I am still trying to figure out why the Air Bags are connected through the ignition switch in the first place. So basically what you are telling me is, if I am stopped on the side of the road eating my lunch or making a phone call, and I have the car turned off, that if I am hit my air bags will NOT work? That would seem to be a safety problem in the first place. Now I admit that may not be the most common way to be hit, but still what is the advantage of turning them off at all?
Bugs, I agree with your main point, but have to strongly disagree with one aspect and with all those who say "this happens all the time" regarding "recycling of part numbers." That is a VERY BAD practice and contrary to any attempt to maintain traceability! The only time a part number should NOT be changed is when there is absolutely no difference in form, fit, or function, AND the parts are both forward AND BACKWARD compatible. Just think about it. There can be hundreds, thousands, or more parts in the supply chain. If both the "good" and "faulty" parts have the same number, it is highly likely that even with a recall, the "replacement" part will also be defective. That is just WRONG. IMHO, anybody who doesn't understand the importance of this does NOT deserve the title "engineer." The other piece of this puzzleis the statement from several that the REPLACEMENT design needs to "have the h... tested out of it" before acceptance. In this case, while that is true, the real source of the problem is the ORIGINAL design was not properly qualified, and in addition there obviously was NO continuing effort to ensure compliance with all design requirements (a basic requirement for QC for many decades now). While there certainly is a portion of blame against the suppliuer, it was still GM's responsibility as the "systems integrator" to perform their part of the QC process. The real question is not why did it take GM a decade to address this problem, but why it's taken them 100+ years to take ownership of product quality!
" If the opinion of the engineer the change is minor (spring length, detent ball material/size, color, something) and the part is interchangeable in all aspects, it will typically keep the same part number."
If there is any change in form, fit or function, the part number must roll, period. In this case, it looks like the fit and function didn't change but the form did. I don't know if Delphi is ISO:9001 certified, but if the part number wasn't changed and tracked, it'a an ISO violation.
The key torque was an issue way back from the 60's through today. It was just not considered the root cause of accidents. I have had three ingnition switches fail down through the years due to over loading with too much weight (keys & etc). This may have been one of the drivers behind push button ingnition design.
I'm still waiting for them to issue the recall on brake lines for 1999+ Silverado / Sierra / Yukon / Suburbans. The brake lines were made of substandard material causing them to dramatically fail. I was lucky that mine failed in my driveway when I stomped the brake pedal to take it out of park, but for many, the lines failed in a panic stop situation, which is about the worst possible time for brake lines to fail! I installed new stainless brake lines myself after they failed, and it was one of the hardest jobs that I've ever done on a vehicle, and I've swapped many motors and transmissions....
Somehow GM has avoided the whole issue and it doesn't appear to be going anywhere:
The part number does matter, because it normally is the part number that is tracked. If a supplier changes a part without identifying the change, the "system" doesn't know that there is a difference. If proper configuration control is done, the part number should be "rolled"/changed, and use of the different part can be tracked.
Once basic qualification tests for a part have been accomplished, the end-product manufacturer should routinely sample and test incoming stock. BUT, if a supplier makes "minor" changes without changing the part identification or otherwise notifying the end-manufacturer, the routine acceptance testing (which normally is not as extensive as qualification testing) may not be sufficient to catch a problem.
Many engineers recognize the value of configuration control; unfortunately, the bean counters often don't, and blow it off. They then will do their best to try to blame the engineers for the resulting problems.
An ignition switch that is used across many models is likely thoroughly qualified when it is first used. When a 'minor' design tweak is suggested by the supplier, someone in the supply chain needs to sign-off on the change. If the opinion of the engineer the change is minor (spring length, detent ball material/size, color, something) and the part is interchangeable in all aspects, it will typically keep the same part number. Recognizing some changes are major and others are minor is one of the jobs of an engineering group and 99 and a lot of decimal places percent of the time they do an excellent job. A whole lot of things need to go wrong in a pretty unique fashion before a faulty ignition switch will cause a death; anything that stops electricity from going to the ignition system could have the same effect. I'm not surprised it took many years to connect the dots.
I am an engineer and I won't say what company this was but I believe it could occur anywhere. I would design a part, choose the supplier, Qualify the part, then see the part placed into production, and after a while My responsibility for that assembly would be passed on to the sustaining engineering group. Later, I would be contatced to find out that that part was failing (or not performing as intended) in the field. now that I am involved again, the investication begins. Long story short it involves a bean counter that changes suppliers, and performs questionable "qualification" of the part. Critical dimensions are ignored, materials are changed, all to save money. Now the assembly is failing in the field and it now becomes my problem. But nobody wants to spend the money to fix it because it is a "legacy" system by now. So who is to blame for the bad part? This may not be the case in this instance, but it is an example of how a large company can produce bad products even though the engineers and designers did a good job. It is also a way that an issue can linger on because the responsibility for the product is shifted, and the path to correcting the issue is not clear.
It doesn't matter who made it or what part number was used. It's up to the final assembly manufacturer to "test the hell out of it" before accepting the part, and certainly BEFORE shipping the final assembled product to the customer.
What the problem appears to have been, although no detailed description has been published, is that the detent that held the rotating portion of the switch driver in the run position was not tight enough. Now consider that with a key alone in the lock it would have been quite well balanced and probably have never had any problems. What caused the failure was peole adding a half pound of keys and other junk to the key ring. And because the hole in the key was not right on the center axis this produced a torque tending to rotate the assembly toward the off position. So what exists is a situation ot the unintended consequences of a user action being fatal on a number of occasions. The maker )GM), failed to consider the probale results of customer abuse and mis-use of the product. This is unfortunate, but if GM had added a warning tag stating that adding weight to the keychain would cause problems, they might have been off the hook about responsibility. Although curent jury tendancies seem to lean more in the direction that even if a supplier warns about the hazards of an action, it is still their fault if a stupid user ignores that warning and is injured as a result.
The problem with a recall is that to change the part in question would require getting past all of the anti-theft efforts added to prevent a potential thief from doing exactly the same thing. So the change of parts would not be a simple fix. You really ought to do a teardown and document just all of the effort required to gain access to change that part. It would be an interesting presentation.
In his keynote address at the RAPID 2015 conference last week, Made In Space CTO Jason Dunn gave an update on how far his company and co-development partner NASA have come in their quest to bring 3D printing to the space station -- and beyond.
On Memorial Day, Americans remember the sacrifices the US armed forces have made, and continue to make, in service to the country. All of us should also consider the developments in technological capabilities and equipment over the years that contribute to the success of our military operations.
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.