TPS40170: Vout delay more than 5sec when -10 degree start-up condition

Kay,

5 seconds is a very long time, even at extreme conditions.

Can you check the EN pin and see what's happening?

If EN looks normal, check FB and COMP.

-Sam

Hi Sam:

1. 5 sec only will happened in low-temp condition only

2.  customer try to connect UVLO & VBP directly, than no delay for 5V output

3. We've an test waveform for EN & SW pin (@ low-temp) as attached.

4. currently COMP pin value would be 3.83K ohm, 8200pF, 220pF (normal temp is OK for this set-up) 

波形圖.pdf

Kay,

Thanks for this information. Interesting that connecting VBP directly to an external supply will cause no delay. I'll talk to the designers to see what could be causing this LDO to delay at cold.

Questions:
1. Does the delay jump from ~0s to ~5s at a certain temperature or is it more gradual? Is there ~2s delay at maybe 0 degrees, ~1s at 10 degrees...?

2. Is this happening on lots of boards/ICs consistently?

3. Your UVLO divider should yield ~1.2V on UVLO but the scope captures show ~1.7V. Is the schematic up-to-date or maybe you're testing at higher VIN.

-Sam

Hi Sam:

thanks again.

Q1:
Test environment: using chamber cooling down to -10 degree and test it.

Q2:
currently known is non specific lots of boards/ICs consistently.

Q3:
customer RD tried to low-down the range of UVLO, still could duplicate the symptom.

and any other though or suggestion from your viewing could make?

Kay,

Have they tried changing the values of the UVLO divider? Instead of 2M and 100K they can try 200K and 10K.

I'd like to clarify Q1. If you measure startup time at T=-10C, -9C, -8C, -7C... does the 5 sec startup go away immediately (5s, 5s, 5s, 0s, 0s...) or gradually (5s, 5s, 5s, 4s, 3s, 2s, 1s, 0s, 0s...)?

-Sam

Hi Sam:

Could you also help to check whether this symptom can duplicate in your side?

1.  RD  try to connect UVLO & VBP pin together, than 5V output won't be delay

2. 5sec startup and go away immediately at -10C chamber (refer attached waveform)

Kay,

I will attempt to reproduce this in the lab.

In the meantime, please change R51 and R53 to 200k and 10k instead of 2M and 100k. That will rule out if UVLO is the issue.

-Sam

Hi Sam:

Thanks again, we'll discuss with RD member about this method. Could we have a schedule to know when will have your thought to reproduce?