LM3431
www.ti.com
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
3-Channel Constant Current LED Driver with Integrated Boost Controller
Check for Samples: LM3431
FEATURES
DESCRIPTION
•
The LM3431 is a 3-channel linear current controller
combined with a boost switching controller ideal for
driving LED backlight panels in space critical
applications. The LM3431 drives 3 external NPN
transistors or MOSFETs to deliver high accuracy
constant current to 3 LED strings. Output current is
adjustable to drive strings in excess of 200 mA. The
LM3431 can be expanded to drive as many as 6 LED
strings.
1
2
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
LM3431Q/LM3431AQ are Automotive Grade
Products that are AEC-Q100 Grade 1 Qualified
(–40°C to 125°C operating junction
temperature)
3-Channel Programmable LED Current
High Accuracy Linear Current Regulation
Analog and Digital PWM Dimming Control
Up to 25kHz Dimming Frequency
>100:1 Contrast Ratio
Integrated Boost Controller
5V-36V Input Voltage Range
Adjustable Switching Frequency up to 1MHz
LED Short and Open Protection
Selectable Fault Shutdown or Automatic
Restart
Programmable Fault Delay
Programmable Cycle by Cycle Current Limit
Output Over Voltage Protection
No Audible Noise
Enable Pin
LED Over-Temperature Shutdown Input
Thermal Shutdown
TSSOP-28 Exposed Pad Package
The boost controller drives an external NFET switch
for step-up regulation from input voltages between 5V
and 36V. The LM3431 features LED cathode
feedback to minimize regulator headroom and
optimize efficiency.
A DIM input pin controls LED brightness from analog
or digital control signals. Dimming frequencies up to
25 kHz are possible with a contrast ratio of 100:1.
Contrast ratios greater than 1000:1 are possible at
lower dimming frequencies.
The LM3431 eliminates audible noise problems by
maintaining constant output voltage regulation during
LED dimming. Additional features include LED short
and open protection, fault delay/error flag, cycle by
cycle current limit, and thermal shutdown for both the
IC and LED array. The enhanced LM3431A features
reduced offset voltage for higher accuracy LED
current.
APPLICATIONS
•
•
Automotive Infotainment Displays
Small to Medium Format Displays
TYPICAL APPLICATION CIRCUIT
Vin: 5V to 36V
VCC
PWM Dim
input
VIN
LG
EN
CS
MODE/F
ILIM
PGND
DIM
THM
LEDOFF
REF
REFIN
COMP
FF
VCC
RT
SS/SH
DLY
SGND
+
LM3431
AFB
SC
CFB
NDRV1
Ch.2/
Ch.3
SNS1
NDRV2
SNS2
NDRV3
SNS3
Ch.2
Ch.3
1
2
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.
All 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 © 2007–2013, Texas Instruments Incorporated
LM3431
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
www.ti.com
CONNECTION DIAGRAM
28 EN
VIN 1
PGND 2
VCC 3
27 DIM
26 THM
25 NDRV1
LG 4
24 SNS1
CS 5
ILIM 6
MODE/F 7
23 NDRV2
22 SNS2
FF 8
21 NDRV3
20 SNS3
RT 9
REF 10
19 LEDOFF
REFIN 11
18 SC
17 CFB
COMP 12
16 DLY
SGND 13
AFB 14
15 SS/SH
Exposed Pad
Connect to SGND
Figure 1. 28 Lead Plastic Exposed Pad TSSOP
Top View
See Package Number PWP0028A
PIN DESCRIPTIONS
2
Pin No.
Pin Name
1
VIN
Description
2
PGND
3
VCC
4
LG
Boost controller gate drive output. Connect to the NFET gate.
5
CS
Boost controller current sense pin. Connect to the top side of the boost current sense resistor.
6
ILIM
Boost controller current limit adjust pin. Connect a resistor from this pin to the Boost current sense resistor to set
the current limit threshold.
7
MODE/F
Dimming mode selection pin. Pull high for digital PWM control. Or connect to a capacitor to GND to set the
internal dimming frequency.
8
FF
Feedforward pin. Connect to a resistor to ground to control the output voltage over/undershoot during PWM
dimming.
9
RT
Frequency adjust pin. Connect a resistor from this pin to ground to set the operating frequency of the boost
controller.
Power supply input.
Power ground pin. Connect to ground.
Internal reference voltage output. Bypass to PGND with a minimum 4.7 µF capacitor.
10
REF
11
REFIN
This pin sets the LED current feedback voltage. Connect to a resistor divider from the REF pin.
12
COMP
Output of the error amplifier. Connect to the compensation network.
13
SGND
Signal ground pin. Connect to ground.
14
AFB
15
SS/SH
16
DLY
Fault delay pin. Connect a capacitor from this pin to ground to set the delay time for shutdown.
17
CFB
Cathode feedback pin. The boost controller voltage feedback. Connect through a diode to the bottom cathode of
each LED string.
18
SC
19
LEDOFF
20
SNS3
21
NDRV3
22
SNS2
23
NDRV2
Reference voltage. Use this pin to provide the REFIN voltage.
Anode feedback pin. The boost controller voltage feedback during LED off time. Connect this pin to a resistor
divider from the output voltage.
Soft-start and sample-hold pin. Connect a capacitor from this pin to ground to set the soft-start time.
LED short circuit detection pin. Connect through a diode to the bottom cathode of each string.
A dual function pin. The LEDOFF signal controls external drivers during PWM dimming. Or connect to ground to
enable automatic fault restart.
Current feedback for channel 3. Connect to the top of the channel 3 current sense resistor.
Base drive for the channel 3 current regulator. Connect to the NPN base or NFET gate.
Current feedback for channel 2.
Base drive for the channel 2 current regulator.
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
LM3431
www.ti.com
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
PIN DESCRIPTIONS (continued)
Pin No.
Pin Name
24
SNS1
25
NDRV1
26
THM
LED thermal monitor input pin. When pulled below 1.2V, device enters standby mode.
27
DIM
PWM dimming input pin. Accepts a digital PWM or analog voltage level input to control LED current duty cycle.
28
EN
Enable pin. Connect to VIN through a resistor divider to set an external UVLO threshold. Pull low to shutdown.
EP
Description
Current feedback for channel 1.
Base drive for the channel 1 current regulator.
Exposed pad. Connect to SGND.
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)
If Military/Aerospace specified devices are required, contact the Texas Instruments Semiconductor Sales Office/
Distributors for availability and specifications.
VALUE / UNIT
Voltages from the indicated pins to SGND:
VIN
–0.3V to 37V
EN
–0.3V to 10V
DIM
–0.3V to 7V
MODE/F
–0.3V to 7V
REFIN
–0.3V to 7V
THM
–0.3V to 7V
DLY
–0.3V to 7V
SNSx
–0.3V to 7V
NDRVx
–0.3V to 7V
CFB
–0.3V to 7V
SC
-0.3V to 40V
AFB
–0.3V to 7V
CS
–0.3V to 7V
VCC
–0.3V to 7V
Storage Temperature
–65°C to +150°C
Soldering Dwell Time, Temperature
Infrared
20sec, 240°C
Vapor Phase
75sec, 219°C
ESD Rating Human Body Model (2)
(1)
(2)
2 kV
Absolute Maximum Ratings are limits beyond which damage to the device may occur. Operating Ratings are conditions under which
operation of the device is intended to be functional. For ensured specifications and test conditions, see the ELECTRICAL
CHARACTERISTICS.
The human body model is a 100pF capacitor discharged through a 1.5 kΩ resistor into each pin.
RECOMMENDED OPERATING CONDITIONS
(1)
VALUE / UNIT
VIN
4.5V to 36V
Junction Temperature Range
(2)
32°C/W
, TSSOP-28
3.1W
Thermal Resistance (θJA)
Power Dissipation
(1)
(2)
(3)
-40°C to +125°C
, TSSOP-28 (0.5W)
(3)
Absolute Maximum Ratings are limits beyond which damage to the device may occur. Operating Ratings are conditions under which
operation of the device is intended to be functional. For specifications and test conditions, see the ELECTRICAL CHARACTERISTICS.
The Thermal Resistance specifications are based on a JEDEC standard 4-layer pcb. θJA will vary with board size and copper area.
The maximum allowable power dissipation is a function of the maximum junction temperature, TJ_MAX, the junction-to-ambient thermal
resistance, θJA, and the ambient temperature, TA. The maximum allowable power dissipation at any ambient temperature is calculated
using: PD_MAX = (TJ_MAX – TA)/θJA. The maximum power dissipation is determined using TA = 25°C, and TJ_MAX = 125°C.
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
3
LM3431
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
www.ti.com
ELECTRICAL CHARACTERISTICS
Specifications in standard type are for TJ = 25°C only, and limits in boldface type apply over the junction temperature (TJ)
range of -40°C to +125°C. Unless otherwise stated, VIN = 12V. Minimum and Maximum limits are specified through test,
design, or statistical correlation. Typical values represent the most likely parametric norm at TJ = 25°C, and are provided for
reference purposes only. (1)
Parameter
Test Conditions
Min
Typ
Max
Units
4.85
mA
SYSTEM
(2)
IQ
Operating VIN Current
DIM = 5V
4.0
IQ_SB
Standby mode VIN current
EN = 1V
3.7
IQ_SD
Shutdown mode VIN Current
EN = 0V, Vin = 36V
VCC
VCC voltage
Iload = 25 mA, Vin = 5.5 to 36V
VCCILIM
VCC current limit
UVLO
UVLO threshold
4.80
mA
15
23
5
5.24
72
VIN rising, measured at VCC
4.36
hysteresis
Enable pin Standby threshold
EN rising
VEN
Enable pin On threshold
EN rising
4.50
hysteresis
V
V
0.75
1.185
V
mA
0.28
VEN_ST
µA
V
1.230
1.275
V
115
165
mV
2.5
2.55
V
80
LINEAR CURRENT CONTROLLER
VREF
Reference Voltage
IREF < 300 µA
2.45
IREFIN
REFIN input bias current
REFIN = 300 mV
14
ΔVREF / ΔVIN
Line regulation
5.5V < VIN < 36V
0.000
1
VNDRV
NDRVx drive voltage capability
INDRVx = 5 mA
INDRV_SK
NDRVx drive sink current
NDRVX = 0.9V
4
6
8
mA
INDRV_SC
NDRVx drive source current
NDRVX = 0.9V
10
15
20
mA
ISNS
SNSx input bias current
SNSx = 300 mV
20
30
µA
VOS
SNSx amp offset voltage
REFIN = 300 mV (LM3431)
-5
+5
mV
VOS
SNSx amp offset voltage
REFIN = 300 mV (LM3431A)
-3
+3
mV
VOS_DELTA
Ch. To Ch. offset voltage mismatch
(3)
REFIN = 300 mV, 25°C (LM3431)
5.5
mV
VOS_DELTA
Ch. To Ch. offset voltage mismatch
(3)
REFIN = 300 mV, -40°C to +125°C
(LM3431)
6
mV
VOS_DELTA
Ch. To Ch. offset voltage mismatch
(3)
REFIN = 300 mV, 25°C (LM3431A)
3.5
mV
VOS_DELTA
Ch. To Ch. offset voltage mismatch
(3)
REFIN = 300 mV, -40°C to +125°C
(LM3431A)
4
mV
bw
SNSx amp bandwidth
At unity gain
2
MHz
VLEDOFF
LEDOFF voltage
DIM low
5
V
VDIM
DIM threshold
MODE/F > 4V
3.7
1.9
hysteresis
(4)
nA
%/V
V
2.3
V
0.8
V
0.4
µs
TDIM
Minimum internal DIM pulse width
DIMDLY_R
DIM to NDRV delay time
DIM rising
100
ns
DIMDLY_F
DIM to NDRV delay time
DIM falling
90
ns
THMODE/F
MODE/F threshold
For Digital Dimming control
3.8
V
IMODE/F
MODE/F source/sink current
40
uA
VMODE_L
MODE/F minimum voltage
Analog dimming mode
0.37
V
VMODE_H
MODE/F peak voltage
Analog dimming mode
2.5
V
(1)
(2)
(3)
(4)
4
All room temperature limits are 100% production tested. All limits at temperature extremes are specified through correlation using
standard Statistical Quality Control (SQC) methods. All limits are used to calculate Average Outgoing Quality Level (AOQL).
IQ specifies the current into the VIN pin and applies to non-switching operation.
VOS_DELTA specifies the maximum absolute difference between the offset of any pair of SNS amplifiers.
The minimum DIM pulse width is an internal signal. Any pulse width may be applied to the DIM pin or generated via analog dimming
mode. A pulse width less than 0.4 µs will be internally extended to 0.4 µs.
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
LM3431
www.ti.com
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
ELECTRICAL CHARACTERISTICS (continued)
Specifications in standard type are for TJ = 25°C only, and limits in boldface type apply over the junction temperature (TJ)
range of -40°C to +125°C. Unless otherwise stated, VIN = 12V. Minimum and Maximum limits are specified through test,
design, or statistical correlation. Typical values represent the most likely parametric norm at TJ = 25°C, and are provided for
reference purposes only. (1)
Parameter
Test Conditions
Min
Typ
Max
Units
PROTECTION
VSC_SHORT
SC high threshold
LED short circuit fault, SC rising
5.7
6
6.2
V
VSC_OPEN
SC open clamp voltage
LED open circuit fault, SC rising
3.16
3.50
3.87
V
IDLY_SC
DLY source current
DLY = 1.0V
57
73
µA
IDLY_SK
DLY sink current
DLY = 1.0V
VDLY
DLY threshold voltage
DLY rising
VDLY_reset
DLY reset threshold voltage
DLY falling
350
TDLY_BLK
DLY blank time
DIM rising
1.6
VTHM
THM threshold
ITHM
THM hysteresis current
THM = 1V
IILIM
ILIM max source current
COMP = 2.0V
VAFB_max
AFB overvoltage threshold
VAFB_UVP
AFB undervoltage threshold
TSD
Thermal shutdown threshold
39
1.8
2.40
1.19
2.8
1.23
µA
3.16
mV
µs
1.27
9.6
AFB falling
V
V
µA
31
40
46
µA
1.87
2.0
2.22
V
0.73
0.85
0.98
V
160
°C
BOOST CONTROLLER
VCFB
CFB voltage
DIM high
1.60
1.71
1.82
ICFB
CFB source current
DIM high
35
CFBTC
CFB temperature coefficient
50
65
ΔVCFB / ΔVIN
CFB Line regulation
5.5V < VIN < 36V
ISS/SH
SS/SH source current
At EN going high
VSS_END
SS/SH voltage
At end of soft-start cycle
VRT
RT voltage
RRT = 34.8 kΩ
FSW
Switching Frequency
RRT = 34.8 kΩ
651
Minimum Switching Frequency
RRT = 130 kΩ
180
200
220
Maximum Switching Frequency
RRT = 22.6 kΩ
900
1000
1100
170
230
-2.6
Ton_min
Minimum on time
DMAX
Maximum duty cycle
ILIMgm
ILIM amplifier transconductance
COMP to ILIM gain
Vslope
Slope compensation
ICOMP_SC
COMP source current
ICOMP_SK
EAgm
V
µA
mV/°C
0.001
%/V
13
19
24
µA
1.80
1.85
1.90
V
749
kHz
1.22
80
700
V
ns
85
%
85
umho
Peak voltage per cycle
75
mV
VCOMP = 1.2V, AFB = 0.5V
155
µA
COMP sink current
VCOMP =1.2V, AFB = 1.5V
150
µA
Error amplifier transconductance
CFB to COMP gain, DIM high
230
umho
RLG
Gate Drive On Resistance
Source Current = 200 mA, VIN = 5.5V
6.4
Ω
Sink Current = 200 mA
2.2
Ω
ILG
Driver Output Current
Source, LG = 2.5V, VIN = 5.5V
0.35
A
Sink, LG = 2.5V
0.70
A
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
5
LM3431
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
www.ti.com
TYPICAL PERFORMANCE CHARACTERISTICS
Unless otherwise specified the following conditions apply: VIN = 12V, TJ = 25°C.
VREF vs. Temperature
CFB Voltage vs. Temperature
2.53
2.50
2.52
2.25
2.51
VCFB (V)
VREF (V)
2.00
2.50
2.49
1.75
1.50
2.48
1.25
2.47
1.00
-40 -20
2.46
-40 -20
0
20
40
60
80 100 120 140
0
20 40
60
80 100 120 140
TEMPERATURE (°C)
TEMPERATURE (°C)
Figure 2.
Figure 3.
SNS 1, 2, 3 VOS vs Temperature (LM3431 or LM3431A)
Delta VOS Max vs Temperature (LM3431 or LM3431A)
1.5
3.0
1.0
DELTA VOS (mV)
2.5
VOS (mV)
0.5
0
-0.5
1.5
1.0
0.5
-1.0
-1.5
-40 -20
2.0
0
20 40
60
0
-40 -20
80 100 120 140
0
20
40
60
80 100 120 140
TEMPERATURE (°C)
TEMPERATURE (°C)
Figure 4.
Figure 5.
IQ_SB vs Temperature
IQ_SD vs Temperature
16.0
3.80
15.5
3.70
IQ_SD (PA)
IQ_SB (mA)
15.0
3.60
3.50
14.5
14.0
3.40
3.30
-40 -20
6
13.5
0
20 40
60
80 100 120 140
13.0
-40 -20
0
20 40
60
80 100 120 140
TEMPERATURE (°C)
TEMPERATURE (°C)
Figure 6.
Figure 7.
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
LM3431
www.ti.com
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
TYPICAL PERFORMANCE CHARACTERISTICS (continued)
Unless otherwise specified the following conditions apply: VIN = 12V, TJ = 25°C.
Normalized Switching Frequency vs. Temperature (700 kHz)
100
1.02
100%
dimming duty
90
EFFICIENCY (%)
NORMALIZED FREQUENCY
1.04
Efficiency vs. Input Voltage LED Current = 140 mA x 3, LED
Vf = 25V
1.00
0.98
0.96
50%
80
10%
70
60
0.94
-40 -20
50
0
20 40
60
80 100 120 140
TEMPERATURE (°C)
5
9
13
17
21
INPUT VOLTAGE (V)
Figure 8.
Figure 9.
Line Transient Response
Dimming Transient Response
Vout
100 mV/Div
ILED
5 mA/Div
Vout
200 mV/Div
Vsw
20V/Div
VIN
5V/Div
ILED
200 mA/Div
400 Ps/DIV
200 Ps/DIV
Figure 10.
Figure 11.
LED Ripple Current
NDRV Waveforms
ILED
5 mA/Div
ILED
50 mA/Div
Vcathode
100 mV/Div
NDRV (FET)
2V/Div
REFIN
10 mV/Div
NDRV (NPN)
1V/Div
DIM
5V/Div
400 ns/DIV
1 Ps/DIV
Figure 12.
Figure 13.
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
7
LM3431
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
www.ti.com
BLOCK DIAGRAM
SC
DLY
60 uA
LED on
open
CFB
override
short
+
FF
THM
FF step scaling
VIN
Linear Reg
vin
ref
LED
on
ovp/uvp
CFB fault
2 uA
3.8V
LED on
+
blank
time
6V
ilimit
ff
+
-
+
-
CS
vcc
ramp
fault T1, T2
2.8V
+
-
VCC
R Q
LG
S
+
fault T3
uvlo
I limit
and COMP
clamp
latch off/restart
logic
set_clk
PGND
+
ILIM
Restart Mode
sd
ff
TSD
EN
+
-
COMP
EA
uvp
ramp
RT
ovp
A
+
-
+
-
CFB
AFB sense
set_clk
B
AFB
4V
A
+
MODE/F
B
LED on = A
+
1.7V
SS/SH
fault
CFB check
ss
ref
DIM
SS/SH logic
Restart Mode
CFB fault
+
-
+
-
+
-
+
-
SGND
LED on
LEDOFF
8
REF
REFIN
SNS3
NDRV3
SNS2
NDRV2
Submit Documentation Feedback
SNS1
NDRV1
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
LM3431
www.ti.com
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
External LED Array
L1
Vin
D1
+ C2
C1
VCC
R2
VIN
LG
EN
CS
R8
R7
DIM
LM3431
MODE/F
Rth
GND
PGND
THM
REF
C9
AFB
Rhys
THM
VCC
R17
REFIN
C4
CFB
FF
C7
SGND
Q2
SNS2
R15
Q3
NDRV3
DLY
C6
+
-
NDRV2
RT
R6
REFIN
Op1
SNS1
SS/SH
Rff
VC4
D6-9
NDRV1
VCC
C3
VC2
SC
COMP
C13
VC3
VC1
D2-5
R9
EP
THM
R18
R3
ILIM
LEDOFF
DIM
C8
R4
EN
R16
RMODE
External
Thermistor
R19
Q1
R1
C5
VA
SNS3
R10
R11
R12
LEDOFF
Q6
Q4
R13
R14
Rrestart
VCC
C15
Figure 14. Typical Application Schematic
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
9
LM3431
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
www.ti.com
OPERATION DESCRIPTION
The LM3431 combines a boost controller and 3 constant current regulator controllers in one device. To simplify
the description, these two blocks will be described separately as Boost Controller and LED Current Regulator. All
descriptions and component numbers refer to the Figure 14 schematic. The LED bottom cathode nodes (VC1 –
VC4) are referred to simply as the cathode.
BOOST CONTROLLER
The LM3431 is a current-mode, PWM boost controller. Although the LM3431 may be operated in either
continuous or discontinuous conduction mode, the following guidelines are designed for continuous conduction
operation. This mode of operation gives lower output ripple and better LED current regulation.
In continuous conduction mode (when the inductor current never reaches zero), the boost regulator operates in
two cycles. In the first cycle of operation, the NFET is turned on and current ramps up and is storing energy in
the inductor. During this cycle, diode D1 is reverse biased and load current is supplied by the output capacitors C8 and C9 in Figure 14.
In the second cycle, the NFET is off and the diode is forward biased. Inductor current is transferred to the load
and output capacitor. The ratio of these two cycles determines the output voltage and is expressed as D or D’:
'¶ = 1-D =
VIN
VOUT
(1)
where D is the duty cycle of the switch.
D=1-
VIN
VOUT
(2)
Maximum duty cycle is limited to 85% typically.
As input voltage approaches the nominal output voltage, duty cycle and switch on time are reduced. When the
on time reaches minimum, pulse skipping will occur. This increases the output ripple voltage and can cause
regulator saturation and poor LED current regulation. If input voltage equals or exceeds the set output voltage,
switching will stop and the output voltage will become unregulated. This will force an increase in the LED
cathode voltage and NPN regulator power dissipation. Although this condition can be tolerated, it is not
recommended.
Therefore, input voltage should be restricted to keep on time above the minimum (see Switching Frequency
section) and at least 1V below the set output voltage.
ENABLE and UVLO
The EN pin is a dual function pin combining both enable and programmable undervoltage lockout (UVLO). The
shutdown threshold is 0.75V. When EN is pulled below this threshold, the LM3431 will shutdown and IQ will be
reduced to 15 µA typically. The typical EN pin UVLO threshold is 1.23V. When the EN voltage is above this
threshold, the LM3431 will begin softstart. Below the UVLO threshold, the LM3431 will remain in standby mode.
A resistor divider, shown as R1 and R2 in Figure 14, can be used to program the UVLO threshold at the EN pin.
This feature is used to shutdown the IC at an input voltage higher than the internal VCC UVLO threshold of 4.4V.
The EN UVLO should be set just below the minimum input voltage for the application.
The internal UVLO is monitored at the VCC pin. When VCC is below the threshold of 4.4V, the LM3431 is in
standby mode (See VCC section).
10
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
LM3431
www.ti.com
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
Soft-Start
The SS/SH pin is a dual function softstart and sample/hold pin. The SH function is described in sections below.
When the EN pin is pulled above the programmable UVLO threshold and VCC rises above the internal UVLO
threshold, the SS/SH pin begins sourcing current. This charges the SS cap (C6) and the SS pin voltage in turn
controls the output voltage ramp-up, sensed via the AFB pin. The softstart capacitor is calculated as shown
below, where 20 µA is the typical softstart source current:
tss x 19 PA
C6 =
1.85V
(3)
The LED current regulators are held off until softstart is completed. During softstart, current limit is active and the
CFB pin is monitored for a cathode short fault (See LED Protection section). When the SS/SH voltage reaches
1.85V, the current regulators are activated, LED current begins flowing, and output voltage control is transferred
to the CFB pin. Typical startup is shown below in Figure 15.
ILED
100 mA/Div
Vout
10V/Div
VSS/SH
1V/Div
VEN
5V/Div
4 ms/DIV
Figure 15. Typical Startup Waveforms (From power-on, DIM = high)
Output Voltage, OVP, and SH
The LM3431 boost controls the LED cathode voltage in order to drive the LED strings with sufficient headroom at
optimum efficiency. When the LED strings are on, voltage is regulated to 1.7V (typical) at the CFB pin, which is
one diode Vf above the LED cathode voltage. Therefore, when the LED strings are on, the output voltage (LED
anodes, shown as VA in Figure 14) will vary according to the Vf of the LED string, while the LED cathode voltage
will be regulated via the CFB pin.
The AFB pin is used to regulate the output voltage when the LED strings are off, which is during startup and
dimming off cycles. During LED-off times, the cathode voltage is not regulated.
AFB should set the initial output voltage to at least 1.0V (CFB voltage minus one diode drop) above the
maximum LED string forward voltage. This ensures that there is enough headroom to drive the LED strings at
startup and keeps the SS/SH voltage below its maximum. The AFB pin voltage at the end of softstart is 1.85V
typically, which determines the ratio of the feedback resistors according to the following equation:
R19 = R18 x =
VOUT(MAX) - 1.85V
1.85V
(4)
The AFB resistors also set the output over-voltage (OVP) threshold. The OVP threshold is monitored during both
LED on and LED off states and protects against any over voltage condition, including all LEDs open (See Open
LED section).The OVP threshold at Vout can be calculated as follows:
2.0 x (R19+ R18)
VOVP =
R18
(5)
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
11
LM3431
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
www.ti.com
Because OVP has a fixed 2.0V threshold sensed at AFB, a larger value for R19 will increase the OVP threshold
of the output voltage. During an open LED fault, the output voltage will increase by 2.6V typically (see LED
Protection section). Therefore, at least this much headroom above the nominal output voltage is required to avoid
a false OVP error. Note that because of the high output voltage setting at the end of softstart, a brief open LED
error may occur during the short time it takes for the cathode voltage to drop to its nominal level. Figure 15
shows a typical startup waveform, where both Vout and the SS/SH voltage reach their peak before the LED
current turns on. Once LED current starts, SS/SH and Vout drop to the nominal operating point.
While the LEDs are on, the AFB voltage is sampled to the SS/SH pin. During LED-off time, this SS/SH voltage is
used as the reference voltage to regulate the output. This allows the output voltage to remain stable between on
and off dimming cycles, even though there may be wide variation in the LED string forward voltage. The SS/SH
pin has a maximum voltage of 1.9V. Therefore, the AFB voltage when the LEDs are on must be below this limit
for proper regulation. This will be ensured by setting the AFB resistors as described above. During LED-off
cycles, there is minimal loading on the output, which forces the boost controller into pulse skipping mode. In this
mode, switching is stopped completely, or for multiple cycles until the AFB feedback voltage falls below the
SS/SH reference level.
Switching Frequency
The switching frequency can be set between 200 kHz and 1 MHz with a resistor from the RT pin to ground. The
frequency setting resistor (R6 in Figure 14) can be determined according to the following empirically derived
equation:
RT = 35403 x fSW-1.06
Where
•
•
fSW is in kHz
the RT result is in kohm.
(6)
1200
FREQUENCY (kHz)
1000
800
600
400
200
0
0
20
40
60
80
100 120 140 160
RT (k:)
Figure 16. Switching Frequency vs RT
For a given application, the maximum switching frequency is limited by the minimum on time. When the LM3431
reaches its minimum on-time, pulse skipping will occur and output ripple will increase. To avoid this, set the
operating frequency below the following maximum setting:
fSW(MAX) =
D
tON(MIN)
(7)
Inductor Selection
Figure 17 shows how the inductor current, IL, varies during a switching cycle.
12
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
LM3431
www.ti.com
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
Vout
200 mV/Div
IL
500 mA/Div
Vsw
10V/Div
400 ns/DIV
Figure 17. Inductor Current, SW Voltage, and VOUT
The important quantities in determining a proper inductance value are IL(AVE) (the average inductor current) and
ΔiL (the peak to peak inductor current ripple). If ΔiL is larger than 2 x IL(AVE), the inductor current will drop to zero
for a portion of the cycle and the converter will operate in discontinuous conduction mode. If ΔiL is smaller than 2
x IL, the inductor current will stay above zero and the converter will operate in continuous conduction mode.
To determine the minimum L, first calculate the IL(AVE) at both minimum and maximum input voltage:
IL(AVE) =
IOUT
'¶
Where
•
IOUT is the sum of all LED string currents at 100% dimming
(8)
IL(AVE) will be highest at the minimum input voltage. Then determine the minimum L based on ΔiL with the
following equation:
L(MIN) =
VIN(MAX) x D(MIN)
'iL x fSW
(9)
A good starting point is to set ΔiL to 150% of the minimum IL(AVE) and calculate using that value. The maximum
recommended ΔiL is 200% of IL(AVE) to maintain continuous current in normal operation. In general a smaller
inductor (higher ripple current) will give a better dimming response due to the higher dI/dt. This is shown
graphically below.
Vcathode
2V/Div
ILED
100 mA/Div
IL
500 mA/Div
1 Ps/DIV
Figure 18. Inductor Current During Dimming
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
13
LM3431
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
www.ti.com
The resulting peak to peak inductor current is:
VIN x D
'iL =
L x fSW
(10)
And the resulting peak inductor current is:
ILPEAK = IL(AVE) +
'iL
2
(11)
Peak inductor current will occur at minimum VIN.
The inductor must be rated to handle both the average current and peak current, which is the same as the peak
switch current. As switching frequency increases, less inductance is required. However, some minimum
inductance value is required to ensure stability at duty cycles greater than 50%. The minimum inductance
required for stability can be calculated as:
L(MIN) =
R3 x (VOUT - 2 x VIN(MIN))
fSW x 75 mV x 2
Where
•
R3 is the sense resistor determined in the next section.
(12)
Although the inductor must be large enough to meet both the stability and the ΔiL requirements, a value close to
minimum will typically give the best performance.
Current Sensing
Switch current is sensed via the sense resistor, R3, while the switch is on and the inductor is charging. The
sensed current is used to control switching and to monitor current limit. To optimize the control signal, a typical
sense voltage between 50mV and 200mV is recommended. The sense resistor can therefore be calculated by
the following equation:
50 mV
200 mV
d R3 d
IL(AVE_MIN)
IL(AVE_MAX)
(13)
Since IL(AVE) will vary with input voltage, R3 should be determined based on the full input voltage range, although
the resulting value may extend somewhat outside the recommended range.
Current Limit
Current limit occurs when the voltage across the sense resistor (measured at the CS pin) equals the current limit
threshold voltage. The current limit threshold is set by R4. This value can be calculated as follows:
R4 =
(IL(LIM) x R3) + (D x 75 mV)
40 PA
(14)
Where 40 µA is the typical ILIM source current in current limit and ILlim is the peak (not average) inductor current
which triggers current limit.
To avoid false triggering, current limit should be set safely above the peak inductor current level. However, the
current limit resistor also has some effect on the control loop as seen in the block diagram. For this reason, R4
should not be set much higher than necessary. When current limit is activated, the NFET will be turned off
immediately until the next cycle. Current limit will typically result in a drop in output and cathode voltage. This will
cause the COMP pin voltage to increase to maximum, which will trigger a fault and start the DLY pin source
current (see LED Protection section). The LM3431 will continue to operate in current limit with reduced on-time
until the DLY pin has reached its threshold. However, the current limit cannot reduce the on-time below the
minimum specification.
In a boost switcher, there is a direct current path between input and output. Therefore, although the LM3431 will
shutdown in a shorted output condition, there are no means to limit the current flowing from input to output.
Note that if the maximum duty cycle of 85% (typical) is reached, the LM3431 will behave as though current limit
has occurred.
14
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
LM3431
www.ti.com
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
VCC
The VCC pin is the output of the internal voltage regulator. It must be bypassed to PGND with a minimum 4.7 µF
ceramic capacitor. Although VCC is capable of supplying up to 72 mA, external loads will increase the power
dissipation and temperature rise within the LM3431. See the TSD section for more detail. Above 72 mA, the VCC
voltage will drop due to current limit. Since the UVLO threshold is monitored at this pin, UVLO may be enabled
by a VCC over current event.
For input voltages between 4.5V and 5.5V, connect VCC to VIN through a 4.7Ω resistor. This will hold VCC
above the UVLO threshold and allow operation at input voltages as low as 4.5V. It may also be necessary to add
additional VIN and VCC capacitance for low VIN operation.
Diode Selection
The average current through D1 is the average load current (total LED current), and the peak current through the
diode is the peak inductor current. Therefore, the diode should be rated to handle more than the peak inductor
current which was calculated earlier. The diode must also be capable of handling the peak reverse voltage,
which is equal to the output voltage (LED Anode voltage). To improve efficiency, a low Vf Schottky diode is
recommended. Diode power loss is calculated as:
PDIODE = Vf x IOUT
(15)
NFET Selection
The drive pin of the LM3431 boost switcher, LG, must be connected to the gate of an external NFET. The NFET
drain is connected to the inductor and the source is connected to the sense resistor. The LG pin will drive the
gate at 5V typically.
The critical parameters for selection of a MOSFET are:
1. Maximum drain current rating, ID(MAX)
2. Maximum drain to source voltage, VDS(MAX)
3. On-resistance, RDS(ON)
4. Total gate charge, Qg
In the on-state, the switch current is equal to the inductor current. Therefore, the maximum drain current, ID, must
be rated higher than the current limit setting. The average switch current (ID(AVE)) is given in the equation below:
ID(AVE) = IL(AVE) x D
(16)
The off-state voltage of the NFET is approximately equal to the output voltage plus the diode Vf. Therefore,
VDS(MAX) of the NFET must be rated higher than the maximum output voltage. The power losses in the NFET can
be separated into conduction losses and switching losses. The conduction loss, Pcond, is the I2R loss across the
NFET. The maximum conduction loss is given by:
PCOND = RDS(ON) x DMAX x IL(AVE)2
where
•
•
DMAX is the maximum duty cycle for the given application
RDS(ON) is the on resistance at high temperature
(17)
wThe switching losses can be roughly calculated by the following equation:
fSW x IL(AVE) x VOUT x (tON + tOFF)
PSW =
2
Where
•
tON and tOFF are the NFET turn-on and turn-off times.
(18)
Power is also consumed in the LM3431 in the form of gate charge losses, Pg. These losses can be calculated
using the formula:
Pg = fSW x Qg x VIN
where
•
Qg is the NFET total gate charge
(19)
Pg adds to the total power dissipation of the LM3431 (See TSD section).
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
15
LM3431
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
www.ti.com
Fast switching FETs can cause noise spikes at the SW node which may affect performance. To reduce these
spikes a drive resistor up to 10Ω can be placed between LG and the NFET gate.
Input Capacitor Selection
Because the inductor is at the input of a boost converter, the input current waveform is continuous and triangular.
The inductor ensures that the input capacitor sees relatively low ripple currents. The rms current in the input
capacitor is given by:
'iL
IRMS_IN =
12
(20)
The input capacitor must be capable of handling this rms current. Input ripple voltage increases with increasing
ESR as well as decreasing input capacitance. A typical value of 10 µF will work well for most applications. For
low input voltages, additional input capacitance may be required to prevent tripping the UVLO. Additionally, a
ceramic capacitor of 1 µF or larger should be placed close to the VIN pin to prevent noise from interfering with
normal device operation.
Output Capacitor Selection
The output capacitor in a boost converter provides all the output current when the switch is on and the inductor is
charging. As a result, the output capacitor sees relatively large ripple currents. The output capacitor must be
capable of handling more than the rms current, which can be estimated as:
2
¶
IRMS_OUT = D x IOUT x
2
'iL
D
2+
12
'¶
(21)
Additionally, the ESR of the output capacitor affects the output ripple and has an effect on transient response
during dimming. For low output ripple voltage, low ESR ceramic capacitors are recommended. Although not a
critical parameter, excessive output ripple can affect LED current.
The output capacitance requirement is somewhat arbitrary and depends mostly on dimming frequency. Although
a minimum value of 4 µF is recommended, at lower dimming frequencies, the longer LED-off times will typically
require more capacitance to reduce output voltage transients.
When ceramic capacitors are used, audible noise may be generated during LED dimming. Audible noise
increases with the amplitude of output voltage transients. To minimize this noise, use the smallest case sizes and
if possible, use a larger number of capacitors in parallel to reduce the case size of each. Output transients are
also minimized via the FF pin (See Setting FF section). Setting the dimming frequency above 18 kHz or below
500 Hz will also help eliminate the audible effects of output voltage transients.
When selecting an output capacitor, always consider the effective capacitance at the output voltage, which can
be less than 50% of the capacitance specified at 0V. Use this effective capacitance value for the compensation
calculations below.
Compensation
Once the output capacitor is selected, the control loop characteristics and compensation can be determined. The
COMP pin is provided to ensure stable operation and optimum transient performance over a wide range of
applications. The following equations define the control-to-output or power stage of the loop:
fP1 =
KD
2S x RL x COUT
fZ1 =
1
2S x ESR x COUT
RHPZ =
VOUT x ('¶)2
2S x IOUT x L
fpn =
fSW
2
(22)
Where RL is the load resistance corresponding to LED current, and Kf is calculated as shown:
16
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
LM3431
www.ti.com
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
RL =
VOUT
IOUT
3
'¶2 x 75 mV '¶ x RL
+
KD = 1+
L x fSW x 2
IOUT x R3
(23)
Since the control-to-output response will shift with input voltage, the compensation should be calculated at both
the minimum and maximum input voltage.
The zero created by the ESR of the output capacitor, fz1, is generally at a very high frequency if the ESR is small.
If low ESR capacitors are used fz1 can be neglected and if high ESR capacitors are used, CC2 can be added (see
below).
GAIN (dB)
A current mode control boost regulator has an inherent right half plane zero, RHPz. This has the effect of a zero
in the gain plot, causing a +20dB/decade increase, but has the effect of a pole in the phase, subtracting 90° in
the phase plot. This can cause instability if the control loop is influenced by this zero. To ensure the RHP zero
does not cause instability, the control loop must be designed to have a bandwidth of less than one third the
frequency of the RHP zero. The regulator also has a double pole, fpn, at one half the switching frequency. The
control loop bandwidth must be lower than 1/5 of fpn. A typical control-to-output gain response is shown in
Figure 19 below.
fp1
RHPz fpn fz1
FREQUENCY
Figure 19. Typical Control-to-Output Bode Plot
Once the control-to-output response has been determined, the compensation components are selected. A series
combination of Rc and Cc is recommended for the compensation network, shown as R9 and C4 in the typical
application circuit. The series combination of Rc and Cc introduces a pole-zero pair according to the following
equations:
fZC =
1
2S x RC x CC
fPC =
1
2S x RO x CC
where
•
RO is the output impedance of the error amplifier, approximately 500 kΩ.
(24)
The initial value of RC is determined based on the required crossover frequency from the following equations
using the maximum input voltage:
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
17
LM3431
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
RC =
www.ti.com
B
EAgm x ILIMgm x R4
B=
Acm =
fCROSS
fP1 x Acm
VIN
R3 x IOUT x KD
Where
•
•
•
•
B is the mid-frequency compensation gain (in v/v)
R4 is the current limit setting resistor
Acm is the control-output DC gain
the gm values are given in the ELECTRICAL CHARACTERISTICS table
(25)
Fcross is the maximum allowable crossover frequency, based on the calculated values of fpn and RHPz. Any Rc
value lower than the value calculated above can be used and will ensure a low enough crossover frequency. Rc
should set the B value typically between 0.01v/v and 0.1v/v (-20db to -40db). Larger values of RC will give a
higher loop bandwidth.
However, because the dynamic response of the LM3431 is enhanced by the FF pin (See Setting FF section) the
RC value can be set conservatively. The typical range for RC is between 300ohm and 3 kΩ. Next, select a value
for Cc to set the compensation zero, fzc, to a frequency greater or equal to the maximum calculated value of fp1
(fzc cancels the power pole, fp1). Since an fzc value of up to a half decade above fp1 is acceptable, choose a
standard capacitor value smaller than calculated. Confirm that fpc, the dominant low frequency pole in the control
loop, is less than 100 Hz and below fp1. The typical range for CC is between 10 nF and 100 nF. The
compensation zero-pole pair is shown graphically below, along with the total control loop, which is the sum of the
compensation and output-control response. Since the calculated crossover frequency is an approximation,
stability should always be verified on the bench.
GAIN (dB)
COMP
LOOP
0
fpc
fzc
RHPz
FREQUENCY (kHz)
Figure 20. Typical Compensation and Total Loop Bode Plots
When using an output capacitor with a high ESR value, another pole, fpc2, may be introduced to cancel the zero
created by the ESR. This is accomplished by adding another capacitor, CC2, shown as C13 in the Figure 14. The
pole should be placed at the same frequency as fz1. This pole can be calculated as:
1
fPC2 =
2S x CC2 x (RC // RO)
(26)
18
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
LM3431
www.ti.com
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
To ensure this equation is valid, and that CC2 can be used without negatively impacting the effects of RC and CC,
fpc2 must be at least 10 times greater than fzc.
LED CURRENT REGULATOR
Setting LED Current
LED current is independently regulated in each of 3 strings by regulating the voltage at the SNS pins. Each SNS
pin is connected to a sense resistor, shown in the typical application schematic as R10 - R13. The sense resistor
value is calculated as follows:
REFIN
RSNS =
ILED + INDRV
Where
•
•
•
ILED is the current in each LED string
REFIN is the regulated voltage at the REFIN pin
INDRV is the NPN base drive current
(27)
If using NFETs, INDRV can be ignored. A minimum REFIN voltage of 100 mV is required, and 200mV to 300mV
is recommended for most applications. The REFIN voltage is set with a resistor divider connected to the REF
pin, shown as R7 and R8 in the typical application schematic. The resistor values are calculated as follows:
2.5V - REFIN
R7 = R8 x
REFIN
(28)
The sum of R7 and R8 should be approximately 100k to avoid excessive loading on the REF pin.
NDRV
The NDRV pins drive the base of the external NPN or N-channel MOSFET current regulators. Each pin is
capable of driving up to 15 mA of base current typically. Therefore, NPN devices with sufficient gain must be
selected. The required NDRV current can be calculated from the following equation, where β is the NPN
transistor gain.
INDRV =
ILED
E
(29)
If NFETs are used, the NDRV current can be ignored. NPN transistors should be selected based on speed and
power handling capability. A fast NPN with short rise time will give the best dimming response. However, if the
rise time is too fast, some ringing may occur in the LED current. This ringing can be improved with a resistor in
series with the NDRV pins. The NPNs must be able to handle a power equal to ILED x NPN voltage. Note that the
NPN voltage can be as high as approximately 5.5V in a fault condition. The NDRV pins have a limited slew rate
capability which can increase the turn-on delay time when driving NFETs. This delay increases the minimum
dimming on-time and can affect the dimming linearity at high dimming frequencies. Low VGS threshold NFETs are
recommended to ensure that they will turn fully on within the required time. At dimming frequencies above 10
kHz, NPN transistors are recommended for the best performance.
CFB and SC Diodes
The bottom of each LED string is connected to the CFB and SC pins through diodes as shown in Figure 14. The
CFB pin receives voltage feedback from the lowest cathode voltage. The other string cathode voltages will vary
above the regulated CFB voltage. The actual cathode voltage on these strings will depend on the LED forward
voltages. This ensures that the lowest cathode voltage (highest Vf) will be regulated with enough headroom for
the NPN regulator. The SC pin monitors for LED fault conditions and limits the maximum cathode voltage (See
LED Protection section). In this way, each LED string’s cathode is maintained within a window between minimum
headroom and fault condition.
Both the CFB and SC diodes must be rated to at least 100 µA, and the CFB diode should have a reverse voltage
rating higher than VOUT. With these requirements in mind, it is best to use the smallest possible case size in
order to minimize diode capacitance which can slow the LED current rise and fall times.
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
19
LM3431
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
www.ti.com
Dimming
The LM3431 is compatible with both analog and digital LED dimming signals. The MODE/F pin is used to select
analog or digital mode. When MODE/F is pulled above 3.8V, digital mode is enabled and a PWM signal up to 25
kHz can be applied to the DIM pin. In this mode, the LED current regulators will be active when DIM is above 2V
(typical) and inactive when DIM is pulled below 1.1V (typical). Although any pulse width may be used at the DIM
pin, 0.4 µs is the minimum LED on time (in either digital or analog mode). This limits the minimum dimming duty
cycle at high dimming frequencies. For example, at 20 kHz, the dimming duty cycle is limited to 0.8% minimum.
At lower dimming frequencies, the dimming duty cycle can be much lower and the minimum depends on the
application conditions including the FF setting (see Setting FF section). In analog dimming mode, the MODE/F
pin is used to set the PWM dimming frequency, and duty cycle is controlled by varying the analog voltage level at
the DIM pin. To operate in analog mode, connect a capacitor from MODE/F to ground, shown as C5 in the
typical application (without the pull-up resistor installed). The dimming frequency is set according to the following
equation:
C5 =
40 PA
2 x fDIM x 2.13
(30)
In analog mode, the MODE/F pin will generate a triangle wave with a peak of 2.5V and minimum of 0.37V. The
DIM pin voltage is compared to the MODE/F voltage to create an internal PWM dimming signal whose duty cycle
is proportional to the DIM voltage. When the DIM voltage is above 2.5V, the duty cycle is 100%. Duty cycle will
vary linearly with DIM voltage as shown in Figure 21. Typical analog dimming waveforms are shown below in
Figure 22.
100
DUTY CYCLE (%)
80
60
40
20
0
0.3
0.9
1.5
2.1
2.7
DIM (V)
Figure 21. Analog Mode Dimming Duty Cycle vs. DIM voltage
ILED
100 mA/Ddiv
LEDOFF
5V/Div
MODE/F
1V/Div
DIM
1V/Div
2 ms/DIV
Figure 22. Analog Dimming Mode Waveforms
20
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
LM3431
www.ti.com
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
In PWM dimming, the average LED current is equal to the set LED current (ILED) multiplied by the dimming duty
cycle. The average LED current tracks the dimming ratio with exceptional linearity. However, the accuracy of
average LED current depends somewhat on the rise and fall times of the external current regulators. This
becomes more apparent with short on-times. To ensure good linearity, select NPN regulators with short and
similar rise and fall times.
Setting FF
To minimize voltage transients during LED dimming, the output voltage is regulated via the AFB pin during LED
off times. However, because the control loop has a limited response time, voltage transients can never be
completely eliminated. If these transients are large enough, LED current will be affected and ceramic output
capacitors may generate audible noise. The FF pin speeds up the loop response time, and thus minimizes output
voltage transients during dimming.
A resistor connected from FF to ground, Rff, sets the FF current which is injected into the control loop at the
rising and falling edge of the dimming signal. In this way, the FF pin creates a correction signal before the control
loop can respond. A smaller FF resistor will generate a larger correction signal. The minimum recommended Rff
value is 10k.
Since the amount of FF correction required for a given application depends on many factors, it is best to
determine a FF resistor value through bench testing. Use the following procedure to determine an optimal Rff
value:
An Rff value of approximately 20k is a good starting point. A 20 kΩ potentiometer in series with a 10 kΩ resistor
works well for bench testing.
The dimming frequency must be selected before setting Rff. Confirm that boost switching operation is stable at
100% dimming duty cycle.
Adjust Rff until the COMP pin voltage is between 0.8V and 0.9V. Next, monitor the cathode voltage response at
a low dimming duty cycle while adjusting Rff until the overshoot and undershoot is minimal or there is a slight
overshoot.
Check the cathode voltage response at the lowest input voltage and lowest dimming duty cycle and adjust Rff if
necessary. This is typically the worst case condition.
The curves in Figure 23 below show the variation in cathode voltage with different Rff settings. Notice that at the
ideal setting, both the cathode voltage and COMP voltage are flat. For clarity, the 3 cathode voltage curves in
this figure have been offset; all FF settings will result in the cathode voltage settling at 1.2V typically.
COMP
500 mV/Div
RFF
low
RFF
ideal
RFF
high
Vcathode
1V/Div
ILED
100 mA/Div
20 Ps/DIV
Figure 23. FF Setting Example
Once an Rff value has been set, check the cathode voltage over the input voltage range and dimming duty
range. Some further adjustment may be necessary.
Submit Documentation Feedback
Copyright © 2007–2013, Texas Instruments Incorporated
Product Folder Links: LM3431
21
LM3431
SNVS547G – NOVEMBER 2007 – REVISED MAY 2013
www.ti.com
In practice the FF pin also has a small effect on the control loop response. As a final step, switching stability at
100% dimming duty should be re-verified once the Rff value has been selected. At the optimal Rff setting, output
voltage transients will be minimized and the cathode voltage will be stable across the range of input voltage and
dimming duty cycle.
The ideal cathode response illustrated in Figure 23 may not be achievable over the entire input voltage range.
However, LED current will not be affected as long as the cathode voltage remains above the regulator saturation
voltage and below the open LED fault threshold (See Open LED section).
A wide input voltage range will cause a wider variation in the feedforward effect, thus making duty cycles less
than 1% more difficult to achieve. For any given application there is a minimum achievable dimming duty cycle.
Below this duty cycle, the cathode voltage will begin to drift higher, eventually appearing as an open LED fault
(See LED Protection section).
During an LED open fault condition, cathode voltage overshoot will tend to increase. If Rff is not set
appropriately, high overshoots may be detected as an LED short fault and lead to shutdown.
LED Protection
Fault Modes and Fault Delay
The LM3431 provides 3 types of protection against several types of potential faults. Table 1 summarizes the fault
protections and groups the fault responses into three types (the auto-restart option is described in the next
section).
Table 1. Fault Mode Summary
Fault
Mechanism
Action
Response
1 LED open
SC > 3.1V
DLY charges
continue to regulate
1 LED short
SC > 3.1V
DLY charges
continue to regulate
All LEDs open
AFB > 2.0V
DLY charges
Shutdown or auto-restart
Output over-voltage
AFB > 2.0V
DLY charges
Shutdown or auto-restart
multiple LED short
SC > 6.0V
DLY charges
Shutdown or auto-restart
Multiple LED short, VIN