You should check www.diyaudio.com
Wow, this a classic problem with these amps, and a fun one for this era of Japanese audio from many manufacturers . . . a bit of a blast from the past.
During assembly, Pioneer used a white, gooey adhesive on the component side of the PC boards, kinda slopped on some of the taller components to keep them in place. With time and heat, this adhesive turns brown, and becomes slightly conductive. When this adhesive is in contact with any of the components in the input differential amp area (a high-impedance matched FET), the amp will intermittently put out a DC offset condition, and the protection relay will activate.
Repair requires desoldering/removal of the affected components, and thoroughly cleaning their leads, then scraping/dissolving the old glue from the board, then reassembly. I traditionally used tolulene as a solvent, it works well and leaves no residue . . . but it's REALLY bad to breathe the fumes, so observe all safety precautions . . . and keep it off of any ABS-plastic parts, as it will dissolve them as well.
Wow - you're kidding right? I would have never been able to figure that one out on my own. I recall seeing some (not a bunch) of that stuff in various places on the PC boards above the power supply and output devices and it is all brown. I'll dig into this today and see if I can alleviate the protection circuit problem.
On another note, since you're so familiar with this amplifier, I'm having trouble locating a replacement relay for the Omrons that switch rail voltages via the 4/8 ohm switch. One of them has contacts that are pitted beyond repair. The Omron part # is G2R-217P-FD-V-US-PF 12VDC. It would seem that they no longer offer this relay. Parts houses like Mouser carry hundreds of similar sized Omrons, but are unable to cross reference a suitable replacement. After looking at a bunch of data sheets on them, I have some ideas, but thought you might have a better one.
Thanks again Kirkus!
No, I'm not kidding . . . the Conductive Goo Phenomonon is commoner [sic] then one would think. Put your DMM on ohms, and you'll probably find a couple of lumps of the stuff that give you a reading of a couple megs . . . or even lower.
It strikes me as odd that Omron no longer manufactures a suitable relay . . . are you sure? I'd start simply by looking for the G2R series, and narrowing it down in the selection tables, rather than simply searching for the whole part number. You can also try some of their competitors' websites (i.e. Magnecraft, Potter & Brumfield, etc.) as they often let you search by another manufacturer's part number, and list equivalents.
The M90/M91 are really beautifully engineered, it's unfortunate that some aspects of their execution were pretty bad. But they can sound VERY nice, when they're working . . .
I found five devices with the goo - painstakingly removed all five, cleaned them and the board, and re-soldered, what a pain. Now, it looks as if I'm going to have to wait for the service manual as I can measure + / - 70 VDC at the speaker outputs of both channels in the 8 ohm setting. I double checked, and this voltage is present on both sides of every output device as well. Naturally, the protection circuit is activated.
Weird thing is that I brought it up on a VARIAC after the work today and its current consumption is nearly identical, 400 mA or so. This points to something simple, but the M91 manual differs enough that I just can't make heads or tails of it.
Thanks for the help thus far! I'll keep you posted.