TPS61165: PWM does not work

hi Leo

Can you help confirm if both old and new devices were purchased from TI or its agents?

would you mind clarifying how PWM doesn't work properly?

HI Mason,

IC was purchased from the distributor Arrow Electronics.

When using the old backlight, the brightness can be controlled normally by PWM (PWM frequency is 10KHz),

The waveform of the CTRL pin is as follow:

But using the new backlight, the PWM doesn't work properly.

The waveform of the CTRL pin is as follow:

What could be the reason for this?

hi Leo

How do you provide the PWM signal, by MCU or wave generator?  compared with old one, it seems there is additional pull-down for new backlight. Is there any regular pattern for this phenomenon? Any difference for the changing of frequency or duty cycle

I provide PWM signal from MCU. The frequency of the new backlight is wrong and irregular during PWM operation.

hi Leo

For new backlight device, please help check following items:

1. condition: MCU PWM applied at ctrl pin, tps61165 does not work, provide waveform of Ctrl pin

2. condition: MCU PWM applied at ctrl pin, tps61165 works, provide waveforms of Ctrl pin, SW pin, VLED, ILED.

also I would like to confirm the application condition: 5VDC input, 3*11 LEDs configuration with LED string voltage 27.9 and total LED current 220mA?

thanks  

Hi Mason,

The application condition is 5VDC input, 3*11(3 LEDs in series & 11 in parallel) LEDs configuration with LED string voltage 9.3 and total LED current 220mA.

1.tps61165 does not work, the waveform of Ctrl pin  as follow:

2.tps61165 works, the waveforms of Ctrl pin, SW pin as follow:.:

hi Leo

From the result, it seems the PWM signal at ctrl pin is not normal even when TPS61165 does not work. can you check the signal from MCU, if it has interference during the path? 

Besides, does the layout of new backlight and old backlight keeps the same? 

Hi Mason,

1. I changed the IC of another brand, and the PWM can work normally, so it doesn't seem to be the problem of the PWM signal or the backlight. What's your opinion?

2. I tried to connect a pull-down resistor to the CTRL pin, which can be dimmed in EASY SCALE mode, but this method seems to be inconsistent with the specification. Is there any doubt about this modification? Or is there any place to adjust on the circuit?

hi Leo

1. How many devices have this issue among newly purchased devices? Have you done the test that changed old TPS61165 on the failure board? what does "another brand" mean?  According to your description in #1, my understanding is that the problem is no longer existing after changing old TPS61165 and keeps other conditions the same, then it seems this is more related to TPS61165 itself rather than PWM signal, backlight, board layout or other factors. 

2. I do not think pull-down resistor can make PWM signal follow the logic of EASY SCALE mode. Normally, this should be pre-programmed in MCU with the timing shown in datasheet

My suggestion is to further confirm this is due to device itself by cross-check test, swapping old device and new device. Besides, it would be helpful that more waveforms can be provided under abnormal condition, SW pin, VLED, ILED. 

Hi Mason,

1. I have replaced the TPS61165 with a compatible IC of another brand. After the IC is replaced, the PWM can operate normally.

Our customer currently has 460 pcs (PWM cannot be dimmed), and the factory still has 600 pcs to be produced. Is it possible to handle urgent case?

2. We tested it, and it is true that a pull-down resistor is connected to CTRL, which can be dimmed in EASY SCALE mode. Will there be any worries in this use? Or is there any other modification method? Our agent is Arrow Electronics, can we send the sample to them for testing?

hi leo

what is the value of pull-down resistor? Please share the new schematic. after adding pull-down resistor, is it available to change LED current by different PWM duty cycle? 

Hi Mason,

pull-down resistor is 1KΩ.

After adding pull-down resistor, it is available to change LED current.

hi Leo

actually not sure why this works as per my understanding this only adds a ~0.5V DC offset on Ctrl pin.

As you mentioned old products works while new products not, I suppose you mean different batches of TPS61165. Then if it is really urgent, pls share waveforms of Vin, Ctrl, SW, VLED and I will check. we may go with this solution first. at the same time, send the issue device back for FA.

Hi Mason,

Now the new backlight uses TPS6115, PWM can't work normally, I replace it with other compatible IC (not TPS6115), PWM can work normally. Not a different batch of TPS6115.

Using the TPS6115 with the new backlight waveform is as follows:

Vin=5V, yellow is CTRL, bule is VLED.

yellow is CTRL, bule is SW.

hi Leo

1. for the waveforms, is it available to zoom in to see details in several switching cycles for SW? What is the duty cycle of input PWM in this waveform, 100%?  in new schematic, current can be controlled. then can you provide waveforms at 50% PWM duty and 20% PWM duty same way as 100%. It would be better to check the current under different duty cycle to see if current follows I=220mA * PWM duty. 

2. would you mind trying the action that remove R40 and R41 at the same time and see results? 

hi Leo

I would like to close this thread if no more update here. thanks

Hi Mason,

Sorry!

The TPS61165(SOT-23 package) I have on hand is damaged, and TI website is not in stock to request samples.

I will ask the factory measurement data for me later.

hi Leo

yes, feel free to discuss for any further information.

Hi Mason,

Please refer to the waveform of ES mode, but the dimming is still not possible using PWM mode. Remove R40 and R41 at the same time, PWM mode and ES mode cannot be dimmed.

waveform.rar

Hi Leo

sorry currently I cannot connect to E2E website on my PC while I fail to decompression the file on phone. I will try again and feedback to you as soon as I fix this

hi Leo

From the result of Vin, SW, VLED, it seems the operation is normal with Vin=~5V, VLED=~9V and SW 1.2MHz frequency shown in the waveform. Any result for the FB voltage (or LED current) under 25%, 50%, 100%?

Hi Mason,

FB voltage 25%, 50%, 100% is as follows:

hi Leo

It looks like the FB voltage follows duty cycle change. can you help clarify again on "dimming is still not possible using PWM mode". So far, the waveforms of FB, SW, Vout seems OK. 

Hi Mason,

At present, only the ES mode can be used to achieve dimming, but the PWM mode still cannot be dimmed.

hi Leo

So the waveforms above (SW, Vout, FB) are all under ES mode? if so, please provide SW, Vout, FB under PWM mode.(25%, 50%,100%)

Hi Mason,

Yes, the waveforms of ES mode was provided before. 

In PWM mode, the backlight will not flicker when the step is 256, but the backlight will flicker when the step is below 255. What is the reason?

hi Leo

From the video, I can only tell that the LED current under 256 is not stable. As I mentioned, waveforms will be helpful. you can also try to increase output capacitor to 10uF or more with 50V voltage rating. 

Hi Mason,

I tried increase the output cap to 10uF, but the backlight still flicker under 256.

hi Leo

Pls share waveforms of FB, Vout, SW, inductor current in PWM mode. waveforms with ES mode are OK and LED current under this condition can be changed according to duty cycle.

If it is not convenient to further get waveforms in PWM mode. You can just go with ES mode. 

Hi Mason,

In PWM mode, FB has an output of 200mV at the 256, but is unstable under the 256.

hi Leo

some of the summary here:

1. operation condition: Vin=5V, Vo=9.3V, ILED=220mA, LED 3s11p, L=10uH. the parameters are OK from my side according to theoretical analysis.  So I recommend to apply for one EVM and compare EVM and your demo  

2. ES mode operates well, PWM mode flicker. input PWM signal has abnormal pull-down that may cause unstable operation in previous waveform. Does this phenomenon still exists?  if yes, try to add high frequency filter at input of PWM input with 1kR 50pF and see any improvement 

3. TPS61165 cannot operate well in PWM mode, other compatible ICs are OK. any difference between them?

4.   waveforms of PWM input, SW, inductor current to help me better check details.

Hi Mason,

The specification of the compatible IC is as follow:

Is there any reason why the TPS61165 cannot operate well in PWM mode?

hi leo

from your previous waveform of input PWM, one possibility is that the PWM signal is interfered and that is why I recommend to add a high frequency filter at PWM pin. 

I can try to verify your operation condition on EVM next week and see if this is related to layout. 

Hi Mason,

If the PWM signal is interfered, why can the TPS61165 work with the old backlight to adjust the PWM dimming normally?

hi Leo

As I mentioned before, have you done ABA sawp test?  I just want to see if this is related to quality. 

ABA means: old backlight board +n new 61165, new backlight board + old 61165.

you mentioned 61165 can work with old backlight, then do they keep the same layout? 

Hi Mason,

Yes, I have done the ABA swap test.

Old backlight + new 61165, PWM can work normally.

New backlight + old 61165, PWM can't work properly.

The PCB layout is all the same.

hi Leo

then I think this should be related to backlight board rather than TPS61165 based on swap test result because both new and old TPS61165 can work on old backlight. You need to compare old and new backlight boards and check the differences.

Hi Mason,

At present, the difference between the new backlight and the old backlight cannot be found, and the new backlight uses pin to pin compatible IC with other brand, PWM can also be dimmed normally.

hi Leo

Basically, without waveforms in PWM mode, I cannot make further analysis. what I got now is the video that shows flicker, I suppose the current is not stable. You may try to change inductance (22uH or 4,7uH) or comp capacitor (100n or 470n) to see any difference. 

Hi Mason,

I changed inductance (22uH) or compensation capacitor (100n or 470n), the screen still flicker.

hi leo

information needed from your side:

1. pls attach PWM mode SW, Vout, FB under PWM mode.(25%, 50%,100%)

2. pls provide layout file for both old and new backlight boards for review

3. try add filter at PWM input pin

I will arrange some time for layout review. I cannot think of more other actions as EVM works well under your operation condition and what I can further check is layout based on swap test. Besides, also not sure why another P2P brand works.

Hi Mason,

 The layout file for both old and new backlight boards are as follow:

OLD backlight.pdf

NEW backlight.pdf

hi leo

thanks for the information, will feedback when I have some results

Hi Mason,

Is there any good news?

hi leo

Pls allow me some delay, I will give you feedback within this week.

hi Leo

not sure if they are right files as looks like not PCB layout file.

Hi Mason,

I only have this PCB layout file in my backlight specification.

hi Leo

based on my understanding, this is actually a PCB manufacture file rather than PCB layout file. I cannot figure out how components including IC are placed and how traces are routing, how GND plane is placed.

Hi Mason,

The PCB layout is as follow:

Top layer:

Bottom layer:

I am out of office until 12th Aug. reply will be delayed. Sorry for that