General Motors a victim of circumstances in failure of ignition switches?
April 22, 2014
News reports recently noted that failure of ignition switches in some General Motors cars caused catastrophic events. The House Committee on Energy and Commerce has been grilling GM CEO Mary Barra on its whys and wherefores.
But the committee, made up largely of lawyers and/or politicians who lack an understanding of the process by which parts and sub-assemblies get into automobiles, were less than satisfied with her testimony.
Initially, the Delphi Company, which supplied the errant switches, must have submitted a preproduction lot to GM for testing and evaluation against the GM requirements. This would have included function, fit, endurance and service condition tests. Acceptance was then contingent upon satisfactory results within this test quantity, as well as Delphi's ability to produce them.
It was the purpose of GM's quality control people that incoming production switches meet GM's detailed specifications. Delphi would have delivered them in batches of hundreds, if not thousands. These batches would be inspected to meet significant, specified requirements, not of each and every switch but by sampling a set quantity from that batch. Then the batch could be accepted only if the number of failures were less than a predetermined maximum.
Thus, even in an accepted batch, there were likely a few that did not meet specifications. If the number of failures were large enough, the type of failure would be evaluated by an engineering representative for its significance. Were the failure nonfunctional, it wouldn't be prudent to hold up the car's production line for, say, the switches being the wrong shade of blue.
Likewise, the batch samples would not be life tested as that would also delay acceptance. Therefore, some switches that have been installed in cars may have had insufficient life.
Additionally, GM relied on Delphi's quality control to monitor the materials and parts that comprise the switches, even with GM oversight present. This applied not only to in-house produced parts but also to those they outsourced, again being accepted by sampling methods. Similarly, other samplings could have been applied down the supply chain.
Those preproduction switches submitted to qualify Delphi were made largely with new fabrication tooling producing parts close to their nominal configuration. But no two switches are completely alike, each having component parts falling into their own areas of manufactured variances.
As more switch parts were produced, the tooling became worn, allowing them to fall near the limits of the acceptable dimensional variation (tolerance) range. But this range was determined from the switch engineer's experience and by economics. It may be that a range was unrealistically large, thus producing parts accepted but with a potential of causing switch failure. They may have been in switches failing in the batch samples or may not have been detected at all.
Even in an accepted batch, it would be known that some switches installed in cars would be faulty, but it would not be known in which cars. Whether that should be reported to the CEO may be a company policy, but certainly if all incidents of concern were reported, it could overwhelm GM's top management. After all, most problems should be resolved at the department head level.
Decisions to inform management would depend on the perceived impact on customer safety and potentially on public relations, i.e., sales. In this instance, because of the furor, that decision itself was initially inappropriate.
Of course, the decision to recall cars with a possibility of having a faulty switch lies with the top management. But relying on reports from subordinates being factual, considering the cost to retrofit thousands of cars when only a few might require it and relying on drivers responding to a failure wisely made it a tough call. In retrospect, from the impact it has had on the company's public image, it shouldn't have been.
It seems that General Motors may be more a victim of circumstances rather than having deliberately avoided responsibility, as most would assume. In either case, the company will doubtlessly review its procedures.
It is of great concern that some drivers have lost their lives due to a failing ignition switch. But some fatalities might have been prevented had these drivers been more astute in reacting to their cars' malfunctions.
As an afterthought, there are thousands of parts in the car that continue to perform reliably day after day. In that light, it is extraordinary that manufacturers have produced cars summoning such high levels of confidence that drivers venture far out on secluded byways with the full expectation of returning safely.
Ed Westervelt lives in Nevada City.