Bingo.

I cleaned the glass. What else to do? If it's not "reading" this pattern,
could indeed be the reason its not "zeroing".

I'd say that's almost certainly the case. You might check of
misalignment of the sensor, or for something interfering with the
light path from the lamp to the pattern, to the sensor.

There's no "sensor", per se, in this design, is there? I mean, the pattern on
the underside of the glass that marks the end of the "page" region is read by
the CCD and interpreted as "home", right? No sensor, cheap to do in software.
Or am I not understanding something? (always a possibility (c: )

I'm guessing the ccd is toast. Or one of the traces on the ribbon cable
(one of those mylar PC affairs) broke.

I've never seen a totally bad CCD. The golden rule of electronic
equipment diagnosis is that 90% of faults are due to bad connections
of some sort, & 90% of those connection faults are in moving parts.
I'd trace the path around the flat cable, starting from the joints on
the main board ribbon connector, right through to the joints on the
corresponding connector on the CCD board.

I ohm'd out the cable. Tests good.

Looks like the CCD has died. Well, the printer is "ok", but not great. It'll
be fine for text (albeit single-sided). Have inkjet for image printing.

Thanks,
Dave