On 16 Sep 2003 19:11:07 -0700, (Alex Wiecek) wrote:
One thing I should mention, and I forgot in my original posting, is
that I had my voltmeter sitting on the filter caps. The measured
voltage was about the same as the one on the dial. While he was
welding, the voltage held up pretty good until the weld started to
sputter and burp. I've noticed DC voltage variations and quite an AC
ripple (about 6VAC). To me this indicates an excessive current flow or
the fact that the microprocessor is firing the SCRs at the wrong time
thus not regulating the output voltage properly.

It is very unlikely to be a microprocessor problem with those symptoms.

To me it indicates a high resistance developing as some poor connection
heats up in the circuitry feeding the filter capacitors. If you have an IR
temperature gun, look for hot spots when the welder malfunctions.
(If you don't have a noncontact temperature probe, unplug the welder
and *immediately* feel around for hot spots when it acts up.)

Likely places to look would be the bolted connections to the SCRs,
primary or secondary connections to the transformer, the circuit breaker,
etc.

I know you said you checked the filter capacitors, but they could still
be at fault. One or more of them could be developing a high ESR as
they heat up under full load. But I'm betting on a loose or corroded
bolted connection somewhere in the circuit before the capacitors.

Gary