Never happened to me on pre-facelift. But admittedly, I hardly use the folding mirrors. Don't really need them they just came with a package of other stuff that I did want.
Here's my guess as to what's going on...
Hardware is fine and software is functioning "fine" (working as expected).
Replacing a mirror "may" fix it, for a while (because it's clean an perfect) but isn't the answer.
Can probably be fixed with software patch to accept wider margins of expected operating conditions.
The mirror doesn't fully extend because the run duration of the motor is fixed but not long enough in all situations. It doesn't matter when the sequence starts w.r.t. other things going on with the car, and unfold executes fine, it just doesn't run the motor long enough. The unfolding sequence is timed with no hard limit feedback switch to inform the application (layer), so mirrors don't "know" if they're extended fully and software can't test that they are. Application software drives these motors brute force and is blind to actual position. It's possible there is some type of limiting switch for power cut-off. (Limiter could be a microswitch or a current sensing feature of the motor driver). These limiters are only being used to protect the hardware from damage.
So the application programmer just picked a reasonable amount of time that mirrors should be unfolded. I bet they set the timing of the motor based on observations in a perfectly clean warm lab on the bench... (not with units in a freezer set to -40C)... and added a bit more time to accommodate different environmental conditions. And so far have been mostly OK and lucky to match most environments. But not you!
Any situation where there's extra resistance will cause drag on the motor and slow down the spin.. but not stop it entirely. These are reasons for mirrors not fully extending:
- cold lubricant in the motor / mechanism, or hinge, or sliding surface
- insufficient lubricant on sliding surfaces, motor / mechanism, hinge
- resistance from obstacles: dirt, ice.. on any sliding surfaces. (Felt being trimmed is this.)
If any of these problems are sufficiently resistant to motion, a current limiting type switch on the motor driving would detect a stall and quit. This is to prevent stripping gears. Anybody accidentally unfold mirrors while parked close to a wall? It happens. If Tesla mirror is not using current sense on the motor, they'll need a gear-slip mechanism to prevent gear stripping for mid-travel blockage.
Anyway, it's just that timing is trimmed too tightly for the tolerances of the mirror assembly (its lube) and the environment it must operate in. Maybe Tesla set and never did adjust the timing from when they went from first to second to third generation of mirrors now on the car.
If Tesla is using current sensing, there's no harm driving these mirrors for an extra whole second in either direction. Mirror motors will cut out when they reach limits. This will be the software fix. The trick might be if some mirror vintages have and use different hardware features... one patch might not work for the fleet. Maybe SC will be able to do a "recalibration" for some but not others .. if you bring it in.