Last week I finally got all the parts to put together the next batch of Ebulobo boards (0.4) but found that the new ESP8266 modules from China were slightly different and were very unstable in being flashed and running code. They were also drawing a lot of current both getting very hot and making the regulator hot. He is a bit of documentation of the various experiments I did to try and track down the problem.
- Removed the antenna to see if there was a short and added a 1/4 wave antenna (I appreciate its not matched) – no change
- Soldered up my last v0.2 Ebulobo board – same problem
- Put together another v0.4 board with another ESP8266 from the batch – same problem
- Measured current draw – ~300mA when getting hot however this was not consistent
- While handling one board (powered) I found that applying pressure to the main ESP chip allowed the ESP8266 to function correctly and draw about 70mA – possibly a manufacturing problem with poor solder contacts?
- Examined the solder points under microscope (well, a reversed telescope eye piece) – appeared to be intact
- Resoldered the passives on the ESP8266 – no change
- Tried with a variety of settings (flash speed/size) – no change
- @ kindly soldered up some similar looking ESP8266-03 modules (with the non-branded chip antennas) and found that they worked fine.
Chip Details (Just for documentation)
- RainSun Chip Antenna
- Main IC: ESP8266EX 522015 P46N9
- Flash IC: AH1521 25041B 26G540
- Crystal 26.000MHz
- Unbranded Chip Antenna
- Main IC: ESP8266EX 482015 P48A15
- Flash IC: BergMicro BG25Q80AISCC 1550
- Crystal – 26.000MHz
- Ordered 2* ESP8266 off a UK seller on ebay, these boards have the original rainsun ceramic antenna. Soldered up another Ebulobo 0.4 board and was able to reflash immediately and the node started up with no issues. – doesn’t appear to be a PCB issue
I suspect that I’ve received a bad batch of ESP8266-03 modules, either a manufacturing problem or a small change which isn’t documented. I’ll probably finish up making this latest batch with ebay supplied ESP8266s and continue looking into why this batch didn’t work as it should. Hopefully its just this batch but if anyone else has similar problems I’d be interested to hear the details.