My client has been using this chip in their current design for two iterations of their current product. The first iteration involved a lot of 25 PCAs where the TPS65133DPDR was functioning properly with no issues. The second iteration of the product involved a lot of 8 PCAs where 4 of the TPS65133DPDRs failed (2 immediately on initial power-up, and 2 on power cycling) and the other 4 are working fine. On failure, the chip begins to smoke and in one case the ferrite bead on the input line blew up. The second iteration of PCAs involved very minor changes from the working first iteration, but nothing directly related to this chip. After taking off the broken TPS65133DPDRs we noticed that there was barely any thermal paste on its thermal pad. We then replaced the part adding an appropriate amount of thermal paste and so far the new chips are working correctly.
Our gut feeling says that this is just an assembly issue. This is due to the minimal amount of paste we saw on the thermal pad and the fact that adding a new chip with more paste fixes these issues. Even though we think this issue MAY be resolved, this IC is going into a class 3 medical device and we want to be absolutely sure that the problem is not going to show up again when we order a new batch of these boards for testing. Our sample size is also only 8 boards which was not enough for me to confidently come to a conclusion.
Have you guys seen issues like this before with this chip? If the thermal pad did not have proper contact with ground would these symptoms make sense? I can provide you with isolated pictures of the schematics and layout around this chip. The capacitors could have been placed better, but I do not think it is enough to make the chip fail.
I appreciate any feedback you can give me.
Thank you,
Nick Boring
