microcontroller crash problem

danadak

Joined Mar 10, 2018
4,057
An excellent tool in tracking down problems like this is a DSO setup for one shot
trigger, on a level, say a diode drop outside rail, ground, Vdd. And see if you can
capture an event. On Power pins or reset pin or GPIO. For starters. Then using
other channels see if you can ID the source when the event happens.

Also use infinite persistence mode (while in auto or normal sweep) to see what power
rails are poor in terms of pk-pk noise.

You can also use advanced triggering to see if you can find any runts or signals out
of spec.


Regards, Dana.
 

jpanhalt

Joined Jan 18, 2008
11,087
If you search on "relay spark MCU reset" you will get lots of hits. Here are two that may be relevant:

1) Almost identical problem solved with ferrite toroids. Author not clear where the toroids were placed, but they appear to be in the load lines:
https://forum.arduino.cc/index.php?topic=316604.0

2) Discusses isolation between mains circuits and MCU's DC circuits:
https://electronics.stackexchange.c...no-resetting-hanging-due-to-sparks-in-ac-line

The fact that without a load on the relay contacts you have no problem certainly makes it look like some unintentional and yet to be discovered coupling is at play.
 

nsaspook

Joined Aug 27, 2009
13,079
Ferrite filters on the load and signal lines work wonders for pesky mcu reset and hang problems when driving inductive loads. Proper grounding and shielding are also important.

PIC based XYZ motor tester prototype with motor control relays and filters.



For production unit one I mainly used snubbers under the relays but had to retrofit snap-on ferrites in a few locations for 100% stable operation in the new design.

The Wolf whistle sound effect are resets.
https://github.com/nsaspook/mandm/tree/master/mandm
 
Last edited:
Top