On Thursday, March 22, 2018 at 12:25:55 AM UTC-4, Ron D. wrote:
I've had bad contacts on relays before. Both audio equipment and a programmable power supply. To fix the power supply, I wrote a program to open/close the relay every 10s overnight and the problem went away.

I'd measure the DC at the speaker terminals before and during a fault first.
I'd expect something initially and probably zero if your loosing contact.

Then do the same, but before the relay.

He won't be picking up any dc on the outputs. The other channel remains. If it was a DC fault detect, the relay would open and both sides would go out.

Besides, the DC detect is hair trigger on these and the meter probably wouldn't resolve the value before the relay opened. A min/max peak recorder feature on the dmm might catch it but even that isn't guaranteed.

The easiest thing to do if the relay is suspected is to thwack it while the channel is out (but not hard enough to bounce to effect a contact bounce). If the contacts are gimpy, something will be heard on the mute channel.

The problem could be anywhere. If he has pre amp in/out jumpers on the back panel, he could remove them and put in short RCA cables and cross them to see if the drop out moves to the other side. If so, it's before the junction, if not, it's after.

The easiest thing to do is trace it with a scope when the channel acts up. The problem is that once you get near an intermittent circuit with a scope probe, the problem clears up..