LM3421-Q1: No current limit

Please find some board pictures.

Hello Ubiratan,

Please allow our expert 48 hours to research and respond to your question.

Best,

Pedro Arango Ramirez. 

What do you mean "connect both circuits in parallel "? Does it means connecting Vsystem together? 

Is there any tested waveform to be shared? 

Regards, 

XG

Hi XG,

YEs, Vsystem from both drivers are connected together. See the picture below. The yellow line shows the copper for Vsystem..

I will send some waveforms in the next reply. 

If you need any specific signal, please let me know.

Ubiratan

HI Ubiratan,

Excellent we will be waiting for your waveforms. 

Best,

Pedro Arango Ramirez

Hi guys. 

please find below some waveforms (for IC pins and also for the load) for Vsystem = 9V (voltage when Iled = 0). When I connect the load of 4 RED Luxeon rebel, limiting Iled in 350mA, the Vsystem rops to 7V.

Vin

Vcomp

Vcsh

Vrct

Vovp

Vndim

Vddrv

Vgate

Vvcc

Vis

Vhsp

Vhsn

Vled+ (V in the positive terminal of LED chain)

Thanks.

Ubiratan

Hi, Ubiratan, 

Do the both circuits work well for case a) and case b) separately? 

I wonder if the circuit can work well for case b).

b) 4 RED Luxeon. - Typical Vf = 1.8V. Total series voltage: 7.2V

As Vsystem = 8.4Vcc is higher than output voltage 7.2V, a buck or buck- boost circuit needs to be considered instead of boost circuit. the circuit you shared should not be able to regulated the LED to target current as HSP/HSN  voltage difference can't be regulated. 

Regards, 

XG 

Hi XG.

Answering your question: Do the both circuits work well for case a) and case b) separately? 

Answer: No, for both circuits the only thing that limits the LED current is my bench power supply current limit. If I increase power supply voltage, LED current also increase.

-----------------------------------------------------------------------------

My fault I should have sent first the waveforms for 8 blue luxeon. Sorry. 

If I add one more RED LED, the total voltage will result in 5 x 1.8V = 9V. Considering two li-ion batteries will apply to the circuit 7,4V (average) would it be a solution to continue using boost controller ?

-----------------------------------------------------------------------------

Please find below the waveforms for 8 Blue luxeon, case a).

a) 8 Blue Luxeon - Typical Vf = 2.95V. Total series voltage: 23.6V

I adjusted the power supply to 8V, and limited the current in 0.35mA for figura 1 to 14.

In figure 15 I increased the power supply current limit to 0.700mA.

When the LED are ON the Vsupply drops to 7.1 due to the current limit of the bench power supply.

Figure 1: Vin

Figure 2: Vcomp

Figure 3:Vcsh

Figure 4: Vrct

Figure 5: Vovp

Figure 6: Vndim

Figure 7: Vddrv

Figure 8: Vgate

Figure 9: Vgate (zoom)

Figure 10: Vvcc

Figure 11: Vis

Figure 12: Vhsp

Figure 13: Vhsn

Figure 14: Vled+

Figure 15: Vled+ for Isupply = 0,70mA

For this signal I adjusted the power supply current limit increasing that to 0,700mA.

Thank you. 

Ubiratan 

Hi, Ubiratan,

Yes, if you add one more LED (or bleed resister), then you can use boost circuit.

According to the schematic, it seems you set LED current to around 1.4A. 

 = (1.24/0.07) *  (1/12.4)

= 1.43A. 

Assuming that power efficiency Eff is 85%, PWM dimming duty cycle =100%, Vin=8V, the input current I_in will be at least ILED * Vo/Vin* (1/Eff) = 1.43 * (23.6+0.124)/8/0.85=5A. 

Please kindly increase the input current limit to provide enough energy and not run into power supply OCP mode. Please test it again and let me know if any issue. 

Regards 

XG 

Hi XG.

Thank you for the explanation.

As I need Iled = 0,7A I have changed Rhsp and Rhsn to 470R. 

= (1,24/0,07)*(470/12,4)

= 0,67A

But when testing the board with 8 BLUE LED in series, Vin = 8V, Iled current is not limited to 0,7A. If I increase power supply current limit this current reaches 2A.

Waiting for a solution.

Ubiratan

Hi ,Ubiratan, 

What do you mean 2A? is it input current or LED current? If Input current, then it's possible according to above calculations. 

Could you further zoom in Figure 9: Vgate (zoom) to like 2us/div? or even 1us/div? 

Also, from Vis and Vcomp waveform, it seems the boost is not regulated. The Vis is reaching 245mV and run into cycle by cycle current limit. Due to it, Vcomp is always high as 4.8V. 

Is there long trace for power supply to the board? Please try increasing power supply current limit and/or more input capacitance so that Vin don't drop much during dimming. 

Can you also use current probe to see current flowing into 33uH inductor and JD2? this helps to know how much current accurately consumes. 

Regards, 

XG

Hi XG.

The driver is working from 8V. I need the driver starts working from 6V.

I tried to remove R18, that is a resistor of the voltage divider of UVLO, to check if it works but the driver did not worked well in voltages lower than 8V. 

What is your suggestion ?

Thank you.

Ubiratan 

Please try reducing the sense resistor R31 0.06 to smaller value. 

Hi XG.

I tried some resistor values.

0.06 ohms and 0.04 ohms > driver turns ON with 8V and turns OFF with 7V (both resistor, same behavior) 

0.02 ohms and 0.01 ohms the driver did not worked well. 

Thanks.

Ubiratan

1. Please add compensation circuit show in Figure 18. Compensation Circuitry. See details in 8.3.8 Control Loop Compensation

2. When you means you need the driver starts working from 6V, do you mean that the device doesn't switch at all when VIN=6 and Vo is close to 6V, not above 20V? 

3. Can you help capture waveform during powering up? Could you further zoom in Figure 9: Vgate (zoom) to like 2us/div? or even 1us/div? 

4. Please also capture inductor current during powering up

Hi XG.

1. What values you suggest I start a test ? 

2. I need the driver to start working with 6V because it is the condition for 2S li-ion batteries when they are almost being considered discharged, according to the manufacturer. So when batteries have a total voltage of 6V in series the driver needs to start working, feeding 8 HBLED in series with 0,7A.

Please find the waveforms for Vgate and Vinductor. I could not acquire inductor current. 

Vin=9,3Vcc   ,   Iled=0,7A    ,  8 Blue HBLED

VGATE

Vinductor

Vinductor - power up

Thank you !

Ubiratan

1. You may try RFS = 10 and CFS = 0.27uF as a start and fine tune them if needed. 

However, it seems the circuit works already and PWM is stable. 

2. Vin=9,3Vcc looks ok. 

If you need the driver to start working with 6V, please set Vin = 6V and test power up waveforms to see if it works. If not work, you may reduce output load first for try. 

Hi XG.

No way to have the driver working with lower voltages. Any other suggestions ?

Thank you.

Ubiratan

LM3421 can support voltage lower to 4.5V. You may use VIN rise edge as trigger to capture VIN/GATE/VLED/SW node etc. to see what happens. 

Regards, 

XG 

XG,

please help us to select other PN that could solve our case. 

We will change the chain of LED to try to find an easier IC to work with. So we will have the following LED series voltages. 

LED case 1: Vseries total = 8,4Vcc (1A max)

LED case 2: Vseries total = 11,8Vcc (1A max)

LED case 3: Vseries total = 11,6Vcc (1A max)

LED case 4: Vseries total = 12,2 Vcc (1A max)

LED case 5: Vseries total = 12,2 Vcc (500mA max)

LED case 6: Vseries total = 9,6 Vcc (500mA max)

We need to consider that the Vin min = 5,5V that comes from two li-ion cells. In this condition we are considering this voltage as the minimum discharging voltage. The voltage when they are fully charged is from 7,4V to 8,4V.

Thank you.

Ubiratan

Actually, LM3421 can be the right device with proper design. Other options like TPS929601, but it's quite similar with LM3421. 

Hi XG,

please find the waveforms for different voltages in the power supply.

Vsource = 8,0V

VIN x VGATE

VIN x VDDRV

VIN x VLED

And now for Vsource = 7,0V

VIN x VGATE

VIN x VDDRV

VIN x VLED

Thank you.

Ubiratan

It seems the circuit didn't switch at all. 

Please increase R18 to around 2.6k to avoid UVLO protection for a try. The below circuit set UVLO to 8V. 

Also, you need to add bulk capacitor at input side to reduce voltage dip during powering up. 

Hi XG. 

I added one more 4.7uF capacitor in the input. Also changed R18 to 3k. Please find the waveforms for different Vsources.

With Vin equal or below 7V the ILED drops to 0.4A and a high frequency noise is heard from the inductor. 

With Vin equal or below 5.5V the driver does not turn ON the LED. 

Thank you.

You need to optimize the input power supply. There is significant voltage dip on Vin that run the device into under voltage protection. 

Please reduce power supply trace, keep power supply/(or battery) as close as possible to input.  Or use a regulated power supply close to input to make sure input voltage is always higher than set UVLO voltage.  

HI XG, 

the circuit is working until 5.5Vcc. Connecting two power supplies in parallel and increasing the cables width solved the problem. The inductor is very hot but I did not perform temperature measurements yet. 

Now we will perform more tests with different LED. 

Another question. What driver IC do you suggest for currents from 100mA to 400mA ?The input voltages are the same. LED chain total series voltages vary from 9.6Vcc to 24Vcc. 

Thanks.

Ubiratan 

May i know what is the application? what dimming mode you expected? 

Hi XG.

The same application but for LED with lower power. The same PWM dimming mode. 

Is it possible, even being more expensive than lower power drivers, to use LM3421 for that current range ?

Thank you.

Ubiratan

You may consider TPS92643 etc. with integrated PFET.  

XG,

sorry but I forgot to inform that the driver needs to be buck/boost and work from input voltage of 5.5Vcc.

Thank you.

Ubiratan

If so, LM3421 should be the good solution.