ESP32 get killed with an external battery
Posted: Tue Feb 03, 2026 11:57 am
Hi everyone,
I use an ESP32-WROOM-32E, integrated on a PCB.
Everything works like a charm when the board is supplied directly with a 12V power supply. On my board, the 12V is converted to 5V and also converted to 3.3V. The 3.3V supplies, among other things, the ESP32.
The issue I'm facing appears when I plug in an external 12V battery. I use a Chinese battery, Nastima (Amazon).
It is very difficult to replicate, but sometimes, for some reason, the ESP32 gets damaged. The symptom is a short circuit between 3.3V and GND inside the ESP32-WROOM module.
My strong suspicion is an overshoot on the 12V line, which propagates to the ESP32. I can observe some overshoots on the 12V when I plug/unplug the cable, but nothing that seems destructive. The issue is that I can't reproduce the failure on my side — only my customer is able to kill an ESP32.
There are some protections on the supplies, such as ferrite beads, decoupling capacitors, and diodes. Only the ESP32 is affected by this setup, while the other components do not seem to suffer from this issue.
Have you already observed such things on your side? Any inputs on how to reproduce the problem? Any hardware recommendations to implement?
Any help is welcome
Best regards
I use an ESP32-WROOM-32E, integrated on a PCB.
Everything works like a charm when the board is supplied directly with a 12V power supply. On my board, the 12V is converted to 5V and also converted to 3.3V. The 3.3V supplies, among other things, the ESP32.
The issue I'm facing appears when I plug in an external 12V battery. I use a Chinese battery, Nastima (Amazon).
It is very difficult to replicate, but sometimes, for some reason, the ESP32 gets damaged. The symptom is a short circuit between 3.3V and GND inside the ESP32-WROOM module.
My strong suspicion is an overshoot on the 12V line, which propagates to the ESP32. I can observe some overshoots on the 12V when I plug/unplug the cable, but nothing that seems destructive. The issue is that I can't reproduce the failure on my side — only my customer is able to kill an ESP32.
There are some protections on the supplies, such as ferrite beads, decoupling capacitors, and diodes. Only the ESP32 is affected by this setup, while the other components do not seem to suffer from this issue.
Have you already observed such things on your side? Any inputs on how to reproduce the problem? Any hardware recommendations to implement?
Any help is welcome
Best regards