Originally Posted by werepossum
If I remember correctly, the Audi problem wasn't a glitch in the electronics so much as in the design. I believe the fix involved replacing a vacuum valve and rewriting the code to range the data, because if that vacuum line stopped up, the computer interpreted the loss of vacuum as an unsatisfied demand for acceleration.
This is exactly what i meant. I will share some history with you.
The history of the widely used 4-20mA current communication system. Simply known as the current loop.
Long ago it seems that Honeywell used pressurized air to control valves.
A pneumatic system to use with thermostats to control heaters to control the temperature in a building. This system always needed a default amount of air pressure. Meaning no signal would mean air pressure. Thus when giving the signal close valve, there would still be pressure but of a different value. When giving the signal open there would be a different amount of air pressure again. But the idea is that the air tubes relaying the thermostat signal are always filled with pressurized air. There is never a zero value. This was probably done for physics and mechanical reasons.
When the electronic revolution started, Honeywell wanted a system that mimicked this exact behavior and was insensitive to electric pulses. Thus the 4 to 20 ma current loop was invented. The funny side effect from this system is that because it is never 0 mA. you can also detect cut wires with it because then the current loop will not function. And you can detect short circuits as well within certain boundaries of detection. This is used for i think about 40 years already. This is all easy to detect. Thus we have a simple system, that is very reliable. It is even used in automated industry processes. This very simple form of reasoning is gone today it seems. The pressure of managers making flash profits and people with dollar signs in their eyes destroys common sense.
The moral of the story is that one never must make use of a system where one cannot detect failure in life threatening situations. Redundancy is the key. And this is all too often seen as a cost factor. Thus flawed mathematical models are used to statistically predict failure and decisions are made to reduce costs. In all honesty, most of the time the mathematical prediction of failures is not even used. A big flaw to not use mathematical prediction wisely is...
Luckily for my uncle and my cousin, their malfunctions were not in braking situations, but were during either cruise control operation or when starting from a dead stop at a red light or stop sign, especially if you shifted into park or neutral while waiting. So every time they had open road ahead. Also, the Audi 5000 had excellent brakes which, if applied quickly, would almost stop the car, but at speed the engine was more powerful than the brakes - something not true under normal acceleration. It's also noteworthy that although the 5000 was not capable of leaving black marks during normal acceleration, it was capable of leaving nice black streaks during abnormal acceleration if from a stop.
I am happy to read it was in relatively "safe" conditions.
There was so much FUD being distributed though that it's hard to be truly sure what was going on. Sixty Minutes for example had an automotive expert figure out what COULD cause the problem, modified the transmission to cause that condition, then passed that off as if it were happening on its own. There's been a fair amount of that going on with Toyota too, as people try to turn their own errors or misfortune into an income opportunity.
Yes, i understand very well :
As example : I had yesterday a discussion with a colleague about the right of the public to know about failure of the government(or at least local semi independent branches of that government creating the failures). It was in this case about the firework disaster that happened in the Netherlands in Enschede. It is obvious for everyone who knows the background that errors have been made in the police investigations and that certain "high placed" people have been protected. However according to the news, the council of the state(Netherlands) has advised that the final investigation report must not be made public to protect the public. This i find disturbing. Because for me this means that serious errors has been made and that corrective actions must be taken to prevent misconduct to happen again. I argued that no matter what , the public needs to know. My colleague argued that it will only will cause mass hysteria. I understand his point very well and partially i agree. However, i feel that the report must be made public but on television and radio and internet broadcast while a group of honest people explain why and what has happened. Thus preventing mass hysteria by explaining the proper causes. Common sense seems to prevail here still, i would argue then that it will not create a mass hysteria.
