It could, and it could be a lot of other things. The bitch about noise is that there is no snakeoil. ANY silly thing can be the cause.

I totally agree! Unfortunately, it's that what we call experience, which is needed to design a noise immune circuit. And, unfortunately, this experience cannot be put into simple rules, which are always valid. There are so many overlaps and exclusions...

If we could have a look at the whole application, we would see the mistake within a few minutes, wouldn't we? But as we cannot do this, unless the original poster gives us some schematics or pictures, we must guess arround. Yes, a varistor could help, a snubber could also help, both together would even more help. But, nevertheless we don't know what's actually the cause of malfunction!

Well, my favorite subject which I put into discussion all the time is: 'Ground Management'. I have seen it many times, that improper ground routing was the cause of malfunction. If spiky currents flow across a poor ground connection, the term 'common ground' looses it's meaning! Means, voltage drop of spiky current across poor ground connection can be so high, that microcontroller can be either being reset, go into a locked status (hang), or even loose the code memory, if a flash type is used.

It cannot be overemphasized how important a proper ground routing is!!! The most relevant mistakes referring to the subject 'ground' are:

1. Omitting PCBs containing a solid ground plane.
2. Connecting noisy loads to ground at unsuited points.
3. Omitting shielding, while the shield has to be referenced to a proper ground.
4. Omitting filtering, while the filter elements have to be referenced to a proper ground.

Makarand, if you could publish a photo of your setup and a schematic we could give much more help...


Kai