By on February 22, 2010

Ever since Toyota’s recent problems hit “frenzy” level on our mainstream media monitoring system, speculation has been rampant that some mysterious electronic problem was at the root of the unintended acceleration scandal. We’ve been wary of jumping on the “ghost-in-the-machine” bandwagon, for a number of reasons, chief among which is the fact that it seems to be the product of an inability to explain specific instances of unintended acceleration, rather than hard evidence. Given that unintended acceleration occurs at the intersection of man and machine, good old-fashioned human error is an easier assumption than mystery software errors. Given the worrying results of our Toyota gas pedal analysis, we’ve been content to explain the situation on a combination of pedals, mats and human error. But now ABC News may just have the first positive evidence of an electronic problem that could explain the mystery behind Toyota’s unintended acceleration problem. Dave Gilbert of Southern Illinois University has found that it’s possible to cause unintended acceleration without it triggering an error code that might give some kind of clue as to its cause. Combined with our finding that Toyota actively conceals data from its black box data recorders (out of line with standard industry practice), this could be some of the first positive evidence that there’s more to the “ghost in the machine” theory than mere panic-driven speculation.

Get the latest TTAC e-Newsletter!

79 Comments on “Has ABC News Found The Ghost In Toyota’s Machine?...”


  • avatar
    twotone

    If Toyota was producing software instead of autos, this would be marketed as a “feature” instead of a “bug”.

    Twotone

  • avatar
    Quentin

    They don’t clearly state what “input” they are adding. I mean, they could have just spliced into the pedal harness and apply voltage, which would give a WOT condition. They don’t physically show what they are shorting nor explain it.

    Show me the short on an affected car.

    • 0 avatar
      JohnAZ

      A first step in any diagnostic procedure is to try to reproduce the symptoms with any appropriate device. The prof seems to have done this with a low resistance short of the gas pedal signal.

      His test showed that UA can be produced without a Check Engine light, while compromising braking power. It is more important to show that the reported symptoms can be reproduced at will rather than it is to prove the actual culprit.

      It is now up to Toyota to follow through and test for how an apparent low resistance short can occur. They should have been working on this diagnostic test from day one.

  • avatar
    grifonik

    ABC (or someone), Please explain the cause of the short. Is this a “wear and tear on wires” type short or is it a short that is caused about by specific inputs to the system?

    • 0 avatar
      JohnAZ

      It does not have to be a short. It could be anything that acts like a low resistance short.

      Could it possibly be a 5v regulator problem that gives a low voltage signal from the gas pedal (like a short would do) and also produces a low voltage signal from the brake pedal producing only a partial application of the brakes. This might feel like a full brake pedal application, but only half the required pressure on the brake pads from the power assist causing them to not overcome the engine but heat up badly and cause the reported odor.

      The main point is that the symptoms have been reproduced, now Toyota has to get off their ass and find why it is happening.

  • avatar
    chuckR

    Remember the 1970’s saddle mounted GM light truck gas tanks? NBC had to apologize on air for triggering the desired gas fire by using model rocket solid rocket motors.
    Remember Audi SUA? One of the three, C-BS, NBC or ABC triggered SUA by tapping a compressed air line into the transmission. Oops.
    Sorry, ABC. If you want anyone with a brain and a memory to believe this, you will have to be completely transparent. I hope your guy stumbled onto the problem, but details please.

    • 0 avatar
      Robert.Walter

      The Audi Sudden Acceleration Fraud of the 1980’s was perpetrated by CBS’ 60 Minutes.

      The Chevy Side-Saddle Fuel Tank Fraud of the 1990’s was perpetrated by NBC’s Dateline.

      I would hope that the Walt Disney Company’s lawyers have a) studied both of the above case-studies, and b) were asked to vet this report before it aired.

    • 0 avatar
      Telegraph Road

      @chuckR–for the Audi SUA media fiasco, it was CBS 60 Minutes with Mike Wallace narrating. I remember the clip of the “induced” SUA as if it were yesterday.

      Concerning the GM side-saddle gas tanks, in 1993 from a TV set within the Glass House I watched Harry Pearce (GM chief atty) give his devastating critique of Dateline NBC’s fraudulent methods. Ford employees literally applauded their arch-rival’s victory.

    • 0 avatar
      JSF22

      Those were the two most egregious frauds by these so-called journalists. But it wasn’t Mike Wallace in the “60 Minutes” segment; it was Ed Bradley, one of the biggest phonies who was ever given airtime.

      Harry Pearce’s surgical strike against NBC, showing how “Dateline” attached incendiary devices to the Chevy pickups, was among the best closing arguments in history. NBC apologized and fired the president of its news operations within the week.

      Toyotas may well have an electronic problem, but the ABC segment surely doesn’t prove it.

    • 0 avatar
      Telegraph Road

      JSF22: Ahh…Ed Bradley…you’re right.

      “Toyotas may well have an electronic problem, but the ABC segment surely doesn’t prove it”. I agree.

      Yet, unless Toyota has another Harry Pearce on staff, this could prove murky.

    • 0 avatar
      chuckR

      Explosives were not used in the GM truck story – too easy to detect and maybe not effective at simulating a simple fire. When NBC issued their retraction/apology (of the sorry you were offended type), Jane Pauley referred to the triggering mechanism as ‘Estes sparkers’. Anyone who as a kid had an interest in model rocketry knew that referred to the shotgun shell-sized solid rocket motors made by Estes, not things like sparklers. As a order of magnitude guess some of these motors could send a 1 oz rocket 1500 feet in the air – expending 100ft-lbs of energy. A 357 magnum round has about 550 ft-lbs so these motors contain a not insubstantial amount of energy. This was a very clever out and out fraud. You could bang away with a 357 and still not do more than have a leaky gas tank. But the motors have a controlled burn of several seconds and by waiting for a leak to occur and vapors to form, you can then trigger the burn at the most (in)opportune time. I completely stopped watching 60 Mins and Dateline after that.

  • avatar
    grifonik

    I was a little skeptical at first… but then I saw all the patches the guy was wearing. He is clearly an automotive Einstein.

  • avatar
    mcs

    Rather than a short, I’d like to see what happens with unclean or out-of-spec power. Usually when I’ve had a strange problem with electronics,that’s been the problem. Learned my lesson – stay away from really cheap electronic components. Especially cheap filter caps.

    • 0 avatar
      Robert.Walter

      I consult on Design, Quality and Liability issues to a Fortune 100 company currently industrializing an advanced mechatronic system for vehicle control (large-scale deliveries begin w/in 2 years).

      I too have counciled my client to ask “what if” questions regarding potential failure modes due to dirty power or signal.

      All kinds of odd things can happen if this is not considered and dealt with.

    • 0 avatar
      JohnAZ

      This is incredibly good news for the rest of the Auto Industry, all except Toyota that is.

      Without this test (and partial evidence), the Washington hearings might have concluded that there could be some form of electrical interference that could potentially affect all auto computers. They could have sent the industry off on a wild goose chase spending billions to find a theoretical ghost that may not be there, while collapsing sales until proof was found.

      The Prof has pointed a very large smoking gun at Toyota and Toyota will now have to successfully reproduce the test themselves, then find the real cause.

      Anyone want to bet that all of the major manufacturers are not hard at work doing this test for themselves. I’ll take your bet. I’ll also bet that mcs is right that it is an unclean power fault induced by a cheap component.

  • avatar
    porschespeed

    We all knew the Audi stuff was operator error, and that the tv reports were rigged – if you know cars there is simply no possible failure mode that could have a K-Jet (CIS) car suddenly floor itself and overcome a fully depressed brake pedal UNLESS there was a total mechanical failure of the brakes and the CIS that would be easy to find after an accident. Natch, none of those failures were ever found. Ever.

    That being said, 100% drive-by-wire throttle worries me. Software in a hybrid is very complex, and the chance of something going wrong is just too likely for my taste. This isn’t a few thousand lines of code to do an EFI box – there’s a lot more to it than that.

    Toyota is doing itself no favors trying to hide the blackbox info, but I need to see something substantive for causality.

    If there’s anything there, I wanna see somebody hood up a 24 channel DA unit and document a failure scenario. Gimme some science and I’ll believe.

  • avatar
    Paul Niedermeyer

    This raises a number of red flags. What they have done is to find a way to send a full acceleration signal in a way that doesn’t create an error code. Is that intrinsically a problem Can a GM or other brand car be made to do the same thing?
    It also doesn’t at all explain how this might have happened in the real world. It just shows that it could have, thanks to carefully engineered human intervention.
    It also doesn’t surprise me that an acceleration doesn’t create an error code. So what have they really proven here? Almost nothing, as far as I can tell.
    In fact, it borders on the 60 Minute piece where they rigged the Audi 5000 to jump forward.

    • 0 avatar
      baldheadeddork

      I don’t know if you read the story accompanying the video, but it does explain how this can happen in the real world:

      “The system is fallible, in fact, it’s got some really troubling design strategies that are employed by Toyota that appear to be outside the norm. And their system clearly has a design strategy that has a very slim margin of safety.”

      Kane said the short circuits introduced by Gilbert in his tests and demonstrations reflect what can happen in the real world because of corrosion, moisture, and manufacturing imperfections.

      It sounds like they’re describing a wiring or board design that can allow this if there’s a short between neighboring wires in a block connector or open solder points on a board. If the wire that uses variable voltage crosses with a wire that carries a constant voltage within the operating range of the first wire, the shorted circuit will take that voltage as a normal signal. For example, if the API off the pedal assembly uses a signal ranging from 0.25v to 3.5v to signal accelerator position and it shorts with a wire carrying 3.5v, that first circuit is going to take the 3.5v and respond to it like a normal signal from the accelerator. It’s within the normal operating range of the circuit so it’s not going to do any damage or register an error code. If the short is caused by moisture, the increased voltage can cook off the water and break the short, creating an intermittent failure.

      About if the problem is intrinsic to Toyota or if it can affect a GM product, again from the story:

      “Other vehicle manufacturers have gone to great extremes,” he said, to prevent the problem he claims to have found in the Toyotas. His tests on GM cars did not find a similar flaw, he said, “not even close.”

      This also points to a problem with the physical layout of the wire connectors or the circuit board.

    • 0 avatar
      Robert.Walter

      Paul: +1

      As I posted above, I hope ABC took lessons from the CBS and NBC debacles.

      Interesting demonstration, but is sorely lacking the context under which such a fault could be generated. Showing potential effect without potential cause is borderline irresponsible.

      Baldy, thanks for the context. ABC still has some work to do by giving more details behind what differentiates Toyota’s design from the other OEM’s and why it is “not even close.”

    • 0 avatar
      baldheadeddork

      @Robert

      ABC has to be careful about what information they put out and how they do it. If they say “shorting this black/yellow wire with that neighboring red/blue wire causes the engine to race uncontrollably”, ABC and SIU are opening themselves to a huge amount of liability if someone does it on their Camry and kills themselves or someone else.

      They’re not putting the reason out there directly, and I think that’s an editorial mistake, but the tone they use in not describing what has to be shorted is almost identical to what Jamie and Adam do on almost every episode of Mythbusters. I think it is the same thing.

    • 0 avatar
      Robert.Walter

      Baldy: I agree. But I’m not looking for a cook-book, rather a broad description of what they did and how the effect they achieved could be caused by the various components (or s/w) in the system and a comparison to other OEM systems (they mentioned GM, so why not start the comparison there?)

    • 0 avatar
      L'avventura

      >>In fact, it borders on the 60 Minute piece where they rigged the Audi 5000 to jump forward.

      The worst is General Motors v. Dateline. They actually rigged a truck with explosives to make it look like GM trucks exploded upon impact.

      This particular article is pretty horrible as well and lacks reproducibility, the accompanying article is incredibly vague, and it doesn’t pinpoint how they reproduced SUA, nor does it get into any technical details actually explain something.

      @Baldheadedork

      There is no liability if ABC is exposing or reporting a real flaw. If its indeed true, and if they legally obtained their information, there don’t need to worry about litigation. First Amendment rights and Free Press laws are very explicit about this. If its a fact, there is no reason why ABC shouldn’t release information.

      More like ABC doesn’t want to be second-guessed or scrutinized on their accusations.

    • 0 avatar
      baldheadeddork

      @L’avventura:

      With all respect, and trying to keep this civil, you’re wrong. To start, the First Amendment only prohibits the government from censoring the press. It doesn’t protect reporters from civil actions, including slander, libel and liability if the information is unique and allows someone to hurt someone else.

      But that doesn’t matter because the legal exposure for ABC has nothing to do with Ross being a reporter or this segment running on their news show. If you figure out how to do this and you show it to someone who then does it with their Avalon and they hit someone else, you are most likely going to end up getting sued because you discovered the process and shared the information that made the act possible. You weren’t at the wheel, but it couldn’t have happened without your actions and that makes you liable.

      This applies to SIU, too. They have some defense if they publish the details in an engineering journal because a layman would have to make extraordinary efforts to find and use that irresponsibly. But if this prof puts up a YouTube video showing exactly how this is done it’s their ass if some kid follows the instructions and ends up wrecking mom and dad’s Camry.

    • 0 avatar
      baldheadeddork

      @ Robert: I can see your point, and I’ll bet anything we’ll see the full details on someone’s blog before the close of business tomorrow. Gilbert left some pretty big clues (at least to my eyes) and it won’t take long at all for someone to duplicate his experiment.

    • 0 avatar
      JohnAZ

      You guys got it all wrong re any similarity to the other events.
      This test was not produced by a news agency.
      The Prof reported it to Toyota before he went to ABC.

      You also miss the point that the Prof has reproduced the symptoms by using a POTENTIAL fault. This story is solely about the reproduction of the mysterious symptoms, NOT about the actual cause. That part is up to Toyota to find.

    • 0 avatar
      L'avventura

      @ Baldheadeddork

      If they are telling the truth there is no fear of slander, and pointing out a flaw in a system does not make libel in the very rare case someone tries to use this elaborate method in a homicide. The civil suit won’t make it far, and ABC already has a team of lawyers. Their primary responsibility is reporting the news. Besides that, there is precedence for judgments in favor news organization for the free decimation of information in publications even when publication really do incite violence such as the Tuner Diaries.

      Also this is all absolutely silly, there are plenty of ways to release technical information without ‘blue wire to green wire’ type step-to-step instructions. Beyond that a kid trying to kill mommy and daddy with this in their Camry is beyond imaginable. First is the high technical level of knowledge required to even implement it, second is that even if its not guaranteed that SUA may not kill or injure anyone, and third, any SUA incident in a Toyota will be tightly scrutinized now and in the future. In fact, the article is so vague it doesn’t point to any specifics what so ever.

      What ABC might be scared of is slander, because they may be staging all of it or they may be bending the facts.

    • 0 avatar
      CamaroKid

      @JohnAZ

      You seem to have some inside information as to what exactly this test was, care to share? Cause in the ABC piece we get basically NO information. In the news report it all looks very suspicious, almost like he is trying to hide what he is doing.

      If it is scientific he should be able to document it for all to see and duplicate. He does not appear willing to do that… So far anyway.

      If he is just shorting the two sensors together and then shorting that to ground…. and a low voltage signal is WOT… then he has discovered NOTHING… anyone of us could do the same thing with a low ohm resistor and a chunk of wire. Again, there is no proof here, are there frayed wires in the cars that experience this problem? Is there a suspect circuit board in the pedals?

    • 0 avatar
      JohnAZ

      @Camarokid,
      Dr Gilbert has said very clearly that what he has done is reproduce the exact symptoms reported by people who have experienced the UA problem. He has not claimed that what he did explains the actual cause of the problem. In fact in his testimony in Congress, he said very clearly that he does not know the actual problem, YET, but he knows that after 30 years of expertise in auto electronic control systems that Toyota’s ETCS system is not responding properly to his induced fault. He suggested 3 or 4 possible causes for the symptoms he reproduced.
      He also said that he has provided all of the specific details of what he did to both Toyota and the NHTSA.

    • 0 avatar
      crash sled

      “In fact in his testimony in Congress, he said very clearly that he does not know the actual problem, YET, but he knows that after 30 years of expertise in auto electronic control systems that Toyota’s ETCS system is not responding properly to his induced fault.

      .
      .

      John, in his testimony today, Toyota’s Lentz claimed that his consultant did Gilbert’s test, and not only reproduced it on a Toyota with his same results, but also reproduced it on other OEM’s vehicles, with the same results.

      If Lentz’ statement is true, then Toyota’s ETCS strategy is in line with those other makers, and this Gilbert guy is mistaken. Somebody’s wrong here. Who?

    • 0 avatar
      CamaroKid

      @JohnAZ

      Clearly? Then you have more information that you aren’t sharing. All he said to congress was the he induced a low resistance short between the two redundant sensors and then applied a 5 volt signal to this “tampered” pedal. That is not at all clear. What is the resistance of the low resistance short? Is there any evidence that this has, could, or can occur? How do you get a 5 Volt signal into this circuit?

      You also left out two more important facts.
      1) Toyota tried to replicate what he said he did and they instantly triggered a ECM code. Unlike David Gilbert’s experiment the Toyota Experiment was documented for scrutiny. and
      2) We now know that David Gilbert is not working alone on this project and in fact is on the take, being paid by the lawyers of one of the class action lawsuits against Toyota, he is NOT independent and he is not untainted… He is in a FULL conflict of interest.

  • avatar
    mcs

    Wait a minute, this whole premise seems flawed. We need more details. If you put a voltage on the sensor wire that indicates full throttle, then the car is going to put the throttle at full. Why would it be an error? Exactly what did he short? If I remember correctly, that unit has two sensors. Did he short both sensors? There’s a lot of missing information here.

  • avatar
    JSF22

    I’m no Toyota fan and I’m astounded how they could have handled this so badly, but I think ABC News is as likely to get to the bottom of this as the U.S. Congress is. I can’t think of two institutions I trust less, or that have more hidden agendas.

    I watched the ABC story very carefully and all it proved to me is that you can rig anything to malfunction, and that the greater the system’s complexity, the more spectacular a failure you can generate. There was no proof whatsoever that this is actually what is happening in these cars.

    Over twenty years ago “60 Minutes” did exactly the same thing to Audi with a crackpot “expert” named Rosenbluth (a plaintiffs’ paid expert witness), who “proved” that if you injected enough compressed air into an automatic transmission, you could make it shift to “Drive” without a driver command. That this had nothing whatsoever to do with anything was completely lost on Ed Bradley and most viewers.

    All that said, Toyota’s legalistic public relations responses have been the most incompetent I’ve ever seen, and as a result of their delay and incompetence they are going to take it right up the a**.

    • 0 avatar
      MBella

      Exactly.

      Using the 60 minutes example, this would be the same as them coming right out and admitting that they rigged it, and afterwards telling people they should be scared because the car can be rigged to do something. With this logic, we should also be afraid of every car the Mythbusters have ever used to do something ridiculous.

    • 0 avatar
      Telegraph Road

      “I’m no Toyota fan and I’m astounded how they could have handled this so badly, but I think ABC News is as likely to get to the bottom of this as the U.S. Congress is. I can’t think of two institutions I trust less, or that have more hidden agendas.”

      I completely agree. Yours is the best summary yet. But I will be glued to C-SPAN tomorrow, sadly.

  • avatar
    BMWfan

    Could this be caused by a battery that is about to die? On BMW’s all sorts of problems can possibly be caused when the battery is past its prime. Warning lights for no reason, limp mode, check engine lights, all sorts of problems. No unintended acceleration, but damn near everything else.

    • 0 avatar
      MBella

      No, it can’t. Depending on the design, low voltage could cause less than desired throttle to be applied, but not more. However, the computer should be designed to compensate for the low battery voltage. The German example is more so designed to get you to the dealer when you need a new battery.

    • 0 avatar
      JohnAZ

      I highly doubt the problem is at that level. Other systems would also fail, and the problem would not likely disappear.

      Cars have various power needs at different voltages, and I wouldn’t be surprised if there is a regulated voltage, separate from the one for the ECU power, that is used for ‘sensor controls’. If there is such a source in the Toyotas, and it failed partially, it might affect the acceleration, and potentially the signal from the brakes leading to UA and the engine overpowering the brakes. Only Toyota can answer that.

      By way of an example, I have a 24 year old 1986 Mustang that had an electromechanical 5V regulator(small silver can behind the dash) that was used to supply power to all the gauges. The regulator was notorious for going out of whack when hot and trying to peg all of the gauge needles. I built an electronic 5v regulator to replace the stock one and the problem ceased.

  • avatar
    ra_pro

    That’s the genius of a BMW; all kinds of things go wrong a lot of the time but nothing so serious as to make it to the front page news.

  • avatar
    Ion

    I saw the video on their website and lol’d at the cheesey cut to the tachometer when he induces sudden acceleration.

  • avatar
    Russycle

    As others noted, ABC has no more credibility than Toyota. Unless they are completely open about how they “recreated” the UA condition, this report is worthless. For all we know, every car on the road with an electronic throttle may be susceptible to this “problem”, given sufficient tweaking. In fairness to the expert, he seems legit, from a very cursory web search.

  • avatar
    crash sled

    Well, here’s the latest scoop from the Detroit News: http://www.detroitnews.com/article/20100222/AUTO01/2220389/House-panel–Toyota-misled-the-public–dismissed-electronic-defects

    The House E&C committee mentioned in this article is the real deal, as the other committee being mentioned elsewhere appears to be a stalking horse for California Congressman Issa to yelp about the NUMMI closure (although I’m sure Cali Rep. Waxman will get his licks in on NUMMI during E&C hearings as well).

    The preliminary report appears to take a whack at Toyota, but offers no proof other than mentioning a “flawed engineering report”. It also takes a whack at NHTSA. No clue from this as to how this will proceed tomorrow. From the tone of this article, I suspect Toyota will have to answer for the hidden black box data. If I was Mr. Toyoda-san, I’d have versions of that blaok box reader for sale on eBay, and by tomorrow morning.

    The House Energy and Commerce Committee and its investigations subcommittee chaired by Rep. Bart Stupak, D-Menominee, outlined their preliminary findings today, one day ahead of a hearing on Toyota’s recall of 8.5 million vehicles worldwide for sudden acceleration issues.

    “Toyota resisted the possibility that electronic defects could cause safety concerns, relied on a flawed engineering report and made misleading public statements,” the committee said in a letter to Toyota Motor Sales USA president Jim Lentz.

    The committee also said the government didn’t adequately investigate.

    “NHTSA’s response to complaints of sudden unintended acceleration in Toyota vehicles appears to have been seriously deficient,” the committee said in a letter to Transportation Secretary Ray LaHood.

    • 0 avatar
      JSF22

      Toyota shows every sign of being in the wrong here, however:

      1) It is possible that a Congressional committee that issues its conclusions before it holds its hearing has something other than my personal well being on its mind.

      2) I will consider Bart Stupak’s views on technical matters after he acknowledges that there just might be something to this human evolution stuff.

      What a collection of fools.

  • avatar
    65corvair

    When you tell the Toyota computer to go and stop at the same time it will do both. It doesn’t turn off the gas imput. The other day at work I thought I would see what would happen if I applied the throttle and brake at the same time on a Toyota electric fork lift. It was like I had taken my foot off the throttle, and the brake stopped it without any problem. The Toyota fork lift division has this safety feature, but not the cars. Must because those fork lifts have such a high top speed. I’ve never been able to it them to go over 8 mph.

  • avatar
    blue adidas

    Are they suggesting that this “is” what likely happened? Or are they saying that this “could” be what happened? Because what I see is a guy that has spliced into the system and rigged it to malfunction without registering a fault. I didn’t hear him suggest at any time in this video that this induced fault represents something that is likely to happen on its own. I don’t trust Toyota. But I also don’t trust highly edited entertainment news programs like this.

  • avatar
    CamaroKid

    Just as all of the rest of the best and brightest have pointed out… We have absolutely nothing to go on here.. we have no idea what he did… we just have a grainy shot of a computer screen with a bunch of squiggly lines… and some lose wires in the passenger seat… This all looks very suspicious.

    If ABC news would like I could also get the wipers to wipe, the radio to turn on an off and the transmission to randomly shift forward gears without generating a code too… Call me, I want to get on TV too.

    What do we all know about automotive sensors? We know that they work in a voltage range usually from 0-10 volts… We don’t know what voltage signal the toyota gas pedals sends for idle and what is sends for “full open” … Why not tell us this simple fact…

    SO Lets guess that 10v = idle and 0 volts = WOT… So he shorts the pedal to ground and ta da the car goes like a bat out of hell… What did you expect? And this guy is some kind of car expert… Is there ANY evidence of any frayed wires or broken connectors on any of the cars? There is a HUGE difference between a theory and a proof.

  • avatar
    Log

    While I agree that the engine control should be fail-safe so as to not allow a single failure to be undetected or unaccommodated — if that’s indeed the case here — however, just because this alleged mode of failure CAN happen doesn’t mean that it HAS happened in any of the accidents. As for Mr. Gilbert, I wonder about his background. While he does teach at SIU, he has a Masters degree in Industrial Arts Education. I would have more confidence in his assessment if he had an engineering background.

    http://automotivetechnology.siuc.edu/d_gilbrt.html

  • avatar
    the_gamper

    As others have pointed out, I think we are missing some key facts to this report. How was the short caused for one? Also, can the UA caused by this type of short be reproduced in other cars from a variety of manufacturers?

    I dont really know what to think at this point about Toyota’s software issues, but one thing I know for sure is that the American media relies heavily on sensationalism. I am not really a fan of Toyota, but in their defense, it is difficult to take ABC’s repot very seriously at this point.

  • avatar
    davejay

    I think any kind of “we found it!” announcement is premature. However, the process this is following is pretty typical for solving engineering solutions:

    1. Suspect there is a problem.
    2. Confirm there is a problem.
    3. Suspect potential causes for the problem.
    4. Confirm potential causes for the problem.
    5. Suspect actual cause of the problem.
    6. Confirm actual cause of the problem.

    We’ve been in the land of #3 for a while now, and we’re transitioning into #4. There has been much suspicion about a software or hardware failure that could manifest as a car ages as an intermittent full-throttle condition with no error codes (#3) and this person appears to have found a design flaw that may validate that suspicion.

    So ABC news jumps the gun for the sake of ratings (surprise!) but this looks like a legitimate mechanism for the behavior being reported. Nothing more, nothing less.

    • 0 avatar
      JohnAZ

      DaveJay,
      I disagree with your list, and I offer my 40 years in IBM as evidence. I spent many years diagnosing hardware, software, and system problems.

      When a likely cause didn’t show up in the early diagnoses, we would look for ways to reproduce the problem. Once you could reproduce the problem at will, you could get a pretty good handle on what was actually causing the problem and further isolate it.

      It didn’t matter what tricks you used to reproduce the symptoms, as long as they were based in reality. What mattered was that you could isolate certain components or causes. The fault was then often easy to find.

      Gilbert has done the hard part – reproduce the reported symptoms. Now all Toyota has to do is the easy part – isolate the failing component.

  • avatar
    GS650G

    If someone can find a way to do this to another companies car, I’d like to see it. You can question the methods and the news organization doing the story, but why doesn’t applying the brakes tell this oh-so-smart computer to drop the engine speed. It needs too. And this is going to cost Toyota deeply.

    After the dust settles and the loyal fans return to buy the New Redesigned Camry with Saf-T-Stop features I’ll be sitting on a nice chunk of Toyota stock ready to make a killing.

    The media is good for one thing, making unaffordable stock affordable for a while.

    • 0 avatar
      CamaroKid

      Give me a wiring diagram, an Ohm meter, a small 12volt battery, some wire, a couple of potentiometers in varying ohm ranges/wattages and a day or two to work on things and I’m confident I could get any and all drive by wire systems to do this.

      Even if he is right, this does not prove a problem with the Toyota ECM, and does not prove a ghost in the machine… It proves that a short in a mechanical sensor can send a bad signal to the ECM. Oh and BTW if we could read the EDR it wouldn’t show anything other then someone pressing hard on the gas.

    • 0 avatar
      Greg Locock

      I tried simultaneous brake and throttle on 4 cars with automatics from different manufacturers this week, none dropped the throttlewhen I applied the brakes. BMW, Ford, Mitsubishi and Toyota. So its absence is not unusual.

      Incidentally they all allowed me to select neutral at 100 kph, and at least two of them were shift by wire.

  • avatar
    niky

    According to Toyota, when Mr. Gilbert contacted them, he mentioned that he had intentionally created a short between the two accelerator sensor wires using a resistor.

    For those of you who know a little bit about automotive modification and/or electronics in general, you’ll know that we use resistors specifically to fool automobile ECUs into thinking one thing or another. It’s a popular and cheap way to change fueling (and is a component of many cheap “performance chip” scams), it’s a way to fool the emissions system and preven MILs or CELs, and apparently, you can fool e-throttles with one, too. Whoop-de-doo.

    A “short” created by soldering two wires together, giving a steady signal, and using a resistor to prevent cross-fire signals from leaving a specified voltage range isn’t quite the same as a short caused by insulation wearing down between two wires, allowing them to rub and causing intermittent, out of spec voltage signals to cross between them.

  • avatar
    CarPerson

    Pedal sensors shorted together do not latch an error code.
    Throttle operation was normal while driven in this fault condition.
    The technician then did something more that caused SUA.

    What additional fault(?) did the technician introduce that caused the SUA???

    If his response was that he connected a frequency generator, whack his pp.

  • avatar
    BlackPope808

    I just wanted to ask if anyone else had read reviews on the particular they were using, an Avalon. I recall while looking at getting a second hand 05-06 I perused some reviews and consumer opinions that stated a lot of people had trouble with engine racing while waiting at a light, and sudden down throttles too. If anyone else recalls these reports it is clear that there is really an electrical problem rather than just stuck pedals

  • avatar
    BuzzDog

    What’s sad is that – if indeed Dave Gilbert is both credible and has actually discovered the cause for Toyota’s woes – his findings may be tainted or ignored by his decision to present them on a “news magazine” show.

    • 0 avatar
      JohnAZ

      He initially presented his findings to Toyota who ignored him. I’d say that makes him pretty credible.

      Also, he did not show the cause of the problem. He showed A POTENTIAL CAUSE, and how to reproduce the reported symptoms, from which he expected Toyota to take over the analysis. They are dropping this ball too.

  • avatar
    ajawamnet

    Anyone actually have any info on what he did? Has he published any papers on this?

    Prof’s usually love publishing papers… typically how they get their money/cred…

    If there’s no published technical paper, ready for peer review – it’s all “marketturds”…

    Anyone help me on this? I know there’s people that hack these ECM/PCM’s; they’d answer these :

    What are the sensors? Are these pots, optical/mech encoders, etc…

    Is it run into A2D’s? If so what kind of conditioning?

    What’s in the ECM? Type of processor? Embedded proc in and ASIC? Simple uC with onboard A2D? or is it a separate system using CAN-bus or similar to talk to the main ECM processor?

    anyone actually dumped and disassembled the code in one of these? I know they can code protect some embedded proc’s but there’s ways around that… sometimes.

    Is the input to the ECM code from whatever the hell these sensors are interrupt driven or polled?

    How is the code written in the uC/uP written? Something like Wind River IDE written in C or are they coding these things in assembly? How did they deal with exceptions?

    I saw that on the Toyota site that it suggests flash based uC/uP “semi-permanent hard coded memory” They mention OS… does that mean it’s MMU based uC/uP’s that can/does run an some sort of OS?

    If so, how well did the guys writing the operating code actually know the underlying hardware…an example of poor understanding how an OS “shields” most crappy coders is most Winblows/Linux/xBSD/OSx code; esp ones that deal with drivers (can you say ATI?)

    They also mention “interference” …. I do know the EMI testing for susceptibility/radiated emissions is tough for automotive… But CAN-bus has inherently high CMR (as long as the PCB is routed/cabled correctly)

    Is the system truly designed to be fault tolerant? Multiple, redundant processors? I recall one of our coders years ago talking about the fire suppression/detection in the wheel wells of aircraft he worked on – three separate systems that vote as to whether there’s really a fire…

    I know embedded stuff somewhat, not a coder but do a lot of hardware design…from simple PIC/Atmel stuff to full blown ARM-based (both mmu and non-mmu types). Lots of I/O both digital and analog…

    Never pissed around with cars much… 5,000 years of fascination with the wheel is enough me thinks…

    lot of unknowns … I have that same resistor sub box (IET Labs) and I’m sure I could cause all kindsa faults that have nothing to do with reality (actually mine’s an RC box – has caps too.. whoo hoo!!!)

    Maybe I’ll go out and get me some banana to alligators and start mucking around with my GM PCM… that could be fun [NOT]…

    (ps to web dude – nice Ajax edit box!!)

    • 0 avatar
      JohnAZ

      You should ask Toyota for the answers to your questions. Gilbert has reported his tests to them.

    • 0 avatar
      crash sled

      ajawamnet,

      That was a stupendous display of technical jargon you wrote there. I was able to pick out and make sense of, oh, about 2% of it. I think I’m better prepared to translate mandarin chinese than your post!

      And in a typical NA vehicle program progress meeting, you’d get through about 2% of your monologue, and the Chief Engineer would stop you, and tell the team to take the discussion “off-line”, partially because even HE/SHE is confused, but also because these issues are so complex that not even senior automotive engineers should spend time dissecting them in their full complexity, as a part of standard team business. They require far more time and discussion than that.

      I think that known complexity should serve to delegitimize the significance of the professor’s minimalist ABC tv work here. I’m sure the trial lawyer guy, Kane, will point to this report in his testimony today, but in all probability, Kane has likely paid this professor or his department in some backhanded way. I’m sure the proffy has good intentions, but those intentions are being abused.

      Still, Toyota should release their black box reader, and pronto. That would be one of the wisest business decisions they could make right now.

  • avatar
    tced2

    Doing engineering work in the media or in the halls of Congress is a waste of time. This demonstration is worthless – it’s “full throttle acceleration theater”. No real engineer can tell exactly what was done from what was shown.

    There does seem to be a secondary problem implied. “the brakes aren’t capable of stopping the car with wide open throttle”. Is this true?

    • 0 avatar
      BuzzDog

      As he was driving in the video, Brian Ross said, “The brakes don’t work, the brakes give up.” But it appeared that he was able to stop the car.

      I wonder if Ross was simply noticing that the brakes become less effective at wide open throttle, due to the loss of vacuum after a few pumps of the pedal.

    • 0 avatar
      prince valiant

      I wouldn’t be so quick to dismiss the ABC story. I’ve spent over ten years investigating vehicle defects and the fact that a short between the two circuits can cause the runaway acceleration would lead me to look at the following.

      1)Proximity of the two terminals in question to each other in the connectors.

      2)Look at the connector on the pedal assembly, the ECU, and any other connectors between the pedal assembly and ECU where the circuits pass through.

      3)Is there grease placed in the connector. I’ve seen instances where minute amounts of copper/gold from the terminals flake off and the grease provides a migration path possibly leading to a short.

      4)Are the connectors waterproof? Perhaps moisture is getting into one of the connectors. Could the pedal connector be under the HVAC unit where moisture could drip onto the connector? Is the ECU mounted in a wet environment? Is it potted?

      5)Check the wiring routing from the pedal assembly to the ECU. Could the wiring be chaffing on something leading to a short. Could a misinstalled fastener in assembly be causing a short potential?

      6)Have any of the vehicles that exhibited the problem had prior service that required opening any part of the circuits in questions(i.e connectors removed and reseated, modules replaced, pedal assy replaced, etc). Could this be a service induced condition?

      After establishing root cause in many of my investigations, and than reviewing the engineering FMEA’s, I’d guess that 90+% of the failure modes I uncovered were never listed on the FMEA’s. Engineers typically address only the most obvious failure modes on the FMEA’s and analyze their system/component as if it existed by itself. As an example, did the engineer who released the throttle assembly list floor mat entrapment on the FMEA as a possible failure mode, my guess not.

    • 0 avatar
      JohnAZ

      This makes sense if the signal from the brake pedal is also compromised at the same time by whatever is causing the problem. My bet is flaky power to both set of sensors.

    • 0 avatar
      JohnAZ

      PrinceValiant,
      I would add two items to your list.

      Investigate what events, other than a cable short can cause the voltage from the gas pedal to drop/rise to the measured level. eg. Regulator output.

      With the short circuit applied, look at the voltage received at the ECU from the gas pedal. Now apply that same voltage to the ECU port for the brake pedal. What happens to braking power?

    • 0 avatar
      FreedMike

      @Buzzdog –

      He was able to slow the car down after the short was discontinued. Not sure whether the brakes were “overpowered” by the engine, or whether there was some kind of electrical fault.

    • 0 avatar
      BuzzDog

      @FreedMike: Makes sense. Not totally apparent from watching the video, at least not to me. Another reason why I’m not a big fan of this sort of “journalism”: Shows all of the drama with little or no methodology.

  • avatar
    FreedMike

    OK, I don’t pretend to be even remotely competent to comment on electrical and engineering questions, but it seems logical for me to make the following conclusions:

    1) There IS a condition under which these cars can have sudden unintended acceleration.

    2) When it happens, the car’s fault sensors will report no problem.

    This explains why dealers who look at the affected cars conclude that there was no problem – the onboard diagnostics don’t report it. That makes sense. If they knew there was a fault, they’d report it to Toyota.

    Most importantly, taking this to its logical conclusion, since the dealers could find no faults, Toyota had no reason to believe that there was a fault. To me, this suggests that this problem – assuming there is one – is the result of a mistake, not a repeat of GM’s “we knew Corvair handling would go all blooey but refused to spend the $10 a car to fix the problem.”

    Now, further testing needs to be done to figure out how that short could happen in the real world.

  • avatar
    bmoredlj

    It did look funny when the Avalon’s tach needle shot up to 6500 RPM exactly and stayed there – with nonhuman precision. But statements like “you want your wife driving this?” aren’t helpful. I don’t believe this report went too far like the 60 Minutes Audi debacle, but they were way too stingy with the details for a report that claims to “challenge Toyota’s denial” of electronic faults. Challenging is one thing, proving they’re wrong is another thing altogether.

  • avatar
    dbatoloco

    I was able to record the video and paused it right about when the rpms went up. It is very strange that the speedmeter was reading 0 and it didn’t move as the revs where going up. The camera didnt move as much when the revs went up. I have own and drive different type of car with an automatic transmission. As you press the pedal all the way down and the revs go up the transmission will downshift to a different gear and it will give you a surge that should have been noticed on the rpm and also on the movement of the camera. I was also able to see that the rpm limiter working as the needle keep going back and fort very sightly. If it were true the rpm needle will have gone over 8000 rpm or at least drop when the car will shift into a higher gear. This was a set up. How come they cut the video rigth before showing the rpm going up and then right after. If it was showing 0 mph while the rpm went up that means they just rev up the engine while in park or neutral. Actually is because the car was in Park. As you are able to see in the video that the Park light is on.
    Can the professor explain why was the speedmeter stuck at 0 mph or that was part of the short. Don’t get me wrong but I also doubt that the car going at 20 MPH will rev up so rapidly to more than 6000 rpm in such a small amount of time. They recorded the reving up the engine while in park and then inserted it into the video to make it more interesting. I also read today that Toyota is also flashin the ECU with a brake overide.
    http://www.just-auto.com/article.aspx?id=103289&lk=dm

    • 0 avatar
      Steven02

      I agree, I think this portion of the video may have been put in for effect and wasn’t part of the actual scenario of the demonstration. But, I also don’t think it makes the entire demonstration false. I hope that more details come out so that we can determine if this is rigged experiment that would never happen, or if the principles that he used to make this happen could actually happen in the car.

    • 0 avatar
      Robert.Walter

      If they are fooling around, cutting in video, as dbatoloco describes, then ABC is unnecessarily practicing dramatic fraud.

      The issue itself is both dramatic and titillating enough on its own to not need such enhancements.

  • avatar
    ASISEEIT

    Well if you haven’t watched the ABC video yet you should and make your own conclusions. Here’s the website–www.autoblog.com/2010/…/video-smoking-gun-abc-news…/2 –I would also like to point out that it takes a HELL of a lot to FINALLY get SOME people to FINALLY admit there might be SOMETHING more to Toyota’s Quality issues than UNION CONSPIRACIES!!!!!!! I also see there are the “Old Stand By’s” that claim this ABC video is a “HOAX”. I wonder if this same crowd believes the world is flat? I’m trying to find out just who this “Crowd” is and I have a few ideas. First they could be Toyota employees,2. Own a large quantity of Toyota stock,3. Silver spoon management employees of any company who have never REALLY PHYSICALLY worked an honest day in their life,4. Narcissists who feel they compare to no one,5. Braindead or a combination of all or part of the aforesaid.

    • 0 avatar
      CamaroKid

      Wow, someone needs to cut back on the coffee,

      Actually what I see on this site is a bunch of people who are trying to figure out what is going on. I don’t see anyone defending Toyota.

      Their cars are accelerating out of control and are killing their customers. That is a given, that is a fact.

      The issue is we still don’t know why. Having an assistant professor sit in the passenger seat and short out two wires to trick the computer into a balls to the wall 0-60 run proves nothing. Just about any technician who knows the wiring diagram for just about any drive by wire system could EASILY accomplish the same feet. So what.

      Until this professor publishes what he did so we can all see what he did and see if it makes sense and has validity, then this is nothing more then someone trying to get their 15 minutes of fame.

  • avatar
    wmba

    Has anybody here looked inside a Toyota ECU? I haven’t personally.

    Reason I ask is that over the last four or five years at our high-end audio store, we have found strange behavior in all manner of Eastern-made electronics.

    It’s caused by dried glue, which becomes conductive, sometimes in a few months, other times years.

    For example, you get a subwoofer or a receiver from a well-known make that goes weird, and upon inspection, you see wires, or a capacitor or some bigger component held in place with a blob of what looks like hot-melt glue. This stuff should be banned, because it causes high resistance shorts and screws things up completely. Wiggle the components around, and the problem will disappear, only to reappear later. Took our tech quite a while to figure out it was the glue itself to blame.

    Now when he sees this stuff, our tech cleans it off and cleans the PCB also, and nine times out of ten, the problem goes away.

    That’s why I ask if anyone has seen the inside of a Toyota ECU and knows whether this hot-melt-like glue is used at all. Would be especially interested if anyone’s been inside a Subaru ECU and seen this stuff, because I drive one!

  • avatar
    Barry K. Nathan

    Here is David W. Gilbert’s prepared testimony, FWIW. I think it has more detail than the ABC News report.


Back to TopLeave a Reply

You must be logged in to post a comment.

Subscribe without commenting

Recent Comments

New Car Research

Get a Free Dealer Quote

Staff

  • Contributing Writers

  • Jack Baruth, United States
  • Brendan McAleer, Canada
  • Marcelo De Vasconcellos, Brazil
  • Vojta Dobes, Czech Republic
  • Matthias Gasnier, Australia
  • W. Christian 'Mental' Ward, Abu Dhabi
  • Mark Stevenson, Canada
  • Cameron Aubernon, United States
  • J Emerson, United States