Toyota's pedal recall: Speculation


  • Discourse touched me in a no-no place

    The more I hear about Toyota's pedal problems (and see the disassemblies of all the affected parts) the more it becomes clear to me that the pedals are probably just an inexepsnvie scapegoat for some fucked ECU programming.

     

    Consider this scenario:

    Due to moisture, intentional design, random electrical voltage variations, whatever, the accelerator pedal "sends" (it's actually a passive magnetic field sensor, but whatever) an out-of-range value.(Let's assume 256)

    The ECU receives this value and sets the throttle position. The code responsible for setting the throttle position (whee, more magnets) dutifully jams the throttle right the hell wide open (hey, you asked for 256, I can only physically go to 255 - so you get 255).

    The ECU then takes the value it received from the pedal and tries to do some calculations with it. This either overflows a buffer somewhere and wedges the thing right away, or produces another garbage, out-of-whack number which in turn gets sent on to another function which wedges when it receives out-of-spec values.

     Your car is now running at wide open throttle, the "read the throttle position and react to it" loop is wedged, and your car is stuck in a runaway situation. Go stomp on the brake pedal, or shift into neutral or something.

     

     

    Now, Toyota's proposed fix is to take the accelerator pedals and stick a plastic shim in them to reduce the range of travel. This works because in 99.999% of the cases, the faulty value is being sent when the pedal is at or really close to the full travel position (when it would send 255) - so lets say this shim reduces the maximum position to 250. That's a lot more room for error caused by bad voltage, condensation, whatever. Yeah, the customer can't intentionally hit that last bit of throttle position, but we're talking about Toyotas here - they aren't exactly sports cars. As a result, the only way anyone is going to send a faulty value to the ECU is with aftermarket equipment (not their problem) or a a seemingly random event which can easily be written off as user error.

     

    Why would Toyota fix the pedals instead of just modifying the code? The unit cost for one of the pedal assemblies is $15 ($120 retail, but that's beside the point).  It can be pulled in about 45 seconds and either replaced with a pre-modified part, or modified in about 5 minutes, and reinstalled in about 45 seconds.

     Fixing the code on the ECU would require fixing it in code, testing the everlasting piss out of it (which takes time and costs money), fixing any issues that arise during testing, etc. - the more they cut corners on processing resources, the longer and more expensive this will be - and then they have to get that code onto ECUs. This, at a minimum, would involve pulling the ECU, taking it into the shop, pulling it apart, attaching an in-place flash harnass (not standard equipment at any car dealership) to the chip in question, flashing it, hoping and praying it worked properly, reassembling it, and reinstalling the ECU. At worst (and this is the most likely case, because if I were going to build a safety-critical system like an ECU, I'd be using burn-once chips so the customer can't fuck with it) it requires replacing the ECU - which costs a damn hell of a lot more than $15.



  •  I have an '85 Volkswagen and what is this



  •  I think we can be glad that this just happened to puny little Toyotas (and technically identical Citroens and Peugots as well). Imagine a BMW or Porsche with a similar problem.



  • @ammoQ said:

     I think we can be glad that this just happened to puny little Toyotas
     

    Dude, just imagine the terrible destruction unleashed by an unstoppable, indestructible Toyota pickup.


  • Discourse touched me in a no-no place

    @ammoQ said:

    puny little Toyotas (and technically identical Citroens and Peugots as well)
    And Pontiacs.



  •  I find it more likely that the US government (which owns GM) is making a big deal out of a small issue.  Most likely all the stuck pedals were caused by chewing gum the owner had dropped down there sticking to the brake pedal.  By putting the plastic on the there they stop the chewing gum from adhering to the pedal.



  • @ammoQ said:

     I think we can be glad that this just happened to puny little Toyotas (and technically identical Citroens and Peugots as well). Imagine a BMW or Porsche with a similar problem.

     

    Cost-benefit as always. I thought software in cars was suposed to be tested to levels that we only read about in fairy tails .

     

    Also why would sending 256 vs 255 be a problem? I mean the car "steps on the gas", but that does not mean you can't remove your foot from the gas pedal. If the value is significantly higher than what you pressed then perhapse it is a hardware issue, if one piece can't fix it they might need another piece of hardware for fault correction.



  • @astonerbum said:

    Also why would sending 256 vs 255 be a problem? I mean the car "steps on the gas", but that does not mean you can't remove your foot from the gas pedal. If the value is significantly higher than what you pressed then perhapse it is a hardware issue, if one piece can't fix it they might need another piece of hardware for fault correction.
    His theory is that anything above 255 triggers a bug that causes the throttle not to be released when the pedal is.



  • @belgariontheking said:

    @astonerbum said:

    Also why would sending 256 vs 255 be a problem? I mean the car "steps on the gas", but that does not mean you can't remove your foot from the gas pedal. If the value is significantly higher than what you pressed then perhapse it is a hardware issue, if one piece can't fix it they might need another piece of hardware for fault correction.
    His theory is that anything above 255 triggers a bug that causes the throttle not to be released when the pedal is.

     

    Hmm... Wow. So just fixing/testing is such a problem :P. I guess its the nature of the beast when your software is responsible for keeping people alive.



  • @astonerbum said:

    Hmm... Wow. So just fixing/testing is such a problem :P. I guess its the nature of the beast when your software is responsible for keeping people alive.
    According to Weng's theory, fixing and testing is not a problem per se, just more expensive and more difficult to implement than the plastic shimmy solution.  The theory further theorizes that the average Toyota dealer does not have this.

    Did you read the OP?


  • Discourse touched me in a no-no place

    @belgariontheking said:

    @astonerbum said:

    Hmm... Wow. So just fixing/testing is such a problem :P. I guess its the nature of the beast when your software is responsible for keeping people alive.
    According to Weng's theory, fixing and testing is not a problem per se, just more expensive and more difficult to implement than the plastic shimmy solution.  The theory further theorizes that the average Toyota dealer does not have this.

    Did you read the OP?

    Apparently my assessment of the cause may be correct, but my analysis of the fix is a little off - Toyota IS issuing an "unrelated" code update while they bust open your pedal assembly and stick some plastic in it. They had the foresight to allow code updates over OBD (which is on one hand ABSOLUTELY TERRIFYING, and on the other a fairly good idea) so tooling and expertise isn't a problem.

     

    So they're trying desperately not to tarnish the public's opinion on drive-by-wire systems, while actually fixing the problem properly (albeit quietly).



  • @Weng said:

    (content omitted due to crap)

    TL;DR, Toyota didnt do their job and do error checking and bounds checking because they made the asspied assumption that nothing would ever go wrong. We've all known Toyota engineers are promoted when they commit hari kari.



  • @Weng said:

    The more I hear about Toyota's pedal problems (and see the disassemblies of all the affected parts) the more it becomes clear to me that the pedals are probably just an inexepsnvie scapegoat for some fucked ECU programming.

    I agree, and I think Steve Wozniak has engaged in similar speculation ( http://news.cnet.com/8301-13924_3-10445564-64.html ).

    Woz: "This is software. It's not a bad accelerator pedal."

    Incidentally, having purchased both Japanese and American cars brand new, my feeling is that the Yanks are much better at getting the firmware right. Japanese ECU programming tends to lug the motor (i.e. select the wrong gear), have trouble in the extreme cold, etc. The program's just not as good. A working Ford EFI system will give a perfect cold start even 10-15 years after the car's manufacture, and will never lug the motor. Say what you want about US manufacturing, the USA remains the world leader in software / firmware.

    Now, there are other areas where the Japanese excel, e.g. achieving both soft ride and good cornering. But I've always considered Toyotas garbage, and I feel a bit vindicated, although tragically so.



  • re. puny Toyotas: I have a toyota. It has twin turbos and produces 330 hp. Puny, it ain't. It starts first time in -13 degree temperatures and has never, ever had any problems whatsoever it over ten years of ownership. The real problem we have here is the owners. Anyone who finds that a stuck throttle is dangerous simply does not know how to drive/think. If they are so incompetent that they then die as a result, then I say "good riddance to your defective genes, and thank you for cleaning the human gene pool". I once had a stuck throttle on a Chevy Nova with a 350 hemi. Full throttle down the I95 for about ten miles controlling the thing by turning the ignition on and off. It was fun!!!


  • Discourse touched me in a no-no place

    @bridget99 said:

    Woz: "This is software. It's not a bad accelerator pedal."
    Actually, he's referring to yet another Toyota issue totally unrelated to the primary "pedal recall". This one is unashamedly a firmware problem within the cruise control programming. It's easily reproducible on every 2010 Prius.

     

    A quick rundown of current major Toyota issues:

    1) FJ Cruiser has structural stability problems with offroading and tends to "snap in half" if you articulate it in certain ways and cracks its own windshield with alarming regularity. This is despite the FJ Cruiser being marketed as the "offroader" of the lineup. (Problem not acknowledged, no recalls issued).
    2) Toyota Urban Cruiser (AKA Scion xD) fails miserably at crash testing in the EU. (Problem denied, "it tested fine in the US. The EU tests are flawed")
    3) Various Toyota and Lexus models feature a stupid floor mat design that can catch pedals (Problem grudgingly acknowledged and two different totally different kludges, one involving sawing off the pedal and another involving securing the floor mat away from the pedals)
    4) Various models in VAST numbers (including many from #3 but including other Toyotas and Lexuses, additionally including various Scions and cars built for Pontiac and Renault among others)  have an unintended acceleration problem, likely a software issue (as discussed above in this thread). (Problem grudgingly acknowledged, scapegoated as a hardware problem. A software fix is also being applied during the hardware recall process with no reason given.)
    5) Pre-2010 Prius regenerative braking/conventional braking changeover lag. A stupid software design decision creates a slight delay between the engagement of one braking system and the disengagement of another. (Problem grudgingly acknowledged, and a software fix is being issued. The fix consists of flipping the order of the two operations, so both brakes now engage concurrently for an instant, creating an equally stupid feel to braking, but in the minds of Toyota "more braking than commanded" is better than "less braking than commanded".
    6) 2010 Prius cruise control runaway. This is what Wozniak was talking about. Readily reproducible. (Problem not acknowledged.)

    In short, it's hard to find a vehicle Toyota's made for the US and EU markets in the past 4 years that doesn't either have a problem known in the automotive community or is under recall.


  • Discourse touched me in a no-no place

    @zedhex said:

    re. puny Toyotas: I have a toyota. It has twin turbos and produces 330 hp.
     

    It also wasn't built in this decade.

     

    Toyota doesn't build cars like that anymore. They killed the Supra. They killed the MR2. They killed the Celica. They killed their entire offroader lineup (by either making them into fat softroaders or making them break). They make nothing of any interest to people who actually like cars.

    They decided they wanted to be the biggest automaker in the world - and in the process made many of the same mistakes GM made - degrading everything except their bread and butter (for GM it was SUVs, for Toyota it's Generic White Goods cars) - foregoing quality because of their reputation, etc. However, at the same time the press was holding them out as an example for what GM should be - so they're paying for it dearly with the seemingly overreactive backlash now that they've been exposed for doing the same damned crap.

     

    Yes, anyone that actually gets killed by a stuck throttle is a moron. However, most drivers are morons - especially the sort of driver that buys a Generic White Goods car - which is all Toyota makes anymore. They don't care how it works. They CERTAINLY don't have a clutch pedal. They have no idea what neutral does (actually, Toyota's shift gate setup is really stupid and it's easy to try to shift to neutral and actually hit the manumatic gate thereby giving you the impression that neutral isn't working). They're confounded by the fact that stabbing the pushbutton ignition doesn't shut the car off while it's in drive - you have to hold the damn thing for 3 seconds - which feels like an eternity in a panic. Hell, holding the power button on my computer for 3 seconds to hard-shutdown feels like a damned eternity, and I'm sitting on my ass doing nothing life threatening. Other manufacturers (Including bloody fucking GM) thought ahead and stabbing the ignition button 3 times, like you would instinctively if just pushing it didn't work and you were panicking does the job.

     

    I also question why your Nova had a 350 hemi. A MOPAR engine in a GM car? A 350 small block perhaps.



  • @Weng said:

    @zedhex said:

    re. puny Toyotas: I have a toyota. It has twin turbos and produces 330 hp.
     

    It also wasn't built in this decade.

    I also question why your Nova had a 350 hemi. A MOPAR engine in a GM car? A 350 small block perhaps.

     You know your cars Weng, it was a small block (and more than fifteen years ago), I stand corrected. My Toyota is a 1994 grey import Soarer - and your right about Toyota giving up on the performance front. Such a shame, as the Soarer, MR2 and Supra were wonderful cars. The only bright spot on the horizon is the Lexus LFA, but who has that much to spend on a car anymore?



  • @Weng said:

    They're confounded by the fact that stabbing the pushbutton ignition doesn't shut the car off while it's in drive - you have to hold the damn thing for 3 seconds - which feels like an eternity in a panic. Hell, holding the power button on my computer for 3 seconds to hard-shutdown feels like a damned eternity, and I'm sitting on my ass doing nothing life threatening. Other manufacturers (Including bloody fucking GM) thought ahead and stabbing the ignition button 3 times, like you would instinctively if just pushing it didn't work and you were panicking does the job.

     

     Why in the name of hell would you want the ignition to turn off while you are driving?   If you are having runaway acceleration, that's certainly a really bad time to lose power steering and power braking.

     NOTE:  I'm not refering to hybrids which I assume have power steering while the engine is "off".  

    NOTE:  I'm further assuming that cars other than hybrids have push button ignitions.



  • @tster said:

    If you are having runaway acceleration, that's certainly a really bad time to lose power steering and power braking.

    If you're traveling with any speed at all, steering isn't that hard without power steering.  It's harder than with power steering, but not by much.  Power steering is really only strictly necessary when you are completely stopped and turning the wheel is really, really difficult.



  • @morbiuswilters said:

    Power steering is really only strictly necessary when you are completely stopped and turning the wheel is really, really difficult.
     

    Such as when driving this.


  • Discourse touched me in a no-no place

    @tster said:

     Why in the name of hell would you want the ignition to turn off while you are driving?
    I never said it was a good idea, but if you're so dumb you've gone and overheated your brakes to the point of uselessness by not stomping them hard enough (brakes on any modern car WILL overpower the engine at WOT, but you only have like 1 and a half chances to get it right - and it requires a DAMN LOT of pedal force to do it), and you're too damned stupid to be able to find neutral on the overcomplicated shift gate, it's just about your only recourse.

    Mind you, the brakes are also brake-by-wire with only a non-boosted backup, and the transmission is completely electronic - so it's entirely possible that moving the shift lever in at least some cases does nothing, and that you're dealing with unboosted brakes anyway - since we're prettymuch accepted we're dealing with shitty software.


Log in to reply
 

Looks like your connection to What the Daily WTF? was lost, please wait while we try to reconnect.