IMS bearing QnA

[IAM_911]

This is everything I learned about IMS (Inter-Mediate Shaft NOT Irritable Man Syndrome) bearing.  In the form of QnA, I have listed all the quotations I received and all the questions I could think off.  These questions are related to the IMS bearing it self. Not to the removal and installation procedure which will come later.

 

The answers are ether from reputable sources or what I have experienced my self. I prefer to stay away from articles found on the web and forum threads with a 12 year old contributes. If it is my opinion or what someone else thinks, I will say so.

 

- What type of bearing does a factory engine have?

 

From the overall reports on the web, the bearing can be any brand. SKF, FAG, NSK and others are noted in engines from the same model type and year.  I personally had an FAG made in Poland.

The single row bearing is a garden variety 6204 double sealed bearing. My MY03 had  FAG 6204-2RSR. I used the exact same one to replace.

 

- How common is IMS failure (IMS bearing failure rate)?

 

I have tirelessly searched for an answer. I did not want to do an open heart surgery on a relatively new car only to later find there was never the need to do so. After some weeks of looking into this issue I concluded: IMS was a real problem. However, the failure rate quoted remained speculative.

 

No reported Failure rate I cam across was backed by real verifiable data.  And no numbers I found had any verifiable reference.  I am yet to be corrected. But until then, I lean towards saying; only Porsche themselves know how many engines have failed as a direct result of IMS bearing failure.

 

-What symptoms does a failing IMS exhibit?

 

It is safe to assume there are no symptoms.  Having said that, you may want to lookout for the following:

 

a) metal shavings in the engine oil and oil filter. Worth inspecting every oil change;

 

distinct bearing noise coming from engine-gearbox area. This is extremely hard to pickup without knowing what this sound is and using a mechanics stethoscope.

 

c) very small changes in relative camshaft timing angle between Bank 1 and Bank 2.  There are suggestions that a bearing on the way out will have larger clearances in it's internal dimensions. These larger clearances causes the bearing to wobble slightly. Which in tern causes the IMS to wobble. And since the two camshafts are connected to both ends of IMS, a wobble in the IMS will result in relative camshaft angle to fluctuate.

 

I was not able to detect any movements in relative camshaft timing angle.  It remains to be seen if the camshaft position sensor or Durametric diagnostic tool for Porsche are able to detect such minute vibrations. After all my IMS bearing was wobbly at the time of replacement.

 

- When should I change the IMS?

 

The Porsche 911 Carrier Service Manual from Bentley Publishers recommends bearing replacement to be mileage based maintenance item between 50km and 100km.

I have changed the IMS bearing when I had to changed the RMS when it started leaking.  I am glad I did. At the time the car was exactly 10 years old and had 72km on the dial. The bearing had no trace of internal grease and had a bit of play in it.

 

- What bearing should I use?

 

I have tried to find an alternative to the common 6204 on my single row engine.  I wanted the best drop-in replacement money can buy.

 

While talking to bearing suppliers, one switched on sales person pointed out that the seal on the 6204 was not design to work in an such environment such as that found inside an engine.  The temperature is too high and the presence of motor oil will not help.

 

Normal 6204 bearings had temperature range between 80 and 120 deg C. This is very close to the Engine operation temperature.  When you add the heat generated internally the bearing may get too hot for it's own seals.

 

Next time the bearing is due for service I will opt for a high temperature bearing (e.g. SKF 6204 2RSH/C3GJN or FAG 6204-C-2Z).

 

- Is IMS special tool required to extract and install the IMS bearing?

In my opinion, YES.

 

There are internal bearing extraction tools. Most I have seen places the stress on the case shell when pulling the IMS outwards.  The IMS special tool will rest on the outer wall of IMS thus minimising stresses.

 

Bearing with seals removed and inside cleaned:

[Peter M]

Ian,

Thanks for posting your thought. Can I ask you a question or two please?

What was your reasoning for not going down the commonly travelled ceramic bearing path?

Also what is your views on the solution where they remove the oil seal from the standard bearing and install an oil feed line to keep the bearing clean and well bathed?

Thanks!

[IAM_911]

Peter,

 

Thanks for asking. I was waiting for a question about ceramic bearing.  The short answer is "conservation of misery".  Bearing design seems to be all about compromises.  This brings us to the next quotion:

 

First, you can by an unsealed bearing for this purpose. Not sure why anyone would pay more for a sealed bearing and then removes the seal!

 

Second, I am not sure engine oil is a good lubricant for this type of bearing.  I think the level of impurities is high enough to reduce the life of steel bearings.  This is why LN engineering uses ceramic bearing. Ceramic, being harder material, has higher resistance to wear from such impurities.

 

I could go more into details. But I must warn you; the long answer contains boring material.

 

I must remind everyone again. I am not a bearing expert or a mechanical engineer. I am only sharing what i have found when I have tried to make sense of the IMS bearing issue. Any helpful thought and hints will be appreciated.

 

Hope this helps,

 

Ammar

[Peter M]

Ammar,
I'm keen for the long answer!  

If you think that a ball bearing of any sort is compromised, why not bite the bullet and install the IMS Solution's plain bearing upgrade? I understand this isn't cheap but it's do and forget which surely isn't a bad thing?

Regards
Peter

[IAM_911]

Peter,

Are you sure IMS solution is a do and forget? I think I read on there website somewhere the bearing is to be changed every 5 years.???

[Peter M]

Ammar,

The solution I'm thinking of is the one where the ball bearing is replaced with a plain bearing: http://theimssolution.com/how-does-the-ims-solution-work/

 

I note that IMS do a comparison: http://theimssolution.com/retrofit-vs-solution/

[jay911]

Hi All,

 

Im new to the forum but found this thread interesting. I thought I would share my recent experience on the IMS bearing issue.

 

I have replaced my bearing with an LN Engineering bearing a few months ago, mostly as a precautionary step while the gearbox was out. We looked at the factory bearing and no issues at all with it. My car has about 104,000 km and has seen some track days in the last few years.

 

I went with LN Engineering bearing because the bearing has been tried and tested and especially in the US is quite a common replacement by many as a precautionary step.

 

I have also learned from some research on forums the IMS issue is not as bad as everyone suggests. It apparently affects less than 5% of all 996's made.

[Jimmy]

Hey Janik

Good to see you on here and welcome to the forum. Good to see you have the IMS sorted, lots of previous discussions on this topic.

Hope to see you at one of the get togethers.

Cheers

Jim

[jay911]

JimTT - Thanks mate i was inspired by our meeting.. Yeah please let me know when you guys get together I keen to catch up with like minded people...

 

Happy to do coffee when your free too mate...just drop me a text or call

 

chat soon

Janik

[HowardMoon]

I'm not a Porsche guy but I have come across this bearing failure mode several times so I thought I'd comment. It isn't all that common (I've encountered it maybe a dozen or so times in 30 odd years) but it certainly isn't unheard of. I have to say I'm surprised that no-one has bothered to contact the engineering guys from the bearing manufacturers - I'm sure they would have identified the problem very quickly. I'm also kind of surprised and disappointed that the guys making the repair kits have also failed to recognize the failure mode. Or maybe they have but it's just that it's in their best interests to ignore it and keep selling parts.

 

Before we go into what really happens let's debunk some of the silly theories. For a start it has nothing to do with lubrication. A little 6000 series bearing like this that is lightly loaded and running at moderate speed really needs only the slightest whiff of an oil mist to survive long term. So to argue that the lubricant is being washed out of the sealed bearing (by more lubricant!) is crazy. So long as a drop of engine oil reaches the bearing every now and then it'll be fine. Engine oil is a perfectly satisfactory lubricant and is sufficiently clean - the lower bearings in most gearboxes have to deal with much more debris. As an aside, despite being a hot topic on most car forums actual lubrication problems are quite rare in the real world, and whenever I hear people blame the type of oil for whatever problem they might have my first thought is that they have no idea what they are talking about. Most rolling element bearings will live a long time on just a little bit of oil of damn near any type.

 

It's nothing to do with heat either; a cooked bearing is obviously so. The amount of heat generated by the bearing in this application would be negligible.

 

It certainly isn't overloaded; the actual loading would be waay under the design specs. A bigger bearing would almost certainly fail even sooner.

 

Whenever I see this failure mode there are always a couple of factors present: one is that the bearing is only lightly and/or intermittently loaded. The other is that there is always some amount of torsional vibration present. Often when the bearing is in some sort of intermediate drive (like an idler gear for example) the loads on it tend to cancel out to some degree, so the actual load on the bearing ends up being quite small indeed, and this is part of the problem. Where I see the problem most often is in gearboxes with vertical shafts, where the bearings don't have the weight of the shafts providing a constant load.

 

As soon as a bearing has run for a couple of hundred hours the balls and tracks burnish to some degree and you end up with a tiny amount of clearance between them. This, coupled with the light loading causes the problem. Because the load is so light, the balls can skid in the tracks rather than roll. Add some TV and an irregular load (such as you might get from a camshaft) and the balls alternate between skidding and gripping/rolling, perhaps several times per revolution.

 

When the load is applied radially though, it's only applied to one, maybe two balls, and if these balls were skidding they are rapidly accelerated when they suddenly gain traction. But because the ball in question is coupled to the other balls via the cage, it also has to accelerate the mass of the other balls as well as the cage itself. A bit of TV (as you'd find in any engine) and the slip/stick cycle means that the balls are constantly tapping away at the cage.

 

The cage is usually pressed from two rings of sheetmetal spot-welded together between the balls. The constant tapping of the balls eventually starts to break the spot-welds and the cage begins to separate. This is the tell-tale sign; if you catch it early enough you'll find the bearing still looks fine but the balls will be all bunched up and the cage will be in two pieces. At this stage it may still function OK but usually there will be some clicking noise as the balls bounce against each other.

 

Depending on how much room is in the housing the cage halves will be either be just rattling around or doing a hula hoop routine. Eventually of course they break up and make their way through the bearing. This is what causes all the real damage. By the time all the bits of cage have been chewed up and passed through the bearing it will look very very sad. Both balls and tracks will be covered in dings and there will be glitter and debris in the sump. Surprisingly, this seems to happen without making a lot of noise. But by this stage of course the failure will be obvious.

 

How do you fix it? The best way, wherever possible, is simply to apply a constant axial load to the bearing. This is why, whenever you take apart something like an electric motor or alternator or whatever there is nearly always a wave washer behind the non-drive-end bearing. The shaft is located by the DE bearing, while the other bearing floats in the housing. The wave washer provides an axial load on both bearings that keeps the balls rolling continuously. The beauty of this is that the axial load is on all the balls simultaneously, so the cage is just along for the ride and only has to deal with its own mass even if there is TV present.

 

Ceramic bearings will last longer simply because the balls have only about 60% of the mass of steel balls, so therefore the cage isn't knocked around so badly. It's notable that this single-row-bearing-in-an-idler is something that you'll never ever find in a long-life engine - something like a Cat, Cummins or Detroit for example will have a pair of bearings pre-loaded against each other or a plain bearing instead, and these engines are often good for 20 - 30,000 hours between rebuilds.

 

I can't say I'm impressed with some of the mickey-mouse solutions from the aftermarket. If it was me I'd be looking into a simple axial preload (though this obviously won't work if the bearing at the other end is a needle or parallel roller for example) or maybe just a plain bush. I have to question the necessity of a dedicated oil supply with the aftermarket solutions too - the idler bearing rotates with a very small load compared with say a little-end bush so would need only a little oil. The little end survives just fine under heavy loads and with just a bit of splash plus it oscillates to boot, so I'm sure the idler bush could be made to live forever off the overflow from the rear main. Some other engines do exactly this.

 

Anyway, that's what happens in failures of this type, they're well know and well understood; I'm just surprised that Porsche used a design that allowed it to happen.

[Steiny]

Informative post Howard, thanks.

[Coastr]

Interesting read, thanks.

 

 

Anyway, that's what happens in failures of this type, they're well know and well understood; I'm just surprised that Porsche used a design that allowed it to happen.

 

They were broke and asked Toyota how to build cars more cheaply.  Toyota said they needed to have a single casting for the heads instead of two separate castings.  This presented the problem of driving the camshaft from different ends of the motor.  The solution was a roller IMS bearing.

 

The same thinking is why the 996/boxster had fried egg headlights.  Because it's one piece, it's easier to install and therefore faster.

 

It all worked-  on the back of the 996/997 Porsche became the most profitable car company in the world.  At the cost of the resale values of their cars and the squandering of the unburstable reputation the later air-cooled cars had developed.

[firstone]

Very well written too. I can picture exactly what you're saying. Thanks for your input.

Edit: The IMS solution (solid bush) seems to be the best answer then IMO.

[Zelrik911]

Thankyou both for those excellent explanations.

 

I am now convinced that all these engines will fail - its just a matter of bad/good luck as to when it happens.

 

As I posted before, 2 of my friends bought new Boxters & one blew at 40K (he was fortunate that he had just got a trade-in valuation on a new 911, so Porsche honoured that & took his dead Boxter away). 

 

But regardless of these problems, they both still drive late model Porsches today. 

[Peter M]

Howard,

 

The failure mode you describe for the IMS bearings sounds the same as the failure mode that afflicted the big end needle bearings on bevel cam drive Ducatis of the 70's and early 80's.  The reason given for these failures was the oil and everyone swore blindly that straight 50W was the only oil that would make them last.  Doesn't that sound similar! 

 

Now for your next topic to separate fact from fiction, I could suggest a few if you're interested.....

[IAM_911]

It isn't all that common, but it certainly isn't unheard of. I have to say I'm surprised that no-one has bothered to contact the engineering guys from the bearing manufacturers

 

So to argue that the lubricant is being washed out of the sealed bearing (by more lubricant!) is crazy. So long as a drop of engine oil reaches the bearing every now and then it'll be fine.

 

It's nothing to do with heat either.

 

It certainly isn't overloaded.

 

I can't say I'm impressed with some of the mickey-mouse solutions from the aftermarket.

 

Anyway, that's what happens in failures of this type, they're well know and well understood; I'm just surprised that Porsche used a design that allowed it to happen.

 

HowardMoon,

 

I congratulate you on your contribution. I was waiting for someone to help me understand more. You finally said what I needed to hear.  Only an engineer will use the phrase "mickey-mouse solutions". So you must know what you talking about.

 

I am glad it is not lubricant, heat or overload.But is it really underloaded?

Porsche changed the bearing to a larger size on the later M97 engines in the 997. Even though M97 are not significantly more powerful than M96 with the 6000 serious bearing. Just though I should mention.

 

I did contact a bearing engineer.  I did not take much of his time, But in short he did not see anything wrong with the current arrangement of IMS bearing.  The only thing he noted was what I mentioned regarding bearing seals.

 

This is why I did not like the mickey-mouse solutions either . However, some are successful in exploiting people fear of losing a $35000 engine in a pop.

 

We need more contributions like yours...so thank you,

[HowardMoon]

I think this shows the dangers in being single-marque-centric. If some of these guys had looked beyond the boundaries of their own little world they'd have quickly found out that other engine makers had encountered the same issues (and overcome them) over half a century ago. If you're having problems with a particular component wouldn't it make sense to see how the rest of the world does it? And if they have a good solution to simply do what they do?

 

Same goes for the doocahtee; the rod bearing problem was well known decades previously, as were the solutions. Why not just do what others did years ago? Or get rid of the damn needles altogether like pretty much everyone (including themselves) eventually did? I don't know but maybe some of those riders liked to pretend that their exotic engines were so powerful that the only way such a thoroughbred could be kept alive would be to feed it with a super-thick oil that may have helped cushion the blows.

 

I think no matter how loyal you are to your favourite brand you should keep your eyes open - sometimes you find the most elegant engineering where you'd least expect it.

 

IAM_911 - not underloaded so much as inconsistently loaded.

[Coastr]

NIH syndrome - 'not invented here'.  I battle it all the time.  Wheel reinvention is a popular pastime amongst some.

[Howe]

Howard, just a thought,

 

So instead of a cage why didn't Porsche put in a full complement ball bearing and not have any cage?

This would also improve the dynamic capacity. 

[Orion03]

well just remember that Porsche "liked" this IMS bearing system so much that they Got Rid of it ASAP

 

The New engines don't have IMS bearings !

 

The Old engines didn't have IMS bearings!

 

So I would say it's an "Oops" 

[Howe]

"I have to question the necessity of a dedicated oil supply with the aftermarket solutions too - the idler bearing rotates with a very small load compared with say a little-end bush so would need only a little oil. The little end survives just fine under heavy loads and with just a bit of splash plus it oscillates to boot, so I'm sure the idler bush could be made to live forever off the overflow from the rear main. Some other engines do exactly this..".............Have a ROTAX TWO STROKE in my plane all crankshaft bearing are lubricated by the 50 to one mix at 6500 rpm, so would have to agree with this.

[Peter M]

A little bit different Howe in that your Rotax would have ball bearing main bearings and needle rollers on the big ends and the gungeons pins.  The mains would only have seals on the outside and the connecting rod would likely have a big slot at each end so the balls and rollers would have sufficient exposure to the fuel/air/oil mixture in the crankcase to be sufficently lubricated and long lasting.

 

What HowardMoon is suggesting is that the IMS bearing in the 986 style engines would have survived if they were a plain bush that had some sort of facility for splash lubrication (ie a catch "tab" that directs the collected splash into the bearing) and that it wouldn't need to be pressurised like the plain bearings on the crankshaft due to the low loads on the IMS bearing.

I guess the IMS solution that involves replacing the pesky ball bearing with a plain bush has gone the pressurised route with an external supply taken off from the oil filter is that it is easier to do this than to modify the crankcase for a splash arrangement. The pressure oil supply method is also consistent with the air cooled cars BTW.

Cheers

Peter    

[tazzieman]

Might as well add this to the thread

 

996 a pinless grenade?

[IAM_911]

Good article. If the study holds up to scrutiny, a sample size of 200 cars is a good size.

The finding in this artical confirms my belief that L10 is over 10 years. That's good enough for me.

[LS1RX7]

I'm not a Porsche guy but I have come across this bearing failure mode several times so I thought I'd comment. It isn't all that common (I've encountered it maybe a dozen or so times in 30 odd years) but it certainly isn't unheard of. I have to say I'm surprised that no-one has bothered to contact the engineering guys from the bearing manufacturers - I'm sure they would have identified the problem very quickly. I'm also kind of surprised and disappointed that the guys making the repair kits have also failed to recognize the failure mode. Or maybe they have but it's just that it's in their best interests to ignore it and keep selling parts.

 

Before we go into what really happens let's debunk some of the silly theories. For a start it has nothing to do with lubrication. A little 6000 series bearing like this that is lightly loaded and running at moderate speed really needs only the slightest whiff of an oil mist to survive long term. So to argue that the lubricant is being washed out of the sealed bearing (by more lubricant!) is crazy. So long as a drop of engine oil reaches the bearing every now and then it'll be fine. Engine oil is a perfectly satisfactory lubricant and is sufficiently clean - the lower bearings in most gearboxes have to deal with much more debris. As an aside, despite being a hot topic on most car forums actual lubrication problems are quite rare in the real world, and whenever I hear people blame the type of oil for whatever problem they might have my first thought is that they have no idea what they are talking about. Most rolling element bearings will live a long time on just a little bit of oil of damn near any type.

 

It's nothing to do with heat either; a cooked bearing is obviously so. The amount of heat generated by the bearing in this application would be negligible.

 

It certainly isn't overloaded; the actual loading would be waay under the design specs. A bigger bearing would almost certainly fail even sooner.

 

Whenever I see this failure mode there are always a couple of factors present: one is that the bearing is only lightly and/or intermittently loaded. The other is that there is always some amount of torsional vibration present. Often when the bearing is in some sort of intermediate drive (like an idler gear for example) the loads on it tend to cancel out to some degree, so the actual load on the bearing ends up being quite small indeed, and this is part of the problem. Where I see the problem most often is in gearboxes with vertical shafts, where the bearings don't have the weight of the shafts providing a constant load.

 

As soon as a bearing has run for a couple of hundred hours the balls and tracks burnish to some degree and you end up with a tiny amount of clearance between them. This, coupled with the light loading causes the problem. Because the load is so light, the balls can skid in the tracks rather than roll. Add some TV and an irregular load (such as you might get from a camshaft) and the balls alternate between skidding and gripping/rolling, perhaps several times per revolution.

 

When the load is applied radially though, it's only applied to one, maybe two balls, and if these balls were skidding they are rapidly accelerated when they suddenly gain traction. But because the ball in question is coupled to the other balls via the cage, it also has to accelerate the mass of the other balls as well as the cage itself. A bit of TV (as you'd find in any engine) and the slip/stick cycle means that the balls are constantly tapping away at the cage.

 

The cage is usually pressed from two rings of sheetmetal spot-welded together between the balls. The constant tapping of the balls eventually starts to break the spot-welds and the cage begins to separate. This is the tell-tale sign; if you catch it early enough you'll find the bearing still looks fine but the balls will be all bunched up and the cage will be in two pieces. At this stage it may still function OK but usually there will be some clicking noise as the balls bounce against each other.

 

Depending on how much room is in the housing the cage halves will be either be just rattling around or doing a hula hoop routine. Eventually of course they break up and make their way through the bearing. This is what causes all the real damage. By the time all the bits of cage have been chewed up and passed through the bearing it will look very very sad. Both balls and tracks will be covered in dings and there will be glitter and debris in the sump. Surprisingly, this seems to happen without making a lot of noise. But by this stage of course the failure will be obvious.

 

How do you fix it? The best way, wherever possible, is simply to apply a constant axial load to the bearing. This is why, whenever you take apart something like an electric motor or alternator or whatever there is nearly always a wave washer behind the non-drive-end bearing. The shaft is located by the DE bearing, while the other bearing floats in the housing. The wave washer provides an axial load on both bearings that keeps the balls rolling continuously. The beauty of this is that the axial load is on all the balls simultaneously, so the cage is just along for the ride and only has to deal with its own mass even if there is TV present.

 

Ceramic bearings will last longer simply because the balls have only about 60% of the mass of steel balls, so therefore the cage isn't knocked around so badly. It's notable that this single-row-bearing-in-an-idler is something that you'll never ever find in a long-life engine - something like a Cat, Cummins or Detroit for example will have a pair of bearings pre-loaded against each other or a plain bearing instead, and these engines are often good for 20 - 30,000 hours between rebuilds.

 

I can't say I'm impressed with some of the mickey-mouse solutions from the aftermarket. If it was me I'd be looking into a simple axial preload (though this obviously won't work if the bearing at the other end is a needle or parallel roller for example) or maybe just a plain bush. I have to question the necessity of a dedicated oil supply with the aftermarket solutions too - the idler bearing rotates with a very small load compared with say a little-end bush so would need only a little oil. The little end survives just fine under heavy loads and with just a bit of splash plus it oscillates to boot, so I'm sure the idler bush could be made to live forever off the overflow from the rear main. Some other engines do exactly this.

 

Anyway, that's what happens in failures of this type, they're well know and well understood; I'm just surprised that Porsche used a design that allowed it to happen.

What you you say makes a lot of sense Howard. I am looking a buying a 1998-2000 996 at the moment, and I intended to do the bearing replacement for peace of mind. Your suggestion of a wave washer is interesting and I was wondering if this is something that should be considered when doing the bearing replacement and where and how should the wave washer be installed? Would it require any other modifications such as extra clearanceing or could it simply be used with the standard replacement bearing?                                    

Cheers, Jeff