09 WRX Engine Failures - Page 15
+ Reply to Thread
Page 15 of 24 FirstFirst ... 56789101112131415161718192021222324 LastLast
Results 211 to 225 of 349

This is a discussion on 09 WRX Engine Failures within the Everyday Impreza Talk forums, part of the Community - Meet other Enthusiasts category; 15k and all is clear but I have a 10/08 build date so I am supposed to be in the ...

  1. #211
    Registered User etothen's Avatar
    Join Date
    Aug 2009
    Location
    Portland
    Posts
    31
    15k and all is clear but I have a 10/08 build date so I am supposed to be in the clear, cant say at the back of my mind I donít occasionally think about it, like is it the last one before the change or the first after? Then I pull away from the house, hit the canyon and start thinking about the next corner...

  2. Remove Advertisements
    ClubWRX.net
    Advertisements
     

  3. #212
    Registered User
    Join Date
    Nov 2009
    Posts
    14
    Engine failures are probably a lot less common they seem from the forum. This forum has how many members with 2009s? A few thousand? Even a one percent failure rate will seem like a lot of cars.

    I conduct a car reliability survey, and have found that engine failures are hardly ever common, even when they seem common on forums.

    25 owners of the 2009 have been participating in the survey. Of these, only one has had an engine failure (engine spun a bearing and threw a rod at 5,300 miles). One more than we usually see with a car this new, but still far from common.

    We'll be updating results next week to reflect this and another reported repair. Current results for the WRX always here:

    Subaru Impreza / WRX reliability comparisons

    The more owners participate, the better the information we can provide.

  4. #213
    Moderating on the run! Big Sky WRX's Avatar
    Join Date
    Jul 2001
    Location
    Plains of Eastern Montana
    Posts
    12,367
    I Support ClubWRX
    might want to read through the thread again

    "common" amongst folks w/ July and August build dates- very much so

    Sept and later- non-existent
    "Simplify and add lightness." - Colin Chapman

  5. #214
    Registered User
    Join Date
    Nov 2009
    Posts
    14
    Quote Originally Posted by Big Sky WRX View Post
    might want to read through the thread again

    "common" amongst folks w/ July and August build dates- very much so

    Sept and later- non-existent
    Good to know. Still, I wonder how common.

    When I first started the survey, I wanted to ask people for build dates because problems are often localized like this. But then I realized that too few people know their build date or would take the time to check it. So I asked for month purchased to get a rough approximation.

    At this point I don't have nearly enough 2009s signed up to split the group by build date anyway.

    But has a poll ever been conducted here, to get a rough idea of the percentages?

  6. #215
    Moderating on the run! Big Sky WRX's Avatar
    Join Date
    Jul 2001
    Location
    Plains of Eastern Montana
    Posts
    12,367
    I Support ClubWRX
    go over to nasioc- there is a 2009 post your vin thread- they quit counting at 100+ vehicles- all July and August build dates

    one recent one went with over 26,000 miles on it!
    "Simplify and add lightness." - Colin Chapman

  7. #216
    Registered User
    Join Date
    Nov 2009
    Posts
    14
    Well, that's certainly a lot of cars.

    If we assume that one in ten owners saw the thread, which might make sense for WRX owners since they tend to visit forums regardless, and that half of all Imprezas sold those months were the WRX...perhaps about a 25% failure rate. Which would be extremely high.

    Failure rates this high usually don't happen because any problem this likely should get caught during development.

    Hopefully Subaru will do the right thing and greatly increase the warranty on the engines of the affected cars. I assume that it's somehow not cost-effective to recall and fix them. Though if the failure rate is anywhere near 25%, you'd think it would be.

    Has anyone asked the dealer to install the updated parts or whatever it takes to prevent this?
    Last edited by mkaresh; 11-18-2009 at 01:49 PM.

  8. #217
    Registered User economatic's Avatar
    Join Date
    May 2009
    Location
    Portland, OR
    Posts
    843
    Quote Originally Posted by mkaresh View Post
    Well, that's certainly a lot of cars.

    If we assume that one in ten owners saw the thread, which might make sense for WRX owners since they tend to visit forums regardless, and that half of all Imprezas sold those months were the WRX...perhaps about a 25% failure rate. Which would be extremely high.

    Failure rates this high usually don't happen because any problem this likely should get caught during development.

    Hopefully Subaru will do the right thing and greatly increase the warranty on the engines of the affected cars. I assume that it's somehow not cost-effective to recall and fix them. Though if the failure rate is anywhere near 25%, you'd think it would be.

    Has anyone asked the dealer to install the updated parts or whatever it takes to prevent this?
    A couple weeks ago I alluded to the same issues with the failure rates of '09s with the magical build date over at nasioc. There are some data on how many have the build date and how many have failed. I actually did a quick analysis in SAS just to satisfy my curiosity since I own one. If I remember correctly the median mileage of failure was around 4,200 (mean was about 5,000....but that really isn't relevant given the distribution). I played around with fitting the data to distributions via MLE and found a lognormal to be about the best (just visually). With some back-of-the-envelope calculations I would say one should be in the clear after about 10,000 miles. That is, about 95% would fail before this value given the distribution and parameters I fit.

    The above certainly isn't a rigorous analysis and I don't trust the data. In fact, there are inherent biases built into the data. One can easily argue that those on a forum may drive their cars harder than the general population leading to a higher failure rate. It was just something I did for entertainment purposes so you can take it with a grain of salt. Probably would have shared it on nasioc too...but they flame you for everything over there, including making valid points.

  9. #218
    Registered User
    Join Date
    Nov 2009
    Posts
    14
    Interesting analysis. I wouldn't be so concerned about how the cars were driven. Many owners here are also going to drive them hard. And you want a sample of people who drive a car like you do.

    The number of cars with well over 10,000 miles might be a bit of a concern, but since they're already about a year old it won't be a concern for long if it still is currently. Adjusting for the current mileage distribution in the population would probably only stretch that tail out a bit.

    Reminds me of the discussions of IMS failures in the Boxster and the insufficiently fastened screw inside a batch of Passat 3.6s. The likelihood of failure in all of these cases definitely goes down as the miles accumulate.

    In all three cases the manufacturer has opted to let the engines fail rather than pre-emptively fix them.
    Last edited by mkaresh; 11-19-2009 at 12:46 AM.

  10. #219
    Admiral Ackbar the 1st mycologist's Avatar
    Join Date
    Oct 2005
    Location
    Late Devonian
    Posts
    13,320
    I Support ClubWRX
    Yeah, as I think you know you just can't do stats the way you guys are saying. Start with a survey that people with both failed and unfailed motors are equally likely to respond and use all demographics.

    The problem with fitting a curve is that the error on the upper end is much larger if not many people have had the car long enough. Extrapolating from that based on the variance in the distribution is way beyond tenuous. If the curve is just based on the number of failure reports and not the actual proportion of cars with that many miles having failures it is way worse.
    "From a little spark may burst a mighty flame." - Dante
    "The stitch is lost unless the thread is knotted." - Italian proverb

  11. #220
    Registered User economatic's Avatar
    Join Date
    May 2009
    Location
    Portland, OR
    Posts
    843
    Quote Originally Posted by mycologist View Post
    Yeah, as I think you know you just can't do stats the way you guys are saying. Start with a survey that people with both failed and unfailed motors are equally likely to respond and use all demographics.

    The problem with fitting a curve is that the error on the upper end is much larger if not many people have had the car long enough. Extrapolating from that based on the variance in the distribution is way beyond tenuous. If the curve is just based on the number of failure reports and not the actual proportion of cars with that many miles having failures it is way worse.
    Yep, I'm well aware of the multitude of problems of the data since I am a statistician by training. The data coming off a forum are inherently biased just by who is responding, e.g. most likely younger than the population of owners and those younger drivers probably push the car harder. Also, if owners of these cars know about the potential problem the may drive the car differently. The list of problems with the data go on and on (like most data sets).

    But you do make a very good point about how many cars haven't hit the higher mileage. Last time I checked mine is around 4,000 miles 15 months after the build date and I'll bet there are many others that are low mileage so the problem may be bigger than the data let us believe. However, I don't think it is as big of a deal most people are making of it. Subaru's willing to fix all of these motors with no questions asked is a sign they are handling it correctly. Also, if you have one of these build dates it has 5 years for a problem to develop.

  12. #221
    Registered User
    Join Date
    Nov 2009
    Posts
    14
    A few months ago I starting inviting Porsche Boxster owners to participate in TrueDelta's Car Reliability Survey to get a better idea of how common engine failures are in those cars due to a known problem with the IMS.

    The way I handle the sampling issue is that people sign up and then continuously participate. Anything that happened in the months before they signed up isn't included in the analysis. This and some other aspects of the research design have kept whatever might have motivated people to participate from distorting the results.

    I'm not aware of any other reliability survey with a prospective longitudinal design, so the others have to (or at least should) worry a great deal about non-response bias.

    It's still not perfect, but the quality of the results have greatly exceeded my expectations. The major limitation hasn't been the nature of the sample but the size of it. Unless a problem is very, very common, it'll take a relatively large sample size to measure its frequency with any precision.

  13. #222
    Admiral Ackbar the 1st mycologist's Avatar
    Join Date
    Oct 2005
    Location
    Late Devonian
    Posts
    13,320
    I Support ClubWRX
    Quote Originally Posted by economatic View Post
    Yep, I'm well aware of the multitude of problems of the data since I am a statistician by training. The data coming off a forum are inherently biased just by who is responding, e.g. most likely younger than the population of owners and those younger drivers probably push the car harder. Also, if owners of these cars know about the potential problem the may drive the car differently. The list of problems with the data go on and on (like most data sets).

    But you do make a very good point about how many cars haven't hit the higher mileage. Last time I checked mine is around 4,000 miles 15 months after the build date and I'll bet there are many others that are low mileage so the problem may be bigger than the data let us believe. However, I don't think it is as big of a deal most people are making of it. Subaru's willing to fix all of these motors with no questions asked is a sign they are handling it correctly. Also, if you have one of these build dates it has 5 years for a problem to develop.
    Yeah, the data really doesn't have the potential to show whether this is a cluster of bad cars or if they are all bad. Given that the signal for actual probability of failure would IMO be washed out by what is actually being tested (i.e. the probability that someone reports positive data vs. negative, and the number of cars that have actually achieved higher miles). I could see this being somewhat isolated but there is still a chance that they are all ticking and hence that the probability of failure will increase with every additional mile.


    My gut take is that this issue exceeds anything I have watched on WRX boards (e.g. "glass" trannies which seem highly dependant on operators) or other subarus (e.g. head gaskets on legacy 2.5ls). One major difference is that in this case people were actually on this thread before they blew, then eventually had it happen to them. Possibly a remarkable coincidence given the low sample size, but I think not.
    "From a little spark may burst a mighty flame." - Dante
    "The stitch is lost unless the thread is knotted." - Italian proverb

  14. #223
    Registered User
    Join Date
    Nov 2009
    Posts
    14
    One thing to keep in mind is that engine failures are very rare these days. I process over 1,000 repair reports a month, and see hardly any engine failures in cars with less than 100,000, even 125,000 miles on them.

    Electrical stuff? This and that sensor? Tons of that. But mechanical failures--very, very few.

    So if over 100 people have reported an engine failure, and they're all within a two-month build window, this is clearly unusually bad. Even if the failure rate is somehow only 10%--and with 100+ reports on a forum out of a population of three to four thousand cars it's at least this high--it would be crazy high by current standards.

    Not that anyone here is arguing otherwise.
    Last edited by mkaresh; 11-19-2009 at 09:44 AM.

  15. #224
    Registered User economatic's Avatar
    Join Date
    May 2009
    Location
    Portland, OR
    Posts
    843
    Interesting: ODI - Office of Defects Investigation

    There have been five complaints about it for the 2009 model.

  16. #225
    Registered User
    Join Date
    Apr 2009
    Location
    Virginia
    Posts
    1
    There are more complaints than 5. Other complaints are under Impreza instead of Impreza WRX.

    I filed a complaint last week. On the options to select a model when creating the complaint there is not an option for Impreza WRX just Impreza and Impreza STI.

+ Reply to Thread

Quick Reply Quick Reply

Register Now

Please enter the name by which you would like to log-in and be known on this site.
If you do not want to register, fill this field only and the name will be used as user name for your post.

Please enter a password for your user account. Note that passwords are case-sensitive.

Please enter a valid email address for yourself. We strongly suggest that you stay away from using aol, yahoo, msn, and hotmail accounts. Sometimes the mail server blocks the emails from our server. As a result you will not receive any notifications including the confirmation email.

Log-in

Human Verification

In order to verify that you are a human and not a spam bot, please enter the answer into the following box below based on the instructions contained in the graphic.


Posting Permissions

  • You may post new threads
  • You may post replies
  • You may not post attachments
  • You may not edit your posts
  •