>I was recently working on a circuit that showed susceptibility
>to a fluorescent workbench light being switched overhead.
>I think this would be a bit more repeatable.

One place I worked used a light box that had fluorescent tubes in it (about
4 12 or 18" tubes IIRC) and the "standard noise test" was to plug this into
the same electrical outlet as the "item under test" and operate the power
switch a fair number of times. Saw it done on a piece of computer gear,
which ended up with so many interrupts from the noise, producing error
messages about strange interrupts that it had to be rebooted.

As to the original problem, it sounds to me as though putting a snubber
across the coil will not fix the problem (along with a heap of comments you
already have). However I suspect that you may need a snubber across the
contacts for when the relays open their contacts, or possibly snubbers in
the light fittings for when the lights turn on.

Otherwise I suspect some form of MOV and line filter on the PIC power feed,
possibly with more heavy filtering after the supply. From skim reading the
posts it seems you have a switch mode supply, and these are not good spike
filters for anything coming down the mains, so you do need a good line
filter before, and supply filtering after if the gear is to withstand mains
spikes. You may even need to do your own line filter to get big enough
inductance to stop large spikes.

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist