Opinions re: debugging/repairing Cambridge Audio 540A amp stage

Hey! My old Cambridge Audio 540A has seen better days. It's been acting up recently. I'm interested in fixing it up myself if I can, but I'm looking for opinions on where to start.

Symptoms: speaker output on *any* channel (L/R, A and B) can drop instantly from the correct level to partial (including zero output). It does not happen consistently: on some audio streams, it seems more stable, but I have no way to reproduce reliably. The drop is instantaneous, not a slow shift. I can reset the audio levels if I switch input selector and back, until it happens again. If I do nothing, sometimes the audio will click back into place on its own.

This can happen on any input. Multimeter suggests that the voltage on the output drops when the audio drops; I don't suspect an issue with the speakers, but I don't have spares to test with. Headphone output is always perfect. My suspicion has been the amp stage in the 540A.

This is not a new unit, and so I popped the top off. My first suspicion was the big capacitors, because of age and the gunk they'd deposited on the board: https://i.imgur.com/u0Qnj8O.jpg  So I've cleaned those and replaced them. The problem persists, but I didn't have replacements to hand for the smaller capacitors.

I'm feeling my way around as I go; I have a copy of the service manual so I can decipher things, but I'm looking for good approaches to debugging these units to home in on the problem.

Note, obviously I can replace this unit easily. But if I can fix it, I'll feel happier, and I'm also happy to get my hands dirty. I'd appreciate pointers from folks on what to attack!
Post removed 
I own a 540A which i am using in a secondary system. I believe it was manufactured in 2004, I bought it used here on AG for $150 shipped to my door.
It is probably worth that today. 

Given the above, if something like what you are experiencing happens to this amp I would be inclined to open it up and look for obvious things however not invest to much time or expense trying to repair. Your amp could easily be replaced via the used integrated market at less expense than a repair and you could own a newer integrated. Just my thoughts.

Good luck which ever way your proceed.
Thanks for your responses!

Without doing anything significant today: if the gunk on the board was glue, it was no longer adhesive. It flaked away when I scraped it; only the solder held those old capacitors in place. Also, if the system's protection feature clips the audio, it never triggers the flashing light that should tell me it hit a failure mode. Also, the system doesn't appear to be super warm; one of the heatsinks accumulates a good amount of heat, but that's where it should go anyway. It's not warm enough that I'm concerned. I don't suspect the volume control since audio via the headphone jack is always perfect.

And yes, I could pick up even a new low-end CA amp to replace this and be quite happy.

But, this amp has been so good that if I could nail down (easily) what's wrong, it'd be great to fix it. I'm more used to electronics boards with chips, though; CA's wiring is a bit of a riot. And since some of the voltages via the amp stage are nontrivial, I'm definitely not poking anything I'm not sure about.
Okay, the relays seem to have been the glitch. Swapped out the old ones for some replacements. The amp seems solid again.

Awesome pointer. Thank you!
to check for relay problems, you let the unit run, usually while listening on headphones. Of course, first ascertaining that the unit has the audio amplifier proper being the source point that powers the headphone jack.

This depends upon the circuit having been run through the relays before it hits the headphone jack. Integrateds/stereo receivers are run this way, big surround amps are generally not. Big surround receivers generally run the headphones off a dedicated op amp.

then use the back end of a old toothbrush (you use old toothbrushes to clean boards and whatnot) or very small paintbrush handles, etc...to ’tap’ the tops of the relays. Generally the sound will cut in and out when you do this.

You tap gently, very gently, and slowly bring the pressure of the taps up, in order to see if the vibration will make the contacts cut in and out. Soft taps are necessary so you can be sure the vibration is not going elsewhere and thus making the problem go back to being ’anywhere’ and thus unknown in source point.

Relays these days, as in the past 15+ years, are junk, compared to the relays in +25 year old gear. Generally speaking.

you can take apart and fix 25-40-45 year old relays. They are built for it. They have removable covers, nonmagnetic conductive parts, silver contact points, etc. Which is part of the problem, silver oxide, but also their gift of quality. You can clean the contacts with 2000 grit sand paper. Just fold a small piece over double, slide it between the contact points, gently push the contacts together and then slide the sandpaper back and forth (up and down, actually) maybe 5-10 times, and you are done. Good for probably another decade.

Headphone testing is many times a good idea as an amp can sometimes be on the verge of a spectacular failure (thar she blows!.....) and you don’t want to blow up a good speaker for no reason, if you can avoid it.
Hi! New in this forum, found in google when facing the same issue in my 540A. First thought the problem was with my newly bought used refurbished B&O Beovox CX100 speakers and was about to call the speaker seller back and complain about a badly distorted speaker... but a quick check turned out that bridging the speaker output relay made the sound crystal clear.
For making the amp a tiny bit more "hi-fi" I decided to simply bridge the relay bypassing the contact paths completely with tiny bits of wire. I understood I would be hearing a loud "thump" in the speakers upon every push of the power button, but that was according to me a cheap price to pay for never having a relay issue again. But how wrong I was! After only a couple of power-ups with the relay bypassed (yes, with audible "thumps"), there was a small "pop" from the power board - two tiny fuses blew. So the amp was definitely not designed to work with a bypassed relay due to the apparent big power surge in the speaker paths.

It appears my 540A has a single, 2-pole relay controlling both channels with one relay coil. The left channel path shows ca. 6 - 7 ohms on the multimeter when the relay is closed, indicating a bad relay contact surface.
Side note:
This page shows that the 640A has two relays in the speaker path(s), either it is one for each channel or it is one for A and B speaker pairs respectively. The 540A has a hardwired front panel pushbutton switch for "Speaker B". I have not seen a 640A myself but I bet the B pair has its own relay and a "soft button" to it.
Side note 2:Could decide to go with connecting my 2 speakers only to the "B" outputs, another way to bypass the relay issue temporarily. Anyone has opinions about that? The B should have identical output ratings  - or not?