Strange problem on VFD board, any ideas?

[lumina2]

Hi I am having a strange circuit problem and I'm wondering if anyone has any ideas how to fix it. I am designing an AC variable frequency drive using a PIC. My PIC is optically isolated from the driver section. The board is mounted in a metal chassis. If I connect the earth ground to the chassis ground and touch the pic ground (isolated from earth) at various places on the board the circuit works erratically. Also if I try to probe the circuit with the scope as soon as I connect the scope ground(earth) to the pic ground the circuit stops working correctly so I can't even look to see whats going on with the scope when this happens. :-( If I remove earth ground from the chassis I can probe with the scope but then the circuit works fine so there's nothing to see.Also if I just attach one lead of my voltmeter to the pic ground the circuit doesn't work the way it should. I think it is an rf problem of some sorts but not sure how to fix it. Anyone have any thoughts on this one??
Jim

 

[jbartos]

Suggestion: The added path by touching the PIC ground creates a apath for the common mode currents and voltages.

 

[lumina2]

Thanks for the suggestion. What is the best way to get rid of this problem. I should be able to connect the PIC ground to earth ground shouldn't I and still have the circuit work? It actually isn't necessary for me to connect my PIC ground to earth ground but devices connecting to my board may have their ground connected to earth. Is an optoisolator the only way?

 

[cbarn24050]

Hi, you should have all your circuits referenced to ground somewhere, then you will be able to measure and find where your problem is.