Well, my success got rained on. I always run a restored deck as much as possible afterwards, because, well, that's what the owner is going to do. If it starts acting up after 20 hours of use, I want to catch it, if I can. Sometimes, that's a problem (like with the reel motors on the Sankyo transports. The more you use them them better they are. let them sit a few weeks, and poof! they start acting up again). For some reason, the left channel has gradually declined to around -6dB at 20kHz, while the right channel remained stellar. Time for some more investigation...
EDIT: Well after a week of delays, I finally got down to more diagnosis on Eriks deck. I recleaned the heads thoroughly, de-magnetized again, and then checked alignment. I was surprised that there was some oxide evident on the swab, as I had done a pretty thorough job cleaning it. Apparantly, not as well as I thought, so I continued a good healthy scrub with swabs and American Recorder S-721H. After this I still found the PB head left channel is down a few dB over the right, at 20kHz, but not enough to warrant replacement especially for the associated cost. I concluded this by simply exchanging L&R connections on the plug from the PB head to the connector on the main board. The right channel "electronics" were now showing -4db, with the left plugged in there, where with the right it was at 0dB (all references from -20dB). Oddly, the PBH right channel showed -2.5dB in the left channel electronics. So the gain on the left channel, AND the PBH left channel are both lower than the right, at high frequency. They are dead even at 400Hz at both -20 and 0dB. I started replacing all the electrolytic caps associated with both channels playback. Gained about a dB there. Then I rechecked the 10kHz level and decided that I would max out the EQ adjustment there, as I had left the originals in and it was low, but within spec. So I replaced the resistance values for the gain amps, and then re-adjusted the LC filters. After all this, I rechecked head height and azimuth, and was able to finally balance L&R to about -2dB at 20kHz, which lessens to about -1dB at 10kHz.
Subsequent replays have not shown any drift, and auto cals are right on the money. It sounds like a Nakamichi should, and recordings are indistinguishable from the sources. Both Dolby B and C perform very well, on pre-recorded tapes, and on recordings. Actually, I never gave it much thought, but on the decks like this, with only one dolby set used for encode, and then decode, it would make sense that their own tapes would be decoded more accurately than a deck with two separate sets doing encoding and decoding. I just need to record a calibration "as left" tape, and the deck can go back to Erik.
Erik, do you have a copy of the user manual? I don't and wouldn't mind one. I bought the service manual, just to do your deck.