点火开关谜团：一个流氓员工搞乱通用汽车公司的故事

    德吉尔吉奥依然否认自己做过任何错事。正如报告中所述，“当被问到点火开关在2009年及之后是否被替换过，德吉尔吉奥说没有。至今为止，在正式面谈和宣誓中，德吉尔吉奥都表示自己不记得授权更换过点火开关，也不记得与此同时他做出过不要改变部件编号的决定。他特意不改变部件编号的决定导致调查者多年来一直都无法弄清事情的真相。”

    自从1991年起，德吉尔吉奥就开始在通用汽车担任设计发布工程师，他的工作主要与汽车开关相关。在1999年10月至2001年3月之间，他主管汽车开关部门，并于2002年从事Cobalt平台的相关工作。

    报告称，作为工作内容之一，德吉尔吉奥最终同意了使用远低于他规定的扭矩和转换压力的最低限度的点火开关。不过，德吉尔吉奥引发开关问题的动机尚不明了。无论如何，这都是案件中的棘手的地方。

    2001年3月到2002年底，点火开关需要经过多次调整，但是从设计到实际做出成品之后，开关仍然不管用。2002年1月，开关供应商德尔福（Delphi）在进行检验测试时，发现所有的样本组都不符合标准。调查报告显示，德吉尔吉奥与德尔福公司就这个问题展开了交流，然后在2002年2月，他做出了抉择：不进行维修与更换，暂缓开关生产。德吉尔吉奥在一封从此变得臭名昭著的电子邮件中对德尔福公司说，他会坚持到底，签上自己的名字。落款：“（无比厌倦开关的）雷•德吉尔吉奥”。

    根据瓦卢卡斯的报告，德吉尔吉奥没有告诉任何人德尔福公司的开关不符合标准。在采访了数以百计的证人后，瓦卢卡斯的团队发现所有其他人都不知道这件事。报告称，开关在2002年获准生产，随后在2003、2004及2005年开始出现问题，但是德吉尔吉奥没有通知任何人开关有问题。这是他自己一个人的秘密。

    2004年，通用汽车的一些工程师屡次发现，司机在驾驶Cobalt汽车时，如果碰到控制引擎开关的车钥匙，汽车就会在行驶中熄火。瓦卢卡斯在报告中写道，一名工程师将情况反馈给了德吉尔吉奥，问他扭矩是否有标准，Cobalt是否达到了这一标准。不过德吉尔吉奥并未给予回复。报告称，“尽管德吉尔吉奥已经获知这个问题，但他并没有对开关做出改动，也没有人得到警报说开关达不到相关标准。”

    到了2005年5月，顾客开始就熄火问题发起投诉。投诉的电子邮件中出现了德吉尔吉奥的名字，但他表示自己没有看到过邮件。其他工程师询问他时，他也说没有接到过投诉。调查报告显示，有证据证实他收到过熄火事件的担保数据报告，但他表示自己不记得有过这类报告。

    调查报告显示，在通用汽车其他人都不知情的情况下，德吉尔吉奥开始私下与德尔福公司讨论如何在以后的车辆中增加扭矩，来修复点火开关问题。2006年4月26日，德吉尔吉奥同意在2007年的汽车中使用更强的弹簧和活塞。利用2001年便可以使用的一个组件，他在没有增加任何成本的情况下完成了这一改动，也没有任何纸面文件记录了这次部件替换行为。调查报告称，之后的调查者也对此一无所知，因为德吉尔吉奥没有跟他们说过这回事。德吉尔吉奥告诉调查者，他不记得自己替换了开关，或是没有更改部件编号。

    这个遗漏至关重要，因为它推迟了事故调查长达七年（！）之久。考虑到这个改变的重要性，部件编号应当升级，以便让Cobalt的事故调查者搞清楚为什么在之后的产品中事故减少了。报告尖锐地指出：“如果其他人知道了这件事，通用汽车就会更快地采取召回行动。”对于不更换部件编号的问题，德吉尔吉奥没有给出任何解释。

    他的阻碍就像水门事件一样。瓦卢卡斯的报告称，2009年，当实地性能评估（Field Performance Assessment）工程师在直接质问德吉尔吉奥是否对开关做出过改动时，他回应说没有任何与熄火问题相关的改动，结果导致调查者多年来误入歧途。有人问过德吉尔吉奥，研发一种新开关需要做什么。调查报告显示，他推荐了一种力矩更大的开关，还说需要18至24个月才能投入使用。而他没有透露的是，实际上点火开关已经得到重新设计并投入生产了。

    DeGiorgio continues to deny he did anything wrong. As the report states, “When asked in 2009 and in the years that followed whether the ignition switch had changed, DeGiorgio said that it had not. To this day, in formal interviews and under oath, DeGiorgio claims not to remember authorizing the change to the ignition switch, or his decision, made at the same time, not to change the switch’s part number. DeGiorgio’s deliberate decision to not change the part number prevented investigators for years from learning what had actually taken place.”

    DeGiorgio had worked at GM as a design release engineer since 1991, and his career focused on vehicle switches. He took over responsibility for vehicle switches between October 1999 and March 2001 and was working on the Cobalt platform in 2002.

    As part of his work, the report states, DeGiorgio ultimately approved an ignition switch that fell well below the lower limit of acceptable torque or turning pressure that he himself had specified. But it is not clear why DeGiorgio was said to be preoccupied with starting problems with the switch. In any event, it was a troublesome part.

    The switch required many fixes between March 2001 and late 2002, and it simply didn’t work when it went from design to production as an actual part. In January 2002 validation-testing by Delphi, the switch supplier, every sample set fell below specification. DeGiorgio discussed the problem with Delphi, according to the report, and in February 2002, he had a choice: do nothing to fix it or change the switch and delay production. In an email to Delphi that has since become notorious, DeGiorgio said he would stay the course, signing the note “Ray (tired of the switch from hell) DeGiorgio.”

    DeGiorgio didn’t tell anybody that the Delphi switch was substandard, according to the Valukas report. And after interviewing hundreds of witnesses, the Valukas team couldn’t find anybody else who knew. The switch was approved for production in 2002, and questions were subsequently raised in 2003, 2004, and 2005, but at no time, the report said, did DeGorgio inform anybody that the switch was out of spec. It was his secret.

    In an incident in 2004, some GM engineers repeatedly experienced a moving stall in a Cobalt when the driver slightly grazed the key, which turned the engine off. The Valukas report said an engineer forwarded the complaint to DeGiorgio, asking if there was a torque spec and if Cobalt was meeting it. But he never responded. “Despite being informed of the problem, DeGiorgio made no changes to the switch, and no one was alerted to the fact that it didn’t meet specification,” the report contends.

    By May 2005, customer complaints about stalling had begun coming in. DeGiorgio’s name appeared on an email chain regarding the complaints, but he claimed not to have seen it. When asked by other engineers, he also claimed there had been no complaints. According to the report, there is evidence that he received warranty data reports with incidents of stalls but claimed he did not recall these reports.

    Unbeknownst to anybody else at GM, the report said that DeGiorgio then began communicating with Delphi about how to fix the ignition switch by increasing the torque in future models. A stronger spring and plunger was approved on April 26, 2006 for use on the 2007 model. The change was made at no cost with a part that had been available in 2001, and no paperwork accompanied the change. Later investigators didn’t know about it because DeGiorgio didn’t tell them, the report said. DeGiorgio told investigators he did not recall changing the switch or failing to change the part number.

    The omission was critical because it delayed the accident investigation for seven (!) years. Given the significance of the change, the part number should have been updated so an investigator looking at Cobalt accidents would understand why they decreased in later models. The report trenchantly observes: “Had others known, the recall would have happened sooner.” DeGiorgio offered no explanation for the omission.

    His stonewalling was Watergate-caliber. When a safety engineer from Field Performance Assessment in 2009 directly asked DeGiorgio whether there had been a change in the switch, the Valukas report said he told him there had been none that would affect the shut-off problem–throwing the investigator off the track for years. DeGiorgio was asked what it would take to develop a new switch. He proposed one with more torque, according to the report, and said it would take 18 to 24 months to get it ready. He didn’t disclose that the switch had already been redesigned and had been put in production.