I am currently working on a project that involves charging 18650 Li-Ion battery through USB-C using TP4056. Nothing really fancy (or at least have I thought!).
The circuit is pretty simple - Battery connected permanently to the TP4056; the rest of the circuit (two boost converters, ESP32 etc.) is separated by toggle switch (SW3).
Now, very weird thing happens when the user tries powering on / off the rest of the device by toggling the switch - the TP4056 starts drawing approximately 1A from the battery and instantly overheats, even though it is not plugged into the USB (+5VP rail is disconnected). This does not stop until I take out the battery.
This problem happens only from time to time, so it is pretty random. However one interesting thing is that when I tried to power the device with my lab PSU instead of the battery, it used to happen a lot more.
I managed to capture the moment, when I toggled the switch and the problem occurred. The screenshot shows voltage on pin 5 of the TP4056 when toggling the switch:
but IMHO there is nothing that special, that should "lockout" the IC until it is reset.
The TP4056 that was causing me problems was ordered from LCSC, PN C16581. Then, I swapped it with another TP4056 (again from LCSC, but PN C382139), made by a different manufacturer and now it works flawlessly. But I am still a little bit worried that this problem could come back even with the "good" TP4056, because I highly doubt that same ICs from different manufacturers should behave so differently.
For those thinking that it might have been only one bad TP4056 - it was not. All C16581 that I had at home were having the same problem.
Has anyone ever experienced something like this? Does anybody know what could be causing such problem?