[Close Window]


Copied from The Roadfly E46 M3 Message Board without permission

Message Title: Still seems to follow current theory (more)...
Posted by: Allen Schiano on 2002-12-13 at 12:19:51
(posted from: Host: palisade.nac.uci.edu IP: 128.200.16.22)
Message:

I'll thry [sic] to enunciate the current theory we're mostly in consensus upon, especially how it relates to this fail:

1) Cars before 11/01-builds can fail. They seem to fail at all mileages. #115 looks like a couple other cases in Jason's data. The projected rate of failure at 50K miles seems to be about 1%. And without an unusual function of mileage to failure points to a failure mode that can occur at any mileage, i.e., it's not a birth defect.

2) Cars built 11/01 to 12/01 fail at a rate about 15 times more likely than previous 01 cars. They also fail with a peak mileage of 7000 miles. So if you have such a car and it has substantially more than 7000 miles your likelihood of failure drops back down to the 1% level. Projections to 50K based on current data are not reliable since this seems like a birth defect that can't be projected in a simple, linear manner from current car mileages.

This failure mode points to problem that has a distribution of likelihood of occurring. Not a pure design problem but more like not meeting tolerance. Some cars do meet the tolerance well, some don't and fail.

3) Cars built in 2002 seems to start falling into groups:

a) cars built in 1/02 to 2/02 have some chance of having the same problems as the 11/01-12/01 cars. I susepct some engines built in the earlier period were placed in newer built 'cars'. There is still a likelihood that a 1/02 and 2/02 car will have a 'bad' engine that fails at 7000 miles. But the underlying problem from earlier 2001 cars still seems to exist.

b) 3/02 to 11/02 cars still seem to fail. But with only two examples - a 3/02 with MANY problems - can you say lemon? - and the recent 6/02 fail - we can't say much yet. But since the major engine internals changes in the 11/02 ETK changes, I suspect that we'll see some difference in the failure mode within 6 months.

c) cars built in 11/02 and on should be different - if they figured out the underlying problem.


------------------------------------------------------------------------





[Close Window]