IMS bearing alternate fix
#31
The failure rate of these cars in the dual row bearing was 1% and the failure rate of the single row AT THE TIME OF THE CLASS ACTION was 8-10% (it was to this larger cadre of failures I was referring). Since that time, nobody is recording how many additional IMS bearings have failed. Nor are they recording how many REPLACEMENT bearings or engines have failed (I'm not talking about aftermarket replacement bearings...I'm talking single row or large captive IMS engine replacements from Porsche). Further, Porsche would not reimburse anyone with over 100k miles or 10 yrs on it at the time of damage...meaning their own confidence in the survival of bearings past that time was also low. You really have no factual basis to assume 8% failure rate at 85k miles assumption 6 years or so after the data was presented in the litigation. You talk about miles driven on a fleet of cars and if the driving habits of all of those drivers was the same, maybe there'd be more validity...but
Last edited by MadIrish; 02-28-2018 at 10:56 PM.
#32
Shameful Thread Killer
Rennlist Member
Rennlist Member
https://rennlist.com/forums/996-foru...30k-miles.html
Judging by this I don't think the $$ bearing kits are any better. A sealed bearing in this application is stupid.
Judging by this I don't think the $$ bearing kits are any better. A sealed bearing in this application is stupid.
Now, having said that, there are some pretty sharp folk who've engineered a good solution. Sadly, the price they charge for the fix-up is pretty high. I"m an amateur in this specific failure mode, but if you can do something less expensive, the proof will be about +200K miles of satisfactory run-in. Some issues you're going to have to confront: Previous less costly fixes have tanked seriously. For various reasons. The exact failure mode(s) are not well defined. I've seen a bunch of theories, and some actual measurements. So far I know we've had serious NRRO with some shafts. There has also been evidence of normalized run-out. There have been issues with foreign contamination of the bearing. With acid etching of the ball or race. Rust/oxidation. Axial skidding, chain woggle, and even cases of unknown seizure of the outer race.
I'm afraid that the true solution would address all these potential failure modes. Which would be to disassemble and mic the shaft, balance, and then do all the needed bearing load and lube design. It seems like there is not going to be a 'quick/cheap fix'. But - I will encourage you to work up anything. More eyes on the issue is always good. Unless you know starting with a good shaft, true gear teeth, correct chain, no contamination, just solving the bearing problem may lead to unpredictable results. A few things we do know. The problem seems to be more common on under-utilized low mileage cars. Single row bearings are more failure prone(I don't know the percent). Note also the intermediate shaft is a non-flow through unit. Having an open cage bearing with full flow of oil onto the cage is a solution, but the shaft needs to be drained as well. A certain amount of drainage back through the ball cage will happen, but I don't know how much outflow is enough for the spray into the bearing.
Good luck, keep us posted.
#33
Track Day
Thread Starter
Bob
Check out the EPS oil feed, might save you some time, it's a replacement oil pump drive shaft with a groove machined down its length , this lets a small amount of oil inside the ims shaft which travels through to lube the bearing at the other end , check out Europeanpartssolution.com
I fitted the full ims kit, roller bearing and oil feed , this was before the ims solution plain bearing by LN was on sale, I would probably go for that now but I've done about 5k on the roller over about two years and no probs up to now .
Good luck
Check out the EPS oil feed, might save you some time, it's a replacement oil pump drive shaft with a groove machined down its length , this lets a small amount of oil inside the ims shaft which travels through to lube the bearing at the other end , check out Europeanpartssolution.com
I fitted the full ims kit, roller bearing and oil feed , this was before the ims solution plain bearing by LN was on sale, I would probably go for that now but I've done about 5k on the roller over about two years and no probs up to now .
Good luck
Glad to hear you're an A&P. We need some of that here.
Now, having said that, there are some pretty sharp folk who've engineered a good solution. Sadly, the price they charge for the fix-up is pretty high. I"m an amateur in this specific failure mode, but if you can do something less expensive, the proof will be about +200K miles of satisfactory run-in. Some issues you're going to have to confront: Previous less costly fixes have tanked seriously. For various reasons. The exact failure mode(s) are not well defined. I've seen a bunch of theories, and some actual measurements. So far I know we've had serious NRRO with some shafts. There has also been evidence of normalized run-out. There have been issues with foreign contamination of the bearing. With acid etching of the ball or race. Rust/oxidation. Axial skidding, chain woggle, and even cases of unknown seizure of the outer race.
I'm afraid that the true solution would address all these potential failure modes. Which would be to disassemble and mic the shaft, balance, and then do all the needed bearing load and lube design. It seems like there is not going to be a 'quick/cheap fix'. But - I will encourage you to work up anything. More eyes on the issue is always good. Unless you know starting with a good shaft, true gear teeth, correct chain, no contamination, just solving the bearing problem may lead to unpredictable results. A few things we do know. The problem seems to be more common on under-utilized low mileage cars. Single row bearings are more failure prone(I don't know the percent). Note also the intermediate shaft is a non-flow through unit. Having an open cage bearing with full flow of oil onto the cage is a solution, but the shaft needs to be drained as well. A certain amount of drainage back through the ball cage will happen, but I don't know how much outflow is enough for the spray into the bearing.
Good luck, keep us posted.
Now, having said that, there are some pretty sharp folk who've engineered a good solution. Sadly, the price they charge for the fix-up is pretty high. I"m an amateur in this specific failure mode, but if you can do something less expensive, the proof will be about +200K miles of satisfactory run-in. Some issues you're going to have to confront: Previous less costly fixes have tanked seriously. For various reasons. The exact failure mode(s) are not well defined. I've seen a bunch of theories, and some actual measurements. So far I know we've had serious NRRO with some shafts. There has also been evidence of normalized run-out. There have been issues with foreign contamination of the bearing. With acid etching of the ball or race. Rust/oxidation. Axial skidding, chain woggle, and even cases of unknown seizure of the outer race.
I'm afraid that the true solution would address all these potential failure modes. Which would be to disassemble and mic the shaft, balance, and then do all the needed bearing load and lube design. It seems like there is not going to be a 'quick/cheap fix'. But - I will encourage you to work up anything. More eyes on the issue is always good. Unless you know starting with a good shaft, true gear teeth, correct chain, no contamination, just solving the bearing problem may lead to unpredictable results. A few things we do know. The problem seems to be more common on under-utilized low mileage cars. Single row bearings are more failure prone(I don't know the percent). Note also the intermediate shaft is a non-flow through unit. Having an open cage bearing with full flow of oil onto the cage is a solution, but the shaft needs to be drained as well. A certain amount of drainage back through the ball cage will happen, but I don't know how much outflow is enough for the spray into the bearing.
Good luck, keep us posted.
Sure, just don't know the extent of damage yet, if any. Will get the car up on the lift this weekend and start poking around. Dropping the motor and transmission is something I was planning on doing anyway to do the IMS bearing fix. I know I don't have to do that but I want to do some other maintenance as well and its just easier if everything's out in the open.
#35
Shameful Thread Killer
Rennlist Member
Rennlist Member
Thank you. I have an IA (inspection authorization) from the FAA so I have a bit of a keen eye for doing modifications in the airplane world. I've had to deal with analyzing ball bearing failures in some Lycoming engines that have a ball bearing support for magnetos so I'm aware of how they fail in bathed applications. In that case the bearings are fed with splash oil and I believe underutilization is a contributing problem. Water and acid gasses are byproducts of combustion which end up in the oil. If the engines sit for long periods of time then these contaminates cause corrosion on steel parts which is obviously hard on contact bearings.
.
Now, I will deny ever writing this(haha) but I have heard - through various grapevines that if the car has a full-flow oil filter, and the failure is caught soon enough through short oil exchange cycles, and careful pleat inspection that an IMS bearing failure in its early stages is not catastrophic. This also is supported by people who have installed a 'chips' detector light in the sump, and stopped at the first indication of metal in the juice. (My most exciting flight was near Ft Lewis in an AH-1L when the chips light came on. I auto-ed into a bean field and they came and got it with a big truck and trailer)
#36
The failure rate of these cars in the dual row bearing was 1% and the failure rate of the single row AT THE TIME OF THE CLASS ACTION was 8-10% (it was to this larger cadre of failures I was referring). Since that time, nobody is recording how many additional IMS bearings have failed. Nor are they recording how many REPLACEMENT bearings or engines have failed (I'm not talking about aftermarket replacement bearings...I'm talking single row or large captive IMS engine replacements from Porsche). Further, Porsche would not reimburse anyone with over 100k miles or 10 yrs on it at the time of damage...meaning their own confidence in the survival of bearings past that time was also low. You really have no factual basis to assume 8% failure rate at 85k miles assumption 6 years or so after the data was presented in the litigation. You talk about miles driven on a fleet of cars and if the driving habits of all of those drivers was the same, maybe there'd be more validity...but
your not talking sense, nore with facts.
have you abetter formula for working it out!!
if not please dont reply
#37
Rennlist Member
Originally Posted by Russ21
Porsche reimbursed on cars upto 130k miles
your not talking sense, nore with facts.
have you abetter formula for working it out!!
if not please dont reply
your not talking sense, nore with facts.
have you abetter formula for working it out!!
if not please dont reply
#38
I am not sure what you tried to calculate, but you did that wrong. 8% of 8,500,000 us 680,000; that's one.
Two, for your group of 100 cars, there will be a bearing failure for every 680,000 miles traveled by the WHOLE GROUP. This is an absolutely meaningless statistic for the IMS situation, as I have yet to meet anyone who would wonder.....what is the failure rate of the IMS bearing for the total miles traveled by all 996s every produced. The answer you get is no way car specific and doesn't give you ANY useful data. So you find out of all 996s ever made, the IMS bearing fails roughly every 1,000,000 miles traveled by ALL OF THE CARS as the group. And that is useful to you how?
Two, for your group of 100 cars, there will be a bearing failure for every 680,000 miles traveled by the WHOLE GROUP. This is an absolutely meaningless statistic for the IMS situation, as I have yet to meet anyone who would wonder.....what is the failure rate of the IMS bearing for the total miles traveled by all 996s every produced. The answer you get is no way car specific and doesn't give you ANY useful data. So you find out of all 996s ever made, the IMS bearing fails roughly every 1,000,000 miles traveled by ALL OF THE CARS as the group. And that is useful to you how?
Regards Russell
#39
Is this how you would talk to someone face to face? In front of your kids? Please take a step back and ask yourself is this a positive addition to forum or am I being a bit toxic? You've obviously read the lawsuit which is more than 98% of us. And I applaud that. But beating people over the head with it provides benefit to no one, not even yourself.
So now people can see how likely it is in the real world.
Some people doubt my equation and call me stupid or they try to dismiss it without a better scenario,
so when they question it I reply.
If it's questioned constructively I will give a constructive reply.
But if people criticise my spelling on a forum when they can't spell well what can I say.
I have had 1 other constructive criticism but his maths was incorrect, but I was extremely polite and explained his error without sarcasm.
If you or anyone comes up with a better equation I will back it, but you lot don't give me a hard time for helping people.
Regards
Russ
#41
Oh, man, Russ, you got me....I made a typo...I feel so humbled and ashamed. Yes, I read the settlement, too...it was 130k miles.
Regardless, you are calculating an incorrect Mean Time Between Failure because you don't know how many miles each of the failed 8 cars in your insurance study (that you referenced elsewhere) traveled before failing. In addition the 100 vehicles/8,500,000 miles driven is not a large enough sample for a good FIT/MTBF calc...more like 1,000,000,000 miles would be a better sample. Nor do I think 85k miles in that survey really expresses the useful life period of the automobile, a 25k to 200K mile survey would be more in the belly of the "bathtub curve", where infant mortality failures and wear out failures are excluded.
Anyway, I see what you are trying to do, but, again, I believe the assumptions are wrong. No attack, no reason to get your back up. Truth is, neither you nor I can calculate a true failure rate because we simply do not have the data required.
Regardless, you are calculating an incorrect Mean Time Between Failure because you don't know how many miles each of the failed 8 cars in your insurance study (that you referenced elsewhere) traveled before failing. In addition the 100 vehicles/8,500,000 miles driven is not a large enough sample for a good FIT/MTBF calc...more like 1,000,000,000 miles would be a better sample. Nor do I think 85k miles in that survey really expresses the useful life period of the automobile, a 25k to 200K mile survey would be more in the belly of the "bathtub curve", where infant mortality failures and wear out failures are excluded.
Anyway, I see what you are trying to do, but, again, I believe the assumptions are wrong. No attack, no reason to get your back up. Truth is, neither you nor I can calculate a true failure rate because we simply do not have the data required.
Last edited by il pirata; 03-02-2018 at 10:37 AM. Reason: taking out the name calling
#42
Former Vendor
Hi folks. I'm a longtime Porsche owner and just picked up a 2003 996 C4 Cabriolet. Brainstorming an alternate means of replacing/fixing the IMS bearing.
1. Replace stock IMS bearing with another stock bearing. I would remove the seals making the bearing open. Replace the bearing with this modified unit.
2. Tig weld a bung on the cover to accept an AN4 aircraft fitting.
3. Notch the case so a 90 degree line will pass thru.
4. Make a spin-on oil adapter fitting which will have a tapped hole to accept another AN4 fitting.
5. Run a braided AN hose between the two, like others have already done.
6. Use a non-bypassing K&P Engineering cleanable filter on the spin-on adapter.
I have a machine shop and its no problem for me to make these components.
Thoughts?
Thanks,
Bob
1. Replace stock IMS bearing with another stock bearing. I would remove the seals making the bearing open. Replace the bearing with this modified unit.
2. Tig weld a bung on the cover to accept an AN4 aircraft fitting.
3. Notch the case so a 90 degree line will pass thru.
4. Make a spin-on oil adapter fitting which will have a tapped hole to accept another AN4 fitting.
5. Run a braided AN hose between the two, like others have already done.
6. Use a non-bypassing K&P Engineering cleanable filter on the spin-on adapter.
I have a machine shop and its no problem for me to make these components.
Thoughts?
Thanks,
Bob
#43
Oh, man, Russ, you got me....I made a typo...I feel so humbled and ashamed. Yes, I read the settlement, too...it was 130k miles.
Man, you are some kinda smug prig. Speaking of spelling, mebbe you should first remove the plank from your own eye (ref. post 36)
Regardless, you are calculating an incorrect Mean Time Between Failure because you don't know how many miles each of the failed 8 cars in your insurance study (that you referenced elsewhere) traveled before failing. In addition the 100 vehicles/8,500,000 miles driven is not a large enough sample for a good FIT/MTBF calc...more like 1,000,000,000 miles would be a better sample. Nor do I think 85k miles in that survey really expresses the useful life period of the automobile, a 25k to 200K mile survey would be more in the belly of the "bathtub curve", where infant mortality failures and wear out failures are excluded.
Anyway, I see what you are trying to do, but, again, I believe the assumptions are wrong. No attack, no reason to get your back up. Truth is, neither you nor I can calculate a true failure rate because we simply do not have the data required.
To borrow a phrase....If you can't reply without being an ****, don't reply.
Man, you are some kinda smug prig. Speaking of spelling, mebbe you should first remove the plank from your own eye (ref. post 36)
Regardless, you are calculating an incorrect Mean Time Between Failure because you don't know how many miles each of the failed 8 cars in your insurance study (that you referenced elsewhere) traveled before failing. In addition the 100 vehicles/8,500,000 miles driven is not a large enough sample for a good FIT/MTBF calc...more like 1,000,000,000 miles would be a better sample. Nor do I think 85k miles in that survey really expresses the useful life period of the automobile, a 25k to 200K mile survey would be more in the belly of the "bathtub curve", where infant mortality failures and wear out failures are excluded.
Anyway, I see what you are trying to do, but, again, I believe the assumptions are wrong. No attack, no reason to get your back up. Truth is, neither you nor I can calculate a true failure rate because we simply do not have the data required.
To borrow a phrase....If you can't reply without being an ****, don't reply.
So 100 cars is a representative of all the cars. DO YOU UNDERSTAND.
I'm really sorry you don't understand but I can't teach you maths I just don't have the time.
But I can tell you every time you answer the credibility you had goes down a few % :-)
#44
LOL. I don't need you to teach me maths, Professor....I've studied probability and statistics and have a passing familiarity with failure analysis.
In another thread, you made the claim that your 8 failures in 100 car of 85k miles each was based on a study you saw by 1 insurance company, not by the overall figures in the settlement. Then you said 8% over the life of the car. Now you are saying you are merely simplifying the example to make it easy for the unwashed masses to understand based on the data from the class action. . Which is it? You also suggested that you can average the failure rates between the 2 classes of cars (dual row models and single row models) to 4%, which you cannot. You must look at them as discrete cases as they are completely different designs and in any engineering study would be analysed separately. .
You are presenting this 1,062,000 miles as a mean time between failure.
.
If you are relying on the insurance company analysis, then the data set is not big enough to do so, nor is it complete enough (ie no detail on when cars failed). If you are basing it on the class action lawsuit, the data is not complete (again, no data on when the cars failed) and does not cover the usable life of the car. To understand MTBF, you need to understand the up time of each car in the dataset....and none of these datasets provide that. These are facts that render your analysis inaccurate. That is all I am saying. . A failure every 1,062,000 miles, or whatever you suggested, is an optimistic view. Do i believe it falls to a MTBF of 100k miles if we had the correct data, NO.. I simply started out questioning your assumptions and you continue with your arrogant silliness..
Truth is, if I had a dual row, I'd leave it alone, myself. If I had a factory single row....I'd change that thing. I have a dual row with a replacement IMS (not The Solution) purchased by the PO, so I will be faced with a decision in a few years as the suggested maintenance interval is reached.
In another thread, you made the claim that your 8 failures in 100 car of 85k miles each was based on a study you saw by 1 insurance company, not by the overall figures in the settlement. Then you said 8% over the life of the car. Now you are saying you are merely simplifying the example to make it easy for the unwashed masses to understand based on the data from the class action. . Which is it? You also suggested that you can average the failure rates between the 2 classes of cars (dual row models and single row models) to 4%, which you cannot. You must look at them as discrete cases as they are completely different designs and in any engineering study would be analysed separately. .
You are presenting this 1,062,000 miles as a mean time between failure.
.
If you are relying on the insurance company analysis, then the data set is not big enough to do so, nor is it complete enough (ie no detail on when cars failed). If you are basing it on the class action lawsuit, the data is not complete (again, no data on when the cars failed) and does not cover the usable life of the car. To understand MTBF, you need to understand the up time of each car in the dataset....and none of these datasets provide that. These are facts that render your analysis inaccurate. That is all I am saying. . A failure every 1,062,000 miles, or whatever you suggested, is an optimistic view. Do i believe it falls to a MTBF of 100k miles if we had the correct data, NO.. I simply started out questioning your assumptions and you continue with your arrogant silliness..
Truth is, if I had a dual row, I'd leave it alone, myself. If I had a factory single row....I'd change that thing. I have a dual row with a replacement IMS (not The Solution) purchased by the PO, so I will be faced with a decision in a few years as the suggested maintenance interval is reached.
#45
Truth is, if I had a dual row, I'd leave it alone, myself. If I had a factory single row....I'd change that thing. I have a dual row with a replacement IMS (not The Solution) purchased by the PO, so I will be faced with a decision in a few years as the suggested maintenance interval is reached.