Sweetbeats may be the best resource on this one.
I've biased a 3 head deck, but generally have avoided it on 2-head decks, rightly or wrongly.
However, the "record, rewind, evaluate, adjust, repeat method, looking for that (X)dB drop on the meters" may not be a practical method on 2 head decks, and therefore the "put in rec-pause and measure TP(X) for (X)mV" method may be the most effective method overall, and was devised in product design to alleviate the obvious downfall of no 'off-tape monitoring' limitation of 2 head decks.
Maybe relevant or not, is what I've observed on 3 head decks is the tiniest little nudge of the trimmer will affect the "needle" a fair amount, so trying to dial this in basically 'blindly', as on a 2 head deck, would seem impractical and prohibitive.
Nowhere in the procedure does it reference any input tones or reference to the tape or meters, but simply that when record is envoked that the "bias section" per the test point has 120mV. I would follow that procedure and move on. I know that's not as satisfactory as "seeing the needle peak and drop".
If Sweetbeats happens upon this thread, perhaps he can enlighten us. Failing that, go dig thru his many paged essay titled "388 Story", where he gives a great amount of detail of this and many other refurbishment procedures he went thru while hot rodding his former 388.
I admire you for attempting to tackle this methodically and with solid logic
, based on what we've all read and know about biasing a 3 head deck, but at a certain point it's apples and oranges. The written procedure may be as close as it gets to "precision", in this case.
My only other random thoughts would be to use 10kHz, as 12.5kHz might be on the upper usable frequency response of the device entirely, @ 7.5ips. Not only that, but to ensure the test signal is fed to the 388 at a verified input voltage that's recommended, whatever that is, either 0.316v or whatever else (i.e., -20dBVU). Pardon me for pulling those random reference
s out of a hat,... but follow whatever trusted literature you have.
Meaning: verify the ouput level in volts from your PT rig, as well as the frequency of the test signal. If the reference
voltage of the test signal is off or unknown, all bets for accuracy are off.
I'm sorry if I could not be any more help than these comments.
PS: If you have any breakthrough or revelation, please drop in and give us an update.