It's getting weirder. Now heads 0 and 1 won't format properly, but apparently (judging from the speed at which it happens) it's now able to format the next six surfaces (and probably the seventh, although the controller doesn't support 9 heads).
Unfortunately that also spells doom for it as it is, since track 0 is obviously unavailable.
It seems to me there's something going on with the PCB; perhaps a slight difference with the voltage as you mentioned, a short circuit, unwanted resistance or the termination.
If I could somehow swap heads on the ribbon cable; I've seen that done on another drive but I won't risk ruining anything on it since I don't know the exact pinouts on that one (and there are more than just for the 9 heads and ground, 32 in total IIRC...
Perhaps I should just ditch the disk, but I have a strong feeling it IS possible to resolve that issue somehow. Seems to me that the PCB swap made it able to format completely different surfaces, and for that reason I've removed and reseated the ribbon cable, cleaned it with IPA, etc... but to no avail.
Covering pin 2 also seem to make no difference whatsoever.
It reminds me of a 251 I bought many many years ago. That too wasn't able to format, so I messed around with it having nothing to lose. It turned out that if I shorted two legs at the termination, it suddenly formatted properly - but I had to wait until the very last moment before I executed the format command, or the program wouldn't recognize the drive. And afterwards the short should be removed right away. But it got it working, and I've used that drive for many years now, formatting it if needed with the above method, haha