TPS35: TPS35AA17AGADDFR TWD way too short

Hi Marcel,

Nice to hear back from you. 

Marcel Robitaille said:

The resistance from RESET to GND is 82.7 kOhm. Unfortunately, I can only measure resistance, not impedance.

Just to make sure, this is measured when the device is on and  RESET is asserted? Because, 82.7kohm is very high if the device is pulling that low. Typically when RESET asserts, the resistance between RESET and GND is closer to short, with few ohms from Rds(on). 

Also at this time the RESET pin does not have a load, correct?

Putting the schematic here for reference in our discussion.

Jesse 

Hi Jesse,

I'm glad you're there and I'm sorry for abandoning the other thread.

Jesse Villanueva said:

Just to make sure, this is measured when the device is on and  RESET is asserted?

No, sorry, I misunderstood. Nothing was powered when I measured 82.7 kOhm. Now with it powered and RESET asserted, I measure 68 Ohm.

Jesse Villanueva said:

Also at this time the RESET pin does not have a load, correct?

the It still has the pullup. I tried to cut the trace between it and the TPS2557, but it's hard to verify that it's completely open since I can't probe under the TPS2557.

Thanks,

Marcel

Hello. Our team will respond tomorrow. Thank you.

Hi Marcel,

Are you able to provide the top marking of the TPS35 device that your are using?

From all the information that you have provided, the device seems to be functioning, but the RESET is asserted as if VDD is still below the threshold. 

I just want to make sure that the device that you have is the proper device.

Jesse 

Hi Jesse,

Unfortunately, I don't have a microscope. Here is the best picture I could take. I hope it is good enough.

Thanks,

Marcel

Marcel,

That is definitely the correct device. 

Just to run down what we have so far.

• The device is not de-asserting the RESET pin even though the VDD is above 1.85V.
• The device does power on since the RESET pin's internal nmos is holding the output LOW. (evidence is LOW impedance between RESET and GND)
• Behavior above can only be due to two reasons, MR held LOW or VDD below threshold. (MR held LOW is not an option since a floating MR is probed and read voltage = to VDD) (VDD below threshold is not an option since VDD is read above 3.0V)
• The CWD and CRST capacitor seems to function properly
• I see that you have a ceramic capacitor on your board, based on the image. Just to double check, are the CWD and CRST are both ceramic?
• Using correct device based on the top marking.

Please let me know if any points is not correct.

Let me run some test on my evm and see if I can reproduce the issue.

Jesse 

Hi Jesse. Thanks for your answer. Everything in your list seems accurate to me. Yes, CWD and CRST and every other cap on the board is ceramic.

Hi Marcel,

Do you happen to have the TPS36Q1EVM, evm for the TPS35, to test your device on the evm board just for sanity check.

https://www.ti.com/tool/TPS36Q1EVM

I have ran some test on my EVM board using the TPS35AA17AGADDFR and everything seems fine on the device side.

Power up with VDD = 1.6V

power up with VDD = 3V

During my testing, the 3 conditions are the state that prevented RESET from going high

1. VDD below 1.75
2. MR low
3. CRST shorted to ground.

Jesse 

Hi Jesse. Thanks for checking those things. No, I don't have the evm board.

Hi Jesse. I don't think the device is damaged. I am able to reproduce the problem on multiple boards. The only explanation I can see is a problem with my design.

Hi Jesse. Actually, even though I already reproduced this on multiple boards, I just tested another one and it is working as expected. TWD is approximately 0.495s. I must have somehow burned all of the boards I tested previously without noticing it.

I am noticing this from the absolute maximum ratings. "CWD, CRST, WD–EN, SETx, WDI, MR (2), WDO (Push Pull): –0.3 VDD+0.3 (3)". I might have pulsed WDI when VDD was not powered on one board, but I find it unlikely that I did it multiple times.

Do you have any idea how they could have got damaged?

Hi Jesse,

Unfortunately no. I tried to desolder one broken chip to try replacing it with a TPS3436, but it was too small for me to do by hand. I am trying to coordinate hiring a company to do this. I can see if they can also put a "broken" chip on a working board.

Thanks,

Marcel