Forums

Tesla's accelerating unexpectedly? has this happened to you?

Tesla's accelerating unexpectedly? has this happened to you?

Never heard of this? has anyone experienced this? I know I have not. you would think we would have heard of this on the forums or?

https://www.reuters.com/article/us-tesla-probe/u-s-to-review-petition-se...

EVRider | 2020年1月17日

There have been a few allegations of unintended acceleration in the forums. How does the investigation of Tesla vehicles compare with other automakers?

The vast majority of unintended acceleration incidents (for any automaker) turn out to be user error.

Magic 8 Ball | 2020年1月17日

'....consumer complaints to NHTSA involving 123 unique vehicles. The reports include 110..."

But honestly I did not cause the crash, the car had a mind of its own and caused the crash. I completely forgot there was brake pedal, it all happened so fast...

Magic 8 Ball | 2020年1月17日

'...consumer complaints to NHTSA involving 123 unique vehicles. The reports include 110 crashes'

Joshan | 2020年1月17日

The one time I saw this I can explain.

If you are using AP and you take it out of AP by using the steering wheel. I had missed that fact and when I let off the go pedal at a red light and the car in front of me moved, the tesla took off!!

Caught me off guard, but it was my fault for still being in TACC and not realizing it.

howard | 2020年1月17日

"The vast majority of unintended acceleration incidents (for any automaker) turn out to be user error."

Love blanket statements!

I had Ford buy back a 7.3 Powerstroke because it would randomly accelerate full throttle when starting to move from a stop until it rev limited. The scariest thing I have experienced. Until it happens to you, you have no idea. They replaced 3 fuel control modules in a row and then bought it back. Turns out Ford was quietly replacing a defective upline module that was damaging the fuel control module. Not everything is user error. It can be an actual defective part.

Just_Ted | 2020年1月17日

This is going to be my go to excuse/reason/man-splanation the next time my wife criticizes my overly aggressive take off from a stop sign....

"But sweety... this must be the unintended acceleration they were talking about... right?"

Bighorn | 2020年1月17日

Tesla can confirm pedal confusion because the sensors log it. No verified case of unintended acceleration due to equipment problems, unlike Toyota which caused multiple deaths including Lexus.

Madatgascar | 2020年1月17日

These claims seem to spike after sudden unexpected acceleration of TSLA.

Rikki-Tikki-Tavi | 2020年1月17日

Haha

Ron.Olsberg | 2020年1月17日

If using Autosteer and you take over control, I believe cruise control is still active or a least it used to be. I was surprised by this the first time or two but I would not consider it the cars fault. Maybe a message on the screen, there might even be one, and a notification sound indicating cruise is still active would help the green horns? Bottom line with the TACC the car should not rear end a car and the brake pedal STILL works.

wayne | 2020年1月17日

This may have happened to me once. I was turning into my alley with foot off the pedals and the car started accelerating rather than slowing. The acceleration was mild and stopped when I pressed the brake pedal. It has not happened again.

Lorenzryanc | 2020年1月17日

Every story I've read the past few years is a person hitting the wrong pedal. This car has more instant torque then people are used to so there's no room for error. 1 second in an ICE car causes engine to rev and car to move a few inches. With these cars, you're in a brick wall. I hope the records will prove every one of these cases. At least we HAVE records to look at. I also agree these stories come up when stock spikes, its a predictable pattern. Also, I've been waiting on Musk to tweet something crazy, but he's been good :)

M3D | 2020年1月17日

+1 Joshan I have experienced this as a result of not knowing that cruise (TACC) was still enabled - then the road opening up ahead. It has happened to me much more in this car because I use Autopilot much much more than I ever used plain old cruise control.

Harriscott | 2020年1月17日

In response to the Audi unintended-acceleration scare back in the '80s, cars now require a foot on the brake before shifting out of Park. (funny how this problem never seems to happen with manual transmission cars...). The members of the Audi Victims Network were outraged by the implication, but it saved many lives.

I want my car to defend against my mistakes, not thumb its nose at me when I make one :-)

apodbdrs | 2020年1月17日

In 19 months of driving my Model 3 I have never experienced unintended acceleration by the car. Now that is not to say, I didn't made a mistake.

Some new drivers just need to go slower, test driving using AP, adjust and understand AP to gain confidence in this new way of driving. I can see how some may think the car accelerated in an unintended fashion, especially when they don't pay attention to the speed set up, they can control with the scroll wheel.

I feel very safe using autopilot on both highway and street driving, just have to still stay alert.

M3phan | 2020年1月17日

@ Madatgascar, my thought exactly. Not a coincidence. Another transparent attempt to (1) excuse ones own poor driving, and (2) attack Tesla after a stunning quarter.
Victim mentality on display.

lbowroom | 2020年1月17日

"It can be an actual defective part"

Yes, of course it CAN be. Especially with a diesel. Likelihood this is a problem with Tesla? Not much.

https://en.wikipedia.org/wiki/Diesel_engine_runaway

bjrosen | 2020年1月17日

Turning on navigate on auto pilot on a street causes unintended acceleration. If you are traveling below the speed limit it immediately accelerates to to the limit and it does it way to quickly. I find thais terrifying so I immediately disable it. Doubt that this is what they are talking about but it could be, if people are engaging auto pilot in the wrong places you would get this effect.

Xerogas | 2020年1月17日

@PTdenver: it’s always operator error https://www.caranddriver.com/features/a15125313/its-all-your-fault-the-d...

Safety tip for anyone who thinks their car won’t stop:

1) lift your foot off the pedal, EVEN IF YOU ARE “SURE” YOU ARE PRESSING ON THE BRAKE
2) carefully press your foot on the brake. The car will stop, always, even if somehow the accelerator is still pressed.
3) remember, brakes are always more powerful than motors. The car will always stop

Bighorn | 2020年1月17日

@Xerogas
The view from 2011 has been debunked. Definitely an electronic glitch.
https://www.google.com/amp/s/abcnews.go.com/amp/Blotter/toyota-pay-12b-h...

infofiles | 2020年1月17日

Folks using basic auto pilot and cruise control sometimes tend to forget that they may have disengaged the auto pilot by turning the wheel but as you approach a curve your car will not slow down, infact if car in front of you if it was more than 50 feet away and your car was following it at its speed that is slower than your cruise set speed, as that vehicle makes the turn your tesla is clear to accelerate to max speed you set, and will not slow down at curve as cruise control does not break for turns. It happened to me twice and I realized that unless you step on break OR there is a car in front of you that is driving slower, your cruise control stays on.

Bighorn | 2020年1月17日

This was a good RadioLab on the underlying electronic etiology of uncontrollable acceleration in Toyota. It’s why redundancy is so critical.
https://www.wnycstudios.org/podcasts/radiolab/articles/bit-flip

jithesh | 2020年1月17日

not yet.. but it has decelerated unexpectedly on rare occasions.

EVRider | 2020年1月17日

@howard: I'm not saying it never happens, but I'm standing by my statement. Plenty of evidence to support it -- see for yourself.

Ron.Olsberg | 2020年1月17日

Cruise control on most if not all cars I have driven before my M3 had two modes on/off and resume. On resume the speed would resume to what it was set at prior to hitting the brake, while just turning on cruise would set it to the current speed. On the M3 autopilot sets the speed to the speed limit, since is know what it is, plus any offset you may have set. So if you are in a 45MPH speed zone and have an offset of 4MPH, the car will quickly accelerate to 49MPH even though you may be only going 25MPH.. I think TACC would also accelerate to 49MPH if memory serves.. This could be taken as unintended acceleration to people not familiar with how Tesla Autopilot and TACC functions.

Joe M | 2020年1月17日

Only thing close that has happened to me in the Tesla was my own failure to push down hard enough on the stalk at which time I got unintended direction. Luckily I have enough respect for the way the car launches to always touch the go peddle lightly after a shift as a last check. It was still scary. Not the car’s fault but driving is a different experience with no detent.

lilbean | 2020年1月17日

Here is a great podcast about sudden acceleration http://revisionisthistory.com/episodes/08-blame-game

beaver | 2020年1月17日

No

1FASTROD | 2020年1月17日

I've had this happen but it was driver error not Tesla's fault.

I came off using AP and pressed on the brakes to stop the car. Once I got it rolling again I accidentally hit the AP stalk one time which caused Cruises control to kick on and started accelerating the car. I wasn't expecting to happen and took me a moment to realize what was happening.

On the other hand,

I've had a old Buick get stuck on WOT leaving a light (turns out the return spring on the carb jammed it open). That was scary as it took both my feet on the brakes to slow it down enough for me to steer from crashing and eventually shift into neutral (bouncing redline). The brakes were smoking when I finally got out to check what happened. Ah, the good ol' days!

neezer | 2020年1月17日

This can only happen if you just simply stink at driving, or you're not admitting it / lying because you're embarrassed or have nefarious purposes. The End.

Quadruple double dare anyone to show any proof of any such incident for even a microsecond where it occured and there was absolutely no throttle pedal depression whatsoever. You will fail.

This BS NHTSA new today is a bunch of a$$hats who thought it would be good to have a jerkaround with their attorneys to pull old complaints of this, which Tesla investigated in each and every single instance and found the evidence of the throttle getting depressed, pooled them together, and throw it at the press. Zero will come out of it.

Next thread people.

TabascoGuy | 2020年1月17日

@neezer - "Zero will come out of it."

Except for the stock price dipping for a few days. Then again, that was the plan all along.

neezer | 2020年1月17日

Meh, it's still the same price as it was on Monday this week. Definitely expect it to go up next week after the long weekend. And anyhow, this is all small potatoes ... Q4 ER is coming real soon....

Bighorn | 2020年1月17日

@lil
Scientists had a lot to say about Malcolm Gladwell’s podcast. I used to believe that too, before I looked at it more in depth.
https://www.safetyresearch.net/blog/articles/tipping-point

canuck55 | 2020年1月17日

just probably other car companies or stock brokers jealous of Tesla stock over $500. not sure why this is just brought up.

bryan.whitton | 2020年1月17日

I have had an unexpected acceleration in my Model 3. I was slowing down as traffic was coming to a stop in front of me. As we were nearly stopped the car in front of me suddenly stopped. As I was not on the brake I now needed to use the brake.

I moved my foot to the brake pedal and being a clumsy oaf I barely touched the brake with my shoe. My foot slipped off and onto the accelerator. As you can belive, I was unexpectedly accelerating. I immediately took my foot off the go pedal and moved it to the brake pedal but not before the car had already started applying the brakes and announcing to me with an emergency clanger that I was driving into the car in front of me. I got my foot on the brake just before I hit the car in front of me. I did feel the ABS work though.

Total damage between both cars was my front license bracket got a crack in it, $42 later it was fixed.

But it was definitely an unexpected acceleration.

rehutton777 | 2020年1月17日

Seems to me that if this was even a very minor problem (reports involving 500,000 vehicles??), someone would have mentioned it on the forum. Given some of the really minisule issues I have seen raised on this forum, I don't recall ever seeing even 1 reporting this concern.

lilbean | 2020年1月17日

@Big Thanks for the link. :)

Rikki-Tikki-Tavi | 2020年1月17日

I have this bs suggested to me by Google by all kind of sources all day today

RedShift | 2020年1月17日

There were some unintended cases of deceleration (on EAP). But not acceleration.

Right foot is usually the culprit when data is analyzed.

kaffine | 2020年1月17日

lbowroom | January 17, 2020
"It can be an actual defective part"

Yes, of course it CAN be. Especially with a diesel. Likelihood this is a problem with Tesla? Not much.

https://en.wikipedia.org/wiki/Diesel_engine_runaway

-----------

With Diesel runaway it is the engine that runs away not the vehicle With how runaway work the engine is normally not making much torque and will stall easily. The ignition of the unintended fuel source normally causes combustion to start way before TDC is reached. While the engine will overspeed to the point of violently coming apart it is if it isn't under a load that it happens. It is also possible for the engine to spin backwards during a runaway as the combustion happens so early compared to TDC the piston is forced down before the crank reaches the breakover point.

When I was a mechanic I only saw one truck that had unintended acceleration that could be shown to be something other than pedal misapplication. It was an issue with the cruise control wiring causing the PCM to think the driver is pushing the cruise control accelerate button. Tapping the brake stopped the acceleration although once you released the brake it may happen again.

The issue I have with the cosmic ray scenario is that the car isn't a voting machine, while the voting machine is just counting so a bit flip causes it to miscount and the next time it counts a vote it adds to the wrong number and the error continues. If it was using parity it should have caused an error but not all systems use parity. In a car computer it is not just counting it is processing the data looking at inputs and adjusting the outputs to the current conditions. So while a bit flip could cause a hiccup and make it think the throttle was just floored and adjust the outputs then next time it scanned the inputs it would go back to having valid data and adjust the outputs to match again. Brakes will overpower and slow a car it will take longer to stop but the brakes will slow the car. Of course this does assume the brakes are in working order and not neglected to the point of failure.

kevin_rf | 2020年1月17日

Everytime I hot the gas pedal... Can't figure out what magic makes it go.

sheldon.mike1010 | 2020年1月17日

OP: Never once.

billstanton | 2020年1月17日

OP: Never once.

Tronguy | 2020年1月17日

So, back when I used to inhabit the Prius forums:
In 2010/2011, there were lots of reports, as @Bighorn points out, of runaway Toyotas. There were some really notable deaths all over. Toyota supposedly got NASA boffins to review their software and attempted to claim it was all about aftermarket floor mats slipping, even to the point of sending out new maintenance documents that made sure that everybody was putting floor clips on the official mats.
It was all a lie and Toyota knew it. In one particular egregious case, a lady got killed being unable to stop coming off an off-ramp on a freeway. Some excellent lawyering managed to get past Toyotas game of keep-away with the design and source code of the motor controllers used on a slew of Toyotas. And then, the breakthrough: They got an excellent PhD in CS/EE who, among other things, was apparently the go-to expert on microcontroller design and feeding. And what this guy found was hair-raising.
Watchdog timers that didn't work. Buggy code. But where it got nasty was in the heap and stack management.
Really important variables are kept in RAM on a controller. Also in RAM is the heap, and then the stack. Every time one goes into a software subroutine the return address of the calling program is pushed onto the stack, along with (typically) space for the variables in that subroutine. When the subroutine exits, all the data is popped off the stack, the return address stuck into the program counter, and the stack pointer popped off to its original value.
Interrupts call subroutines; subroutines call other subroutines; and ad infiinitum. So, the stack grows and shrinks like mad as the CPU does all that's asked of it. So, those of us who do microcontrollers for a living spend some time making damn sure just how big that stack gets and making sure that it doesn't work its way out of the stack area. The heap is similar, managed the same way.
Toyota's software allowed system variables to get overwritten by the stack. Oh, sheesh. People who build safety critical software often put in hardware bounds so, if this happens, the CPU performs an in-place reset. Toyota didn't do this. It gets nastier. When software goes out to lunch like this the idea is to have a watchdog that one periodically punts into submission. If the punt doesn't occur on schedule or often enough, the watchdog (which is partly in hardware) wakes up and resets the CPU. The design was just.. Wrong.
The original CPU board barely worked on a Corolla back in the 2000 time frame. Trying to cut costs, Toyota kept on piling more and more software on this board. Our heroes finally figured out, and realized: Just driving straight down the road, the main motor controller board was resetting itself a few times an hour. Most of the time it would do its reset, recover, and the customer driving the car wouldn't notice the tenth-of-a-second glitch as it came back up. But sometimes, that wouldn't happen.
The engineers and his friends (who weren't allowed to take notes, just inspect the code in a clean room, more lawyering on Toyota's part) realized how bad it all was and was able to make can't-change-the-throttle-setting faults occur on demand, on a dynanometer. But that was just _that_ fault: When one overwrites system variables like this, random behavior is guaranteed to occur, and there was lots, geez, lots wrong.
Toyota lost that lawsuit, big time. Weirdly enough, to my knowledge, the cars themselves weren't recalled; my wife had one of them, a Toyota Sienna. Instead, we got $50 checks for "diminished value".
Turns out that Priuses use a different motor controller with completely different software and were never affected by this. But this didn't stop one enterprising soul in California from driving down a freeway with his Gen 3 Prius with his feet on both the brakes and accelerator at the same time, then launching a lawsuit and press conference at the same time. Unluckily and unexpected by him, people with cell phones took videos of the event and, if memory serves, there was an off-duty cop in the train who detected enough fishy stuff to blow holes in his attempt.
Finally: As previously mentioned, there were a number of German autos that were accused, mainly by a number of elderly drivers, that there was unwanted acceleration. A thorough study showed what happened: With slightly unfamiliar pedal placement, the driver would hit the gas instead of the brake. The car's moving! Stop it! Press the Brake Some More! And, zoom, right through the shop window.
And finally: The SO and I were in her secondhand Japanese Dodge Colt one year and, after accelerating a bit from a stop, realized that the engine wasn't slowing down. Being younger with fast reflexes, reached over and turned off the car via the key, which stopped the runaway. After a couple of pumps to free up the throttle, we restarted the car and parked. Found that the throttle cable was getting hung on the carb body. Moved the errant wire and went on our way.
So: 1. Bad design and a coverup. 2. Malicious intent. 3. Driver Error. 4. Mechanical malfunction. Whee!
For this one with Tesla: I'm leaning towards #'s 2 and 3. And my favorite is #2 because Tesla makes a big, inviting target and there's entities willing to fund any attempt they can think of to slow Tesla down.

Tronguy | 2020年1月17日

Oh, yeah. About that heap and stack: The system variables being overwritten by Toyota's motor controller were things like the throttle angle, gas volume.. stuff like that. Urgh. The testimony by the CS/EE type in court, in front of the jury, is something to behold.

Bighorn | 2020年1月17日

@Tronguy
Thanks for corroborating with facts. I think people got the Audi 5000 facts hardwired and then gave Toyota the benefit of the doubt with ossified thinking.

Bighorn | 2020年1月17日

Even Audi wasn’t as straightforward as what we heard from the MSM.
https://embeddedgurus.com/barr-code/2014/03/a-look-back-at-the-audi-5000...

dave | 2020年1月17日

You don't have permission to access

dave | 2020年1月17日

... sorry, I was trying to post a link, but I keep getting that error:
> You don't have permission to access "http://forums.tesla.com/forum/forums/teslas-accelerating-unexpectedly-ha..." on this server.
I posted something else as a test, and it worked.
Here is what I was trying to post:

https://electrek.co/2020/01/17/tesla-scrutiny-over-127-claims-sudden-uni...
> In every case, including that one, Tesla claimed that the car’s log showed that it was a user mistake due to pedal misapplication.

> In one case, we were even able to have Tesla’s log be verified by a third-party and it supported the automaker’s claims that it showed the driver pressed on the accelerator.

Bighorn | 2020年1月17日

@dave
A space or return in front of a link will help you.

Pages