GeorgesOatesLarsen
commented
3 years ago DUT8 draws 0.4A on 12V, and thermal pad gets hot very fast. Similar behavior observed on USB. Not even booting in programming mode rectifies. DUT6 has no such issue. Neither does DUT5
So DUT7 is stuck in programming mode, and DUT8 can be programmed (and communicated with) just fine, but actively tries to set itself on fire.
Both problems are indicative of some kind of short with the ESP32.
it is, once again, possible this was caused by our misaligned ESP32 package. I'm not sure how we went this long without seeing this program, though.
YIKES. 3v3 shows an anomalous short to GND. Furthermore, it turns out I can no longer contact the ESP32. What happened?! I was able to program it and get serial data just fine. I expose it to 12V, and now I cannot? I will perform the same test on DUT5. It cannot be programmed, but perhaps it will also immediately stop functioning after exposure to 12V.
No heat on thermal pad with USB power...
High current NOT drawn when on 12V.
I will see if I can bypass the autoreset for DUT6. I am realizing it may have been too early to give up on it just yet.
No luck.
I will need to think about this. The shorts we are getting, seemingly at random, and in response to no stimulus at all, are very perplexing.
DUT7 is stuck in programming mode due to an apparently irreparable short of the BOOTMODE pin to GND. It programs just fine, but literally cannot execute its program.
It is unclear how this short is even possible, but I did find some slight misalignment in the ESP32 package, as well as a via to GND that was perhaps too close to the BOOTMODE pin.
It is impossible for me to say which of these two issues is the cause of our short just yet. Our best bet would be to use a manual mill to cut the trace to BOOTMODE as close to the ESP32 as possible, and observe whether the short still exists. I cannot do that here.
However, I have already corrected the misalignment in the package, as well as greatly increased the margin between under-ESP32 vias and their pins (except where a short to the corresponding pin is intentional). Hopefully both of these issues will prevent this problem from occurring in the future.
I was able to remove the autoreset circuitry and manually program DUT8, so DUT8 is able to be load tested. Results here soon.