GMs 'switch from hell'
Discussion
Out of Spec switch, cultural issues at GM and one Engineer's failure to documents cost 13 lives - Link and different site
fourpointsixgt said:
I'm guessing the 13 lives lost were due to the car stalling whilst being driven.
Following a link or two to find out some more, it seems that the lock was too easy to turn, meaning it could easily be knocked off - by a knee, or possibly even just by jiggling about of a heavy keyring. Or in a collision. Trouble is, turning the key off also disabled the airbags...WTF?
I thought airbags had a power backup, in the form of big capacitors, in case the earliest part of a collision severed the battery cables? Unless I'm misunderstanding something, I don't think the lock's much more than a symptom.
TooMany2cvs said:
Trouble is, turning the key off also disabled the airbags...
WTF?
I thought airbags had a power backup, in the form of big capacitors, in case the earliest part of a collision severed the battery cables? Unless I'm misunderstanding something, I don't think the lock's much more than a symptom.
I suspect there is some logic when ignition is off there is no need to deploy air-bags and it might increase risk of injury to do so (as most likely when people getting in/out of car). So there could be a signal sent to WTF?
I thought airbags had a power backup, in the form of big capacitors, in case the earliest part of a collision severed the battery cables? Unless I'm misunderstanding something, I don't think the lock's much more than a symptom.
air-bag saying don't deploy when ignition switch is off - if this signal is missing it will deploy when needed using backup power if required.
JimmyTheHand said:
I suspect there is some logic when ignition is off there is no need to deploy air-bags and it might increase risk of injury to do so (as most likely when people getting in/out of car). So there could be a signal sent to
air-bag saying don't deploy when ignition switch is off - if this signal is missing it will deploy when needed using backup power if required.
Because having the key knocked in a collision is utterly unforeseeable... Perhaps such a system could be a composite? If the ignition was turned off when the parking brake was on, and the car was stationary before, and the door is open - THEN disable the airbag...air-bag saying don't deploy when ignition switch is off - if this signal is missing it will deploy when needed using backup power if required.
TooMany2cvs said:
Because having the key knocked in a collision is utterly unforeseeable... Perhaps such a system could be a composite? If the ignition was turned off when the parking brake was on, and the car was stationary before, and the door is open - THEN disable the airbag...
I am not saying that was the reason, but when designing systems/vehicles/etc. you have to balance lots of conflicting requirements and it was one of the potential conflicts I could think of. I suspect a lot of design effort went into safety during accidents and they would look at whether the key was likely to be knocked before air-bags fired, sadly sounds compromised by a poor company culture and judgement by a engineer.Cars are getting much "smarter" now, so I'd expect much more being taken into account about when best to disable such things and even down things like weight of passenger and whether seat belt on to determine how much to deploy.
They have come a long was since I first started driving - when even the radios didn't have computers in - though being someone who writes software for a living software bugs in the drive by systems is quite a scary thought.
TooMany2cvs said:
Does somebody want to tell this man about aeroplanes?
Funny enough I started my working life on avionics and related equipment.The difference is anything to do with aeroplanes (especially passenger planes) is considerably more regulated and tested. Any crash or near miss will be investigated by people specially trained to do so - under such gaze the makers aren't just going to do their own thing with programmers inexperienced in the field - unlike say Toyota.
Well designed software can be far more reliable than than hardware solutions - trouble is you can't see how well built software is as all you see is the User Interface/actions, where as you can inspect and pull the hardware apart to asses how well it is built.
Bottom line well designed, properly tested and independently audited software is generally low risk - but does this apply to car manufacturers
TooMany2cvs said:
Trouble is, turning the key off also disabled the airbags...
Safety-critical systems should never be disabled while the vehicle is moving - regardless of the state of the ignition switch or anything else.And if the vehicle speed signal is missing then it should be assumed that the vehicle is moving.
In the case of airbags it could further be argued that they should never be disabled even if the vehicle is stationary with the ignition off, as other vehicles could still collide with it.
If there's a risk of the airbags deploying from the actions of people getting in and out of the vehicle, then perhaps the trigger mechanism needs looking at, rather than simply disabling them.
Edited by Phatboy317 on Sunday 22 June 11:30
JimmyTheHand said:
The difference is anything to do with aeroplanes (especially passenger planes) is considerably more regulated and tested. Any crash or near miss will be investigated by people specially trained to do so - under such gaze the makers aren't just going to do their own thing with programmers inexperienced in the field - unlike say Toyota.
I suspect that article is highly biased. Remember the US automotive industry feels very threatened by the Japanese cars which are generally considered higher status (along with German cars) than the domestic ones. The US car-makers acted with unconcealed glee when Toyota recalled over a less serious problem than GM have now with a reported 16 deaths.Edited by UnderTheRadar on Tuesday 1st July 09:57
Gassing Station | Motoring News | Top of Page | What's New | My Stuff