Autosteer would not disengage

[l1steven1l]

While on the freeway today, I tried to disengage the autosteer but it would not let me. Pressing up on the right lever did not work. Forcefully turning the wheel did not work. Stepping on the brake only slowed the car down but was still stuck in autosteer. The car allowed me to change lanes only when I used on the turn signal but the wheel was pretty stiff to turn. I had to pull over on the side of the freeway and physically get out to allow the car to reset itself back to normal.

This was a first for me after 2 years of ownership. Is there anything else I can do besides pulling over if it happens again?

 

[Daniel in SD]

OP stated that neither manual steering, brake application nor upstalk would disengage AP. Assuming the OP is not making things up (this is the internet) then he appears to have a failure in the computer. I would hope it's a hardware issue as the AP disengagement software should be very well debugged!

If this happened to me, I would be very leery of driving the car before Tesla service looked at it and would have likely called roadside assistance immediately.

OP did not provide model and software version.

Yet somehow he was able to steer on to the shoulder. He didn't describe anything seen on the screen. He described the steering as very "stiff to turn" which sounds like a power steering failure to me.

 

[Supcom]

Yet somehow he was able to steer on to the shoulder. He didn't describe anything seen on the screen. He described the steering as very "stiff to turn" which sounds like a power steering failure to me.

You can probably overpower the steering motor even if it does not disengage. It would likely resemble a power steering failure.

Of course, had this happened to me, I probably wouldn't be looking at the display until I had the car safely stopped.

 

[Daniel in SD]

You can probably overpower the steering motor even if it does not disengage. It would likely resemble a power steering failure.

Of course, had this happened to me, I probably wouldn't be looking at the display until I had the car safely stopped.

I think it's unlikely most people could overpower the steering motor if it were resisting. Try turning the wheel while parked, the motor has more torque than that (I verified that the Model 3 has no steering lock mechanism, it turns freely when the wheels are off the ground).

 

[Supcom]

I think it's unlikely most people could overpower the steering motor if it were resisting. Try turning the wheel while parked, the motor has more torque than that (I verified that the Model 3 has no steering lock mechanism, it turns freely when the wheels are off the ground).

Turning the wheels on a stationary vehicle is much harder than on one that is moving. Having had cars with no power steering, I have experience with that.

 

[ItsNotAboutTheMoney]

Report it to Tesla.
Report it to NHTSA.

 

[Daniel in SD]

Turning the wheels on a stationary vehicle is much harder than on one that is moving. Having had cars with no power steering, I have experience with that.

Of course! I was just pointing out how much torque the power steering motor is capable of generating. It's capable of generating that much torque plus whatever additional margin is built in to the system (2x?).

 

[KArnold]

I tried to disengage the autosteer but it would not let me. Pressing up on the right lever did not work. Forcefully turning the wheel did not work. Stepping on the brake only slowed the car down but was still stuck in autosteer. The car allowed me to change lanes only when I used on the turn signal but the wheel was pretty stiff to turn.

So...AP disengagement at switch, steering, wheel, and brake all failed. Power steering itself failed. Turn signal allowed 'lane change' into the breakdown lane, which it's not supposed to be capable of. Even if you assume AP disengagement has a potential single point of failure somewhere, the steering is fully independent. An impossible lane change is just weird. So five (or three, at least) major, critical systems all failed simultaneously and without any apparent error messages?

If true I don't think I'd ever drive the car again, at least not without a technical rationale for such an event. I assume some hacker doesn't have a contract out on the OP.

This should concern us all.

 

[Supcom]

So...AP disengagement at switch, steering, wheel, and brake all failed. Power steering itself failed. Turn signal allowed 'lane change' into the breakdown lane, which it's not supposed to be capable of. Even if you assume AP disengagement has a potential single point of failure somewhere, the steering is fully independent. An impossible lane change is just weird. So five (or three, at least) major, critical systems all failed simultaneously and without any apparent error messages?

If true I don't think I'd ever drive the car again, at least not without a technical rationale for such an event. I assume some hacker doesn't have a contract out on the OP.

This should concern us all.

Multiple systems did not fail. There is almost certainly a single point failure. It could be a hardware failure in the computer or even a software failure triggered by some rare combination of factors. My money would be on a hardware failure. If software, this would likely be happening to many people.

 

[electric-donut]

If EAP was autonomous, me and my wife would be dead as well. Coming back from Ohio on I-95 in my 2018-S this past week, I was using EAP in the rain. I was closing in on a semi from the left and the semi hit a puddle and splashed water on me. EAP shut down without warning. The car was performing a left turn in the left lane when this occurred. I was able to take immediate control and keep the car in its lane.

Not being part of the FSD beta program, I can only hope that FSD software is better on the highway than EAP. I also hope there is better redundancy in the hardware/software than what I experienced. If not, the hardware/software is nowhere close to being where it needs to be.

 

[clydeiii]

I don’t think OP is suggesting AP changed lanes onto the shoulder. He’s saying AP “let” him change lanes after he used turn indicator.

Something is majorly wrong here, up to and including OP’s understanding of how AP works.

 

[dgatwood]

While on the freeway today, I tried to disengage the autosteer but it would not let me. Pressing up on the right lever did not work. Forcefully turning the wheel did not work. Stepping on the brake only slowed the car down but was still stuck in autosteer. The car allowed me to change lanes only when I used on the turn signal but the wheel was pretty stiff to turn. I had to pull over on the side of the freeway and physically get out to allow the car to reset itself back to normal.

This was a first for me after 2 years of ownership. Is there anything else I can do besides pulling over if it happens again?

Oh! You found the new murder easter egg! Congratulations!

 

[DrChaos]

Multiple systems did not fail. There is almost certainly a single point failure. It could be a hardware failure in the computer or even a software failure triggered by some rare combination of factors. My money would be on a hardware failure. If software, this would likely be happening to many people.

On the BMW i3 forums, there is a meme (I wish I could find the picture) used for many weird and otherwise inexplicable hardware failures: "It's the 12V".

 

[Twiglett]

why not call roadside assistance immediately ?

 

[DerbyCityY]

Why did you come to us with such a serious problem with your first post in the TMC?

 

[thesmokingman]

I don’t think OP is suggesting AP changed lanes onto the shoulder. He’s saying AP “let” him change lanes after he used turn indicator.

Something is majorly wrong here, up to and including OP’s understanding of how AP works.

OP has lane assist on and is confused by it.

 

[clydeiii]

Unclear since OP has owned car for two years, but doesn’t explain how often they’ve used AP.

 

[thesmokingman]

Unclear since OP has owned car for two years, but doesn’t explain how often they’ve used AP.

The turning with blinker on is a dead giveaway. This is more like trololo than real.

 

[dtdtdt]

Here’s my guess. Sometimes when I disengage AP via a drastic turning of the steering wheel, the wheel becomes stiffer to turn, until I re-engage.

Suppose this happened to the OP, except the Tesla made steering extra, extra stiff. If he didn’t actually check the display, this might make it seem that AP did not turn off and was still guiding the car. Now, suppose he has Lane Departure Avoidance set to Assist. If he tried to change lanes to pull off, but forgot to use the blinkers, his attempt to steer would be resisted by Assist and the extra stiffness. He says he was able to pull off once he used the turn signal. So, that part is explained.

 

[l1steven1l]

So it allowed an automatic lane change on to the shoulder? It's not supposed to allow lane changes on to the shoulder at all.
The car is back to normal now?
Maybe it was a temporary failure of the power steering system. That would make the steering quite stiff (and disable autosteer of course since it's the same power steering motor). I would think that would throw an error code though.

I don't have have FSD so it did not lane change for me. Turning the signal on allowed me to move the wheel enough to change lanes onto the shoulder.

 

[l1steven1l]

OP stated that neither manual steering, brake application nor upstalk would disengage AP. Assuming the OP is not making things up (this is the internet) then he appears to have a failure in the computer. I would hope it's a hardware issue as the AP disengagement software should be very well debugged!

If this happened to me, I would be very leery of driving the car before Tesla service looked at it and would have likely called roadside assistance immediately.

OP did not provide model and software version.

I have a model Y and up to date on software.

I have an appointment scheduled with tesla service so we'll see what they say.