I think your conclusion makes sense on the failure mode. I suspect typical operation has the assembly operating in a fairly narrow band and over time can wear the feedback strip on the contacts of the board.So, being curious like a cat I decided to open up my bad one to see how it works. There is a worm gear on a small electric motor that drives a gear, that drives another gear, that drives another gear, that drives another gear with copper contacts on it. That is a lot of reduction there. I'll assume for very fine grained adjustments.
You'll see that those contacts slide across those black semi-circles on the circuit board. I'm guessing that is some kind of resistive material. Because of the limited resolution, what you can't see is that right in the middle of the range of motion the black stuff is very glossy. Like it is worn. I suspect that contact in that area is marginal causing the motor to drift.
That assumption makes sense as some members here have reported cycling the temp up and down through the range seems to help. I believe that is true but only a temporary fix. That original part number has been superseded twice already. Perhaps the third revision that I now have in my car will outlast me.
View attachment 94343
I've run into similar failures on linear transducers where I work. To solve that problem, the sensing portion was changed to a non-contact technology.