Toyota stop sale

[Urbane Guerrilla]

Originally Posted by tw

That 'lock the key inside' was a common problem with all GM products. Even the original Honda's made that problem not possible. If you locked the driver's door, then shut it, the door would unlock.

Unfortunately that solution no longer exists with Hondas. Now you can accident[all]y lock yourself out of a Honda.

AFAIK Volkswagen still idiot-proofs their driver's doors with that feature. The lock button will disengage if the driver uses it and then shuts the door. You have to leave by some other door to lock the keys in the car. AAA has a card-key gizmo for getting back into your car; it's like a thick credit card, you keep it in your wallet. At need, pull it out, flex the plastic key out of the plane of the card, get in your car.

[classicman]

Honda Motor Co. says it is recalling about 410,000 Odyssey minivans and Element small trucks because of problems with the brake pedals.

The recall includes 344,000 Odysseys and 68,000 Elements from the 2007 and 2008 model years.

Honda says the brake pedals can feel soft to drivers and must be depressed closer to the floor than usual before the vehicles will fully stop. The condition worsens over time but affects only some of the vehicles.

The company says the problem happens because a part in the stability control system can let a small amount of air into the braking system.

Honda says owners should take their vehicles to a dealer as soon as they get notification from Honda. Letters will go out at the end of April.

Link
Now its Honda's turn . . .
Can they do anything else to make GM's look more attractive?

[Shawnee123]

glatt predicted this would keep going and affect other carmakers.

[Pie]

Speaking as a firmware engineer, I have been waiting for this sort of computer-initiated failure for a looooooong time. It's not mechanical; it's the almost mystical capability for computers, sensors & computer code to interact in a non-reproducible, non-deterministic fashion. We call it 'emergent behavior'.

Acceleration problems likely are caused by numerous factors, said Ingolf Krueger, a professor of engineering and computer science at the University of California San Diego. “It’s possibly an entire sequence of events that has to come together for the conditions to actually occur,” he said. “There are literally millions of possible combinations for these signals.”

When your computer blue-screens, you can reboot it. (Ask lj.) When your car insists that nothing is wrong and refuses to be rebooted, while doing 90mph down the interstate... well then, you're screwed.

CAVANAUGH: I’m interested, when you’re trying to find a mechanical problem, you know, you can do a crash test or you can do a stress test, how do you find a problem in the software?

DR. KRUEGER: That’s a very deep question. So, you know, there are many things we actively do in our software, whether it goes into the car, and I have colleagues at the computer science department who are working specifically on finding bugs in software. What do you do? You actually – you can test it but you can also work with tools that check every possible state of the system it can be in. Whether it is in a valid state or whether it’s in a state that will lead to a failure. And, you know, there are computer science researchers who are developing these tools to actually do these checks very efficiently. It’s very complex because you have 80 electronic control units who all can interact with one another, so the number of possible states the system can be in is – it explodes and so you have to apply strong mathematical reasoning to find these errors.

Then add the fact that mathematical modeling cannot possibly encompass everything that might happen to your particular car -- crosstalk on cables, manufacturing irregularities, cosmic rays fer chrissakes! Battery levels, failing alternators. Squirrel nests.

I think I better buy a bicycle.

[jinx]

"Can't be reproduced" is what they always say to me the first time something fucks up in my jeep. I don't think they try very hard.

[glatt]

It would be helpful if the media, you know, told people how to stop a car that was having this happen. The media is happy to get all breathless saying how terrible these incidents are, but I haven't heard anyone say to just switch it into neutral, pull over, stop, and turn it off.

I can understand parking lot accidents where there isn't time to react before hitting anything, but some dude on the highway who can't stop his car is incompetent.

[Pie]

Glatt -- they have tried this. The car 'believes' it is not in drive, and won't respond to the shifter. Since everything is computer controlled (steering, gas, brake, transmission), there is no direct way for the operator to force the car to stop.

If, however, the problem is mechanical (stuck gas pedal or floor mat jammed) this will work just fine.

[glatt]

Originally Posted by Pie

Glatt -- they have tried this. The car 'believes' it is not in drive, and won't respond to the shifter. Since everything is computer controlled (steering, gas, brake, transmission), there is no direct way for the operator to force the car to stop.

If that's the case, then why does Toyota's recall fix this by mechanically adjusting the gas pedal?

[Pie]

Because they're in denial, and can't afford to track down all the code problems. They're hoping people will go away mollified and won't sue. It's called the placebo effect.

[glatt]

Are you basing what you are saying on the ABC/David Gilbert test where they rigged a car to suddenly accelerate by doing very specific things to the wiring? If so, I thought that had been generally debunked.

Or are you talking about something else?

I didn't realize transmissions in cars were controlled by wire these days, but I'm out of my element there, driving an older model manual transmission car.

[classicman]

Are these computers different than those in other makes? Do they all have their own brand of computer or is this more of a systemic problem?

[Pie]

In the Prius, the transmission is electronic, not mechanical.

Glatt, I was referring to the San Diego story from last week, though that's shaping up to be uncreditable.

Your 'debunking' article also has this conclusion:

In no way does any of this exonerate Toyota with regard to a potential software problem. It's too early tell if a errant black box is the cause and we still have no evidence one way or the other on this count. This also doesn't discount the possibility of some other electrical (such as electromagnetic interference) or mechanical problem that we aren't aware of at this time. So the mystery continues...

Toyota stands to lose billions if they don't control the spin of this story.

Drivers Complain That Toyota's Fixes Didn't Work
At least 15 Toyota drivers have complained to U.S. safety officials that their cars sped up by themselves even after being fixed under recalls for sticky gas pedals or floor mat problems, according to an Associated Press analysis.
The development raises questions about whether Toyota's repairs will bring an end to the cases of wild, uncontrolled acceleration or if there may be electronic causes behind the complaints that have dogged the automaker.

[glatt]

Yeah, Toyota is certainly biased. No argument here. But so is the media. They are desperate for profits, and nothing sells papers or tv advertising like a good scandal.

[xoxoxoBruce]

Why is this not happening in other countries?

[tw]

Originally Posted by classicman

Are these computers different than those in other makes? Do they all have their own brand of computer or is this more of a systemic problem?

Every automaker has its own computers. GM used a stripped down variation of the Motorola - forgot if it was the PowerPC or 68000. Many of the early problems because GM used a measuring standard that was different from Motorola and other semiconductor industry benchmarks.

Ford used a customized design from Intel. One of the computers I once tried to design with was the Intel 80196 - a variation of the computer also found in Fords then.

It always amazed me. Japanese automakers used computers about as powerful as the computer inside every keyboard. GM used a more powerful computer equivalent to that in Apple computers. So Japanese engines therefore had superior control and performance. Go figure.

One feature finally appearing in cars is the CAN bus - pioneered by Intel, if I remember, in the 1980s. It finally begin appearing because Japanese automakers tend to use stifled American innovations ten and twenty years later.