On Thu, 03 Dec 2015 14:08:34 -0700, Don Y
wrote:

No, that's the lazy approach. That's the way auto mechanics start swapping
things (charging you for each "new replacement" -- even if it didn't FIX
the problem) out until they stumble on the "solution".

Since you are not paying for things you try that is not really a good
analogy but without the right test equipment, eliminating things in
the path is probably the only real way to go.

If I was really willing to "diagnose" this I would get a device type
GFCI, Disable the trip mechanism and look at the output of the
differential amplifier with a scope as I plugged in the lights, cords
etc looking for the one that is the offender. You could calibrate your
result using a pot and introducing a known fault value.
My bet is you will see this thing cruising in the 3-4 ma range so any
little glitch pushes it over.