Phantom braking so bad I want to return my car

[Fourdoor]

here is another youtube video - for the most part they are a 5 mph decrease which is fine but some from 55 -> 30

here is another:

Actually, NO... a 5 mph decrease is NOT fine. It means the car is seeing a threat that is not real and taking minor action on that imaginary threat. If it were a real threat of some kind the 5 mph drop in speed would be useless, so why is the car doing it? If I am alone in the car this kind of PB is annoying, not dangerous like the "slam on the brakes" PB (that I also experience). When I am not alone it is not just annoying it is infuriating, making my wife look over at me and ask "why did you do that"? My wife eventually accepted that it is the car doing this not me doing it, but it is still unacceptable behavior in a $60,000+ car.

Keith

 

[tm1v2]

I have received similar dismissals from Tesla Service like this. It is frustrating when they say they they have remotely inspected your car and determined everything is working as expected so you must not really be having any problem. So we’ve gone ahead and cancelled your service appointment for you

As best I can tell there's essentially nothing Tesla Service can do about software problems affecting every car.

They do have escalation routes, and if it's just your car exhibiting some weird behavior they will use them. But from Service team's perspective, there really is nothing they can do about "VO TACC sucks," this is a higher level product culture problem of pretending a theoretical driver assist future (VO TACC works everywhere so well you don't need basic CC) is here now (it's not, but all we need is to retain access to basic CC in addition to TACC).

You can see this pattern time and again with VO rain sensing, FSD marketing, AP2 launch, etc.

I'll note that even though I haven't experienced phantom braking with TACC yet (VO or otherwise), my M3P's VO TACC is clearly worse even on divided highways (all I've used it on) than AP1 or current radar-based AP2.5 or a competing EV's radar-based ACC that I recently tested. The inconsistent following distances, late braking, and excessively slow+delayed acceleration / gap closing of current VO ACC make for an unnecessarily stressful ride.

On a side note I will say Autosteer is leaps and bounds better these days than years ago. That's not affected by radar vs VO of course. Also the highways around me have gone through some repaving and have more consistent lane markings now, that helps too, but I can tell Autosteer itself is much improved. I used to HATE it for its first couple years, it was nauseating and tried to kill me. Now I'm finding myself actually liking and using it sometimes, a big change. Maybe, just maybe, Tesla will get there with VO TACC...in a couple years...

 

[bodosom]

That's not true on my 2017 Outback with Eyesight -- you can switch to "conventional cruise control" if the Eyesight system isn't working.

Well live and learn. It's on page 68 in my 2015 manual and 85 in my 2019 manual.

 

[stopcrazypp]

As best I can tell there's essentially nothing Tesla Service can do about software problems affecting every car.

They do have escalation routes, and if it's just your car exhibiting some weird behavior they will use them. But from Service team's perspective, there really is nothing they can do about "VO TACC sucks," this is a higher level product culture problem of pretending a theoretical driver assist future (VO TACC works everywhere so well you don't need basic CC) is here now (it's not, but all we need is to retain access to basic CC in addition to TACC).

You can see this pattern time and again with VO rain sensing, FSD marketing, AP2 launch, etc.

I'll note that even though I haven't experienced phantom braking with TACC yet (VO or otherwise), my M3P's VO TACC is clearly worse even on divided highways (all I've used it on) than AP1 or current radar-based AP2.5 or a competing EV's radar-based ACC that I recently tested. The inconsistent following distances, late braking, and excessively slow+delayed acceleration / gap closing of current VO ACC make for an unnecessarily stressful ride.

On a side note I will say Autosteer is leaps and bounds better these days than years ago. That's not affected by radar vs VO of course. Also the highways around me have gone through some repaving and have more consistent lane markings now, that helps too, but I can tell Autosteer itself is much improved. I used to HATE it for its first couple years, it was nauseating and tried to kill me. Now I'm finding myself actually liking and using it sometimes, a big change. Maybe, just maybe, Tesla will get there with VO TACC...in a couple years...

Yeah you put it well, I was about to say the same thing. There is absolutely nothing service can do about common software false positives, especially if you don't even have specific timestamps for them to examine (like if you activated the bug report). The only thing service can really address is if your issues stem from a hardware issue (like bad camera calibration or a defective camera, which is not impossible).

 

[stopcrazypp]

Actually, NO... a 5 mph decrease is NOT fine. It means the car is seeing a threat that is not real and taking minor action on that imaginary threat. If it were a real threat of some kind the 5 mph drop in speed would be useless, so why is the car doing it? If I am alone in the car this kind of PB is annoying, not dangerous like the "slam on the brakes" PB (that I also experience). When I am not alone it is not just annoying it is infuriating, making my wife look over at me and ask "why did you do that"? My wife eventually accepted that it is the car doing this not me doing it, but it is still unacceptable behavior in a $60,000+ car.

Keith

In terms of the effect on damage in an impact maybe 5 mph doesn't, but in terms of reaction time, it does make a significant difference, especially if the car is at the edge of its perception. A drop from 65 mph to 60 mph can give the car 2-3 more frames (given 36 fps camera pipeline) before reaching a target, which can make a huge difference.

It's the same reason why Vision is still at 80mph max vs 90 mph, that 10mph does make a difference.

 

[MikeR99]

Don't know if anyone has said it yet, and I'm not discounting how people feel when this happens, but all of this software is beta and this should be very well known to anyone even considering buying a Tesla. I knew it before I made the purchase. It's annoying and I don't like it, but it's beta. On the freeways/highways, where I believe it is intended to be used (standard autopilot anyway), it has worked almost flawlessly for me. Even FSD beta is impressive for what it is currently. I've never had an issue with TACC.

 

[tm1v2]

Don't know if anyone has said it yet, and I'm not discounting how people feel when this happens, but all of this software is beta and this should be very well known to anyone even considering buying a Tesla. I knew it before I made the purchase. It's annoying and I don't like it, but it's beta. On the freeways/highways, where I believe it is intended to be used (standard autopilot anyway), it has worked almost flawlessly for me. Even FSD beta is impressive for what it is currently. I've never had an issue with TACC.

I think this perspective would be okay if basic CC was also available alongside TACC on Autopilot cars. Which it totally could be, if Tesla wanted it to.

I don't drive much on straight 2 lane country roads these days, but I used to years ago and I would regularly use CC on them. I don't think the "Beta" label absolves anything for those who drive such roads when Tesla doesn't allow basic CC on Autopilot cars.

 

[MikeR99]

I think this perspective would be okay if basic CC was also available alongside TACC on Autopilot cars. Which it totally could be, if Tesla wanted it to.

I don't drive much on straight 2 lane country roads these days, but I used to years ago and I would regularly use CC on them. I don't think the "Beta" label absolves anything for those who drive such roads when Tesla doesn't allow basic CC on Autopilot cars.

I must have missed someone saying they wanted standard CC. It does seem like something that could be done easily. But, then again, I don't have any issues with TACC (barring some of the 15 or 20 mph curves on a 40 mph zone) and I regularly drive on some pretty narrow, twisty, two lane roads, and I don't have any phantom braking until I turn on FSD (or standard autopilot, before I got the FSD beta). I wonder how much variability there is between different cars?

 

[TheSlenderman6]

Yeah you put it well, I was about to say the same thing. There is absolutely nothing service can do about common software false positives, especially if you don't even have specific timestamps for them to examine (like if you activated the bug report). The only thing service can really address is if your issues stem from a hardware issue (like bad camera calibration or a defective camera, which is not impossible).

This is only partially correct. Tesla does not want to acknowledge something isn't right. Tesla's position is that Vision only is the way we're going. If you do not have the FSD Beta, and have Radar in your car, Radar works and does not have 99% of these issues. The solution is to add Radar back into the car.

Anyone that has this issue, and Tesla is dragging feet: review lemon law in your state. Contact them via certified mail to notify of their last chance to fix your car. Given the chance to spend $1,000 on installing radar in your car vs refunding/replacing the car - you might just get lucky.

 

[drtimhill]

If anyone has an idea how to get through to Tesla that this needs more attention, Ill be happy to spearhead an effort.

It IS getting more attention .. it's called FSD beta.

 

[MP3Mike]

It IS getting more attention .. it's called FSD beta.

Yep. The FSD Beta 10.7 release notes show that it appears that they have made significant progress on the "false slowdowns":

Now the question is when will those updates make it in to the public Tesla Vision releases?

 

[stopcrazypp]

This is only partially correct. Tesla does not want to acknowledge something isn't right. Tesla's position is that Vision only is the way we're going. If you do not have the FSD Beta, and have Radar in your car, Radar works and does not have 99% of these issues. The solution is to add Radar back into the car.

Anyone that has this issue, and Tesla is dragging feet: review lemon law in your state. Contact them via certified mail to notify of their last chance to fix your car. Given the chance to spend $1,000 on installing radar in your car vs refunding/replacing the car - you might just get lucky.

What you describe are legal matters, which once again service can do nothing to help you with. Any lemon law claim would be handled with Tesla corporate or the processing center of your state. Below is the generic list for where to send your lemon law claim, but you should check the documents in your online account, which should have the one most applicable to your state.

https://www.tesla.com/sites/default/files/blog_attachments/ownersrightsnotification.pdf

Also it is a known fact that having the radar does not eliminate phantom braking. Phantom braking existed in all versions of Teslas and is also warned about in the manual. So I doubt you would have a case to ask them to install radar, especially without documented evidence.

 

[TheSlenderman6]

What you describe are legal matters, which once again service can do nothing to help you with. Any lemon law claim would be handled with Tesla corporate or the processing center of your state. Below is the generic list for where to send your lemon law claim, but you should check the documents in your online account, which should have the one most applicable to your state.

https://www.tesla.com/sites/default/files/blog_attachments/ownersrightsnotification.pdf

Also it is a known fact that having the radar does not eliminate phantom braking. Phantom braking existed in all versions of Teslas and is also warned about in the manual. So I doubt you would have a case to ask them to install radar, especially without documented evidence.

Radar does not eliminate every instance of false braking. However, it does drastically reduce the number of false positives.

A friend has a MYLR with Radar, 2021. He and I are setting up some times to make the same run 2-3 minutes apart to determine the amount of differences between the two.

 

[linux-works]

its not super clear, but it seems by so many reports that vision-only cars NEVER do better than radar cars.

at best, they do almost as well, in good daylight, no rain, etc.

am I wrong? are there documented tests or even anecdotes of where vision-only outperformed a radar car in the same conditions?

 

[thesmokingman]

its not super clear, but it seems by so many reports that vision-only cars NEVER do better than radar cars.

at best, they do almost as well, in good daylight, no rain, etc.

am I wrong? are there documented tests or even anecdotes of where vision-only outperformed a radar car in the same conditions?

I don't think it really matters either way as nothing that's debated ad nasuem on a forum will sway Musk's pov. When ppl bring up radar if its some panacea it's kind of BS, we're here because of radar's limits to begin with. And thus here we are in vision only, there's no sense in looking bad with rose colored glasses at radar.

 

[linux-works]

careful, be more precise.

you probably meant to say that we're here because of tesla's implementation limits using fusion of both radar+vision.

its a bit of a stretch to say: "elon's boys gave up, therefore everyone else should/will since it wont pay off". I just dont believe that and I dont think the industry believes that, either.

tesla is just a small part of the industry. others are also playing the lane-following game; this is not a tesla only trick and as tesla's code base and approach ages, their advantage gets less and less. and they wont be able to buy or plugin other offerings (if they are better) since tesla rolled their own everything, thus making it so much harder to integrate standard parts in their designs (ie, external adas systems).

tesla did get the world's attention, but now that everyone wants in, tesla's edge is going to lessen over time.

the cheapness of reducing sensor richness is not going to help, in the short or long run. I think its a huge error. time will tell.

 

[MP3Mike]

its not super clear, but it seems by so many reports that vision-only cars NEVER do better than radar cars.

at best, they do almost as well, in good daylight, no rain, etc.

am I wrong? are there documented tests or even anecdotes of where vision-only outperformed a radar car in the same conditions?

Well Tesla has shown that Tesla Vision responds better to partial lane obstructions than their radar solution does. I don't know that anyone has reported that themselves, but I also don't think we have seen a Tesla Vision vehicle run into the back of a firetruck.

So, yeah, I think it is better in some situations. (And will get even better rapidly.)

 

[drtimhill]

careful, be more precise.

you probably meant to say that we're here because of tesla's implementation limits using fusion of both radar+vision.

its a bit of a stretch to say: "elon's boys gave up, therefore everyone else should/will since it wont pay off". I just dont believe that and I dont think the industry believes that, either.

tesla is just a small part of the industry. others are also playing the lane-following game; this is not a tesla only trick and as tesla's code base and approach ages, their advantage gets less and less. and they wont be able to buy or plugin other offerings (if they are better) since tesla rolled their own everything, thus making it so much harder to integrate standard parts in their designs (ie, external adas systems).

tesla did get the world's attention, but now that everyone wants in, tesla's edge is going to lessen over time.

the cheapness of reducing sensor richness is not going to help, in the short or long run. I think its a huge error. time will tell.

The trouble is, this is the old "once the REAL guys move in next year, Tesla is toast" argument that has been stated again and again, with the date (conveniently) moving forward a year every time the old deadline for the "big takeover" date comes and goes (for at least the last 3-4 years).

And, if you hadn't noticed, Tesla is now rather a large part of the industry, and, so far, vertical integration seems to be a rather successful strategy. As you note, time will tell, but based on the current track record, time has "told" for Tesla.

Actually, in the very long-term, as the various EV technologies do become commoditized, things most probably will pivot to a parts-bin mindset, but that is decades away imho. Right now though, things in the EV industry are much like they were for the computer industry in the 1960's to 1970's, with the rise of a single dominant power, IBM, who beat everyone else by a direct marketing model and heavy vertical integration with intense investment in the base underlying technologies in-house. Sound familiar?

 

[Fourdoor]

Well Tesla has shown that Tesla Vision responds better to partial lane obstructions than their radar solution does. I don't know that anyone has reported that themselves, but I also don't think we have seen a Tesla Vision vehicle run into the back of a firetruck.

So, yeah, I think it is better in some situations. (And will get even better rapidly.)

You have never seen a Tesla radar equipped car where the driver was paying attention run into the back of a fire truck either. People wanting the car to "do it all" while they pay no attention are the problem. I want a car that centers itself in the lane of travel and maintains speed in that lane unless the car in front of me slows down. I want it to do this while I look out the windshield / windows monitoring for dangers that are not directly ahead of me. Wanting the car to monitor things outside of those parameters is what causes the problems. Oncoming traffic in the other lane? I want the car to ignore that. Speed limit signs or lack of signage? I want the car to ignore that. On rural highways you may not see a speed limit sign for 15 or more miles... if a Tesla travels more than 5 miles without seeing a speed limit sign it defaults to 45 mph speed limit... in my area by law a rural highway without speed limit signs by default has a speed limit of 65 mph. The car should not be able to arbitrarily limit the speed the owner of the car can select on the cruise control system. A 15 mph differential between what I can legally drive on a road and what the car will allow me to set the cruise control too (5 mph higher than it thinks the speed limit is) is ridiculous. I think the move to vision only was a mistake that will be mitigated by software improvements. If Tesla just said "screw it, here is adaptive cruise and lane centering and that is all you get on basic auto pilot" they could implement that and it would be virtually perfect within weeks, possibly days in the vision only cars. Trying to teach the car to save everyone's ass when they drive without paying any attention will take years. That is fine, but it should be limited to the FSD cars, while the rest of us get functional cruise control and lane centering.

Keith

 

[Fourdoor]

The trouble is, this is the old "once the REAL guys move in next year, Tesla is toast" argument that has been stated again and again, with the date (conveniently) moving forward a year every time the old deadline for the "big takeover" date comes and goes (for at least the last 3-4 years).

And, if you hadn't noticed, Tesla is now rather a large part of the industry, and, so far, vertical integration seems to be a rather successful strategy. As you note, time will tell, but based on the current track record, time has "told" for Tesla.

Actually, in the very long-term, as the various EV technologies do become commoditized, things most probably will pivot to a parts-bin mindset, but that is decades away imho. Right now though, things in the EV industry are much like they were for the computer industry in the 1960's to 1970's, with the rise of a single dominant power, IBM, who beat everyone else by a direct marketing model and heavy vertical integration with intense investment in the base underlying technologies in-house. Sound familiar?

Nobody I have seen in this thread is claiming that any other manufacturer is a "Tesla killer"... we all want Tesla to thrive. But keeping the blinders on about other companies with lane centering and adaptive cruise control systems that work properly, and pretending that Tesla's system is as good as (let alone better than) other manufacturers will hurt Tesla in the long run. Tesla is pursuing edge case self driving while ignoring basic adaptive cruise control and lane centering. This isn't customer driven, it is an obsession of the founder of the company with massive amounts of ego invested in being right about not needing anything other than cameras, and being the first to have real self driving cars.

As to your computer industry analogy, look at the flip side. Computers (cars) have existed for a century and then a small innovative company comes alone that makes them better, faster, and more efficient. Do the existing computer companies go belly up and die? Or do they adapt? Tesla will never be "killed" by the legacy auto makers... but in reality Tesla will not "kill" the legacy auto makers either. It most likely will become just another car maker amongst the crowd.

Keith