This guy sounds like a typical old cretin who has lived past his useful life on this planet and has nothing meaningful to contribute to humanity. What is wrong with the reliability of computers? They are brilliantly reliable. Maybe old uncle Eugene is in the habit of installing a lot of crapware (meganfoxsexvideo.exe), which thankfully is not an option in cars.

And what is your certainty based upon?

My lack of ignorance on matters pertaining them computers.

I don't have quite that amount of faith in computers. And I've had an awful lot in the last 30 years. The software is only as reliable as the programmers can make it. Ditto the firmware. And the hardware itself can cause problems. The computer I'm typing this on occasionally, for no obvious reason, would shut down. No pattern, could go for months without a problem, then it could shut down four times in two hours. Not fun.

The cure in the short term was, as always, reboot.

Eventually traced to problems with hardware (motherboard) and formware (dicey programme on chip) not quite enjoying each other's company. New MB and we're friends again.

You must be kidding. This is an actual joke that went around the Internet a few years ago: What if your car worked like Windows? Twice a week or twice a day (depending) the instrument panel would go all blue and the car would just stop in the middle of the street. You would have to pull over to the side, shut everything off, turn everything back on, and see if it started to work again. If not, call tech support and wait on hold 45 minutes, whereupon they would tell you to turn everything off, turn everything on, etc.

A MS Windows machine and a Lexus IS350 do have this in common: to shut them off you push the start button.

Now, a story. A few years ago my next-door neighbor (retired engineering VP/COO, major Tier 1 OE supplier) bought a new MB SL600. He noticed a number of apparent hinks or at least oddities in the car's operation; the three he cared about related to transmission shifting, cruise control, and nav if I recall properly. Didn't represent a safety hazard or prevent driving the car, just small yet clear shortcomings in the vehicle's operation/performance. He reported them on his first service appointment with the response being NPF (no problem found). He next went to the service manager, then the area service rep, then the MB district manager, then the head of NA field service engineering, who brought two of his people to drive and inspect the car. On the last two appointments he asked me to come with him. Now here is the interesting part: None of these totally competent people (nor me, nor the car owner for that matter) could at all determine if these were normal operating characteristics of the vehicle, problems that could be fixed, or built-in glitches with no known correction. Yes, the car does this. We don't know why. Couldn't say. Beats us. Threw up their hands.

When you build cars at the current level of complexity, there will be ghosts in the machine.

New article on worldwide repercussions of this situation:

http://www.sniffpetrol.com/2010/02/05/god-...-recall-humans/

I've never owned a car that I have wanted to attack with a pick-axe, unlike my computer, which much of the time behaves as though someone else was in charge. Perhaps Macs are different, but my Dell PC with the disgaceful Vista crap is woeful.

I agree with McG. You're nuts, or you're trolling the forum.

This is why windoze is not part of the flight control software on a passenger jet, nor is it in the control system of a destroyer or nuclear submarine.

"Hey, we have a problem."

"Reboot the computer"

<big smoking hole in the ground>

Not to say windows is in the ECUs of the Toyotas, but the point being the computer "system" is only as good as the programming of firmware/software, system-level FMEA (including effects of hardware, software, and firmware), and the test protocols to validate the design.

This piece reminds me of the 15 year-old who thinks you are a moron because you never heard of his favorite band.

Well, what about the cases in which there were no floor mats in the car?

What about the New Jersey case in which Toyota dealer personnel reached in and manipulated the pedal by hand and the engine stayed at ~WOT?

That's what the widow of the guy hit by a meteor while juggling ducks would say, "We have a very real being hit a meteor while juggling ducks problem." Don't laugh, it is certainly real for guy hit by the meteor.

While i have owned cars that I do wish to attack with an axe my PC and internet connection causes more grief than any car. While trolling this forum this morning it has frozen twice and is running very slowly again. Combination of lousy programmes and junk connection. My so called hi speed broadband drops out, goes slow half of the time. And I live 4km as the crow flies from the centre of Adelaide, the capital of South Australia.
It makes Toyotas so reliable in comparison!!

Ah good.
(Quick thread sidetrack to make a point sorry guys)

If you're good with computers, then you should be able to explain how the Qantas Airbus A330 did two huge uncontrollable nose-dives from high altitude when flying from Asia to the east coast of Australia, around Learmonth on the west coast.
Without warning the aeroplane had the nose pitch down between 15° to 20° and no amount of fiddling with the controls by the crew could stop it. It pulled out of the dive twice all by itself fortunately and the crew elected to land at Learmonth instead of continuing another four hours to Sydney.
http://blogs.crikey.com.au/planetalking/20...c-interference/

My point being that even supposedly extremely reliable systems still fail no matter how safe they are designed to be. Fair enough it's pretty rare on the FBW aeroplanes though.

When humans are involved in the design process, the product will invariably prove to be fallible. There is some evidence to suggest the same is true even when the product is designed by divinity.

10 years back we were lead program on a new iteration of engine management computer. The proto installations were very prone to EMC problems, to the extent that the cars were virtually undrivable, which rather reduced their usefulness!.

As a quick hack the programmers inserted a very fast reboot routine. So although the computers fell over a lot (once a second or so in some circumstances) the car kept going. As a result a whole generation of development engineers learned what an engine computer rebooting sounds like - basically a misfire. To this day when I hear a misfire I suspect that it is the computer rebooting.

The reason I am inclined to blame the software is because writing robust time critical software is very difficult, and all the safety related software is carefully designed, and then tested very very thoroughly, for hardware interactions. My guess is that throttle software is not regarded as safety related, and would not be tested as thoroughly. Rig time is expensive, things have to be prioritised.

It would be strange if those remote odds did prevent it from happening, because then they would be zero and not remote.

I wonder if there's any statistical evidence to show that the inevitable malfunctions of TBW (given that no technology is perfect) are any more common than those for purely mechanical throttles? If the TBWs malfunction statistically less often than mechanical throttles did perhaps we should be lauding them for making us safer than demonizing them for their inevitable imperfections.

If even one life- wait no- one child's life can be saved by throwing billions at these vanishingly rare malfunctions...

This may be little more than pure irrational hysteria.

Oh, is now a good time to mention that some car makers (Toyota included) are mucking around with fully-electric steering? That is steering with no mechanical connections at all.
*shudder*

Might be safer than a mechanical steering system. Desmo's point is spot on. It's the same problem with autopilots for cars, they'll be held to a much higher standard than human drivers.

Until a floor mat gets stuck under it.
I doubt it'll be safer, but I hope I'm wrong.

Hmmm. One in 'x' random problems is one thing, but if it is the same problem recurring, that is a systemic problem that is susceptible to examination, resolution and solution.

Is this symptomatic of the F1 debacle - we know best, don't tell us how to do our business we are TOYOTA (cue: all bow)

There sure is. Toyota's SUA incidence rate spiked with the introduction of ETC. With several separate models. For example: on the Tacoma, SUAs increased 20x when ETC was adopted. There is a clear, unmistakable pattern. Toyota's rates spiked again for MY2008, which is partly the basis for NHTSA's concern.

How? The current system has an enormous experience base, failure rates are extremely low, and sudden catastrophic failure is extremely difficult to achieve and nearly unheard of during the vehicle's normal service life. Now let's replace it with a new, unproven system with which the auto industry has zero expertise, no experience base, and with none of the manufacturing and maintenance controls mandated by law in every other industry that employs such critical systems. In high-volume, mass production (often without adult supervision). And here, catastrophic failure is easy as pie to achieve. So how can fly-by-steering possibly be safer, in theory or in practice?

Give me mechanical steering at all times. Basically trouble free and simple and when it goes wrong it normally gives notice. Unlike electric models. They are seeming to have a lot of problems with that type of thing on aircraft where the standards are far higher than the motorindustry,
As this discussiion is about is that Toyota [and others][ cannot make a simple throttle that works fly by wire.Which most of the time when it fails makes the vehicle stop, unlike an electric stering which will cause it too crash!
Just because it can be done is no reason to take away a simple reliable proven system and go hi tech.

No, you are reversing incidence and causality. Your math is ok but your logic sucks.

Not only that but the consequences of a failure with a mechanical system is fairly limited and immediately recognisable; either the steering jams or it fails to respond to input. DBW steering can do all sorts of crazy, subtle things when it fails. In fact it may fail so often and so subtly that nobody even knows there are any failures at all. And if a failure ends up by killing the driver then there will be no evidence left behind to suggest a problem with the DBW system. Maybe the OEs view that as convenient side benefit, I don't know.

My first exposure to DBW throttle resulted in what I would call a failure; it responded inversely to my input for a few seconds. I will never knowingly drive a car with DBW steering.

Suck it up McG, the logical failure was all yours;

Didn't Mercedes Benz have on some of its models a few years bake brake-by-wire?

Me too - As I said early, use the most simple machine to do the job.
There's no doubt technical merit in having electric steering but is it worth all the extra modes in which it can fail?

I was just making fun of your statement that was devoid of any meaning (as is your rebuttal). If some event is one in a million, then noting that this "supposedly" rare event happened to ten people (out of ten million) is not a very useful observation.

All cars equipped with an anti-skid control like ESP have a sort of brake-by-wire built in. And of course, there are already cases where this system actually led to fatal accidents, be it that the spin-sensor was connected wrongly or simply acted up....

Zoe

As for electrical steering; there are also reports that some Volkswagens showed spurious steering activity - was traced down to faulty capacitors in the control unit.
Having seen a few of VWs electronic units myself, I am scared to death by the idea, that a control unit with such a lousy quality has total control over the steering of the car....

Zoe

The way I understood it was that M-B had some cars where there was no direct hydraulic link between the pedal and the brakes. That the computer sensed the the amount of brake being applied and then calculated the amount of pressure to apply to the system.

Or was I just mmisinterpreting what they said about their ABS/ESP system?

No, the point went straight over your head. The odds of a particular Toyota -- yours, let's say -- exhibiting SUA cannot be determined as the cause has not been identified. Without causality the incidence rate is meaningless. The possibility of your car having the defect could be zero or it could be 100 percent.

I can assert without fear of rational contradiction that every single car equipped with the necessary defect (component, configuration, calibration, whatever) is capable of the failure, without exception, and that those cars built without the defect are not. So until they identify the cause of the failure along with which years, which models, which plants, which vendors, etc, Toyota can't tell you that your car is or is not safe to drive without fear of SUA. They don't know that. They have no freakin' idea. For all they know, the next guy on the telephone could be a primary candidate.

Another way to underline the logical fallacy on parade here: in their calculations, people have just sort of taken the assumption that all 14 zillion Toyotas produced in the last n years are equally capable of exhibiting the failure. No, it is silly to assume the defect is evenly distributed across the entire fleet. In the first place, the defects would be far more prevalent. Equally silly to assume they are randomly distributed, as patterns of greater incidence have already emerged, in Tacomas and Camrys for example. There are also the problems of multiple failure paths and system noise, which are always an issue in these deals.

My advice to Toyota owners is not to worry but to be mindful and know how to handle an SUA. The last is more important than any of the rest and no matter what make of car you drive.

So, by your logic, the probability of drawing an ace of spades from a shuffled deck is not 1/52. It's either 0 or 1, depending on how the deck is shuffled.

It is perfectly valid, with limited information, to determine the odds from the incidence rates without regard to causality. It's done in insurance, which is my field of expertise. If you have no way of knowing whether you got the good Toyota with zero odds of failure, or a lemon Toyota which is certain to fail, it's is perfectly valid from the statistical point of view to use the probability of you having a lemon car as your probability of failure. There is a potential to refine those odds once you can tell a lemon car from the non-lemon one, but until you can do that, there is nothing wrong with doing the first thing.

Good to see y'all missed the point on SteerBW.

The car is a system. It is perfectly possible to install a complex system that is more unreliable and still end up with a safer car. Think ESC, it MUST be more unreliable than not having ESC. Yet according to the people who care (the insurers) ESC cars are cheaper to insure.

In the case of SBW you eliminate the steering column. This means the drivers side airbag can be dialled back to become as gentle as the PAB. That's a plus. You also have the possibility of allowing the car to steer itself, something which Volvo have identified as a necessary technology if they are to achieve a zero fatality rate. So, you need to balance the additional deaths from the car failing to steer, vs the reduction in fatalities by being able to exploit advantages from the new system.

The Mercedes failed demonstration of brake by wire has effectively killed that tech for about 5 years.

BTW I'm not saying that I am necessarily in favour of either of these ideas, but the reactionary "it is more complicated therefore it is stupid" response is just plain weird, in this crowd.

How Greg? McGuire's question is still relevant;


There is not enough experience with SBW to be able to get reliable statistical information on its risks. Even once you have that data, AND insurers find that SBW is worth the trouble (not necessarily to my benefit), do you still think it is OK for manufacturers to accept to be the cause of an increased incidence of crashes in exchange for a better statistical record of reduced injuries?

Depends on how the money flows doesn't it?

I don't think that's a good point as such, because in that particular case if the ESC fails then you have a basically 'normal' car. With the ESC working the car is safer so there's no real way that the ESC can make the car less safe than the base standard if it fails. I can see how some odd failure mode could make it worse but I'm assuming a simple overall work/no work type of failure.

Assuming the hypotheticals work that way, of course. It's a no brainer really. The idea of safety is fewer and less severe injuries. If that can be accomplished by TWB/STW, it'd be stupid to oppose its widespread implementation due to irrational ideological or technophobic emotional reasons.

Why not, what's the problem? Why should any new invention be required to introduce absolutely zero new risks, regardless of how much it mitigates the old risks that are so accepted that no one thinks twice about them?

If improved safety is the only upside of SBW then it better damn well be there!!!

To be honest I think if Mercedes hadn't queered the pitch then BBW would have probably snuck in, as people still don't understand how ABS really works, and were probably getting comfortable with the concept. However that's all a bit hypothetical now. The way the rules are written at the moment SBW is verboten, off limits, shall not be done.

OK, another example -airbags. There's a risk of one going off in your face and smashing your glasses into your eyes and breaking your thumbs even if you aren't in an accident. But they were introduced anyway, because people carefully worked out that there was a net benefit, even though there was no large scale experience that all the interlocks would work.

Here's some more food for thought http://news.discovery.com/tech/toyota-reca...tware-code.html

I don't think anybody cares how DBW works. The concern throughout this thread is about how it fails.

A general point that some posters have not considered is that DBW, SBW, BBW and all their relations are not really alternatives to traditional manual systems in the sense that the parts being replaced are not failure prone components (Throttle cable or linkage rod, steering column etc). In each case a simple mechanical component is being replaced with a complex system of sensors, actuators, wiring, control electronics and software. The rest of the system (throttle pedal, return springs, butterfly and shaft etc) is still there and still a source of failure.

What is the probability that the new system will have lower failure rates?

Unless someone here can identify how a 99.9990% reliable system materially differs from a 99.9998% reliable one then that is a bit of a tricky question to answer. Yet one will cause 5 times as many failures as the other.

Another issue is that modern cars need controllable overrides over the manual inputs (ISCs, cruise control and the like). so you don't really gain in simplicity, you actually have parallel and series systems working with or against each other even with the original throttle cable and carby setup. Which is easier to debug - a system that uses a stepper motor to control a throttle blade in response to cruise, throttle pedal, and emissions requirements, or a throttle cable that then has separate cruise control, ISC and so on, controls?

The way we look at this stuff is an FMEA, until you've got a good quality FMEA of both systems then it is very hard to decide which is more reliable.

Adding to the "software" argument. Whether or not the failures can be attributed to SW, the fact remains that some failure modes can be almost eliminated with better programming eg

WOT + Brake input => Idle mode or engine cut.

Or how about

WOT + Brief press on start button => engine cut.

"B" because it is easy to disable the cruise and ISC to test what remains.

I take your point on cruise, but ISC doesn't normally need sufficient authority to produce SUA's of the magnitude we are discussing.

I agree, you just need to look at the number of re-calls related to sticking throttle/cruise control issues in recent years. The Subaru one a couple of years back was a good example. When using the cruise control the main throttle cable could jump off the side of the guide wheel and jam the throttle when the cruise backed off or was turned off. I have seen a number of cars using a similar guided wheel with two cables side by side.
I have never had an electronic throttle stick (yet!) But I have had a lot of cable and linkage throttles jam.

A funny one.... My old Ford F100 suddenly had the throttle stick and the motor cut out at the same time. A frayed wire on the throttle cable poked out of the outer cable and shorted on the back of the fusebox. It welded to a fuse which jammed the throttle and then blew the fuse.
The funny bit was when I reached under the dash to see why the throttle was stuck and burned my fingers on the glowing red cable.

That comes close to calling me an out and out liar, which I'm not...

In the case of the car my father bought new and my sister now drives, there is definitely a problem which has nothing to do with floor mats.

Not only that, it would be very difficult for any Corolla with RHD to have a floor mat problem (of the entrapment kind Toymotor describe) as the pedal goes into a recess in the floor when depressed.

Nor is it an issue where the pedal simply failed to return, as it has repeatedly accelerated of its own accord. No foot on the pedal, it just took off.



Did the police investigation stop when the guy from the dealership said he'd had that problem with that car the day before? Or when they allied that 'information' with Toymotor saying this was a common complaint?