Unintended Acceleration - Find the Cause

plekto

As I pointed out (I think it was on page 2?), the contest was clearly rigged such that only a company with a test lab, a dyno, cameras, and so on could even claim the "prize". The fine print resulted in it being nearly impossible for normal people or even automotive mechanics to compete.

And, as I pointed out, the *real* problem is the drive-by-wire system itself and reliance on computers instead of a simple cable. As more and more auto makers jump on the video game car bandwagon, expect to see problems every few years.

wwest

But you overlook...

Absent DBW we woudln't have FWD vehicles with more than ~200HP...!!

DBW allows high HP FWD engines to be detuned during low speed acceleration for safe reasons, let alone total detrottling for TC/VSC without damaging the catalyst.

Let's go FORWARD...

wwest

"..contest was clearly rigged..such that only a company..."

Not by Edmunds, just by the circumstances.

plekto

No, look at the fine print. It had to be repeatable and essentially a full lab type setup typical of a major defense contractor or auto maker or similar to be accepted.

As for the other comment, my dad's Park Avenue has ~240HP, FWD, a throttle cable, and it drives actually pretty slow and old-man like. You really don't need an electronic throttle except for maybe professional racing.

frankok1

And I bet he always hits the correct pedals.
I repeat:
There is no reason why drivers are misapplying pedals in so many different models of Toyotas, but not in many other brands such as GM's that have had very very low numbers of cases per vehicle sold. I would bet there still is an electronics / computer error.

Brake pedal override will help overcome glitches.

I believe they were putting in brake override for free even on NOT recalled models. Ask your Toyota service rep.

wwest

The more important point is how the brake override is being implemented.

Just a firmware change (reflash/update) in the engine/transaxle controlling ECU, an ECU that is already suspect for being the base cause of SUA isn't very creative and may well prove to be absolutely WORTHLESS.

plekto

That's the thing, though. With a physical cable, it's impossible for the software or computers to ever have such a problem. Even if the car caught on fire or was under water or whatever else happened to it.

It's just plain smart to rely on low tech solutions whenever you can if it's safety-related. Sometimes a simple hammer is the best solution to your problem, as it were.

wwest

Sorry, my company builds computers that control, amongst a myriad of other things, nuclear reactors, so I just can't see your point.

plekto

Well, yeah... That's a multi-million dollar control system that's custom built to be fail-safe and redundant barring a direct missile strike or a major earthquake or something ripping the place clean in half.

What happens if there is a voltage spike and the computers in a car lock up? Or there's a short and something fries itself? Relying on software when mechanical devices work just as well is silly. And yes, that plant should have a *few* manual overrides. I can't imagine a coolant system without a few manual valves in strategic places.

In a dirve by wire car, there's no backup. And no redundant system, sensors, or a second main computer. Nada. Stuff that breaks simply stops working. And if that's your throttle, steering, or brakes, well, have fun with that.

houdini1

I agree with you on simple solutions. The U.S. spent millions developing a pen that our astronauts could write with while in space. The Russians just used a pencil !!

stever

Better check out the full story at Snopes. We used pencils in Mercury and Gemini but the problem with pencils is broken lead floating around the cabin. One little piece could get stuck under the throttle and send you to Mars.

And the million bucks was spent by Fisher, not the taxpayers. We paid $2.95 each for the pens that went on Apollo VII. The Russians switched to Fisher pens too.

You can buy your own for $50. Cheaper than a Mont Blanc.

houdini1

Snopes???

Sorry, I don't believe a word they say. Just a liberal guy and his liberal wife putting the liberal spin on everything they can.

CLAIM: Barry and Michelle spent $90,000. on room service at a hotel over the weekend.

FALSE: Actually it was only $87,000.

stever

Fine, go to About.com. :shades:

houdini1

About.com is much more fair and balanced.

stever

I avoid 'em as much as possible but at least they don't put frames around sites they link to anymore.

(btw, you know they're owned by the right wing New York Times, right? :shades: ).

k, back to SUA stuff:

"One of lessons of the Toyota sudden acceleration hysteria should have been that a little data is a dangerous thing. Unfortunately it doesn’t seem that anyone has learned that lesson. Public misconceptions about runaway Toyotas were due in no small part to the NHTSA’s complaint database. Countless media stories were based on a supposed analysis of the complaint database leaving the public with the impression that there was well-documented proof of defects in Toyota vehicles."

Do Complaint Databases Contribute to Safety? (AutoObserver)

stever

"U.S. auto-safety regulators proposed standardizing keyless ignitions to allow drivers to turn off cars faster and more easily in incidents of unintended acceleration following Toyota Motor Corp.’s record recalls.

“These are the kinds of things you never think to read up on when you’re in a new vehicle or a rental vehicle,” said Henry Jasny, vice president of Advocates for Highway and Auto Safety, which pushed for the standard. “It’s better that it’s standardized.”

Standard Panic Stop for Keyless-Ignition Cars Sought by U.S. (Bloomberg)

srs_49

How 'bout a big red button on the dash that says "push to stop" - A panic button?

stever

And when your knee or your 9 year old hits it, then what? Twenty car pile-up? :-)

wwest

The simple solution is to have the same "delay" requirements for starting or stopping the engine. PB must be held down for 3 seconds.

stever

3 seconds is a long time when you can't get your car to stop. You can reach for the key and turn it off in a second.

la4mead

I don't have one, but I thought I saw in print somewhere that some vehicles with keyless systems use a "press and hold to stop" already. Perhaps they just don't print it on the button (how much can you print on the button, and expect operators to read).
Great that these systems have a soft "off" switch, supposing there's even a real off switch, they will however not prevent a real danger and common cause of unintended acceleration, that of driver error.

carguyfrank

As I have said for the past two years: Just put the car into neutral. then apply the brakes, it will stop! duh!

wwest

We MUST assume that some rather serious failure has occurred in which case the engine/transaxle control computer may be "out to lunch" and thereby not responding to ANY driver control inputs, including a 3 second depression of the Start/Stop PB. That might be the case even of actually having an ignition key to switch OFF.

In the case of some of the new hybrids with E-braking (a spring is used to simulate brake pedal foot [back]pressure) even the brake pedal might prove to be inoperative.

stever

Better spring for the ejection seat option.

plekto

Considering that a typical PC power supply uses a similar "push for three seconds to power off" circuit in it, and often it will simply NOT respond and require you to manually hit the power switch on the back of the computer, the only proper solution would be to install a switch under the dashboard that kills the power to the entire computer. Of course, it should be a big 10+lb hard to move switch like you see on old vacuum cleaners and the like.

The other option, of course, is to just drive a manual. Clutch in, move to neutral - problem solved, even if the engine is still going crazy.

houdini1

Don't manuals have computers in there somewhere too?

Many here have said that in Toyota's case, the computer would not let you shift into neutral...so I guess the computer could also stop you from shifting into neutral with a manual.

The darn thing would also probably prevent you from pushing the kill switch.
This whole thing was a tempest in a teapot. Driver error explains everything.

plekto

A manual transmission has a clutch, though, which acts as a purely mechanical means of physically pulling the gears apart. And most manuals have a similar setup for neutral as well where it physically moves the gears out of contact with each other.

You'll note that not a single case of UA involved a manual transmission vehicle, though there were a few reports of engines racing out of control.

kernick

Agree that a manual, mechanical device is the best way to assure the driver has the ultimate control. Electronics can freeze-up for no apparent reason. The best example of electronic controls not working at all times, is waiting for you to type on it. How many devices and times have you found the best way to fix an electronic problem is to reboot or power-down?

My suggestion was a lever or handle under the dashboard, attached to a rod to a ball-valve in the fuel-line. Normally it is fully open. You need to kill the engine in an emergency you turn the handle 90-degrees. With no new fuel, the engine dies in a few seconds, no ifs ands, or buts.

Sensors and other electronics are convenient, nice, and usually work. When they don't, the average person is going to have a hard time making the right choices.

carguyfrank

Why not just shift into neutral then apply the brakes....

houdini1

Their claim is that the computer sometimes can prevent you from shifting into neutral with an automatic transmission.

Does anyone know of a documented situation where someone was prevented from shifting into neutral due to computer malfunction?

wwest

A documented sitution is not really required...

The "shifter" does not directly shift gears, it only changes the open/closure of a switch set/group that directs the ECU, computer, as to which gear (or not) you wish to be in.

Beyond that all you need to know is that computers often "crash" or get locked in executing a "deadly embrace.

No computer "monitoring", "polling" those switches, no shift into neutral.

wwest

"..My suggestion..."

No, the simple, simplest, solution is to have an independent computer, separate control channel, watching the EFI injectors' PWM and the brake light switch. If the EFI's PWM doesn't drop to idle level with the brake light switch "on" then the "separate channel/path" control computer will open (relay) the EFI circuit.

houdini1

In your personal experience, have you ever known of this to actually happen?

I can give you several instances which happened to me back in the '60's where various MANUAL transmissions got hung up and would not shift at all. Our family had a fairly new International pick-up where gears would hang up. We also had a Nash Rambler that this would happen to. My brother-in-law had one of those small Metropolitans, and the gears would frequently hang up.

All manuals.

kernick

Why not just shift into neutral then apply the brakes....
Older cars - you're right. As soon as you talk about modern cars, you're talking about systems that are either networked or rely on the same processor boards. In that case it is just like hitting different keys on your keyboard; they should do different things, but they rely on some key component that is down. Or think of an electrical grid; it is not necessary for many individual things to fail, it just requires 1 critical item to fail, and then the rest of the grid is down.

kernick

Don't like it. You must be an IT or EE guy. I'm a ChE process eng. and I will not let my backups, when possible, be another electronic system.

I have a machine here that sorts bottles with 1 subsystem using an optical sensor to detect the open end of the bottle. On the other end of the machine is a subsystem that sorts the up-and-down of a cap by using the weight differential - invoking gravity. The cap subsystem has never failed. The Bottle sorter fails due to dust, misalignment and such. Maybe I'm still influenced by how many times the Robinsons could only stop the Robot by pulling his battery-pack.

wwest

Then were I you I wouldn't go flying on one of these newer airplanes.

carguyfrank

I would love to hear from someone that actually had the problem of unintended acceleration who then shifted the vehicle into neutral and then stopped.. Anyone out there?

houdini1

Actually it did happen to me back around 1957. I was driving an almost new 1957 Mercury Monterey. My mother was a passenger. The gas pedal stuck, I stomped on it a couple of times, did not work, so I shifted the auto to neutral, turned the ignition off, and coasted to the side of the road. No problem and no big deal. I was 15 years old at the time.

It started right up afterwards and drove normally. Turned out it was a bad spring or something like that.

kernick

Sort of like the Air France crew that didn't realize their air speed sensors had failed, and proceeded to stall the plane over the Atlantic?

wwest

The latest news I have seen indicated the airplane never stalled, the crew simply continued to "fly" in a descending attitude, unknowingly, until they contacted the ocean surface.

box1car

When Rhonda Smith's Toyota was in runaway mode SUA, she used her blue
tooth cell phone to call her hub- The car slowed to 35mph. Question not asked
is "What NUMBERS did you dial?" The CRUISE CONTROL light was on, so it was
active. Her RF generator (cell phone) operates at about 14" wave length where any conductor of 7" would be fine tuned to receive that digital signal
Can we say Turn Signal Arm which connects into the Cruise Control. Operating
at 800 MHz, there's no way to get a card in edgewise from continuous signal
If she dialed 35, it wasn't God who intervened, it was her RF Cell Phone- duh

kernick

Here's what I found:

Two hours and 10 minutes into the flight, the computers controlling the flight switched off the autopilot after becoming confused by conflicting speed readings, caused by the icing up of pitot tubes monitoring the plane's velocity.

"There was an inconsistency between the speeds displayed on the left side and the integrated standby instrument system (ISIS). This lasted for less than one minute," the BEA document said.

And:

By the time the Dubois appeared, just over a minute later, and as the plane began its fatal descent, another stall warning had been issued.

With the plane now rocking and falling at 10,000ft a minute, the pilot acknowledged the terrifying speed of the descent, saying "we're going to arrive at level 100", meaning 10,000ft.

At that point, just over a minute before the recordings stopped, the control sticks were used simultaneously, indicating the battle to control the plane had reached a frantic pitch. The pilot handed control to an unnamed colleague, presumed to be Dubois.

By now the "angle of attack", a critical indication of airflow over the wings, was at more than 35 degrees – nearly triple the outer limits for safe flight.

The BEA said the plane remained stalled throughout its three and a half minute descent, with the last recorded measurement showing the plane plummeting at 10,912ft per minute.

http://www.guardian.co.uk/world/2011/may/27/air-france-crash-inquiry

Summary: sensor failure, computer software can not comprehend the error, and professionals looking at the situation can not figure out the cause, system solution or engage a manual operation that is successful.

Everyone on that plane knew they were going down and in serious trouble.

I'm not saying all technology is useless; but there is an infinite number of chaotic events which can cause it to fail. And of course there are the systemic design flaws that all products have - as they are designed by humans.

wwest

It is very difficult, VERY, to push the nose down when you can see that the airplane is already descending at ~10,000FPM.

srs_49

Baloney. Takes A lot more thaan just the rf freq being tuned to match the hypothetical length of the wire to the turn signal stalk to couple a cmd into the ECU.

andres3

Beyond that all you need to know is that computers often "crash" or get locked in executing a "deadly embrace.

I've had home computers lock up and "crash" more times than I can count. Not hardly at all on modern (last 7 years) computers, but that's another argument.

I've never, not once, had a computer in an automobile lockup in anyway that impaired any function in my control. It just doesn't happen.

vulpine

Car computers have been problematical from the beginning. I'll grant that they are better, but they still fail in unexpected ways simply because extreme weather conditions wreak havoc with electronics. Something as simple as a cold solder joint will crack and break, given enough hot and cold cycles. Even components that 'live' in the cabin of the car and not under the hood can fail. A car's environment is nothing like the one in which your home or office computer works.

Never had one lock up that impaired function or control? Never had a radio fail? Never had your computerized air conditioning system fail? Never had that rear-window defogger fail? Any one of those is a control system. Maybe not in the same sense as the ECM or BCM, but those, too, fail. I should know--I've had more than one fail in more than one brand of automobile.

andres3

Never had one lock up that impaired function or control? Never had a radio fail? Never had your computerized air conditioning system fail?

Never had a radio fail.

My AC has failed, but it wasn't due to the computer, it was due to the compressor blowing up. It still blew air, but no cold air.

Never had that rear-window defogger fail? Nope. Although the rear windshield washer fluid sprayer has failed, but it is probably either the pump for the rear washer fluid, or most likely, a hole/leak in the tubing to get water all the way to the back from the front that has failed.

srs_49

...finds tin whiskers in a Camry Accelerator Pedal Position Sensor that could cause erratic throttle response. They do not go as far as to say it is the cause of Toyota's UAEs.

EDN (Electronic Design News) Article

NASA Paper

stever

What a difference a year makes? Has NASA gone full circle? (no, that'd be an orbit).

Should have entered my guess in the million dollar contest.

wwest

"...It just doesn't happen.."

Not to you, anyway. I'll put your vote in with the "pile" of >10,000,000.