Unintended Acceleration - Find the Cause

wwest · January 2012

"..My AC failed, but it wasn't due to the computer..."

How would YOU know...?

And....which computer...?

The engine/transmission ECU that is supposed to disable the A/C clutch during WOT, or closely nearby, engine operation...?

Or was it the one (or 2) in the A/C control "head" that didn't fail...?

andres3 · January 2012

The engine/transmission ECU that is supposed to disable the A/C clutch during WOT, or closely nearby, engine operation...?

Or was it the one (or 2) in the A/C control "head" that didn't fail...?

if the A/C can't handle WOT without destroying itself, that sounds like a design and engineering issue, not a computer issue.

wwest · January 2012

No argument there, the design may not be robust enough for operating above ~4,000 RPM. But still, it wouldn't have failed had the ECU not left it enabled.

andres3 · January 2012

But still, it wouldn't have failed had the ECU not left it enabled.

But how would we ever know that, couldn't it have just been a poorly assembled compressor?

Wouldn't an ECU malfunction trigger a Check Engine light or some other warning light or readable error code.

wwest · January 2012

"...couldn't it have been..."

Yes, but I was "rebuting" the statement that there was no computer "failure" associated with the compressor failure.

"...Would an ECU malfunction trigger..."

Possibly not, the A/C monitor firmware will provide an indication and throw a code if the compressor is enabled but NOT rotating cognizant with engine RPM. The inverse case is a "corner case" and would be mostly needless.

houdini1 · January 2012

I'm with andres3, all of this is pure speculation...if not outright hogwash !!

wwest · January 2012

Speculation...Yes, YES.

That was the actual intention.

Hogwash !!

NOT.

la4mead · January 2012

Yes, I agree. This discussion frequently takes turns away from the topic.

stever · April 2012

"Federal regulators plan to require automakers to design a brake-throttle override system into future vehicles to reduce the risks of high-speed, unintended acceleration.

The proposal by the Department of Transportations National Highway Traffic Safety Administration is a result of a highly publicized 2009 crash of a Lexus ES 350 and a subsequent flood of complaints about incidents of unintended acceleration in Toyota and Lexus vehicles."

Federal regulators want brake override systems in all cars (LA Times)

wwest · April 2012

If they expect it to be functional then it will need to be implemented separately from the engine/transaxle control ECU.

stever · August 2012

Maybe floor mats will just get banned.

"Mercedes-Benz is recalling all-season accessory floor mats sold in model year 2012 and 2013 ML-Class vehicles because they could cause the car's gas pedal to get stuck."

Mercedes-Benz recalls floor mats from ML-Class (Detroit News)

greg128 · August 2012

just did a bit of math and found the ratio of complaints to NHTSA for Unintended Acceleration to the total number of that car's sales for the last seven years (2006-2012)

The results for 4 cars..Sante Fe, Camry, Accord, and Malibu.

# of complaints : total sales
Sante Fe - 1 : 6,404
Camry - 1 : 2,960
Accord - 1 : 31,360
Malibu - 1 : 60,640

The complaint rate for the Sante Fe is 10x that of the Malibu and the Camry is 20x.
There are plenty of reports coming in for Toyota even after their recall, and Hyundai isn't trending very well either

greg128 · August 2012

If driver error was responsible for UA cases, then it would follow that older drivers would have a higher report rate. However I checked the complaint rate for a car that probably has one of the highest average age buyer - Ford Crown Vic, and a car that should have a relatively much lower age driver - Toyota Prius.

#complaints : total sales
Ford Crown Vic - 1 : 26,450
Toyota Prius - 1 : 3,060

These complaints are made mostlly only after an accident of some kind and most reported multiple instances before the accident where the accelerator pedal stuck or responded abnormally. I am sure these are only a small percentage of actual occurances of UA.

I think it is safe to say that there are definitely instances of driver error where the accelerator is mistaken for the brake, but that does not explain the much higher rate in some makes, even when pedal positions and floor mat placement is taken into account. Luckily this is rare, but apparently real, and in all probabiltiy having something to do the the electronics and computer control.

This happened to a couple in their 60's in Korea in a Hyundai Sonata:

Sonata UA Video

stever · February 2013

"The lawsuit claims that certain Toyota, Scion and Lexus vehicles equipped with electronic throttle control systems (ETCS) are defective and can experience unintended acceleration. As a result, the suit pursues claims for breach of warranties, unjust enrichment, and violations of various state consumer protection statutes. Toyota denies that it has violated any law, denies that it engaged in any and all wrongdoing, and denies that its ETCS is defective. The parties agreed to resolve these matters before these issues were decided by the Court. This settlement does not involve claims of personal injury or property damage."

Toyota SUA settlement website

plekto · February 2013

And so no evidence is presented. Everything is under NDA agreements and swept under the rug. This surely doesn't seem to me to be the behavior of a company that's got nothing to hide.

BTW, the most likely actual cause was not covered at all in the filing. It's not a bad sensor or bad code. It's nothing mechanical. It's simply that the computer froze up and got locked into doing the last thing it was doing. You didn't have acceleration so much as the throttle stayed exactly in the same position where it was before the computer froze up.

This happens to PCs, industrial equipment, and even aerospace components. Sensors and microchips often get stuck in a "livelock" (endless loop of repeating code) scenario when they unexpectedly crash.

stever · February 2013

You can always opt out and file your own suit.

I'm still holding out for tin whiskers myself. :shades:

srs_49 · February 2013

For the past year or so, there has been an examination of Toyota's ECU code by some independent third parties (or maybe they were parties to some of the pending lawsuits, not exactly sure here). This examination took place under a very strict Non Disclosure Agreement (NDA), with very stringent security and associated monitoring and logging arrangements in place so that it would be clear who accessed what. The computer system that provided this service was isolated from the rest of the world - no internet access, email services (except within the secured environment).

While the the examination showed no smoking gun, there were many instances of poor programming practices - something that you would not expect to find in code as safety critical as controlling the throttle. From what I heard, the code was certainly not anywhere the robustness that you find on critical flight software for an airliner, for instance.

I think that it was because of these findings that Toyota is caving in to the inevitable.

None of this is general public knowledge, BTW.

plekto · February 2013

BUT... the livelock scenario has absolutely nothing to do with the code.(bad coding aside, of course)

Nobody ever has tested the thing for abuse. As in literally hit the thing with a stun gun or physical shock and crash(software/hardware, not THE car) the thing while the car is running. What happens?

Think of is as closer to the power supply on your PC. how often has it crashed where doing the three second reset hasn't worked? I bet it has happened to you at least once in your lifetime where you had to unplug the computer from the wall and restart it manually. The biggest tip-off is the start button not working to turn off the car. I suspect that the start button is really a power supply switch and it simply froze up.

I do know that if a car was having UA, unplugging the battery lead would kill the engine as it would physically disable the injectors and coil packs regardless of whatever the computer might be trying to tell it to do.

Is this Toyota's fault, though? Likely not. Computers crash for all sorts of non-code related reasons and none of them are covered under any warranty or service plan that I know of. So why does this matter to me, then? Because I see the same idiot designs in multiple cars and UA isn't confined to just a few Toyotas, either. The vehicles have to be designed to be fail-safe when it comes to the computers freezing while the car is running.

srs_49 · February 2013

Computers crash for all sorts of non-code related reasons

Maybe, but i bet the large majority of PC crashes/hangups/BSDs are cause by software problems. Improper garbage collection, errant pointers accessing memory it shouldn't, etc are probably at the root of most crashes.

none of them are covered under any warranty or service plan that I know of

The typical shrink-wrap disclosure you're probably thinking of I don't think applies here. Did your car come with a lengthy EUA (End User Agreement) that says that none of the software on the vehicle is guaranteed to do anything correctly, and that if it does something wrong that causes loss of property or life that the SW vendor is not responsible?

plekto · February 2013

No, but hardware faults such as bad memory modules are simply covered under the basic warranty. That you lose your data, well, it's never covered.

Toyota can't really be sued because of outside influences, corroded wiring harnesses, vibration and shock, and so on. At most, they would be forced to change their design, though, which would be a good thing. But there's no money in that, really, so the lawyers don't bother.

ie - what this entire "challenge" was about was not about finding the overall cause (no proper fail-safe designs in any of the drive-by-wire systems), but finding a cause that could end up in Toyota being sued for damages.

houdini1 · October 2013

In the first big case to come to trial, Toyota was found not liable in a 20 million dollar unintended acceleration lawsuit. Looks like the cause all along was poor drivers and greedy lawyers. At least that is what the jury said in this bell weather case.

stever · October 2013

Here's a blurb with some other details. Be curious to see how many of the remaining ~300 or so cases wind up in trial.

Mr_Shiftright · October 2013

I always thought the scenario of a car going full throttle while loosing its brakes completely and negating the ignition shut-off completely and THEN coming completely back to normal by the time it gets to the dealer was simply UN-credible.

houdini1 · October 2013

Also throw not being able to shift into neutral in the mix.

I can see where a loose floor mat or other loose object might interfere with the gas pedal, but even in that unlikely event you would still have a lot of options.

elias · November 2013

you guys must somehow be missing all the reports of toyota settling some of these cases and paying out big $ . such as to the estate of the cop & his family who died.
hence the idea that 'cause was poor drivers and greedy lawyers' does not seem consistent with the court results!

Mr_Shiftright · November 2013

You don't go to court for the truth or for justice. You go to court to outmaneuver the other guy. Whoever can best work the system and is clever enough, wins.

So the settlement has nothing to do with presuming that electronics caused the UA. The two have no direct relationship, only a correlation in the lawsuit.

fintail · November 2013

Or it was cheaper than paying an overpaid legal team to work for months on a case decided by a jury of rubes.

The cop case doesn't seem connected to the usually over-50 "drivers" of runaway Camrys and Prius.

Mr_Shiftright · November 2013

no but I still find the odds of such a claim astronomical in unlikelihood.

fintail · November 2013

I can probably buy floormat + panic in the cop tragedy, but in the rest, it all seems like driver error.

srs_49 · November 2013

Have you read some of the reports from the Oklahoma trial that Toyota lost? Some heavily redacted parts of the testimony by Michael Barr (who's company was one of the ones that examined the source code for the ECU in a highly secured clean room environment) state that they were able to duplicate a UAE event with a Camry running on a treadmill or dynamometer with a simple single bit-flip in one of the critical software routines.

fintail · November 2013

Did they get this on video, and if so, is it available for public viewing? I'd think such data would be valuable to the market in general. In this day and age when anything can be recorded in numerous ways, data counts.

I'm very leery of trials for such things when the people making the decision usually have zero knowledge of cars, driving, engineering, et al.

srs_49 · November 2013

I'm sure they got it on video, but almost certainly it's not for public viewing. Most of what came out of that clean room study of Toyota's (Denso's actually, I believe) software is under heavy wraps. Toyota/Denso is claiming IP rights and protection.

Evidently, the courtroom in Oklahoma had to be emptied at times when testimony was going to cross certain lines. Even much of the publicly available testimony has been heavily redacted in some places to protect the guilty, and corny terms such as "Software Routine X" used in place of the real name, real variable names replaced with a generic VarY, that sort of stuff.

stevedebi · November 2013

"Have you read some of the reports from the Oklahoma trial that Toyota lost? Some heavily redacted parts of the testimony by Michael Barr (who's company was one of the ones that examined the source code for the ECU in a highly secured clean room environment) state that they were able to duplicate a UAE event with a Camry running on a treadmill or dynamometer with a simple single bit-flip in one of the critical software routines. "

I'm afraid I find your comment unclear. Are you saying the bit flipped due to software error, or they manually flipped the bit to cause the UAE? The first would be a major problem, the second simply a side note that the code is (correctly) not flipping the bit. But bits don't flip themselves, and if they observed that under debugging conditions, they could easily find out what happened - and issue a software fix. I have not read of any software fixes...

srs_49 · November 2013

No. A bit flip caused by an external event - a soft error. They simulated such a event by manually flipping the bit.

Bit flips are not all that uncommon given the small feature sizes of todays memories. The problem is that Toyota/Denso did not provide error detection and correction (EDC/ECC) on their memories or registers, even those holding critical variables. This is a no-no for safety critical applications, particularly when those techniques are widely known and employed in other areas.

The bit-flip mentioned caused the software module/routine to die.

Mr_Shiftright · November 2013

And causes the brakes to fail and the ignition to not shut off, all simultaneously?

I'm all ears to hear how that happened.

Have you all noticed that it isn't happening anymore? Now why is that?

busiris · November 2013

In my mind, this topic is approaching the level of the JFK probe.

srs_49 · November 2013

All of this came from EETimes.com - a trade newsletter for electrical and software engineers. You can probably go to that site, do a search for "Toyota" and you'll find hits to their articles on that case.

Here's the one that discussed the dynamometer test of an '05 and '08 Camry.

http://www.eetimes.com/document.asp?doc_id=1319966

And one of the more telling excerpts from that article:

However, we have confirmed in other vehicle testing that I'll talk about later, that if the incident begins with the peddle, [sic] brake peddle [sic] pressed at all, even lightly then the unintended acceleration will continue, potentially, forever unless the driver tries the risky thing of letting go of the brake while the car is driving away with him.

In this case, the driver has to do something counterintuitive - release the brake, then reapply it while the car is accelerating.

srs_49 · November 2013

In my mind, this topic is approaching the level of the JFK probe.

I find it rather fascinating!

Mr_Shiftright · November 2013

Still doesn't explain the brake *failure*. You can stop a car that has the pedal pressed all the way to the floor. It may not be pretty but you can do it.

Nor does it explain the inability to shut off the ignition.

srs_49 · November 2013

The brake failure and inability to shut off the ignition are almost moot points now. The bottom line from a legal standpoint is that the testimony showed that there were flaws in the design of the ECU, and that those flaws could lead to unintended acceleration. Whether or not there are ways the UA condition could have have been ameliorated/worked around is almost beside the point.

busiris · November 2013

"The brake failure and inability to shut off the ignition are almost moot points now. The bottom line from a legal standpoint is that the testimony showed that there were flaws in the design of the ECU, and that those flaws could lead to unintended acceleration. Whether or not there are ways the UA condition could have have been ameliorated/worked around is almost beside the point."

Well, if we're looking at it from a legal standpoint, by which I mean convincing non-technically oriented people of something, the question remains...

Why did the UA incidents go away?

Certainly the recall could have an impact, but no recall even gets close to 100% coverage of the recalled vehicles being repaired/modified. And, since we are talking about the possibility of hundreds of thousands of "potentially" affected vehicles still running around without any fix being applied, it seems to me that one would need a bit more proof to convince people on a jury that a real problem exists, or ever existed.

Just because something "could" happen in no way means it "will", or "has" happened.

stever · November 2013

Did the UA incidents go away or just fall off the news cycle?

Tesla Model S Involved In 'Unintended Acceleration Incident (autoblog.com - 9/25/13)

Unintended acceleration is claimed in New York accident (consumeraffairs.com - 11/26/13)

Oh, just got sidetracked - the tin whiskers theory lives on:

Genesee County judge orders Toyota to turn over documents in fatal unintended acceleration crash case (mlive.com/)

The Audi case lives on - this one isn't going anywhere fast either.

Mr_Shiftright · November 2013

Well if I read everything correctly, the testimony proves no such thing--it proves the *potential* UA from a design defect (a "what if" scenario) but in fact, in reality, no critic has been able to just hop in a Toyota, not touch it or manipulate it, and make this actually happen.

And it doesn't even touch on what level of competence we expect from a "driver" of a modern vehicle.

Based on this "evidence", what prevents someone from suing a car company that built a car with a big blind spot, thereby "causing" the driver to have an accident?

How To Spot a Conspiracy Theory:

1. The "official story" from the suspected party is always a lie

2. The alternative explanations, no matter how convoluted, are always more credible than the official story

3. Any evidence that contradicts the alternative explanations are only proof of the depth and cleverness of the conspiracy.

busiris · November 2013

"Did the UA incidents go away or just fall off the news cycle? "

I don't know if it makes much of a difference in the end.

The legal system is all about opinion, and much less about the actual facts.

Just ask anyone freed by the Innocence Project that were previously convicted because a small group of individuals found them "guilty beyond a reasonable doubt".

We will all be long dead and gone before the final chapter is written on Toyota and UA.

Mr_Shiftright · November 2013

Drivers should be able to know what to do in a UA incident. This should be a part of any basic training. There are people out there operating 1.5 tons of hurtling steel who probably could mess up a paperclip while trying to use it.

busiris · November 2013

"Drivers should be able to know what to do in a UA incident. This should be a part of any basic training. There are people out there operating 1.5 tons of hurtling steel who probably could mess up a paperclip while trying to use it."

I completely agree.

Who does history hold to blame for the Titanic sinking, the Captain or the shipbuilders?

Most historians would place the blame squarely on Cap'n Smith's shoulders, even though there were (clearly by today's standards, anyway) "issues" with the ship's construction.

IMO, it seems that some believe there will be a magical "Aha!" moment, at which time all facts will be revealed, clearly identifying Toyota (in this case, anyway) as knowingly negligent in manufacturing cars with UA an inevitability. Even if that were to happen, I can see them looking like a deer in the headlights of an oncoming 18-wheeler after it gets dragged through the US court system, wondering what happened...

fintail · November 2013

Brakes fail easily when you hit the gas pedal instead

Still curious to me that most of these incidents happened to older drivers in cars usually not aimed at those who are "with it", so to speak.

gagrice · December 2013

The driver of the Lexus SUV was parking her vehicle when the vehicle accelerated over a parking barrier onto East Bidwell Street, then back toward the building, said Andrew Bates, Folsom Police officer. The car crashed through a side window and into the main area of the shop.

The driver and her passenger were not injured, but eight customers inside the store were hurt.

Although the investigation is ongoing, there is no indication that impairment, age or cell phone usage was a factor in the crash.

http://www.sacbee.com/2013/12/10/5989688/suv-crashes-into-starbucks-in.html

fintail · December 2013

Given the extreme situational awareness of Lexus SUV drivers, I am skeptical.

"there is no indication that impairment, age or cell phone usage was a factor in the crash. "

According to the badge carrying demographic, the Paul Walker/Roger Rodas death car was only going 40-45, too.

The Lexus is also a first gen RX, no newer than 2003, if it matters.

box1car · December 2013

When Rhonda Smith testified before congress about her runaway Lexus as it was going ~100mph, that she dialed up her husband on her cell phone and
"God Intervened" to slow her car to ~35mph. Someone should have asked
"What number did you dial? did it contain the digits 3 and 5?" no one did

Cell phone frequency ranges of 800 MHz to 2100 MHz give a wave length of
WL=Light Speed / Frequency, where W.L. varies from 14" at 800 Mhz to
5.3" at 2100 MHz. Now an FM "T" antenna of half wave wire lengths will
RECEIVE these RF signals. So any wire length of 7" protruding into cockpit
will pick up 800 MHz, can we say CRUISE CONTROL lever, using turn signal

So when Rhonda Smith dialed her phone she sent a digital signal thru the
turn signal arm directly into the CPU for Cruise Control and it responded
Why did gov't DOT NOT say anything? Last thing gov't needs is another
car company to bail out, so better to say nothing, just like when they
finally realized Tobacco fields were full of poisonous Arsenate of Leadd
Just zip you lip and let tobacco plants suck up all the poison, hey, the
stoopid smokers are gonna' die from Nicotine and Tar poison anyway so
who cares?

Unintended Acceleration - Find the Cause

Comments

Helpful Tools

Most popular vehicles