Thanks for confirming & thereby warning others. It would be exceedingly dangerous if the disengagement occurred on your curve at the exact same time as oncoming traffic Which model do you have? My fault occurs on X, but not 3. Both have FSD computers
Mine is a model 3. The curve is to the right, so if i don't take over quickly, the car is in the trees, not oncoming traffic.
My parents knocked up a FSD/robo taxi computer over 50 years ago that I'm still using today, I'll stick with that for a while longer yet.
Thanks. Just had a 350km drive today. Repeatedly happened on curves. Lean towards torque sensor being the issue, but I am interested in your experience as it seems your problem started recently also?
Mine only had disengagements over the last week or so, but only in one place that I can recall. It happened at that one place multiple times though. I updated to 2021.4.10 yesterday, and it didn't disengage there on my way home yesterday. Fingers crossed.
I posted on another thread. My AP was engaged for about 6 miles. Just AP, one hand on lower left steering wheel, Hard breaking and loud warning out of nowhere. I shook the wheel and hit breaks to disengage. No one around except for a car 500 yards back. I have version 2021 4.6. Maybe mine is something different,
Hint: What time of day? I find that driving into the low bright sun (AM/PM) causes problems with lane detection and occassional "here, you take over" behaviour. Perhaps you can beep the horn to save the dash cam footage? Interesting to note the captured view of front windscreen camera.
Warnings should come on while it meanders, not disengage autosteer altogether as if the steering wheel or brake pedal has been moved
Probably the third time I've had to mention it in this thread, but we're not talking about "take over immediately" alarm, we're talking about the "ding-dong", as if you just moved the steering wheel to disengage AP, but tacc is still on.
Just got 2021.4.11 Hasn’t fixed the problem. So must be steering torque sensor & hence pure coincidence
I haven't driven much over the last few days, a Bugger. I haven't driven much in the last week, and haven't been on the road where my disengagements were happening, so don't have any new feedback. Wouldn't a steering torque sensor issue result in more nags to apply more force to the steering wheel rather than just disengaging without warning?
I’d prefer a nag, rather than the dangerous disengagement. @immolated do you happen to know the thread where others had an issue with the torque sensor? (I’ll do a forum search)
Could be my imagination, but I've noticed since recent updates my steering wheel sensor seems more sensitive - it disengages AP with the slightest tug. It's an AP1 Model S, 2020.48.37.1, so probably not relevant to the OP's problem, but it means they might be tweaking things related to the sensor.
Tesla Service Centre technician told me 2 years ago that the torque input sensor sensitivity is factory set, after I requested them to increase the sensitivity (the Model X loaner they provided required less input). Maybe I was lied to.
Maybe you weren't. The steering system has 4 torque sensors forming two redundant systems. They are not serviceable and if faulty the steering rack needs replacing. Tesla can sure as hell modify the threshold at which any of the steering torque related functions occur by software without touching the hardware.