TPS61500: PWM dimming - Boost stays on when EN is low - until OVP

Hi Jim,

There is a 10ms shut down delay time from EN high to low. I think the behavior on your board is normal.

On EVM board, here is a NFET Q1 to discharge VOUT to GND that shut down the device immediately.

Thanks for your reply, Hardy.

On the normal EVM, Q1 is turned on when EN is high, so the 2K load on VOUT goes away when EN is low. It doesn't discharge VOUT then. I think Q1 just pulls the gate of Q2 low to turn it on. Besides, how would adding a 2K load shut down the device?

But my EVM is not connected that way. I only have one NFET as shown below, the same configuration as on my board. But the boost converter only switches when EN is high.

I found a mistake. My EVM was not connected as I thought. I connected FB to the drain of the NFET by mistake, so the FET was included in the feedback current sense resistance. When the FET was off, FB went high and shut down the converter. I can see that's not safe since FB will go above it's 3V max.

Now I fixed that mistake, so FB is connected directly to R8 and the converter keeps switching between EN pulses. It doesn't go up to OVP as on my board, so there is still some difference I don't understand.

Robin, please keep following this case.

Hi Jim,

From your schematic, it seems that FB is actually connected with the source of FET. Why did you say it's a mistake? Anyway, pls send your final schematic to us. Thanks!

BR,

Robin

Sorry for the confusion. I am comparing the EVM to my own board. The schematic you marked is my board, which has FB connected to source of FET. But when I modified the EVM, I connected FB to drain of FET by mistake, as shown in the second schematic of EVM that I posted. But even when I change the EVM to match my board, they still behave differently.

It seems like the 'tuning' of the device may have changed between the 83T chip on the EVM and the 85T part on my board.

Can you explain what EN actually does in the PWM mode? If it doesn't force the boost to stop when the FET is turned off, what does it do?

Thanks for your effort, Robin. I found another device that is spec'd for my low duty cycle and it's working on the bench, so we can close this issue.

Hi Jim,

Would you mind sharing the P/N? Thanks a lot!

Robin

MIC3223. Sorry - not TI!

Hi Jim,

Thanks for your info.

One more question: Comparing to analog dimming, why PWM dimming is preferred? For PWM dimming frequency and lowest duty cycle, do you have any clear requirements?

So far, our most products are analog dimming and we're researching to PWM dimming marker situation. We're appreciated for your valuable info. Thanks!

Robin

The application is illumination for machine vision, requiring variable exposure times of 200-400 usec at 120 fps to stop motion. The pulses must be synchronized with the camera, hence the need for digital PWM at low duty cycle.

Hi Jim,

I think TPS61500 is able to handle this requirements although the minimum PWM frequency is 200Hz in datasheet. 120Hz should work and I'll test it. 200us for 120 fps means the minimum duty cycle is 2.4%, which is also acceptable for this device.

I have received some questions about machine vision, but they all asked about BUCK LED driver. So I wonder the quantity of LED in series in your application.

Robin

I'm using a 5 LED string - about 3.5V per LED, so 17.5V at 600 mA.

Note that I had it working on the EVM, but the same circuit on my own PCB behaves differently. That's why I asked about a change between IC versions. Anyway, the MIC part datasheet explicitly covers my application parameters and it's working well on the bench. I will test on my own PCB next week.

Robin, please help to close this thread.