Chat & Social > The Bar - General Chat

$1 Million Prize For Anyone To Prove Cause of Toyota's Runaway Cars

<< < (2/4) > >>

topless matt:
I have a friend that works at toyota and the interesting thing is that the components that have failed on the cars are NOT actually made by toyota.
They are a product bought in that has malfunctioned, many other companies buy thiese parts in that have failed but toyota have taken it upon themselves to not try and shift the blame and try to help their customers.
All good grace to them for fixing a problem with a part that they dont produce, only fit!

dxmedia:
10 000 000 lines of computer code to read (that's about correct for a modern cars computer systems - yota's are closed source so you'd need to reverse engineer it to begin with) to find the glitch.

To put the problem into perspective, you press the window up button and your seat moves forward.  Go on, everything is wired correctly, go fix...  Oh only does it once every 100 000 000 times.

Range Rover Blues:
They are drive by wire throttle, so it's probably an electricery fault or an electro-mechanical problem like pressing the gas pedal slightly sideways because the car isn't actually big enough to sit with your legs straight out.

Or maybe the drivers just awoke form a coma and realised where they were, "OH MY GOD, I'M IN A TOYOTA AAAAARRGHHH".

dxmedia:
Far more compliacted than that,

Pot on the top of the throttle pedal sends a reading to the EUC.

The ECU measures that, the engine speed, the car speed, the gearbox output... and then tells the engine to increase it's revs by 'x' amount, whether there is a load on the engine, how quick the throttle was depressed to decide if a down shift is needed. If a down shift is needed, a gentle one, 1 gear, 2 gears, how much use of the torque convertor... If it's a prius, is the electric motor added to the acceleration, is the regenerative charging circuit disabled for extra power...


1 line of code wrong in that, such as if this = this then this +that does this in the case of that+ that -that < all logic statments.

It's not the cable getting stuck, the mat riding over the pedal, or a sticky butterfly valve. it's far far more complicated.


This is why aeronautical computer systems are soooooo expensive - they have been tested for all faults. Windows NT4 is still commonly used due to there being variations where the bugs have all been removed.  Windows Vista isn't an option ;)

Saffy:
Who says it's a coding error? why not random corruption of a system from stray RF noise from a passing transmitter or a random thermal effect degrading a signal path or other such random things.
Anyhow... they not asking to find cause of bug be it a piece of code or what ever, they are asking for the test environment where the symptoms of the bug can replicated on demand. You *could* hit upon that without looking at one scrap of decompiled code but it might cost more that the 'prize' money to set up for it if seriously going to have a stab at it... but ya be famous.

Navigation

[0] Message Index

[#] Next page

[*] Previous page