Phantom Braking

[AndreP]

Everyone take a step back and look at this logically. We live in a free market. If there were enough people unhappy with the performance of their Tesla purchase, they would not make a repeat purchase, or there would be a flood of aftermarket Tesla's available which would drive down the used market price. Neither seems to be the case. According to Forbes, Tesla is #1 for repeat buyers (as of December 2020). According to InsideEV and Kelly Blue Book, Tesla's resale value is highest in its class (compared to Audi, BMW, etc). This tells us that people are happy with their purchase of Tesla. This also contradicts views that Tesla creates inferior products purely for the benefit of shareholders. Many non-professional investors (casual investors) purchase stock from a company they feel positive about, which means there are likely many Tesla stockholders who also own a Tesla car and purchased that stock based on their experience of the vehicle.

Are there problems with the software, fit and finish, styling, or drive experience? For some people, yes. However, the above statistics show that they are either willing to look past those negatives, not experiencing those negatives, able to have Tesla repair/modify enough to satisfy their needs, or not in a position to extricate themselves from the situation.

Try to keep an objective view, look at the free market statistics, weigh your needs with what the market provides, and make the best choice for yourself. If you are very upset (which it seems some of you are), you have every right to be. However, analyze your experience and make adjustments accordingly. Is Tesla not repairing your vehicle or updating their software fast enough to match your needs? Then look at the after-market value for your Tesla and consider a private party sale which may surprise you and allow you to come out of the experience with minimal downside. Then research and purchase a vehicle from another manufacturer that does meet your needs.

There are plethora potential influences here. Tesla pioneered the consumer EV and still dominates that corner of the auto industry, so there haven't been many alternatives and those alternatives are still lacking. And that could be contributing to repeat purchases.

Ownership in the company could make people more inclined to be repeat buyers, especially pertaining to those who have paid for FSD multiple times on many new Tesla vehicles they've churned. I have a hard time imagining why people would pay for it multiple times otherwise. I'd wager that shareholders are also more inclined to endure the Beta testing with the hope that their efforts bring the company closer to realizing revenue and profits from the service, thus benefitting their investment.

Employee ownership is a great way to incentivize performance in a company and this feels like a 4D-chess evolution of that concept, and Tesla isn't alone in it. Rivian set aside 7% of their IPO share allocation for people who pre-ordered their vehicles, and it doesn't take a SpaceX scientist to see why they might do that aside from just wanting to benefit their early adopters.

All of this is super fun to watch, contemplate, and discuss

 

[sleepydoc]

my Lexus got totaled

You forgot to slow down when driving past a truck?

Regarding your “as advertised “ I don’t remember seeing any advertising claiming anything they do is better than anybody. It is not even mentioned in their Order options.

'as advertised' isn't a difficult concept for cruise control - you set the speed, the car maintains the speed. Bouncing around from 55 to 45 to 55 to 50 to 55 to 35 is not expected or acceptable behavior.

Something I didn't realize until after I bought my MY is that TACC is considered beta. You can fault me for not checking before I purchased, however the manual wasn't even available when I ordered. The feature list on the web site was (and still is) very abbreviated. (among other things not listed are power windows, auto dimming rear view mirror, power trunk, etc) Since cruise control is a standard feature on every car sold I incorrectly (perhaps naively) assumed that Tesla would have feature parity with a Honda Civic or Toyota Prius.

Many others have mentioned this but I'll reiterate it, if Tesla is going to make TACC a 'beta' feature then why can't it be disabled to at least give 'normal' cruise control?

 

[enemji]

You forgot to slow down when driving past a truck?


'?

sht. You got me there I told you u were smart, didn’t I?

 

[sleepydoc]

Everyone take a step back and look at this logically. We live in a free market. If there were enough people unhappy with the performance of their Tesla purchase, they would not make a repeat purchase, or there would be a flood of aftermarket Tesla's available which would drive down the used market price. Neither seems to be the case. According to Forbes, Tesla is #1 for repeat buyers (as of December 2020). According to InsideEV and Kelly Blue Book, Tesla's resale value is highest in its class (compared to Audi, BMW, etc). This tells us that people are happy with their purchase of Tesla. This also contradicts views that Tesla creates inferior products purely for the benefit of shareholders. Many non-professional investors (casual investors) purchase stock from a company they feel positive about, which means there are likely many Tesla stockholders who also own a Tesla car and purchased that stock based on their experience of the vehicle.

Are there problems with the software, fit and finish, styling, or drive experience? For some people, yes. However, the above statistics show that they are either willing to look past those negatives, not experiencing those negatives, able to have Tesla repair/modify enough to satisfy their needs, or not in a position to extricate themselves from the situation.

Try to keep an objective view, look at the free market statistics, weigh your needs with what the market provides, and make the best choice for yourself. If you are very upset (which it seems some of you are), you have every right to be. However, analyze your experience and make adjustments accordingly. Is Tesla not repairing your vehicle or updating their software fast enough to match your needs? Then look at the after-market value for your Tesla and consider a private party sale which may surprise you and allow you to come out of the experience with minimal downside. Then research and purchase a vehicle from another manufacturer that does meet your needs.

@AndreP is very correct. For much of Tesla's history it has had no real competition in the EV market. Even now the competition is just starting to come. When you purchased a Tesla in 2014 they were ahead of the curve and had features no one else had. You were also an early adopter so you were willing to put up with the quirks and foibles of a new car company creating a new market. We're 10 years on now. Tesla is no longer a new company, there are other options available with more on the way and other companies have caught up or are catching up in technology.

Newer Tesla buyers are less likely to be early adopters and more likely to be the 'average' buyer that wants a car that works. The free market is also not quite so free right now because of supply chain issues, etc.

Tesla needs to keep its focus on improving the car and improving the quality. If they keep screwing with customers, messing up the UI and delivering a sub-par product, the good will of the early adopters will disappear as will their reputation and they risk becoming another Blackberry.

 

[enemji]

RTFM

 

[sleepydoc]

TACC matches is supposed to match the speed of your car to that of the surrounding traffic.

Fixed that for you.

I took a 150 mile trip today. The first half of the trip was a 2 lane state highway, the 2nd half was an interstate. On the interstate it was practically perfect. On the 2 lane highway it would randomly slow down. Most of the time it was just 2-3 MPH. One time it was 10 MPH. The vast majority of the time there were no other cars in sight. The times I passed cars, trucks and semis going the opposite direction it generally did not slow at all. The time it did slow it typically did so a few hundred feet early then sped up as we passed.

It is quite clarifying that it does not just adapt to the vehicle(s) ahead of you.

??? Not sure what you mean here, but if one lane is going 40 MPH and the lane you're in is going 50 MPH the proper speed is not 40 mph unless you're merging. If the cars in the oncoming/opposite lane are approaching at 55 mph, the correct action is not slowing down.

 

[enemji]

Fixed that for you.

I took a 150 mile trip today. The first half of the trip was a 2 lane state highway, the 2nd half was an interstate. On the interstate it was practically perfect. On the 2 lane highway it would randomly slow down. Most of the time it was just 2-3 MPH. One time it was 10 MPH. The vast majority of the time there were no other cars in sight. The times I passed cars, trucks and semis going the opposite direction it generally did not slow at all. The time it did slow it typically did so a few hundred feet early then sped up as we passed.

??? Not sure what you mean here, but if one lane is going 40 MPH and the lane you're in is going 50 MPH the proper speed is not 40 mph unless you're merging. If the cars in the oncoming/opposite lane are approaching at 55 mph, the correct action is not slowing down.

I deleted my post as it was inaccurate.

 

[sleepydoc]

View attachment 766194

RTFM

"TACC may... brake when not required" that pretty much says it all, doesn't it? That's the entire point of this thread. Braking when not required means it's not working as designed/expected. The entire goal of adaptive cruise control is to brake when required and not brake when not required.

"this can be caused by closely following a vehicle ahead" - except if you're using TACC, the car is setting the following distance, not you, so it still comes back to the car, doesn't it? Besides, none of the cases people are complaining about are cases where there's an obstacle.

"Detecting vehicles or objects in adjacent lates..." - see my last point. And the point I keep making about every other ****ing car on the road with adaptive cruise control doing this without a problem. Except Tesla's 'advanced optical tracking technology.'

Your post implies that putting it in the manual makes the fact that their system sucks acceptable. It's still not acceptable, It just means they know it sucks.

 

[AndreP]

I think the real problem is that people are buying into these systems with unreasonably high expectations, likely because the capabilities have been massively oversold. If people understood all the warnings and limitations described in the manuals, they would likely go in with a much more realistic opinion of the functions.

People don’t buy Autopilot thinking that they're told to not use it on winding roads with sharp curves, that it will occasionally brake when detecting vehicles in adjacent lanes and especially in curves, and that it won’t adjust speed to match driving conditions — these are actual warnings in the manual. And the limitations seem incongruent with a system called “Autopilot”, so who can blame people for being disappointed when they use the system in an area and find that it falls short of their expectations?

 

[enemji]

I think the real problem is that people are buying into these systems with unreasonably high expectations, likely because the capabilities have been massively oversold. If people understood all the warnings and limitations described in the manuals, they would likely go in with a much more realistic opinion of the functions.

People don’t buy Autopilot thinking that they're told to not use it on winding roads with sharp curves, that it will occasionally brake when detecting vehicles in adjacent lanes and especially in curves, and that it won’t adjust speed to match driving conditions — these are actual warnings in the manual. And the limitations seem incongruent with a system called “Autopilot”, so who can blame people for being disappointed when they use the system in an area and find that it falls short of their expectations?

I guess some are thinking Elon can land rockets so adaptive cruise should be easy to do. Maybe it is. Maybe it is not as they may be attempting to do something more than just maintain a speed and slow down to match the speed of the vehicle ahead of you.

 

[Knightshade]

"TACC may... brake when not required" that pretty much says it all, doesn't it? That's the entire point of this thread. Braking when not required means it's not working as designed/expected. The entire goal of adaptive cruise control is to brake when required and not brake when not required.

Except, as pointed out in the last 9736 threads about this, every brand of car has the same disclaimers in the manual because every brand has unintentional braking

"Detecting vehicles or objects in adjacent lates..." - see my last point. And the point I keep making about every other ****ing car on the road with adaptive cruise control doing this without a problem. Except Tesla's 'advanced optical tracking technology.'

This is outright false.

every brand has this issue- which is why every brand has similar warnings in their manuals for these features.

It's inherent to the system

And the limitations seem incongruent with a system called “Autopilot”, so who can blame people for being disappointed when they use the system in an area and find that it falls short of their expectations?

Having actually flown aircraft, and knowing what the word autopilot actually means- I can blame them.


Especially when other brands use names like "copilot" which, unlike autopilot, is the name of something that can operate the vehicle without extra help from a human (but of course the car version does not)

 

[OncomingStorm]

I think the real problem is that people are buying into these systems with unreasonably high expectations, likely because the capabilities have been massively oversold. If people understood all the warnings and limitations described in the manuals, they would likely go in with a much more realistic opinion of the functions.

People don’t buy Autopilot thinking that they're told to not use it on winding roads with sharp curves, that it will occasionally brake when detecting vehicles in adjacent lanes and especially in curves, and that it won’t adjust speed to match driving conditions — these are actual warnings in the manual. And the limitations seem incongruent with a system called “Autopilot”, so who can blame people for being disappointed when they use the system in an area and find that it falls short of their expectations?

agree 100% however Autopilot in planes can't take off or land or even do complex maneuvers, its pretty much for flying straight for a long time with no obstacles (kind of like an interstate highway), imho its a perfect name. and honestly i'm not even sure its fair to say it's Tesla that "oversells" their product, really its the userbase (or mis-userbase lol) that really gives people the wrong idea of what it can and can't do.

really at the end of the day TACC / AP could perform flawlessly 99.9% of the time and there would still be people who freak out about it. its like the idea of adding a "dumb" cruise control, aside from the obvious issues of people using it thinking they are using TACC and get into accidents, my guess is there would be far more people bitching about confusing the options and causing actual accidents than the handful of people who complain about so called phantom breaking (which i still maintain those who call it phantom simply aren't paying attention and missed the thing that caused it, personally i'd prefer my car to err on the side of caution but thats just me).

 

[AndreP]

Except, as pointed out in the last 9736 threads about this, every brand of car has the same disclaimers in the manual because every brand has unintentional braking

This is outright false.

every brand has this issue- which is why every brand has similar warnings in their manuals for these features.

It's inherent to the system

Having actually flown aircraft, and knowing what the word autopilot actually means- I can blame them.

Especially when other brands use names like "copilot" which, unlike autopilot, is the name of something that can operate the vehicle without extra help from a human (but of course the car version does not)

Any brand doing this type of stuff should be held accountable, and I think there‘s plenty of it to go around. The whole AV space is full of massive promises and very little delivery.

I wouldn’t pay one additional red cent for the Lane Keeping technology on my work rental 2021 Ford F-150, even the auto high beams are flawed in many situations.

 

[S4WRXTTCS]

I think the real problem is that people are buying into these systems with unreasonably high expectations, likely because the capabilities have been massively oversold. If people understood all the warnings and limitations described in the manuals, they would likely go in with a much more realistic opinion of the functions.

People don’t buy Autopilot thinking that they're told to not use it on winding roads with sharp curves, that it will occasionally brake when detecting vehicles in adjacent lanes and especially in curves, and that it won’t adjust speed to match driving conditions — these are actual warnings in the manual. And the limitations seem incongruent with a system called “Autopilot”, so who can blame people for being disappointed when they use the system in an area and find that it falls short of their expectations?

I think there are a few problem.

1.) Much (but, not all) of the phantom braking that occurs with AP also occurs with TACC. This is a mistake because TACC should be more about smoothness, and less about safety. It doesn't need to be as sensitive as AP because the driver is fully engaged. I don't want TACC to slow down because it thinks the car in the lane over might come across. I also have to make sure to turn off traffic light response or it will sometimes get fooled by flashing lights on nearby roads.

2.) Tesla absolutely refuses to make any effort whatsoever when it comes to allowing the user to report map/navigation issues. Most of the phantom braking I encountered before Tesla Vision was due to the car suddenly thinking it was on a different road type because of maps issues. It was pretty consistent about exactly where it would do it. We would have significantly less AP braking events if users could report errors.

3.) Tesla vision when released to my car was an absolute pile of garage when it came to phantom braking on the freeway. I wasn't even that early as I had 10.3.1. It has improved a lot from 10.3.1 to 10.9. I believe Tesla didn't do a good enough job validating Tesla vision before they removed the radar from vehicles. The survey on this TMC section also indicates that Tesla Vision was/is problematic with phantom braking.

So in summary I do expect phantom braking to happen on a rare occasion, but I expect it to be an extremely rare event with TACC. I had a 2015 Model S that very rarely had phantom braking on TACC, and my Jeep Wrangler Unlimited has never had a single phantom braking event on the freeway. My brothers Subaru with eyesight doesn't have a phantom braking issue and neither does his Honda Ridgeline.

0 phantom braking events with something like FSD beta would be an insane ask, but <1 phantom braking event per 5K miles on average under TACC isn't too much to ask.

 

[nvx1977]

I think there are a few problem.

1.) Much (but, not all) of the phantom braking that occurs with AP also occurs with TACC. This is a mistake because TACC should be more about smoothness, and less about safety. It doesn't need to be as sensitive as AP because the driver is fully engaged. I don't want TACC to slow down because it thinks the car in the lane over might come across. I also have to make sure to turn off traffic light response or it will sometimes get fooled by flashing lights on nearby roads.

2.) Tesla absolutely refuses to make any effort whatsoever when it comes to allowing the user to report map/navigation issues. Most of the phantom braking I encountered before Tesla Vision was due to the car suddenly thinking it was on a different road type because of maps issues. It was pretty consistent about exactly where it would do it. We would have significantly less AP braking events if users could report errors.

3.) Tesla vision when released to my car was an absolute pile of garage when it came to phantom braking on the freeway. I wasn't even that early as I had 10.3.1. It has improved a lot from 10.3.1 to 10.9. I believe Tesla didn't do a good enough job validating Tesla vision before they removed the radar from vehicles. The survey on this TMC section also indicates that Tesla Vision was/is problematic with phantom braking.

So in summary I do expect phantom braking to happen on a rare occasion, but I expect it to be an extremely rare event with TACC. I had a 2015 Model S that very rarely had phantom braking on TACC, and my Jeep Wrangler Unlimited has never had a single phantom braking event on the freeway. My brothers Subaru with eyesight doesn't have a phantom braking issue and neither does his Honda Ridgeline.

0 phantom braking events with something like FSD beta would be an insane ask, but <1 phantom braking event per 5K miles on average under TACC isn't too much to ask.

wide variation of experiences. I very rarely have any sort of unexpected braking event on the highway with FSD beta. People who glorify radar, I can assure you, it caused phantom braking with overpasses and signs on the highway. Not sure why people want to go back to that. It doesn't happen anymore with my car.

It's not all rainbows for me. NoA has never worked well when attempting to take off-ramps. Over 50% of the time, it doesn't work and goes by the exit unless I intervene. Yet some people think NoA is amazing. *shrug*

 

[RowdyMY]

Any brand doing this type of stuff should be held accountable, and I think there‘s plenty of it to go around. The whole AV space is full of massive promises and very little delivery.

I wouldn’t pay one additional red cent for the Lane Keeping technology on my work rental 2021 Ford F-150, even the auto high beams are flawed in many situations.

Not being sarcastic or disrespectful here, as I tend to agree with your arguments.

Isn’t “I wouldn’t pay one additional red cent” a textbook definition of accountability in a free market? We shouldn’t be regulating every in and out of a company that implements a sub-standard iteration of a product unless it actually proves to be inherently dangerous, which this endeavor hasn’t panned out to be.

If the software sucks, the economy will vote with its feet. Not like this is a gov subsidized effort, in so far as the car division is concerned anyway.

 

[sleepydoc]

Except, as pointed out in the last 9736 threads about this, every brand of car has the same disclaimers in the manual because every brand has unintentional braking

False - I haven’t reviewed every other brand of car but our Subaru Forester has no such warning. Beyond the warnings, I care about the level of function and accuracy. See below.

This is outright false.

every brand has this issue- which is why every brand has similar warnings in their manuals for these features.

It's inherent to the system

Again, false. I’ve driven several other cars with adaptive cruise control and had exactly zero problems with phantom braking. Zero. nada. Zilch. Null. Nothing. 1/∞. Also see @S4WRXTTCS’s experience above.

Trying to claim ‘inherent to the system’ is simply minimizing and covering up the extent of the problems with Tesla’s adaptive cruise control. I don’t expect prefection. I do expect performance comparable to other cars on the market. Like @S4WRXTTCS says, 1 event per 5k miles is certainly reasonable. Heck, even 1 event per 1000 miles would be a significant improvement.

 

[AndreP]

Not being sarcastic or disrespectful here, as I tend to agree with your arguments.

Isn’t “I wouldn’t pay one additional red cent” a textbook definition of accountability in a free market? We shouldn’t be regulating every in and out of a company that implements a sub-standard iteration of a product unless it actually proves to be inherently dangerous, which this endeavor hasn’t panned out to be.

If the software sucks, the economy will vote with its feet. Not like this is a gov subsidized effort, in so far as the car division is concerned anyway.

I’d agree in general as well, although this implementation is a bit different from
what people are experiencing with Autopilot. Lane Keeping here is entirely on the side of Cruise Control, and not using it has no impact on the driving experience otherwise. It does some odd stuff but generally just doesn’t add any value, I’m not even sure what it costs but just know it wouldn’t be worth additional money to me.

I don’t think the regulators should take any of Tesla’s tech off the roads though and don’t think they would, but they can help force the companies to work on issues that might not be at the top of their priority list despite negatively impacting users. If my truck required this Lane Keeping active with Cruise Control and added any significant stress, I’d be requesting a different truck from the rental company because cruise control is absolutely essential for my ridiculous highway commutes.

If someone can provide a system that will allow me to check out during 10hr drives on major divided highways, I’m all over that.

 

[sleepydoc]

wide variation of experiences. I very rarely have any sort of unexpected braking event on the highway with FSD beta. People who glorify radar, I can assure you, it caused phantom braking with overpasses and signs on the highway. Not sure why people want to go back to that. It doesn't happen anymore with my car.

It's not all rainbows for me. NoA has never worked well when attempting to take off-ramps. Over 50% of the time, it doesn't work and goes by the exit unless I intervene. Yet some people think NoA is amazing. *shrug*

One minor technical correction - currently my understanding is Tesla uses autopilot on the highways and FSD on other roads.

I’ll also clarify - our Forester manual does state that winding roads can cause an issue for its adaptive cruise control and that’s a shortcoming I can understand. The overwhelming majority of phantom braking events I experience can’t be explained by this. (Like events on an empty straight road)

 

[hatch]

TACC should be more about smoothness, and less about safety.

I believe i'm about to get my car soon. So i will get to check out this problem - or lack of - myself. I saw this comment above and a light bulb went off. I just had flashbacks of me driving on the 24 hr demo Model Y and never really being smooth.

I have a theory why Tesla is more jerky than any other car. It's got to be the fixed regen and them not being concerned at all about ever varying the level of regen. My corolla never has a phantom slowdown when i'm on TACC. Because it's an ICE car and it has no regen. It lets off the gas plenty of times, but that just makes it coast - nice and smooth. If it needs to slow down, it applies the brakes smoothly. There is no "coast" on a Tesla, outside of shifting to neutral.

Other EV's have varying regen. I think i even saw the ioniq 5 has dynamic regen just for the purpose of cruise control. I'll bet other EV's vary the regen as well and can coast if they want.

So i don't think a Tesla can coast, and I'll bet it can't even use the physical brakes during a normal cruise cycle. If it does use the brakes, it would be emergency braking as a part of accident avoidance. They could certainly program it to do anything they want. But if they strongly hold onto the stubborn view that the regen level should never change, then TACC will never get any better. And we KNOW how stubborn they are.

I'll bet if you set regen to zero, or really low, and also used the physical brakes lightly in some scenarios, they could make it really smooth on TACC.

I completely agree that the goal of TACC should be comfort. I don't believe great comfort can ever be achieved with fixed heavy regen.

I am SOO going to get the S3XY buttons for the purpose of varying regen. Whether TACC will override the regen i set - well that's another story.