Hacker News Reader

Comment by tragomaskhalos | Hacker News Reader

I had an "unintended acceleration case" in my old Austin Morris 1300; the cable connecting the pedal to the throttle snapped, jamming it at a fixed (fairly high revs) level, requiring me to control the speed using the brake.

The solution was to pop open the bonnet and swap in a replacement cable, which probably cost a couple of quid.

This recollection combined with the Toyota story merely convinces me that automobile automation has got completely out of control.

View on HN

I had almost the same problem and just hit the brake and pulled the car out of the road. Then I turned the engine off and saw that the plastic end of the cable is blocking the wire. I just removed it. No danger at all. It could be deadly if you have weak breaks and can't remember which pedal is the clutch :D

I think this issue is more to do with complexity than anything else; my car has an ECU, it's based on an 8-bit CPU from the 70s, never had any problems with it.

So does mine, and it is fuel injected. A cable controls a plate that controls the amount of air that gets into the engine. The only way for that design to fail WFO is for the return spring that keeps the plate closed to snap. At that point one would press in the clutch and turn off the ignition.

This kind of user corrective action is not possible on modern cars which I consider a huge engineering flaw.

I'm not tremendously knowledgeable about automobiles, but in a car with an automatic transaxle, shouldn't dropping it in neutral and switching off the ignition do essentially likewise?

I ask because that's how I practice responding to uncommanded acceleration, which I've done on occasion since I first heard of the failure mode. I've done this a few times a year in each of several cars, and as far as I can tell it's had no ill effect, but if "as far as I can tell" isn't far enough then I'd like to know it.

I know of no specific danger in shifting an AT into neutral. You need to be EXTREMELY careful when switching off the ignition that you don't lock the steering column. It would be better to

* switch to neutral

* safely exit the road and stop

* turn off engine

I am glad you have practiced it.

My concern is over the code running tiptronic transmissions, they are computer controlled manual transmissions where you no longer have an effective physical connection.

There was a programme on TV here in the UK a while back which featured an incredibly cost-conscious (read: mean) guy who turned off the ignition of his holiday hire car "to save money" while going down a winding mountain road in Spain. The steering locked and the car went over the side. Luckily he and his family survived and they were able to slaughter him on national TV over the incident.

My experience has been that switching the ignition from "run" to "accessory" doesn't risk locking the column, but now that I think about it, dropping the transmission into neutral should suffice -- the only concern I'd have there, with a runaway throttle, would be that it'd wreck the engine to have it running flat out with no load.

The engine will only be run at max for 15-30 seconds until you pull off, it will be fine. Better that than a potentially fatal accident. Only hardware after all.

Switching off ignition is usually also done in software.

Many ECUs store data (user settings, logging and fault codes) to non volatile memory before shutting down so the key only sends a signal. It's not a hard switch. But it could still help if all other ECUs except the faulty does shut down correctly.

How do you figure? Mechanical failures like that are no better than ECU-based failures. Indeed, doesn't matter how old the car is, the software in the ECU isn't going to develop new bugs over time. But mechanical pieces fail over time, as your story demonstrates.

Actually, the ECU might very well do that if it gets in field updates :)

Mechanical pieces fail over time, but we know a whole lot more about inspecting for, finding and preventing mechanical faults than we know how to create defect free software.

Exactly; and the mechanical case is often "field fixable" by a person with a modicum of experience learned while tinkering with cars (a mate of mine in a few minutes in my case), whereas now a tech needs to connect a laptop somewhere on the car to even begin a diagnosis.

Marginally relevant sidepoint: I remember reading that the one of the advantages enjoyed by the Americans over the Japanese in WWII was the formers' significantly greater expertise in mechanical maintenance of front-line equipment gleaned from their culture of fixing up and maintaining automobiles at home.

Reply on news.ycombinator.com