ESP32 restart on mechanical shock

Projects Programming
1

Hi

I have a project to measure velocity of door. So a have a Wemos ESP32 D1 Minikit connected to a GY-571 giroscope.
The circuit runs very nicely until a have velocities above certain value. From this values above the ESP32 just restart itself, with this reset message below:

21:03:52.969 ->
21:03:52.969 -> Stack smashing protect failure!
21:03:52.969 ->
21:03:52.969 -> abort() was called at PC 0x40149c6f on core 1
21:03:52.969 ->
21:03:52.969 -> Backtrace: 0x400927dc:0x3ffcc480 0x40092a0d:0x3ffcc4a0 0x40149c6f:0x3ffcc4c0 0x400d1acd:0x3ffcc4e0 0x400d2a16:0x3ffcc530 0x400d5061:0x3ffcc550 0x4008e6fd:0x3ffcc570
21:03:52.969 ->
21:03:52.969 -> Rebooting...
21:03:53.016 -> ets Jun 8 2016 00:22:57
21:03:53.016 ->
21:03:53.016 -> rst:0xc (SW_CPU_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT)
21:03:53.016 -> configsip: 0, SPIWP:0xee
21:03:53.016 -> clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00
21:03:53.016 -> mode:DIO, clock div:2
21:03:53.016 -> load:0x3fff0018,len:4
21:03:53.016 -> load:0x3fff001c,len:1100
21:03:53.016 -> load:0x40078000,len:9232
21:03:53.016 -> load:0x40080400,len:6412
21:03:53.016 -> entry 0x400806a8

After some tracing it was clear that ESP32 resets during the

SerialBT.print(txMsg);

instruction. AND it occurs (the restart) only when bluetooth module is PARIED with my phone (running the master app). I already change the battery used by USB powering.

Uhh... I forgot to say that the restart occurs when the door (where ESP32 is attached to) knocks...

Did anybody see something like this?
Any help will be great.

Gustavo

2

From this the "Stack smashing protect failure!" sounds related to an overflow. Which might happen is you somewhere in your code try to put that large value into a to small spot. Just cropping will happen when you try to put an int into a byte for example. But things get different when you convert your value to text for example. You need more room for "10003449934" compared to "123".

3

Running freeRTOS, I'd move the suspected task to the other core to see if the issue follows.

I have an ESP32 on the X axis of a X/Y stabilized platform with the Z axis following a scanning route. The application starts off with a hard slam of the X/Y/Z axis during the magnetometer calibration. I do not experience resets during such movements.

Would you happen to be using a Adafruit library?