According to the N23-related TSBs which are floating around, the only changes to the C140F triggers are two conditions both essentially relating to the transfer-case actuator/motor position aka "mode" sensor - Which is referred to simply as an "encoder" in all of the factory documentation I've found.
If it was an optical encoder (measuring shaft rotations for example.), that'd point to fluctuations in the circuit voltage being more likely.
But considering the reports of replacement transfer-case motors temporarily "fixing" the problem; IMO that points towards it using some kind of linear encoder, (maybe contact, such as a potentiometer.) which would be more susceptible to to corrosion/aging (that an optical or most other non-contact encoders.).
So trying to clean-up the supply circuit(s) is definitely worth a shot, but with caution IMO (I wouldn't spend too much cash experimenting with it.).
I don't have proof, but the 5V circuit should be quite well regulated already, just because regulating 5V from ~10-30V is very 'tried and tested', there are so many good cheap & easy regulators available, they would've almost had to go out of their way to regulate it badly
.
Anyway, with Clarky's N23-infected FDCM on the way, I'll hopefully have something interesting to report within a few weeks
I've basically just been experimenting with a few techniques to fool the FDCM into thinking that it's getting the values it wants from the transfer-case motor.