Thursday, 25 February 2010

Toyota Recall - Part III

Finally! A calm analysis of the facts at hand.

Interesting highlights:
  1. Ford has more complaints than Toyota about "unintended acceleration"
  2. The vehicle that was the subject of the emotional testimony before congress has had no further incidents after 27,000 more miles of usage.
  3. Subsequent to 2: Safety regulators, human-error experts and auto makers say driver error is the primary cause of sudden accelerations, and if there are no error codes in the electronics, there is no evidence to support an electronic failure.
(Disclaimer: I do not own a Toyota, have never owned a Toyota, and my next vehicle is unlikely to be a Toyota.)

I feel for the Toyota engineers tasked with finding the "problem". I would bet that in meetings more than one engineer has screamed in frustration "For fucks sakes! There is nothing wrong with the accelerator pedal or the electronics!"

My job involves troubleshooting problems, which is why I sympathize with the engineers troubleshooting this problem, and the first rule of trouble-shooting is, you have to able to replicate the problem to identify the cause and apply the solution. The evidence shows so far that this issue happens only once in a vehicle and has happened in 2,515 Toyota vehicles. 8.5 million potentially affected vehicles have been recalled worldwide. Even assuming only 1 million of those vehicles are American owned, that works out to an incidence of 0.25% of vehicles having this issue. Let's assume that this issue arises in the first 3000 miles driven. That works out to a "sudden acceleration" event happening once every 1,192,843 miles driven. Talk about your needles in a haystack. And I've deliberately kept the service numbers low.

Is there any chance that the vehicle is contributing to the events reported? A very, very slight chance, but yes the chance is always there. But based on the low incidence it is not an endemic failure of design. It could be a problem at point of assembly where human error on the line causes improper installation.

But, even the incidence is too low for that. Realistically? It would have to be a combination of factors. When car z is driven at speed y at temperature x at humidity w at elevation v with a steering wheel angle of u, the problem may occur.

Good luck finding that combination though. But hey, if you're worried about the unknowns, GM (according to the WSJ) has the lowest reported incidence of unintended acceleration.

No comments: