Closing out this thread with final results as follows
(delays due to state-mandated stay-at-home orders for 3 months ....)
- Obtained replacement board for the one with the damaged socket (my fault for being "ham-fisted", as my dad would say).
- Installed latest chip for ho106 , build date 8-1-09, in both locomotives (using PLCC chip-puller)
- Reconfigure both decoders on programming track using JMRI.
- Short-hood-forward operation achieved through setting CV29 bit zero to 1 ("reversed").
- Locomotive runs correctly as a single.
- Build consist (NCE Advanced Consist) with locomotives in back-to-back lashup
(This is accomplished by setting CV19 bit D7 to "1" for the reverse-facing locomotive during consist setup.)
- Rear-facing locomotive does not move in correct direction when in consist
Sadly, the newest version of firmware available (thanks to Kelly Dorf for getting the chips to me) does not fix the issue.
Only solution appears to be to NOT use CV29 bit D0 for "reversed", and instead to swap motor leads and adjust function mapping.
Kelly, there should be enough information here to pass along to the engineers if they want to track the problem.