Welcome to Bimmerfest -- The #1 Online Community for BMW related information! Please enjoy the discussion forums below and share your experiences with the 200,000 current, new and past BMW owners. The forums are broken out by car model and into other special interest sections such as BMW European Delivery and a special forum to voice your questions to the many BMW dealers on the site to assist our members!

Please follow the links below to help get you started!

Go Back   Bimmerfest - BMW Forums > BMW Model Discussions > 5 Series > E39 (1997 - 2003)

E39 (1997 - 2003)
The BMW 5-Series (E39 chassis) was introduced in the United States as a 1997 model year car and lasted until the 2004 when the E60 chassis was released. The United States saw several variations including the 525i, 528i, 530i and 540i. -- View the E39 Wiki

Reply
 
Thread Tools Display Modes
  #1  
Old 11-08-2005, 02:26 PM
rtw_travel rtw_travel is offline
Officially Welcomed to the 'Fest
Location: Vancouver
 
Join Date: Feb 2005
Posts: 427
Mein Auto: '98 540i/6
V-8 Rad failure data needed

I am doing a Weibull statistical analysis of the e39 540 rad and expansion tank failures... and I need your data. Believe it or not this is for fun, and not a PhD thesis.

Please pm me (do not post here) with data similar to the following which is for my car:

Model: 1998 540
Build date: Sep 1997
delivery date: 10 Dec 1997

Expansion tank failure 1 Aug 2005 113,495 km
Rad replaced before failure 1 Aug 2005 113,495 km


Notes:

Miles or km acceptable. Please identify which you are using.

Please identify whether the component failed or whether you pre-emptively replaced it before failure (as I did with my rad).

If you have gone through more than one rad, please let me know which brand you installed after OEM.


I ~know~ that all of you 540 owners have this data somewhere. Please dig it out and send it to me via pm. Any data you send will be combined with other's data so that it is entirely anonymous.

I'll post the results when I figure it all out.

Thanks in advance!!
Reply With Quote
Advertisement
  #2  
Old 11-08-2005, 02:30 PM
texbid's Avatar
texbid texbid is offline
Officially Welcomed to the 'Fest
Location: Scotland
 
Join Date: Aug 2005
Posts: 174
Mein Auto: 2002 540i6, 2002 525d
are you using a two or three parameter weibull fit
Reply With Quote
  #3  
Old 11-08-2005, 02:38 PM
rtw_travel rtw_travel is offline
Officially Welcomed to the 'Fest
Location: Vancouver
 
Join Date: Feb 2005
Posts: 427
Mein Auto: '98 540i/6
Quote:
Originally Posted by texbid
are you using a two or three parameter weibull fit
Two should do it for standard failure analysis. I'll know for sure when I see the data.

By the way, any 540's who have never had a rad or expansion tank failure, please send me that as a pm too (with date of delivery etc).

Last edited by rtw_travel; 11-08-2005 at 02:41 PM.
Reply With Quote
  #4  
Old 11-08-2005, 03:04 PM
BillP BillP is offline
Officially Welcomed to the 'Fest
Location: Northern California
 
Join Date: Jan 2002
Posts: 1,564
Mein Auto: Cars
Someone already did this. I seem to remember hundreds of entries in the table.

I'll see what I can dig up ...

OK, I think Dave Kowached from roadfly.com did the survey, but the link I found is now dead.


Cheers,
Bill
__________________
'00 540iA/Sport, Anthracite/Black - Euro Delivery in Spring 2000
'06 Honda S2000, Black/Black (my go-kart) - GONE
'03 Z4 3.0 6-speed/Sport, Sterling Gray/Black - GONE
'71 2002 Green/Black - GONE

Last edited by BillP; 11-08-2005 at 03:08 PM.
Reply With Quote
  #5  
Old 11-08-2005, 03:54 PM
NFS NFS is offline
Officially Welcomed to the 'Fest
Location: Cali
 
Join Date: May 2005
Posts: 629
Mein Auto: E39 540I
Quote:
Originally Posted by BillP
Someone already did this. I seem to remember hundreds of entries in the table.

I'll see what I can dig up ...

OK, I think Dave Kowached from roadfly.com did the survey, but the link I found is now dead.


Cheers,
Bill
I think you're referring to 540i6.com survey but the link is down.
__________________
Reply With Quote
  #6  
Old 11-08-2005, 09:43 PM
rtw_travel rtw_travel is offline
Officially Welcomed to the 'Fest
Location: Vancouver
 
Join Date: Feb 2005
Posts: 427
Mein Auto: '98 540i/6
If someone can find the table, then that would be terrific. Otherwise please send the data. 5 people have come forward so far - I'm hoping for at least 20.

Weibull is the next step after a survey - I'll be able to predict failure probabilities, and calculate an optimum replacement time/ mileage.
Reply With Quote
  #7  
Old 11-09-2005, 10:10 AM
Zigzag Zigzag is offline
Registered User
Location: Phila. Pa.
 
Join Date: Aug 2005
Posts: 5
Mein Auto: 97 540i
My radiator started to leak (hairline crack) just above the upper hose inlet. Car is a 97 540 with 62k.
Reply With Quote
  #8  
Old 11-12-2005, 07:37 AM
rtw_travel rtw_travel is offline
Officially Welcomed to the 'Fest
Location: Vancouver
 
Join Date: Feb 2005
Posts: 427
Mein Auto: '98 540i/6
Quote:
Originally Posted by Zigzag
My radiator started to leak (hairline crack) just above the upper hose inlet. Car is a 97 540 with 62k.

Thanks Zigzag - i need a bit more than that. Please see post #1 for details.

I also need about 10 more 540 drivers to send me the info before I get something I can work with.

C'mon guys - its the weekend. You have time to do this now!


Thanks in advance
Reply With Quote
  #9  
Old 11-12-2005, 08:40 AM
Steve D Steve D is offline
Officially Welcomed to the 'Fest
Location: New Jersey
 
Join Date: Oct 2002
Posts: 243
Mein Auto: 2000 540i/6sp
I just posted my radiator failure data on the Yahoo E39 board. It will be interesting to see if the data shows an exponential distribution (constant failure rate), or an increasing rate which may indicate a premature end of life failure rate.

Steve D
__________________
Reply With Quote
  #10  
Old 11-12-2005, 09:12 AM
1Dreamer's Avatar
1Dreamer 1Dreamer is offline
Super Moderator
Location: A block off Sunset Blvd.
 
Join Date: Oct 2004
Posts: 7,419
Mein Auto: Fiver
Since your survey is lmited to radiator and expansion tank failures, I'm afraid I can't help. I replaced both as preventive maintenance when the upper hose assembly blew - which is another known weakness in the cooling system. See this thread where others have had the same problem.

The new assembly looked different and when I asked about it, I was told BMW redesigned it after the weakness in the old design became known, but it seems pretty common that once one component of the cooling system fails, the others soon follow and after hearing too many stories of engines needing to be replaced before people were able to get their cars safely off the road, I decided to just replace everything rather than take that chance.

Although you don't hear about it as much, apparently the hose assembly was a big enough problem for BMW to redesign it - which makes me wonder if the previous design was an even bigger problem than the radiator. If you expand your survey, let me know and I'll send you the info.

EDIT: By the way, the guy in the link said the cost for the piece of hose was $38, but to have the hose assembly replaced with the newer design was between $150 to $200 at an indy shop. It would be interesting if the newer design somehow caused less stress on the radiator making them last longer. I kind of doubt it, but I'm hoping.

Last edited by 1Dreamer; 11-12-2005 at 09:29 AM.
Reply With Quote
  #11  
Old 11-12-2005, 01:01 PM
2001 540 AZ 2001 540 AZ is offline
Officially Welcomed to the 'Fest
Location: Gilbert, AZ
 
Join Date: Oct 2005
Posts: 141
Mein Auto: 2001 540i
Quote:
Originally Posted by rtw_travel
I am doing a Weibull statistical analysis of the e39 540 rad and expansion tank failures... and I need your data. Believe it or not this is for fun, and not a PhD thesis.

Please pm me (do not post here) with data similar to the following which is for my car:

Model: 1998 540
Build date: Sep 1997
delivery date: 10 Dec 1997

Expansion tank failure 1 Aug 2005 113,495 km
Rad replaced before failure 1 Aug 2005 113,495 km


Notes:

Miles or km acceptable. Please identify which you are using.

Please identify whether the component failed or whether you pre-emptively replaced it before failure (as I did with my rad).

If you have gone through more than one rad, please let me know which brand you installed after OEM.


I ~know~ that all of you 540 owners have this data somewhere. Please dig it out and send it to me via pm. Any data you send will be combined with other's data so that it is entirely anonymous.

I'll post the results when I figure it all out.

Thanks in advance!!
RTW travel How's the data input coming?
Reply With Quote
  #12  
Old 11-24-2005, 05:20 PM
rtw_travel rtw_travel is offline
Officially Welcomed to the 'Fest
Location: Vancouver
 
Join Date: Feb 2005
Posts: 427
Mein Auto: '98 540i/6
Arrow Rad failure results

There was data from 12 cars: 7 failures and 5 ‘suspensions’. A suspension is where either the rad was replaced before failure, or the original rad is still running with no problems.

I did the analysis two ways:
  • rad failure vs. mileage at time of failure; and
  • rad failure vs. years of service at time of failure

The results are:

For Mileage
Beta 3.09
Eta 74,686
R^2 89%

For years of service
Beta 4.17
Eta 7.25
R^2 95%



A couple of comments:
1) According to the R^2 values, “years of service” is a better predictor of failures than mileage. In fact, my guess is that the real engineering reason for failure is either drive cycles (ie. Start up from cold and bring to operating temperature then cool back down), or engine hours. We can’t measure these, so we’re stuck with mileage or years of service as a proxy.

2) the high Beta values indicate that this component is in ‘wear out’ phase. This is no surprise to V8 owners. However, every well designed component will eventually wear out. The real problem with our e39 is the low Eta value. Eta is the value at which 63% of the components have failed, and 7.25 years is quite short for a rad life. I would have hoped for 12 or more.

The Weibull curves are shown below. The left hand ‘y’ axis is cumulative distribution function. This is the percentage of units that have failed. It is a log scale running from 1% to 100%. The bottom ‘x’ axis is either in miles or years since in-service depending on the graph.

Just to give you a frame of reference, if you look at the mileage graph, you can see that 10% of the rads will fail by about 36,000 miles.

So what can you do with a Weibull curve?
1) You can use it to calculate the optimum replacement time. In my case, it costs me about $200 to do a DIY replacement of my rad – assuming I have the parts and time and don’t charge for my labour. If I let it fail, then I have the cost to tow the car plus typically I would have a dealer do the repairs because I would be too busy to do breakdown maintenance. So the cost of letting the rad fail might be $600.

If I decide to replace the rad too soon, then I am spending money needlessly. If I replace it too late, then I run too high a risk of an in-service failure. There is an optimum time for replacement – doing the calculations, this turns out to be around 5 years in my case assuming the $200/ $600 cost differential. i.e. after 5 years, I should replace my rad if it has not already failed. But you always have to tie in statistics with engineering – if you believe that you can tell when the rad is about to blow by inspecting for small amounts of dried coolant, then you may decide to inspect frequently to try and catch the leak before it blows.

2) If we had Weibull graphs on all the parts, then we could make more informed maintenance decisions. For example, suppose your rad was being replaced and the service advisor suggested you also replace the water pump. Simply look on the water pump Weibull curve and calculate whether you should based on the probability of failure of the water pump.

3) Weibulls are specific to a single part. If we had enough data, we could compare weibulls from all the replacement rads and figure out which is the best.

4) With weibulls of all components, you can estimate maintenance costs and decide if it really is worth buying the extended maintenance… or when it is time to sell the car.

5) You could really make a much more informed decision of a used car purchase – you would know exactly the failure modes of the car. It would make inspection of the car much easier, and you would certainly understand likely maintenance required before purchase.

6) If you were a supplier of rads, you could use the weibull to predict how many rad failures there would be every month given the current population and age of e39 V8's in the US or Canada.

Questions or comments?
Attached Thumbnails
Click image for larger version

Name:	radmile.png
Views:	123
Size:	4.5 KB
ID:	62179   Click image for larger version

Name:	radyear.png
Views:	92
Size:	4.5 KB
ID:	62180  

Last edited by rtw_travel; 11-24-2005 at 11:54 PM.
Reply With Quote
  #13  
Old 11-24-2005, 05:26 PM
bmw325 bmw325 is online now
Officially Welcomed to the 'Fest
Location: New York, NY
 
Join Date: Dec 2001
Posts: 8,446
Mein Auto: 528i
This is very cool. We need to do something like this for e46 window regulators- although radiator failure is much more serious.
__________________
Reply With Quote
Reply

Bookmarks


Forum Navigation
Go Back   Bimmerfest - BMW Forums > BMW Model Discussions > 5 Series > E39 (1997 - 2003)
Today's Posts Search
Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is On



Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
BMW ConnectedDrive: ChrisTO General BMW 12 11-08-2005 06:41 AM
NE Hacks for the NAV Data CD-55 3 Series / 4 Series 5 12-31-2002 07:22 PM
computer / data buses used in new bimmers? mwette General BMW 4 09-30-2002 08:45 PM
Non-magnetic data archiving (50gb+) question alee Off-Topic 10 02-28-2002 07:36 AM


All times are GMT -7. The time now is 12:15 PM.


Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2014, vBulletin Solutions, Inc.
© 2001-2011 performanceIX, Inc. All Rights Reserved .: guidelines .:. privacy .:. terms