGE command is used to precharge or close a bank that has been activated. The PRECHARGE command is initiated with CKE, CS, and CA[5:0] in the proper
state (see Command Truth Table). The PRECHARGE command can be used to precharge each bank independently or all banks simultaneously. The all banks (AB) flag
and the bank address bit are used to determine which bank(s) to precharge. The precharged bank(s) will be available for subsequent row access tRPab after an all-bank
PRECHARGE command is issued, or tRPpb after a single-bank PRECHARGE command
is issued.
To ensure that the device can meet the instantaneous current demands, the row precharge time for an all-bank PRECHARGE ( tRPab) is longer than the per-bank precharge
time (tRPpb).
Table 94: Precharge Bank Selection
AB (CA[5], R1)
BA2 (CA[2], R2)
BA1 (CA[1], R2)
BA0 (CA[0], R2)
Precharged Bank
0
0
0
0
Bank 0 only
0
0
0
1
Bank 1 only
0
0
1
0
Bank 2 only
0
0
1
1
Bank 3 only
0
1
0
0
Bank 4 only
0
1
0
1
Bank 5 only
0
1
1
0
Bank 6 only
0
1
1
1
Bank 7 only
1
Don't Care
Don't Care
Don't Care
All banks
Burst READ Operation Followed by Precharge
The PRECHARGE command can be issued as early as BL/2 clock cycles after a READ
command, but the PRECHARGE command cannot be issued until after tRAS is satisfied.
A new bank ACTIVATE command can be issued to the same bank after the row precharge time (tRP) has elapsed. The minimum read-to-precharge time must also satisfy a
minimum analog time from the second rising clock edge of the CAS-2 command. tRTP
begins BL/2 - 8 clock cycles after the READ command.
109
200b: x32 LPDDR4 SDRAM
PRECHARGE Operation
Figure 60: Burst READ Followed by Precharge – BL16, Toggling Preamble, 0.5nCK Postamble
T0
T1
T2
T3
T4
Tx
Tx+1 Tx+2 Tx+3 Tx+4 Tx+5 Tx+6 Tx+7
Valid
Valid
Valid
Valid
Valid
Valid
Ty
Ty+1 Ty+2 Ty+3 Ty+4
CK_c
CK_t
CA[5:0]
Valid
Valid
Valid
Valid
Valid
Valid
tRTP
READ-1
Command
CAS-2
Valid
Valid
Valid
Valid
Valid
tRP
Valid
PRECHARGE
Valid
Valid
Valid
Valid
ACT-2
ACT-1
DQS_t
DQS_c
DQ[15:0]
DMI[1:0]
Valid
Don’t Care
Transitioning Data
Figure 61: Burst READ Followed by Precharge – BL32, 2tCK, 0.5nCK Postamble
T0
T1
T2
T3
T4
T5
T10
T11
T12
Tx
Tx+1 Tx+2 Tx+3 Tx+4 Tx+5 Tx+6 Tx+7
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Ty
Ty+1 Ty+2 Ty+3 Ty+4
CK_c
CK_t
CA[5:0]
Valid
Valid
Valid
Valid
Valid
tRTP
Command
READ-1
CAS-2
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
tRP
PRECHARGE
Valid
Valid
Valid
ACT-1
ACT-2
DQS_t
DQS_c
DQ[15:0]
DMI[1:0]
Valid
Transitioning Data
Don’t Care
Burst WRITE Followed by Precharge
A write recovery time (tWR) must be provided before a PRECHARGE command may be
issued. This delay is referenced from the next rising edge of CK after the last valid DQS
clock of the burst.
Devices write data to the memory array in prefetch multiples (prefetch = 16). An internal WRITE operation can only begin after a prefetch group has been clocked; therefore,
tWR starts at the prefetch boundaries. The minimum write-to-precharge time for commands to the same bank is WL + BL/2 + 1 + RU( tWR /tCK) clock cycles.
110
200b: x32 LPDDR4 SDRAM
Auto Precharge
Figure 62: Burst WRITE Followed by PRECHARGE – BL16, 2nCK Preamble, 0.5nCK Postamble
T0
T1
T2
T3
T4
Tx
Tx+1
Tx+2
Tx+3
Tx+4
Tx+5
Tx+6
Ta
Ta+1
Ta+2
Tn
Tn+1
Tn+2
Tn+3
Ty
Ty+1
Ty+2
Ty+3
Ty+4
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
PRECHARGE
Valid
CK_c
CK_t
CA
WL
Command
WRITE-1
CAS-2
Valid
Valid
Valid
tDQSS
Valid
Valid
Valid
tWR
(MAX)
ACT-1
ACT-2
tRP
DQS_c
DQS_t
tDQS2DQ
DQ
DMI
Valid
Transitioning Data
Don’t Care
Auto Precharge
Before a new row can be opened in an active bank, the active bank must be precharged
using either the PRECHARGE command or the auto precharge (AP) function. When a
READ or a WRITE command is issued to the device, the AP bit (CA5) can be set to enable the active bank to automatically begin precharge at the earliest possible moment
during the burst READ or WRITE cycle.
If AP is LOW when the READ or WRITE command is issued, the normal READ or WRITE
burst operation is executed, and the bank remains active at the completion of the burst.
If AP is HIGH when the READ or WRITE command is issued, the auto PRECHARGE
function is engaged. This feature enables the PRECHARGE operation to be partially or
completely hidden during burst READ cycles (dependent upon READ or WRITE latency), thus improving system performance for random data access.
Burst READ With Auto Precharge
If AP is HIGH when a READ command is issued, the READ with AUTO PRECHARGE
function is engaged. The devices start an AUTO PRECHARGE operation on the rising
edge of the clock at BL/2 after the second beat of the READ w/AP command, or BL/4 - 4
+ RU (tRTP/tCK) clock cycles after the second beat of the READ w/AP command, whichever is greater. Following an AUTO PRECHARGE operation, an ACTIVATE command can
be issued to the same bank if the following two conditions are both satisfied:
1. The RAS precharge time (tRP) has been satisfied from the clock at which the auto
precharge began, and
2. The RAS cycle time (tRC) from the previous bank activation has been satisfied.
111
200b: x32 LPDDR4 SDRAM
Auto Precharge
Figure 63: Burst READ With Auto Precharge – BL16, Non-Toggling Preamble, 0.5nCK Postamble
T0
T1
T2
T3
T4
Tx
Tx+1 Tx+2 Tx+3 Tx+4 Tx+5 Tx+6 Tx+7
Valid
Valid
Valid
Valid
Valid
Valid
Ty
Ty+1 Ty+2 Ty+3 Ty+4
CK_c
CK_t
CA[5:0]
Valid
Valid
Valid
Valid
Valid
Valid
tRTP
READ-1
w/AP
Command
CAS-2
Valid
Valid
Valid
Valid
Valid
tRPpb
Valid
Valid
Valid
Valid
Valid
Valid
ACT-2
ACT-1
DQS_t
DQS_c
DQ[15:0]
DMI[1:0]
Valid
Don’t Care
Transitioning Data
Figure 64: Burst READ With Auto Precharge – BL32, Toggling Preamble, 1.5nCK Postamble
T0
T1
T2
T3
T4
T5
T10
T11
T12
T13
T14
Tx
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Tx+1 Tx+2 Tx+3 Tx+4 Tx+5
Ty
Ty+1 Ty+2 Ty+3 Ty+4
CK_c
CK_t
CA[5:0]
Valid
Valid
Valid
tRTP
Command
READ-1
CAS-2
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
tRP
Valid
Valid
Valid
Valid
Valid
ACT-1
ACT-2
DQS_t
DQS_c
DQ[15:0]
DMI[1:0]
Valid
Transitioning Data
Don’t Care
Burst WRITE With Auto Precharge
If AP is HIGH when a WRITE command is issued, the WRITE with AUTO PRECHARGE
function is engaged. The device starts an auto precharge on the rising edge tWR cycles
after the completion of the burst WRITE.
Following a WRITE with AUTO PRECHARGE, an ACTIVATE command can be issued to
the same bank if the following conditions are met:
1. The RAS precharge time (tRP) has been satisfied from the clock at which the auto
precharge began, and
112
200b: x32 LPDDR4 SDRAM
Auto Precharge
2. The RAS cycle time (tRC) from the previous bank activation has been satisfied.
Figure 65: Burst WRITE With Auto Precharge – BL16, 2nCK Preamble, 0.5nCK Postamble
T0
T1
T2
T3
T4
Valid
Valid
Valid
Valid
Valid
Tx
Tx+1
Tx+2
Tx+3
Tx+4
Tx+5
Tx+6
Ta
Ta+1
Ta+2
Tn
Tn+1
Tn+2
Tn+3
Ty
Ty+1
Ty+2
Ty+3
Ty+4
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
CK_c
CK_t
CA
WL
Command
WRITE-1
CAS-2
Valid
Valid
Valid
tDQSS
Valid
Valid
Valid
Valid
Valid
tWR
(MAX)
ACT-1
ACT-2
tRP
DQS_c
DQS_t
tDQS2DQ
DQ[15:0]
DMI[1:0]
Valid
Transitioning Data
Don’t Care
Table 95: Timing Between Commands (PRECHARGE and AUTO PRECHARGE): DQ ODT is Disable
Minimum Delay Between
"From Command" and "To Command"
Unit
Notes
PRECHARGE
(to same bank as READ)
tRTP
tCK
1, 6
PRECHARGE ALL
tRTP
tCK
1, 6
PRECHARGE
(to same bank as READ)
8tCK + tRTP
tCK
1, 6
PRECHARGE ALL
8tCK + tRTP
tCK
1, 6
PRECHARGE
(to same bank as READ w/AP)
nRTP
tCK
1, 10
PRECHARGE ALL
nRTP
tCK
1, 10
nRTP + tRPpb
tCK
1, 8, 10
WRITE or WRITE w/AP
(same bank)
Illegal
–
MASK-WR or MASK-WR w/AP
(same bank)
Illegal
–
WRITE or WRITE w/AP
(different bank)
RL + RU (tDQSCK MAX/tCK) + BL/2
+ RD (tRPST) - WL + tWPRE
tCK
3, 4, 5
MASK-WR or MASK-WR w/AP
(different bank)
RL + RU (tDQSCK MAX/tCK) + BL/2
+ RD (tRPST) - WL + tWPRE
tCK
3, 4, 5
READ or READ w/AP
(same bank)
Illegal
–
READ or READ w/AP
(different bank)
BL/2
tCK
From Command To Command
READ
BL = 16
READ
BL = 32
READ w/AP
BL = 16
ACTIVATE
(to same bank as READ w/AP)
113
3
200b: x32 LPDDR4 SDRAM
Auto Precharge
Table 95: Timing Between Commands (PRECHARGE and AUTO PRECHARGE): DQ ODT is Disable (Continued)
Minimum Delay Between
"From Command" and "To Command"
Unit
Notes
PRECHARGE
(to same bank as READ w/AP)
8tCK + nRTP
tCK
1, 10
PRECHARGE ALL
8tCK + nRTP
tCK
1, 10
tCK
1, 8, 10
From Command To Command
READ w/AP
BL = 32
8tCK
ACTIVATE
(to same bank as READ w/AP)
WRITE
BL = 16 and 32
MASK-WR
BL = 16
WRITE w/AP
BL = 16 and 32
+ nRTP +
tRPpb
WRITE or WRITE w/AP
(same bank)
Illegal
–
MASK-WR or MASK-WR w/AP
(same bank)
Illegal
–
WRITE or WRITE w/AP
(different bank)
RL + RU (tDQSCK MAX/tCK) + BL/2
+ RD (tRPST) - WL + tWPRE
tCK
3, 4, 5
MASK-WR or MASK-WR w/AP
(different bank)
RL + RU (tDQSCK MAX/tCK) + BL/2
+ RD (tRPST) - WL + tWPRE
tCK
3, 4, 5
READ or READ w/AP
(same bank)
Illegal
–
READ or READ w/AP
(different bank)
BL/2
tCK
3
PRECHARGE
(to same bank as WRITE)
WL + BL/2 + tWR + 1
tCK
1, 7
PRECHARGE ALL
WL + BL/2 + tWR + 1
tCK
1, 7
tCK
1, 7
tWR
+1
PRECHARGE
(to same bank as MASK-WR)
WL + BL/2 +
PRECHARGE ALL
WL + BL/2 + tWR + 1
tCK
1, 7
PRECHARGE
(to same bank as WRITE w/AP)
WL + BL/2 + nWR + 1
tCK
1, 11
PRECHARGE ALL
WL + BL/2 + nWR + 1
tCK
1, 11
tCK
1, 8, 11
ACTIVATE
(to same bank as WRITE w/AP)
WL + BL/2 + nWR + 1 +
tRPpb
WRITE or WRITE w/AP
(same bank)
Illegal
–
READ or READ w/AP
(same bank)
Illegal
–
WRITE or WRITE w/AP
(different bank)
BL/2
tCK
3
MASK-WR or MASK-WR w/AP
(different bank)
BL/2
tCK
3
WL + BL/2 + tWTR + 1
tCK
3, 9
READ or READ w/AP
(different bank)
114
200b: x32 LPDDR4 SDRAM
Auto Precharge
Table 95: Timing Between Commands (PRECHARGE and AUTO PRECHARGE): DQ ODT is Disable (Continued)
Minimum Delay Between
"From Command" and "To Command"
Unit
Notes
PRECHARGE
(to same bank as MASK-WR
w/AP)
WL + BL/2 + nWR +1
tCK
1, 11
PRECHARGE ALL
WL + BL/2 + nWR + 1
tCK
1, 11
tCK
1, 8, 11
From Command To Command
MASK-WR w/AP
BL = 16
WL + BL/2 + nWR + 1 +
ACTIVATE
(to same bank as MASK-WR
w/AP)
PRECHARGE
PRECHARGE ALL
tRPpb
WRITE or WRITE w/AP
(same bank)
Illegal
–
3
MASK-WR or MASK-WR w/AP
(same bank)
Illegal
–
3
WRITE or WRITE w/AP
(different bank)
BL/2
tCK
3
MASK-WR or MASK-WR w/AP
(different bank)
BL/2
tCK
3
READ or READ w/AP
(same bank)
Illegal
–
3
READ or READ w/AP
(different bank)
WL + BL/2 + tWTR + 1
tCK
3, 9
PRECHARGE
(to same bank as PRECHARGE)
4
tCK
1
PRECHARGE ALL
4
tCK
1
PRECHARGE
4
tCK
1
PRECHARGE ALL
4
tCK
1
Notes:
1. For a given bank, the precharge period should be counted from the latest PRECHARGE
command, whether per-bank or all-bank, issued to that bank. The precharge period is
satisfied tRP after that latest PRECHARGE command.
2. Any command issued during the minimum delay time as specified in the table above is
illegal.
3. After READ w/AP, seamless READ operations to different banks are supported. After
WRITE w/AP or MASK-WR w/AP, seamless WRITE operations to different banks are supported. READ, WRITE, and MASK-WR operations may not be truncated or interrupted.
4. tRPST values depend on MR1 OP[7] respectively.
5. tWPRE values depend on MR1 OP[2] respectively.
6. Minimum delay between "from command" and "to command" in clock cycle is calculated by dividing tRTP (in ns) by tCK (in ns) and rounding up to the next integer: Minimum
delay [cycles] = roundup (tRTP [ns]/tCK [ns]).
7. Minimum delay between "from command" and "to command" in clock cycle is calculated by dividing tWR (in ns) by tCK (in ns) and rounding up to the next integer: Minimum
delay [cycles] = roundup (tWR [ns]/tCK [ns]).
115
200b: x32 LPDDR4 SDRAM
Auto Precharge
8. Minimum delay between "from command" and "to command" in clock cycle is calculated by dividing tRPpb (in ns) by tCK (in ns) and rounding up to the next integer: Minimum delay [cycles] = roundup (tRPpb [ns]/tCK [ns]).
9. Minimum delay between "from command" and "to command" in clock cycle is calculated by dividing tWTR (in ns) by tCK (in ns) and rounding up to the next integer: Minimum delay [cycles] = roundup (tWTR [ns]/tCK [ns]).
10. For READ w/AP the value is nRTP, which is defined in mode register 2.
11. For WRITE w/AP the value is nWR, which is defined in mode register 1.
Table 96: Timing Between Commands (PRECHARGE and AUTO PRECHARGE): DQ ODT is Enable
From Command To Command
READ w/AP
BL = 16
WRITE or WRITE w/AP
(different bank)
MASK-WR or MASK-WR w/AP
(different bank)
READ w/AP
BL = 32
WRITE or WRITE w/AP
(different bank)
MASK-WR or MASK-WR w/AP
(different bank)
Notes:
Minimum Delay Between
"From Command" and "To Command"
Unit
Notes
RL + RU (tDQSCK MAX/tCK) + BL/2
+ RD (tRPST) - ODTLon - RD (tODTon MIN/tCK) + 1
tCK
2, 3
RL + RU (tDQSCK MAX/tCK) + BL/2
- ODTLon - RD (tODTon MIN/tCK) + 1
tCK
2, 3
RL + RU (tDQSCK MAX/tCK) + BL/2
- ODTLon - RD (tODTon MIN/tCK) + 1
tCK
2, 3
RL + RU (tDQSCK MAX/tCK) + BL/2
- ODTLon - RD (tODTon MIN/tCK) + 1
tCK
2, 3
+ RD
(tRPST)
+ RD
(tRPST)
+ RD
(tRPST)
1. The rest of the timing about PRECHARGE and AUTO PRECHARGE is same as DQ ODT is
disable case.
2. After READ w/AP, seamless read operations to different banks are supported. READ,
WRITE, and MASK-WR operations may not be truncated or interrupted.
3. tRPST values depend on MR1 OP[7] respectively.
RAS Lock Function
READ with AUTO PRECHARGE or WRITE/MASK WRITE with AUTO PRECHARGE commands may be issued after tRCD has been satisfied. The LPDDR4 SDRAM RAS lockout
feature will schedule the internal precharge to assure that tRAS is satisfied. tRC needs to
be satisfied prior to issuing subsequent ACTIVATE commands to the same bank.
The figure below shows example of RAS lock function.
116
200b: x32 LPDDR4 SDRAM
Auto Precharge
Figure 66: Command Input Timing with RAS Lock
T0
T1
T2
T3
RA
RA
BA0
RA
RA
T4
T19
T20
T21
T22
T23
T24
T25
T31
T32
T38
T39
T47
T48
Ta0
Ta1
Ta2
Ta3
Ta4
RA
RA
BA0
RA
Ta5
CK_c
CK_t
CKE
CS
CA
Command
ACTIVATE-1
ACTIVATE-2
Valid
DES
DES
BA0
CA
RDA-1
tRCD
CA
CAS-2
DES
= 20nCK
DES
DES
8nCK
DES
DES
DES
DES
DES
DES
DES
ACTIVATE-1
RA
ACTIVATE-2
nRTP = 8nCK
tRAS
tRC
Don’t Care
Notes:
1. tCK (AVG) = 0.938ns, Data rate = 2133 Mbps, tRCD (MIN) = MAX (18ns, 4nCK), tRAS (MIN)
= MAX (42ns, 3nCK), nRTP = 8nCK, BL = 32.
2. tRCD = 20nCK comes from roundup (18ns/0.938ns).
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Delay Time From WRITE-to-READ with Auto Precharge
In the case of WRITE command followed by READ with AUTO PRECHARGE, controller
must satisfy tWR for the WRITE command before initiating the device internal auto-precharge. It means that (tWTR + nRTP) should be equal or longer than (tWR) when BL setting is 16, as well as (tWTR + nRTP + 8nCK) should be equal or longer than (tWR) when
BL setting is 32. Refer to the following figure for details.
Figure 67: Delay Time From WRITE-to-READ with Auto Precharge
T0
T1
T2
T3
BL
BA0
CA
CA
CA
T4
Ta0
Ta1
Ta2
Tb0
Tb1
Tb2
Tb3
Tc0
Tc1
Tc2
Tc3
Tc4
Tc5
Tc6
Td0
Td1
Td2
DES
DES
DES
DES
Td3
Td4
CK_c
CK_t
CKE
CS
CA
Command
WRITE-1
CAS-2
Valid
DES
DES
WL
DES
DES
DES
DES
DES
DES
DES
DES
BA0
CA
RDA-1
CAS-2
tWTR
BL/2 + 1 clock
CA
DES
DES
nRTP
tWR
Don’t Care
Notes:
1. Burst length at read = 16.
117
200b: x32 LPDDR4 SDRAM
REFRESH Command
2. DES commands are shown for ease of illustration; other commands may be valid at
these times.
REFRESH Command
The REFRESH command is initiated with CS HIGH, CA0 LOW, CA1 LOW, CA2 LOW, CA3
HIGH and CA4 LOW at the first rising edge of clock. Per bank REFRESH is initiated with
CA5 LOW at the first rising edge of the clock. The all-bank REFRESH is initiated with
CA5 HIGH at the first rising edge of clock.
A per bank REFRESH command (REFpb) is performed to the bank address as transferred on CA0, CA1, and CA2 on the second rising edge of the clock. Bank address BA0 is
transferred on CA0, bank address BA1 is transferred on CA1, and bank address BA2 is
transferred on CA2. A per bank REFRESH command (REFpb) to the eight banks can be
issued in any order. For example, REFpb commands may be issued in the following order: 1-3-0-2-4-7-5-6. After the eight banks have been refreshed using the per bank REFRESH command, the controller can send another set of per bank REFRESH commands
in the same order or a different order. One possible order can be a sequential round
robin: 0-1-2-3-4-5-6-7. It is illegal to send a per bank REFRESH command to the same
bank unless all eight banks have been refreshed using the per bank REFRESH command. The count of eight REFpb commands starts with the first REFpb command after
a synchronization event.
The bank count is synchronized between the controller and the device by resetting the
bank count to zero. Synchronization can occur upon reset procedure or at every exit
from self refresh. The REFab command also synchronizes the counter between the controller and the device to zero. The device can be placed in self refresh, or a REFab command can be issued at any time without cycling through all eight banks using per bank
REFRESH command. After the bank count is synchronized to zero, the controller can
issue per bank REFRESH commands in any order, as described above.
A REFab command issued when the bank counter is not zero will reset the bank counter
to zero and the device will perform refreshes to all banks as indicated by the row counter. If another REFRESH command (REFab or REFpb) is issued after the REFab command then it uses an incremented value of the row counter.
The table below shows examples of both bank and refresh counter increment behavior.
Table 97: Bank and Refresh Counter Increment Behavior
#
Command
BA0
0
BA1
BA2
Refresh
Bank #
Reset, SRX, or REFab
Bank
Counter #
Ref. Conter #
(Row Address #)
To 0
–
n
1
REFpb
0
0
0
0
0 to 1
2
REFpb
0
0
1
1
1 to 2
3
REFpb
0
1
0
2
2 to 3
4
REFpb
0
1
1
3
3 to 4
5
REFpb
1
0
0
4
4 to 5
6
REFpb
1
0
1
5
5 to 6
7
REFpb
1
1
0
6
6 to 7
8
REFpb
1
1
1
7
7 to 0
118
200b: x32 LPDDR4 SDRAM
REFRESH Command
Table 97: Bank and Refresh Counter Increment Behavior (Continued)
#
Command
BA0
BA1
BA2
Refresh
Bank #
Bank
Counter #
Ref. Conter #
(Row Address #)
9
REFpb
1
1
0
6
0 to 1
n+1
10
REFpb
1
1
1
7
1 to 2
11
REFpb
0
0
1
1
2 to 3
12
REFpb
0
1
1
3
3 to 4
13
REFpb
1
0
1
5
4 to 5
14
REFpb
0
1
0
2
5 to 6
15
REFpb
0
0
0
0
6 to 7
16
REFpb
1
0
0
4
7 to 0
17
REFpb
0
0
0
0
0 to 1
18
REFpb
0
0
1
1
1 to 2
19
REFpb
0
1
0
2
2 to 3
20
REFab
V
V
V
0 to 7
To 0
n+2
21
REFpb
1
1
0
6
0 to 1
n+3
22
REFpb
1
1
1
7
1 to 2
n+2
Snip
A bank must be idle before it can be refreshed. The controller must track the bank being
refreshed by the per bank REFRESH command.
The REFpb command must not be issued to the device until the following conditions
have been met:
•
•
•
•
tRFCab
has been satisfied after the prior REFab command
has been satisfied after the prior REFpb command
tRP has been satisfied after the prior PRECHARGE command to that bank
tRRD has been satisfied after the prior ACTIVATE command (for example, after activating a row in a different bank than the one affected by the REFpb command)
tRFCpb
The target bank is inaccessible during per bank REFRESH cycle time (tRFCpb). However, other banks within the device are accessible and can be addressed during the cycle.
During the REFpb operation, any of the banks other than the one being refreshed can
be maintained in an active state or accessed by a READ or a WRITE command. When
the per bank REFRESH cycle has completed, the affected bank will be in the idle state.
After issuing REFpb, the following conditions must be met:
•
•
•
•
tRFCpb
must be satisfied before issuing a REFab command
must be satisfied before issuing an ACTIVATE command to the same bank
tRRD must be satisfied before issuing an ACTIVATE command to a different bank
tRFCpb must be satisfied before issuing another REFpb command
tRFCpb
An all-bank REFRESH command (REFab) issues a REFRESH command to every bank in
a channel. All banks must be idle when REFab is issued (for example, by issuing a PRECHARGE ALL command prior to issuing an all-bank REFRESH command). The REFab
119
200b: x32 LPDDR4 SDRAM
REFRESH Command
command must not be issued to the device until the following conditions have been
met:
• tRFCab has been satisfied following the prior REFab command
• tRFCpb has been satisfied following the prior REFpb command
• tRP has been satisfied following the prior PRECHARGE command
When an all-bank REFRESH cycle has completed, all banks will be idle. After issuing REFab:
• RFCab latency must be satisfied before issuing an ACTIVATE command,
• RFCab latency must be satisfied before issuing a REFab or REFpb command
Table 98: REFRESH Command Timing Constraints
Symbol
Minimum
Delay From...
To
tRFCab
REFab
REFab
Notes
ACTIVATE command to any bank
REFpb
tRFCpb
REFpb
REFab
ACTIVATE command to same bank as REFpb
REFpb
tRRD
REFpb
ACTIVATE command to a different bank than REFpb
ACTIVATE
REFpb
1
ACTIVATE command to a different bank than the prior ACTIVATE command
1. A bank must be in the idle state before it is refreshed; therefore, REFab is prohibited
following an ACTIVATE command. REFpb is supported only if it affects a bank that is in
the idle state.
Note:
Figure 68: All-Bank REFRESH Operation
T0
T1
T2
T3
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Tc0
Tc1
Tc2
Tc3
CK_c
CK_t
CKE
CS
CA Valid
t
Command PRECHARGE
ALL bank
Valid Valid
Valid Valid
Valid
DES
DES
t
RPab
DES
DES
All bank
REFRESH
DES
DES
Valid Valid
t
RFCab
DES
DES
All bank
REFRESH
DES
DES
RFCab
DES
DES
Any command
DES
Don’t Care
120
200b: x32 LPDDR4 SDRAM
REFRESH Command
Figure 69: Per Bank REFRESH Operation
T0
T1
T2
T3
Ta0
Ta1
Ta2
Valid
BA0
Ta3
Ta4
Tb0
Tb1
Tb2
Valid
BA1
Tb3
Tb4
Tc0
Tc1
Tc2
Tc3
Tc4
Valid
BA1
Valid Valid
Tc5
Tc6
DES
DES
CK_c
CK_t
CKE
CS
CA Valid
Valid
t
Command PRECHARGE
ALL bank
DES
t
RPab
DES
DES
Per bank
REFRESH
DES
t
RFCpb
DES
DES
Per bank
REFRESH
DES
RFCpb
DES
DES
ACTIVATE-1 ACTIVATE-2
Don’t Care
Notes:
1. In the beginning of this example, the REFpb bank is pointing to bank 0.
2. Operations to banks other than the bank being refreshed are supported during the
tRFCpb period.
In general, a REFRESH command needs to be issued to the device regularly every tREFI
interval. To allow for improved efficiency in scheduling and switching between tasks,
some flexibility in the absolute refresh interval is provided. A maximum of eight REFRESH commands can be postponed during operation of the device, but at no point in
time are more than a total of eight REFRESH commands allowed to be postponed. And
a maximum number of pulled-in or postponed REF command is dependent on refresh
rate. It is described in the table below. In the case where eight REFRESH commands are
postponed in a row, the resulting maximum interval between the surrounding REFRESH commands is limited to 9 × tREFI. A maximum of eight additional REFRESH
commands can be issued in advance (pulled in), with each one reducing the number of
regular REFRESH commands required later by one. Note that pulling in more than eight
REFRESH commands in advance does not reduce the number of regular REFRESH
commands required later; therefore, the resulting maximum interval between two surrounding REFRESH commands is limited to 9 × tREFI. At any given time, a maximum of
16 REFRESH commands can be issued within 2 × tREFI.
Self refresh mode may be entered with a maximum of eight REFRESH commands being
postponed. After exiting self refresh mode with one or more REFRESH commands postponed, additional REFRESH commands may be postponed to the extent that the total
number of postponed REFRESH commands (before and after self refresh) will never exceed eight. During self refresh mode, the number of postponed or pulled-in REFRESH
commands does not change.
And for per bank refresh, a maximum of 8 x 8 per bank REFRESH commands can be
postponed or pulled in for scheduling efficiency. At any given time, a maximum of 2 x 8
x 8 per bank REFRESH commands can be issued within 2 x tREFI.
121
200b: x32 LPDDR4 SDRAM
REFRESH Command
Table 99: Legacy REFRESH Command Timing Constraints
MR4
OP[2:0]
000b
Refresh Rate
Max. No. of
pulled-in or
postponed REFab
Max. Interval
between
Two REFab
Max. No. of
REFab1
Per Bank Refresh
Low Temp. Limit
N/A
N/A
N/A
N/A
001b
4×
tREFI
010b
2 × tREFI
011b
1 × tREFI
100b
tREFI
0.5 ×
101b
0.25 ×
tREFI
110b
0.25 × tREFI
111b
High Temp. Limit
Note:
8
9×4×
tREFI
16
1/8 of REFab
8
9 × 2 × tREFI
16
1/8 of REFab
8
9 × tREFI
16
1/8 of REFab
8
9 × 0.5 ×
tREFI
16
1/8 of REFab
8
9 × 0.25 ×
tREFI
16
1/8 of REFab
8
9 × 0.25 × tREFI
16
1/8 of REFab
N/A
N/A
N/A
N/A
1. Maximum number of REFab within MAX(2 × tREFI × refresh rate multiplier, 16 × tRFC).
Table 100: Modified REFRESH Command Timing Constraints
MR4
OP[2:0]
Refresh Rate
Max. No. of
pulled-in or
postponed REFab
000B
LOW Temp. Limit
001B
4 × tREFI
010B
2 × tREFI
011B
tREFI
1×
Max. Interval
between
Two REFab
Max. No. of
REFab1
N/A
N/A
N/A
N/A
2
3 × 4 × tREFI
4
1/8 of REFab
4
5 × 2 × tREFI
8
1/8 of REFab
16
1/8 of REFab
8
tREFI
9×
tREFI
tREFI
Per Bank Refresh
100B
0.5 ×
8
9 × 0.5 ×
16
1/8 of REFab
101B
0.25 × tREFI
8
9 × 0.25 × tREFI
16
1/8 of REFab
110B
0.25 × tREFI
8
9 × 0.25 × tREFI
16
1/8 of REFab
111B
HIGH Temp. Limit
N/A
N/A
N/A
N/A
Notes:
1. For any thermal transition phase where refresh mode is transitioned to either 2 × tREFI
or 4 × tREFI, LPDDR4 devices will support the previous postponed refresh requirement
provided the number of postponed refreshes is monotonically reduced to meet the new
requirement. However, the pulled-in REFRESH commands in previous thermal phase are
not applied in new thermal phase. Entering new thermal phase the controller must
count the number of pulled-in REFRESH commands as zero, regardless of remaining
pulled-in REFRESH commands in previous thermal phase.
2. LPDDR4 devices are refreshed properly if memory controller issues REFRESH commands
with same or shorter refresh period than reported by MR4 OP[2:0]. If shorter refresh period is applied, the corresponding requirements from Table apply. For example, when
MR4 OP[2:0] = 001b, controller can be in any refresh rate from 4 × tREFI to 0.25 × tREFI.
When MR4 OP[2:0] = 010b, the only prohibited refresh rate is 4 × tREFI.
122
200b: x32 LPDDR4 SDRAM
Refresh Requirement
Figure 70: Postponing REFRESH Commands (Example)
tREFI
9 tREFI
t
tRFC
8 REFRESH commands postponed
Figure 71: Pulling In REFRESH Commands (Example)
tREFI
9 tREFI
t
tRFC
8 REFRESH commands pulled in
Refresh Requirement
Between the SRX command and SRE command, at least one extra REFRESH command
is required. After the SELF REFRESH EXIT command, in addition to the normal REFRESH command at tREFI interval, the device requires a minimum of one extra REFRESH command prior to the SELF REFRESH ENTRY command.
Table 101: Refresh Requirement Parameters
Density (per channel)
Parameter
Symbol
12Gb
16Gb
Unit
–
8
TBD
TBD
–
(tREFW):
tREFW
32
TBD
TBD
ms
Refresh window (tREFW):
1/2 rate refresh
tREFW
16
TBD
TBD
ms
Number of banks per channel
Refresh window
TCASE ≤ 85°C
2Gb
3Gb
123
4Gb
6Gb
8Gb
200b: x32 LPDDR4 SDRAM
SELF REFRESH Operation
Table 101: Refresh Requirement Parameters (Continued)
Density (per channel)
Parameter
Refresh window
1/4 rate refresh
Symbol
(tREFW):
Required number of REFRESH
commands in tREFW window
Average refresh interval REFab
REFpb
12Gb
16Gb
Unit
8
TBD
TBD
ms
R
8192
TBD
TBD
–
tREFI
3.904
TBD
TBD
μs
tREFIpb
488
TBD
TBD
ns
tREFW
2Gb
3Gb
4Gb
6Gb
8Gb
REFRESH cycle time (all banks)
tRFCab
130
180
280
TBD
TBD
ns
REFRESH cycle time (per bank)
tRFCpb
60
90
140
TBD
TBD
ns
Notes:
1. Refresh for each channel is independent of the other channel on the die, or other channels in a package. Power delivery in the user’s system should be verified to make sure
the DC operating conditions are maintained when multiple channels are refreshed simultaneously.
2. Self refresh abort feature is available for higher density devices starting with 6Gb density per channel device and tXSR_abort(min) is defined as tRFCpb + 17.5ns.
SELF REFRESH Operation
Self Refresh Entry and Exit
The SELF REFRESH command can be used to retain data in the device without external
REFRESH commands. The device has a built-in timer to accommodate SELF REFRESH
operation. Self refresh is entered by the SELF REFRESH ENTRY command defined by
having CS HIGH, CA0 LOW, CA1 LOW, CA2 LOW, CA3 HIGH, CA4 HIGH, and CA5 valid
(valid meaning that it is at a logic level HIGH or LOW) for the first rising edge, and CS
LOW, CA0 valid, CA1 valid, CA2 valid, CA3 valid, CA4 valid, and CA5 valid at the second
rising edge of clock. The SELF REFRESH command is only allowed when READ DATA
burst is completed and the device is in the idle state.
During self refresh mode, external clock input is needed and all input pins of the device
are activated. The device can accept the following commands: MRR-1, CAS-2, DES, SRX,
MPC, MRW-1, and MRW-2, except PASR bank/segment setting.
The device can operate in self refresh mode within the standard and elevated temperature ranges. It also manages self refresh power consumption when the operating temperature changes: lower at low temperatures and higher at high temperatures.
For proper SELF REFRESH operation, power supply pins (VDD1, V DD2, and V DDQ) must
be at valid levels. V DDQ can be turned off during self refresh with power-down after
tCKELCK is satisfied. (Refer to the Self Refresh Entry/Exit Timing with Power-Down Entry/Exit figure.) Prior to exiting self refresh with power-down, V DDQ must be within
specified limits. The minimum time that the device must remain in self refresh mode is
tSR (MIN). After self refresh exit is registered, only MRR-1, CAS-2, DES, MPC, MRW-1,
and MRW-2 except PASR bank/segment setting are allowed until tXSR is satisfied.
The use of self refresh mode introduces the possibility that an internally timed refresh
event can be missed when self refresh exit is registered. Upon exit from self refresh, it is
124
200b: x32 LPDDR4 SDRAM
SELF REFRESH Operation
required that at least one REFRESH command (8 per-bank or 1 all-bank) is issued before entry into a subsequent self refresh. This REFRESH command is not included in the
count of regular REFRESH commands required by the tREFI interval, and does not
modify the postponed or pulled-in refresh counts; the REFRESH command does count
toward the maximum refreshes permitted within 2 × tREFI.
Figure 72: Self Refresh Entry/Exit Timing
T0
T1
T2
T3
T4
T5
T6
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Tb0
Tb1
Tb2
Tb3 Tb4
Tb5
Tb6
Tb7
DES Any command Any command DES
DES
CK_c
CK_t
CKE
CS
CA
Valid Valid
Valid Valid
Valid Valid Valid Valid
t SR
Command
DES
SELF REFRESH DES
ENTRY
Enter self refresh
DES
DES
t XSR
DES
DES
DES
SELF REFRESH DES
EXIT
DES
DES
Exit self refresh
Don’t Care
Notes:
1. MRR-1, CAS-2, DES, SRX, MPC, MRW-1, and MRW-2 commands (except PASR bank/
segment setting) are allowed during self refresh.
2. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Power-Down Entry and Exit During Self Refresh
Entering/exiting power-down mode is allowed during self refresh mode. The related
timing parameters between self refresh entry/exit and power-down entry/exit are
shown below.
125
200b: x32 LPDDR4 SDRAM
SELF REFRESH Operation
Figure 73: Self Refresh Entry/Exit Timing with Power-Down Entry/Exit
T0
T1
T2
T3
Ta0
Tb0
Tb1
Tc0
Td1
Te0
Tf0
Tf1
Tg0
Tg1
Th0
Tk0
Tk1
Tk2
Tk3
CK_c
Note 2
CK_t
t
CKE
t
t
CKELCK
CKCKEH
CKE
t
CSCKE tCKELCS
t
CSCKEH
t
CKEHCS
CS
t
ESCKE
t
CMDCKE
CA
Valid
Valid
Valid
t
XP
Valid
Valid
Valid
t SR
Note 3
Command
SELFWrite-2
REFRESH
Any command
MR
ENTRY
Enter self refresh
DES
DES
DES
SELF REFRESH
DES
EXIT
Exit self refresh
Don’t Care
Notes:
1. MRR-1, CAS-2, DES, SRX, MPC, MRW-1, and MRW-2 commands (except PASR bank/
segment setting) are allowed during self refresh.
2. Input clock frequency can be changed, or the input clock can be stopped, or floated after tCKELCK satisfied and during power-down, provided that upon exiting power-down,
the clock is stable and within specified limits for a minimum of tCKCKEH of stable clock
prior to power-down exit and the clock frequency is between the minimum and maximum specified frequency for the speed grade in use.
3. Two clock command for example.
Command Input Timing After Power-Down Exit
Command input timings after power-down exit during self refresh mode are shown below.
126
200b: x32 LPDDR4 SDRAM
SELF REFRESH Operation
Figure 74: Command Input Timings after Power-Down Exit During Self Refresh
T0
T1
T2
T3
Ta0
Tb0
Tb1
Tc0
Td1
Te0
Tf0
Tf1
Tg0
Tg1
Th0
Tk0
Tk1
Tk2
Tk3
CK_c
Note 2
CK_t
t
CKE
t
t
CKELCK
CKCKEH
CKE
t
CSCKE tCKELCS
t
CSCKEH
t
CKEHCS
CS
t
ESCKE
t
CMDCKE
CA
Valid
Valid
Valid
t
XP
Valid
Valid
Valid
t SR
Note 3
Note 3
Command
SELFWrite-2
REFRESH
Any command
MR
ENTRY
Enter self refresh
DES
DES
DES
Any command
DES
Exit self refresh
Don’t Care
Notes:
1. MRR-1, CAS-2, DES, SRX, MPC, MRW-1, and MRW-2 commands (except PASR bank/
segment setting) are allowed during self refresh.
2. Input clock frequency can be changed or the input clock can be stopped or floated after
tCKELCK satisfied and during power-down, provided that upon exiting power-down, the
clock is stable and within specified limits for a minimum of tCKCKEH of stable clock prior
to power-down exit and the clock frequency is between the minimum and maximum
specified frequency for the speed grade in use.
3. Two clock command for example.
Self Refresh Abort
If MR4 OP[3] is enabled, the device aborts any ongoing refresh during self refresh exit
and does not increment the internal refresh counter. The controller can issue a valid
command after a delay of tXSR_abort instead of tXSR.
The value of tXSR_abort (MIN) is defined as tRFCpb + 17.5ns.
Upon exit from self refresh mode, the device requires a minimum of one extra refresh
(eight per bank or one for the entire bank) before entering a subsequent self refresh
mode. This requirement remains the same irrespective of the setting of the MR bit for
self refresh abort.
Self refresh abort feature is valid for 6Gb density per channel and larger densities only.
MRR, MRW, MPC Commands During tXSR, tRFC
MODE REGISTER READ (MRR) command, MODE REGISTER WRITE (MRW) command
and MULTI PURPOSE command (MPC) can be issued during tXSR period.
127
200b: x32 LPDDR4 SDRAM
SELF REFRESH Operation
Figure 75: MRR, MRW, and MPC Commands Issuing Timing During tXSR
T0
T1
T2
T3
T4
T5
T6
T7
Ta0 Ta1 Ta2 Ta3 Ta4 Ta5 Ta6 Tb0 Tb1 Tb2 Tb3 Tb4 Tb5
CK_c
CK_t
CKE
CS “H” for case 2
CS
CA
Valid Valid Valid Valid
Valid Valid Valid Valid
Valid Valid
t
MRD
Command DES
(Case 1)
SRX
MPC
(2 clock command)
DES
DES
DES DES
MRW-1
MRW-2
DES
DES
DES
Note 3
DES Any command
t
MRD
Command DES
(Case 2)
SRX
MPC
(4 clock command)
CAS-2
DES DES
MRW-1
MRW-2
DES
DES
DES
Note 3
DES
Any command
XSR Note 2
t
Don’t Care
Notes:
1. MPC and MRW commands are shown. Any combination of MRR, MRW, and MPC is allowed during tXSR period.
2. "Any command" includes MRR, MRW, and all MPC commands.
MRR, MRW, and MPC can be issued during tRFC period.
128
200b: x32 LPDDR4 SDRAM
SELF REFRESH Operation
Figure 76: MRR, MRW, and MPC Commands Issuing Timing During tRFC
T0
T1
T2
T3
T4
T5
T6
T7
Ta0 Ta1 Ta2 Ta3 Ta4 Ta5 Ta6 Tb0 Tb1 Tb2 Tb3 Tb4 Tb5
CK_c
CK_t
CKE
CS “H” for case 2
CS
CA
Valid Valid Valid Valid
Valid Valid Valid Valid
Valid Valid
t
MRD
Command DES REF all bank
(Case 1)
MPC
(2 clock command)
DES
DES
DES DES
MRW-1
MRW-2
DES
DES
DES
Note 3
DES Any command
t
MRD
Command DES REF all bank
(Case 2)
MPC
(4 clock command)
CAS-2
DES DES
MRW-1
MRW-2
DES
DES
DES
Note 3
DES
Any command
RFCab Note 2
t
Don’t Care
Notes:
1. MPC and MRW commands are shown. Any combination of MRR, MRW, and MPC is allowed during tRFCab or tRFCpb period.
2. REFRESH cycle time depends on REFRESH command. In the case of per bank REFRESH
command issued, REFRESH cycle time will be tRFCpb.
3. "Any command" includes MRR, MRW, and all MPC commands.
129
200b: x32 LPDDR4 SDRAM
Power-Down Mode
Power-Down Mode
Power-Down Entry and Exit
Power-down is asynchronously entered when CKE is driven LOW. CKE must not go LOW
while the following operations are in progress:
•
•
•
•
•
•
•
•
Mode register read
Mode register write
Read
Write
VREF(CA) range and value setting via MRW
VREF(DQ) range and value setting via MRW
Command bus training mode entering/exiting via MRW
VRCG HIGH current mode entering/exiting via MRW
CKE can go LOW while any other operations such as row activation, precharge, auto
precharge, or refresh are in progress. The power-down I DD specification will not be applied until such operations are complete. Power-down entry and exit are shown below.
Entering power-down deactivates the input and output buffers, excluding CKE and RESET_n. To ensure that there is enough time to account for internal delay on the CKE signal path, CS input is required stable LOW level and CA input level is "Don’t Care" after
CKE is driven LOW, this timing period is defined as tCKELCS. Clock input is required after CKE is driven LOW, this timing period is defined as tCKELCK. CKE LOW will result in
deactivation of all input receivers except RESET_n after tCKELCK has expired. In powerdown mode, CKE must be held LOW; all other input signals except RESET_n are "Don't
Care." CKE LOW must be maintained until tCKE(MIN) is satisfied.
VDDQ can be turned off during power-down after tCKELCK is satisfied. Prior to exiting
power-down, V DDQ must be within its minimum/maximum operating range. No REFRESH operations are performed in power-down mode except self refresh power-down.
The maximum duration in non-self-refresh power-down mode is only limited by the refresh requirements outlined in the REFRESH command section.
The power-down state is asynchronously exited when CKE is driven HIGH. CKE HIGH
must be maintained until tCKE(MIN) is satisfied. A valid, executable command can be
applied with power-down exit latency tXP after CKE goes HIGH. Power-down exit latency is defined in the AC timing parameter table.
Clock frequency change or clock stop is inhibited during tCMDCKE, tCKELCK,
tCKCKEH, tXP, tMRWCKEL, and tZQCKE periods.
If power-down occurs when all banks are idle, this mode is referred to as idle powerdown. if power-down occurs when there is a row active in any bank, this mode is referred to as active power-down. And If power-down occurs when self refresh is in progress,
this mode is referred to as self refresh power-down in which the internal refresh is continuing in the same way as self refresh mode.
When CA, CK, and/or CS ODT is enabled via MR11 OP[6:4] and also via MR22 or CAODT pad setting, the rank providing ODT will continue to terminate the command bus
in all DRAM states including power-down when V DDQ is stable and within its minimum/maximum operating range.
130
200b: x32 LPDDR4 SDRAM
Power-Down Mode
The LPDDR4 DRAM cannot be placed in power-down state during start DQS interval
oscillator operation.
Figure 77: Basic Power-Down Entry and Exit Timing
T0
T1
Ta0
Tb0
Tb1
Tc0
CK_c
Tc1
Td0
Te0
Te1
Tf0
Tf1
Tg0
Th0
Th1
Th2
Th3 Tk0
Tk1
Tk2
Note 1
CK_t
tCKE
t CMDCKE
t CKELCK
t CKE
t CKCKEH
t XP
CKE
t CSCKE t CKELCS
t CSCKEH t CKEHCS
CS
CA Valid Valid
Command
Valid
Valid Valid
DES
DES
DES
Valid
DES
DES
Don’t Care
Note:
1. Input clock frequency can be changed or the input clock can be stopped or floated during power-down, provided that upon exiting power-down, the clock is stable and within
specified limits for a minimum of tCKCKEH of stable clock prior to power-down exit and
the clock frequency is between the minimum and maximum specified frequency for the
speed grade in use.
131
200b: x32 LPDDR4 SDRAM
Power-Down Mode
Figure 78: Read and Read with Auto Precharge to Power-Down Entry
T0
T1
T2
T3
T4
Ta0
Ta1
Ta2
Tb0
Tb1
Tb2
Tb3
Tc0
Tc1
DES
DES
DES
Tc2
Tc3
Tc4
Td0
Td1
CK_c
CK_t
CKE
See Note 2
CS
CA Valid Valid Valid Valid
Command
READ-1
CAS-2
DES
DES
DES
DES
DES
DES
t
RL
DES
DQSCK
DQS_c
DQS_t
t
t
DQ
DMI
RPRE
RPST
DO DO DO DO DO DO
n0 n1 n2 n13 n14 n15
Don’t Care
Notes:
1. CKE must be held HIGH until the end of the burst operation.
2. Minimum delay time from READ command or READ with AUTO PRECHARGE command
to falling edge of CKE signal is as follows:
When read postamble = 0.5nCK (MR1 OP[7] = [0]),
(RL × tCK) + tDQSCK(MAX) + ((BL/2) × tCK) + 1tCK
When read postamble = 1.5nCK (MR1 OP[7] = [1]),
(RL × tCK) + tDQSCK(MAX) + ((BL/2) × tCK) + 2tCK
132
200b: x32 LPDDR4 SDRAM
Power-Down Mode
Figure 79: Write and Mask Write to Power-Down Entry
T0
T1
T2
T3
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Tb0
Tb1
Tb2
Tc0
DES
DES
DES
DES
Tc1
Tc2
Td0
Td1
Td2
CK_c
CK_t
CKE
See Note 2
CS
CA
Command
Valid Valid Valid Valid
WRITE-1
MASK WRITE-1
CAS-2
DES
DES
DES
WL
t
DES
DES
t
DQSS
WPRE
t
DES
t
DQS2DQ
DI
n0
DI
n1
WPST
BL/2
t
WR
DI DI DI DI
n2 n13 n14 n15
Don’t Care
Notes:
1. CKE must be held HIGH until the end of the burst operation.
2. Minimum delay time from WRITE command or MASK WRITE command to falling edge
of CKE signal is as follows:
(WL × tCK) + tDQSS(MAX) + tDQS2DQ(MAX) + ((BL/2) × tCK) + tWR
3. This timing is applied regardless of DQ ODT disable/enable setting: MR11 OP[2:0].
4. This timing diagram only applies to the WRITE and MASK WRITE commands without auto precharge.
133
200b: x32 LPDDR4 SDRAM
Power-Down Mode
Figure 80: Write With Auto Precharge and Mask Write With Auto Precharge to Power-Down Entry
T0
T1
T2
T3
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Tb0
Tb1
Tb2
Tc0
Tc1
Tc2
Tc3
DES
DES
DES
DES
DES
DES
DES
Tc4
Td0
CK_c
CK_t
CKE
See Note 2
CS
CA Valid Valid Valid Valid
Command
WRITE-1
MASK WRITE-1
CAS-2
DES
DES
DES
DES
WL
t
DES
t
DQSS
WPRE
t
DES
t
DQS2DQ
DI
n0
DI
n1
WPST
BL/2
DI DI DI DI
n2 n13 n14 n15
Don’t Care
Notes:
1. CKE must be held HIGH until the end of the burst operation.
2. Delay time from WRITE with AUTO PRECHARGE command or MASK WRITE with AUTO
PRECHARGE command to falling edge of CKE signal is more than
(WL × tCK) + tDQSS(MAX) + tDQS2DQ(MAX) + ((BL/2) × tCK) + (nWR × tCK) + (2 × tCK)
3. This timing is applied regardless of DQ ODT disable/enable setting: MR11 OP[2:0].
134
200b: x32 LPDDR4 SDRAM
Power-Down Mode
Figure 81: Refresh Entry to Power-Down Entry
T0
T1
T2
Ta0
Ta1
Ta2
Ta3
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Tb6
Tb7
Tb8
Tb9
Tb10
CK_c
CK_t
CKE
t
CMDCKE
CS
CA Valid Valid
Command
REFRESH
DES
DES
Don’t Care
Note:
1. CKE must be held HIGH until tCMDCKE is satisfied.
Figure 82: ACTIVATE Command to Power-Down Entry
T0
T1
T2
T3
T4
Ta0
Ta1
Ta2
Ta3
Tb0
Tb1
Tb2
Tb4
Tb5
Tb6
Tb7
Tb8
Tb9
CK_c
CK_t
CKE
t
CMDCKE
CS
CA Valid Valid Valid Valid
Command ACTIVATE-1
ACTIVATE-2
DES
DES
Don’t Care
Note:
1. CKE must be held HIGH until tCMDCKE is satisfied.
135
200b: x32 LPDDR4 SDRAM
Power-Down Mode
Figure 83: PRECHARGE Command to Power-Down Entry
T0
T1
T2
Ta0
Ta1
Ta2
Ta3
Tb1
Tb2
Tb3
Tb4
Tb5
Tb6
Tb7
Tb8
Tb9
Tb10 Tb11
CK_c
CK_t
CKE
t
CMDCKE
CS
CA Valid Valid
Command PRECHARGE
DES
DES
Don’t Care
Note:
1. CKE must be held HIGH until tCMDCKE is satisfied.
136
200b: x32 LPDDR4 SDRAM
Power-Down Mode
Figure 84: Mode Register Read to Power-Down Entry
T0
T1
T2
T3
T4
Ta0
Ta1
Ta2
Tb1
Tb0
Tb2
Tb3
Tc0
Tc1
DES
DES
DES
Tc2
Tc3
Tc4
Td0
Td1
CK_c
CK_t
CKE
See Note 2
CS
CA Valid Valid Valid Valid
Command
MR READ-1
CAS-2
DES
DES
DES
DES
DES
DES
t
RL
DES
DQSCK
DQS_c
DQS_t
t
t
DQ
DMI
RPRE
RPST
DO DO DO DO DO DO
n0 n1 n2 n13 n14 n15
Don’t Care
Notes:
1. CKE must be held HIGH until the end of the burst operation.
2. Minimum delay time from MODE REGISTER READ command to falling edge of CKE signal is as follows:
When read postamble = 0.5nCK ( MR1 OP[7] = [0]),
(RL × tCK) + tDQSCK(MAX) + ((BL/2) × tCK) + 1tCK
When read postamble = 1.5nCK (MR1 OP[7] = [1]),
(RL × tCK) + tDQSCK(MAX) + ((BL/2) × tCK) + 2tCK
137
200b: x32 LPDDR4 SDRAM
Power-Down Mode
Figure 85: Mode Register Write to Power-Down Entry
T0
T1
T2
T3
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Tb6
Tb7
CK_c
CK_t
CKE
t
MRWCKEL
CS
CA Valid Valid Valid Valid
Command MR WRITE-1 MR WRITE-2
DES
DES
DES
Don’t Care
Notes:
1. CKE must be held HIGH until tMRWCKEL is satisfied.
2. This timing is the general definition for power-down entry after MODE REGISTER WRITE
command. When a MODE REGISTER WRITE command changes a parameter or starts an
operation that requires special timing longer than tMRWCKEL, that timing must be satisfied before CKE is driven LOW. Changing the VREF(DQ) value is one example, in this case
the appropriate tVREF-SHORT/MIDDLE/LONG must be satisfied.
138
200b: x32 LPDDR4 SDRAM
Power-Down Mode
Figure 86: MULTI PURPOSE Command for ZQCAL Start to Power-Down Entry
T0
T1
T2
Ta0
Ta1
Ta2
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Tb6
Tb7
Tb8
Tb9
Tb10 Tb11
CK_c
CK_t
CKE
t
ZQCKE
CS
CA Valid Valid
MPC
Command [ZQCAL
START]
DES
DES
ZQ calibration progresses
ZQ Cal Status
t
ZQCAL
Don’t Care
Note:
1. ZQ calibration continues if CKE goes LOW after tZQCKE is satisfied.
139
200b: x32 LPDDR4 SDRAM
Input Clock Stop and Frequency Change
Input Clock Stop and Frequency Change
Clock Frequency Change – CKE LOW
During CKE LOW, the device supports input clock frequency changes under the following conditions:
•
•
•
•
tCK(abs)
(MIN) is met for each clock cycle
Refresh requirements apply during clock frequency change
During clock frequency change, only REFab or REFpb commands may be executing
Any ACTIVATE or PRECHARGE commands have completed prior to changing the frequency
• Related timing conditions, tRCD and tRP, have been met prior to changing the frequency
• The initial clock frequency must be maintained for a minimum of tCKELCK after CKE
goes LOW
• The clock satisfies tCH(abs) and tCL(abs) for a minimum of tCKCKEH prior to CKE going HIGH
After the input clock frequency changes and CKE is held HIGH, additional MRW commands may be required to set the WR, RL, and so forth. These settings may require adjustment to meet minimum timing requirements at the target clock frequency.
Clock Stop – CKE LOW
During CKE LOW, the device supports clock stop under the following conditions:
•
•
•
•
CK_t and CK_c are don't care during clock stop
Refresh requirements apply during clock stop
During clock stop, only REFab or REFpb commands may be executing
Any ACTIVATE or PRECHARGE commands have completed prior to stopping the
clock
• Related timing conditions, tRCD and tRP, have been met prior to stopping the clock
• The initial clock frequency must be maintained for a minimum of tCKELCK after CKE
goes LOW
• The clock satisfies tCH(abs) and tCL(abs) for a minimum of tCKCKEH prior to CKE going HIGH
Clock Frequency Change – CKE HIGH
During CKE HIGH, the device supports input clock frequency change under the following conditions:
•
•
•
•
tCK(abs)
(MIN) is met for each clock cycle
Refresh requirements apply during clock frequency change
During clock frequency change, only REFab or REFpb commands may be executing
Any ACTIVATE, READ, WRITE, PRECHARGE, MODE REGISTER WRITE, or MODE
REGISTER READ commands (and any associated data bursts) have completed prior
to changing the frequency
• Related timing conditions (tRCD, tWR, tRP, tMRW, and tMRR) have been met prior to
changing the frequency
140
200b: x32 LPDDR4 SDRAM
Input Clock Stop and Frequency Change
• During clock frequency change, CS is held LOW
• The device is ready for normal operation after the clock satisfies tCH(abs) and
tCL(abs) for a minimum of 2 × tCK + tXP
After the input clock frequency is changed, additional MRW commands may be required to set the WR, RL, and so forth. These settings may need to be adjusted to meet
minimum timing requirements at the target clock frequency.
Clock Stop – CKE HIGH
During CKE HIGH, the device supports clock stop under the following conditions:
•
•
•
•
•
•
•
•
•
CK_t is held LOW and CK_c is held HIGH during clock stop
During clock stop, CS is held LOW
Refresh requirements apply during clock stop
During clock stop, only REFab or REFpb commands may be executing
Any ACTIVATE, READ, WRITE, MPC (WRITE-FIFO, READ-FIFO, READ DQ CALIBRATION), PRECHARGE, MODE REGISTER WRITE, or MODE REGISTER READ commands have completed, including any associated data bursts and extra 4 clock cycles
must be provided prior to stopping the clock
Related timing conditions (tRCD, tWR, tRP, tMRW, tMRR, tZQLAT, and so forth) have
been met prior to stopping the clock
READ with AUTO PRECHARGE and WRITE with AUTO PRECHARGE commands need
extra 4 clock cycles in addition to the related timing constraints, nWR and nRTP, to
complete the operations
REFab, REFpb, SRE, SRX, and MPC[ZQCAL START] commands are required to have
extra 4 clock cycles prior to stopping the clock
The device is ready for normal operation after the clock is restarted and satisfies
tCH(abs) and tCL(abs) for a minimum of 2 × tCK + tXP
141
200b: x32 LPDDR4 SDRAM
MODE REGISTER READ Operation
MODE REGISTER READ Operation
The MODE REGISTER READ (MRR) command is used to read configuration and status
data from the device registers. The MRR command is initiated with CS and CA[5:0] in
the proper state as defined by the Command Truth Table. The mode register address
operands (MA[5:0]) enable the user to select one of 64 registers. The mode register contents are available on the first four UI data bits of DQ[7:0] after RL × tCK + tDQSCK +
tDQSQ following the MRR command. Subsequent data bits contain valid but undefined
content. DQS is toggled for the duration of the MODE REGISTER READ burst. The MRR
has a command burst length of 16. MRR operation must not be interrupted.
Table 102: MRR
UI
0
1
2
3
4
5
6
7
8
9
10
DQ0
OP0
V
DQ1
OP1
V
DQ2
OP2
V
DQ3
OP3
V
DQ4
OP4
V
DQ5
OP5
V
DQ6
OP6
V
DQ7
OP7
12
13
14
15
V
DQ8–
DQ15
V
DMI0–
DMI1
V
Notes:
11
1. MRR data are extended to the first 4 UIs, allowing the LPDRAM controller to sample data easily.
2. DBI may apply or may not apply during normal MRR. It’s vendor-specific. If read DBI is
enable with MRS and vendor cannot support the DBI during MRR, DMI pin status should
be LOW.
3. The read preamble and postamble of MRR are the same as for a normal read.
142
200b: x32 LPDDR4 SDRAM
MODE REGISTER READ Operation
Figure 87: MODE REGISTER READ Operation
T0
T1
T2
T3
MA
CAn
CAn
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Tc0
Tc1
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA Valid
Command
MR READ-1
CAS-2
Valid Valid
DES
DES
Valid Valid
Any command Any command
t
MRR
t
RL
DQSCK
t
BL/2 = 8
RPRE
DQS_c
DQS_t
t
t
DQSQ
OP Code out
DQ7:0
RPST
Va- Va- Va- Valid lid lid lid
Va- Va- Va- Va- Va- Va- Va- Valid lid lid lid lid lid lid lid
DQ15:8
DMI1:0
Don’t Care
Notes:
1. Only BL = 16 is supported.
2. Only DESELECT is allowed during tMRR period.
3. There are some exceptions about issuing commands after tMRR. Refer to MRR/MRW
Timing Constraints Table for detail.
4. DBI is disable mode.
5. DES commands except tMRR period are shown for ease of illustration; other commands
may be valid at these times.
6. DQ/DQS: VSSQ termination
MRR After a READ and WRITE Command
After a prior READ command, the MRR command must not be issued earlier than BL/2
clock cycles, in a similar way WL + BL/2 + 1 + RU ( tWTR/tCK) clock cycles after a PRIOR
WRITE, WRITE with AP, MASK WRITE, MASK WRITE with AP, and MPC[WRITE-FIFO]
command in order to avoid the collision of READ and WRITE burst data on device internal data bus.
143
200b: x32 LPDDR4 SDRAM
MODE REGISTER READ Operation
Figure 88: READ-to-MRR Timing
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
T15
T16
T17
T18
T19
Valid
MA
CAn
CAn
T20
T21
T33
T34
T35
T36
T37
T43
DES
DES
DES
DES
DES
DES
DES
DES
T44
CK_c
CK_t
CS
CA
Command
READ-1
CAS-2
DES
DES
MRR-1
CAS-2
RL/2 = 14
BL/2
RL = 14
t
DQSCK
t
t
DQSCK
BL/2 = 8
BL/2 = 16
RPRE
t
RPST
DQS_c
DQS_t
t
DQSQ
t
DQSQ
Va- Valid lid
DQ7:0
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT
n0
n1
n2 n3
n26 n27 n28 n29 n30 n31
DQ15:8
DMI1:0
DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT DOUT Va- Va- Va- Va- Va- Van0
n1
n2 n3
n26 n27 n28 n29 n30 n31 lid lid lid lid lid lid
OP Code out
Don’t Care
Notes:
1. The minimum number of clock cycles from the burst READ command to the MRR command is BL/2.
2. Read BL = 32, MRR BL = 16, RL = 14, Preamble = Toggle, Postamble = 0.5nCK, DBI = Disable, DQ/DQS: VSSQ termination.
3. DOUT n = data-out to column n.
4. DES commands except tMRR period are shown for ease of illustration; other commands
may be valid at these times.
144
200b: x32 LPDDR4 SDRAM
MODE REGISTER READ Operation
Figure 89: WRITE-to-MRR Timing
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Tb0
Tb1
Tb2
Tc0
Tc1
Tc2
Tc3
Tc4
Valid
MA
CAn
CAn
Tc5
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
DES
DES
DES
DES
WL
DES
DES
DES
DES
DES
t
WPRE
MRR-1
t
BL/2 + 1 clock
t
DQS_c
DES
WTR
CAS-2
DES
t
MMR
WPST
DQS_t
t
DQ
DMI
DQS2DQ
DOUT DOUT DOUT DOUT DOUT DOUT
n0 n1 n12 n13 n14 n15
Don’t Care
Notes:
1.
2.
3.
4.
Write BL=16, Write postamble = 0.5nCK, DQ/DQS: VSSQ termination.
Only DES is allowed during tMRR period.
DOUT n = data-out to column n.
The minimum number of clock cycles from the BURST WRITE command to MRR command is WL + BL/2 + 1 + RU(tWTR/tCK).
t
5. WTR starts at the rising edge of CK after the last latching edge of DQS.
6. DES commands except tMRR period are shown for ease of illustration; other commands
may be valid at these times.
MRR After Power-Down Exit
Following the power-down state, an additional time, tMRRI, is required prior to issuing
the MODE REGISTER READ (MRR) command. This additional time (equivalent to
tRCD) is required in order to maximize power-down current savings by allowing more
power-up time for the MRR data path after exit from power-down mode.
145
200b: x32 LPDDR4 SDRAM
MODE REGISTER WRITE
Figure 90: MRR Following Power-Down
T0
Ta0
Tb0
Tb1
Tb2
Tc0
Tc1
Tc2
Tc3
Tc4
Tc5
Td1
Td0
Td2
Td3
Td4
Td5
Td6
Td7
Td8
Td9
CK_c
CK_t
t CKCKEH
CKE
t
t
XP
t
MRRI
MMR
CS
CA
Valid Valid
Command
DES
DES
Valid
Valid Valid
Any command Any command
DES
DES
DES
MA
MRR-1
DES
CAn
CAn
CAS-2
DES
DES
DES
Don't Care
1. Only DES is allowed during tMRR period.
2. DES commands except tMRR period are shown for ease of illustration; other commands
may be valid at these times.
Notes:
MODE REGISTER WRITE
The MODE REGISTER WRITE (MRW) writes configuration data to the mode registers.
The MRW command is initiated with CKE, CS, and CA[5:0] to valid levels at the rising
edge of the clock. The mode register address and the data written to it is contained in
CA[5:0] according to the Command Truth Table. The MRW command period is defined
by tMRW. Mode register WRITEs to read-only registers have no impact on the functionality of the device.
Figure 91: MODE REGISTER WRITE Timing
T0
T1
T2
T3
OPn
MA
OPn
OPn
T4
Ta0
Ta1
Ta2
Ta3
Ta4
OPn
MA
OPn
OPn
Ta5
Tb0
Tb1
Tb2
Tb3
Tb4
Valid
Valid
Valid
Valid
Tb5
Tb6
DES
DES
Tb7
CK_c
CK_t
CS
CA
tMRW
Command
MRW-1
MRW-2
DES
DES
MRW-1
tMRD
MRW-2
DES
DES
Any command
Any command
DES
Don’t Care
146
200b: x32 LPDDR4 SDRAM
MODE REGISTER WRITE
Mode Register Write States
MRW can be issued from either a bank-idle or a bank-active state. Certain restrictions
may apply for MRW from an active state.
Table 103: Truth Table for MRR and MRW
Current State
All banks idle
Bank(s) active
Command
Intermediate State
Next State
MRR
Reading mode register, all banks idle
All banks idle
MRW
Writing mode register, all banks idle
All banks idle
MRR
Reading mode register
Bank(s) active
MRW
Writing mode register
Bank(s) active
Table 104: MRR/MRW Timing Constraints: DQ ODT is Disable
Minimum Delay Between
"From Command" and "To Command"
Unit
MRR
tMRR
–
RD/RDA
tMRR
From Command To Command
MRR
WR/WRA/MWR/MWRA
MRW
RD/RDA
RL +
+ BL/2 -WL +
RD(tRPST)
RD with AUTO
PRECHARGE
WR/
MWR/
WR-FIFO
WR/MWR with
AUTO PRECHARGE
nCK
nCK
BL/2
nCK
nCK
tMRD
–
tXP
POWER-DOWN
EXIT
RD/
RD-FIFO/
READ DQ CAL
+
WL + 1 + BL/2 + RU(tWTR/tCK)
MRW
MRW
tWPRE
RL + RU(tDQSCK(MAX)/tCK) + BL/2 + 3
MRR
WR/WRA/MWR/
MWRA
–
RU(tDQSCK(MAX)/tCK)
+
tMRRI
–
RD/RDA
tMRD
–
WR/WRA/MWR/MWRA
tMRD
–
MRW
tMRW
MRW
RU(tDQSCK(MAX)/tCK)
–
RD(tRPST)
RL + BL/2 +
+
MAX(RU(7.5ns/tCK), 8nCK)
+
nCK
RL + BL/2 + RU(tDQSCK(MAX)/tCK) + RD(tRPST) +
MAX(RU(7.5ns/tCK), 8nCK) + nRTP - 8
nCK
WL + 1 + BL/2 + MAX(RU(7.5ns/tCK), 8nCK)
nCK
WL + 1 + BL/2 + MAX(RU(7.5ns/tCK), 8nCK) + nWR
nCK
147
Notes
200b: x32 LPDDR4 SDRAM
VREF Current Generator (VRCG)
Table 105: MRR/MRW Timing Constraints: DQ ODT is Enable
Minimum Delay Between
"From Command" and "To Command"
Unit
MRR
tMRR
–
RD/RDA
tMRR
–
From Command To Command
MRR
WR/WRA/MWR/MWRA
RD/RDA
RU(tDQSCK(MAX)/tCK)
+ BL/2 - ODTLon RL +
RD(tODTon(MIN)/tCK) + RD(tRPST) + 1
RL + RU(tDQSCK(MAX)/tCK) + BL/2 + 3
nCK
MRR
BL/2
nCK
WL + 1 + BL/2 + RU(tWTR/tCK)
nCK
tMRD
–
MRW
tXP
POWER-DOWN
EXIT
MRW
RD/
RD-FIFO/
READ DQ CAL
nCK
MRW
WR/WRA/MWR/
MWRA
+
tMRRI
–
RD/RDA
tMRD
–
WR/WRA/MWR/MWRA
tMRD
–
MRW
tMRW
RU(tDQSCK(max)/tCK)
MRW
–
RD(tRPST)
RL + BL/2 +
+
MAX(RU(7.5ns/tCK), 8nCK)
RD with AUTO
PRECHARGE
WR/
MWR/
WR-FIFO
WR/MWR with
AUTO PRECHARGE
Notes
+
nCK
RL + BL/2 + RU(tDQSCK(MAX)/tCK) + RD(tRPST) +
MAX(RU(7.5ns/tCK), 8nCK) + nRTP - 8
nCK
WL + 1 + BL/2 + MAX(RU(7.5ns/tCK), 8nCK)
nCK
WL + 1 + BL/2 + MAX(RU(7.5ns/tCK), 8nCK) + nWR
nCK
VREF Current Generator (VRCG)
LPDDR4 SDRAM V REF current generators (VRCG) incorporate a high current mode to
reduce the settling time of the internal V REF(DQ) and V REF(CA) levels during training and
when changing frequency set points during operation. The high current mode is enabled by setting MR13[OP3] = 1. Only DESELECT commands may be issued until
tVRCG_ENABLE is satisfied. tVRCG_ENABLE timing is shown below.
148
200b: x32 LPDDR4 SDRAM
VREF Current Generator (VRCG)
Figure 92: VRCG Enable Timing
T0
T1
T2
T3
T4
T5
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Tb6
DES
DES
Valid
Valid
Valid
Valid
DES
DES
Valid
Valid
Valid
Valid
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CKE
CS
CA
DES
MRW1 MRW1 MRW2 MRW2
Command
DES
VRCG enable: MR13 [OP3] = 1
t
Valid
Valid
VRCG_ENABLE
VRCG high current mode is disabled by setting MR13[OP3] = 0. Only DESELECT commands may be issued until tVRCG_DISABLE is satisfied. tVRCG_DISABLE timing is
shown below.
Figure 93: VRCG Disable Timing
T0
T1
T2
T3
T4
T5
Ta0
Ta1
Ta2
Ta3
Ta4
CA
DES
Valid
Valid
Valid
Valid
DES
DES
MRW1 MRW1 MRW2 MRW2
Command
DES
DES
DES
VRCG disable: MR13 [OP3] = 0
Ta5
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Tb6
DES
DES
Valid
Valid
Valid
Valid
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CKE
CS
Valid
t
Valid
VRCG_DISABLE
Note that LPDDR4 SDRAM devices support V FER(CA) and V REF(DQ) range and value
changes without enabling VRCG high current mode.
Table 106: VRCG Enable/Disable Timing
Parameter
Symbol
Min
Max
VREF high current mode enable time
tVRCG_ENABLE
–
200
ns
VREF high current mode disable time
tVRCG_DISABLE
–
100
ns
149
Unit
200b: x32 LPDDR4 SDRAM
VREF Training
VREF Training
VREF(CA) Training
The device's internal V REF(CA) specification parameters are operating voltage range, step
size, V REF step time, V REF full-range step time, and V REF valid level.
The voltage operating range specifies the minimum required V REF setting range for
LPDDR4 devices. The minimum range is defined by V REF,max and V REF,min.
Figure 94: VREF Operating Range (VREF,max, VREF,min)
VDD2
VIN(DC)max
VREF,max
VREF
range
VREF,min
VIN(DC)min
VSWING large
System variance
VSWING small
Total range
The V REF step size is defined as the step size between adjacent steps. However, for a given design, the device has one value for V REF step size that falls within the given range.
The V REF set tolerance is the variation in the V REF voltage from the ideal setting. This accounts for accumulated error over multiple steps. There are two ranges for V REF set tolerance uncertainty. The range of V REF set tolerance uncertainty is a function of the
number of steps n.
The V REF set tolerance is measured with respect to the ideal line that is based on the two
endpoints, where the endpoints are at the minimum and maximum V REF values for a
specified range.
150
200b: x32 LPDDR4 SDRAM
VREF Training
Figure 95: VREF Set-Point Tolerance and Step Size
VREF
level
Actual VREF
output
VREF set-point
tolerance
Straight line endpoint fit
VREF
step size
VREF step setting
The V REF increment/decrement step times are defined by tVREF_TIME-SHORT,
t
REF_TIME-MIDDLE, and VREF_TIME-LONG. The parameters are defined from TS to
TE as shown below, where TE is referenced to when the V REF voltage is at the final DC
level within the V REF valid tolerance (VREF,val_tol).
tV
The V REF valid level is defined by V REF,val_tol to qualify the step time TE (see the following
figures). This parameter is used to ensure an adequate RC time constant behavior of the
voltage level change after any V REF increment/decrement adjustment. This parameter is
only applicable for LPDDR4 component level validation/characterization.
tV
REF_TIME-SHORT is for a single step size increment/decrement change in the V REF
voltage.
tV
REF_TIME-MIDDLE is at least two stepsizes increment/decrement change within the
same V REF(CA) range in V REF voltage.
tV
REF_TIME-LONG is the time including up to V REF,min to V REF,max or V REF,max to V REF,min
change across the V REF(CA) range in V REF voltage.
TS is referenced to MRW command clock.
TE is referenced to V REF_val_tol.
151
200b: x32 LPDDR4 SDRAM
VREF Training
Figure 96: tVref for Short, Middle, and Long Timing Diagram
T0
T1
T2
T3
T4
T5
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Ta6
Ta7
Ta8
Ta9
Ta10
Ta11
Ta12
CK_c
CK_t
CKE
CS
CA
DES
MRW-1 MRW-1 MRW-2 MRW-2
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
Command
DES
VRFF(CA) value/range set
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
VREF_time – short/middle/long
New VREF setting
Updating VRFF(CA) setting
Old VREF setting
TS
VREF setting
adjustment
TE
The MRW command to the mode register bits are as follows;
MR12 OP[5:0] : V REF(CA) Setting
MR12 OP[6] : V REF(CA) Range
The minimum time required between two V REF MRW commands is tVREF_TIME-SHORT
for a single step and tVREF_TIME-MIDDLE for a full voltage range step.
Figure 97: VREF(CA) Single-Step Increment
VREF
voltage
VREF(DC)
VREF_val_tol
step size
t1
Time
152
200b: x32 LPDDR4 SDRAM
VREF Training
Figure 98: VREF(CA) Single-Step Decrement
VREF
voltage
t1
stepsize
VREF_val_tol
VREF(DC)
Time
Figure 99: VREF(CA) Full Step from VREF,min to VREF,max
VREF
voltage
VREF(DC)
VREF,max
VREF_val_tol
Full range
step
t1
VREF,min
Time
Figure 100: VREF(CA) Full Step from VREF,max to VREF,min
VREF VREF,max
voltage
Full range
step
t1
VREF_val_tol
VREF,min
VREF(DC)
Time
The following table contains the CA internal V REF specification that will be characterized at the component level for compliance.
153
200b: x32 LPDDR4 SDRAM
VREF Training
Table 107: Internal VREF(CA) Specifications
Min
Typ
Max
Unit
Notes
VREF(CA),max_r0
Symbol
VREF(CA) range-0 MAX operating point
–
–
30%
VDD2
1, 11
VREF(CA),min_r0
VREF(CA) range-0 MIN operating point
10%
–
–
VDD2
1, 11
VREF(CA),max_r1
VREF(CA) range-1 MAX operating point
–
–
42%
VDD2
1, 11
VREF(CA),min_r1
VREF(CA) range-1 MIN operating point
22%
–
–
VDD2
1, 11
VREF(CA) step size
0.30%
0.40%
0.50%
VDD2
2
VREF(CA) set tolerance
–1.00%
0.00%
1.00%
VDD2
3, 4, 6
–0.10%
0.00%
0.10%
VDD2
3, 5, 7
VREF(CA),step
VREF(CA),set_tol
tV
REF_TIME-SHORT
Parameter
–
–
100
ns
8
tV
REF_TIME-MIDDLE
–
–
200
ns
12
tV
REF_TIME-LONG
–
–
250
ns
9
tV
REF_time_weak
–
–
1
ms
13, 14
–0.10%
0.00%
0.10%
VDD2
10
VREF(CA)_val_tol
VREF(CA) step time
VREF(CA) valid tolerance
Notes:
1. VREF(CA) DC voltage referenced to VDD2(DC).
2. VREF(CA) step size increment/decrement range. VREF(CA) at DC level.
3. VREF(CA),new = VREF(CA),old + n × VREF(CA),step; n = number of steps; if increment, use "+"; if
decrement, use "–".
4. The minimum value of VREF(CA) setting tolerance = VREF(CA),new - 1.0% × VDD2. The maximum value of VREF(CA) setting tolerance = VREF(CA),new + 1.0% × VDD2. For n > 4.
5. The minimum value of VREF(CA) setting tolerance = VREF(CA),new - 0.10% × VDD2. The maximum value of VREF(CA) setting tolerance = VREF(CA),new + 0.10% × VDD2. For n < 4.
6. Measured by recording the minimum and maximum values of the VREF(CA) output over
the range, drawing a straight line between those points and comparing all other
VREF(CA) output settings to that line.
7. Measured by recording the minimum and maximum values of the VREF(CA) output across
four consecutive steps (n = 4), drawing a straight line between those points and comparing all other VREF(CA) output settings to that line.
8. Time from MRW command to increment or decrement one step size for VREF(CA) .
9. Time from MRW command to increment or decrement VREF,min to VREF,max or VREF,max to
VREF,min change across the VREF(CA) range in VREF voltage.
10. Only applicable for DRAM component level test/characterization purposes. Not applicable for normal mode of operation. VREF valid is to qualify the step times which will be
characterized at the component level.
11. DRAM range-0 or range-1 set by MR12 OP[6].
12. Time from MRW command to increment or decrement more than one step size up to a
full range of VREF voltage within the same VREF(CA) range.
13. Applies when VRCG high current mode is not enabled, specified by MR13 [OP3] = 0b.
14. tVREF_time_weak covers all VREF(CA) range and value change conditions are applied to
tV
REF_TIME-SHORT/MIDDLE/LONG.
154
200b: x32 LPDDR4 SDRAM
VREF Training
VREF(DQ) Training
The device's internal V REF(DQ) specification parameters are operating voltage range, step
size, V REF step tolerance, V REF step time and V REF valid level.
The voltage operating range specifies the minimum required V REF setting range for
LPDDR4 devices. The minimum range is defined by V REF,max and V REF,min.
Figure 101: VREF Operating Range (VREF,max, VREF,min)
VDDQ
VIN(DC)max
VREF,max
VREF
range
VREF,min
VIN(DC)min
VSWING large
System variance
VSWING small
Total range
The V REF step size is defined as the step size between adjacent steps. However, for a given design, the device has one value for V REF step size that falls within the given range.
The V REF set tolerance is the variation in the V REF voltage from the ideal setting. This accounts for accumulated error over multiple steps. There are two ranges for V REF set tolerance uncertainty. The range of V REF set tolerance uncertainty is a function of the
number of steps n.
The V REF set tolerance is measured with respect to the ideal line that is based on the two
endpoints, where the endpoints are at the minimum and maximum V REF values for a
specified range.
155
200b: x32 LPDDR4 SDRAM
VREF Training
Figure 102: VREF Set Tolerance and Step Size
VREF
level
Actual VREF
output
VREF set-point
tolerance
Straight line endpoint fit
VREF
step size
VREF step setting
The V REF increment/decrement step times are defined by tVREF_TIME-SHORT,
t
t
t
REF_TIME-MIDDLE and VREF_TIME-LONG. The VREF_TIME-SHORT, VREF_TIMEt
MIDDLE and VREF_TIME-LONG times are defined from TS to TE in the following figure
where TE is referenced to when the V REF voltage is at the final DC level within the V REF
valid tolerance (VREF,VAL_TOL).
tV
The V REF valid level is defined by V REF,VAL_TOL to qualify the step time TE (see the figure
below). This parameter is used to ensure an adequate RC time constant behavior of the
voltage level change after any V REF increment/decrement adjustment. This parameter is
only applicable for DRAM component level validation/characterization.
tV
REF_TIME-SHORT is for a single step size increment/decrement change in the V REF
voltage.
tV
REF_TIME-MIDDLE is at least two step sizes of increment/decrement change in the
VREF(DQ) range in the V REFvoltage.
tV
REF_TIME-LONG is the time including and up to the full range of V REF (MIN to MAX or
MAX to MIN) across the V REF(DQ) range in V REF voltage.
156
200b: x32 LPDDR4 SDRAM
VREF Training
Figure 103: VREF(DQ) Transition Time for Short, Middle, or Long Changes
T0
T1
T2
T3
T4
T5
Ta
Ta+1
Ta+2
Ta+3
Ta+4
Ta+5
Ta+6
Ta+7
Ta+8
Ta+9
T+10
T+11
T+12
CK_c
CK_t
CKE
CS
CA[5:0]
DES
MRW-1 MRW-1 MRW-2 MRW-2
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
Command
DES
VREF(DQ) value/range set
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
VREF_time – short/middle/long
VREF
Old VREF setting
Updating VREF(DQ) setting
New VREF setting
TS
VREF setting
adjustment
Notes:
TE
1. TS is referenced to MRW command clock.
2. TE is referenced to VREF,VAL_TOL.
The MRW command to the mode register bits are defined as:
MR14 OP[5:0]: V REF(DQ) setting
MR14 OP[6]: V REF(DQ) range
The minimum time required between two V REF MRW commands is tVREF_TIME-SHORT
for a single step and tVREF_TIME-MIDDLE for a full voltage range step.
Figure 104: VREF(DQ) Single-Step Size Increment
VREF
voltage
VREF(DC)
VREF_val_tol
step size
t1
Time
157
200b: x32 LPDDR4 SDRAM
VREF Training
Figure 105: VREF(DQ) Single-Step Size Decrement
VREF
voltage
t1
stepsize
VREF_val_tol
VREF(DC)
Time
Figure 106: VREF(DQ) Full Step from VREF,min to VREF,max
VREF
voltage
VREF(DC)
VREF,max
VREF_val_tol
Full range
step
t1
VREF,min
Time
Figure 107: VREF(DQ) Full Step from VREF,max to VREF,min
VREF VREF,max
voltage
Full range
step
t1
VREF_val_tol
VREF,min
VREF(DC)
Time
The following table contains the DQ internal V REF specification that will be characterized at the component level for compliance.
158
200b: x32 LPDDR4 SDRAM
VREF Training
Table 108: Internal VREF(DQ) Specifications
Min
Typ
Max
Unit
Notes
VREF(DQ),max_r0
Symbol
VREF MAX operating point
Range-0
–
–
30%
VDDQ
1, 11
VREF(DQ),min_r0
VREF MIN operating point
Range-0
10%
–
–
VDDQ
1, 11
VREF(DQ),max_r1
VREF MAX operating point
Range-1
–
–
42%
VDDQ
1, 11
VREF(DQ),min_r1
VREF MIN operating point
Range-1
22%
–
–
VDDQ
1, 11
VREF(DQ) step size
0.30%
0.40%
0.50%
VDDQ
2
VREF(DQ) set tolerance
–1.00%
0.00%
1.00%
VDDQ
3, 4, 6
–0.10%
0.00%
0.10%
VDDQ
3, 5, 7
–
–
100
ns
8
–
–
200
ns
12
REF_TIME-LONG
–
–
250
ns
9
tV
–
–
1
ms
13, 14
–0.10%
0.00%
0.10%
VDDQ
10
VREF(DQ),step
VREF(DQ),set_tol
tV
tV
REF_TIME-SHORT
Parameter
VREF(DQ) step time
REF_TIME-MIDDLE
tV
REF_time_weak
VREF(DQ),val_tol
VREF(DQ) valid tolerance
Notes:
1. VREF(DQ) DC voltage referenced to VDDQ(DC).
2. VREF(DQ) step size increment/decrement range. VREF(DQ) at DC level.
3. VREF(DQ),new = VREF(DQ),old + n × VREF(DQ),step; n = number of steps; if increment, use "+"; if
decrement, use "–".
4. The minimum value of VREF(DQ) setting tolerance = VREF(DQ),new - 1.0% × VDDQ. The maximum value of VREF(DQ) setting tolerance = VREF(DQ),new + 1.0% × VDDQ. For n > 4.
5. The minimum value of VREF(DQ)setting tolerance = VREF(DQ),new - 0.10% × VDDQ. The maximum value of VREF(DQ) setting tolerance = VREF(DQ),new + 0.10% × VDDQ. For n < 4.
6. Measured by recording the minimum and maximum values of the VREF(DQ) output over
the range, drawing a straight line between those points and comparing all other
VREF(DQ) output settings to that line.
7. Measured by recording the minimum and maximum values of the VREF(DQ) output across
four consecutive steps (n = 4), drawing a straight line between those points and comparing all other VREF(DQ) output settings to that line.
8. Time from MRW command to increment or decrement one step size for VREF(DQ) .
9. Time from MRW command to increment or decrement VREF,min to VREF,max or VREF,max to
VREF,min change across the VREF(DQ) Range in VREF(DQ) Voltage.
10. Only applicable for DRAM component level test/characterization purposes. Not applicable for normal mode of operation. VREF valid is to qualify the step times which will be
characterized at the component level.
11. DRAM range-0 or range-1 set by MR14 OP[6].
12. Time from MRW command to increment or decrement more than one step size up to a
full range of VREF voltage within the same VREF(DQ) range.
13. Applies when VRCG high current mode is not enabled, specified by MR13 [OP3] = 0.
14. tVREF_time_weak covers all VREF(DQ) Range and Value change conditions are applied to
tV
REF_TIME-SHOR/MIDDLE/LONG.
159
200b: x32 LPDDR4 SDRAM
Command Bus Training
Command Bus Training
Command Bus Training Mode
The command bus must be trained before enabling termination for high-frequency operation. The device provides an internal V REF(CA) that defaults to a level suitable for unterminated, low-frequency operation, but the V REF(CA) must be trained to achieve suitable receiver voltage margin for terminated, high-frequency operation.
The training mode described here centers the internal V REF(CA) in the CA data eye and at
the same time allows for timing adjustments of the CS and CA signals to meet setup/
hold requirements. Because it can be difficult to capture commands prior to training
the CA inputs, the training mode described here uses a minimum of external commands to enter, train, and exit the CA bus training mode.
The die has a bond-pad (ODT_CA) to control the command bus termination for multirank operation. Other mode register bits are provided to fine tune termination control
in a variety of system configuration. See On-Die Termination for more information.
The device uses frequency set points to enable multiple operating settings for the die.
The device defaults to FSP-OP[0] at power-up, which has the default settings to operate
in un-terminated, low-frequency environments. Prior to training, the termination
should be enabled for one die in each channel by setting MR13 OP[6] = 1b (FSP-WR[1])
and setting all other mode register bits for FSP-OP[1] to the desired settings for highfrequency operation. Upon training entry, the device will automatically switch to FSPOP[1] and use the high-frequency settings during training (See the Command Bus
Training Entry Timing figure for more information on FSP-OP register sets). Upon training exit, the device will automatically switch back to FSP-OP[0], returning to a "knowngood" state for unterminated, low-frequency operation.
To enter command bus training mode, issue a MRW-1 command followed by a MRW-2
command to set MR13 OP[0] = 1b (command bus training mode enabled).
After time tMRD, CKE may be set LOW, causing the device to switch to FSP-OP[1], and
completing the entry into command bus training mode.
A status DQS_t, DQS_c, DQ, and DMI are as noted below; the DQ ODT state will be followed by FREQUENCY SET POINT function except in the case of output pins.
•
•
•
•
•
DQS_t[0], DQS_c[0] become input pins for capturing DQ[6:0] levels by toggling.
DQ[5:0] become input pins for setting V REF(CA) level.
DQ[6] becomes an input pin for setting V REF(CA) range.
DQ[7] and DMI[0] become input pins, and their input level is valid or floating.
DQ[13:8] become output pins to feedback, capturing value via the command bus using the CS signal.
• DQS_t[1], DQS_c[1], DMI[1], and DQ[15:14] become output pins or are disabled,
meaning the device may be driven to a valid level or may be left floating.
At time tCAENT later, the device may change its V REF(CA) range and value using input
signals DQS_t[0], DQS_c[0], and DQ[6:0] from existing value that is set via MR12
OP[6:0]. The mapping between MR12 OP code and DQs is shown below. At least one
VREF(CA) setting is required before proceeding to the next training step.
160
200b: x32 LPDDR4 SDRAM
Command Bus Training
Table 109: Mapping MR12 Op Code and DQ Numbers
Mapping
MR12 OP code
OP6
OP5
OP4
OP3
OP2
OP1
OP0
DQ number
DQ6
DQ5
DQ4
DQ3
DQ2
DQ1
DQ0
The new V REF(CA) value must "settle" for time tVREFCA_Long before attempting to latch
CA information.
Note: If DQ ODT is enabled in MR11-OP[2:0], then the SDRAM will terminate the DQ
lanes during command bus training when entering V REF(CA) range and values on
DQ[6:0].
To verify that the receiver has the correct V REF(CA) setting, and to further train the CA eye
relative to clock (CK), values latched at the receiver on the CA bus are asynchronously
output to the DQ bus.
To exit command bus training mode, drive CKE HIGH, and after time tVREFCA_Long,
issue the MRW-1 command followed by the MRW-2 command to set MR13 OP[0] = 0b.
After time tMRW, the device is ready for normal operation. After training exit, the device
will automatically switch back to the FSP-OP registers that were in use prior to training.
Command bus training (CBT) may be executed from the idle or self refresh state. When
executing CBT within the self refresh state, the device must not be in a power-down
state (for example, CKE must be HIGH prior to training entry). CBT entry and exit is the
same, regardless of the state from which CBT is initiated.
Training Sequence for Single-Rank Systems
The sequence example shown here assumes an initial low-frequency, non-terminating
operating point training a high-frequency, terminating operating point. The bold text
shows high-frequency instructions. Any operating point may be trained from any
known good operating point.
1. Set MR13 OP[6] = 1b to enable writing to frequency set point 1 (FSP-WR[1]) (or
FSP-OP[0]).
2. Write FSP-WR[1] (or FSP-WR[0]) registers for all channels to set up high-frequency
operating parameters.
3. Issue MRW-1 and MRW-2 commands to enter command bus training mode.
4. Drive CKE LOW, and change CK frequency to the high-frequency operating
point.
5. Perform command bus training (VREF(CA), CS, and CA).
6. Exit training by driving CKE HIGH, change CK frequency to the low-frequency
operating point, and issue MRW-1 and MRW-2 commands. When CKE is driven
HIGH, the device will automatically switch back to the FSP-OP registers that were
in use prior to training (trained values are not retained).
7. Write the trained values to FSP-WR[1] (or FSP-WR[0]) by issuing MRW-1 and
MRW-2 commands to the SDRAM and setting all applicable mode register parameters.
8. Issue MRW-1 and MRW-2 commands to switch the terminating rank to FSP-OP[1]
(or FSP-OP[0]), to turn on termination, and change CK frequency to the high-frequency operating point. At this point the command bus is trained and you may
proceed to other training or normal operation.
161
200b: x32 LPDDR4 SDRAM
Command Bus Training
Training Sequence for Multiple-Rank Systems
The sequence example shown here is assuming an initial low-frequency operating
point, training a high-frequency operating point. The bold text shows high-frequency
instructions. Any operating point may be trained from any known good operating point.
1. Set MR13 OP[6] = 1b to enable writing to frequency set point 1 (FSP-WR[1]) (or
FSP-WR[0]).
2. Write FSP-WR[1] (or FSP-WR[0]) registers for all channels and ranks to set up
high-frequency operating parameters.
3. Read MR0 OP[7] on all channels and ranks to determine which die are terminating, signified by MR0 OP[7] = 1b.
4. Issue MRW-1 and MRW-2 commands to enter command bus training mode on
the terminating rank.
5. Drive CKE LOW on the terminating rank (or all ranks), and change CK frequency
to the high-frequency operating point.
6. Perform command bus training on the terminating rank (VREF(CA), CS, and CA).
7. Exit training by driving CKE HIGH, change CK frequency to the low-frequency
operating point, and issue MRW-1 and MRW-2 commands to write the trained
values to FSP-WR[1] (or FSP-WR[0]). When CKE is driven HIGH, the SDRAM will
automatically switch back to the FSP-OP registers that were in use prior to training (trained values are not retained by the device).
8. Issue MRW-1 and MRW-2 commands to enter training mode on the non-terminating rank (but keep CKE HIGH).
9. Issue MRW-1 and MRW-2 commands to switch the terminating rank to FSP-OP[1]
(or FSP-OP[0]), to turn on termination, and change CK frequency to the high-frequency operating point.
10. Drive CKE LOW on the non-terminating (or all) ranks. The non-terminating
rank(s) will now be using FSP-OP[1] (or FSP-OP[0]).
11. Perform command bus training on the non-terminating rank (VREF(CA), CS, and
CA).
12. Issue MRW-1 and MRW-2 commands to switch the terminating rank to FSPOP[0] (or FSP-OP[1]) to turn off termination.
13. Exit training by driving CKE HIGH on the non-terminating rank, change CK frequency to the low-frequency operating point, and issue MRW-1 and MRW-2 commands. When CKE is driven HIGH, the device will automatically switch back to
the FSP-OP registers that were in use prior to training (that is, trained values are
not retained by the device).
14. Write the trained values to FSP-WR[1] (or FSP-WR[0]) by issuing MRW-1 and
MRW-2 commands and setting all applicable mode register parameters.
15. Issue MRW-1 and MRW-2 commands to switch the terminating rank to FSP-OP[1]
(or FSP-OP[0]), to turn on termination, and change CK frequency to the high-frequency operating point. At this point the command bus is trained for both ranks
and the user may proceed to other training or normal operation.
162
200b: x32 LPDDR4 SDRAM
Command Bus Training
Relation Between CA Input Pin and DQ Output Pin
Table 110: Mapping of CA Input Pin and DQ Output Pin
Mapping
CA number
CA5
CA4
CA3
CA2
CA1
CA0
DQ number
DQ13
DQ12
DQ11
DQ10
DQ9
DQ8
Figure 108: Command Bus Training Mode Entry – CA Training Pattern I/O with VREF(CA) Value Update
T0
T1
T2
T3
T4
T5
Ta
Tb
Tb+1
Tc See Note 1
Td
Te
Te+1
Te+2
Tf
Tg
Th
Th+1
Th+2
CK_c
CK_t
tCKPRECS
See Note 7
tCKPSTCS
See Note 2
CKE
tMRD
tCKELCK
See Note 3
tCACD
CS
CA
DES
MRW1
MRW1
MRW2
MRW2
DES
DES
DES
DES
Valid
DES
Valid
CA training
pattern B
Command
DES
Enter command bus training mode
DES
DES
DES
DES
CA training
pattern A
DES
tDQSCKE
tVREFCA_Long
tCAENT
(see Note 5)
tADR
See Note 4
DQS0_t
DQS0_c
tDS,train
tDH,train
Valid
DQ[6:0]
DQ7
DMI0
DQ[13:8]
Pattern A
DQ[15:14]
DMI1
DQS1_t
DQS1_c
VREF(CA)
(reference)
Setting value of MR X (Y)
Updating setting from FSP switching
tCKELODTon
ODT_CA
(reference)
Mode register X (Y)
Updating setting
Temporary setting value
(see Note 6)
Switching MR
Mode register X (Y)
Don’t Care
Notes:
1. After tCKELCK, the clock can be stopped or the frequency changed any time.
2. The input clock condition should be satisfied tCKPRECS and tCKPSTCS.
3. Continue to drive CK, and hold CA and CS LOW, until tCKELCK after CKE is LOW (which
disables command decoding).
4. The device may or may not capture the first rising edge of DQS_t/DQS_c due to an unstable first rising edge. Therefore, at least two consecutive pulses of DQS signal input is
required every for DQS input signal while capturing DQ[6:0] signals. The captured value
of the DQ[6:0] signal level by each DQS edge may be overwritten at any time and the
device will temporarily update the VREF(CA) setting of MR12 after time tVREFCA_Long.
5. tVREFCA_Long may be reduced to tVREFCA_Short if the following conditions are met: 1)
The new VREF setting is a single step above or below the old VREF setting; 2) The DQS
pulses a single time, or the new VREF setting value on DQ[6:0] is static and meets
tDS,train/ tDH,train for every DQS pulse applied.
163
200b: x32 LPDDR4 SDRAM
Command Bus Training
6. When CKE is driven LOW, the device will switch its FSP-OP registers to use the alternate
(non-active) set. For example, if the device is currently using FSP-OP[0], then it will
switch to FSP-OP[1] when CKE is driven LOW. All operating parameters should be written to the alternate mode registers before entering command bus training to ensure
that ODT settings, RL/WL/nWR setting, and so forth, are set to the correct values. If the
alternate FSP-OP has ODT_CA disabled, then termination will not be enabled in command bus training mode. If the ODT_CA pad is bonded to VSS or floating, ODT_CA termination will never enable for that die.
7. When CKE is driven LOW in command bus training mode, the device will change operation to the alternate FSP, that is, the inverse of the FSP programmed in the FSP-OP mode
register.
Figure 109: Consecutive VREF(CA) Value Update
T0
Ta
Tb
Tb+1
Tc See Note 1
Td
Te
Te+1
Te+2
Te+3
Te+4
Te+5
Te+6
Te+7
Te+8
Te+9 Te+10
Tf
Tf+1
Tf+2
Tf+3
CK_c
CK_t
tCKELCK
(see Note 2)
See Note 7
CKE
tCKELCK
(see Note 3)
CS
CA
DES
DES
Command
DES
DES
DES
DES
DES
DES
DES
DES
tDQSCKE
Valid
CA training
pattern A
tVREFCA_Long
tCAENT
(see Note 5)
tCS_VREF
See Note 4
tVREFCA_Long
(see Note 5)
See Note 4
DQS0_t
DQS0_c
tDS,train
tDH,train
tDS,train
Valid
Valid
DQ[6:0]
tDH,train
DQ7
DMI0
tADR
DQ[13:8]
Pattern A
DQ[15:14]
DMI1
DQS1_t
DQS1_c
VREF(CA)
(reference)
Setting value of MR X (Y)
Updating setting from FSP switching
tCKELODTon
ODT_CA
(reference)
Mode register X (Y)
Updating setting
Temporary setting value
Updating setting
(see Note 6)
Switching MR
Mode register X (Y)
Don’t Care
Notes:
1. After tCKELCK, the clock can be stopped or the frequency changed any time.
2. The input clock condition should be satisfied tCKPRECS and tCKPSTCS.
3. Continue to drive CK, and hold CA and CS LOW, until tCKELCK after CKE is LOW (which
disables command decoding).
4. The device may or may not capture the first rising edge of DQS_t/DQS_c due to an unstable first rising edge. Therefore, at least two consecutive pulses of DQS signal input is
required every for DQS input signal while capturing DQ[6:0] signals. The captured value
of the DQ[6:0] signal level by each DQS edge may be overwritten at any time and the
device will temporarily update the VREF(CA) setting of MR12 after time tVREFCA_Long.
5. tVREFCA_Long may be reduced to tVREFCA_Short if the following conditions are met: 1)
The new VREF setting is a single step above or below the old VREF setting; 2) The DQS
164
200b: x32 LPDDR4 SDRAM
Command Bus Training
pulses a single time, or the new VREF setting value on DQ[6:0] is static and meets
tDS,train/ tDH,train for every DQS pulse applied.
6. When CKE is driven LOW, the device will switch its FSP-OP registers to use the alternate
(non-active) set. For example, if the device is currently using FSP-OP[0], then it will
switch to FSP-OP[1] when CKE is driven LOW. All operating parameters should be written to the alternate mode registers before entering command bus training to ensure
that ODT settings, RL/WL/nWR setting, and so forth, are set to the correct values. If the
alternate FSP-OP has ODT_CA disabled, then termination will not be enabled in command bus training mode. If the ODT_CA pad is bonded to VSS or floating, ODT_CA termination will never enable for that die.
7. When CKE is driven LOW in command bus training mode, the device will change operation to the alternate FSP, that is, the inverse of the FSP programmed in the FSP-OP mode
register.
Figure 110: Command Bus Training Mode Exit with Valid Command
T0
T1
T2
Ta
Ta+1
Ta+2
Tb
Tc
Td
Td+1
Te
Te+1
Tf
Tf+1
Tf+2
Tf+3
Tf+4
Tg
Tg+1
Tg+2
Tg+3
Tg+4
Tg+5
Valid
Valid
Valid
DES
CK_c
CK_t
tCPSTCS
See Note 5
CKE
tCACD
tCCKCKEH
tFC
(see Note 1)
tMRW
CS
See Note 2
CA
Valid
Command
Valid
Valid
Valid
tADR
tADR
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
MRW1
MRW1
MRW2
MRW2
Exiting command bus training mode
DES
DES
Valid
Valid
DES
tCKEHDQS
DQS0_t
DQS0_c
DQ[6:0]
DQ7
DMI0
tMRZ
DQ[13:8]
Pattern A
Pattern B
DQ[15:14]
DMI1
DQS1_t
DQS1_c
VREF(CA)
(reference)
Pattern C
See Note 4
Temporary setting value
Switching MR
tCKELODToff
ODT_CA
(reference)
Mode register X (Y)
Setting value of MR X (Y)
(see Note 3)
Switching MR
Mode register X (Y)
Don’t Care
Notes:
1. The clock can be stopped or the frequency changed any time before tCKCKEH. CK must
meet tCKCKEH before CKE is driven HIGH. When CKE is driven HIGH, the clock frequency
must be returned to the original frequency (that is, the frequency corresponding to the
FSP at command bus training mode entry.
2. CS and CA[5:0] must be deselected (LOW) tCKCKEH before CKE is driven HIGH.
3. When CKE is driven HIGH, ODT_CA will revert to the state/value defined by FSP-OP prior
to command bus training mode entry, that is, the original frequency set point (FSP-OP,
MR13-OP[7]). For example, if the device was using FSP-OP[1] for training, then it will
switch to FSP-OP[0] when CKE is driven HIGH.
165
200b: x32 LPDDR4 SDRAM
Command Bus Training
4. Training values are not retained by the device and must be written to the FSP-OP register set before returning to operation at the trained frequency. For example, VREF(CA) will
return to the value programmed in the original set point.
5. When CKE is driven HIGH, the device will revert to the FSP in operation at command bus
training mode entry.
Figure 111: Command Bus Training Mode Exit with Power-Down Entry
T0
T1
T2
Ta
Ta+1
Ta+2
Tb
Tc
Td
Td+1
Te
Te+1
Tf
Tf+1
Tf+2
Tf+3
Tf+4
Tg
Tg+1
Tg+2
Tg+3
Th
Tk
CK_c
CK_t
tCPSTCS
tCKELCK
See Note 5
CKE
tCACD
tCCKCKEH
tFC
(see Note 1)
tMRD
tCKELCMD
CS
See Note 2
CA
Valid
Valid
CA Pattern B
CA Pattern C
Command
tADR
tADR
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
MRW1
MRW1
MRW2
MRW2
Exiting command bus training mode
DES
Valid
Valid
DES
DES
DES
Power-down entry
DES
DES
tCKEHDQS
DQS0_t
DQS0_c
DQ[6:0]
DQ7
DMI0
tMRZ
DQ[13:8]
Pattern A
Pattern B
DQ[15:14]
DMI1
DQS1_t
DQS1_c
VREF(CA)
(reference)
Pattern C
See Note 4
Temporary setting value
Switching MR
tCKEHODToff
ODT_CA
(reference)
Mode register X (Y)
Setting value of MR X (Y)
(see Note 3)
Switching MR
Mode register X (Y)
Don’t Care
Notes:
1. The clock can be stopped or the frequency changed any time before tCKCKEH. CK must
meet tCKCKEH before CKE is driven HIGH. When CKE is driven HIGH, the clock frequency
must be returned to the original frequency (that is, the frequency corresponding to the
FSP at command bus training mode entry.
2. CS and CA[5:0] must be deselected (LOW) tCKCKEH before CKE is driven HIGH.
3. When CKE is driven HIGH, ODT_CA will revert to the state/value defined by FSP-OP prior
to command bus training mode entry, that is, the original frequency set point (FSP-OP,
MR13-OP[7]). For example, if the device was using FSP-OP[1] for training, then it will
switch to FSP-OP[0] when CKE is driven HIGH.
4. Training values are not retained by the device and must be written to the FSP-OP register set before returning to operation at the trained frequency. For example, VREF(CA) will
return to the value programmed in the original set point.
5. When CKE is driven HIGH, the device will revert to the FSP in operation at command bus
training mode entry.
166
200b: x32 LPDDR4 SDRAM
Write Leveling
Write Leveling
Mode Register Write-WR Leveling Mode
To improve signal-integrity performance, the device provides a write leveling feature to
compensate for CK-to-DQS timing skew, affecting timing parameters such as tDQSS,
tDSS, and tDSH. The memory controller uses the write leveling feature to receive feedback from the device, enabling it to adjust the clock-to-data strobe signal relationship
for each DQS_t/DQS_c signal pair. The device samples the clock state with the rising
edge of DQS signals and asynchronously feeds back to the memory controller. The
memory controller references this feedback to adjust the clock-to-data strobe signal relationship for each DQS_t/DQS_c signal pair.
All data bits (DQ[7:0] for DQS[0] and DQ[15:8] for DQS[1]) carry the training feedback to
the controller. Both DQS signals in each channel must be leveled independently. Write
leveling entry/exit is independent between channels for dual channel devices.
The device enters write leveling mode when mode register MR2-OP[7] is set HIGH.
When entering write leveling mode, the state of the DQ pins is undefined. During write
leveling mode, only DESELECT commands, or a MRW command to exit the WRITE
LEVELING operation, are allowed. Depending on the absolute values of tQSL and tQSH
in the application, the value of tDQSS may have to be better than the limits provided in
the AC Timing Parameters section in order to satisfy the tDSS and tDSH specifications.
Upon completion of the WRITE LEVELING operation, the device exits write leveling
mode when MR2-OP[7] is reset LOW.
Write leveling should be performed before write training (DQS2DQ training).
Write Leveling Procedure
1. Enter write leveling mode by setting MR2-OP[7]=1.
2. Once in write leveling mode, DQS_t must be driven LOW and DQS_c HIGH after a
delay of tWLDQSEN.
3. Wait for a time tWLDQSEN before providing the first DQS signal input. The delay
time tWLMRD(MAX) is controller-dependent.
4. The device may or may not capture the first rising edge of DQS_t due to an unstable first rising edge; therefore, at least two consecutive pulses of DQS signal input
is required for every DQS input signal during write training mode. The captured
clock level for each DQS edge is overwritten, and the device provides asynchronous feedback on all DQ bits after time tWLO.
5. The feedback provided by the device is referenced by the controller to increment
or decrement the DQS_t and/or DQS_c delay settings.
6. Repeat steps 4 and 5 until the proper DQS_t/DQS_c delay is established.
7. Exit write leveling mode by setting MR2-OP[7] = 0.
167
200b: x32 LPDDR4 SDRAM
Write Leveling
Figure 112: Write Leveling Timing – tDQSL (MAX)
T0
T1
T2
T3
T4
Ta
Ta1
Tb
Tb1
Tc
Tc1
MRW
MA
MRW
MA
MRW
OP
MRW
OP
DES
DES
DES
DES
DES
DES
DES
Td
Td1
Td2
Td3
Te
DES
DES
DES
DES
DES
Te1
Tf
Tf1
Tf2
Tf3
Tf4
Tg
Tg1
Tg2
Tg3
Tg4
MRW
MA
MRW
MA
MRW
OP
MRW
OP
DES
DES
Valid
Valid
Valid
Valid
MRW-2 WR
LEVELING exit
DES
CK_c
CK_t
CA[5:0]
Command
MRW-1
WR LEVELING
MRW-2
WR LEVELING
DES
DES
DES
DES
DES
DES
DES
MRW-1 WR
LEVELING exit
DES
DES
Valid
Valid
tDQSH
tWLDQSEN
tDQSL
tWLWPRE
DQS_c
DQS_t
tWLMRD
tWLO
tWLO
tWLO
tMRD
tWLO
DQ[15:0]
DMI[1:0]
Don’t Care
1. Clock can be stopped except during DQS toggle period (CK_t = LOW, CK_c = HIGH).
However, a stable clock prior to sampling is required to ensure timing accuracy.
Note:
Figure 113: Write Leveling Timing – tDQSL (MIN)
T0
T1
T2
T3
T4
Ta
Ta1
Tb
Tb1
Tc
Tc1
MRW
MA
MRW
MA
MRW
OP
MRW
OP
DES
DES
DES
DES
DES
DES
DES
Td
Td1
Td2
Td3
Te
DES
DES
DES
DES
DES
Te1
Tf
Tf1
Tf2
Tf3
Tf4
Tg
Tg1
Tg2
Tg3
MRW
MA
MRW
MA
MRW
OP
MRW
OP
DES
DES
Valid
Valid
Valid
Valid
MRW-2 WR
LEVELING exit
DES
CK_c
CK_t
CA[5:0]
Command
MRW-1
WR LEVELING
MRW-2
WR LEVELING
DES
tWLDQSEN
DES
tWLWPRE
DES
DES
DES
DES
DES
tDQSH
DES
DES
MRW1 WR
LEVELING exit
Valid
Valid
tDQSL
DQS_c
DQS_t
tWLMRD
tWLO
tWLO
tMRD
tWLO
DQ[15:0]
DMI[1:0]
Don’t Care
Note:
1. Clock can be stopped except during DQS toggle period (CK_t = LOW, CK_c = HIGH).
However, a stable clock prior to sampling is required to ensure timing accuracy.
Input Clock Frequency Stop and Change
The input clock frequency can be stopped or changed from one stable clock rate to another stable clock rate during write leveling mode. The frequency stop or change timing
is shown below.
168
200b: x32 LPDDR4 SDRAM
Write Leveling
Figure 114: Clock Stop and Timing During Write Leveling
T0
T1
T2
T3
T4
Ta0
Ta1
Tb0
Tb1
Tb2
Tc0
Td0
Te0
Te1
Te2
Te3
Te4
Tf0
Tf1
Tf2
Tf3
DES
DES
DES
DES
DES
CK_c
CK_t
t CKPSTDQS
t CKPRDQS
CS
CKE
CA
Command
MRW
MA
MRW
MA
MRW-1
WR LEVELING
MRW
OP
MRW
OP
MRW-2
WR LEVELING
DES
DES
DES
DESELECT
DESELECT
t WLDQSEN
DES
DES
DES
DES
DES
DESELECT
DES
DES
DES
DESELECT
DESELECT
t DQSH
t WLWPRE
DES
DESELECT
DESELECT
DESELECT
t DQSL
DQS_t
DQS_c
t WLMRD
t WLO
t WLO
t WLO
t WLO
DQ
DMI
Don’t Care
Notes:
1. CK_t is held LOW and CK_c is held HIGH during clock stop.
2. CS will be held LOW during clock clock stop.
Table 111: Write Leveling Timing Parameters
Parameter
DQS_t/DQS_c delay after write leveling mode is
programmed
Symbol
Min/Max
Value
Units
tWLDQSEN
MIN
20
tCK
MAX
–
MIN
20
Write preamble for write leveling
tWLWPRE
First DQS_t/DQS_c edge after write leveling
mode is programmed
tWLMRD
Write leveling output delay
tWLO
MODE REGISTER SET command delay
tMRD
Valid clock requirement before DQS toggle
tCKPRDQS
Valid clock requirement after DQS toggle
tCKPSTDQS
MAX
–
MIN
40
MAX
–
MIN
0
MAX
20
MIN
MAX (14ns, 10nCK)
MAX
–
MIN
MAX (7.5ns, 4nCK)
MAX
–
MIN
MAX (7.5ns, 4nCK)
MAX
–
tCK
tCK
ns
ns
–
–
Table 112: Write Leveling Setup and Hold Timing
Data Rate
Parameter
Write leveling hold time
Symbol
Min/Max
1600
2400
3200
3733
4267
Unit
tWLH
MIN
150
100
75
62.5
50
ps
169
200b: x32 LPDDR4 SDRAM
Write Leveling
Table 112: Write Leveling Setup and Hold Timing (Continued)
Data Rate
Parameter
Symbol
Min/Max
1600
2400
3200
3733
4267
Unit
Write leveling setup time
tWLS
MIN
150
100
75
62.5
50
ps
Write leveling input valid
window
tWLIVW
MIN
240
160
120
105
90
ps
Notes:
1. In addition to the traditional setup and hold time specifications, there is value in a invalid window-based specification for write leveling training. As the training is based on
each device, worst case process skews for setup and hold do not make sense to close
timing between CK and DQS.
2. tWLIVW is defined in a similar manner to TdIVW_total, except that here it is a DQS invalid window with respect to CK. This would need to account for all voltage and temperature (VT) drift terms between CK and DQS within the device that affect the write leveling invalid window.
The figure below shows the DQS input mask for timing with respect to CK. The “total”
mask (tWLIVW) defines the time the input signal must not encroach in order for the
DQS input to be successfully captured by CK. The mask is a receiver property and it is
not the valid data-eye.
Figure 115: DQS_t/DQS_c to CK_t/CK_c Timings at the Pins Referenced from the Internal Latch
Internal composite DQS eye
center aligned to CK
CK_c
CK_t
DQ_diff =
DQS_t–DQS_c
tWLIVW
170
200b: x32 LPDDR4 SDRAM
MULTIPURPOSE Operation
MULTIPURPOSE Operation
The device uses the MULTIPURPOSE command to issue a NO OPERATION (NOP) command and to access various training modes. The MPC command is initiated with CS,
and CA[5:0] asserted to the proper state at the rising edge of CK, as defined by the Command Truth Table. The MPC command has seven operands (OP[6:0]) that are decoded
to execute specific commands in the SDRAM. OP[6] is a special bit that is decoded on
the first rising CK edge of the MPC command. When OP[6] = 0, the device executes a
NOP command, and when OP[6] = 1, the device further decodes one of several training
commands.
When OP[6] = 1 and the training command includes a READ or WRITE operation, the
MPC command must be followed immediately by a CAS-2 command. For training commands that read or write, READ latency (RL) and WRITE latency (WL) are counted from
the second rising CK edge of the CAS-2 command with the same timing relationship as
a typical READ or WRITE command. The operands of the CAS-2 command following a
MPC READ/WRITE command must be driven LOW. The following MPC commands
must be followed by a CAS-2 command:
• WRITE-FIFO
• READ-FIFO
• READ DQ CALIBRATION
All other MPC commands do not require a CAS-2 command, including the following:
•
•
•
•
•
NOP
START DQS OSCILLATOR
STOP DQS OSCILLATOR
ZQCAL START
ZQCAL LATCH
Table 113: MPC Command Definition
SDR Command Pins
SDR CA Pins
CKE
SDR Command
MPC
(Train, NOP)
CK_t
(n-1)
CK_t(n)
CS
CA0
CA1
CA2
CA3
CA4
CA5
H
H
H
L
L
L
L
L
OP6
1
L
OP0
OP1
OP2
OP3
OP4
OP5
2
Notes:
CK_t
Edge
Notes
1, 2
1. See the Command Truth Table for more information.
2. MPC commands for READ or WRITE TRAINING operations must be immediately followed
by the CAS-2 command, consecutively, without any other commands in between. The
MPC command must be issued before issuing the CAS-2 command.
171
200b: x32 LPDDR4 SDRAM
MULTIPURPOSE Operation
Table 114: MPC Commands
Function
Operand
Training Modes
OP[6:0]
Data
0XXXXXXb: NOP
1000001b: READ-FIFO: READ-FIFO supports only BL16 operation
1000011b: READ DQ CALIBRATION (MR32/MR40)
1000101b: RFU
1000111b: WRITE-FIFO: WRITE-FIFO supports only BL16 operation
1001001b: RFU
1001011b: START DQS OSCILLATOR
1001101b: STOP DQS OSCILLATOR
1001111b: ZQCAL START
1010001b: ZQCAL LATCH
All Others: Reserved
1. See command truth table for more information.
2. MPC commands for READ or WRITE TRAINING operations must be immediately followed
by CAS-2 command consecutively without any other commands in-between. MPC command must be issued first before issuing the CAS-2 command.
3. WRITE-FIFO and READ-FIFO commands will only operate as BL16, ignoring the burst
length selected by MR1 OP[1:0].
Notes:
Figure 116: WRITE-FIFO – tWPRE = 2nCK, tWPST = 0.5nCK
T0
T1
T2
T3
BL
BA0,
CA
CAn
CAn
T4
Ta0
Ta1
Ta2
Tb0
Tb1
Tc0
Tc1
Tc2
Tc3
Tc4
Valid
Valid
Valid
Valid
Tc5
Td0
Td1
Td2
Td3
Td4
Valid
Valid
Valid
Valid
Td5
Te0
Te1
Te2
Tf0
Tf1
Tg0
Tg1
Tg2
DES
DES
DES
DES
DES
DES
DES
DES
Tg3
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
DES
DES
DES
DES
DES
DES
DES
MPC
[WRITE-FIFO]
CAS-2
DES
DES
MPC
[WRITE-FIFO]
tCCD
t WRWTR
WL
=8
CAS-2
DES
DES
WL
WL
tDQSS
t WPRE
tDQSS
tDQSS
t WPST
WPRE
t WPST
DQS_c
DQS_t
tDQS2DQ
tDQS2DQ
DQ[15:0]
DMI[1:0]
n0 n13 n14 n15
tDQS2DQ
n0 n13 n14 n15 n0 n13 n14 n15
Don’t Care
Notes:
1. MPC[WRITE-FIFO] can be executed with a single bank or multiple banks active, during
refresh or during self refresh, with CKE HIGH.
2. Write-1 to MPC is shown as an example of command-to-command timing for MPC. Timing from Write-1 to MPC[WRITE-FIFO] is tWRWTR.
3. Seamless MPC[WRITE-FIFO] commands may be executed by repeating the command every tCCD time.
4. MPC[WRITE-FIFO] uses the same command-to-data timing relationship (WL, tDQSS,
tDQS2DQ) as a WRITE-1 command.
5. A maximum of five MPC[WRITE-FIFO] commands may be executed consecutively without
corrupting FIFO data. The sixth MPC[WRITE-FIFO] command will overwrite the FIFO data
172
200b: x32 LPDDR4 SDRAM
MULTIPURPOSE Operation
from the first command. If fewer than five MPC[WRITE-FIFO] commands are executed,
then the remaining FIFO locations will contain undefined data.
6. For the CAS-2 command following an MPC command, the CAS-2 operands must be driven LOW.
7. To avoid corrupting the FIFO contents, MPC[READ-FIFO] must immediately follow
MPC[WRITE-FIFO]/CAS-2 without any other commands in-between. See Write Training
section for more information on FIFO pointer behavior.
Figure 117: READ-FIFO – tWPRE = 2nCK, tWPST = 0.5nCK, tRPRE = Toggling, tRPST = 1.5nCK
T0
T1
T2
T3
T4
Ta
Ta+1
Ta+2
Ta+3
Tb
Tb+1 Tb+2
Tb+3
Tb+4
Tb+5
Tc+2
Tc+3
Tc+4
Tc+5
Tc+6
Tc+7
Td
Td+1
Td+2 Td+3
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Td+4
Td+5
Valid
Valid
Te
Te+1
Tf
Tf+1
Tf+2
Tf+3
Tf+4
Valid
Valid
Valid
Valid
Valid
Valid
Valid
CK_c
CK_t
tCCD
CA[5:0]
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
WL
Command
MPC
[WRITE-FIFO]
CAS-2
Valid
RL
Valid
Valid
Valid
MPC
[READ-FIFO]
tDQSS
CAS-2
Valid
MPC
[READ-FIFO]
tWTR
CAS-2
Valid
Valid
Valid
Valid
Valid
Valid
Valid
tDQSCK
DQS_t
DQS_c
tDQS2DQ
DQ[15:0]
DMI[1:0]
tRPRE
D0 D1 D12 D13 D14 D15
tRPST
D0 D1 D2 D13 D14 D15 D0 D11 D12 D13 D14 D15
Don’t Care
Notes:
1. MPC[WRITE-FIFO] can be executed with a single bank or multiple banks active, during
refresh or during self refresh with CKE HIGH.
2. Seamless MPC[READ-FIFO] commands may be executed by repeating the command every tCCD time.
3. MPC[READ-FIFO] uses the same command-to-data timing relationship (RL, tDQSCK) as a
READ-1 command.
4. Data may be continuously read from the FIFO without any data corruption. After five
MPC[READ-FIFO] commands, the FIFO pointer will wrap back to the first FIFO and continue advancing. If fewer than five MPC[WRITE-FIFO] commands were executed, then
the MPC[READ-FIFO] commands to those FIFO locations will return undefined data. See
Write Training for more information on the FIFO pointer behavior.
5. For the CAS-2 command immediately following an MPC command, the CAS-2 operands
must be driven LOW.
6. DMI[1:0] signals will be driven if WR-DBI, RD-DBI, or DM is enabled in the mode registers. See Write Training for more information on DMI behavior.
173
200b: x32 LPDDR4 SDRAM
MULTIPURPOSE Operation
Figure 118: READ-FIFO – tRPRE = Toggling, tRPST = 1.5nCK
T0
T1
T2
T3
T4
Ta
Ta+1
Ta+2
Ta+3
Ta+4
Ta+5
Tb
Valid
Valid
Valid
Valid
Tb+1 Tb+2
Tb+3
Tb+4
Valid
Valid
Valid
Tc
Tc+1
Tc+2
Td
Td+1
Td+2
Td+3
Te
Te+1
Te+2
Te+3
Te+4
Te+5
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
CK_c
CK_t
tRTRRD
CA[5:0]
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
RL
RL
Command
MPC
[READ-FIFO]
CAS-2
Valid
READ-1
CAS-2
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
tDQSCK
Valid
Valid
tDQSCK
DQS_t
DQS_c
tRPST
tRPRE
DQ[15:0]
DMI[1:0]
tRPST
tRPRE
D0 D1 D12 D13 D14 D15
D0 D1 D12 D13 D14 D15
Don’t Care
Notes:
1. MPC[WRITE-FIFO] can be executed with a single bank or multiple banks active, during
refresh or during self refresh with CKE HIGH.
2. MPC[READ-FIFO] to READ-1 operation is shown as an example of command-to-command timing for MPC. Timing from MPC[READ-FIFO] command to read is tRTRRD.
3. Seamless MPC[READ-FIFO] commands may be executed by repeating the command every tCCD time.
4. MPC[READ-FIFO] uses the same command-to-data timing relationship (RL, tDQSCK) as a
READ-1 command.
5. Data may be continuously read from the FIFO without any data corruption. After five
MPC[READ-FIFO] commands, the FIFO pointer will wrap back to the first FIFO and continue advancing. If fewer than five MPC[WRITE-FIFO] commands are executed, then the
MPC[READ-FIFO] commands to those FIFO locations will return undefined data. See
Write Training for more information on the FIFO pointer behavior.
6. For the CAS-2 command immediately following an MPC command, the CAS-2 operands
must be driven LOW.
7. DMI[1:0] signals will be driven if WR-DBI, RD-DBI, or DM is enabled in the mode registers. See Write Training for more information on DMI behavior.
Table 115: Timing Constraints for Training Commands
Previous Command
Next Command
Minimum Delay
Unit
Notes
WR/MWR
MPC[WRITE-FIFO]
tWRWTR
nCK
1
MPC[READ-FIFO]
Not allowed
–
2
RU(tDQSS(MAX)/tCK)
MPC[READ DQ CALIBRATION]
RD/MRR
WL +
BL/2 + RU(tWTR/tCK)
+
nCK
MPC[WRITE-FIFO]
tRTW
nCK
4
MPC[READ-FIFO]
Not allowed
–
2
MPC[READ DQ CALIBRATION]
tRTRRD
nCK
3
174
200b: x32 LPDDR4 SDRAM
MULTIPURPOSE Operation
Table 115: Timing Constraints for Training Commands (Continued)
Previous Command
Next Command
Minimum Delay
Unit
Notes
MPC[WRITE-FIFO]
WR/MWR
Not allowed
–
2
MPC[WRITE-FIFO]
tCCD
nCK
RD/MRR
Not allowed
–
RU(tDQSS(MAX)/tCK)
MPC[READ-FIFO]
Notes:
+
nCK
MPC[READ DQ CALIBRATION]
Not allowed
–
2
WR/MWR
tRTW
nCK
4
MPC[WRITE-FIFO]
tRTW
nCK
4
RD/MRR
tRTRRD
nCK
3
MPC[READ-FIFO]
tCCD
nCK
MPC[READ DQ CALIBRATION]
tRTRRD
nCK
3
WR/MWR
tRTW
nCK
4
MPC[WRITE-FIFO]
tRTW
nCK
4
RD/MRR
tRTRRD
nCK
3
MPC[READ-FIFO]
Not allowed
–
2
MPC[READ DQ CALIBRATION]
tCCD
nCK
MPC[READ-FIFO]
MPC[READ DQ CALIBRATION]
WL +
BL/2 + RU(tWTR/tCK)
2
1. tWRWTR = WL + BL/2 + RU(tDQSS(MAX)/tCK) + MAX(RU(7.5ns/tCK), 8nCK).
2. No commands are allowed between MPC[WRITE-FIFO] and MPC[READ-FIFO] except the
MRW commands related to training parameters.
3. tRTRRD = RL + RU(tDQSCK(MAX)/tCK) + BL/2 + tRPST + 0.5 + MAX(RU(7.5ns/tCK), 8nCK).
4. In case of DQ ODT disable MR11 OP[2:0] = 000b,
tRTW
= RL + RU(tDQSCK(MAX)/tCK) + BL/2 - WL + tWPRE + RD(tRPST).
In case of DQ ODT enable MR11 OP[2:0] ≠ 000b,
tRTW
= RL + RU(tDQSCK(MAX)/tCK) + BL/2 + tRPST - ODTLon - RD(tODTon(Min)/tCK) + 1.
175
200b: x32 LPDDR4 SDRAM
Read DQ Calibration Training
Read DQ Calibration Training
The READ DQ CALIBRATION TRAINING function outputs a 16-bit, user-defined pattern on the DQ pins. Read DQ calibration is initiated by issuing a MPC[READ DQ CALIBRATION] command followed by a CAS-2 command, which causes the device to drive
the contents of MR32, followed by the contents of MR40 on each of DQ[15:0] and
DMI[1:0]. The pattern can be inverted on selected DQ pins according to user-defined
invert masks written to MR15 and MR20.
Read DQ Calibration Training Procedure
1. Issue MRW commands to write MR32 (first eight bits), MR40 (second eight bits),
MR15 (eight-bit invert mask for byte 0), and MR20 (eight-bit invert mask for byte
1).
In the alternative, this step could be replaced with the default pattern:
• MR32 default = 5Ah
• MR40 default = 3Ch
• MR15 default = 55h
• MR20 default = 55h
2. Issue an MPC command, followed immediately by a CAS-2 command.
• Each time an MPC command, followed by a CAS-2, is received by the device, a
16-bit data burst will drive the eight bits programmed in MR32 followed by the
eight bits programmed in MR40 on all I/O pins after the currently set RL.
• The data pattern will be inverted for I/O pins with a 1 programmed in the corresponding invert mask mode register bit (see table below).
• The pattern is driven on the DMI pins, but no DATA BUS INVERSION function is
enabled, even if read DBI is enabled in the mode register.
• The MPC command can be issued every tCCD seamlessly, and tRTRRD delay is
required between ARRAY READ command and the MPC command as well the
delay required between the MPC command and an ARRAY READ.
• The operands received with the CAS-2 command must be driven LOW.
3. DQ
Read DQ calibration training can be performed with any or no banks active during
refresh or during self refresh with CKE HIGH.
Table 116: Invert Mask Assignments
DQ pin
0
1
2
3
DMI0
4
5
6
7
MR15 bit
0
1
2
3
N/A
4
5
6
7
DQ pin
8
9
10
11
DMI1
12
13
14
15
MR20 bit
0
1
2
3
N/A
4
5
6
7
176
200b: x32 LPDDR4 SDRAM
Read DQ Calibration Training
Figure 119: Read DQ Calibration Training Timing: Read-to-Read DQ Calibration
T0
T1
T2
T3
BL
BA0,
CA
CAn
CAn
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Tb0
Tb1
Tb2
Tc1
Tc2
Tc3
Tc4
Tc5
Tc6
Valid
Valid
Tc7
Td1
Td2
Td3
Td4
Td5
Td6
Te0
DES
DES
DES
DES
DES
DES
DES
Te1
Te2
Te3
CK_c
CK_t
CS
CA
Command
Valid
CAS-2
READ-1
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
Valid
MPC
[RD DQ CAL]
DES
CAS-2
DES
tRTRRD
t
RL
RL
DES
DES
DQSCK
t RPST
t RPRE
t RPST
High-Z
High-Z
tDQSQ
DQ
DMI
DES
DQSCK
t RPRE
DQS_c
DQS_t
t
High-Z
tDQSQ
High-Z
n0 n13 n14 n15
n0 n13 n14 n15
Don’t Care
1. Read-1 to MPC operation is shown as an example of command-to-command timing. Timing from Read-1 to MPC command is tRTRRD.
2. MPC uses the same command-to-data timing relationship (RL, tDQSCK, tDQSQ) as a
Read-1 command.
3. BL = 16, Read preamble: Toggle, Read postamble: 0.5nCK.
4. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Notes:
Figure 120: Read DQ Calibration Training Timing: Read DQ Calibration to Read DQ Calibration/Read
T0
T1
T2
T3
T8
T9
T10
T11
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
T12
T13
T14
Ta0
Ta1
Ta2
Ta3
Tb0
Tb1
Tc0
Tc1
Tc2
Tc3
Tc4
BA0,
CA
CAn
CAn
Tc5
Td0
DES
DES
Td1
Td2
Td3
Td4
DES
DES
DES
Te0
Te1
CK_c
CK_t
CS
CA
Command
MPC
[RD DQ CAL]
CAS-2
DES
MPC
[RD DQ CAL]
CAS-2
BL
DES
DES
DES
DES
DES
DES
t CCD
DES
DES
READ-1
CAS-2
RL
DES
DES
DES
tDQSCK
tDQSCK
RL
t RPST
t RPRE
High-Z
High-Z
t RPRE
t RPST
High-Z
tDQSQ
DQ
DMI
DES
tDQSCK
RL
DQS_c
DQS_t
DES
tRTRRD
t DQSQ
n0
n9 n10 n11 n12 n13 n14 n15 n0 n13 n14 n15
tDQSQ
High-Z
n0 n13 n14 n15
Don’t Care
Notes:
1. MPC[READ DQ CALIBRATION] to MPC[READ DQ CALIBRATION] operation is shown as an
example of command-to-command timing.
2. MPC[READ DQ CALIBRATION] to READ-1 operation is shown as an example of command-to-command timing.
3. MPC[READ DQ CALIBRATION] uses the same command-to-data timing relationship (RL,
tDQSCK, tDQSQ) as a READ-1 command.
4. Seamless MPC[READ DQ CALIBRATION] commands may be executed by repeating the
command every tCCD time.
5. Timing from MPC[READ DQ CALIBRATION] command to READ-1 is tRTRRD.
177
200b: x32 LPDDR4 SDRAM
Read DQ Calibration Training
6. BL = 16, Read preamble: Toggle, Read postamble: 0.5nCK.
7. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Read DQ Calibration Training Example
An example of read DQ calibration training output is shown in table below. This shows
the 16-bit data pattern that will be driven on each DQ in byte 0 when one READ DQ
CALIBRATION TRAINING command is executed. This output assumes the following
mode register values are used:
•
•
•
•
MR32 = 1CH
MR40 = 59H
MR15 = 55H
MR20 = 55H
Table 117: Read DQ Calibration Bit Ordering and Inversion Example
Bit Sequence →
Pin
Invert
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
DQ0
Yes
1
0
1
0
0
1
1
0
1
1
1
0
0
0
1
1
DQ1
No
0
1
0
1
1
0
0
1
0
0
0
1
1
1
0
0
DQ2
Yes
1
0
1
0
0
1
1
0
1
1
1
0
0
0
1
1
DQ3
No
0
1
0
1
1
0
0
1
0
0
0
1
1
1
0
0
DMI0
Never
0
1
0
1
1
0
0
1
0
0
0
1
1
1
0
0
DQ4
Yes
1
0
1
0
0
1
1
0
1
1
1
0
0
0
1
1
DQ5
No
0
1
0
1
1
0
0
1
0
0
0
1
1
1
0
0
DQ6
Yes
1
0
1
0
0
1
1
0
1
1
1
0
0
0
1
1
DQ7
No
0
1
0
1
1
0
0
1
0
0
0
1
1
1
0
0
DQ8
Yes
1
0
1
0
0
1
1
0
1
1
1
0
0
0
1
1
DQ9
No
0
1
0
1
1
0
0
1
0
0
0
1
1
1
0
0
DQ10
Yes
1
0
1
0
0
1
1
0
1
1
1
0
0
0
1
1
DQ11
No
0
1
0
1
1
0
0
1
0
0
0
1
1
1
0
0
DMI1
Never
0
1
0
1
1
0
0
1
0
0
0
1
1
1
0
0
DQ12
Yes
1
0
1
0
0
1
1
0
1
1
1
0
0
0
1
1
DQ13
No
0
1
0
1
1
0
0
1
0
0
0
1
1
1
0
0
DQ14
Yes
1
0
1
0
0
1
1
0
1
1
1
0
0
0
1
1
DQ15
No
0
1
0
1
1
0
0
1
0
0
0
1
1
1
0
0
Notes:
1. The patterns contained in MR32 and MR40 are transmitted on DQ[15:0] and DMI[1:0]
when read DQ calibration is initiated via a MPC[READ DQ CALIBRATION] command. The
pattern transmitted serially on each data lane, organized little endian such that the loworder bit in a byte is transmitted first. If the data pattern is 27H, then the first bit transmitted with be a 1, followed by 1, 1, 0, 0, 1, 0, and 0. The bit stream will be 00100111 →.
2. MR15 and MR20 may be used to invert the MR32/MR40 data pattern on the DQ pins.
See MR15 and MR20 for more information. Data is never inverted on the DMI[1:0] pins.
3. DMI [1:0] outputs status follows MR Setting vs. DMI Status table.
178
200b: x32 LPDDR4 SDRAM
Write Training
4. No DATA BUS INVERSION (DBI) function is enacted during read DQ calibration, even if
DBI is enabled in MR3-OP[6].
Table 118: MR Setting vs. DMI Status
DM Function
MR13 OP[5]
WRITE DBIdc Function
MR3 OP[7]
READ DBIdc Function
MR3 OP[6] DMI
Status
1: Disable
0: Disable
0: Disable
High-Z
1: Disable
1: Enable
0: Disable
The data pattern is transmitted
1: Disable
0: Disable
1: Enable
The data pattern is transmitted
1: Disable
1: Enable
1: Enable
The data pattern is transmitted
0: Enable
0: Disable
0: Disable
The data pattern is transmitted
0: Enable
1: Enable
0: Disable
The data pattern is transmitted
0: Enable
0: Disable
1: Enable
The data pattern is transmitted
0: Enable
1: Enable
1: Enable
The data pattern is transmitted
MPC[READ DQ CALIBRATION] After Power-Down Exit
Following the power-down state, an additional time, tMRRI, is required prior to issuing
the MPC[READ DQ CALIBRATION] command. This additional time (equivalent to
tRCD) is required in order to be able to maximize power-down current savings by allowing more power-up time for the read DQ data in MR32 and MR40 data path after exit
from standby, power-down mode.
Figure 121: MPC[READ DQ CALIBRATION] Following Power-Down State
T0
Ta0
Tb0
Tb1
Tb2
Tc0
Tc1
Tc2
Tc3
Tc4
Tc5
Td0
Td1
Td2
Td3
Td4
Td5
Td6
Td7
Td8
Td9
DES
DES
DES
CK_c
CK_t
t CKCKEH
CKE
t MRRI
t XP
CS
CA
Command
Valid Valid Valid Valid
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
MPC
[READ DQ CAL]
CAS-2
Don’t Care
Write Training
The device uses an unmatched DQS-DQ path to enable high-speed performance and
save power. As a result, the DQS strobe must be trained to arrive at the DQ latch centeraligned with the data eye. The DQ receiver is located at the DQ pad and has a shorter
internal delay than the DQS signal. The DQ receiver will latch the data present on the
179
200b: x32 LPDDR4 SDRAM
Write Training
DQ bus when DQS reaches the latch, and training is accomplished by delaying the DQ
signals relative to DQS such that the data eye arrives at the receiver latch centered on
the DQS transition.
Two modes of training are available:
• Command-based FIFO WR/RD with user patterns
• An internal DQS clock-tree oscillator, which determines the need for, and the magnitude of, required training
The command-based FIFO WR/RD uses the MPC command with operands to enable
this special mode of operation. When issuing the MPC command, if CA[5] is set LOW
(OP[6] = 0), then the device will perform a NOP command. When CA[5] is set HIGH, the
CA[4:0] pins enable training functions or are reserved for future use (RFU). MPC commands that initiate a read or write to the device must be followed immediately by a
CAS-2 command. See the MPC Operation section for more information.
To perform write training, the controller can issue an MPC[WRITE-FIFO] command
with OP[6:0] set, followed immediately by a CAS-2 command (CAS-2 operands should
be driven LOW) to initiate a WRITE-FIFO. Timings for MPC[WRITE-FIFO] are identical
to WRITE commands, with WL timed from the second rising clock edge of the CAS-2
command. Up to five consecutive MPC[WRITE-FIFO] commands with user-defined
patterns may be issued to the device, which will store up to 80 values (BL16 × 5) per pin
that can be read back via the MPC[READ-FIFO] command. (The WRITE/READ-FIFO
POINTER operation is described in a different section.
After writing data with the MPC[WRITE-FIFO] command, the data can be read back
with the MPC[READ-FIFO] command and results can be compared with "expected" data to determine whether further training (DQ delay) is needed. MPC[READ-FIFO] is initiated by issuing an MPC command, as described in the MPC Operation section, followed immediately by a CAS-2 command (CAS-2 operands must be driven LOW). Timings for the MPC[READ-FIFO] command are identical to READ commands, with RL
timed from the second rising clock edge of the CAS-2 command.
READ-FIFO is nondestructive to the data captured in the FIFO; data may be read continuously until it is disturbed by another command, such as a READ, WRITE, or another
MPC[WRITE-FIFO]. If fewer than five WRITE-FIFO commands are executed, unwritten
registers will have undefined (but valid) data when read back.
For example: If five WRITE-FIFO commands are executed sequentially, then a series of
READ-FIFO commands will read valid data from FIFO[0], FIFO[1]….FIFO[4] and then
wrap back to FIFO[0] on the next READ-FIFO. However, if fewer than five WRITE-FIFO
commands are executed sequentially (example = 3), then a series of READ-FIFO commands will return valid data for FIFO[0], FIFO[1], and FIFO[2], but the next two READFIFO commands will return undefined data for FIFO[3] and FIFO[4] before wrapping
back to the valid data in FIFO[0].
The READ-FIFO pointer and WRITE-FIFO pointer are reset under the following conditions:
•
•
•
•
Power-up initialization
RESET_n asserted
Power-down entry
Self refresh power-down entry
180
200b: x32 LPDDR4 SDRAM
Write Training
The MPC[WRITE-FIFO] command advances the WRITE-FIFO pointer, and the
MPC[READ-FIFO] advances the READ-FIFO pointer. Also any normal (non-FIFO) READ
operation (RD, RDA) advances both WRITE-FIFO pointer and READ-FIFO pointer. Issuing (non-FIFO) READ operation command is inhibited during write training period. To
keep the pointers aligned, the SoC memory controller must adhere to the following restriction at the end of Write training period:
b = a + (n × c)
Where:
'a' is the number of MPC[WRITE-FIFO] commands
'b' is the number of MPC[READ-FIFO] commands
'c' is the FIFO depth (= 5 for LPDDR4)
'n' is a positive integer, 0
Figure 122: WRITE-to-MPC[WRITE-FIFO] Operation Timing
T0
T1
T2
T3
BL
BA0,
CA
CAn
CAn
T4
Ta0
Ta1
Ta2
Tb0
Tb1
Tc0
Tc1
Tc2
Tc3
Tc4
Valid
Valid
Valid
Valid
Tc5
Td0
Td1
Td2
Td3
Td4
Valid
Valid
Valid
Valid
Td5
Te0
Te1
Te2
Tf0
Tf1
Tg0
Tg1
Tg2
DES
DES
DES
DES
DES
DES
DES
Tg3
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
DES
DES
DES
DES
DES
DES
DES
MPC
[WRITE-FIFO]
CAS-2
DES
DES
MPC
[WRITE-FIFO]
tCCD
t WRWTR
WL
=8
CAS-2
DES
DES
DES
WL
WL
tDQSS
t WPRE
tDQSS
tDQSS
t WPST
WPRE
t WPST
DQS_c
DQS_t
tDQS2DQ
tDQS2DQ
DQ[15:0]
DMI[1:0]
n0 n13 n14 n15
tDQS2DQ
n0 n13 n14 n15 n0 n13 n14 n15
Don’t Care
Notes:
1. MPC[WRITE-FIFO] can be executed with a single bank or multiple banks active during
REFRESH or during SELF REFRESH with CKE HIGH.
2. Write-1 to MPC is shown as an example of command-to-command timing for MPC. Timing from Write-1 to MPC[WRITE-FIFO] is tWRWTR.
3. Seamless MPC[WR-FIFO] commands may be executed by repeating the command every
tCCD time.
4. MPC[WRITE-FIFO] uses the same command-to-data timing relationship (WL, tDQSS,
tDQS2DQ) as a WRITE-1 command.
5. A maximum of five MPC[WRITE-FIFO] commands may be executed consecutively without
corrupting FIFO data. The sixth MPC[WRITE-FIFO] command will overwrite the FIFO data
from the first command. If fewer than five MPC[WRITE-FIFO] commands are executed,
then the remaining FIFO locations will contain undefined data.
6. For the CAS-2 command following an MPC command, the CAS-2 operands must be driven LOW.
7. To avoid corrupting the FIFO contents, MPC[READ-FIFO] must immediately follow
MPC[WRITE-FIFO]/CAS-2 without any other commands disturbing FIFO pointers in between. FIFO pointers are disturbed by CKE LOW, WRITE, MASKED WRITE, READ, READ
DQ CALIBRATION, and MRR.
8. BL = 16, Write postamble = 0.5nCK.
9. DES commands are shown for ease of illustration; other commands may be valid at
these times.
181
200b: x32 LPDDR4 SDRAM
Write Training
Figure 123: MPC[WRITE-FIFO]-to-MPC[READ-FIFO] Timing
T0
T1
T2
T3
BL
BA0,
CA
CAn
CAn
T4
Ta0
Ta1
Ta2
Tb0
Tb1
Tb2
Tc0
Tc1
Tc2
Tc3
Tc4
Valid
Valid
Valid
Valid
Tc5
Td0
Td1
Td2
Td3
Td4
Valid
Valid
Td5
Te0
Te1
Te2
Te3
Tf0
Tf1
Tg0
Tg1
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
MPC
[WRITE-FIFO]
CAS-2
DES
DES
DES
WL
DES
DES
DES
DES
DES
MPC
[READ-FIFO]
CAS-2
DES
DES
Valid
MPC
[READ-FIFO]
tCCD
tWTR
BL/2 + 1 clock
Valid
=8
RL
tDQSS
t WPRE
CAS-2
tDQSCK
t RPRE
t WPST
t RPST
DQS_c
DQS_t
tDQS2DQ
DQ
DMI
tDQSQ
n0 n13 n14 n15
n0 n13 n14 n15 n0 n13 n14 n15
Don’t Care
Notes:
1. MPC[WRITE-FIFO] can be executed with a single bank or multiple banks active during refresh or during self refresh with CKE HIGH.
2. MPC[WRITE-FIFO] to MPC[READ-FIFO] is shown as an example of command-to-command
timing for MPC. Timing from MPC[WRITE-FIFO] to MPC[READ-FIFO] is specified in the
command-to-command timing table.
3. Seamless MPC[READ-FIFO] commands may be executed by repeating the command every tCCD time.
4. MPC[READ-FIFO] uses the same command-to-data timing relationship (RL, tDQSCK,
tDQSQ ) as a READ-1 command.
5. Data may be continuously read from the FIFO without any data corruption. After five
MPC[READ-FIFO] commands, the FIFO pointer will wrap back to the first FIFO and continue advancing. If fewer than five MPC[WRITE-FIFO] commands were executed, then
the MPC[READ-FIFO] commands to those FIFO locations will return undefined data. See
Write Training for more information on the FIFO pointer behavior.
6. For the CAS-2 command immediately following an MPC command, the CAS-2 operands
must be driven LOW.
7. DMI[1:0] signals will be driven if WR-DBI, RD-DBI, or DM is enabled in the mode registers. See Write Training section for more information on DMI behavior.
8. BL = 16, Write postamble = 0.5nCK, Read preamble: Toggle, Read postamble: 0.5nCK.
9. DES commands are shown for ease of illustration; other commands may be valid at
these times.
182
200b: x32 LPDDR4 SDRAM
Write Training
Figure 124: MPC[READ-FIFO] to Read Timing
T0
T1
T2
T3
Valid
Valid
CAn
CAn
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Tb0
Tb1
Tb2
Tc1
Tc2
Tc3
Tc4
Tc5
Tc6
BL
BA0,
CA
CAn
CAn
Tc7
Td1
Td2
Td3
Td4
Td5
Td6
Te0
DES
DES
DES
DES
DES
DES
DES
Te1
Te2
Te3
CK_c
CK_t
CS
CA
Command
MPC
RAD FIFO
CAS-2
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
tRTRRD
RL
t
DES
READ-1
CAS-2
DES
RL
t
DES
DES
DES
DQSCK
DQSCK
t RPRE
t RPRE
t RPST
t RPST
DQS_c
DQS_t
tDQSQ
DQ
DMI
tDQSQ
n0 n13 n14 n15
n0 n13 n14 n15
Don’t Care
Notes:
1. MPC[WRITE-FIFO] can be executed with a single bank or multiple banks active during refresh or during self refresh with CKE HIGH.
2. MPC[READ-FIFO] to READ-1 operation is shown as an example of command-to-command timing for MPC. Timing from MPC[READ-FIFO] command to READ is tRTRRD.
3. Seamless MPC[READ-FIFO] commands may be executed by repeating the command every tCCD time.
4. MPC[READ-FIFO] uses the same command-to-data timing relationship (RL, tDQSCK,
tDQSQ ) as a READ-1 command.
5. Data may be continuously read from the FIFO without any data corruption. After five
MPC[READ-FIFO] commands, the FIFO pointer will wrap back to the first FIFO and continue advancing. If fewer than five MPC[WRITE-FIFO] commands were executed, then
the MPC[READ-FIFO] commands to those FIFO locations will return undefined data. See
Write Training for more information on the FIFO pointer behavior.
6. For the CAS-2 command immediately following an MPC command, the CAS-2 operands
must be driven LOW.
7. DMI[1:0] signals will be driven if WR-DBI, RD-DBI, or DM is enabled in the mode registers. See Write Training for more information on DMI behavior.
8. BL = 16, Read preamble: Toggle, Read postamble: 0.5nCK
9. DES commands are shown for ease of illustration; other commands may be valid at
these times.
183
200b: x32 LPDDR4 SDRAM
Write Training
Figure 125: MPC[WRITE-FIFO] with DQ ODT Timing
T0
T1
T2
T3
Valid
Valid
Valid
Valid
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Ta6
Ta7
Ta8
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Tb6
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
CK_c
CK_t
CS
CA
Command
MPC
[WRITE-FIFO]
CAS-2
WL
tDQSS
t WPRE
t WPST
DQS_c
DQS_t
tDQS2DQ
DQ
DMI
n0
ODTLon
t
ODTon(MAX)
t
DRAM RTT
ODT High-Z
n1 n2 n13 n14 n15
ODTon(MIN)
Transition
Transition
ODT On
ODTLoff
t
t
ODT High-Z
ODToff(MIN)
ODToff(MAX)
Don’t Care
Notes:
1. MPC[WRITE-FIFO] can be executed with a single bank or multiple banks active during refresh or during self refresh with CKE HIGH.
2. MPC[WRITE-FIFO] uses the same command-to-data/ODT timing relationship (RL, tDQSCK,
tDQS2DQ, ODTLon, ODTLoff, tODTon, tODToff) as a WRITE-1 command.
3. For the CAS-2 command immediately following an MPC command, the CAS-2 operands
must be driven LOW.
4. BL = 16, Write postamble = 0.5nCK.
5. DES commands are shown for ease of illustration; other commands may be valid at
these times.
184
200b: x32 LPDDR4 SDRAM
Write Training
Figure 126: Power-Down Exit to MPC[WRITE-FIFO] Timing
T0
Ta0
Tb0
Tb1
Tb2
Tc0
Tc1
Tc2
Tc3
Tc4
Tc5
Td0
Td1
Td2
Td3
Td4
Td5
Td6
Td7
Td8
Td9
CK_c
CK_t
t
CKCKEH
CKE
t
t
XP
MPCWR (= t RCD + 3nCK)
WL
CS
CA
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Note 1
DES
Command
DES
Any command Any command
DES
DES
DES
DES
MPC
[WRITE-FIFO]
CAS-2
DES
DES
DES
Don’t Care
Notes:
1. Any commands except MPC[WRITE-FIFO] and other exception commands defined other
section in this document (for example. MPC[READ DQ CALIBRATION]).
2. DES commands are shown for ease of illustration; other commands may be valid at
these times.
Table 119: MPC[WRITE-FIFO] AC Timing
Parameter
Symbol
tXP
has expired until
Additional time after
MPC[WRITE-FIFO] command may be issued
tMPCWR
MIN/MAX
MIN
Value
tRCD
+ 3nCK
Unit
–
Internal Interval Timer
As voltage and temperature change on the device, the DQS clock-tree delay will shift,
requiring retraining. The device includes an internal DQS clock-tree oscillator to measure the amount of delay over a given time interval (determined by the controller), allowing the controller to compare the trained delay value to the delay value seen at a later
time. The DQS oscillator will provide the controller with important information regarding the need to retrain and the magnitude of potential error.
The DQS interval oscillator is started by issuing an MPC command with OP[5:0] =
101011b, which will start an internal ring oscillator that counts the number of time a
signal propagates through a copy of the DQS clock tree.
The DQS oscillator may be stopped by issuing an MPC[STOP DQS OSCILLATOR] command with OP[6:0] set, as described in MPC Operation, or the controller may instruct
the SDRAM to count for a specific number of clocks and then stop automatically (See
MR23 for more information). If MR23 is set to automatically stop the DQS oscillator,
then the MPC[STOP DQS OSCILLATOR] command should not be used (illegal). When
the DQS oscillator is stopped by either method, the result of the oscillator counter is automatically stored in MR18 and MR19.
The controller may adjust the accuracy of the result by running the DQS interval oscillator for shorter (less accurate) or longer (more accurate) duration. The accuracy of the
185
200b: x32 LPDDR4 SDRAM
Write Training
result for a given temperature and voltage is determined by the following equation,
where run time = total time between START and STOP commands and DQS delay = the
value of the DQS clock tree delay (tDQS2DQ MIN/MAX):
DQS oscillator granularity error =
2 x (DQS delay)
run time
Additional matching error must be included, which is the difference between DQS
training circuit and the actual DQS clock tree across voltage and temperature. The
matching error is vendor specific. Therefore, the total accuracy of the DQS oscillator
counter is given by:
DQS oscillator accuracy = 1 - granularity error - matching error
For example, if the total time between START and STOP commands is 100ns, and the
maximum DQS clock tree delay is 800ps (tDQS2DQ MAX), then the DQS oscillator granularity error is:
DQS oscillator granularity error =
2 x (0.8ns)
= 1.6%
100ns
This equates to a granularity timing error of 12.8ps. Assuming a circuit matching error
of 5.5ps across voltage and temperature, the accuracy is:
DQS oscillator accuracy = 1 -
12.8 + 5.5
= 97.7%
800
For example, running the DQS oscillator for a longer period improves the accuracy. If
the total time between START and STOP commands is 500ns, and the maximum DQS
clock tree delay is 800ps (tDQS2DQ MAX), then the DQS oscillator granularity error is:
DQS oscillator granularity error =
2 x (0.8ns)
= 0.32%
500ns
This equates to a granularity timing error or 2.56ps. Assuming a circuit matching error
of 5.5ps across voltage and temperature, the accuracy is:
DQS oscillator accuracy = 1 -
2.56 + 5.5
= 99.0%
800
The result of the DQS interval oscillator is defined as the number of DQS clock tree delays that can be counted within the run time, determined by the controller. The result is
stored in MR18-OP[7:0] and MR19-OP[7:0].
MR18 contains the least significant bits (LSB) of the result, and MR19 contains the most
significant bits (MSB) of the result. MR18 and MR19 are overwritten by the SDRAM
when a MPC[STOP DQS OSCILLATOR] command is received.
The SDRAM counter will count to its maximum value (= 2^16) and stop. If the maximum value is read from the mode registers, the memory controller must assume that
the counter overflowed the register and therefore discard the result. The longest run
time for the oscillator that will not overflow the counter registers can be calculated as
follows:
Longest runtime interval = 216 x tDQS2DQ(MIN) = 216 × 0.2ns = 13.1μs
186
200b: x32 LPDDR4 SDRAM
Write Training
DQS Interval Oscillator Matching Error
The interval oscillator matching error is defined as the difference between the DQS
training ckt (interval oscillator) and the actual DQS clock tree across voltage and temperature.
Parameters:
tDQS2DQ: Actual DQS clock tree delay
tDQS
OSC: Training ckt (interval oscillator) delay
OSCOffset: Average delay difference over voltage and temperature (shown below)
OSCMatch: DQS oscillator matching error
Figure 127: Interval Oscillator Offset – OSCoffset
Offset 2
tDQS2DQ
tDQS
OSC
Time
(ps)
OSC offset = AVG(offset1, offset2)
Offset 1 (at end point) = tDQS2DQ(V,T) – tDQSOSC(V,T)
Offset 2 (at end point) = tDQS2DQ(V,T) – tDQSOSC(V,T)
Offset 1
Temperature(T)/Voltage(V)
OSCMatch :
OSCMatch = [ tDQS2DQ(V,T) - tDQSOSC (V,T) - OSCoffset ]
tDQS
tDQS
OSC:
OSC(V,T) =
[
Runtime
2 × Count
]
Table 120: DQS Oscillator Matching Error Specification
Parameter
Symbol
MIN
MAX
Unit
Notes
DQS oscillator matching error
OSCMatch
–20
20
ps
1, 2, 3, 4,
5, 6, 7, 8
DQS oscillator offset
OSCoffset
–100
100
ps
2, 4. 7
Notes:
1. The OSCMatch is the matching error per between the actual DQS and DQS interval oscillator over voltage and temperature.
187
200b: x32 LPDDR4 SDRAM
Write Training
2. This parameter will be characterized or guaranteed by design.
3. The OSCMatch is defined as the following:
OSCMatch =
[ tDQS2DQ(V, T) - tDQSOSC(V, T) - OSCoffset ]
Where tDQS2DQ(V,T) and tDQSOSC(V,T) are determined over the same voltage and temperature conditions.
4. The runtime of the oscillator must be at least 200ns for determining tDQSOSC(V,T).
tDQS
OSC(V,T) =
[
Runtime
2 × Count
]
5. The input stimulus for tDQS2DQ will be consistent over voltage and temperature conditions.
6. The OSCoffset is the average difference of the endpoints across voltage and temperature.
7. These parameters are defined per channel.
8. tDQS2DQ(V,T) delay will be the average of DQS-to-DQ delay over the runtime period.
OSC Count Readout Time
OSC Stop to its counting value readout timing is shown in following figures.
Figure 128: In Case of DQS Interval Oscillator is Stopped by MPC Command
T0
T1
T2
Valid
Valid
T3
T4
T5
Ta0
Ta1
Ta2
Valid
Valid
Ta3
Ta4
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Valid
Valid
Valid
Valid
Tb6
CK_c
CK_t
CKE
CS
CA
Command
DES
MPC :START
MROSCILLATOR
Write-2
DQS
DES
DES
DES
DES
MPC :STOP
DQS OSCILLATOR
DES
DES
DES
DES
MRR-1
MR18/MR19
CAS-2
t OSCO
Don’t Care
Note:
1. DQS interval timer run time setting :MR23 OP[7:0] = 00000000b.
188
200b: x32 LPDDR4 SDRAM
Write Training
Figure 129: In Case of DQS Interval Oscillator is Stopped by DQS Interval Timer
T0
T1
T2
Valid
Valid
T3
T4
T5
Ta0
Ta1
Ta2
Ta3
Ta4
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Valid
Valid
Valid
Valid
Tb6
CK_c
CK_t
CKE
CS
CA
Command
DES
MPC :START
MROSCILLATOR
Write-2
DQS
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
MRR-1
MR18/MR19
CAS-2
t OSCO
See Note 2
Don’t Care
Notes:
1. DQS interval timer run time setting: MR23 OP[7:0] ≠ 00000000b.
2. Setting counts of MR23.
Table 121: DQS Interval Oscillator AC Timing
Parameter
Delay time from OSC stop to mode register readout
Note:
Symbol
MIN/MAX
Value
Unit
tOSCO
MIN
MAX (40ns,
8nCK)
ns
1. START DQS OSCILLATOR command is prohibited until tOSCO(MIN) is satisfied.
189
200b: x32 LPDDR4 SDRAM
Thermal Offset
Thermal Offset
Because of tight thermal coupling, hot spots on an SOC can induce thermal gradients
across the device. Because these hot spots may not be located near the thermal sensor,
the temperature compensated self refresh (TCSR) circuit may not generate enough refresh cycles to guarantee memory retention. To address this shortcoming, the controller
can provide a thermal offset that the memory can use to adjust its TCSR circuit to ensure reliable operation.
This thermal offset is provided through MR4 OP[6:5] to either or both channels (dual
channel devices). This temperature offset may modify refresh behaviour for the channel
to which the offset is provided. It will take a maximum of 200μs to have the change reflected in MR4 OP[2:0] for the channel to which the offset is provided. If the induced
thermal gradient from the device temperature sensor location to the hot spot location
of the controller is greater than 15°C, self refresh mode will not reliably maintain memory contents.
To accurately determine the temperature gradient between the memory thermal sensor
and the induced hot spot, the memory thermal sensor location must be provided to the
controller.
Support of the thermal offset function is optional. Please refer to the vendor data sheet
to determine if this function is supported.
Temperature Sensor
The device has a temperature sensor that can be read from MR4. This sensor can be
used to determine the appropriate refresh rate, to determine whether AC timing de-rating is required at an elevated temperature range, and to monitor the operating temperature. Either the temperature sensor or the device T OPER can be used to determine if operating temperature requirements are being met.
The device monitors device temperature and updates MR4 according to tTSI. Upon exiting self refresh or power-down, the device temperature status bits shall be no older than
tTSI.
When using the temperature sensor, the actual device case temperature may be higher
than the T OPER specification that applies to standard or elevated temperature ranges.
For example, T CASE may be above 85°C when MR4[2:0] = b011. The device enables a 2°C
temperature margin between the point when the device updates the MR4 value and the
point when the controller reconfigures the system accordingly. When performing tight
thermal coupling of the device to external hot spots, the maximum device temperature
may be higher than indicated by MR4.
To ensure proper operation when using the temperature sensor, consider the following:
• TempGradient is the maximum temperature gradient experienced by the device at the
temperature of interest over a range of 2°C.
• ReadInterval is the time period between MR4 reads from the system.
• TempSensorInterval (tTSI) is the maximum delay between the internal updates of
MR4.
• SysRespDelay is the maximum time between a read of MR4 and a response from the
system.
190
200b: x32 LPDDR4 SDRAM
ZQ Calibration
In order to determine the required frequency of polling MR4, the system uses the TempGradient and the maximum response time of the system in the following equation:
TempGradient × (ReadInterval + tTSI + SysRespDelay) 2°C
Table 122: Temperature Sensor
Parameter
Symbol
Max/Min
Value
Unit
TempGradient
MAX
System Dependent
°C/s
ReadInterval
MAX
System Dependent
ms
tTSI
MAX
32
ms
System response delay
SysRespDelay
MAX
System Dependent
ms
Device temperature margin
TempMargin
MAX
2
°C
System temperature gradient
MR4 read interval
Temperature sensor interval
For example, if TempGradient is 10°C/s and the SysRespDelay is 1ms:
(10°C/s) x (ReadInterval + 32ms + 1ms) 2°C
In this case, ReadInterval shall be no greater than 167ms.
Figure 130: Temperature Sensor Timing
Temperature
< [tTSI + ReadInterval + SysRespDelay]
Device
temperature
margin
ient
Grad
2°
p
Tem
MR4
trip level
tTSI
MR4 = 0x03
MR4 = 0x06
Temperature
sensor
update
Host
MR4 read
MR4 = 0x06
MR4 = 0x06
ReadInterval
MRR MR4 = 0x06
MR4 = 0x06
MR4 = 0x06
Time
SysRespDelay
MRR MR4 = 0x06
ZQ Calibration
The MPC command is used to initiate ZQ calibration, which calibrates the output driver
impedance and CA/DQ ODT impedance across process, temperature, and voltage. ZQ
calibration occurs in the background of device operation and is designed to eliminate
any need for coordination between channels (that is, it allows for channel independence). ZQ calibration is required each time that the PU-Cal value (MR3-OP[0]) is
changed. Additional ZQ CALIBRATION commands may be required as the voltage and
temperature change in the system environment. CA ODT values (MR11-OP[6:4]) and
191
200b: x32 LPDDR4 SDRAM
ZQ Calibration
DQ ODT values (MR11-OP[2:0]) may be changed without performing ZQ calibration, as
long as the PU-Cal value doesn’t change.
There are two ZQ calibration modes initiated with the MPC command: ZQCAL START
and ZQCAL LATCH. ZQCAL START initiates the calibration procedure, and ZQCAL
LATCH captures the result and loads it into the drivers.
A ZQCAL START command may be issued anytime the device is not in a power-down
state. A ZQCAL LATCH command may be issued anytime outside of power-down after
tZQCAL has expired and all DQ bus operations have completed. The CA bus must maintain a deselect state during tZQLAT to allow CA ODT calibration settings to be updated.
The DQ calibration value will not be updated until ZQCAL LATCH is performed and
tZQLAT has been met. The following mode register fields that modify I/O parameters
cannot be changed following a ZQCAL START command and before tZQCAL has expired:
•
•
•
•
PU-Cal (pull-up calibration V OH point)
PDDS (pull-down drive strength and Rx termination)
DQ ODT (DQ ODT value)
CA ODT (CA ODT value)
ZQCAL Reset
The ZQCAL RESET command resets the output impedance calibration to a default accuracy of ±30% across process, voltage, and temperature. This command is used to ensure output impedance accuracy to ±30% when ZQCAL START and ZQCAL LATCH commands are not used.
The ZQCAL RESET command is executed by writing MR10-OP[0] = 1B.
Table 123: ZQ Calibration Parameters
Parameter
Symbol
Min/Max
Value
Unit
ZQCAL START to ZQCAL LATCH command interval
tZQCAL
MIN
1
μs
ZQCAL LATCH to next valid command interval
tZQLAT
MIN
MAX (30ns, 8nCK)
ns
ZQCAL RESET to next valid command interval
tZQRESET
MIN
MAX (50ns, 3nCK)
ns
192
200b: x32 LPDDR4 SDRAM
ZQ Calibration
Figure 131: ZQCAL Timing
T0
T1
T2
T3
T4
T5
T6
T7
T8
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Ta6
Ta7
Tb0
Tb1
Tb2
Tb3
Tb4
Tb5
Tb6
Tb7
Tc0
Tc1
Tc2
Tc3
Tc4
Valid
Valid
Tc5
Tc6
DES
DES
CK_c
CK_t
tZQLAT
tZQCAL
CA
ZQCAL ZQCAL
START START
WR
WR
CAS
ZQCAL ZQCAL
LATCH LATCH
CAS
WL
Command
MPC
TRAIN/CAL
DES
DES
WRITE
CAS2
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
MPC
TRAIN/CAL
DES
DES
DES
DES
PRECHARGE
tDQSS
DQS_t
DQS_c
tWPRE
tWPST
tDQS2DQ
DQ[15:0]
Transitioning Data
Notes:
Don’t Care
1. WRITE and PRECHARGE operations are shown for illustrative purposes. Any single or
multiple valid commands may be executed within the tZQCAL time and prior to latching
the results.
2. Before the ZQCAL LATCH command can be executed, any prior commands that utilize
the DQ bus must have completed. WRITE commands with DQ termination must be given
enough time to turn off the DQ ODT before issuing the ZQCAL LATCH command. See
the ODT section for ODT timing.
Multichannel Considerations
The device includes a single ZQ pin and associated ZQ calibration circuitry. Calibration
values from this circuit will be used by both channels according to the following protocol:
• The ZQCAL START command can be issued to either or both channels.
• The ZQCAL START command can be issued when either or both channels are executing other commands, and other commands can be issued during tZQCAL.
• The ZQCAL START command can be issued to both channels simultaneously.
• The ZQCAL START command will begin the calibration unless a previously requested
ZQ calibration is in progress.
• If the ZQCAL START command is received while a ZQ calibration is in progress, the
command will be ignored and the in-progress calibration will not be interrupted.
• The ZQCAL LATCH command is required for each channel.
• The ZQCAL LATCH command can be issued to both channels simultaneously.
• The ZQCAL LATCH command will latch results of the most recent ZQCAL START
command provided tZQCAL has been met.
• ZQCAL LATCH commands that do not meet tZQCAL will latch the results of the most
recently completed ZQ calibration.
• The ZQRESET MRW commands will only reset the calibration values for the channel
issuing the command.
In compliance with complete channel independence, either channel may issue ZQCAL
START and ZQCAL LATCH commands as needed without regard to the state of the other
channel.
193
200b: x32 LPDDR4 SDRAM
ZQ Calibration
ZQ External Resistor, Tolerance, and Capacitive Loading
To use the ZQ CALIBRATION function, a 240 ohm, ±1% tolerance external resistor must
be connected between the ZQ pin and V DDQ.
If the system configuration shares the CA bus to form a x32 (or wider) channel, the ZQ
pin of each die’s x16 channel must use a separate ZQCAL resistor.
If the system configuration has more than one rank, and if the ZQ pins of both ranks are
attached to a single resistor, then the SDRAM controller must ensure that the ZQCAL's
don’t overlap.
The total capacitive loading on the ZQ pin must be limited to 25pF. For example, if a system configuration shares a CA bus between n channels to form an n x16 wide bus, and
no means are available to control the ZQCAL separately for each channel (that is, separate CS, CKE, or CK), then each x16 channel must have a separate ZQCAL resistor. For a
x32, two-rank system, each x16 channel must have its own ZQCAL resistor, but the
ZQCAL resistor can be shared between ranks on each x16 channel. In this configuration,
the CS signal can be used to ensure that the ZQCAL commands for Rank[0] and Rank[1]
don’t overlap.
194
200b: x32 LPDDR4 SDRAM
Frequency Set Points
Frequency Set Points
Frequency set points enable the CA bus to be switched between two differing operating
frequencies with changes in voltage swings and termination values, without ever being
in an untrained state, which could result in a loss of communication to the device. This
is accomplished by duplicating all CA bus mode register parameters, as well as other
mode register parameters commonly changed with operating frequency.
These duplicated registers form two sets that use the same mode register addresses,
with read/write access controlled by MR bit FSP-WR (frequency set point write/read)
and the operating point controlled by MR bit FSP-OP (FREQUENCY SET POINT operation). Changing the FSP-WR bit enables MR parameters to be changed for an alternate
frequency set-point without affecting the current operation.
Once all necessary parameters have been written to the alternate set point, changing
the FSP-OP bit will switch operation to use all of the new parameters simultaneously
(within tFC), eliminating the possibility of a loss of communication that could be
caused by a partial configuration change.
Parameters which have two physical registers controlled by FSP-WR and FSP-OP include:
Table 124: Mode Register Function with Two Physical Registers
MR Number
Operand
MR1
OP[1:0]
WR-PRE (Write preamble length)
RD-PRE (Read preamble type)
MR12
RD-PST (Read postamble length)
RL (READ latency)
OP[5:3]
WL (WRITE latency)
OP[6]
WLS (WRITE latency set)
OP[0]
PU-CAL (Pull-up calibration point)
OP[1]
WR-PST (Write postamble length)
PDDS (Pull-down drive strength)
OP[6]
DBI-RD (DBI-read enable)
OP[7]
DBI-WR (DBI-write enable)
OP[2:0]
DQ ODT (DQ bus receiver on-die termination)
OP[6:4]
CA ODT (CA bus receiver on-die termination)
OP[5:0]
VREF(CA) (VREF(CA) setting)
OP[6]
MR14
nWR (Write-recovery for AUTO PRECHARGE command)
OP[2:0]
OP[5:3]
MR11
BL (Burst length)
OP[3]
OP[7]
MR3
Notes
OP[2]
OP[6:4]
MR2
Function
OP[5:0]
OP[6]
VRCA (VREF(CA) range)
VREF(DQ) (VREF(DQ) setting)
VRDQ (VREF(DQ) range)
195
1
200b: x32 LPDDR4 SDRAM
Frequency Set Points
Table 124: Mode Register Function with Two Physical Registers (Continued)
MR Number
Operand
MR22
OP[2:0]
Function
Notes
SOC ODT (Controller ODT value for VOH calibration)
OP[3]
ODTE-CK (CK ODT enabled for non-terminating rank)
OP[4]
ODTE-CS (CS ODT enable for non-terminating rank)
OP[5]
ODTD-CA (CA ODT termination disable)
1. PU-CAL setting is required as the same value for both Ch. A and Ch. B before issuing
ZQCAL START command. See Mode Register Definition section for more details.
Note:
Table below shows how the two mode registers for each of the parameters above can be
modified by setting the appropriate FSP-WR value, and how device operation can be
switched between operating points by setting the appropriate FSP-OP value. The FSPWR and FSP-OP functions operate completely independently.
Table 125: Relation Between MR Setting and DRAM Operation
Function
MR# and Operand
Data
FSP-WR
MR13 OP[6]
0 (default)
Data write to mode register N for FSP-OP[0] by MRW command.
1
Data write to mode register N for FSP-OP[1] by MRW command.
FSP-OP
MR13 OP[7]
Notes:
Operation
Notes
0 (default)
DRAM operates with mode register N for FSP-OP[0] setting.
1
DRAM operates with mode register N for FSP-OP[1] setting.
1
2
1. FSP-WR stands for frequency set point write/read.
2. FSP-OP stands for frequency set point operating point.
Frequency Set Point Update Timing
The frequency set point update timing is shown below. When changing the frequency
set point via MR13 OP[7], the V RCG setting: MR13 OP[3] have to be changed into V REF
fast response (high current) mode at the same time. After frequency change time (tFC) is
satisfied. V RCG can be changed into normal operation mode via MR13 OP[3].
196
200b: x32 LPDDR4 SDRAM
Frequency Set Points
Figure 132: Frequency Set Point Switching Timing
.
T0
T1
T2
T3
T4
T5
Ta0
Ta1
Tb0
Tb1
Tc0
Tc1
Tc2
Tc3
Tc4
Tc5
Tc6
CK_c
CK_t
Note 1
Frequency
change
tCKFSPE
tVRCG_DISABLE
tCKFSPX
CKE
CS
CA
DES
Command
DES
MRW-1
MRW-1
MRW-2
MRW-2
FSP changes from 0 to 1
VRCG changes from normal to HIGH current
DES
DES
DES
DES
DES
MRW-1 MRW-1 MRW-2 MRW-2
DES
DES
DES
DES
DES
DES
VRCG changes
from HIGH current to normal
DES
tFC_short/middle/long
Applicable
mode register
Switching mode register
Mode register for FSP-OP0
Mode register for FSP-OP1
Don’t Care
Note:
1. For frequency change during frequency set point switching, refer to input Clock Stop
and Frequency Change section.
Table 126: Frequency Set Point AC Timing
Parameter
Symbol
Min/
Max
Frequency set point switching time
tFC_short
MIN
tFC_middle
MIN
tFC_long
Valid clock requirement after entering FSP
change
tCKFSPE
Valid clock requirement before first valid command after FSP change
tCKFSPX
Note:
Data Rate
1600
3200
3733
4267
Unit
Notes
200
ns
1
200
ns
MIN
250
ns
MIN
MAX(7.5ns, 4nCK)
–
MIN
MAX(7.5ns, 4nCK)
–
1. Frequency set point switching time depends on value of VREF(CA) setting: MR12 OP[5:0]
and VREF(CA) range: MR12 OP[6] of FSP-OP 0 and 1. The details are shown in table below.
Additionally change of frequency set point may affect VREF(DQ) setting. Settling time of
VREF(DQ) level is the same as VREF(CA) level.
Table 127: tFC Value Mapping
Step Size
Range
Application
From FSP-OP0
To FSP-OP1
From FSP-OP0
To FSP-OP1
tFC_short
Base
A single step size increment/decrement
Base
No change
197
200b: x32 LPDDR4 SDRAM
Frequency Set Points
Table 127: tFC Value Mapping (Continued)
Step Size
Range
Application
From FSP-OP0
To FSP-OP1
From FSP-OP0
To FSP-OP1
tFC_middle
Base
Two or more step size increment/decrement
Base
No change
tFC_long
–
–
Base
Change
Note:
1. As well as change from FSP-OP1 to FSP-OP0.
Table 128: tFC Value Mapping: Example
Case
From/To
FSP-OP:
MR13 OP[7]
VREF(CA) Setting:
MR12: OP[5:0]
VREF(CA) Range:
MR12 OP[6]
Application
Notes
1
From
0
001100
0
tFC_short
1
To
1
001101
0
2
From
0
001100
0
tFC_middle
2
To
1
001110
0
3
From
0
Don't Care
0
tFC_long
3
To
1
Don't Care
1
Notes:
1. A single step size increment/decrement for VREF(CA) setting value.
2. Two or more step size increment/decrement for VREF(CA) setting value.
3. VREF(CA) range is changed. In this case, changing VREF(CA) setting doesn’t affect tFC value.
The LPDDR4 SDRAM defaults to FSP-OP[0] at power-up. Both set points default to settings needed to operate in un-terminated, low-frequency environments. To enable the
device to operate at higher frequencies, Command bus training mode should be utilized to train the alternate frequency set point. See Command Bus Training section for
more details on this training mode.
198
200b: x32 LPDDR4 SDRAM
Frequency Set Points
Figure 133: Training for Two Frequency Set Points
Power-up/
Initialization
Prepare for CA bus
training of FSP1 for
high frequency
FSP-OP = 0
FSP-WR = 0
Freq. = Boot
FSP-OP = 0
FSP-WR = 1
Freq. = Boot
CA bus training,
FSP-OP1
CKE High to Low
FSP-WR = 1
Freq. = High
CKE LOW to HIGH
Exit CA bus
training
Switch to highspeed mode
FSP-OP = 0
FSP-WR = 1
Freq. = Boot
FSP-OP = 1
FSP-WR = 1
Freq. = High
Prepare for CA bus
training of FSP0 for
medium frequency
FSP-OP = 1
FSP-WR = 0
Freq. = High
CKE HIGH to LOW
CA bus training,
FSP-OP0
Exit CA bus
training
FSP-WR = 0
Freq. = Medium
FSP-OP = 1
FSP-WR = 0
Freq. = High
CKE LOW to HIGH
Operate at
high speed
Once both of the frequency set points have been trained, switching between points can
be performed with a single MRW followed by waiting for time tFC.
Figure 134: Example of Switching Between Two Trained Frequency Set Points
State n-1:
FSP-OP = 1
MRW command
State n: FSP-OP = 0
Operate at
high speed
State n-1:
FSP-OP = 0
MRW command
State n: FSP-OP = 1
tFC
Operate at
medium speed
Operate at
high speed
tFC
Switching to a third (or more) set point can be accomplished if the memory controller
has stored the previously-trained values (in particular the V REF(CA) calibration value)
and rewrites these to the alternate set point before switching FSP-OP.
199
200b: x32 LPDDR4 SDRAM
Pull-Up and Pull-Down Characteristics and Calibration
Figure 135: Example of Switching to a Third Trained Frequency Set Point
State n-1:
FSP-WR = 1
MRW command
State n: FSP-WR = 0
Operate at
high speed
State n-1:
FSP-OP = 1
MRW command
State n: FSP-OP = 0
MRW command
{VREF(CA)
CA ODT, DQ ODT,
RL, WL, VREF(DQ),
ODTD-CA...}
tFC
Operate at
third speed
tFC
Pull-Up and Pull-Down Characteristics and Calibration
Table 129: Pull-Down Driver Characteristics – ZQ Calibration
RONPD,nom
Register
Min
Nom
Max
Unit
40 Ohm
RON40PD
0.90
1.0
1.10
RZQ/6
48 Ohm
RON48PD
0.90
1.0
1.10
RZQ/5
60 Ohm
RON60PD
0.90
1.0
1.10
RZQ/4
80 Ohm
RON80PD
0.90
1.0
1.10
RZQ/3
120 Ohm
RON120PD
0.90
1.0
1.10
RZQ/2
240 Ohm
RON240PD
0.90
1.0
1.10
RZQ/1
1. All value are after ZQ calibration. Without ZQ calibration, RONPD values are ±30%.
Note:
Table 130: Pull-Up Characteristics – ZQ Calibration
VOHPU,nom
VOH,nom
Min
Nom
Max
Unit
VDDQ/2.5
440
0.90
1.0
1.10
VOH,nom
VDDQ/3
367
0.90
1.0
1.10
VOH,nom
Notes:
1. All value are after ZQ calibration. Without ZQ calibration, RONPD values are ±30%.
2. VOH,nom (mV) values are based on a nominal VDDQ = 1.1V.
Table 131: Valid Calibration Points
ODT Value
VOHPU
240
120
80
60
48
40
VDDQ/2.5
Valid
Valid
Valid
DNU
DNU
DNU
200
200b: x32 LPDDR4 SDRAM
On-Die Termination for the Command/Address Bus
Table 131: Valid Calibration Points (Continued)
ODT Value
VOHPU
240
120
80
60
48
40
VDDQ/3
Valid
Valid
Valid
Valid
Valid
Valid
Notes:
1. Once the output is calibrated for a given VOH(nom) calibration point, the ODT value may
be changed without recalibration.
2. If the VOH(nom) calibration point is changed, then recalibration is required.
3. DNU = Do not use.
On-Die Termination for the Command/Address Bus
The on-die termination (ODT) feature allows the device to turn on/off termination resistance for CK_t, CK_c, CS, and CA[5:0] signals without the ODT control pin. The ODT
feature is designed to improve signal integrity of the memory channel by allowing the
DRAM controller to turn on and off termination resistance for any target DRAM devices
via the mode register setting.
A simple functional representation of the DRAM ODT feature is shown below.
Figure 136: ODT for CA
RTT = VOUT
|IOUT|
VDD2
To other
circuitry
like RCV, ...
ODT
CA
IOUT
RTT
VOUT
VSS
ODT Mode Register and ODT State Table
ODT termination values are set and enabled via MR11. The CA bus (CK_t, CK_c, CS,
CA[5:0]) ODT resistance values are set by MR11 OP[6:4]. The default state for the CA is
ODT disabled.
ODT is applied on the CA bus to the CK_t, CK_c, CS, and CA signals. The CA ODT of the
device is designed to enable one rank to terminate the entire command bus in a multirank system, so only one termination load will be present even if multiple devices are
201
200b: x32 LPDDR4 SDRAM
On-Die Termination for the Command/Address Bus
sharing the command signals. For this reason, CA ODT remains on, even when the device is in the power-down or self refresh power-down state.
The die has a bond pad (ODT_CA) for multirank operations. When the ODT_CA pad is
LOW, the die will not terminate the CA bus regardless of the state of the mode register
CA ODT bits (MR11 OP[6:4]). If, however, the ODT_CA bond pad is HIGH and the mode
register CA ODT bits are enabled, the die will terminate the CA bus with the ODT values
found in MR11 OP[6:4]. In a multirank system, the terminating rank should be trained
first, followed by the non-terminating rank(s).
Table 132: Command Bus ODT State
CA ODT
MR11[6:4]
ODT_CA
Bond Pad
ODTD-CA
MR22 OP[5]
ODTE-CK
MR22 OP[3]
ODTE-CS
MR22 OP[4]
ODT State
for CA
ODT State
for CK
ODT State
for CS
Disabled1
Valid2
Valid3
Valid3
Valid3
Off
Off
Off
0
Valid3
0
0
Off
Off
Off
Valid
3
Valid
3
0
Valid3
0
1
Off
Off
On
Valid
3
0
Valid3
1
0
Off
On
Off
Valid 3
0
Valid3
1
1
Off
On
On
0
Valid3
Valid3
On
On
On
1
Valid3
Valid3
Off
On
On
Valid
3
Valid
3
1
1
Notes:
1.
2.
3.
4.
Default value.
Valid = H or L (a defined logic level)
Valid = 0 or 1.
The state of ODT_CA is not changed when the device enters power-down mode. This
maintains termination for alternate ranks in multirank systems.
ODT Mode Register and ODT Characteristics
Table 133: ODT DC Electrical Characteristics – up to 3200 Mbps
RZQ Ω ±1% over entire operating range after calibration
MR11 OP[6:4]
RTT
VOUT
001b
010b
011b
100b
Ω
Ω
Ω
Ω
Min
Nom
Max
Unit
Notes
VOL(DC) = 0.1 × VDD2
0.8
1.0
1.1
RZQ/1
1, 2
VOM(DC) = 0.33 × VDD2
0.9
1.0
1.1
VOH(DC) = 0.5 × VDD2
0.9
1.0
1.2
RZQ/2
1, 2
RZQ/3
1, 2
RZQ/4
1, 2
VOL(DC) = 0.1 × VDD2
0.8
1.0
1.1
VOM(DC) = 0.33 × VDD2
0.9
1.0
1.1
VOH(DC) = 0.5 × VDD2
0.9
1.0
1.2
VOL(DC) = 0.1 × VDD2
0.8
1.0
1.1
VOM(DC) = 0.33 × VDD2
0.9
1.0
1.1
VOH(DC) = 0.5 × VDD2
0.9
1.0
1.2
VOL(DC) = 0.1 × VDD2
0.8
1.0
1.1
V OM(DC) = 0.33 × VDD2
0.9
1.0
1.1
VOH(DC) = 0.5 × VDD2
0.9
1.0
1.2
202
200b: x32 LPDDR4 SDRAM
On-Die Termination for the Command/Address Bus
Table 133: ODT DC Electrical Characteristics – up to 3200 Mbps (Continued)
RZQ Ω ±1% over entire operating range after calibration
MR11 OP[6:4]
RTT
VOUT
101b
110b
Ω
Ω
Mismatch, CA -CA within clock
group
Notes:
Min
Nom
Max
Unit
Notes
VOL(DC) = 0.1 × VDD2
0.8
1.0
1.1
RZQ/5
1, 2
VOM(DC) = 0.33 × VDD2
0.9
1.0
1.1
VOH(DC) = 0.5 × VDD2
0.9
1.0
1.2
VOL(DC) = 0.1 × VDD2
0.8
1.0
1.1
RZQ/6
1, 2
VOM(DC) = 0.33 × VDD2
0.9
1.0
1.1
VOH(DC) = 0.5 × VDD2
0.9
1.0
1.2
0.33 × VDD2
–
–
2
%
1, 2, 3
1. The tolerance limits are specified after calibration with stable temperature and voltage.
To understand the behavior of the tolerance limits when voltage or temperature
changes after calibration, see the section on voltage and temperature sensitivity.
2. Pull-down ODT resistors are recommended to be calibrated at 0.33 × VDD2. Other calibration points may be required to achieve the linearity specification shown above, for
example, calibration at 0.5 × VDD2 and 0.1 × VDD2.
3. CA to CA mismatch within clock group variation for a given component including CK_t,
CK_c ,and CS (characterized).
CA-to-CA mismatch = RODT (MAX) - RODT (MIN)
RODT (AVG)
Table 134: ODT DC Electrical Characteristics – Beyond 3200 Mbps
RZQ Ω ±1% over entire operating range after calibration
MR11 OP[6:4]
RTT
VOUT
001b
010b
011b
100b
101b
Ω
Ω
Ω
Ω
Ω
Min
Nom
Max
Unit
Notes
RZQ/1
1, 2
RZQ/2
1, 2
RZQ/3
1, 2
RZQ/4
1, 2
RZQ/5
1, 2
VOL(DC) = 0.1 × VDD2
0.8
1.0
1.1
VOM(DC) = 0.33 × VDD2
0.9
1.0
1.1
VOH(DC) = 0.5 × VDD2
0.9
1.0
1.3
VOL(DC) = 0.1 × VDD2
0.8
1.0
1.1
VOM(DC) = 0.33 × VDD2
0.9
1.0
1.1
VOH(DC) = 0.5 × VDD2
0.9
1.0
1.3
VOL(DC) = 0.1 × VDD2
0.8
1.0
1.1
VOM(DC) = 0.33 × VDD2
0.9
1.0
1.1
VOH(DC) = 0.5 × VDD2
0.9
1.0
1.3
VOL(DC) = 0.1 × VDD2
0.8
1.0
1.1
V OM(DC) = 0.33 × VDD2
0.9
1.0
1.1
VOH(DC) = 0.5 × VDD2
0.9
1.0
1.3
VOL(DC) = 0.1 × VDD2
0.8
1.0
1.1
VOM(DC) = 0.33 × VDD2
0.9
1.0
1.1
VOH(DC) = 0.5 × VDD2
0.9
1.0
1.3
203
200b: x32 LPDDR4 SDRAM
On-Die Termination for the Command/Address Bus
Table 134: ODT DC Electrical Characteristics – Beyond 3200 Mbps (Continued)
RZQ Ω ±1% over entire operating range after calibration
MR11 OP[6:4]
RTT
VOUT
Min
Nom
Max
Unit
Notes
VOL(DC) = 0.1 × VDD2
0.8
1.0
1.1
RZQ/6
1, 2
VOM(DC) = 0.33 × VDD2
0.9
1.0
1.1
VOH(DC) = 0.5 × VDD2
0.9
1.0
1.3
0.33 × VDD2
–
–
2
%
1, 2, 3
Ω
110b
Mismatch, CA -CA within clock
group
Notes:
1. The tolerance limits are specified after calibration with stable temperature and voltage.
To understand the behavior of the tolerance limits when voltage or temperature
changes after calibration, see the section on voltage and temperature sensitivity.
2. Pull-down ODT resistors are recommended to be calibrated at 0.33 × VDD2. Other calibration points may be required to achieve the linearity specification shown above, e.g.
calibration at 0.5 × VDD2 and 0.1 × VDD2.
3. CA to CA mismatch within clock group variation for a given component including CK_t,
CK_c ,and CS (characterized).
CA-to-CA mismatch = RODT (MAX) - RODT (MIN)
RODT (AVG)
ODT for CA Update Time
Figure 137: ODT for CA Setting Update Timing in 4-Clock Cycle Command
T0
T1
T2
T3
T4
T5
Ta
Ta1
Ta2
Ta3
Ta4
Ta5
Ta6
Ta7
Ta8
DES
MRW1
MRW1
MRW2
MRW2
DES
DES
DES
DES
DES
Valid 1
Valid 1
Valid 1
Valid 1
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
Valid
CK_c
CK_t
CKE
CS_n
Command
CA[5:0]
CA ODT
Old setting value
Updating setting
New setting value
tODTUP
Don’t Care
204
200b: x32 LPDDR4 SDRAM
DQ On-Die Termination
DQ On-Die Termination
On-die termination (ODT) is a feature that allows the device to turn on/off termination
resistance for each DQ, DQS, and DMI signal without the ODT control pin. The ODT
feature is designed to improve signal integrity of the memory channel by allowing the
DRAM controller to turn on and off termination resistance for any target DRAM devices
during WRITE or MASK WRITE operation.
The ODT feature is off and cannot be supported in power-down and self refresh modes.
The switch is enabled by the internal ODT control logic, which uses the WRITE-1 or
MASK WRITE-1 command and other mode register control information. The value of
RTT is determined by the MR bits.
RTT = VOUT
|IOUT|
Figure 138: Functional Representation of DQ ODT
VDDQ
To other
circuitry
like RCV, ...
ODT
DQ
IOUT
VOUT
RTT
VSSQ
Table 135: ODT DC Electrical Characteristics – up to 3200 Mbps
RZQ Ω ±1% over entire operating range after calibration
MR11 OP[2:0]
RTT
VOUT
001b
010b
011b
Ω
Ω
Ω
Min
Nom
Max
Unit
Notes
VOL(DC) = 0.1 × VDDQ
0.8
1.0
1.1
RZQ/1
1, 2
VOM(DC) = 0.33 × VDDQ
0.9
1.0
1.1
VOH(DC) = 0.5 × VDDQ
0.9
1.0
1.2
VOL(DC) = 0.1 × VDDQ
0.8
1.0
1.1
RZQ/2
1, 2
VOM(DC) = 0.33 × VDDQ
0.9
1.0
1.1
RZQ/3
1, 2
VOH(DC) = 0.5 × VDDQ
0.9
1.0
1.2
VOL(DC) = 0.1 × VDDQ
0.8
1.0
1.1
VOM(DC) = 0.33 × VDDQ
0.9
1.0
1.1
VOH(DC) = 0.5 × VDDQ
0.9
1.0
1.2
205
200b: x32 LPDDR4 SDRAM
DQ On-Die Termination
Table 135: ODT DC Electrical Characteristics – up to 3200 Mbps (Continued)
RZQ Ω ±1% over entire operating range after calibration
MR11 OP[2:0]
RTT
VOUT
100b
101b
110b
Ω
Ω
Ω
Mismatch error, DQ-to-DQ within a channel
Notes:
Min
Nom
Max
Unit
Notes
VOL(DC) = 0.1 × VDDQ
0.8
1.0
1.1
RZQ/4
1, 2
V OM(DC) = 0.33 × VDDQ
0.9
1.0
1.1
RZQ/5
1, 2
RZQ/6
1, 2
%
1, 2, 3
VOH(DC) = 0.5 × VDDQ
0.9
1.0
1.2
VOL(DC) = 0.1 × VDDQ
0.8
1.0
1.1
VOM(DC) = 0.33 × VDDQ
0.9
1.0
1.1
VOH(DC) = 0.5 × VDDQ
0.9
1.0
1.2
VOL(DC) = 0.1 × VDDQ
0.8
1.0
1.1
VOM(DC) = 0.33 × VDDQ
0.9
1.0
1.1
VOH(DC) = 0.5 × VDDQ
0.9
1.0
1.2
0.33 × VDDQ
–
–
2
1. The ODT tolerance limits are specified after calibration with stable temperature and
voltage. To understand the behavior of the tolerance limits when voltage or temperature changes after calibration, see the following section on voltage and temperature
sensitivity.
2. Pull-down ODT resistors are recommended to be calibrated at 0.33 × VDDQ. Other calibration points may be required to achieve the linearity specification shown above, (for
example, calibration at 0.5 × VDDQ and –0.1 × VDDQ.
3. DQ-to-DQ mismatch within byte variation for a given component, including DQS (characterized).
DQ-to-DQ mismatch= RODT (MAX) - RODT (MIN)
RODT (AVG)
Table 136: ODT DC Electrical Characteristics – Beyond 3200 Mbps
RZQ Ω ±1% over entire operating range after calibration
MR11 OP[2:0]
RTT
VOUT
001b
010b
011b
100b
Ω
Ω
Ω
Ω
Min
Nom
Max
Unit
Notes
VOL(DC) = 0.1 × VDDQ
0.8
1.0
1.1
RZQ/1
1, 2
VOM(DC) = 0.33 × VDDQ
0.9
1.0
1.1
VOH(DC) = 0.5 × VDDQ
0.9
1.0
1.3
RZQ/2
1, 2
RZQ/3
1, 2
RZQ/4
1, 2
VOL(DC) = 0.1 × VDDQ
0.8
1.0
1.1
VOM(DC) = 0.33 × VDDQ
0.9
1.0
1.1
VOH(DC) = 0.5 × VDDQ
0.9
1.0
1.3
VOL(DC) = 0.1 × VDDQ
0.8
1.0
1.1
VOM(DC) = 0.33 × VDDQ
0.9
1.0
1.1
VOH(DC) = 0.5 × VDDQ
0.9
1.0
1.3
VOL(DC) = 0.1 × VDDQ
0.8
1.0
1.1
V OM(DC) = 0.33 × VDDQ
0.9
1.0
1.1
VOH(DC) = 0.5 × VDDQ
0.9
1.0
1.3
206
200b: x32 LPDDR4 SDRAM
DQ On-Die Termination
Table 136: ODT DC Electrical Characteristics – Beyond 3200 Mbps (Continued)
RZQ Ω ±1% over entire operating range after calibration
MR11 OP[2:0]
RTT
VOUT
Ω
101b
Ω
110b
Mismatch error, DQ-to-DQ within a channel
Notes:
Min
Nom
Max
Unit
Notes
VOL(DC) = 0.1 × VDDQ
0.8
1.0
1.1
RZQ/5
1, 2
VOM(DC) = 0.33 × VDDQ
0.9
1.0
1.1
RZQ/6
1, 2
%
1, 2, 3
VOH(DC) = 0.5 × VDDQ
0.9
1.0
1.3
VOL(DC) = 0.1 × VDDQ
0.8
1.0
1.1
VOM(DC) = 0.33 × VDDQ
0.9
1.0
1.1
VOH(DC) = 0.5 × VDDQ
0.9
1.0
1.3
0.33 × VDDQ
–
–
2
1. The ODT tolerance limits are specified after calibration with stable temperature and
voltage. To understand the behavior of the tolerance limits when voltage or temperature changes after calibration, see the following section on voltage and temperature
sensitivity.
2. Pull-down ODT resistors are recommended to be calibrated at 0.33 × VDDQ. Other calibration points may be required to achieve the linearity specification shown above, for
example, calibration at 0.5 × VDDQ and –0.1 × VDDQ.
3. DQ-to-DQ mismatch within byte variation for a given component, including DQS (characterized).
DQ-to-DQ mismatch= RODT (MAX) - RODT (MIN)
RODT (AVG)
Output Driver and Termination Register Temperature and Voltage Sensitivity
When temperature and/or voltage change after calibration, the tolerance limits are widen according to the tables below.
Table 137: Output Driver and Termination Register Sensitivity Definition
Resistor
Definition
Point
Min
Max
Unit
Notes
RONPD
0.33 × VDDQ
90 - (dRONdT ΔT|) - (dRONdV ΔV|)
110 + (dRONdT ΔT|) + (dRONdV ΔV|)
%
1, 2
VOHPU
0.33 × VDDQ
90 - (dVOHdT ΔT|) - (dVOHdV ΔV|)
110 + (dVOHdT ΔT|) + (dVOHdV ΔV|)
1, 2, 5
RTT(I/O)
0.33 × VDDQ
90 - (dRONdT ΔT|) - (dRONdV ΔV|)
110 + (dRONdT ΔT|) + (dRONdV ΔV|)
1, 2, 3
RTT(IN)
0.33 × VDD2
90 - (dRONdT ΔT|) - (dRONdVΔV|)
110 + (dRONdT ΔT|) + (dRONdV ΔV|)
1, 2, 4
Notes:
1. ΔT = T - T(@calibration), ΔV = V - V(@calibration)
2. dRONdT, dRONdV, dVOHdT, dVOHdV, dRTTdV, and dRTTdT are not subject to production test
but are verified by design and characterization.
3. This parameter applies to input/output pin such as DQS, DQ, and DMI.
4. This parameter applies to input pin such as CK, CA, and CS.
5. Refer to Pull-up/Pull-down Driver Characteristics for VOHPU.
207
200b: x32 LPDDR4 SDRAM
DQ On-Die Termination
Table 138: Output Driver and Termination Register Temperature and Voltage Sensitivity
Symbol
Min
Max
Unit
dRONdT
Parameter
RON temperature sensitivity
0
0.75
%/˚C
dRONdV
RON voltage sensitivity
0
0.20
%/mV
dVOHdT
VOH temperature sensitivity
0
0.75
%/˚C
dVOHdV
VOH voltage sensitivity
0
0.35
%/mV
dRTTdT
RTT temperature sensitivity
0
0.75
%/˚C
dRTTdV
RTT voltage sensitivity
0
0.20
%/mV
ODT Mode Register
The ODT mode is enabled if MR11 OP[2:0] are non-zero. In this case, the value of RTT is
determined by the settings of those bits. The ODT mode is disabled if MR11 OP[2:0] = 0.
Asynchronous ODT
When ODT mode is enabled in MR11 OP[2:0], DRAM ODT is always High-Z. The DRAM
ODT feature is automatically turned ON asynchronously after a WRITE-1, MASK
WRITE-1, or MPC[WRITE-FIFO] command. After the burst write is complete, the DRAM
ODT turns OFF asynchronously. The DQ bus ODT control is automatic and will turn the
ODT resistance on/off if DQ ODT is enabled in the mode register.
The following timing parameters apply when the DQ bus ODT is enabled:
• ODTLon, tODTon (MIN), tODTon (MAX)
• ODTLoff, tODToff (MIN), tODToff (MAX)
ODTLON is a synchronous parameter and is the latency from a CAS-2 command to the
tODTon reference. ODTL
ON latency is a fixed latency value for each speed bin. Each
speed bin has a different ODTLON latency.
Minimum RTT turn-on time ( tODTon (MIN)) is the point in time when the device termination circuit leaves High-Z and ODT resistance begins to turn on.
Maximum RTT turn on time ( tODTon (MAX)) is the point in time when the ODT resistance is fully on.
tODTon
(MIN) and tODTon (MAX) are measured once ODTLon latency is satisfied from
CAS-2 command.
ODTLOFF is a synchronous parameter and it is the latency from CAS-2 command to
tODToff reference. ODTL
OFF latency is a fixed latency value for each speed bin. Each
speed bin has a different ODTLOFF latency.
Minimum RTT turn-off time ( tODToff (MIN)) is the point in time when the device termination circuit starts to turn off the ODT resistance.
Maximum ODT turn off time ( tODToff (MAX)) is the point in time when the on-die termination has reached High-Z.
tODToff
(MIN) and tODToff (MAX) are measured once ODTLoff latency is satisfied from
CAS-2 command.
208
200b: x32 LPDDR4 SDRAM
DQ On-Die Termination
Table 139: ODTLON and ODTLOFF Latency Values
ODTLON Latency1
WL Set A (nCK)
WL Set B (nCK)
WL Set A (nCK)
WL Set B (nCK)
Lower
Frequency Limit
(>) (MHz)
N/A
N/A
N/A
N/A
10
266
N/A
N/A
N/A
N/A
266
533
N/A
6
N/A
22
533
800
4
12
20
28
800
1066
4
14
22
32
1066
1333
6
18
24
36
1333
1600
6
20
26
40
1600
1866
8
24
28
44
1866
2133
tWPRE
= 2 tCK
ODTLOFF Latency2
Upper
Frequency Limit
(≤
≤) (MHz)
1. ODTLON is referenced from CAS-2 command.
2. ODTLOFF as shown in table assumes BL = 16. For BL32, 8 tCK should be added.
Notes:
Figure 139: Asynchronous ODTon/ODToff Timing
T0
T1
T2
T3
BL
BA0,
CA, AP
CAn
CAn
T4
Ta0
Ta1
Ta2
Ta3
Ta4
Ta5
Ta6
Ta7
Ta8
Ta9
Ta10
Ta11
Ta12 Ta13
Ta14
Ta15
Ta16
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
DES
Ta17 Ta18
Ta19 Ta20
Ta21
DES
DES
CK_c
CK_t
CS
CA
Command
WRITE-1
CAS-2
t
WL
DES
DES
DES
DES
DQSS(MIN)
t
t
WPRE
WPST
DQS_c
DQS_t
t
DQ
DMI
DIN
n0
t
DQS2DQ
DIN DIN
n1 n2
DIN
n3
DIN
n4
DIN
n5
DIN
n6
DIN
n7
DIN
n8
DIN
n9
DIN DIN DIN DIN DIN DIN
n10 n11 n12 n13 n14 n15
DQSS(MAX)
t
t
WPRE
WPST
DQS_c
DQS_t
t
DQ
DMI
DIN
n0
t
ODTLon
t
DRAM RTT
DQS2DQ
DIN DIN
n1 n2
DIN
n3
DIN
n4
DIN
n5
DIN
n6
DIN
n8
DIN
n9
DIN DIN DIN DIN DIN DIN
n10 n11 n12 n13 n14 n15
t
ODTon(MAX)
t
ODTon(MIN)
ODT High-Z
DIN
n7
ODTL2Q
Transition
ODToff(MAX)
ODToff(MIN)
Transition
ODT High-Z
ODTLoff
Don’t Care
Notes:
1. BL = 16, Write postamble = 0.5nCK, DQ/DQS: VSSQ termination
2. Din n = data-in to column n.
209
200b: x32 LPDDR4 SDRAM
DQ On-Die Termination
3. DES commands are shown for ease of illustration; other commands may be valid at
these times.
DQ ODT During Power-Down and Self Refresh Modes
DQ bus ODT will be disabled in power-down mode. In self refresh mode, the ODT will
be turned off when CKE is LOW but will be enabled if CKE is HIGH and DQ ODT is enabled in the mode register.
ODT During Write Leveling Mode
If ODT is enabled in MR11 OP[2:0] in write leveling mode, the device always provides
the termination on DQS signals. DQ termination is always off in write leveling mode.
Table 140: Termination State in Write Leveling Mode
DQS Termination
DQ[15:0]/DMI[1:0]
Termination
Disabled
Off
Off
Enabled
On
Off
ODT State in MR11 OP[2:0]
210
200b: x32 LPDDR4 SDRAM
Target Row Refresh Mode
Target Row Refresh Mode
The device limits the number of times that a given row can be accessed within a refresh
period (tREFW × 2) prior to requiring adjacent rows to be refreshed. The maximum activate count (MAC) is the maximum number of activates that a single row can sustain
within a refresh period before the adjacent rows need to be refreshed. The row receiving
the excessive actives is the target row (TRn), the adjacent rows to be refreshed are the
victim rows. When the MAC limit is reached on TRn, either the device receives all (R × 2)
REFRESH commands before another row activate is issued, or the device should be
placed into targeted row refresh (TRR) mode. The TRR mode will refresh the rows adjacent to the TRn that encountered tMAC limit.
If the device supports unlimited MAC value: MR24 OP[2:0] = 000 and MR24 OP[3] = 1,
TARGET ROW REFRESH operation is not required. Even though the device allows to set
MR24 OP[7] = 1: TRR mode enable, in this case the device behavior is vendor specific.
For example, a certain device may ignore MRW command for entering/exiting TRR
mode or a certain device may support commands related TRR mode. See vendor device
datasheets for details about TRR mode definition at supporting unlimited MAC value
case.
There could be a maximum of two target rows to a victim row in a bank. The cumulative
value of the activates from the two target rows on a victim row in a bank should not exceed MAC value.
MR24 fields are required to support the new TRR settings. Setting MR24 OP[7] = 1 enables TRR mode and setting MR24 OP[7] = 0 disables TRR mode. MR24 OP[6:4] defines
which bank (BAn) the target row is located in (refer to MR24 table for details).
The TRR mode must be disabled during initialization as well as any other device calibration modes. The TRR mode is entered from a DRAM idle state, once TRR mode has
been entered, no other mode register commands are allowed until TRR mode is completed; however, setting MR24 OP[7] = 0 to interrupt and reissue the TRR mode is allowed.
When enabled, TRR mode is self-clearing. the mode will be disabled automatically after
the completion of defined TRR flow (after the third BAn precharge has completed plus
tMRD). Optionally, the TRR mode can also be exited via another MRS command at the
completion of TRR by setting MR24 OP[7] = 0. If the TRR is exited via another MRS command, the value written to MR24 OP[6:4] are "Don’t Care."
TRR Mode Operation
1. The timing diagram depicts TRR mode. The following steps must be performed
when TRR mode is enabled. This mode requires all three ACT (ACT1, ACT2, and
ACT3) and three corresponding PRE commands (PRE1, PRE2, and PRE3) to complete TRR mode. PRECHARGE All (PREA) commands issued while the device is in
TRR mode will also perform precharge to BAn and counts towards PREn command.
2. Prior to issuing the MRW command to enter TRR mode, the device should be in
the idle state. MRW command must be issued with MR24 OP[7] = 1 and MR24
OP[6:4] defining the bank in which the targeted row is located. All other MR24 bits
should remain unchanged.
3. No activity is to occur with the device until tMRD has been satisfied. When tMRD
has been satisfied, the only commands allowed BAn, until TRR mode has completed, are ACT and PRE.
211
200b: x32 LPDDR4 SDRAM
Target Row Refresh Mode
4. The first ACT to the BAn with the TRn address can now be applied; no other command is allowed at this point. All other banks must remain inactive from when the
first BAn ACT command is issued until [(1.5 x tRAS) + tRP] is satisfied.
5. After the first ACT to the BAn with the TRn address is issued, PRE to BAn is to be
issued (1.5 × tRAS) later; and then followed tRP later by the second ACT to the BAn
with the TRn address.
6. After the second ACT to the BAn with the TRn address is issued, PRE to BAn is to
be issued tRAS later and then followed tRP later by the third ACT to the BAn with
the TRn address.
7. After the third ACT to the BAn with the TRn address is issued, PRE to BAn would
be issued tRAS later. TRR mode is completed once tRP plus tMRD is satisfied.
8. TRR mode must be completed as specified to guarantee that adjacent rows are refreshed. Anytime the TRR mode is interrupted and not completed, the interrupted
TRR mode must be cleared and then subsequently performed again. To clear an
interrupted TRR mode, MR24 change is required with setting MR24 OP[7] = 0,
MR24 OP[6:4] are "Don’t care," followed by three PRE to BAn, with tRP time in between each PRE command. The complete TRR sequence (steps 2–7) must be then
reissued and completed to guarantee that the adjacent rows are refreshed.
9. A REFRESH command to the device, or entering self refresh mode, is not allowed
while the device is in TRR mode.
Figure 140: Target Row Refresh Mode
T0
T1
T2
T3
Ta0 Ta1
Ta2
Ta3
Tb0
Tb1
Tc0
Tc1
Tc2 Tc3
Td0
Td1 Td2 Td3
Te0
Te1
Tf0 Tf1
Tf2
Tf3
Tg0 Tg1 Tg2
Tg3
Th0 Th1 Th2
Th3
Tk0
Tk1 Tk2
Tm0 Tm1 Tm2 Tm3 Tm4 Tm5
CK_c
CK_t
CKE
CS
ACT1
Command
MRW-1
MRW-2
DES
PRE1
ACT-2
ACT-1
1st ACT
TRR entry
N/A
N/A
N/A
N/A
N/A
BAn
N/A
N/A
Address
OP
MA
OP
OP
TRn
TRn
TRn
TRn
tMRD
BAn
BAn
in idle
ACT-1
PRE2
ACT-2
DES
CMD-1
CMD-2
2nd ACT
BAn
N/A
BAn
N/A
N/A
V
Non
BAn
N/A
N/A
TRn
TRn
TRn
TRn
V
V
tRP
DES
PRE
ACT3
DES
CMD-1
CMD-2
DES
V
N/A
BAn
V
Non
BAn
V
V
V
N/A
N/A
V
V
V
CMD-1
CMD-2
N/A
BAn
N/A
N/A
V
Non
BAn
V
TRn
TRn
TRn
TRn
V
V
9
BAn TRR operation allowed
DES
PRE
V
V
N/A
BAn
V
V
N/A
N/A
tRAS
Activity allowed
No activity allowed in other banks (Banks closed)
DES
DES
CMD-2
CMD-1
DES
3rd PRE
V
t RP
No activity allowed (Banks closed)
DES
3rd ACT
V
tRAS
PRE3
ACT-2
ACT-1
2nd PRE
N/A
1.5 × tRAS
Non BAn
in idle
ACT2
DES
PRE
1st PRE
Bank
Address
Non BAn
DES
tRP
V
Any
BAn
V
V
V
V
V
V
+ tMRD
No activity allowed
(may have bank(s) open)
Activity
allowed
Activity
allowed
Don’t Care
Notes:
1. TRn is the targeted row.
2. Bank BAn represents the bank in which the targeted row is located.
3. TRR mode self-clears after tMRD + tRP measured from the third BAn precharge PRE3 at
clock edge Th4.
4. TRR mode or any other activity can be re-engaged after tRP + tMRD from the third BAn
precharge PRE3. PRE_ALL also counts if it is issued instead of PREn. TRR mode is cleared
by the device after PRE3 to the BAn bank.
5. ACTIVATE commands to BAn during TRR mode do not provide refresh support (the refresh counter is unaffected).
6. The device must restore the degraded row(s) caused by excessive activation of the targeted row (TRn) necessary to meet refresh requirements.
7. A new TRR mode must wait tMRD + tRP time after the third precharge.
8. BAn may not be used with any other command.
9. ACT and PRE are the only allowed commands to BAn during TRR mode.
10. REFRESH commands are not allowed during TRR mode.
212
200b: x32 LPDDR4 SDRAM
Post-Package Repair
11. All timings are to be met by DRAM during TRR mode, such as tFAW. Issuing ACT1, ACT2,
and ACT3 counts towards tFAW budget.
Post-Package Repair
The device has fail row address repair as an optional post-package repair (PPR) feature
and it is readable through MR25 OP[7:0].
PPR provides simple and easy repair method in the system and fail row address can be
repaired by the electrical programming of Electrical-fuse scheme. The device can correct one row per bank with PPR.
Electrical-fuse cannot be switched back to un-fused states once it is programmed. The
controller should prevent unintended PPR mode entry and repair.
Failed Row Address Repair
1.
2.
3.
4.
5.
6.
7.
8.
Before entering PPR mode, all banks must be precharged.
Enable PPR using MR4 OP[4] = 1 and wait tMRD.
Issue ACT command with fail row address.
Wait tPGM to allow the device repair target row address internally then issue PRECHARGE
Wait tPGM_EXIT after PRECHARGE, which allows the device to recognize repaired
row address RAn.
Exit PPR mode with setting MR4 OP[4] = 0.
The device is ready for any valid command after tPGMPST.
In more than one fail address repair case, repeat step 2 to 7.
Once PPR mode is exited, to confirm whether the target row has correctly repaired, the
host can verify the repair by writing data into the target row and reading it back after
PPR exit with MR4 OP[4] = 0 and tPGMPST.
The following timing diagram shows PPR operation.
213
200b: x32 LPDDR4 SDRAM
Post-Package Repair
Figure 141: Post-Package Repair Timing
T0 T1
T2
T3 T4
T5 Ta0 Ta1 Ta2 Ta3 Ta4 Ta5 Ta6 Tb0 Tb1 Tb2 Tb3 Tb4 Tc0 Tc1 Tc2 Tc3 Tc4 Tc5 Td0 Td1 Td2 Td3 Td4 Td5
CK_c
CK_t
CKE
CS
Command
DES MR WRITE-1 MR WRITE-2 DES
BA
Address
PPR
staus
ACT-1
ACT-2
DES
PRE
DES
DES
Any
command
MR WRITE-1 MR WRITE-2 DES
Any
command
N/A N/A N/A N/A
Valid BA Valid Valid
Valid Valid
N/A N/A N/A N/A
Valid Valid Valid Valid
MA OP
RAn RAn RAn RAn
Valid Valid
OP
MA OP
Valid Valid Valid Valid
OP
OP
Normal mode
(All banks must be idle)
Move to PPR mode
PPR repair
t MRD
OP
Move to PPR mode
PPR recognition
t PGM
t PGM_Exit
Normal
mode
t PGMPST
Don’t Care
Notes:
1.
2.
3.
4.
During tPGM, any other commands (including refresh) are not allowed on each die.
With one PPR command, only one row can be repaired at one time per die.
When PPR procedure completes, reset procedure is required before normal operation.
During PPR, memory contents are not refreshed and may be lost.
Table 141: Post-Package Repair Timing Parameters
Parameter
PPR programming time
PPR exit time
New address setting time
Symbol
Min
Max
Units
tPGM
1000
–
ms
tPGM_EXIT
15
–
ns
tPGMPST
50
–
μs
214
200b: x32 LPDDR4 SDRAM
Read Preamble Training
Read Preamble Training
Read preamble training is supported through the MPC function.
This mode can be used to train or read level the DQS receivers. After read preamble
training is enabled by MR13 OP[1] = 1, the device will drive DQS_t LOW and DQS_c
HIGH within tSDO and remain at these levels until an MPC[READ DQ CALIBRATION]
command is issued.
During read preamble training, the DQS preamble provided during normal operation
will not be driven by the device. After the MPC[READ DQ CALIBRATION] command is
issued, the device will drive DQS_t/DQS_c and DQ like a normal READ burst after RL
and tDQSCK. Prior to the MPC[READ DQ CALIBRATION] command, the device may or
may not drive DQ[15:0] in this mode.
While in read preamble training mode, only READ DQ CALIBRATION commands may
be issued.
• Issue an MPC[READ DQ CALIBRATION] command followed immediately by a CAS-2
command.
• Each time an MPC[READ DQ CALIBRATION] command followed by a CAS-2 is received by the device, a 16-bit data burst will, after the currently set RL, drive the eight
bits programmed in MR32 followed by the eight bits programmed in MR40 on all I/O
pins.
• The data pattern will be inverted for I/O pins with a 1 programmed in the corresponding invert mask mode register bit.
• Note that the pattern is driven on the DMI pins, but no DATA BUS INVERSION function is enabled, even if read DBI is enabled in the DRAM mode register.
• This command can be issued every tCCD seamlessly.
• The operands received with the CAS-2 command must be driven LOW.
Read preamble training is exited within tSDO after setting MR13 OP[1] = 0.
The device supports the READ preamble training as optional feature. Refer to vendorspecific data sheets.
Figure 142: Read Preamble Training
T0
T1
T2
T3
Ta2
Ta3
Ta4
Ta6
T4
Ta0
Ta1
Ta5
DES
DES
MPC
MPC
DES [RD DQ CAL] [RD DQ CAL] CAS-2 CAS-2 DES
Tb0
Tb1
Tc0
Tc1
DES
DES
Tc2
Tc3
Tc4
Td0
Td1
Td2
Td3
Td4
Td5
Te0
Te1
CK_c
CK_t
CS
Command MRW-1 MRW-1 MRW-2 MRW-2
tSDO
DES
DES
DES
DES
DES
DES MRW-1 MRW-1 MRW-2 MRW-2
tDQSCK
RL
DES
DES
DES
tSDO
Read preamble training mode =
Enable: MR13[OP1] = 1
Read preamble training mode =
Enable: MR13[OP1] = 0
DQS_c
DQS_t
t
DQSQ
DQ
DMI
DOUT DOUT DOUT DOUT DOUT DOUT
n0
n1
n12 n13 n14 n15
DQ (High-Z or Driven )
DQ (High-Z or Driven )
Don’t Care
Note:
1. Read DQ calibration supports only BL16 operation.
215
200b: x32 LPDDR4 SDRAM
Electrical Specifications
Electrical Specifications
Absolute Maximum Ratings
Stresses greater than those listed in the table below may cause permanent damage to
the device. This is a stress rating only, and functional operation of the device at these
conditions, or any other conditions outside those indicated in the operational sections
of this document, is not implied. Exposure to absolute maximum rating conditions for
extended periods may adversely affect reliability.
Table 142: Absolute Maximum DC Ratings
Parameter
Symbol
Min
Max
Unit
Notes
VDD1
–0.4
2.1
V
1
VDD2 supply voltage relative to VSS
VDD2
–0.4
1.5
V
1
VDDQ supply voltage relative to VSS
VDDQ
–0.4
1.5
V
1
Voltage on any ball relative to VSS
VIN, VOUT
–0.4
1.5
V
TSTG
–55
125
˚C
VDD1 supply voltage relative to VSS
Storage temperature
Notes:
2
1. For information about relationships between power supplies, see the Voltage Ramp and
Device Initialization section.
2. Storage temperature is the case surface temperature on the center/top side of the device. For measurement conditions, refer to the JESD51-2 standard.
AC and DC Operating Conditions
Operation or timing that is not specified is illegal. To ensure proper operation, the device must be initialized properly.
Table 143: Recommended DC Operating Conditions
Symbol
Min
Typ
Max
DRAM
Unit
Notes
VDD1
1.7
1.8
1.95
Core 1 power
V
1, 2
VDD2
1.06
1.1
1.17
Core 2 power/Input buffer
power
V
1, 2, 3
VDDQ
1.06
1.1
1.17
I/O buffer power
V
2, 3
Notes:
1. VDD1 uses significantly less power than VDD2.
2. The voltage range is for DC voltage only. DC voltage is the voltage supplied at the
DRAM and is inclusive of all noise up to 20 MHz at the DRAM package ball.
3. The voltage noise tolerance from DC to 20 MHz exceeding a peak-to-peak tolerance of
45mV at the DRAM ball is not included in the TdIVW.
216
200b: x32 LPDDR4 SDRAM
AC and DC Operating Conditions
Table 144: Input Leakage Current
Parameter/Condition
Symbol
Min
Max
Unit
Notes
IL
–4
4
μA
1, 2
Input leakage current
Notes:
1. For CK_t, CK_c, CKE, CS, CA, ODT_CA, and RESET_n. Any input 0V ≤ VIN ≤ VDD2. (All other
pins not under test = 0V.
2. CA ODT is disabled for CK_t, CK_c, CS, and CA.
Table 145: Input/Output Leakage Current
Parameter/Condition
Symbol
Min
Max
Unit
Notes
IOZ
–5
5
μA
1, 2
Input/Output leakage current
Notes:
1. For DQ, DQS_t, DQS_c, and DMI. Any I/O 0V ≤ VOUT ≤ VDDQ.
2. I/Os status are disabled: High Impedance and ODT off.
Table 146: Operating Temperature Range
Parameter/Condition
Standard
Symbol
Min
Max
Unit
TOPER
Note 4
85
˚C
85
Note 4
˚C
Elevated
Notes:
1. Operating temperature is the case surface temperature at the center of the top side of
the device. For measurement conditions, refer to the JESD51-2 standard.
2. When using the device in the elevated temperature range, some derating may be required. See Mode Registers for vendor-specific derating.
3. Either the device case temperature rating or the temperature sensor can be used to set
an appropriate refresh rate, determine the need for AC timing derating, and/or monitor
the operating temperature (see Temperature Sensor). When using the temperature sensor, the actual device case temperature may be higher than the TOPER rating that applies
for the standard or elevated temperature range. For example, TCASE could be above
+85˚C when the temperature sensor indicates a temperature of less than +85˚C.
4. Refer to Operating temperature range on top page.
217
200b: x32 LPDDR4 SDRAM
AC and DC Input Measurement Levels
AC and DC Input Measurement Levels
Input Levels for CKE
Table 147: Input Levels
Parameter
Symbol
Min
Max
Unit
Notes
Input HIGH level (AC)
VIH(AC)
0.75 × VDD2
VDD2 + 0.2
V
1
Input LOW level (AC)
VIL(AC)
–0.2
0.25 × VDD2
V
1
Input HIGH level (DC)
VIH(DC)
0.65 × VDD2
VDD2 + 0.2
V
Input LOW level (DC)
VIL(DC)
–0.2
0.35 × VDD2
V
Note:
1. See the AC Overshoot and Undershoot section.
Figure 143: Input Timing Definition for CKE
VIH(AC)
VIH(DC)
VIL(AC)
VIL(DC)
Input
level
Don’t Care
Input Levels for RESET_n and ODT_CA
Table 148: Input Levels
Parameter
Symbol
Min
Max
Unit
Notes
Input HIGH level
VIH
0.80 × VDD2
VDD2 + 0.2
V
1
Input LOW level
VIL
–0.2
0.20 × VDD2
V
1
Note:
1. See the AC Overshoot and Undershoot section.
Figure 144: Input Timing Definition for RESET_n and ODT_CA
VIH
VIH
VIL
VIL
Input
level
Don’t Care
Differential Input Voltage for CK
The minimum input voltage needs to satisfy both V indiff_CK and V indiff_CK/2 specification
at input receiver and their measurement period is 1tCK. V indiff_CK is the peak-to-peak
218
200b: x32 LPDDR4 SDRAM
AC and DC Input Measurement Levels
voltage centered on 0 volts differential and V indiff_CK/2 is maximum and minimum peak
voltage from 0 volts.
0.0
Vindiff_CK
Vindiff_CK/2
Peak
voltage
Vindiff_CK/2
Differential Input Voltage : CK_t - CK_c
Figure 145: CK Differential Input Voltage
Peak
voltage
Time
Table 149: CK Differential Input Voltage
1600/1867
2133/2400/3200
3733/4267
Parameter
Symbol
Min
Max
Min
Max
Min
Max
Unit
Note
CK differential input voltage
Vindiff_CK
420
–
380
–
360
–
mV
1
Note:
1. The peak voltage of differential CK signals is calculated in a following equation.
•
•
•
•
Vindiff_CK = (Maximum peak voltage) - (Minimum peak voltage)
Maximum peak voltage = MAX(f(t))
Minimum peak voltage = MIN(f(t))
f(t) = VCK_t - VCK_c
Peak Voltage Calculation Method
The peak voltage of differential clock signals are calculated in a following equation.
• VIH.DIFF.peak voltage = MAX(f(t))
• VIL.DIFF.peak voltage = MIN(f(t))
• f(t) = V CK_t - V CK_c
219
200b: x32 LPDDR4 SDRAM
AC and DC Input Measurement Levels
Figure 146: Definition of Differential Clock Peak Voltage
Single Ended Input Voltage
CK_t
Min(f(t))
Max(f(t))
VREF(CA)
CK_c
Time
1. VREF(CA) is device internal setting value by VREF training.
Note:
Single-Ended Input Voltage for Clock
The minimum input voltage need to satisfy V inse_CK, V inse_CK_HIGH, and V inse_CK_LOW
specification at input receiver.
Figure 147: Clock Single-Ended Input Voltage
Vinse_CK_LOW
Vinse_CK
Vinse_CK_HIGH
Vinse_CK_HIGH
Vinse_CK_LOW
VREF(CA)
Vinse_CK
Single Ended Input Voltage
CK_t
CK_c
Time
Note:
1. VREF(CA) is device internal setting value by VREF training.
220
200b: x32 LPDDR4 SDRAM
AC and DC Input Measurement Levels
Table 150: Clock Single-Ended Input Voltage
1600/1867
2133/2400/3200
3733/4267
Parameter
Symbol
Min
Max
Min
Max
Min
Max
Unit
Clock single-ended
input voltage
Vinse_CK
210
–
190
–
180
–
mV
Clock single-ended input
voltage HIGH from VREF(CA)
Vinse_CK_HIGH
105
–
95
–
90
–
mV
Clock single-ended input
voltage LOW from VREF(CA)
Vinse_CK_LOW
105
–
95
–
90
–
mV
Differential Input Slew Rate Definition for Clock
Input slew rate for differential signals (CK_t, CK_c) are defined and measured as shown
below in figure and the tables.
Figure 148: Differential Input Slew Rate Definition for CK_t, CK_c
Differential Input Voltage : f(t) = CK_t - CK_c
Peak
Voltage
VIHdiff_CK
0.0
VILdiff_CK
Peak
Voltage
Delta TFdiff
Delta TRdiff
Time
Notes:
1. Differential signal rising edge from VILdiff_CK to VIHdiff_CK must be monotonic slope.
2. Differential signal falling edge from VIHdiff_CK to VILdiff_CK must be monotonic slope.
Table 151: Differential Input Slew Rate Definition for CK_t, CK_c
Description
From
To
Differential input slew rate for
rising edge (CK_t - CK_c)
VILdiff_CK
VIHdiff_CK
|VILdiff_CK - VIHdiff_CKΔTRdiff
Differential input slew rate for
falling edge (CK_t - CK_c)
VIHdiff_CK
VILdiff_CK
|VILdiff_CK - VIHdiff_CKΔTFdiff
221
Defined by
200b: x32 LPDDR4 SDRAM
AC and DC Input Measurement Levels
Table 152: Differential Input Level for CK_t, CK_c
1600/1867
2133/2400/3200
3733/4267
Parameter
Symbol
Min
Max
Min
Max
Min
Max
Unit
Differential Input HIGH
VIHdiff_CK
175
–
155
–
145
–
mV
Differential Input LOW
VILdiff_CK
–
–175
–
–155
–
–145
mV
Table 153: Differential Input Slew Rate for CK_t, CK_c
1600/1867
Parameter
Differential input
slew rate for clock
2133/2400/3200
3733/4267
Symbol
Min
Max
Min
Max
Min
Max
Unit
SRIdiff_CK
2
14
2
14
2
14
V/ns
Differential Input Cross-Point Voltage
The cross-point voltage of differential input signals (CK_t, CK_c) must meet the requirements in table below. The differential input cross-point voltage V IX is measured from
the actual cross-point of true and complement signals to the mid level that is V REF(CA).
Figure 149: Vix Definition (Clock)
VDD
Single-Ended Input Voltage
CK_t
Max(f(t))
ViX_CK_RF
ViX_CK_FR
VREF(CA)
ViX_CK_RF
Min(f(t))
ViX_CK_FR
CK_c
VSS
Time
Note:
1. The base levels of Vix_CK_FR and Vix_CK_RF are VREF(CA) that is device internal setting value
by VREF training.
222
200b: x32 LPDDR4 SDRAM
AC and DC Input Measurement Levels
Table 154: Cross-Point Voltage for Differential Input Signals (Clock)
Notes 1 and 2 apply to entire table
1600/1867
Parameter
Clock single-ended
cross-point voltage ratio
2133/2400/3200
3733/4267
Symbol
Min
Max
Min
Max
Min
Max
Unit
Vix_CK_ratio
–
25
–
25
–
25
%
1. Vix_CK_ratio is defined by this equation: Vix_CK_ratio = Vix_CK_FR/|MIN(f(t))|
2. Vix_CK_ratio is defined by this equation: Vix_CK_ratio = Vix_CK_RF/MAX(f(t))
Notes:
Differential Input Voltage for DQS
The minimum input voltage needs to satisfy both V indiff_DQS and V indiff_DQS/2 specification at input receiver and their measurement period is 1UI (tCK/2). V indiff_DQS is the
peak to peak voltage centered on 0 volts differential and V indiff_DQS/2 is maximum and
minimum peak voltage from 0 volts.
0.0
Vindiff_DQS
Vindiff_DQS /2
Peak
voltage
Vindiff_DQS /2
Differential Input Voltage : DQS_t - DQS_c
Figure 150: DQS Differential Input Voltage
Peak
voltage
Time
Table 155: DQS Differential Input Voltage
1600/1867
2133/2400/3200
3733/4267
Parameter
Symbol
Min
Max
Min
Max
Min
Max
Unit
Note
DQS differential input voltage
Vindiff_DQS
360
–
360
–
340
–
mV
1
Note:
1. The peak voltage of differential DQS signals is calculated in a following equation.
•
•
•
•
Vindiff_DQS = (Maximum peak voltage) - (Minimum peak voltage)
Maximum peak voltage = MAX(f(t))
Minimum peak voltage = MIN(f(t))
f(t) = VDQS_t - VDQS_c
Peak Voltage Calculation Method
The peak voltage of differential DQS signals are calculated in a following equation.
223
200b: x32 LPDDR4 SDRAM
AC and DC Input Measurement Levels
• VIH.DIFF.peak voltage = MAX(f(t))
• VIL.DIFF.peak voltage = MIN(f(t))
• f(t) = V DQS_t - V DQS_c
Figure 151: Definition of Differential DQS Peak Voltage
Single Ended Input Voltage
DQS_t
Min(f(t))
Max(f(t))
VREF(DQ)
DQS_c
Time
1. VREF(DQ) is device internal setting value by VREF training.
Note:
Single-Ended Input Voltage for DQS
The minimum input voltage need to satisfy V inse_DQS, V inse_DQS_HIGH, and V inse_DQS_LOW
specification at input receiver.
Figure 152: DQS Single-Ended Input Voltage
Vinse_DQS_LOW
Vinse_DQS
Vinse_DQS_HIGH
Vinse_DQS_HIGH
Vinse_DQS_LOW
VREF(DQ)
Vinse_DQS
Single Ended Input Voltage
DQS_t
DQS_c
Time
Note:
1. VREF(DQ) is device internal setting value by VREF training.
224
200b: x32 LPDDR4 SDRAM
AC and DC Input Measurement Levels
Table 156: DQS Single-Ended Input Voltage
1600/1867
2133/2400/3200
3733/4267
Parameter
Symbol
Min
Max
Min
Max
Min
Max
Unit
DQS single-ended
input voltage
Vinse_DQS
180
–
180
–
170
–
mV
DQS single-ended input
voltage HIGH from VREF(DQ)
Vinse_DQS_HIGH
90
–
90
–
85
–
mV
DQS single-ended input
voltage LOW from VREF(DQ)
Vinse_DQS_LOW
90
–
90
–
85
–
mV
Differential Input Slew Rate Definition for DQS
Input slew rate for differential signals (DQS_t, DQS_c) are defined and measured as
shown below in figure and the tables.
Figure 153: Differential Input Slew Rate Definition for DQS_t, DQS_c
Differential Input Voltage : f(t) = CK_t - CK_c
Peak
Voltage
VIHdiff_CK
0.0
VILdiff_CK
Peak
Voltage
Delta TFdiff
Delta TRdiff
Time
Notes:
1. Differential signal rising edge from VILdiff_DQS to VIHdiff_DQS must be monotonic slope.
2. Differential signal falling edge from VIHdiff_DQS to VILdiff_DQS must be monotonic slope.
Table 157: Differential Input Slew Rate Definition for DQS_t, DQS_c
Description
From
To
Differential input slew rate for
rising edge (DQS_t - DQS_c)
VILdiff_DQS
VIHdiff_DQS
|VILdiff_DQS - VIHdiff_DQSΔTRdiff
Differential input slew rate for
falling edge (DQS_t - DQS_c)
VIHdiff_DQS
VILdiff_DQS
|VILdiff_DQS - VIHdiff_DQSΔTFdiff
225
Defined by
200b: x32 LPDDR4 SDRAM
AC and DC Input Measurement Levels
Table 158: Differential Input Level for DQS_t, DQS_c
1600/1867
Parameter
2133/2400/3200
3733/4267
Symbol
Min
Max
Min
Max
Min
Max
Unit
Differential Input HIGH
VIHdiff_DQS
140
–
140
–
120
–
mV
Differential Input LOW
VILdiff_DQS
–
–140
–
–140
–
–120
mV
Table 159: Differential Input Slew Rate for DQS_t, DQS_c
1600/1867
Parameter
2133/2400/3200
3733/4267
Symbol
Min
Max
Min
Max
Min
Max
Unit
SRIdiff
2
14
2
14
2
14
V/ns
Differential input slew rate
Differential Input Cross-Point Voltage
The cross-point voltage of differential input signals (DQS_t, DQS_c) must meet the requirements in table below. The differential input cross-point voltage V IX is measured
from the actual cross-point of true and complement signals to the mid level that is
VREF(DQ).
Figure 154: Vix Definition (DQS)
VDDQ
Single-Ended Input Voltage
DQS_t
Max(f(t))
ViX_DQS_RF
ViX_DQS_FR
VREF(DQ)
ViX_DQS_RF
Min(f(t))
ViX_DQS_FR
DQS_c
VSSQ
Time
Note:
1. The base levels of Vix_DQS_FR and Vix_DQS_RF are VREF(DQ) that is device internal setting value by VREF training.
226
200b: x32 LPDDR4 SDRAM
Output Slew Rate and Overshoot/Undershoot specifications
Table 160: Cross-Point Voltage for Differential Input Signals (DQS)
Notes 1 and 2 apply to entire table
1600/1867
Parameter
Clock single-ended
cross-point voltage ratio
Notes:
2133/2400/3200
3733/4267
Symbol
Min
Max
Min
Max
Min
Max
Unit
Vix_DQS_ratio
–
20
–
20
–
20
%
1. Vix_DQS_ratio is defined by this equation: Vix_DQS_ratio = Vix_DQS_FR/|MIN(f(t))|
2. Vix_DQS_ratio is defined by this equation: Vix_DQS_ratio = Vix_DQS_RF/MAX(f(t))
Input Levels for ODT
Table 161: Input Levels
Parameter
Symbol
Min
ODT input HIGH level (AC)
VIHODT(AC)
ODT input LOW level (AC)
VILODT(AC)
ODT input HIGH level (DC)
ODT input LOW level (DC)
Note:
Max
Unit
Notes
0.75 × VDD2
VDD2 + 0.2
V
1
–0.2
0.25 × VDD2
V
1
VIHODT(DC)
0.65 × VDD2
VDD2 + 0.2
V
VILODT(DC)
–0.2
0.35 × VDD2
V
1. See the Overshoot and Undershoot section.
Output Slew Rate and Overshoot/Undershoot specifications
Single-Ended Output Slew Rate
Table 162: Single-Ended Output Slew Rate
Note 1-5 applies to entire table
Value
Symbol
Min
Max
Units
Single-ended output slew rate (VOH = VDDQ/3)
Parameter
SRQse
3.5
9.0
V/ns
Output slew rate matching ratio (rise to fall)
–
0.8
1.2
–
Notes:
1. SR = Slew rate; Q = Query output; se = Single-ended signal
2. Measured with output reference load.
3. The ratio of pull-up to pull-down slew rate is specified for the same temperature and
voltage, over the entire temperature and voltage range. For a given output, it represents the maximum difference between pull-up and pull-down drivers due to process
variation.
4. The output slew rate for falling and rising edges is defined and measured between
VOL(AC) = 0.2 × VOH(DC) and VOH(AC) = 0.8 × VOH(DC).
5. Slew rates are measured under average SSO conditions with 50% of the DQ signals per
data byte switching.
227
200b: x32 LPDDR4 SDRAM
Output Slew Rate and Overshoot/Undershoot specifications
Figure 155: Single-Ended Output Slew Rate Definition
Single-Ended Output Voltage (DQ)
¨TRSE
VOH(AC)
VCENT
VOL(AC)
¨TFSE
Time
Differential Output Slew Rate
Table 163: Differential Output Slew Rate
Note 1-4 applies to entire table
Value
Parameter
Symbol
Min
Max
Units
Differential output slew rate (VOH = VDDQ/3)
SRQdiff
7
18
V/ns
Notes:
1. SR = Slew rate; Q = Query output; se = Differential signal
2. Measured with output reference load.
3. The output slew rate for falling and rising edges is defined and measured between
VOL(AC) = 0.2 × VOH(DC) and VOH(AC) = 0.8 × VOH(DC).
4. Slew rates are measured under average SSO conditions with 50% of the DQ signals per
data byte switching.
Figure 156: Differential Output Slew Rate Definition
Differential Output Voltage (DQ)
¨TRdiff
0
¨TFdiff
Time
228
200b: x32 LPDDR4 SDRAM
Driver Output Timing Reference Load
Overshoot and Undershoot Specifications
Table 164: AC Overshoot/Undershoot Specifications
Parameter
1600
1866
3200
3733
4267
Unit
Maximum peak amplitude provided for overshoot area
MAX
0.3
0.3
0.3
0.3
0.3
V
Maximum peak amplitude provided for undershoot area
MAX
0.3
0.3
0.3
0.3
0.3
V
Maximum area above VDD/ VDDQ
MAX
0.1
0.1
0.1
0.1
0.1
V-ns
Maximum area below VSS/ VSSQ
MAX
0.1
0.1
0.1
0.1
0.1
V-ns
Notes:
1. VDD stands for VDD2 for CA[5:0], CK_t, CS_n, CKE, and ODT. VDD stands for VDDQ for DQ,
DMI, DQS_t, and DQS_c.
2. VSS stands for VSS for CA[5:0], CK_t, CK_c, CS_n, CKE, and ODT. VSS stands for VSSQ for
DQ, DMI, DQS_t, and DQS_c.
3. Maximum peak amplitude values are referenced from actual VDD and VSS values.
4. Maximum area values are referenced from maximum VDD and VSS values.
Table 165: Overshoot/Undershoot Specification for CKE and RESET
Parameter
Specification
Maximum peak amplitude provided for overshoot area
0.35V
Maximum peak amplitude provided for undershoot area
0.35V
Maximum area above VDD
0.8 V-ns
Maximum area below VSS
0.8 V-ns
Figure 157: Overshoot and Undershoot Definition
Maximum amplitude
Volts (V)
Overshoot area
VDD
Time (ns)
VSS
Undershoot area
Maximum amplitude
Driver Output Timing Reference Load
Timing reference loads are not intended as a precise representation of any particular
system environment or depiction of an actual load presented by a production tester.
System designers should use IBIS or other simulation tools to correlate the timing reference load to a system environment. Manufacturers correlate to their production test
conditions, generally one or more coaxial transmission lines terminated at the tester
electronics.
229
200b: x32 LPDDR4 SDRAM
LVSTL I/O System
Figure 158: Driver Output Timing Reference Load
DRAM
50 Ohms
1. All output timing parameter values are reported with respect to this reference load; this
reference load is also used to report slew rate.
Note:
LVSTL I/O System
LVSTL I/O cells are comprised of a driver pull-up and pull-down and a terminator.
Figure 159: LVSTL I/O Cell
VDDQ
Pull-Up
DQ
ODT
Enabled when receiving
Pull-Down
VSSQ
VSSQ
To ensure that the target impedance is achieved, calibrate the LVSTL I/O cell as following example:
1. Calibrate the pull-down device against a 240 ohm resistor to V DDQ via the ZQ pin.
• Set strength control to minimum setting
• Increase drive strength until comparator detects data bit is less than V DDQ/3
• NMOS pull-down device is calibrated to 120 ohms
2. Calibrate the pull-up device against the calibrated pull-down device.
• Set V OH target and NMOS controller ODT replica via MRS (VOH can be automatically
controlled by ODT MRS)
230
200b: x32 LPDDR4 SDRAM
Input/Output Capacitance
• Set strength control to minimum setting
• Increase drive strength until comparator detects data bit is greater than V OH target
• NMOS pull-up device is calibrated to V OH target
Figure 160: Pull-Up Calibration
VDDQ
Strength contol [N-1:0]
N
Comparator
VOH target
Controller ODT replica could be
60 Ohm, 120 Ohm, ... via MRS setting
Calibrated NMOS PD
control + ODT information
VSSQ
Input/Output Capacitance
Table 166: Input/Output Capacitance
Notes 1 and 2 apply to entire table
Parameter
Symbol
Min
Max
Unit
Input capacitance, CK_t and CK_c
CCK
0.5
0.9
Input capacitance delta, CK_t and CK_c
CDCK
0
0.09
3
Input capacitance, all other input-only pins
CI
0.5
0.9
4
Input capacitance delta, all other input-only pins
CDI
–0.1
0.1
pF
Notes
5
CIO
0.7
1.3
CDDQS
0
0.1
7
Input/output capacitance delta, DQ, DMI
CDIO
–0.1
0.1
8
Input/output capacitance, ZQ pin
CZQ
0
5.0
Input/output capacitance, DQ, DMI, DQS_t, DQS_c
Input/output capacitance delta, DQS_t, DQS_c
Notes:
6
1. This parameter applies to LPDDR4 die only (does not include package capacitance).
2. This parameter is not subject to production testing; it is verified by design and characterization. The capacitance is measured according to JEP147 (procedure for measuring input capacitance using a vector network analyzer), with VDD1, VDD2, VDDQ, and VSS applied; all other pins are left floating.
3. Absolute value of CCK_t – CCK_c.
4. CI applies to CS, CKE, and CA[5:0].
5. CDI = CI – 0.5 × (CCK_t + CCK_c); it does not apply to CKE.
6. DMI loading matches DQ and DQS.
7. Absolute value of CDQS_t and CDQS_c.
231
200b: x32 LPDDR4 SDRAM
IDD Specification Parameters and Test Conditions
8. CDIO = CIO – 0.5 × (CDQS_t + CDQS_c) in byte-lane.
IDD Specification Parameters and Test Conditions
Table 167: IDD Measurement Conditions
Switching for CA
CK_t edge
R1
R2
R3
R4
R5
R6
R7
R8
CKE
HIGH
HIGH
HIGH
HIGH
HIGH
HIGH
HIGH
HIGH
CS
LOW
LOW
LOW
LOW
LOW
LOW
LOW
LOW
CA0
HIGH
LOW
LOW
LOW
LOW
HIGH
HIGH
HIGH
CA1
HIGH
HIGH
HIGH
LOW
LOW
LOW
LOW
HIGH
CA2
HIGH
LOW
LOW
LOW
LOW
HIGH
HIGH
HIGH
CA3
HIGH
HIGH
HIGH
LOW
LOW
LOW
LOW
HIGH
CA4
HIGH
LOW
LOW
LOW
LOW
HIGH
HIGH
HIGH
CA5
HIGH
HIGH
HIGH
LOW
LOW
LOW
LOW
HIGH
Notes:
1. LOW = VIN ≤ VIL(DC) MAX
HIGH = VIN ≥ VIH(DC) MIN
STABLE = Inputs are stable at a HIGH or LOW level
2. CS must always be driven LOW.
3. 50% of CA bus is changing between HIGH and LOW once per clock for the CA bus.
4. The pattern is used continuously during IDD measurement for IDD values that require
switching on the CA bus.
Table 168: CA Pattern for IDD4R
Clock Cycle
Number
CKE
CS
Command
CA0
CA1
CA2
CA3
CA4
CA5
N
HIGH
HIGH
Read-1
L
H
L
L
L
L
N+1
HIGH
LOW
L
H
L
L
L
L
N+2
HIGH
HIGH
L
H
L
L
H
L
N+3
HIGH
LOW
L
L
L
L
L
L
N+4
HIGH
LOW
DES
L
L
L
L
L
L
N+5
HIGH
LOW
DES
L
L
L
L
L
L
N+6
HIGH
LOW
DES
L
L
L
L
L
L
N+7
HIGH
LOW
DES
L
L
L
L
L
L
N+8
HIGH
HIGH
Read-1
L
H
L
L
L
L
L
H
L
L
H
L
L
H
L
L
H
H
H
H
H
H
H
H
CAS-2
N+9
HIGH
LOW
N+10
HIGH
HIGH
N+11
HIGH
LOW
N+12
HIGH
LOW
DES
L
L
L
L
L
L
N+13
HIGH
LOW
DES
L
L
L
L
L
L
N+14
HIGH
LOW
DES
L
L
L
L
L
L
CAS-2
232
200b: x32 LPDDR4 SDRAM
IDD Specification Parameters and Test Conditions
Table 168: CA Pattern for IDD4R (Continued)
Clock Cycle
Number
CKE
CS
Command
CA0
CA1
CA2
CA3
CA4
CA5
N+15
HIGH
LOW
DES
L
L
L
L
L
L
Notes:
1. BA[2:0] = 010; CA[9:4] = 000000 OR 111111; Burst order CA[3:2] = 00 or 11 (same as
LPDDR3 IDDR3 specification).
2. CA pins are kept LOW with DES CMD to reduce ODT current (different from LPDDR3
IDDR3 specification).
Table 169: CA Pattern for IDD4W
Clock Cycle
Number
CKE
CS
Command
CA0
CA1
CA2
CA3
CA4
CA5
N
HIGH
HIGH
Write-1
L
L
H
L
L
L
N+1
HIGH
LOW
L
H
L
L
L
L
N+2
HIGH
HIGH
N+3
HIGH
LOW
N+4
HIGH
LOW
N+5
HIGH
N+6
N+7
CAS-2
L
H
L
L
H
L
L
L
L
L
L
L
DES
L
L
L
L
L
L
LOW
DES
L
L
L
L
L
L
HIGH
LOW
DES
L
L
L
L
L
L
HIGH
LOW
DES
L
L
L
L
L
L
N+8
HIGH
HIGH
Write-1
L
L
H
L
L
L
N+9
HIGH
LOW
L
H
L
L
H
L
N+10
HIGH
HIGH
N+11
HIGH
LOW
N+12
HIGH
LOW
N+13
HIGH
N+14
N+15
L
H
L
L
H
H
L
L
H
H
H
H
DES
L
L
L
L
L
L
LOW
DES
L
L
L
L
L
L
HIGH
LOW
DES
L
L
L
L
L
L
HIGH
LOW
DES
L
L
L
L
L
L
Notes:
CAS-2
1. BA[2:0] = 010; CA[9:4] = 000000 or 111111 (same as LPDDR3 specification).
2. No burst ordering (different from LPDDR3 specification).
3. CA pins are kept LOW with DES CMD to reduce ODT current (different from LPDDR3
specification).
Table 170: Data Pattern for IDD4W (DBI Off)
DBI Off Case
DQ[7]
DQ[6]
DQ[5]
DQ[4]
DQ[3]
DQ[2]
DQ[1]
DQ[0]
DBI
# of 1s
BL0
1
1
1
1
1
1
1
1
0
8
BL1
1
1
1
1
0
0
0
0
0
4
BL2
0
0
0
0
0
0
0
0
0
0
BL3
0
0
0
0
1
1
1
1
0
4
233
200b: x32 LPDDR4 SDRAM
IDD Specification Parameters and Test Conditions
Table 170: Data Pattern for IDD4W (DBI Off) (Continued)
DBI Off Case
DQ[7]
DQ[6]
DQ[5]
DQ[4]
DQ[3]
DQ[2]
DQ[1]
DQ[0]
DBI
# of 1s
BL4
0
0
0
0
0
0
1
1
0
2
BL5
0
0
0
0
1
1
1
1
0
4
BL6
1
1
1
1
1
1
0
0
0
6
BL7
1
1
1
1
0
0
0
0
0
4
BL8
1
1
1
1
1
1
1
1
0
8
BL9
1
1
1
1
0
0
0
0
0
4
BL10
0
0
0
0
0
0
0
0
0
0
BL11
0
0
0
0
1
1
1
1
0
4
BL12
0
0
0
0
0
0
1
1
0
2
BL13
0
0
0
0
1
1
1
1
0
4
BL14
1
1
1
1
1
1
0
0
0
6
BL15
1
1
1
1
0
0
0
0
0
4
BL16
1
1
1
1
1
1
0
0
0
6
BL17
1
1
1
1
0
0
0
0
0
4
BL18
0
0
0
0
0
0
1
1
0
2
BL19
0
0
0
0
1
1
1
1
0
4
BL20
0
0
0
0
0
0
0
0
0
0
BL21
0
0
0
0
1
1
1
1
0
4
BL22
1
1
1
1
1
1
1
1
0
8
BL23
1
1
1
1
0
0
0
0
0
4
BL24
0
0
0
0
0
0
1
1
0
2
BL25
0
0
0
0
1
1
1
1
0
4
BL26
1
1
1
1
1
1
0
0
0
6
BL27
1
1
1
1
0
0
0
0
0
4
BL28
1
1
1
1
1
1
1
1
0
8
BL29
1
1
1
1
0
0
0
0
0
4
BL30
0
0
0
0
0
0
0
0
0
0
BL31
0
0
0
0
1
1
1
1
0
4
# of 1s
16
16
16
16
16
16
16
16
Note:
1. Simplified pattern; same data pattern was applied to DQ[4], DQ[5], DQ[6], and DQ[7] to
reduce complexity for IDD4W pattern programming.
Table 171: Data Pattern for IDD4R (DBI Off)
DBI Off Case
BL0
DQ[7]
DQ[6]
DQ[5]
DQ[4]
DQ[3]
DQ[2]
DQ[1]
DQ[0]
DBI
# of 1s
1
1
1
1
1
1
1
1
0
8
234
200b: x32 LPDDR4 SDRAM
IDD Specification Parameters and Test Conditions
Table 171: Data Pattern for IDD4R (DBI Off) (Continued)
DBI Off Case
DQ[7]
DQ[6]
DQ[5]
DQ[4]
DQ[3]
DQ[2]
DQ[1]
DQ[0]
DBI
# of 1s
BL1
1
1
1
1
0
0
0
0
0
4
BL2
0
0
0
0
0
0
0
0
0
0
BL3
0
0
0
0
1
1
1
1
0
4
BL4
0
0
0
0
0
0
1
1
0
2
BL5
0
0
0
0
1
1
1
1
0
4
BL6
1
1
1
1
1
1
0
0
0
6
BL7
1
1
1
1
0
0
0
0
0
4
BL8
1
1
1
1
1
1
1
1
0
8
BL9
1
1
1
1
0
0
0
0
0
4
BL10
0
0
0
0
0
0
0
0
0
0
BL11
0
0
0
0
1
1
1
1
0
4
BL12
0
0
0
0
0
0
1
1
0
2
BL13
0
0
0
0
1
1
1
1
0
4
BL14
1
1
1
1
1
1
0
0
0
6
BL15
1
1
1
1
0
0
0
0
0
4
BL16
1
1
1
1
1
1
1
1
0
8
BL17
1
1
1
1
0
0
0
0
0
4
BL18
0
0
0
0
0
0
0
0
0
0
BL19
0
0
0
0
1
1
1
1
0
4
BL20
1
1
1
1
1
1
0
0
0
6
BL21
0
0
0
0
1
1
1
1
0
4
BL22
0
0
0
0
0
0
1
1
0
2
BL23
1
1
1
1
0
0
0
0
0
4
BL24
0
0
0
0
0
0
0
0
0
0
BL25
0
0
0
0
1
1
1
1
0
4
BL26
1
1
1
1
1
1
1
1
0
8
BL27
1
1
1
1
0
0
0
0
0
4
BL28
0
0
0
0
0
0
1
1
0
2
BL29
1
1
1
1
0
0
0
0
0
4
BL30
1
1
1
1
1
1
0
0
0
6
BL31
0
0
0
0
1
1
1
1
0
4
# of 1s
16
16
16
16
16
16
16
16
Note:
1. Simplified pattern; same data pattern was applied to DQ[4], DQ[5], DQ[6], and DQ[7] to
reduce complexity for IDD4W pattern programming.
235
200b: x32 LPDDR4 SDRAM
IDD Specification Parameters and Test Conditions
Table 172: Data Pattern for IDD4W (DBI On)
DBI On Case
DQ[7]
DQ[6]
DQ[5]
DQ[4]
DQ[3]
DQ[2]
DQ[1]
DQ[0]
DBI
# of 1s
BL0
0
0
0
0
0
0
0
0
1
1
BL1
1
1
1
1
0
0
0
0
0
4
BL2
0
0
0
0
0
0
0
0
0
0
BL3
0
0
0
0
1
1
1
1
0
4
BL4
0
0
0
0
0
0
1
1
0
2
BL5
0
0
0
0
1
1
1
1
0
4
BL6
0
0
0
0
0
0
1
1
1
3
BL7
1
1
1
1
0
0
0
0
0
4
BL8
0
0
0
0
0
0
0
0
1
1
BL9
1
1
1
1
0
0
0
0
0
4
BL10
0
0
0
0
0
0
0
0
0
0
BL11
0
0
0
0
1
1
1
1
0
4
BL12
0
0
0
0
0
0
1
1
0
2
BL13
0
0
0
0
1
1
1
1
0
4
BL14
0
0
0
0
0
0
1
1
1
3
BL15
1
1
1
1
0
0
0
0
0
4
BL16
0
0
0
0
0
0
1
1
1
3
BL17
1
1
1
1
0
0
0
0
0
4
BL18
0
0
0
0
0
0
1
1
0
2
BL19
0
0
0
0
1
1
1
1
0
4
BL20
0
0
0
0
0
0
0
0
0
0
BL21
0
0
0
0
1
1
1
1
0
4
BL22
0
0
0
0
0
0
0
0
1
1
BL23
1
1
1
1
0
0
0
0
0
4
BL24
0
0
0
0
0
0
1
1
0
2
BL25
0
0
0
0
1
1
1
1
0
4
BL26
0
0
0
0
0
0
1
1
1
3
BL27
1
1
1
1
0
0
0
0
0
4
BL28
0
0
0
0
0
0
0
0
1
1
BL29
1
1
1
1
0
0
0
0
0
4
BL30
0
0
0
0
0
0
0
0
0
0
BL31
0
0
0
0
1
1
1
1
0
4
# of 1s
8
8
8
8
8
8
16
16
8
236
200b: x32 LPDDR4 SDRAM
IDD Specification Parameters and Test Conditions
Table 173: Data Pattern for IDD4R (DBI On)
DBI On Case
DQ[7]
DQ[6]
DQ[5]
DQ[4]
DQ[3]
DQ[2]
DQ[1]
DQ[0]
DBI
# of 1s
BL0
0
0
0
0
0
0
0
0
1
1
BL1
1
1
1
1
0
0
0
0
0
4
BL2
0
0
0
0
0
0
0
0
0
0
BL3
0
0
0
0
1
1
1
1
0
4
BL4
0
0
0
0
0
0
1
1
0
2
BL5
0
0
0
0
1
1
1
1
0
4
BL6
0
0
0
0
0
0
1
1
1
3
BL7
1
1
1
1
0
0
0
0
0
4
BL8
0
0
0
0
0
0
0
0
1
1
BL9
1
1
1
1
0
0
0
0
0
4
BL10
0
0
0
0
0
0
0
0
0
0
BL11
0
0
0
0
1
1
1
1
0
4
BL12
0
0
0
0
0
0
1
1
0
2
BL13
0
0
0
0
1
1
1
1
0
4
BL14
0
0
0
0
0
0
1
1
1
3
BL15
1
1
1
1
0
0
0
0
0
4
BL16
0
0
0
0
0
0
0
0
1
1
BL17
1
1
1
1
0
0
0
0
0
4
BL18
0
0
0
0
0
0
0
0
0
0
BL19
0
0
0
0
1
1
1
1
0
4
BL20
0
0
0
0
0
0
1
1
1
3
BL21
0
0
0
0
1
1
1
1
0
4
BL22
0
0
0
0
0
0
1
1
0
2
BL23
1
1
1
1
0
0
0
0
0
4
BL24
0
0
0
0
0
0
0
0
0
0
BL25
0
0
0
0
1
1
1
1
0
4
BL26
0
0
0
0
0
0
0
0
1
1
BL27
1
1
1
1
0
0
0
0
0
4
BL28
0
0
0
0
0
0
1
1
0
2
BL29
1
1
1
1
0
0
0
0
0
4
BL30
0
0
0
0
0
0
1
1
1
3
BL31
0
0
0
0
1
1
1
1
0
4
# of 1s
8
8
8
8
8
8
16
16
8
237
200b: x32 LPDDR4 SDRAM
IDD Specification Parameters and Test Conditions
IDD Specifications
IDD values are for the entire operating voltage range, and all of them are for the entire
standard temperature range.
Table 174: IDD Specification Parameters and Operating Conditions
VDD2, VDDQ = 1.06–1.17V; VDD1 = 1.70–1.95V
Parameter/Condition
tCK
tCK
=
Operating one bank active-precharge current:
(MIN); tRC = tRC (MIN); CKE is HIGH; CS is LOW between valid commands; CA bus inputs are switching; Data bus inputs are stable;
ODT is disabled
Idle power-down standby current: tCK = tCK (MIN); CKE is
LOW; CS is LOW; All banks are idle; CA bus inputs are switching;
Data bus inputs are stable; ODT is disabled
Idle power-down standby current with clock stop: CK_t =
LOW, CK_c = HIGH; CKE is LOW; CS is LOW; All banks are idle; CA
bus inputs are stable; Data bus inputs are stable; ODT is disabled
tCK
tCK
=
(MIN); CKE is
Idle non-power-down standby current:
HIGH; CS is LOW; All banks are idle; CA bus inputs are switching;
Data bus inputs are stable; ODT is disabled
Idle non-power-down standby current with clock stopped:
CK_t = LOW; CK_c = HIGH; CKE is HIGH; CS is LOW; All banks are
idle; CA bus inputs are stable; Data bus inputs are stable; ODT is
disabled
Active power-down standby current: tCK = tCK (MIN); CKE is
LOW; CS is LOW; One bank is active; CA bus inputs are switching;
Data bus inputs are stable; ODT is disabled
Active power-down standby current with clock stop: CK_t =
LOW, CK_c = HIGH; CKE is LOW; CS is LOW; One bank is active; CA
bus inputs are stable; Data bus inputs are stable; ODT is disabled
Active non-power-down standby current: tCK = tCK (MIN);
CKE is HIGH; CS is LOW; One bank is active; CA bus inputs are
switching; Data bus inputs are stable; ODT is disabled
Active non-power-down standby current with clock stopped: CK_t = LOW, CK_c = HIGH; CKE is HIGH; CS is LOW; One bank
is active; CA bus inputs are stable; Data bus inputs are stable; ODT
is disabled
Operating burst READ current: tCK = tCK (MIN); CS is LOW between valid commands; One bank is active; BL = 16 or 32; RL = RL
(MIN); CA bus inputs are switching; 50% data change each burst
transfer; ODT is disabled
238
Symbol
Power
Supply
IDD01
VDD1
IDD02
VDD2
IDD0Q
VDDQ
IDD2P1
VDD1
IDD2P2
VDD2
IDD2PQ
VDDQ
IDD2PS1
VDD1
IDD2PS2
VDD2
IDD2PSQ
VDDQ
IDD2N1
VDD1
IDD2N2
VDD2
IDD2NQ
VDDQ
IDD2NS1
VDD1
IDD2NS2
VDD2
IDD2NSQ
VDDQ
IDD3P1
VDD1
IDD3P2
VDD2
IDD3PQ
VDDQ
IDD3PS1
VDD1
IDD3PS2
VDD2
IDD3PSQ
VDDQ
IDD3N1
VDD1
IDD3N2
VDD2
IDD3NQ
VDDQ
IDD3NS1
VDD1
IDD3NS2
VDD2
IDD3NSQ
VDDQ
IDD4R1
VDD1
IDD4R2
VDD2
IDD4RQ
VDDQ
Notes
2
2
2
2
2
2
3
3
3
4
200b: x32 LPDDR4 SDRAM
IDD Specification Parameters and Test Conditions
Table 174: IDD Specification Parameters and Operating Conditions (Continued)
VDD2, VDDQ = 1.06–1.17V; VDD1 = 1.70–1.95V
Parameter/Condition
tCK
tCK
=
(MIN); CS is LOW beOperating burst WRITE current:
tween valid commands; One bank is active; BL = 16 or 32; WL =
WL (MIN); CA bus inputs are switching; 50% data change each
burst transfer; ODT is disabled
All-bank REFRESH burst current: tCK = tCK (MIN); CKE is HIGH
between valid commands; tRC = tRFCab (MIN); Burst refresh; CA
bus inputs are switching; Data bus inputs are stable; ODT is disabled
All-bank REFRESH average current: tCK = tCK (MIN); CKE is
HIGH between valid commands; tRC = tREFI; CA bus inputs are
switching; Data bus inputs are stable; ODT is disabled
tCK
tCK
=
(MIN); CKE is
Per-bank REFRESH average current:
HIGH between valid commands; tRC = tREFI/8; CA bus inputs are
switching; Data bus inputs are stable; ODT is disabled
Power-down self refresh current: CK_t = LOW, CK_c = HIGH;
CKE is LOW; CA bus inputs are stable; Data bus inputs are stable;
Maximum 1x self refresh rate; ODT is disabled
Notes:
1.
2.
3.
4.
5.
Symbol
Power
Supply
IDD4W1
VDD1
IDD4W2
VDD2
IDD4WQ
VDDQ
IDD51
VDD1
IDD52
VDD2
IDD5Q
VDDQ
IDD5AB1
VDD1
IDD5AB2
VDD2
IDD5ABQ
VDDQ
IDD5PB1
VDD1
Notes
3
3
3
IDD5PB2
VDD2
IDD5PBQ
VDDQ
3
IDD61
VDD1
5, 6
IDD62
VDD2
5, 6
IDD6Q
VDDQ
3, 5, 6
ODT disabled: MR11[2:0] = 000b.
IDD current specifications are tested after the device is properly initialized.
Measured currents are the summation of VDDQ and VDD2.
Guaranteed by design with output load = 5pF and RON = 40 ohm.
The 1x self refresh rate is the rate at which the device is refreshed internally during self
refresh before going into the elevated temperature range.
6. This is the general definition that applies to full-array self refresh.
7. For all IDD measurements, VIHCKE = 0.8 × VDD2; VILCKE = 0.2 × VDD2.
239
200b: x32 LPDDR4 SDRAM
AC Timing
AC Timing
Table 175: Clock Timing
Parameter
Symbol
Average clock period
tCK(avg)
Average HIGH pulse width
tCH(avg)
Average LOW pulse width
tCL(avg)
Absolute clock period
tCK(abs)
Absolute clock HIGH pulse width
tCH(abs)
Absolute clock LOW pulse width
tCL(abs)
Clock period jitter
Maximum clock jitter between two
consecutive clock cycles (includes
clock period jitter)
Data Rate
Min/
Max
1600
Min
1250
625
535
468
ps
Max
100
100
100
100
ns
3200
3733
Min
0.46
Max
0.54
Min
0.46
Max
0.54
tCK(avg)min
Min
4267
Unit
tCK(avg)
tCK(avg)
+ tJIT(per)min
Min
0.43
Max
0.57
Min
0.43
Max
0.57
ps
tCK(avg)
tCK(avg)
tJIT(per)al-
Min
–70
–40
–34
–30
lowed
Max
70
40
34
30
tJIT(cc)allowed
Max
140
80
68
60
ps
ps
Table 176: Read Output Timing
Parameter
Symbol
DQS output access time
from CK_t/CK_c
tDQSCK
DQS output access time
from CK_t/CK_c - voltage
variation
DQS output access time
from CK_t/CK_c - temperature variation
CK to DQS rank to rank
variation
DQS_t, DQS_c to DQ skew
total, per group, per access (DBI Disabled)
DQ output hold time total from DQS_t, DQS_c
(DBI Disabled)
Min/
Max
Data Rate
533
1066 1600 2133 2667 3200 3733 4267
Unit
Notes
ps
1
Min
1500
Max
3500
Max
7
ps/mV
2
Max
4
ps°/C
3
Max
1.0
ns
4, 5
tDQSQ
Max
0.18
UI
6
tQH
Min
MIN(tQSH, tQSL)
ps
6
tDQSCK_
VOLT
tDQSCK_
TEMP
tDQSCK_r
ank2rank
240
200b: x32 LPDDR4 SDRAM
AC Timing
Table 176: Read Output Timing (Continued)
Parameter
Symbol
Data output valid window time total, per pin
(DBI-Disabled)
DQS_t, DQS_c to DQ skew
total, per group, per access (DBI-Enabled)
DQ output hold time total from DQS_t, DQS_c
(DBI-Enabled)
Data output valid window time total, per pin
(DBI-Enabled)
tQW_to-
tal
tDQSQ_D
BI
tQH_DBI
tQW_to-
tal_DBI
Min/
Max
Data Rate
533
Min
1066 1600 2133 2667 3200 3733 4267
0.75
0.73
0.68
Unit
Notes
UI
6, 11
Max
0.18
UI
6
Min
MIN(tQSH_DBI, tQSL_DBI)
ps
6
UI
6, 11
Min
0.75
0.73
0.68
DQS_t, DQS_c differential
output LOW time (DBIDisabled)
tQSL
Min
tCL(abs)
- 0.05
tCK(avg)
9, 11
DQS_t, DQS_c differential
output HIGH time (DBIDisabled)
tQSH
Min
tCH(abs)
- 0.05
tCK(avg)
10, 11
DQS_t, DQS_c differential
output LOW time (DBIEnabled)
tQSL-DBI
Min
tCL(abs)
- 0.045
tCK(avg)
9, 11
DQS_t, DQS_c differential
output HIGH time (DBIEnabled)
tQSH-DBI
Min
tCH(abs)
- 0.045
tCK(avg)
10, 11
Read preamble
tRPRE
Min
1.8
tCK(avg)
Read postamble
tRPST
Min
0.4 (or 1.4 if extra postamble is programmed in MR)
tCK(avg)
tLZ(DQS)
Min
(RL x tCK) + tDQSCK(Min) - (tRPRE(Max) x tCK) - 200ps
ps
tLZ(DQ)
Min
(RL x tCK) + tDQSCK(Min) - 200ps
ps
DQS Low-Z from clock
DQ Low-Z from clock
DQS High-Z from clock
DQ High-Z from clock
tHZ(DQS)
Min
tHZ(DQ)
Min
Notes:
(RL x
tCK)
+
tDQSCK(Max)+(BL/2
tCK)
x
- 100ps
tCK)
+
(tRPST(Max)
x
(RL x tCK) + tDQSCK(Max) + tDQSQ(Max) + (BL/2 x tCK) 100ps
ps
ps
1. This parameter includes DRAM process, voltage, and temperature variation. It also includes the AC noise impact for frequencies >20 MHz and a MAX voltage of 45mV peakto-peak from DC-20 MHz at a fixed temperature on the package. The voltage supply
noise must comply with the component MIN/MAX DC operating conditions.
2. tDQSCK_volt MAX delay variation as a function of DC voltage variation for VDDQ and
VDD2. The voltage supply noise must comply with the component MIN/MAX DC operating conditions. The voltage variation is defined as the MAX[abs(tDQSCKmin@V1 tDQSCKmax@V2), abs(tDQSCKmax@V1 - tDQSCKmin@V2)]/abs(V1 - V2). For tester measurement VDDQ = VDD2 is assumed.
t
3. DQSCK_temp MAX delay variation as a function of temperature.
4. The same voltage and temperature are applied to tDQSCK_rank2rank.
241
200b: x32 LPDDR4 SDRAM
AC Timing
5. tDQSCK_rank2rank parameter is applied to multi-ranks per byte lane within a package
consisting of the same design dies.
6. DQ-to-DQS differential jitter where the total includes the sum of deterministic and random timing terms for a specified BER.
7. The deterministic component of the total timing.
8. This parameter will be characterized and guaranteed by design.
9. tQSL describes the instantaneous differential output low pulse width on DQS_t - DQS_c,
as measured from one falling edge to the next consecutive rising edge.
10. tQSH describes the instantaneous differential output high pulse width on DQS_t DQS_c, as measured from one falling edge to the next consecutive rising edge.
11. This parameter is a function of input clock jitter. These values assume MIN tCH(abs) and
tCL(abs). When the input clock jitter MIN tCH(abs) and tCL(abs) is 0.44 or greater than
tCK(avg), the MIN value of tQSL will be tCL(abs) - 0.04 and tQSH will be tCH(abs) - 0.04.
Table 177: Write Voltage and Timing
Note UI = tCK(AVG)(MIN)/2
Min/
Max
Data Rate
Parameter
Symbol
Rx timing window total
at VdIVW voltage levels
TdIVW_t
otal
Max
DQ and DMI input pulse
width (at VCENT_DQ)
TdIPW
Min
0.45
Min
200
Max
800
Max
tDQS2DQ
DQ-to-DQS offset
tDQDQ
DQ-to-DQ offset
DQ-to-DQS offset temperature variation
tDQS2DQ
DQ-to-DQS offset voltage
variation
tDQS2DQ
_temp
_volt
533
1066 1600 2133 2667 3200 3733 4267
Unit
Notes
UI
1, 2, 3
UI
7
ps
6
30
ps
7
Max
0.6
ps/°C
8
Max
33
ps/50mV
9
Min
0.75
Max
1.25
0.22
0.25
WRITE command to first
DQS transition
tDQSS
DQS input HIGH-level
width
tDQSH
–
0.4
tCK(avg)
DQS input LOW-level
width
tDQSL
Min
0.4
tCK(avg)
DQS falling edge to CK
setup time
tDSS
Min
0.2
tCK(avg)
DQS falling edge from CK
hold time
tDSH
Min
0.2
tCK(avg)
Write postamble
tWPST
Min
0.4 (or 1.4 if extra postamble is programmed in MR)
tCK(avg)
Write preamble
tWPRE
Min
1.8
tCK(avg)
Notes:
tCK(avg)
1. Data Rx mask voltage and timing parameters are applied per pin and include the DRAM
DQ-to-DQS voltage AC noise impact for frequencies >20 MHz with a maximum voltage
of 45mV peak-to-peak at a fixed temperature on the package. The voltage supply noise
must comply to the component MIN/MAX DC operating conditions.
242
200b: x32 LPDDR4 SDRAM
AC Timing
2. Rx differential DQ-to-DQS jitter total timing window at the VdIVW voltage levels.
3. Defined over the DQ internal VREF range. The Rx mask at the pin must be within the internal VREF DQ range irrespective of the input signal common mode.
4. Rx mask defined for one pin toggling with other DQ signals in a steady state.
5. DQ-only minimum input pulse width defined at the VCENT_DQ(pin_mid).
6. DQ-to-DQS offset is within byte from DRAM pin to DRAM internal latch. Includes all
DRAM process, voltage, and temperature variations.
7. DQ-to-DQ offset defined within byte from DRAM pin to DRAM internal latch for a given
component.
8. tDQS2DQ (MAX) delay variation as a function of temperature.
9. tDQS2DQ (MAX) delay variation as a function of the DC voltage variation for VDDQ and
VDD2. It includes the VDDQ and VDD2 AC noise impact for frequencies >20 MHz and MAX
voltage of 45mV peak-to-peak from DC-20 MHz at a fixed temperature on the package.
For tester measurement, VDDQ = VDD2 is assumed.
Table 178: CKE Input Timing
Symbol
Min/
Max
tCKE
Min
Delay from valid command to CKE input LOW
tCMDCKE
Valid clock requirement after CKE input LOW
tCKELCK
Valid CS requirement before CKE input LOW
tCSCKE
Valid CS requirement after CKE input LOW
tCKELCS
Valid Clock requirement before CKE Input
HIGH
Parameter
Data Rate
Unit
Notes
MAX(7.5ns, 4nCK)
ns
1
Min
MAX(1.75ns, 3nCK)
ns
1
Min
MAX(5ns, 5nCK)
ns
1
Min
1.75
ns
Min
MAX(5ns, 5nCK)
ns
1
tCKCKEH
Min
MAX(1.75ns, 3nCK)
ns
1
Exit power-down to next valid command delay
tXP
Min
MAX(7.5ns, 5nCK)
ns
1
Valid CS requirement before CKE input HIGH
tCSCKEH
Min
1.75
ns
Valid CS requirement after CKE input HIGH
tCKEHCS
Min
MAX(7.5ns, 5nCK)
ns
1
Valid clock and CS requirement after CKE input LOW after MRW command
tMRWCKEL
Min
MAX(14ns, 10nCK)
ns
1
Valid clock and CS requirement after CKE input LOW after ZQ calibration start command
tZQCKE
Min
MAX(1.75ns, 3nCK)
ns
1
CKE minimum pulse width (HIGH and LOW
pulse width)
Note:
1600
3200
3733
4267
1. Delay time has to satisfy both analog time(ns) and clock count(nCK). For example,
tCMDCKE will not expire until CK has toggled through at least 3 full cycles (3tCK) and
3.75ns has transpired. The case which 3nCK is applied to is shown below.
243
200b: x32 LPDDR4 SDRAM
AC Timing
Figure 161: tCMDCKE Timing
T-1
T0
T1
T2
T3
T4
CK_c
CK_t
tCMDCKE
CKE
CS
CA Valid Valid
Command
Valid
DES
Don’t Care
Table 179: Command Address Input Timing
Symbol
Min/
Max
Command/address valid
window (referenced from
CA VIL/VIH to CK VIX)
tcIVW
Min
Address and control input
pulse width (referenced
to VREF)
tcIPW
Min
Parameter
Notes:
Data Rate
533
1066 1600 2133 2667 3200 3733 4267
0.3
0.55
0.55
0.55
0.6
0.6
0.6
Unit
Notes
0.35
0.4
tCK(avg)
1, 2, 3
0.7
0.8
tCK(avg)
4
1. CA Rx mask timing parameters at the pin including voltage and temperature drift.
2. Rx differential CA to CK jitter total timing window at the VcIVW voltage levels.
3. Defined over the CA internal VREF range. The Rx mask at the pin must be within the internal VREF(CA) range irrespective of the input signal common mode.
4. CA only minimum input pulse width defined at the VCENT_CA(pin mid).
Table 180: Boot Timing Parameters (10–55 MHz)
Parameter
Clock cycle time
DQS output data acess time
from CK
DQS edge to output data
edge
Symbol
tCKb
tDQSCKb
tDQSQb
Min/
Max
Value
Min
18
Max
100
Min
1.0
Max
10.0
Max
1.2
244
Unit
ns
ns
ns
200b: x32 LPDDR4 SDRAM
AC Timing
Table 181: Mode Register Timing Parameters
Data Rate
Symbol
Min/
Max
MODE REGISTER WRITE (MRW) command period
tMRW
Min
MAX(10ns, 10nCK)
ns
MODE REGISTER READ (MRR) command period
tMRR
Min
8
tCK(avg)
Additional time after tXP has expired
until the MRR command may be issued
tMRRI
Min
tRCD(min)
+ 3nCK
ns
Delay from MRW command to DQS
driven out
tSDO
Max
MAX(12nCK, 20ns)
ns
Parameter
1600
3200
3733
4267
Unit
Table 182: Core Timing Parameters
Refresh rate is determined by the value in MR4 OP[2:0]
Data Rate
Symbol
Min/
Max
READ latency (DBI disabled)
RL-A
Min
6
10
14
20
24
28
32
36
tCK(avg)
READ latency (DBI enabled)
RL-B
Min
6
12
16
22
28
32
36
40
tCK(avg)
WRITE latency (Set A)
WL-A
Min
4
6
8
10
12
14
16
18
tCK(avg)
WRITE latency (Set B)
WL-B
Min
4
8
12
18
22
26
30
34
tCK(avg)
Parameter
533
1066 1600 2133 2667 3200 3733 4267
Unit
Min
tRAS + tRPab
(with all-bank precharge)
tRAS + tRPpb
(with per-bank precharge)
ns
tSR
Min
MAX(15ns, 3nCK)
ns
Self refresh exit to next
valid command delay
tXSR
Min
MAX(tRFCab + 7.5ns, 2nCK)
ns
CAS-to-CAS delay
tCCD
Min
8
tCK(avg)
CAS-to-CAS delay masked
write
tCCDMW
MIN
32
tCK(avg)
Internal READ-to-PRECHARGE command delay
tRTP
Min
MAX(7.5ns, 8nCK)
ns
RAS-to-CAS delay
tRCD
Min
MAX(18ns, 4nCK)
ns
Row precharge time (single bank)
tRPpb
Min
MAX(18ns, 3nCK)
ns
Row precharge time (all
banks)
tRPab
Min
MAX(21ns, 3nCK)
ns
ACTIVATE-to-ACTIVATE
command period (same
bank)
tRC
Minimum self refresh
time (entry to exit)
245
Notes
1
200b: x32 LPDDR4 SDRAM
AC Timing
Table 182: Core Timing Parameters (Continued)
Refresh rate is determined by the value in MR4 OP[2:0]
Parameter
Symbol
Min/
Max
Data Rate
533
1066 1600 2133 2667 3200 3733 4267
Min
MAX(42ns, 3nCK)
Max
tREFI
Unit
Notes
ns
1
Row active time
tRAS
Write recovery time
tWR
Min
MAX(18ns, 4nCK)
ns
Write-to-read delay
tWTR
Min
MAX(10ns, 8nCK)
ns
Active bank A to active
bank B
tRRD
Min
MAX(10ns, 4nCK)
ns
Precharge-to-precharge
delay
tPPD
Min
4
tCK(avg)
Four-bank activate window
tFAW
Min
40
ns
tESCKE
Min
MAX(1.75ns, 3nCK)
–
Delay from SRE command
to CKE input LOW
Notes:
MIN(9 ×
× Refresh
Rate1,
70.2)
μs
2
3
1. DRAM devices should be evenly addressed when being accessed. Disproportionate accesses to a particular row address may result in reduction of the product lifetime.
2. Precharge to precharge timing restriction does not apply to AUTO PRECHARGE commands.
3. Delay time has to satisfy both analog time(ns) and clock count (nCK). It means that
tESCKE will not expire until CK has toggled through at least three full cycles (3 tCK) and
1.75ns has transpired. The case which 3nCK is applied to is shown below.
Figure 162: tESCKE Timing
T-1
T0
T1
T2
T3
T4
CK_c
CK_t
t ESCKE
CKE
CS
CA Valid Valid
REFRESH
Command SELF Entry
DES
Don’t Care
246
200b: x32 LPDDR4 SDRAM
AC Timing
Table 183: CA Bus ODT Timing
Symbol
Min/
Max
Data Rate
Parameter
CA ODT value update time
tODTUP
Min
RU(20ns/tCK,avg)
533-4267
Table 184: CA Bus Training Parameters
Data Rate
Parameter
Symbol
Min/
Max
Valid clock requirement after CKE
Input LOW
tCKELCK
Min
MAX(5ns, 5nCK)
tCK
Data setup for VREF training mode
tDStrain
Min
2
ns
Data hold for VREF training mode
tDHtrain
Min
2
ns
tADR
Max
20
ns
CA BUS TRAINING command-tocommand delay
tCACD
Min
RU(tADR/tCK)
tCK
Valid strobe requirement before
CKE LOW
tDQSCKE
Min
10
ns
First CA BUS TRAINING command
following CKE LOW
tCAENT
Min
250
ns
Max
250
ns
Asynchronous data read
tVREF-
VREF step time – multiple steps
ca_LONG
1600
3200
3733
4267
Unit
VREF step time – one step
tVREFca_SHORT
Max
80
ns
Valid clock requirement before CS
HIGH
tCKPRECS
Min
2tCK + tXP
–
Valid clock requirement after CS
HIGH
tCKPSTCS
Min
MAX(7.5ns, 5nCK)
–
Minimum delay from CS to DQS
toggle in command bus training
tCS_VREF
Min
2
tCK
tCKEHDQS
Min
10
ns
tMRZ
Min
1.5
ns
tCKELODTon
Min
20
ns
Min
20
ns
Minimum delay from CKE HIGH to
strobe High-Z
CA bus training CKE HIGH to DQ
tri-state
ODT turn-on latency from CKE
ODT turn-off latency from CKE
Note:
tCKEHODT-
off
Notes
1
1. If tCACD is violated, the data for samples that violate tCACD will not be available, except
for the last sample (where tCACD after this sample is met). Valid data for the last sample
will be available after tADR.
247
200b: x32 LPDDR4 SDRAM
AC Timing
Table 185: Asynchronous ODT Turn On and Turn Off Timing
Symbol
800–2133 MHz
Unit
tODTon,min
1.5
ns
tODTon,max
3.5
ns
tODToff,min
1.5
ns
tODToff,max
3.5
ns
Table 186: Temperature Derating Parameters
Data Rate
Parameter
Symbol
Min/
Max
DQS output access time from CK_t/
CK_c (derated)
tDQSCKd
Max
tRCDd
Min
tRCD
tRCd
Min
tRC
tRASd
Min
tRAS
Row precharge time (derated)
tRPd
Min
tRP
Active bank A to active bank B (derated)
tRRD
Min
tRRD
RAS-to-CAS delay (derated)
ACTIVATE-to-ACTIVATE command period (same bank, derated)
Row active time (derated)
Note:
1600
3200
3733
3600
Unit
ps
+ 1.875
ns
+ 3.75
ns
+ 1.875
ns
+ 1.875
+ 1.875
1. Timing derating applies for the operation at 85°C to 105°C.
248
4267
ns
ns
200b: x32 LPDDR4 SDRAM
CA Rx Voltage and Timing
CA Rx Voltage and Timing
The command and address (CA), including CS input receiver compliance mask for voltage and timing, is shown in the CA Receiver (Rx) Mask figure below. All CA and CS signals apply the same compliance mask and operate in single data rate mode.
The CA input Rx mask for voltage and timing is applied across all pins, as shown in the
figure below. The Rx mask defines the area that the input signal must not encroach if
the DRAM input receiver is expected to successfully capture a valid input signal; it is not
the valid data eye.
Figure 163: CA Receiver (Rx) Mask
tcIVW_total
Rx Mask
VCENT_CA(pin mid)
VcIVW
Figure 164: Across Pin VREF (CA) Voltage Variation
CAx
CAy
CAz
VCENT_CAz
VCENT_CAx
VCENT_CAy
VREF variation
(component)
VCENT_CA(pin mid) is defined as the midpoint between the largest V CENT_CA voltage level
and the smallest V CENT_CA voltage level across all CA and CS pins for a given DRAM
component. Each CA V CENT level is defined by the center, which is, the widest opening
of the cumulative data input eye, as depicted in the figure above. This clarifies that any
DRAM component level variation must be accounted for within the CA Rx mask. The
component-level V REF will be set by the system to account for RON and ODT settings.
249
200b: x32 LPDDR4 SDRAM
CA Rx Voltage and Timing
Figure 165: CA Timings at the DRAM Pins
CK, CK Data-in at DRAM Pin
Minimum CA eye center aligned
CK_c
VcIVW
CK_t
Rx mask
DRAM pin
CA
tcIVW
TcIVW for all CA signals is defined as centered on
the CK_t/CK_c crossing at the DRAM pin.
Note:
1. All of the timing terms in above figure are measured from the CK_t/CK_c to the center(midpoint) of the TcIVW window taken at the VcIVW_total voltage levels centered
around VCENT_CA(pin mid).
Figure 166: CA tcIPW and SRIN_cIVW Definition (for Each Input Pulse)
tr
tf
Rx Mask
VcIVW
VCENT_CA(pin mid)
tcIPW
Note:
1. SRIN_cIVW = VdIVW_total/(tr or tf); signal must be monotonic within tr and tf range.
Figure 167: CA VIHL_AC Definition (for Each Input Pulse)
VIHL(AC)min/2
VCENT_CA
Rx Mask
Rx Mask
Rx Mask
VcIVW
VIHL(AC)min/2
250
200b: x32 LPDDR4 SDRAM
CA Rx Voltage and Timing
Table 187: DRAM CMD/ADR, CS
UI = tCK(AVG)MIN
DQ – 13337
DQ –
1600/1867
DQ –
3200/3733
DQ – 4267
Symbol
Parameter
Min
Max
Min
Max
Min
Max
Min
Max
Unit
Notes
VclVW
Rx mask voltage peak-topeak
–
175
–
175
–
155
–
145
mV
1, 2, 3
VIHL(AC)
CA AC input pulse amplitude peak-to-peak
210
–
210
–
190
–
180
–
mV
4, 6
1
7
1
7
1
7
1
7
V/ns
5
SRIN_clVW Input slew rate over VclVW
Notes:
1. CA Rx mask voltage and timing parameters at the pin, including voltage and temperature drift.
2. Rx mask voltage VcIVW total (MAX) must be centered around VCENT_CA(pin mid).
3. Defined over the CA internal VREF range. The Rx mask at the pin must be within the internal VREF(CA) range irrespective of the input signal common mode.
4. CA-only input pulse signal amplitude into the receiver must meet or exceed VIHL(AC) at
any point over the total UI. No timing requirement above level. VIHL(AC) is the peak-topeak voltage centered around VCENT_CA(pin mid), such that VIHL(AC)/2 (MIN) must be met
both above and below VCENT_CA.
5. Input slew rate over VcIVW mask is centered at VCENT_CA(pin mid).
6. VIHL(AC) does not have to be met when no transitions are occurring.
7. The Rx voltage and absolute timing requirements apply for DQ operating frequencies at
or below 1333 for all speed bins. For example the tcIVW(ps) = 450ps at or below 1333
operating frequencies.
251
200b: x32 LPDDR4 SDRAM
DQ Tx Voltage and Timing
DQ Tx Voltage and Timing
DRAM Data Timing
Figure 168: Read Data Timing Definitions – tQH and tDQSQ Across DQ Signals per DQS Group
tQSH(DQS_t)
DQS_c
DQS_t
tQH
tDQSQ
Associated
DQ pins
DQS_c
DQS_t
DQx
tQW
tQW
DQy
DQz
tQW
252
tQSL(DQS_t)
200b: x32 LPDDR4 SDRAM
DQ Rx Voltage and Timing
DQ Rx Voltage and Timing
The DQ input receiver mask for voltage and timing is applied per pin, as shown in the
DQ Receiver (Rx) Mask figure below. The total mask (V dIVW_total, TdIVW_total) defines
the area that the input signal must not encroach in order for the DQ input receiver to
successfully capture an input signal. The mask is a receiver property, and it is not the
valid data eye.
Figure 169: DQ Receiver (Rx) Mask
TdIVW_total
Rx Mask
VdIVW
VCENT_DQ(pin mid)
Figure 170: Across Pin VREF DQ Voltage Variation
DQx
DQz
DQy
VCENT_DQz
VCENT_DQx
VCENT_DQy
VREF variation
(component)
VCENT_DQ(pin_mid) is defined as the midpoint between the largest V CENT_DQ voltage level
and the smallest V CENT_DQ voltage level across all DQ pins for a given DRAM component. Each V CENT_DQ is defined by the center, which is the widest opening of the cumulative data input eye as shown in the figure above. This clarifies that any DRAM component level variation must be accounted for within the DRAM Rx mask. The componentlevel V REF will be set by the system to account for RON and ODT settings.
253
200b: x32 LPDDR4 SDRAM
DQ Rx Voltage and Timing
Figure 171: DQ-to-DQS tDQS2DQ and tDQDQ
DQ, DQS Data-in at DRAM Latch
DQS, DQs Data-in Skews at DRAM
Internal componsite data-eye center aligned to DQS
Nonminimum data-eye/maximum Rx mask
DQS_c
DQS_c
DQS_t
DQS_t
DQx
Rx mask
DRAM pin
tDQS2DQy2
All DQ signals center aligned to the
strobe at the device internal latch
DQy
Rx mask
DRAM pin
VdIVW_total
DQx, y, z
VdIVW_total
tDQS2DQ2
DQz
Rx mask
DRAM pin
VdIVW_total
tDQS2DQz2
tDQDQ
Notes:
1.
2.
3.
4.
These timings at the DRAM pins are referenced from the internal latch.
is measured at the center (midpoint) of the TdIVW window.
DQz represents the MAX tDQS2DQ in this example.
DQy represents the MIN tDQS2DQ in this example.
tDQS2DQ
All of the timing terms in DQ to DQS_t are measured from the DQS_t/DQS_c to the center (midpoint) of the TdIVW window taken at the V dIVW_total voltage levels centered
around V CENT_DQ(pin_mid). In figure above, the timings at the pins are referenced with respect to all DQ signals center-aligned to the DRAM internal latch. The data-to-data offset is defined as the difference between the MIN and MAX tDQS2DQ for a given component.
254
200b: x32 LPDDR4 SDRAM
DQ Rx Voltage and Timing
Figure 172: DQ tDIPW and SRIN_dIVW Definition for Each Input Pulse
UI = tCK(AVG) MIN/2
tr
tf
Rx Mask
VDIVW_total
VCENT_DQ(pin mid)
tDIPW
Note:
1. SRIN_dIVW = VdIVW_total/(tr or tf) signal must be monotonic within tr and tf range.
Figure 173: DQ VIHL(AC) Definition (for Each Input Pulse)
VIHL(AC)min/2
VCENT_DQ
Rx Mask
Rx Mask
VdIVW_total
Rx Mask
VIHL(AC)min/2
Table 188: DQs In Receive Mode
Note UI = tCK(AVG)(MIN)/2
Symbol
Parameter
VdIVW_total
Rx mask voltage – peak-topeak
VIHL(AC)
DQ AC input pulse amplitude peak-to-peak
SRIN_dIVW
Input slew rate over
VdIVW_total
Notes:
1600/1867
2133/2400
3200/3733
4267
Min
Max
Min
Max
Min
Max
Min
Max
Unit
Notes
–
140
–
140
–
140
–
120
mV
1, 2, 3
180
–
180
–
180
–
170
–
mV
5, 7
1
7
1
7
1
7
1
7
V/ns
6
1. Data Rx mask voltage and timing parameters are applied per pin and include the DRAM
DQ-to-DQS voltage AC noise impact for frequencies >20 MHz with a maximum voltage
of 45mV peak-to-peak at a fixed temperature on the package. The voltage supply noise
must comply to the component MIN/MAX DC operating conditions.
2. Rx mask voltage VdIVW_total (MAX) must be centered around VCENT_DQ(pin_mid).
3. Defined over the DQ internal VREF range. The Rx mask at the pin must be within the internal VREF DQ range irrespective of the input signal common mode.
4. Deterministic component of the total Rx mask voltage or timing. Parameter will be characterized and guaranteed by design.
5. DQ-only input pulse amplitude into the receiver must meet or exceed VIHL(AC) at any
point over the total UI. No timing requirement above level. VIHL(AC) is the peak-to-peak
voltage centered around VCENT_DQ(pin_mid), such that VIHL(AC)/2 (MIN) must be met both
above and below VCENT_DQ.
6. Input slew rate over VdIVW mask centered at VCENT_DQ(pin_mid).
255
200b: x32 LPDDR4 SDRAM
Clock Specification
7. VIHL(AC) does not have to be met when no transitions are occurring.
Clock Specification
The specified clock jitter is a random jitter with Gaussian distribution. Input clocks violating minimum or maximum values may result in device malfunction.
Table 189: Definitions and Calculations
Symbol
tCK(avg)
Description
and
nCK
Calculation
The average clock period across any consecutive
200-cycle window. Each clock period is calculated tCK(avg) =
from rising clock edge to rising clock edge.
Unit tCK(avg) represents the actual clock average
tCK(avg) of the input clock under operation. Unit
nCK represents one clock cycle of the input clock,
counting from actual clock edge to actual clock
edge.
Notes
N
Ȉ tCKj /N
j=1
Where N = 200
tCK(avg)
can change no more than ±1% within a
100-clock-cycle window, provided that all jitter
and timing specifications are met.
tCK(abs)
The absolute clock period, as measured from one
rising clock edge to the next consecutive rising
clock edge.
tCH(avg)
The average HIGH pulse width, as calculated
across any 200 consecutive HIGH pulses.
1
N
tCH(avg) =
Ȉ tCHj
/(N × tCK(avg))
j=1
Where N = 200
tCL(avg)
The average LOW pulse width, as calculated
across any 200 consecutive LOW pulses.
N
tCL(avg) =
Ȉ tCL
j
/(N × tCK(avg))
j=1
Where N = 200
tJIT(per)
The single-period jitter defined as the largest detJIT(per) = min/max of tCK – tCK(avg)
i
viation of any signal tCK from tCK(avg).
1
Where i = 1 to 200
tJIT(per),act
The actual clock jitter for a given system.
tJIT(per),
The specified clock period jitter allowance.
allowed
tJIT(cc)
The absolute difference in clock periods between
t
tJIT(cc) = max of tCK
i + 1 – CKi
two consecutive clock cycles. tJIT(cc) defines the
cycle-to-cycle jitter.
1
tERR(nper)
The cumulative error across n multiple consecutive cycles from tCK(avg).
1
i+n–1
tERR(nper) =
Ȉ
j=i
tERR(nper),act
The actual clock jitter over n cycles for a given
system.
256
tCK – (n × tCK(avg))
j
200b: x32 LPDDR4 SDRAM
Clock Period Jitter
Table 189: Definitions and Calculations (Continued)
Symbol
Description
tERR(nper),
The specified clock jitter allowance over n cycles.
Calculation
Notes
allowed
tERR(nper),min
The minimum tERR(nper).
tERR(nper),min = (1 + 0.68LN(n)) × tJIT(per),min
2
tERR(nper),max
The maximum tERR(nper).
tERR(nper),max = (1 + 0.68LN(n)) × tJIT(per),max
2
tJIT(duty)
Defined with absolute and average specifications tJIT(duty),min =
for tCH and tCL, respectively.
MIN((tCH(abs),min – tCH(avg),min),
(tCL(abs),min – tCL(avg),min)) × tCK(avg)
tJIT(duty),max
=
MAX((tCH(abs),max – tCH(avg),max),
(tCL(abs),max – tCL(avg),max)) × tCK(avg)
Notes:
tCK(abs), tCH(abs),
1. Not subject to production testing.
2. Using these equations, tERR(nper) tables can be generated for each tJIT(per),act value.
and tCL(abs)
These parameters are specified with their average values; however, the relationship between the average timing and the absolute instantaneous timing (defined in the following table) is applicable at all times.
Table 190: tCK(abs), tCH(abs), and tCL(abs) Definitions
Parameter
Symbol
Minimum
Absolute clock period
tCK(abs)
tCK(avg),min
+
Absolute clock HIGH pulse width
tCH(abs)
tCH(avg),min
+ tJIT(duty),min2/tCK(avg)min
tCK(avg)
Absolute clock LOW pulse width
tCL(abs)
tCL(avg),min
+ tJIT(duty),min2/tCK(avg)min
tCK(avg)
Notes:
Unit
tJIT(per),min
ps1
1. tCK(avg),min is expressed in ps for this table.
2. tJIT(duty),min is a negative value.
Clock Period Jitter
LPDDR4 devices can tolerate some clock period jitter without core timing parameter
derating. This section describes device timing requirements with clock period jitter
(tJIT(per)) in excess of the values found in the AC Timing table. Calculating cycle time
derating and clock cycle derating are also described.
Clock Period Jitter Effects on Core Timing Parameters
Core timing parameters (tRCD, tRP, tRTP, tWR, tWRA, tWTR, tRC, tRAS, tRRD, tFAW) extend across multiple clock cycles. Clock period jitter impacts these parameters when
measured in numbers of clock cycles. Within the specification limits, the device is characterized and verified to support tnPARAM = RU[tPARAM/tCK(avg)]. During device operation where clock jitter is outside specification limits, the number of clocks, or
tCK(avg), may need to be increased based on the values for each core timing parameter.
257
200b: x32 LPDDR4 SDRAM
Clock Period Jitter
Cycle Time Derating for Core Timing Parameters
For a given number of clocks (tnPARAM), when tCK(avg) and tERR(tnPARAM),act exceed
cycle time derating may be required for core timing parameters.
tERR(tnPARAM),allowed,
t
t
t
t
t
CycleTimeDerating = max PARAM + ERR( nPARAM),act – ERR( nPARAM),allowed – tCK(avg) , 0
tnPARAM
Cycle time derating analysis should be conducted for each core timing parameter. The
amount of cycle time derating required is the maximum of the cycle time deratings determined for each individual core timing parameter.
Clock Cycle Derating for Core Timing Parameters
For each core timing parameter and a given number of clocks (tnPARAM), clock cycle
derating should be specified with tJIT(per).
For a given number of clocks (tnPARAM), when tCK(avg) plus (tERR(tnPARAM),act) exceed the supported cumulative tERR(tnPARAM),allowed, derating is required. If the
equation below results in a positive value for a core timing parameter (tCORE), the required clock cycle derating will be that positive value (in clocks).
t
t
t
t
t
ClockCycleDerating = RU PARAM + ERR( nPARAM),act – ERR( nPARAM),allowed – tnPARAM
tCK(avg)
Cycle-time derating analysis should be conducted for each core timing parameter.
Clock Jitter Effects on Command/Address Timing Parameters
Command/address timing parameters (tIS, tIH, tISb, tIHb) are measured from a command/address signal (CS or CA[5:0]) transition edge to its respective clock signal (CK_t/
CK_c) crossing. The specification values are not affected by the tJIT(per) applied, because the setup and hold times are relative to the clock signal crossing that latches the
command/address. Regardless of clock jitter values, these values must be met.
Clock Jitter Effects on READ Timing Parameters
tRPRE
When the device is operated with input clock jitter, tRPRE must be derated by the
of the input clock that exceeds tJIT(per),allowed,max. Output deratings are relative to the input clock:
tJIT(per),act,max
tRPRE(min,derated) = 0.9 – tJIT(per),act,max – tJIT(per),allowed,max
tCK(avg)
For example, if the measured jitter into a LPDDR4 device has tCK(avg) = 625ps,
tJIT(per),act,min = –xx, and tJIT(per),act,max = +xx ps, then tRPRE,min,derated = 0.9 (tJIT(per),act,max - tJIT(per),allowed,max)/tCK(avg) = 0.9 - (xx - xx)/xx = yy tCK(avg).
258
200b: x32 LPDDR4 SDRAM
Clock Period Jitter
tLZ(DQ), tHZ(DQ), tDQSCK, tLZ(DQS), tHZ(DQS)
These parameters are measured from a specific clock edge to a data signal transition
(DMn or DQm, where: n = 0,1; and m = 0–15, and specified timings must be met with
respect to that clock edge. Therefore, they are not affected by tJIT(per).
tQSH, tQSL
These parameters are affected by duty cycle jitter, represented by tCH(abs)min and
tCL(abs)min. These parameters determine the absolute data-valid window at the device
pin. The absolute minimum data-valid window at the device pin = MIN {( tQSH(abs)min
- tDQSQmax), (tQSL(abs)min - tDQSQmax)}. This minimum data valid window must be
met at the target frequency regardless of clock jitter.
tRPST
tRPST
is affected by duty cycle jitter, represented by tCL(abs). Therefore, tRPST(abs)min
can be specified by tCL(abs)min. tRPST(abs)min = tCL(abs)min - 0.05 = tQSL(abs)min.
Clock Jitter Effects on WRITE Timing Parameters
tDS, tDH
These parameters are measured from a data signal (DMIn or DQm, where n = 0, 1 and m
= 0–15) transition edge to its respective data strobe signal (DQSn_t, DQSn_c: n = 0,1)
crossing. The specification values are not affected by the amount of tJIT(per) applied,
because the setup and hold times are relative to the data strobe signal crossing that
latches the command/address. Regardless of clock jitter values, these values must be
met.
tDSS, tDSH
These parameters are measured from a data signal (DQS_t, DQSn_c) crossing to its respective clock signal (CK_t, CK_c) crossing. When the device is operated with input
clock jitter, this parameter needs to be derated by the actual tJIT(per)act of the input clock
in excess of the allowed period jitter tJIT(per)allowed.
tDQSS
tDQSS
is measured from a data strobe signal (DQSn_t, DQSn_c) crossing to its respective clock signal (CK_t, CK_c) crossing. When the device is operated with input clock jitter, this parameter must be derated by the actual tJIT(per),act of the input clock in excess of tJIT(per)allowed.
tDQSS(min,derated) = 0.75 - tJIT(per),act,min – tJIT(per),allowed, min
tCK(avg)
tDQSS(max,derated) = 1.25 – tJIT(per),act,max – tJIT(per),allowed, max
tCK(avg)
For example, if the measured jitter into an LPDDR4 device has tCK(avg) = 625ps,
= -xxps, and tJIT(per),act,max = +xx ps, then:
tJIT(per),act,min
tDQSS,(min,derated)
= 0.75 - (-xx + yy)/625 = xxxx tCK(avg)
tDQSS,(max,derated)
= 1.25 - (xx – yy)/625 = xxxx tCK(avg)
259
200b: x32 LPDDR4 SDRAM
Revision History
Revision History
Rev. D – 3/17
• Added QDP die
• Added DS, DT, and NQ package codes
• Added XT temperature range
Rev. C – 11/16
• Added -053 speed grade
Rev. B – 8/16
• Production release
Rev. A – 5/16
• Initial Preliminary release (reference document: PDF: 09005aef865c23a1
366b_z0am_qdp_mobile_lpddr4.pdf – Rev. B 3/16 EN)
260