LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
LM10520 Single-Phase Buck Controller for AVS Systems
Check for Samples: LM10520
FEATURES
DESCRIPTION
•
•
•
The LM10520 is a single-phase Energy Management
Unit (EMU) that actively reduces system-level power
consumption by utilizing a continuous, real-time,
closed-loop Adaptive Voltage Scaling (AVS) scheme.
The AVS technology enables optimum energy
management delivery to the load in order to maximize
system-level energy savings.
1
23
•
•
•
•
•
•
•
•
•
•
•
•
•
Typical Power Savings with AVS: 20 to 50%
PWI 2.0 Interface
7-Bit AVS Control for One Output (Typical
Range of 0.6V to 1.2V)
Precision Enable
Integrated, Non-Overlapping NFET Drivers
Switching Frequency Over 50 kHz to 1MHz
Switching Frequency Synchronize Range from
250 kHz to 1MHz
Startup into Pre-Biased Output
Power Stage Input from 1V to 14V
Control stage Input from 3V to 6V
Power Good Flag and Shutdown
Output Over-Voltage and Under-Voltage
Detection
Low-Side Adjustable Current Sensing
Adjustable Soft Start
Tracking and Sequencing with Shutdown and
Soft-Start Pins
TSSOP-28 Package
APPLICATIONS
•
•
AVS-Enabled FPGAs
AVS-Enabled ASICs
The
LM10520
operates
cooperatively
with
PowerWise™ AVS-compatible ASICs, SoCs, and
processors to optimize supply voltages adaptively
over process and temperature variations. The device
is controlled via the high-speed serial PWI 2.0 openstandard interface. It also supports Dynamic Voltage
Scaling (DVS) using frequency-voltage pairs from
pre-characterized lookup tables.
The LM10520 features a fixed-frequency voltagemode PWM control architecture which is adjustable
from 50 kHz to 1MHz with one external resistor. In
addition, the LM10520 allows the switching frequency
to be synchronized to an external clock signal over
the range of 250 kHz to 1MHz. This wide range of
switching frequency gives the power supply designer
the flexibility to make better tradeoffs between
component size, cost and efficiency.
Features include the ability to startup with a prebiased load on the output, soft-start, input undervoltage lockout (UVLO) and Power good (based on
both under-voltage and over-voltage detection). In
addition, the soft-start pin can be used for
implementing precise tracking, for the purpose of
sequencing with respect to an external rail.
1
2
3
Please be aware that an important notice concerning availability, standard warranty, and use in critical applications of
Texas Instruments semiconductor products and disclaimers thereto appears at the end of this data sheet.
PowerWise is a trademark of Texas Instruments.
All other trademarks are the property of their respective owners.
PRODUCTION DATA information is current as of publication date.
Products conform to specifications per the terms of the Texas
Instruments standard warranty. Production processing does not
necessarily include testing of all parameters.
Copyright © 2010, Texas Instruments Incorporated
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
Typical Application Circuit
ASIC/FPGA
VIN
3.3 ± 14V
LM10520
VIN
VCC5
VCC
VDD
ON/OFF
EN_BIAS
CONTROL
BOOT
PWGD
FLT_N
RESET_N
CNTL_EN
SD
FREQ/SYNC
SS/TRACK
ADDR
VOUT
0.6 ± 1.2V, 20A
HG
ISEN
LG
CORE
HPM
PGND
FB
VPWI
EAO
SPWI
SCLK
IOUT
AVS
2
2
Submit Documentation Feedback
APC 2
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
Connection Diagram
GND (DAP)
BOOT 1
28 HG
LG 2
27 PGND
PGND 3
26 SD
VCC 4
25 FREQ/SYNC
24 FB
PWGD 5
23 SS/TRACK
ISEN 6
VIN 7
22 EAO
NC 8
21 ON/OFF
VCC5 9
20 SCLK
FLT_N 10
19 SPWI
VDD 11
18 VPWI
ADDR 12
17 CONTL_EN
ENBIAS 13
16 RESET_N
CONTROL 14
15 IAVS
Figure 1. 28-Lead Plastic TSSOP
Package Number PW0028A
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
3
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
PIN DESCRIPTIONS
4
Number
Pad Name
Type
Pad Description
1
DAP
GND
Connect Die Attach Pad to ground
BOOT
Analog
2
LG
Boot cap voltage. Connect boot capacitor to this pin.
Output
Low Side MOSFET gate drive.
3
PGND
GND
4
VCC
Power
5
PWGD
GND
Power good signal
6
ISEN
Analog
Current limit sense
7
VIN
Power
High voltage bias input
8
NC
9
VCC5
Power
10
FLT_N
I/O
Power ground.
5V bias input
No Connect
5V bias output
External fault input, active low. Causes output to be disabled and
resets R0 (output voltage register)
11
VDD
Power
Digital circuitry bias
12
ADDR
Analog
Connect a resistor from this to ground to set the PWI address
13
ENBIAS
Input
Enable for digital circuitry
14
CONTROL
Input
Enable for output voltage
15
IOUT1
Analog
16
RESET_N
Input
17
CNTL_EN
Output
Digital circuitry output which control Vout enable/disable
18
VPWI
Power
PWI I/O bias input
Connect this pin the FB pin
Digital circuitry reset, active low
19
SPWI
I/O
PWI signal
20
SCLK
Output
PWI clock
21
ON/OFF
Input
22
EAO
Analog
Error amp output
23
SS/TRACK
Analog
Connecting a capacitor to ground will set the softstart time. Optionally,
if this pin is driven externally the output will track the voltage at this pin
24
FB
25
FREQ/SYNC
Analog
26
SD
Input
Shutdown for the analog circuitry
27
PGND
GND
Power ground
28
HG
Analog
Enable for internal 5V LDO
Feedback connection
Connecting a resistor from this pin to ground will set the switching
frequency. Alternatively, a clock source can drive this pin, and the
LM10520 will synchronize to the clock frequency.
High side MOSFET drive
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
These devices have limited built-in ESD protection. The leads should be shorted together or the device placed in conductive foam
during storage or handling to prevent electrostatic damage to the MOS gates.
ABSOLUTE MAXIMUM RATINGS (1) (2)
ON/OFF, VIN
-0.3V to 16V
VCC
-0.3V to 6V
LG, PGND, SGND, PWGD, HG, SD, FB, SS/TRACK, EAO, FREQ/SYNC
-0.3V to VCC + 0.3V
BOOT
-0.3V to 18V
ISEN
-0.3V to 14V
VPWI
-0.2V to VDD
SPWI, SCLK
-0.2 to VPWI
All other pins
-0.2V to 6V
Junction Temperature
150°C
Storage Temperature
ESD Tolerance
-45°C to 150°C
(3)
Human Body Model
1.5 kV
Soldering Information
See product folder at www.ti.com and literature number SNOA549
(1)
(2)
(3)
Absolute maximum ratings indicate limits beyond which damage to the device may occur. Operating ratings indicate conditions for which
the device operates correctly. Operating Ratings do not imply ensured performance limits.
If Military/Aerospace specified devices are required, please contact the Texas Instruments Sales Office/Distributors for availability and
specifications.
ESD using the human body model which is a 100 pF capacitor discharged through a 1.5 kΩ resistor into each pin.
OPERATING RATINGS
VIN
3.5V to 16V
VCC, VDD
3V to 5.5V
VPWI (1)
1.6V to 3.6V
BOOT Voltage
1V to 17V
Junction Temperature
(1)
-40°C to 125°C
Note: VPWI cannot be higher than VDD
THERMAL PROPERTIES
Junction-to-Ambient Thermal Resistance
26°C/W
ELECTRICAL CHARACTERISTICS
VCC = 3.3V unless otherwise indicated. Typicals and limits appearing in plain type apply for TA= TJ= 25°C. Limits appearing in
boldface type apply over full Operating Temperature Range. Datasheet min/max specification limits are specified by design,
test, or statistical analysis.
Symbol
Parameter
Conditions
Min
Typ
Max
Units
0.591
0.6
0.609
V
VFB
FB Pin Voltage
VCC = 3V to 6V
VON
UVLO Thresholds
VCC Rising
VCC Falling
2.79
2.42
VCC = 3.3V, VSD = 3.3V
fSW = 600 kHz, VCC connected to
VDD
2.15
VCC = VDD = 5V, VSD = 5V
fSW = 600 kHz
2.315
3.01
Operating VCC/VDD Current
IQ
V
2.71
mA
Shutdown VCC/VDD Current
VCC = VDD = 3.3V,
VSD = EN_BIAS= 0V
2.5
13
Shutdown VIN Current
ON/OFF = 0V
0.05
2
tPWGD1
PWGD Pin Response Time
VFB Rising
tPWGD2
PWGD Pin Response Time
VFB Falling
ISS-ON
SS Pin Source Current
VSS = 0V
µA
10
µs
10
7
10
µs
14
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
µA
5
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
ELECTRICAL CHARACTERISTICS (continued)
VCC = 3.3V unless otherwise indicated. Typicals and limits appearing in plain type apply for TA= TJ= 25°C. Limits appearing in
boldface type apply over full Operating Temperature Range. Datasheet min/max specification limits are specified by design,
test, or statistical analysis.
Symbol
Parameter
ISS-OC
SS Pin Sink Current During Over
Current
ISEN-TH
ISEN Pin Source Current Trip Point
IFB
FB Pin Current
IADDR
Address pin source current
Conditions
Min
VSS = 2.0V
Typ
Max
90
25
Sourcing
40
Units
µA
55
µA
20
nA
7.8
µA
IAVS
LSB
IDAC-MAX / 2n (1 ≤ n ≤ 7)
DAC Step size
Resolution
FS
Full Scale
INL
Integral Non-Linearity
DNL
Differential Non-Linearity
ZE
Zero Code Error/Offset Error
470
nA
7
Bits
56.69
µA
−2
2
−0.5
0.5
LSB
57
nA
9
MHz
ERROR AMPLIFIER
GBW
Error Amplifier Unity Gain Bandwidth
G
Error Amplifier DC Gain
SR
118
dB
Error Amplifier Slew Rate
2
V/µs
IEAO
EAO Pin Current Sourcing and
Sinking Capability
14
16
mA
VEAO
Error Amplifier Output Voltage
Minimum
1
V
Maximum
2.2
V
GATE DRIVE
IQ-BOOT
BOOT Pin Quiescent Current
VBOOT = 12V, VSD = 0
18
RHG_UP
High-Side MOSFET Driver Pull-Up
ON resistance
VBOOT = 5V @ 350 mA Sourcing
2.7
Ω
RHG_DN
High-Side MOSFET Driver Pull-Down
ON resistance
350 mA Sinking
0.8
Ω
RLG_UP
Low-Side MOSFET Driver Pull-Up ON
VBOOT = 5V @ 350 mA Sourcing
resistance
2.7
Ω
RLG_DN
Low-Side MOSFET Driver Pull-Down
ON resistance
0.8
Ω
350 mA Sinking
90
µA
OSCILLATOR
PWM Frequency
RFADJ = 750 kΩ
50
RFADJ = 100 kΩ
300
RFADJ = 42.2 kΩ
fSW
475
RFADJ = 18.7 kΩ
LM10520 External Synchronizing
Signal Frequency
Voltage Swing = 0V to VCC
SYNCL
LM10520 Synchronization Signal Low
Threshold
fSW = 250 kHz to 1 MHz
SYNCH
LM10520 Synchronization Signal High
fSW = 250 kHz to 1 MHz
Threshold
DMAX
Max High-Side Duty Cycle
6
fSW = 300 kHz
fSW = 600 kHz
fSW = 1 MHz
Submit Documentation Feedback
600
725
kHz
1000
250
1000
1
2
V
V
86
78
67
%
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
ELECTRICAL CHARACTERISTICS (continued)
VCC = 3.3V unless otherwise indicated. Typicals and limits appearing in plain type apply for TA= TJ= 25°C. Limits appearing in
boldface type apply over full Operating Temperature Range. Datasheet min/max specification limits are specified by design,
test, or statistical analysis.
Symbol
Parameter
Conditions
Min
Typ
Max
1.32
1.4
Units
LOGIC AND CONTROL INPUTS
EN_BIASTH
Precision enable threshold
IIL
Input Current Low
Input Current High
IIH
Rising
Falling
1.09
FLT_N, RESET_N
−10
ENBIAS, CONTROL
−1
SPWI, SCLK
−1
1
ENBIAS, CONTROL
10
SPWI, SCLK
5
Input Low Voltage
CONTROL, FLT_N, RESET_N
VIH
Input High Voltage
CONTROL, FLT_N, RESET_N
VIL PWI
Input Low Voltage, PWI
SPWI, SCLK, 1.5 < VPWI < 3.3
VIH
Input High Voltage, PWI
SPWI, SCLK, 1.5 < VPWI < 3.3
70
fSCLK
PWI2 SCLK
**DC useful for testing/debug
0
VSTBY-IH
Standby High Trip Point
VFB = 0.575V, VBOOT = 3.3V
VSD Rising
VSTBY-IL
Standby Low Trip Point
VFB = 0.575V, VBOOT = 3.3V
VSD Falling
VSD-IH
SD Pin Logic High Trip Point
VSD Rising
VSD-IL
SD Pin Logic Low Trip Point
VSD Falling
PWI
µA
FLT_N, RESET_N
VIL
V
1.19
µA
0.5
V
1.1
30
% of
VPWI
15M
Hz
1.1
0.232
V
V
1.3
0.8
V
V
LOGIC AND CONTROL OUTPUTS
VOL
Output Low Level
CNTL_EN, ISINK ≤ 1mA
VOH
Output High Level
CNTL_EN, ISINK ≤ 1mA
VDD 0.4
Output High Level, PWI
SPWI, ISOURCE ≤ 1mA
VPWI 0.4
VPWGD-TH-LO
PWGD Pin Trip Points
VFB Falling
0.408
0.434
0.457
V
VPWGD-TH-HI
PWGD Pin Trip Points
VFB Rising
0.677
0.710
0.742
V
VPWGD-HYS
PWGD Hysteresis
VFB Falling
VFB Rising
VOH
PWI
0.4
V
60
90
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
mV
7
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
TYPICAL PERFORMANCE CHARACTERISTICS
8
Internal Reference Voltage
vs
Temperature
Frequency
vs
Temperature
Figure 2.
Figure 3.
Output Voltage
vs
Output Current
Switch Waveforms
VCC = 3.3V, VIN = 5V, VOUT = 1.2V
IOUT = 3A, CSS = 12 nF, fSW = 1MHz
Figure 4.
Figure 5.
Start-Up (Full-Load)
VCC = 3.3V, VIN = 5V, VOUT = 1.2V
IOUT = 3A, CSS = 12 nF, fSW = 1MHz
Start-Up (No-Load)
VCC = 3.3V, VIN = 5V, VOUT = 1.2V
CSS = 12 nF, fSW = 1MHz
Figure 6.
Figure 7.
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
TYPICAL PERFORMANCE CHARACTERISTICS (continued)
Shutdown (Full-Load)
VCC = 3.3V, VIN = 5V, VOUT = 1.2V
IOUT = 3A, CSS = 12 nF, fSW = 1MHz
Load Transient Response
PVIN = 12V, AVIN = 4.5V,
VOUT = 1.2V, FSW = 300 kHz
Figure 8.
Figure 9.
Line Transient Response (VIN = 3V to 9V)
VCC = 3.3V, VOUT = 1.2V
IOUT = 2A, fSW = 1 MHz
Frequency
vs.
Frequency Adjust Resistor
Figure 10.
Figure 11.
Maximum Duty Cycle
vs
Frequency
VCC = 3.3V
Maximum Duty Cycle
vs
VCC
fSW = 600 kHz
Figure 12.
Figure 13.
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
9
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
TYPICAL PERFORMANCE CHARACTERISTICS (continued)
Maximum Duty Cycle
vs
VCC
fSW = 1 MHz
Figure 14.
10
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
BLOCK DIAGRAM
FREQ/SYNC*
PLL*
VCC
SD
SHUT DOWN
LOGIC
CLOCK &
RAMP
PGND
PGND
UVLO
BOOT
10 Ps
DELAY
HG
SSDONE
PWGD
SYNCHRONOUS
DRIVER LOGIC
OV
UV
LG
0.71V
0.434V
10 PA
SS/TRACK
Zero Detector
Soft-Start
Comparator
+ Logic
1VPP
REF
90 PA
PWM LOGIC
+
PWM
40 PA
EA
VREF=0.6V
-
ISEN
ILIM
+
FB
ON/OFF
EAO
VIN
LDO
VCC5
VDD
CNTL_EN
POR
IDAC
IAVS
VPWI
CONTROL
FLT_N
Slave Power
Controller
(SPC 2)
SPWI
SCLK
RESET_N
ADDR
EN_BIAS
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
11
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
LM10520 PWI REGISTER MAP
The PWI 2.0 standard defines 32 8-bit base registers, and up to 256 8-bit extended registers, on each PWI slave.
The table below summarizes these registers and shows default register bit values after reset, as programmed by
the factory. The following sub-sections provide additional details on the use of each individual register.
Table 1. SUMMARY
Base Registers
Register
Address
Register
Name
Register Usage
Type
Reset Default Value
7
6
5
4
R/W
0* (1)
R/O
0
0
0
Device Capability
R/O
0
0
0
IAVS Default
R/W
R10
Ramp Control
R/W
1
0
0
1
0x0F
R15
Revision ID
N/A
0
0
0
0x1F
R31
Reserved
Do not write to
R/W
-
-
-
0x00
R0
0x03
R3
0x04
R4
0x09
R9
0x0A
(1)
IAVS
3
2
1
0
Configured by R9
0
1
1
1
FLT_N
0
0
0
1
0
1
1
0
0
0
0
0
0
0
-
-
-
-
-
I AVS Default Code
Note: A bit with an asterisk (*) denotes a register bit that is always read as a fixed value. Writes to these bits will be ignored. A bit with a
hyphen (-) denotes a bit in an unimplemented register location. A write into unimplemented register(s) will be ignored. A read of an
unimplemented register(s) will produce a “No response frame”. Please refer to PWI specification version 2.0 for further information.
Table 2. R0 - IAVS
AVS Feedback Current Injection
Address
0x00
Type
R/W
Reset Default
8h'7F
Bit
Field Name
7
6:0
Description or Comment
Sign
This bit is fixed to '0'. Reading this bit will result in a '0'. Any data written into this bit position
using the Register Write command is ignored.
IAVS Sourcing Current
Programmed voltage value. Default value is in bold.
Current Data Code [6:0]
Current (µA)
7h'00
60
7h'xx
Linear Scaling
7h'7F
0 (default)
Table 3. R3 - Status
LM10520 Status
Address
0x03
Type
R/O
Reset Default
–
Bit
Description or Comment
7:4
Not Used
Always read back 0
3:1
Not Used
Always read back 1
FLT_N
1: FLT_N is high (no fault)
0: FLT_N is low (fault)
0
12
Field Name
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
Table 4. R4 - Device Capability Register
Address
0x04
Type
R/O
Reset Default
8h'02
Bit
Field Name
Description or Comment
7:3
Always read back 0
2:0
Always read back 010, specifying PWI 2.0
Table 5. R9 - IAVS Default Register
Address
0x09
Type
R/W
Reset Default
8h'7F
Bit
7
6:0
Field Name
Description or Comment
Sign
Always read back 0.
IAVS Default
Current Data Code [6:0]
Current (µA)
7h'00
60
7h'xx
Linear Scaling
7h'7F
0 (default)
Table 6. R10 - Ramp Control
Address
0x0A
Type
R/W
Reset Default
8h'9C
Bit
Field Name
Description or Comment
7
Ramp Control Enable
1: Enabled
0: Disabled
6
Not Used
Always read 0
Ramp Time Step Control
Ramp Time Step Control
5:3
0
0
0
0
1
1
1
1
2:0
Ramp Code Step Control
0
0
1
1
0
0
1
1
Ramp Time Step (µs)
0
1
0
1
0
1
0
1
Ramp Code Step
0
0
0
0
1
1
1
1
1
2
3
4
6
8
12
16
Rising Step (LSB) Falling Step
(LSB)
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
1
2
3
4
6
8
12
16
1
1
2
3
4
5
6
8
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
13
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
Table 7. R15 - Revision ID Register
Address
0x7F
Type
R/O
Reset Default
8h'00
Bit
Field Name
7:0
14
Description or Comment
Always read back 0
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
APPLICATION INFORMATION
The device is a PowerWise Interface (PWI) compliant energy management unit (EMU). It operates cooperatively
with processors using Texas Instruments' Advanced Power Controller (APC) to provide Adaptive or Dynamic
Voltage Scaling (AVS, DVS) which drastically improves processor efficiencies compared to conventional power
delivery methods. The device consists of PWI registers, logic, and a switching DC/DC buck controller to supply
the AVS or DVS voltage domain.
VOLTAGE SCALING
The device is designed to be used in a voltage scaling system to lower the power dissipation of SoC or ASICs.
By scaling the supply voltage with the clock frequency and process variations, dramatic power savings can be
achieved. Two types of voltage scaling are supported, dynamic voltage scaling (DVS) and adaptive voltage
scaling (AVS). DVS systems switch between pre-characterized voltages which are paired to clock frequencies
used for frequency scaling in the ASIC. AVS systems track the ASIC’s performance and optimizes the supply
voltage to the required performance. AVS is a closed loop system that provides process and temperature
compensation such that for any given process, temperature, or clock frequency, the minimum supply voltage is
delivered.
DIGITALLY ASSISTED VOLTAGE SCALING
The device delivers fast, controlled voltage scaling transients with the help of a digital state machine. The state
machine automatically optimizes the slew rate of the output to provide large signal transients with minimal overand undershoot. This is an important characteristic for voltage scaling systems that rely on minimal over- and
undershoot to set voltages as low as possible and save energy.
POWERWISE INTERFACE
The device is programmable via the low-power, 2-wire PowerWise Interface (PWI). This serial interface controls
the various voltages and states of the regulator in the device. The output voltage is programmable with 7-bit
resolution and an adjustable range, set by the feedback resistors, from 0.6V – Vin*Dmax (see ELECTRICAL
CHARACTERISTICS for Dmax). This high resolution voltage control affords accurate temperature and process
compensation in AVS. The device supports the full command set as described in PWI 1.0/2.0 specification:
• Core Voltage Adjust
• Reset
• Sleep
• Shutdown
• Wakeup
• Register Read
• Register Write
• Authenticate
• Synchronize
The output voltage of the switching regulator is programmed via the Core Voltage Adjust command.
PWI ADDRESS
A resistor from the ADDR pin to ground sets the device's PWI address. The device senses the resistance as it is
initializing from the shutdown state. The device will not update the address until it cycles through shutdown
again. Use the table below to choose the appropriate resistor to place form ADDR pin to ground.
PWI Address
Resistance (± 1% tolerance)
0
≤ 40.2 kΩ
1
60.4 kΩ
2
80.6 kΩ
3
100 kΩ
4
120 kΩ
5
140 kΩ
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
15
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
PWI Address
Resistance (± 1% tolerance)
6
160 kΩ
7
180 kΩ
INPUTS: ON/OFF, ENBIAS, CONTROL, FLT_N, RESET_N, SCLK, SPWI
•
•
•
•
•
•
•
ON/OFF:
– The ON/OFF logic input enables the internal LDO (VCC5 output pin).
ENBIAS:
– The ENBIAS logic input enables the internal logic of the LM10520. The LM10520 goes through an
initialization procedure upon the rising edge of ENBIAS. Initialization is complete within 100 µsec, after
which the device is ready to be used. If at any point ENBIAS goes low, the device enters a low Iq
shutdown state.
– The ENBIAS input is buffered internally by a Schmitt trigger with precision thresholds to allow accurate
output voltage sequencing off of the input rail.
CONTROL:
– The CONTROL logic input allows control of the CNTL_EN output without incurring delays associated with
initialization. This signal is effectively ANDed with the internal ‘ready’ signal, which is high once
initialization is complete.
– The CONTROL pin level toggles the device between Active and Sleep states, and will reset the R0
register.
FLT_N:
– The FLT_N logic input resets and holds the R0 register when its input signal is low. It has no effect on
CNTL_EN. This provides a convenient way to support automatic fault recovery modes in the slave power
regulator. When connected to a standard PWGD pin of a DC/DC regulator, FLT_N will reset and hold R0
as long as PWGD is low, allowing the slave regulator to recover from the fault by returning to the default
voltage. Once FLT_N returns high, R0 can be written to.
RESET_N:
– The RESET_N provides a separate, level controlled logic reset.
SCLK and SPWI:
– SCLK and SPWI provide serial PWI communication.
CNTL_EN:
– The CNTL_EN output connects to the SD# pin. CNTL_EN allows power state control via the PWI interface
or ENBIAS/CONTROL logic inputs. LM10520 will drive CNTL_EN to the VDD voltage to enable the buck
controller circuitry, and to 0 V to disable the circuitry.
IAVS OUTPUT CURRENT: CONTROLLING THE OUTPUT VOLTAGE
The LM10520 uses a 7-bit current DAC to control the output voltage. Since it is a current output, IAVS can be
connected directly to the feedback node. IAVS has a range of 0 – 60 µA with 7 bits of resolution, or a 0.469 µA
LSB.
IAVS should be connected to the feedback node of LM10520 as shown in Figure 15. The output voltage VOUT is
expressed as:
RFB1
- IAVS x RFB1
VOUT = VFB x §1+
R
© FB2
§
©
where
•
VFB = the regulated feedback voltage of the slave regulator
(1)
This equation is valid for VOUT > VFB.
16
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
VOUT
LM10520
EAO
IAVS
FB
Z1
RFB1
Z2
RFB2
Figure 15. Connecting IAVS to the Feedback Node
Using Register R9 to Change the Default Output Voltage
The LM10520 default IAVS current is set by R9. R9 is trimmed to 0x7F, so that IAVS = 0µA when power is
applied to LM10520. Between power cycling, R9 can be changed so that IAVS defaults to values between 0 - 60
μA. This can be useful for software trim of the default output voltage of the LM10520 controlled regulator. In
order to do this, the system must take care to write to R9 before enabling the output. (The output can be
enabled/disabled while keeping the LM10520 logic on via the CONTROL input.) Therefore, R9 must be written to
by some system controller that is on a different power domain than that provided by LM10520. In addition, the
“INITIAL_VDD” register in the Advanced Power Controller (APC) must have the same value as R9 so that the
APC and LM10520 default to the same voltage code.
Digital Slew Rate Control
The IAVS and IAVS Mirror outputs have an adjustable, digital slew rate control. The slew rate control is
programmed in register R10.
Single PWI Core Voltage
Adjust Command (value)
Code Step
Time Step
Figure 16. Digital Slew Rate Control
LM10520 effectively overrides the internal reference of the slave regulator to allow it to operate in an AVS
system. The amount of current drawn from the AVS enabled power supply when scaling voltage depends on
several factors determined by the well known equation for current in a capacitor:
IC = Cdv/dt
where
•
•
•
C is the total output capacitance seen by the power supply
dv is the voltage step
dt is the step time
(2)
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
17
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
The digital slew rate control of LM10520 allows independent manipulation of the dv and dt terms to
accommodate a wide range of output capacitances.
STATES
Startup
During the startupt state, the LM10520 initializes all its registers and enables its bandgap. This process typically
takes 1.116 msec. CNTL_EN is low during startup.
Active
During the active state, CNTL_EN is is high, the IAVS DACs are enabled, and PWI registers can be accessed.
Sleep
During the sleep state, CNTL_EN is low, the IAVS DACs are disabled, and PWI registers can be accessed.
Fault
During the fault state, the IAVS current register (R0) is reset, after which the LM10520 automatically returns to its
previous state.
Shutdown
During the shutdown state, CNTL_EN is low, the IAVS DACs are disabled, and most internal circuitry is disabled.
Only the PWI state machine is biased to allow register access.
Pin State
ENBIAS Falling Edge
ENBIAS Rising Edge
Shutdown
CNTL_EN = 0
Low Iq state
RESET Cmd
Startup
CNTL_EN = 0
Shutdown
Cmd
Shutdown
Cmd
CONTROL = 0
CONTROL = 0
Wakeup Cmd
(Reset R0)
CONTROL = 1
(Reset R0)
SLEEP
CNTL_EN = 0
CONTROL = 1
Active
CNTL_EN = 1
CONTROL = 0
Sleep Cmd
FLTN = 1
FLTN = 0
FLTN = 0
FLTN = 1
Fault
Reset and hold
R0
Figure 17. LM10520 State Diagram
18
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
VOLTAGE MODE CONTROLLER
The LM10520 incorporates control circuitry and drivers for synchronous buck PWM regulation. It uses voltage
mode control to achieve the low duty cycles necessary for the low conversion ratios in an AVS system. It has
flexible input enable controls to allow logic level control of output voltage enable, bias enable, and internal LDO
enable. In addition, an active low fault input can be used for system fault response sequencing. These inputs
allow simple, system level control of the device state. The LM10520 also includes input under-voltage lock-out
(UVLO) and a power good (PWGD) flag (based on over-voltage and under-voltage detection). The over-voltage
and under-voltage signals are OR-gated to drive the power good signal and provide a logic signal to the system if
the output voltage goes out of regulation. Current limit is achieved by sensing the voltage VDS across the low
side MOSFET. The LM10520 is also able to stat-up with the output pre-biased with a load. The LM10520 also
allows the switching frequency to be a synchronized with an external clock source.
START UP/SOFT-START
When VCC exceeds 2.79V and the shutdown pin (SD) sees a logic high, the soft-start period begins. Then an
internal, fixed 10 µA source begins charging the soft-start capacitor. During soft-start the voltage on the soft-start
capacitor CSS is connected internally to the non-inverting input of the error amplifier. The soft-start period lasts
until the voltage on the soft-start capacitor exceeds the LM10520 reference voltage of 0.6V. At this point the
reference voltage takes over at the non-inverting error amplifier input. The capacitance of CSS determines the
length of the soft-start period, and can be approximated by:
CSS =
tSS
60
where
•
•
CSS is in µF
tSS is in ms
(3)
During soft start the Power Good flag is forced low and it is released when the FB pin voltage reaches 70% of
0.6V. At this point the chip enters normal operation mode, and the output overvoltage and undervoltage
monitoring starts.
SETTING THE DEFAULT AND PROGRAMMABILITY RANGE OF THE OUTPUT VOLTAGE
The LM10520 has a flexible output voltage range control. When the system is starting up, the output voltage
exits soft-start and AVS has not been enabled by the load (the load is the AVS enabled processor, and is booting
up). This is the default voltage of the LM10520, and is set by the feedback resistor divider ratio. Once the
automatic AVS authentication process has successfully completed, the AVS loop is engaged and the LM10520
automatically reduces the voltage. A 7-bit current source is injected into the feedback resistor node to achieve
voltage scaling. Therefore, the range and resolution of the device is adjustable via the top feedback resistor.
SETTING THE SWITCHING FREQUENCY
During fixed-frequency mode of operation the PWM frequency is adjustable between 50 kHz and 1 MHz and is
set by an external resistor, RFADJ, between the FREQ/SYNC pin and ground. The resistance needed for a
desired frequency is approximated by the curve Frequency vs. Frequency Adjust Resistor in the Typical
Performance Characteristics.
When it is desired to synchronize the switching frequency with an external clock source, the LM10520 has the
unique ability to synchronize from this external source within the range of 250 kHz to 1MHz. The external clock
signal should be AC coupled to the FREQ/SYNC pin as shown below in Figure 18, where the RFADJ is chosen so
that the fixed frequency is approximately within ±30% of the external synchronizing clock frequency. An internal
protection diode clamps the low level of the synchronizing signal to approximately -0.5V. The internal clock
sychronizes to the rising edge of the external clock.
CCLK
External Clock
Signal
To FREQ/SYNC Pin
RFADJ
Figure 18. AC Coupled Clock
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
19
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
It is recommended to choose an AC coupling capacitance in the range of 50 pF to 100 pF. Exceeding the
recommended capacitance may inject excessive energy through the internal clamping diode structure present on
the FREQ/SYNC pin.
The typical trip level of the synchronization pin is 1.5V. To ensure proper synchronization and to avoid damaging
the IC, the peak-to-peak value (amplitude) should be between 2.5V and VCC. The minimum width of this pulse
must be greater than 100 ns, and it's maximum width must be 100 ns less than the period of the switching cycle.
The external clock synchronization process begins once the LM10520 is enabled and an external clock signal is
detected. During the external clock synchronization process the internal clock initially switches at approximately
1.5 MHz and decreases until it has matched the external clock’s frequency. The lock-in period is approximately
30 µs if the external clock is switching at 1MHz, and about 100 µs if the external clock is at 200 kHz. When there
is no clock signal present, the LM10520 enters into fixed-frequency mode and begins switching at the frequency
set by the RFADJ resistor. If the external clock signal is removed after frequency synchronization, the LM10520
will enter fixed-frequency mode within two clock cycles. If the external clock is removed within the 30 µs lock-in
period, the LM10520 will re-enter fixed-frequency mode within two internal clock cycles after the lock-in period.
OUTPUT PRE-BIAS STARTUP
If there is a pre-biased load on the output of the LM10520 during startup, the IC will disable switching of the lowside MOSFET and monitor the SW node voltage during the off-time of the high-side MOSFET. There is no load
current sensing while in pre-bias mode because the low-side MOSFET never turns on. The IC will remain in this
pre-bias mode until it sees the SW node stays below 0V during the entire high-side MOSFET's off-time. Once it
is determined that the SW node remained below 0V during the high-side off-time, the low-side MOSFET begins
switching during the next switching cycle. Figure 19 shows the SW node, HG, and LG signals during pre-bias
startup. The pre-biased output voltage should not exceed VCC + VGS of the external High-Side MOSFET to
ensure that the High-Side MOSFET will be able to switch during startup.
Figure 19. Output Pre-Bias Mode Waveforms
TRACKING A VOLTAGE LEVEL
The LM10520 can track the output of a master power supply during soft-start by connecting a resistor divider to
the SS/TRACK pin. In this way, the output voltage slew rate of the LM10520 will be controlled by the master
supply for loads that require precise sequencing. When the tracking function is used no soft-start capacitor
should be connected to the SS/TRACK pin. However in all other cases, a CSS value of at least 1nF between the
soft-start pin and ground should be used.
20
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
Master Power
Supply
VOUT1 = 5V
RT2
1 k:
VOUT2 = 1.8V
SS/TRACK
VSS = 0.65V
RT1
150:
LM10520
FB
RFB2
10 k:
VFB
RFB1
5 k:
Figure 20. Tracking Circuit
One way to use the tracking feature is to design the tracking resistor divider so that the master supply’s output
voltage (VOUT1) and the LM10520’s output voltage (represented symbolically in Figure 20 as VOUT2, i.e. without
explicitly showing the power components) both rise together and reach their target values at the same time. For
this case, the equation governing the values of the tracking divider resistors RT1 and RT2 is:
0.65 = VOUT1
RT1
RT1 + RT2
(4)
The current through RT1 should be about 4mA for precise tracking. The final voltage of the SS/TRACK pin should
be set higher than the feedback voltage of 0.6V (say about 0.65V as in the above equation). If the master supply
voltage was 5V and the LM10520 output voltage was 1.8V, for example, then the value of RT1 needed to give the
two supplies identical soft-start times would be 150Ω. A timing diagram for the equal soft-start time case is
shown in Figure 21.
5V
VOUT1
1.8V
VOUT2
Figure 21. Tracking with Equal Soft-Start Time
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
21
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
TRACKING A VOLTAGE SLEW RATE
The tracking feature can alternatively be used not to make both rails reach regulation at the same time but rather
to have similar rise rates (in terms of output dV/dt). This method ensures that the output voltage of the LM10520
always reaches regulation before the output voltage of the master supply. In this case, the tracking resistors can
be determined based on the following equation:
0.65 = VOUT2
RT1
RT1 + RT2
(5)
For the example case of VOUT1 = 5V and VOUT2 = 1.8V, with RT1 set to 150Ω as before, RT2 is calculated from the
above equation to be 265Ω. A timing diagram for the case of equal slew rates is shown in Figure 22.
5V
1.8V
VOUT1
1.8V
VOUT2
Figure 22. Tracking with Equal Slew Rates
MOSFET GATE DRIVERS
The LM10520 has two gate drivers designed for driving N-channel MOSFETs in a synchronous mode. Note that
unlike most other synchronous controllers, the bootstrap capacitor of the LM10520 provides power not only to the
driver of the upper MOSFET, but the lower MOSFET driver too (both drivers are ground referenced, i.e. no
floating driver).
Two things must be kept in mind here. First, the BOOT pin has an absolute maximum rating of 18V. This must
never be exceeded, even momentarily. Since the bootstrap capacitor is connected to the SW node, the peak
voltage impressed on the BOOT pin is the sum of the input voltage (VIN) plus the voltage across the bootstrap
capacitor (ignoring any forward drop across the bootstrap diode). The bootstrap capacitor is charged up by a
given rail (called VBOOT_DC here) whenever the upper MOSFET turns off. This rail can be the same as VCC or it
can be any external ground-referenced DC rail. But care has to be exercised when choosing this bootstrap DC
rail that the BOOT pin is not damaged. For example, if the desired maximum VIN is 14V, and VBOOT_DC is chosen
to be the same as VCC, then clearly if the VCC rail is 6V, the peak voltage on the BOOT pin is 14V + 6V = 20V.
This is unacceptable, as it is in excess of the rating of the BOOT pin. A VCC of 3V would be acceptable in this
case. Or the VIN range must be reduced accordingly. There is also the option of deriving the bootstrap DC rail
from another 3V external rail, independent of VCC.
The second thing to be kept in mind here is that the output of the low-side driver swings between the bootstrap
DC rail level of VBOOT_DC and Ground, whereas the output of the high-side driver swings between VIN+ VBOOT_DC
and Ground. To keep the high-side MOSFET fully on when desired, the Gate pin voltage of the MOSFET must
be higher than its instantaneous Source pin voltage by an amount equal to the 'Miller plateau'. It can be shown
that this plateau is equal to the threshold voltage of the chosen MOSFET plus a small amount equal to Io/g. Here
Io is the maximum load current of the application, and g is the transconductance of this MOSFET (typically about
100 for logic-level devices). That means we must choose VBOOT_DC to at least exceed the Miller plateau level.
This may therefore affect the choice of the threshold voltage of the external MOSFETs, and that in turn may
depend on the chosen VBOOT_DC rail.
22
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
So far, in the discussion above, the forward drop across the bootstrap diode has been ignored. But since that
does affect the output of the driver somewhat, it is a good idea to include this drop in the following examples.
Looking at the Typical Application schematic, this means that the difference voltage VCC - VD1, which is the
voltage the bootstrap capacitor charges up to, must always be greater than the maximum tolerance limit of the
threshold voltage of the upper MOSFET. Here VD1 is the forward voltage drop across the bootstrap diode D1.
This may place restrictions on the minimum input voltage and/or type of MOSFET used.
A basic bootstrap circuit can be built using one Schottky diode and a small capacitor, as shown in Figure 23. The
capacitor CBOOT serves to maintain enough voltage between the top MOSFET gate and source to control the
device even when the top MOSFET is on and its source has risen up to the input voltage level. The charge pump
circuitry is fed from VCC, which can operate over a range from 3.0V to 6.0V. Using this basic method the voltage
applied to the gates of both high-side and low-side MOSFETs is VCC - VD. This method works well when VCC is
5V±10%, because the gate drives will get at least 4.0V of drive voltage during the worst case of VCC-MIN = 4.5V
and VD-MAX = 0.5V. Logic level MOSFETs generally specify their on-resistance at VGS = 4.5V. When VCC =
3.3V±10%, the gate drive at worst case could go as low as 2.5V. Logic level MOSFETs are not ensured to turn
on, or may have much higher on-resistance at 2.5V. Sub-logic level MOSFETs, usually specified at VGS = 2.5V,
will work, but are more expensive, and tend to have higher on-resistance. The circuit in Figure 23 works well for
input voltages ranging from 1V up to 14V and VCC = 5V±10%, because the drive voltage depends only on VCC.
LM10520
BOOT
D1
VCC
CBOOT
VIN
HG
+
VO
+
LG
Figure 23. Basic Charge Pump (Bootstrap)
Note that the LM10520 can be paired with a low cost linear regulator like the LM78L05 to run from a single input
rail between 6.0 and 14V. The 5V output of the linear regulator powers both the VCC and the bootstrap circuit,
providing efficient drive for logic level MOSFETs. An example of this circuit is shown in Figure 24.
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
23
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
LM10520
VCC
5V
LM78L05
D1
BOOT
CBOOT
VIN
+
HG
VO
LG
+
Figure 24. LM78L05 Feeding Basic Charge Pump
Figure 25 shows a second possibility for bootstrapping the MOSFET drives using a doubler. This circuit provides
an equal voltage drive of VCC - 3VD + VIN to both the high-side and low-side MOSFET drives. This method should
only be used in circuits that use 3.3V for both VCC and VIN. Even with VIN = VCC = 3.0V (10% lower tolerance on
3.3V) and VD = 0.5V both high-side and low-side gates will have at least 4.5V of drive. The power dissipation of
the gate drive circuitry is directly proportional to gate drive voltage, hence the thermal limits of the LM10520 IC
will quickly be reached if this circuit is used with VCC or VIN voltages over 5V.
LM10520
BOOT D3
D2
HG
D1
VCC
VIN
+
VO
LG
+
Figure 25. Charge Pump with Added Gate Drive
All the gate drive circuits shown in the above figures typically use 100 nF ceramic capacitors in the bootstrap
locations.
POWER GOOD SIGNAL
The open drain output on the Power Good pin needs a pull-up resistor to a low voltage source. The pull-up
resistor should be chosen so that the current going into the Power Good pin is less than 1mA. A 100 kΩ resistor
is recommended for most applications.
24
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
The Power Good signal is an OR-gated flag which takes into account both output over-voltage and under-voltage
conditions. If the feedback pin (FB) voltage is 18% above its nominal value (118% x VFB = 0.708V) or falls 28%
below that value (72% x VFB = 0.42V) the Power Good flag goes low. The Power Good flag can be used to signal
other circuits that the output voltage has fallen out of regulation, however the switching of the LM10520
continues regardless of the state of the Power Good signal. The Power Good flag will return to logic high
whenever the feedback pin voltage is between 72% and 118% of 0.6V.
CURRENT LIMIT
Current limit is realized by sensing the voltage across the low-side MOSFET while it is on. The RDSON of the
MOSFET is a known value; hence the current through the MOSFET can be determined as:
VDS = IOUT x RDSON
(6)
The current through the low-side MOSFET while it is on is also the falling portion of the inductor current. The
current limit threshold is determined by an external resistor, RCS, connected between the switching node and the
ISEN pin. A constant current (ISEN-TH) of 40 µA typical is forced through RCS, causing a fixed voltage drop. This
fixed voltage is compared against VDS and if the latter is higher, the current limit of the chip has been reached.
To obtain a more accurate value for RCS you must consider the operating values of RDSON and ISEN-TH at their
operating temperatures in your application and the effect of slight parameter differences from part to part. RCS
can be found by using the following equation using the RDSON value of the low side MOSFET at it's expected hot
temperature and the absolute minimum value expected over the full temperature range for the for the ISEN-TH
which is 25 µA:
RCS = RDSON-HOT x ILIM / ISEN-TH
(7)
For example, a conservative 15A current limit in a 10A design with a RDSON-HOT of 10 mΩ would require a 6kΩ
resistor. The minimum value for RCS in any application is 1kΩ. Because current sensing is done across the lowside MOSFET, no minimum high-side on-time is necessary. The LM10520 enters current limit mode if the
inductor current exceeds the current limit threshold at the point where the high-side MOSFET turns off and the
low-side MOSFET turns on. (The point of peak inductor current, see Figure 26). Note that in normal operation
mode the high-side MOSFET always turns on at the beginning of a clock cycle. In current limit mode, by contrast,
the high-side MOSFET on-pulse is skipped. This causes inductor current to fall. Unlike a normal operation
switching cycle, however, in a current limit mode switching cycle the high-side MOSFET will turn on as soon as
inductor current has fallen to the current limit threshold. The LM10520 will continue to skip high-side MOSFET
pulses until the inductor current peak is below the current limit threshold, at which point the system resumes
normal operation.
Normal Operation
Current Limit
ILIM
IL
D
Figure 26. Current Limit Threshold
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
25
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
Unlike a high-side MOSFET current sensing scheme, which limits the peaks of inductor current, low-side current
sensing is only allowed to limit the current during the converter off-time, when inductor current is falling.
Therefore in a typical current limit plot the valleys are normally well defined, but the peaks are variable, according
to the duty cycle. The PWM error amplifier and comparator control the off-pulse of the high-side MOSFET, even
during current limit mode, meaning that peak inductor current can exceed the current limit threshold. Assuming
that the output inductor does not saturate, the maximum peak inductor current during current limit mode can be
calculated with the following equation:
ICL x RDS(ON)max
RLIM (Tj) =
ILIM-TH (Tj)
where
•
TSW is the inverse of switching frequency fSW
(8)
The 200 ns term represents the minimum off-time of the duty cycle, which ensures enough time for correct
operation of the current sensing circuitry.
In order to minimize the time period in which peak inductor current exceeds the current limit threshold, the IC
also discharges the soft-start capacitor through a fixed 90 µA sink. The output of the LM10520 internal error
amplifier is limited by the voltage on the soft-start capacitor. Hence, discharging the soft-start capacitor reduces
the maximum duty cycle D of the controller. During severe current limit this reduction in duty cycle will reduce the
output voltage if the current limit conditions last for an extended time. Output inductor current will be reduced in
turn to a flat level equal to the current limit threshold. The third benefit of the soft-start capacitor discharge is a
smooth, controlled ramp of output voltage when the current limit condition is cleared.
DESIGN CONSIDERATIONS
The following is a design procedure for all the components needed to create the Typical Application Circuit
shown on the front page. This design converts 3.3V (VIN) to 1.2V (VOUT) at a maximum load of 4A with an
efficiency of 89% and a switching frequency of 300 kHz. The same procedures can be followed to create many
other designs with varying input voltages, output voltages, and load currents.
Input Capacitor
The input capacitors in a Buck converter are subjected to high stress due to the input current trapezoidal
waveform. Input capacitors are selected for their ripple current capability and their ability to withstand the heat
generated since that ripple current passes through their ESR. Input rms ripple current is approximately:
IRMS_RIP = IOUT x
D(1 - D)
where
•
duty cycle D = VOUT/VIN
(9)
The power dissipated by each input capacitor is:
2
(IRMS_RIP) x ESR
PCAP =
n
2
where
•
•
n is the number of paralleled capacitors
ESR is the equivalent series resistance of each capacitor
(10)
The equation above indicates that power loss in each capacitor decreases rapidly as the number of input
capacitors increases. The worst-case ripple for a Buck converter occurs during full load and when the duty cycle
(D) is 0.5. For this 3.3V to 1.2V design the duty cycle is 0.364. For a 4A maximum load the ripple current is
1.92A.
26
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
Output Inductor
The output inductor forms the first half of the power stage in a Buck converter. It is responsible for smoothing the
square wave created by the switching action and for controlling the output current ripple (ΔIOUT). The inductance
is chosen by selecting between tradeoffs in efficiency and response time. The smaller the output inductor, the
more quickly the converter can respond to transients in the load current. However, as shown in the efficiency
calculations, a smaller inductor requires a higher switching frequency to maintain the same level of output current
ripple. An increase in frequency can mean increasing loss in the MOSFETs due to the charging and discharging
of the gates. Generally the switching frequency is chosen so that conduction loss outweighs switching loss. The
equation for output inductor selection is:
VIN - VOUT
L=
L=
xD
'IOUT x fSW
3.3V - 1.2V
1.2V
x
3.3V
0.4 x 4A x 300 kHz
L = 1.6 µH
(11)
Here we have plugged in the values for output current ripple, input voltage, output voltage, switching frequency,
and assumed a 40% peak-to-peak output current ripple. This yields an inductance of 1.6 µH. The output inductor
must be rated to handle the peak current (also equal to the peak switch current), which is (IOUT + (0.5 x ΔIOUT)) =
4.8A, for a 4A design.
The Coilcraft DO3316P-222P is 2.2 µH, is rated to 7.4A peak, and has a direct current resistance (DCR) of 12
mΩ. After selecting the Coilcraft DO3316P-222P for the output inductor, actual inductor current ripple should be
re-calculated with the selected inductance value, as this information is needed to select the output capacitor. Rearranging the equation used to select inductance yields the following:
VIN(MAX) - VO
'IOUT =
fSW x LACTUAL
xD
where
•
VIN(MAX) is assumed to be 10% above the steady state input voltage, or 3.6V at VIN = 3.3V
(12)
The re-calculated current ripple will then be 1.2A. This gives a peak inductor/switch current will be 4.6A.
Output Capacitor
The output capacitor forms the second half of the power stage of a Buck switching converter. It is used to control
the output voltage ripple (ΔVOUT) and to supply load current during fast load transients.
In this example the output current is 4A and the expected type of capacitor is an aluminum electrolytic, as with
the input capacitors. Other possibilities include ceramic, tantalum, and solid electrolyte capacitors, however the
ceramic type often do not have the large capacitance needed to supply current for load transients, and tantalums
tend to be more expensive than aluminum electrolytic. Aluminum capacitors tend to have very high capacitance
and fairly low ESR, meaning that the ESR zero, which affects system stability, will be much lower than the
switching frequency. The large capacitance means that at the switching frequency, the ESR is dominant, hence
the type and number of output capacitors is selected on the basis of ESR. One simple formula to find the
maximum ESR based on the desired output voltage ripple, ΔVOUT and the designed output current ripple, ΔIOUT,
is:
ESRMAX =
'VOUT
'IOUT
(13)
In this example, in order to maintain a 2% peak-to-peak output voltage ripple and a 40% peak-to-peak inductor
current ripple, the required maximum ESR is 20 mΩ. The Sanyo 4SP560M electrolytic capacitor will give an
equivalent ESR of 14 mΩ. The capacitance of 560 µF is enough to supply energy even to meet severe load
transient demands.
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
27
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
MOSFETs
Selection of the power MOSFETs is governed by a trade-off between cost, size, and efficiency. One method is to
determine the maximum cost that can be endured, and then select the most efficient device that fits that price.
Breaking down the losses in the high-side and low-side MOSFETs and then creating spreadsheets is one way to
determine relative efficiencies between different MOSFETs. Good correlation between the prediction and the
bench result is not ensured, however. Single-channel buck regulators that use a controller IC and discrete
MOSFETs tend to be most efficient for output currents of 2 to 10A.
Losses in the high-side MOSFET can be broken down into conduction loss, gate charging loss, and switching
loss. Conduction, or I2R loss, is approximately:
PC = D (IO2 x RDSON-HI x 1.3) (High-Side MOSFET)
PC = (1 - D) x (IO2 x RDSON-LO x 1.3) (Low-Side MOSFET)
(14)
(15)
In the above equations the factor 1.3 accounts for the increase in MOSFET RDSON due to heating. Alternatively,
the 1.3 can be ignored and the RDSON of the MOSFET estimated using the RDSON Vs. Temperature curves in the
MOSFET datasheets.
Gate charging loss results from the current driving the gate capacitance of the power MOSFETs, and is
approximated as:
PGC = n x (VDD) x QG x fSW
where
•
•
•
‘n’ is the number of MOSFETs (if multiple devices have been placed in parallel)
VDD is the driving voltage (see MOSFET GATE DRIVERS)
QGS is the gate charge of the MOSFET
(16)
If different types of MOSFETs are used, the ‘n’ term can be ignored and their gate charges simply summed to
form a cumulative QG. Gate charge loss differs from conduction and switching losses in that the actual
dissipation occurs in the LM10520, and not in the MOSFET itself.
Switching loss occurs during the brief transition period as the high-side MOSFET turns on and off, during which
both current and voltage are present in the channel of the MOSFET. It can be approximated as:
PSW = 0.5 x VIN x IO x (tr + tf) x fSW
where
•
tr and tf are the rise and fall times of the MOSFET
(17)
Switching loss occurs in the high-side MOSFET only.
For this example, the maximum drain-to-source voltage applied to either MOSFET is 3.6V. The maximum drive
voltage at the gate of the high-side MOSFET is 3.1V, and the maximum drive voltage for the low-side MOSFET
is 3.3V. Due to the low drive voltages in this example, a MOSFET that turns on fully with 3.1V of gate drive is
needed. For designs of 5A and under, dual MOSFETs in SO-8 provide a good trade-off between size, cost, and
efficiency.
Support Components
CIN2 - A small (0.1 to 1 µF) ceramic capacitor should be placed as close as possible to the drain of the high-side
MOSFET and source of the low-side MOSFET (dual MOSFETs make this easy). This capacitor should be X5R
type dielectric or better.
RCC, CCC- These are standard filter components designed to ensure smooth DC voltage for the chip supply. RCC
should be 1 to 10Ω. CCC should 1µF, X5R type or better.
CBOOT- Bootstrap capacitor, typically 100 nF.
RPULL-UP – This is a standard pull-up resistor for the open-drain power good signal (PWGD). The recommended
value is 100 kΩ connected to VCC. If this feature is not necessary, the resistor can be omitted.
D1 - A small Schottky diode should be used for the bootstrap. It allows for a minimum drop for both high and lowside drivers. The MBR0520 or BAT54 work well in most designs.
28
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
RCS - Resistor used to set the current limit. Since the design calls for a peak current magnitude (IOUT + (0.5 x
ΔIOUT)) of 4.8A, a safe setting would be 6A. (This is below the saturation current of the output inductor, which is
7A.) Following the equation from the Current Limit section, a 1.3 kΩ resistor should be used.
RFADJ - This resistor is used to set the switching frequency of the chip. The resistor value is approximated from
the Frequency vs. Frequency Adjust Resistor curve in the Typical Performance Characteristics. For 300 kHz
operation, a 100 kΩ resistor should be used.
CSS - The soft-start capacitor depends on the user requirements and is calculated based on the equation given in
the section titled START UP/SOFT-START. Therefore, for a 7 ms delay, a 12 nF capacitor is suitable.
Control Loop Compensation
The LM10520 uses voltage-mode (‘VM’) PWM control to correct changes in output voltage due to line and load
transients. VM requires careful small signal compensation of the control loop for achieving high bandwidth and
good phase margin.
The control loop is comprised of two parts. The first is the power stage, which consists of the duty cycle
modulator, output inductor, output capacitor, and load. The second part is the error amplifier, which for the
LM10520 is a 9MHz op-amp used in the classic inverting configuration. Figure 27 shows the regulator and
control loop components.
L
RL
+ C
O
VIN
RO
+
RC
+
VRAMP
RC2
CC2
RC1
RFB2
CC3
CC1
+
RFB1
+
-
VREF
Figure 27. Power Stage and Error Amp
One popular method for selecting the compensation components is to create Bode plots of gain and phase for
the power stage and error amplifier. Combined, they make the overall bandwidth and phase margin of the
regulator easy to see. Software tools such as Excel, MathCAD, and Matlab are useful for showing how changes
in compensation or the power stage affect system gain and phase.
The power stage modulator provides a DC gain ADC that is equal to the input voltage divided by the peak-to-peak
value of the PWM ramp. This ramp is 1.0Vpk-pk for the LM10520. The inductor and output capacitor create a
double pole at frequency fDP, and the capacitor ESR and capacitance create a single zero at frequency fESR. For
this example, with VIN = 3.3V, these quantities are:
ADC =
fDP =
VIN
VRAMP
1
2S
=
3.3
= 10.4 dB
1.0
RO + RL
LCO(RO + ESR)
(18)
= 4.5 kHz
(19)
1
= 20.3 kHz
fESR =
2SCOESR
(20)
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
29
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
In the equation for fDP, the variable RL is the power stage resistance, and represents the inductor DCR plus the
on resistance of the top power MOSFET. RO is the output voltage divided by output current. The power stage
transfer function GPS is given by the following equation, and Figure 28 shows Bode plots of the phase and gain in
this example.
VIN x RO
GPS =
VRAMP
x
sCORC + 1
2
as + bs + c
where
a = LCO(RO + RC)
b = L + CO(RORL + RORC + RCRL)
c = RO + RL
(21)
20
0
4
-30
PHASE (o)
GAIN (dB)
•
•
•
-12
-28
-60
-90
-120
-44
-60
100
1k
10k
100k
1M
-150
100
1k
FREQUENCY (Hz)
10k
100k
1M
FREQUENCY (Hz)
Figure 28. Power Stage Gain and Phase
The double pole at 4.5 kHz causes the phase to drop to approximately -130° at around 10 kHz. The ESR zero, at
20.3 kHz, provides a +90° boost that prevents the phase from dropping to -180°. If this loop were left
uncompensated, the bandwidth would be approximately 10 kHz and the phase margin 53°. In theory, the loop
would be stable, but would suffer from poor DC regulation (due to the low DC gain) and would be slow to
respond to load transients (due to the low bandwidth.) In practice, the loop could easily become unstable due to
tolerances in the output inductor, capacitor, or changes in output current, or input voltage. Therefore, the loop is
compensated using the error amplifier and a few passive components.
For this example, a Type III, or three-pole-two-zero approach gives optimal bandwidth and phase.
In most voltage mode compensation schemes, including Type III, a single pole is placed at the origin to boost DC
gain as high as possible. Two zeroes fZ1 and fZ2 are placed at the double pole frequency to cancel the double
pole phase lag. Then, a pole, fP1 is placed at the frequency of the ESR zero. A final pole fP2 is placed at one-half
of the switching frequency. The gain of the error amplifier transfer function is selected to give the best bandwidth
possible without violating the Nyquist stability criteria. In practice, a good crossover point is one-fifth of the
switching frequency, or 60 kHz for this example. The generic equation for the error amplifier transfer function is:
s
+1
2SfZ1
GEA = AEA x
s
s
+1
2SfP1
s
+1
2SfZ2
s
+1
2SfP2
(22)
In this equation the variable AEA is a ratio of the values of the capacitance and resistance of the compensation
components, arranged as shown in Figure 27. AEA is selected to provide the desired bandwidth. A starting value
of 80,000 for AEA should give a conservative bandwidth. Increasing the value will increase the bandwidth, but will
also decrease phase margin. Designs with 45-60° are usually best because they represent a good trade-off
between bandwidth and phase margin. In general, phase margin is lowest and gain highest (worst-case) for
maximum input voltage and minimum output current. One method to select AEA is to use an iterative process
beginning with these worst-case conditions.
30
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
1.
2.
3.
4.
SNVS638 – NOVEMBER 2010
Increase AEA
Check overall bandwidth and phase margin
Change VIN to minimum and recheck overall bandwidth and phase margin
Change IO to maximum and recheck overall bandwidth and phase margin
The process ends when the both bandwidth and the phase margin are sufficiently high. For this example input
voltage can vary from 3.0 to 3.6V and output current can vary from 0 to 4A, and after a few iterations a moderate
gain factor of 101 dB is used.
The error amplifier of the LM10520 has a unity-gain bandwidth of 9 MHz. In order to model the effect of this
limitation, the open-loop gain can be calculated as:
OPG =
2S x 9 MHz
s
(23)
The new error amplifier transfer function that takes into account unity-gain bandwidth is:
HEA =
GEA x OPG
1 + GEA + OPG
(24)
60
50
48
20
36
PHASE (o)
GAIN (dB)
The gain and phase of the error amplifier are shown in Figure 29.
24
-10
-40
-70
12
0
100
1k
10k
100k
1M
-100
100
FREQUENCY (Hz)
1k
10k
100k
1M
FREQUENCY (Hz)
Figure 29. Error Amp. Gain and Phase
In VM regulators, the top feedback resistor RFB2 forms a part of the compensation. Setting RFB2 to 10 kΩ±1%,
usually gives values for the other compensation resistors and capacitors that fall within a reasonable range.
(Capacitances > 1pF, resistances < 1MΩ) CC1, CC2, CC3, RC1, and RC2 are selected to provide the poles and
zeroes at the desired frequencies, using the following equations:
fZ1
CC1 =
CC2 =
AEA x 10,000 x fP2
1
AEA x 10,000
= 27 pF
(25)
- CC1 = 882 pF
(26)
1
1
1
x
CC3 =
= 2.73 nF
fZ2 fP1
2S x 10,000
(27)
1
= 39.8 k:
RC1 =
2S x CC2 x fZ1
RC2 =
(28)
1
= 2.55 k:
2S x CC3 x fP1
(29)
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
31
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
In practice, a good tradeoff between phase margin and bandwidth can be obtained by selecting the closest ±10%
capacitor values above what are suggested for CC1 and CC2, the closest ±10% capacitor value below the
suggestion for CC3, and the closest ±1% resistor values below the suggestions for RC1, RC2. Note that if the
suggested value for RC2 is less than 100Ω, it should be replaced by a short circuit. Following this guideline, the
compensation components will be:
CC1 = 27 pF±10%, CC2 = 820 pF±10%
CC3 = 2.7 nF±10%, RC1 = 39.2 kΩ±1%
RC2 = 2.55 kΩ±1%
The transfer function of the compensation block can be derived by considering the compensation components as
impedance blocks ZF and ZI around an inverting op-amp:
ZF
GEA-ACTUAL =
ZI
(30)
1
1
x 10,000 +
sCC2
sCC1
ZF =
10,000 +
1
1
+
sCC1 sCC2
RC1 RC2 +
(31)
1
sCC3
Z1 =
RC1 + RC2 +
1
sCC3
(32)
As with the generic equation, GEA-ACTUAL must be modified to take into account the limited bandwidth of the error
amplifier. The result is:
GEA-ACTUAL x OPG
HEA =
1 + GEA-ACTUAL + OPG
(33)
The total control loop transfer function H is equal to the power stage transfer function multiplied by the error
amplifier transfer function.
H = GPS x HEA
(34)
60
-60
40
-84
20
PHASE (o)
GAIN (dB)
The bandwidth and phase margin can be read graphically from Bode plots of HEA as shown in Figure 30.
0
-20
-40
100
-108
-132
-156
1k
10k
100k
1M
-180
100
1k
FREQUENCY (Hz)
10k
100k
1M
FREQUENCY (Hz)
Figure 30. Overall Loop Gain and Phase
The bandwidth of this example circuit is 59 kHz, with a phase margin of 60°.
32
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
LM10520
www.ti.com
SNVS638 – NOVEMBER 2010
EFFICIENCY CALCULATIONS
The following is a sample calculation.
A reasonable estimation of the efficiency of a switching buck controller can be obtained by adding together the
Output Power (POUT) loss and the Total Power (PTOTAL) loss:
POUT
K=
POUT + PTOTAL
x 100%
(35)
The Output Power (POUT) for the Typical Application Circuit design is (1.2V x 4A) = 4.8W. The Total Power
(PTOTAL), with an efficiency calculation to complement the design, is shown below.
The majority of the power losses are due to the low side and high side MOSFET’s losses. The losses in any
MOSFET are group of switching (PSW) and conduction losses (PCND).
PFET = PSW + PCND = 61.38 mW + 270.42 mW
PFET = 331.8 mW
(36)
FET Switching Loss (PSW)
PSW =
PSW =
PSW =
PSW =
PSW(ON) + PSW(OFF)
0.5 x VIN x IOUT x (tr + tf) x fSW
0.5 x 3.3V x 4A x 300 kHz x 31 ns
61.38 mW
(37)
The FDS6898A has a typical turn-on rise time tr and turn-off fall time tf of 15 ns and 16 ns, respectively. The
switching losses for this type of dual N-Channel MOSFETs are 0.061W.
FET Conduction Loss (PCND)
PCND = PCND1 + PCND2
PCND1 = I2OUT x RDS(ON) x k x D
PCND2 = I2OUT x RDS(ON) x k x (1-D)
RDS(ON) = 13 mΩ and the factor is a constant value (k = 1.3) to account for the increasing RDS(ON) of a FET due to
heating.
PCND1 = (4A)2 x 13 mΩ x 1.3 x 0.364
PCND2 = (4A)2 x 13 mΩ x 1.3 x (1 - 0.364)
PCND = 98.42 mW + 172 mW = 270.42 mW
(38)
There are few additional losses that are taken into account:
IC Operating Loss (PIC)
PIC = IQ_VCC x VCC
where
• IQ-VCC is the typical operating VCC current
PIC= 1.7 mA x 3.3V = 5.61 mW
(39)
FET Gate Charging Loss (PGATE)
PGATE = n x VCC x QGS x fSW
PGATE = 2 x 3.3V x 3 nC x 300 kHz
PGATE = 5.94 mW
(40)
The value n is the total number of FETs used and QGS is the typical gate-source charge value, which is 3 nC. For
the FDS6898A the gate charging loss is 5.94 mW.
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
33
LM10520
SNVS638 – NOVEMBER 2010
www.ti.com
Input Capacitor Loss (PCAP)
2
(IRMS_RIP) x ESR
PCAP =
n
2
where
•
IRMS_RIP = IOUT x
D(1 - D)
(41)
Here n is the number of paralleled capacitors, ESR is the equivalent series resistance of each, and PCAP is the
dissipation in each. So for example if we use only one input capacitor of 24 mΩ.
2
PCAP =
(1.924A) x 24 m:
1
2
PCAP = 88.8 mW
(42)
Output Inductor Loss (PIND)
PIND = I2OUT x DCR
where
•
DCR is the DC resistance
Therefore, for example
PIND = (4A)2 x 11 mΩ
PIND = 176 mW
(43)
Total System Efficiency
PTOTAL = PFET + PIC + PGATE + PCAP + PIND
(44)
POUT
K=
POUT + PTOTAL
x 100%
(45)
(RFB1 + RFB2)
VOUT = VFB x
RFB1
34
(46)
Submit Documentation Feedback
Copyright © 2010, Texas Instruments Incorporated
Product Folder Links: LM10520
MECHANICAL DATA
PWP0028B
MXA28B (Rev A)
www.ti.com
PACKAGE OPTION ADDENDUM
www.ti.com
30-May-2018
PACKAGING INFORMATION
Orderable Device
Status
(1)
LM10520MH/NOPB
NRND
Package Type Package Pins Package
Drawing
Qty
HTSSOP
PWP
28
48
Eco Plan
Lead/Ball Finish
MSL Peak Temp
(2)
(6)
(3)
Green (RoHS
& no Sb/Br)
Call TI
Level-3-260C-168 HR
Op Temp (°C)
Device Marking
(4/5)
-40 to 125
LM10520
(1)
The marketing status values are defined as follows:
ACTIVE: Product device recommended for new designs.
LIFEBUY: TI has announced that the device will be discontinued, and a lifetime-buy period is in effect.
NRND: Not recommended for new designs. Device is in production to support existing customers, but TI does not recommend using this part in a new design.
PREVIEW: Device has been announced but is not in production. Samples may or may not be available.
OBSOLETE: TI has discontinued the production of the device.
(2)
RoHS: TI defines "RoHS" to mean semiconductor products that are compliant with the current EU RoHS requirements for all 10 RoHS substances, including the requirement that RoHS substance
do not exceed 0.1% by weight in homogeneous materials. Where designed to be soldered at high temperatures, "RoHS" products are suitable for use in specified lead-free processes. TI may
reference these types of products as "Pb-Free".
RoHS Exempt: TI defines "RoHS Exempt" to mean products that contain lead but are compliant with EU RoHS pursuant to a specific EU RoHS exemption.
Green: TI defines "Green" to mean the content of Chlorine (Cl) and Bromine (Br) based flame retardants meet JS709B low halogen requirements of