Spline problem smoking gun?

[mbohn]

There have been many threads on this and other oilhead forums about premature spline failures. I’m talking about the gearbox input shaft here.

 

I recently responded to one thread with a brief description of my work on this problem and another person, nrp, said I may have the smoking gun. So, I thought I’d explain what I had done even though the jury is still out until I get a few more miles on the bike.

 

This is a very long, complex, painful story that is best described in detail some other time. Short version: my ’98 R11R began exhibiting shifting problems 8 months after new purchase, at about 14,000 miles. The problem is that after any engine braking the shifter lever locks up. BMW tried to fix the problem several times but never achieved a lasting fix. Even after I diagnosed the problem for them in August 2003 ago their subsequent 2 repairs also failed.

 

My diagnosis in August of 2003 proved that the crankshaft and the gearbox input shaft are not aligned. In making my diagnosis I first found that one of two alignment pins was missing. These are the two steel pins, about 0.5” diameter, that stick out of the end of the engine case and mate with sockets on the gearbox cover to force alignment between the engine and gearbox. If you’ve ever had your gearbox off the bike you know what I’m talking about. I also found that even _with_ the pin in place there was no way to align the crankshaft and the gearbox input shaft.

 

After that diagnosis BMW made repairs (new input shaft, clutch disk and some other related bits) but that failed after a few months. They then replaced the entire gearbox and that repair also failed after a few months. The second alignment pin was finally installed during one of these two repairs.

 

This fall I decided to fix this once and for all. I measured the misalignment again. I then calculated how much the gearbox cover alignment sockets would have to move to bring the crankshaft and gearbox input shaft into alignment. I took the gearbox cover to a local machine shop and asked them to move one of the alignment sockets in the cover. I asked for 0.010” movement and they delivered. When I re-measured the misalignment I found that the machine shop work did exactly what I needed but I now had a misalignment in the opposite direction. So I trimmed both alignment pins until I had very good alignment.

 

My goal was a misalignment of well under 0.003” starting from a misalignment of about twice that. I had been quoted the 0.003” tolerance by a machine shop for DIN 5481 splines, which I believe is the standard to which our gearbox input shaft splines comply. I achieved about 0.0005” misalignment.

 

I presently have about 600 miles on this latest repair and the bike shifts beautifully. What a thrill it is to be able to shift this bike whenever I want! My hope is to get to 1,540 miles before I report back here on a possible success. The reason for that mileage is that the first time BMW tried to repair the bike they just did a spline lube and that worked for 1,540 miles before the problem arose again. So, I assume that a spline lube will fix the problem for about that long and I need to go well beyond that before I can declare success.

 

If I can get well beyond the 1,540 miles I’ll report back. If people are interested at that point I can describe in detail how I did the repair and some alternatives.

 

Cross your fingers for me.

[russell_bynum]

Cool! Let us know how it goes.

[Rider1200RT]

Wow, what a nightmare to go through! Thanks for posting. I really would like to read the update too, when you have the miles on the repair. I am sure there are quite a few riders here who would like to as well.

[DogGone]

How did you measure the (mis)alignment?

 

How did the machine shop move the pin?

 

How did you trim the pins?

 

I'm not doubting your work, but I know how difficult it is to make these kinds of measurements. Did you build some kind of a jig?

 

pete

[nrp]

Don't take it apart if you got the alignment within .0005 inch runnout. I'll bet your problem will remain solved.

 

I peeked inside my 2000 R1100RT bell housing by taking the starter out @ 19000 miles. It was whistle clean inside & had no appreciable spline slop - something yours was not based on your pictures. So I have my fingers crossed.

[Clive Liddell]

Mark,

I'm also looking forward to your further report(s) back.

 

It went through my mind that eccentrically machined locating pins may be the way to go?

[mbohn]

How did you measure the (mis)alignment?

 

How did the machine shop move the pin?

 

How did you trim the pins?

 

I'm not doubting your work, but I know how difficult it is to make these kinds of measurements. Did you build some kind of a jig?

 

pete

Pete:

 

Here is what I had to say about the measurement method in the "other" thread:

 

==================================================

I built a tool that bolts to the end of the crankshaft using the flywheel bolt holes. The tool protrudes through the gearbox front cover (which has been removed from the gearbox). I then use a telescoping gauge between the tool and the input shaft front bearing seat to measure the radial distance from the tool to the bearing seat. I make this measurement at four locations roughly 10, 2, 5, 8 oclock. I'll try to attach a photo of the tool. The red radial line shows one such gaps I measure.

==================================================

A photo of the tool is attached.

 

The machine shop did not move the pins, they moved one of the mating sockets in the gearbox cover.

 

I trimmed the pins by hand with a small file, sandpaper, etc, about 0.001" at a time. (I was on a mission.)

[mbohn]

Mark,

I'm also looking forward to your further report(s) back.

 

It went through my mind that eccentrically machined locating pins may be the way to go?

 

Clive:

 

It is great to hear from you. We used to both participate in the old Cinammon list. Remember?

 

I discussed an eccentric pin with the machine shop and a few other ideas but we converged on the simplest one, just move the hole.

[DogGone]

Thanks for the clarification. Did it measure 10 thou out radially?

 

Did you repeat the measurements after rotating the crank?

 

Did you factor in the tool's tolerance?

 

pete

[mbohn]

Thanks for the clarification. Did it measure 10 thou out radially?

 

Did you repeat the measurements after rotating the crank?

 

Did you factor in the tool's tolerance?

 

pete

 

Pete:

 

It measured about 0.006" radial offset. This is between the crank axis and the input shaft axis.

 

I first measured the tool runout and documented that so I could correct the actual measurements. The runout was less than 0.001" so I felt pretty good about that.

 

Yes, I repeated the measurement many many times rotating the crank thru four positions. I knew the machining work would be costly so I didn't want to go forward until I was very confident of my measurements.

[nrp]

I think we're getting near critical mass here -

 

How did your machine shop move the holes? Did they swage them shut slightly (I think it is called Pittsburghing) & then bore them etc? Or did they just bore them oversize & use special dowel pins?

 

I like the idea from C Lidell on making custom dowel pins. Then it would only require a TIR measurement setup & the time to machine up new pins from screw stock to a particular offset spec. At least it would permanently take care of that particular combination of engine - transmission assembly. The user would have to reinstall the new set of pins in a correct orientation of course.

[SWB]

Mark,

 

I'm looking forward to your pictures, tool specs, and detailed diagrams. I know how to use a dial indicator, have measured "run out" on a spoked motorcycle wheel before, and have seen the BMW R/clutch/final drive vicariously through "Ted's Spline Lube DVD", but that's about the extent of my experience. I can understand what you're talking about, but can't quite visualize the whole thing.

 

If you're experiment works out, it's going to help out a lot of Beemer owners. (You're going to get some sort of whole new "oilheader" award on IBMWR.)

[DogGone]

Mark, very impressive work. I can't think of anything you haven't taien into account. Many people have suggested the alignment as a potential problem, but you are the first I'm aware of to measure it.

 

In one of the other threads on this topic I tried to point out the costs to bmw, even if they knew what caused this, for them to inspect all potentially affected bikes. It is my opinion that bmw is of the opinion it's cheaper to fix the ones that fail, than to inspect all of them. Especially when they would have to fix the failing ones anyway.

 

How many hours to disassemble, inspect, measure and reassemble? Times how many bikes? And on the ones that are out of spec you know that bmw isn't just going to remachine the case, they would have to replace it. If it isn't the transmission case that's off then they would be replacing the engine case

 

My only remaining question is, are there bikes with misalignment that do not suffer a failure?

 

Again, well done.

 

pete

[Stan Walker]

My only remaining question is, are there bikes with misalignment that do not suffer a failure?

 

And my only remaining question is, will a simple radial misalignment cause the uneven wear patterns that we have seen time and time again.

 

I have my own opinion as to what the wear pattern would look like given only a simple radial misalignment, but I make no claims to being a mechanical engineer, nor do I play one on TV.

 

I am most interested in hearing other opinions on this, all ears so to speak.

 

Stan

[SWB]

In one of the other threads on this topic I tried to point out the costs to bmw, even if they knew what caused this, for them to inspect all potentially affected bikes. It is my opinion that bmw is of the opinion it's cheaper to fix the ones that fail, than to inspect all of them. Especially when they would have to fix the failing ones anyway.

 

How many hours to disassemble, inspect, measure and reassemble? Times how many bikes? And on the ones that are out of spec you know that bmw isn't just going to remachine the case, they would have to replace it. If it isn't the transmission case that's off then they would be replacing the engine case.

 

I don't inspecting every bike would be an appropriate, or cost-effective solution. The bikes that are breaking are still a minority of the Beemers.

 

First off, if BMW ACKNOWLEDGED the problem, it could FIX THE PROBLEM in either the engineering or manufacturing process, improving the reliability of their machines and solidifying their customer base. Second, it could set the standards by where they'd fix the problem bikes. They could extend the drive train warranty in the area of the splines to 100K miles. They could develop a procedure for service departments to measure the spline alignment (if that in fact is the problem) on every bike that's apart for service, and fix those found to have problems. Certainly, such a procedure would improve the chances that a transmission/final drive/spline failure would reoccur.

 

Instead, BMW has responded that there have been "no warranty related failures". I think the problem is Teutonic pride first, and cost avoidance second.

 

Someone at BMW is burying their head in the sand. We've had 3 or 4 spline failures brought to our attention IN THE PAST MONTH ALONE! As I've said before, BMW supplies the parts. They know precisely how many failures have occurred. That is what ticks me off most. I bought my BMW because I'd heard for years (back to the 70's) of their legendary reliability. Now, I'm finding out that what is legendary at BMW is stubbornness, NOT reliability.

 

I'm not trying to (mis-)use this forum to berate BMW, but with every additional spline failure I hear off (particularly with NEW, LOW MILEAGE bikes!), I get just a little more ticked off.

[DogGone]

...

I'm not trying to (mis-)use this forum to berate BMW, but with every additional spline failure I hear off (particularly with NEW, LOW MILEAGE bikes!), I get just a little more ticked off.

 

Which of the recent failures have been reported on 2006 or even 2005 models? I've read the posts and must have missed that.

 

pete

[mbohn]

I think we're getting near critical mass here -

 

How did your machine shop move the holes? Did they swage them shut slightly (I think it is called Pittsburghing) & then bore them etc? Or did they just bore them oversize & use special dowel pins?

 

I like the idea from C Lidell on making custom dowel pins. Then it would only require a TIR measurement setup & the time to machine up new pins from screw stock to a particular offset spec. At least it would permanently take care of that particular combination of engine - transmission assembly. The user would have to reinstall the new set of pins in a correct orientation of course.

 

nrp:

 

First of all I apologize for dribbling out info on what I did. It was my intention to gain confidence in the fix before I opened my big mouth. But there has been so much activity on this subject in this forum I decided to start this thread. I'm presently installing a new alternator on my bike (yes, it is a cursed bike) and I'll soon be able to put some miles on it.

 

I've attached a photo of the moved socket. The machine shop located the original hole, moved 0.010" in the direction I asked, drilled a new hole about 0.040" oversize, and fabricated a brass bushing to an interference fit in that oversize hole and with an i.d. appropriate for the steel pin diameter. I heated the gearbox cover and pushed the bushing in place.

[mbohn]

My only remaining question is, are there bikes with misalignment that do not suffer a failure?

 

And my only remaining question is, will a simple radial misalignment cause the uneven wear patterns that we have seen time and time again.

 

I have my own opinion as to what the wear pattern would look like given only a simple radial misalignment, but I make no claims to being a mechanical engineer, nor do I play one on TV.

 

I am most interested in hearing other opinions on this, all ears so to speak.

 

Stan

 

Stan: I believe it does. When I first diagnosed the problem the clutch disk really wobbled on the input shaft. I recall the outer edge of the clutch disk moved something like 3/8" in a direction parallel to the shaft axis. I've attached a photo showing what those splines looked like.

 

Mark

[mbohn]

Instead, BMW has responded that there have been "no warranty related failures".

I'm not sure where they said that but it is not true. I've documented 6 warranty repairs on my bike for this problem. That doesn't include my recent repair, #7.

 

Mark

[Stan Walker]

I've attached a photo showing what those splines looked like.

 

I'm sorry, but I've lost the thread again. From which of the many repairs is this photo from? Before the missing pin was installed, or after?

 

The wear on your pictured splines looks fairly even along the length of the engaged spline area. That is what I would expect of a radial misalignment. Many of the spline problems pictured here have had wear that was pronounced at one end of the splines and almost absent at the other. That's the case that worries me.

 

Stan

[russell_bynum]

Many of the spline problems pictured here have had wear that was pronounced at one end of the splines and almost absent at the other. That's the case that worries me.

 

That would indicate the input shaft (or the clutch) is wobbling or that the clutch and the input shaft are not parallel. Right?

[nrp]

I would speculate that the input shaft is being continuously flexed by the runnout. The deflections involved are comparatively small (like .005 inch) but bearings are very stiff radially.

[Stan Walker]

That would indicate the input shaft (or the clutch) is wobbling or that the clutch and the input shaft are not parallel. Right?

 

Once the disc is clamped between the pressure plates I don't see any way it can wobble short of some really bad bearings!!!!.

 

I have long suspected that the axis of the two shafts are not parallel on those bikes where the wear on the input splines is not even.

 

I bet this just frustrates the h@ll out of the BMW engineering group. They redesigned the splines for the 1100 series to prevent problems they had on airheads. The did it again for the 1150's to solve failures they had on the 1100's and made it even worse. Now they have done it a third time for the 1200's. Maybe this time they will get it right. Maybe!!! What's so funny is that there are three other sets of splines in the drive shaft and they virtually never seem to fail even though they carry heavier loads. (hint, they are not locked into a rigid alignment to each other).

 

Stan

[Paul Mihalka]

IF BMW found out what the problem "was", as to them it is past, '04 is two years ago, their chance to fix it was the completely redesigned hexheads. Let's hope they fixed it and did not build in new catastrophic failure opportunities.

[Stan Walker]

Let's hope they fixed it and did not build in new catastrophic failure opportunities.

 

We will forget all about splines when the counter balancers start exiting through the side of the crankcase.........

 

Stan

 

oops, forgot a smiley, how else would you know I was only kidding!!!

[nrp]

I really don't see this so much as a BMW engineering mistake. Rather it is becoming a substantial marketing screwup.

 

Given the concept of a dry clutch assembly (and I realize that is not necessarily a given) and the short transmission input shaft extension, there really isn't any other way to do this alignment task other than being darn careful in manufacturing process. Properly fixtured, modern numerical manufacturing centers should be able to hold the necessary accuracies on these parts.

 

It appears the accumulation of manufacturing tolerances have gotten out of hand. But looking at it from BMW's standpoint, they may have considered it most economical to fix what they thought were a few badies enough to get thru warranty & lean on Fritz on the shop floor to watch what he is doing. However, they (and us for that matter) may not realize extent of the problem. Certainly tearing every bike apart to make runnout measurements would be incredibly expensive.

 

It still bothers me though that it isn't just early bikes that are having this problem. Manufacturing should have gotten this straightened out by now, and BMW customer service should be screaming for field runnout data on failed units.

[Silvereagle]

Mark---Looking at the picture of your worn input splines, it looks to me that there is indeed more wear in the spline contact area closest to the tranny cover. That would seem to indicate a slight tilt between the two axis.

 

My '02 RT suffered a spline failure on 4 Nov 05 at about 35,800 miles while at 65 mph highway cruise with no prior indication of any problem. I was on an out of state trip at the time. It is still at the dealer awaiting repair.

 

Many thanks for your diligence pursing this dilemma.

 

John

[russell_bynum]

Given the concept of a dry clutch assembly (and I realize that is not necessarily a given) and the short transmission input shaft extension, there really isn't any other way to do this alignment task other than being darn careful in manufacturing process.

 

How about a pilot bearing supporting the front of the input shaft...like cars have?

[mbohn]

Mark---Looking at the picture of your worn input splines, it looks to me that there is indeed more wear in the spline contact area closest to the tranny cover. That would seem to indicate a slight tilt between the two axis.

 

My '02 RT suffered a spline failure on 4 Nov 05 at about 35,800 miles while at 65 mph highway cruise with no prior indication of any problem. I was on an out of state trip at the time. It is still at the dealer awaiting repair.

 

Many thanks for your diligence pursing this dilemma.

 

John

John:

The first time I disassembled the bike to diagnose the problem I put the gearbox front cover on a milling machine to determine if there was a significant tilt between the surface that mates with the engine and the surface that mates with the gearbox. It looked good enough that I pursued other sources of misalignment.

[mbohn]

I really don't see this so much as a BMW engineering mistake. Rather it is becoming a substantial marketing screwup.

 

Given the concept of a dry clutch assembly (and I realize that is not necessarily a given) and the short transmission input shaft extension, there really isn't any other way to do this alignment task other than being darn careful in manufacturing process. Properly fixtured, modern numerical manufacturing centers should be able to hold the necessary accuracies on these parts.

 

It appears the accumulation of manufacturing tolerances have gotten out of hand. But looking at it from BMW's standpoint, they may have considered it most economical to fix what they thought were a few badies enough to get thru warranty & lean on Fritz on the shop floor to watch what he is doing. However, they (and us for that matter) may not realize extent of the problem. Certainly tearing every bike apart to make runnout measurements would be incredibly expensive.

 

It still bothers me though that it isn't just early bikes that are having this problem. Manufacturing should have gotten this straightened out by now, and BMW customer service should be screaming for field runnout data on failed units.

nrp: I agree. In my case it was especially interesting that there was one of the two alignment pins missing from the factory and there was a misalignment between the engine and gearbox. I don't see how these two are unrelated but I'll leave it to you to draw your own conclusions. I've drawn mine and it is not a pretty pricture re BMW QA.

[nrp]

Having a pilot bearing would force the input shaft to flex enough to handle any misalignment. At say .005 inch runnout (runout?) that would be a lot of flex to accommodate over say only a 2 inch shaft extension. The radial loading of the spline would still be substantial, and a similar locked-in radial load would be experienced by the three bearings involved. I would expect the transmission front bearing to fail first. Three bearings in a row is an engineering no-no unless very controlled or very compliant.

 

Another way would be to eliminate all dowelling or piloting except for adding the pilot bearing, and float the transmission into a zero stress position on assembly. I think however that would still have a lot of error if the assembly was pulled up (i. e. bolts torqued up) & done in an improper order. (Others - do cars do this?)

 

I strongly feel that owners of spline eating bikes are going to have to figure out a way to get runout (I think that is the proper spelling) info & compensate accordingly.

 

I am going to take my 1975 R90/6 (it ate its first spline etc at 20000 miles) apart again real soon & I'm gonna check runout this time assuming the splines are really worn.

[mbohn]

Having a pilot bearing would force the input shaft to flex enough to handle any misalignment. At say .005 inch runnout (runout?) that would be a lot of flex to accommodate over say only a 2 inch shaft extension. The radial loading of the spline would still be substantial, and a similar locked-in radial load would be experienced by the three bearings involved. I would expect the transmission front bearing to fail first. Three bearings in a row is an engineering no-no unless very controlled or very compliant.

 

Another way would be to eliminate all dowelling or piloting except for adding the pilot bearing, and float the transmission into a zero stress position on assembly. I think however that would still have a lot of error if the assembly was pulled up (i. e. bolts torqued up) & done in an improper order. (Others - do cars do this?)

 

I strongly feel that owners of spline eating bikes are going to have to figure out a way to get runout (I think that is the proper spelling) info & compensate accordingly.

 

I am going to take my 1975 R90/6 (it ate its first spline etc at 20000 miles) apart again real soon & I'm gonna check runout this time assuming the splines are really worn.

There seems to be a tendency to overheat the clutch disk/pressure plate for these misaligned bikes. That is based on a sample of two, mine and a riding buddies'. His required a new clutch disk every 8000 miles. Mine showed blue marks on the pressure plate. This may indicate the stress the clutch disk is under when the engine/gearbox are out of alignment.

[nrp]

Could this be from the clutch disk orbiting around the pressure plate as the clutch is being forced to rotate about a different center? Remember there would also be a slight simultaneous rotational slipping that would not be apparent to the rider. It could be the heat source.

[mbohn]

Could this be from the clutch disk orbiting around the pressure plate as the clutch is being forced to rotate about a different center? Remember there would also be a slight simultaneous rotational slipping that would not be apparent to the rider. It could be the heat source.

Exactly. At my first disassembly I took a video of my clutch disk rotating with the crank (gearbox removed of course). It wobbled very noticeably, about 0.020" total radial runout. Imagine the stress this must put on the input shaft. I still have the video if anyone is interested.

[bmwmick]

Exactly. At my first disassembly I took a video of my clutch disk rotating with the crank (gearbox removed of course). It wobbled very noticeably, about 0.020" total radial runout. Imagine the stress this must put on the input shaft. I still have the video if anyone is interested.

 

Mark,

Looking at your input shaft wear, you have slightly more wear at the rear of the contact area and that indicates a wobble (which I think is the more common failure) but your splines seen to be worn more all over than ones I've seen in the past.

If the axis was off radially, I would expect blue pressure plates. If the axis were off at some angle (which I think happens in almost ALL the failures) there would be flexing of the friction disc at all times when the clutch is engaged. I think THAT is what is at play when the center of the engagement area of the splines looks almost new but the splines nearest the transmission are knife-edge thin.

Yours would have been using the friction disc like an orbital sander on the pressure plates.

 

Can't wait to see that video......

 

Mick

[mbohn]

The video can be accessed here:

 

http://www.wispertel.com/~mbohn/mov00375.mpg

 

The flywheel is being turned by the starter. The dial indicator is touching the outer machined edge of the clutch disk hub. Dial indicator movement you see is total of about 0.020" . This radial position of the clutch disk is where it was left the last time the bike was operated.

[nrp]

Dramatic video! I have to assume the diameter of the indication is representative of the spline axis.

 

The extra near-the-transmission input shaft spline wear is probably due to input shaft deflection under the radial load applied to it by the clutch disc. Deflection would create the heaviest unit radial loading of the spline immediately adjacent to the transmission. Eventually wear would remove most of the spline teeth with the debris as fretting corrosion until there isn't enough left of them to carry the torque.

 

The error probably isn't angular as that sort of thing would be pretty easy to control in a production environment. To get a remotely similar spline fretting situation as say a .005 inch radial runout would require on the order of 1/16 inch out-of-parallel error across the bell housing face. Not impossible - but I'd think unlikely & it would be easy to measure without special tools.

 

THX

[chrisx]

i phone the local bmw dealer ( name left out ) and they gave me the blink blink empty stare response whatcha talkin bout willis. acted like he had never heard of the issue. gee what a suprise. i wonder how one can verify the offset(s) without the trans case front as per mbohn. the dealer indicated that engineering drawings were "not" available. no suprise on that answer either. any suggestions on how to get the "blueprint" for those mating surfaces? that seems to be the only way to set it up correctly with a jig .vs. runout on the trans bearing bore.

[SWB]

i phone the local bmw dealer ( name left out ) and they gave me the blink blink empty stare response whatcha talkin bout willis.

 

I think this is a bit beyond the local bmw dealers.

 

I sent a note (independently) to BMW Motorrad Germany, BMW Motorrad USA, and a "cc" to both my local dealers' service writers. I explained the "concerns that owners have had regarding premature spline failures", indicated that some are trying to diagnose and solve the problem on their own, stressed that this is a good-will effort on the part of owners, and suggested that BMW might find it in their interest to have an engineer log in and participate in this discussion.

 

If the company is actually interested in improving their product's reliability or retaining their owner base, they'd be crazy not to jump in. Liability issues (if there are any) should be secondary to diagnosing the problem, repairing the problem, preventing it in future bikes, and shoring up the confidence of their existing owner base in their product.

 

We'll see how BMW responds. We have all expressed our confidence in the company with our pocket books. We should expect a good response from BMW.

[SWB]

...

I'm not trying to (mis-)use this forum to berate BMW, but with every additional spline failure I hear off (particularly with NEW, LOW MILEAGE bikes!), I get just a little more ticked off.

 

Which of the recent failures have been reported on 2006 or even 2005 models? I've read the posts and must have missed that.

 

pete

 

I'm speaking of new being 2004 models with low mileage. There are still a few never-registered 2004's at dealers around the country. No, I've not yet heard of 2005/2005 models with the problem, but then again, it may take a while for some ex-Harley riders to get 6K miles on their new Beemers.

 

It will be interesting to hear of any failures with the new bikes, and even more interesting if the problem ends with the 2004 models. Again, only BMW knows if it's replacing splines at the same rate on the newer model; they have the parts-failure database.

[SWB]

Instead, BMW has responded that there have been "no warranty related failures".

I'm not sure where they said that but it is not true. I've documented 6 warranty repairs on my bike for this problem. That doesn't include my recent repair, #7.

 

Mark

 

Sorry, I missed this post earlier. (I still don't know how to search this site for all responses to my last post, as is the case on other forums. Unless I check back on a thread I've been active on, I don't know if someone's responded to what I posted. )

 

peterbulgar quoted a respones from BMW here.

 

"We are aware of the concerns being expressed through various websites. However, BMW is unaware of any cases which have been confirmed as a warrantable repair (a result of a defect in materials or workmanship). To date, we have found the vast majority of these complaints to be attributed to normal wear or external influence, such as misuse or neglect. However, we are still investigating this matter. We would strongly recommend you stay in contact with the service team at your BMW Retailer for further updates."

 

Of course, the dealerships know nothing, because BMW Engineering has told them nothing. Disappointing.

 

More disappointing is that BMW seems to be indicating that you caused your own bike's spline failure problem.

[SWB]

The flywheel is being turned by the starter. The dial indicator is touching the outer machined edge of the clutch disk hub. Dial indicator movement you see is total of about 0.020" . This radial position of the clutch disk is where it was left the last time the bike was operated.

 

This is a significant video, only because the clutch wobble is so well documented.

 

I just looked at the pictures of your tool, the failed spline, and your case adjustment. I can't quite get my "head" around all of this, probably because I've yet to tear down the bike. I'm wondering if either the crankshaft or gearbox input shaft (probably the later) could have been damaged because the bike was originally built out of alignment, due to the missing alignment pin. Is it possible that your case/alignment pin adjustment was compensating for damage shaft(s)?

 

We'll know how your situation works out as the bike wears. It would seem that a good way to begin verifying your theory is to put together a kit with instructions and the shaft alignment tool to calculate runout for every bike which is torn down for spline lube, clutch replacement, and so forth. A high precentage of coincidence between damaged and undamaged splines with out-of-alignment and alignment-within-tolerance shafts would gradually prove your theory.

 

Going one step further, it's be nice to refine the case re-engineering so that twiddling with the alignment pins isn't necessary. Ideally, the case would be ready to install, and the pins correctly set up to align the case and shafts to within tolorances, upon return from the machine shop.

 

My bike's ready for a new clutch (and maybe spline lube). How did you make the tool, and could I pay you to have another one made and mailed to me? (also, diagrams and/or detailed pix on how you set up your dial indicator to measure the clutch disk hub wobble would be great too).

 

I realize that this is putting a lot on your shoulders, but thanks in any case.

[DogGone]

i phone the local bmw dealer ( name left out ) and they gave me the blink blink empty stare response whatcha talkin bout willis. acted like he had never heard of the issue. gee what a suprise. i wonder how one can verify the offset(s) without the trans case front as per mbohn. the dealer indicated that engineering drawings were "not" available. no suprise on that answer either. any suggestions on how to get the "blueprint" for those mating surfaces? that seems to be the only way to set it up correctly with a jig .vs. runout on the trans bearing bore.

 

Well, I can say that the dealer will not get the blueprints or engineering docs, never happen. And neither will you without some kind of court order. Companies guard that info as trade secrets.

 

As far as the dealer towing the company line, they don't have much of a choice since bmw doesn't acknowledge there is a problem. What's the dealer supposed to do, fix it on his dime? bmw won't reimburse him for fixing a problem that doesn't officially exist.

 

pete

[mbohn]

I just looked at the pictures of your tool, the failed spline, and your case adjustment. I can't quite get my "head" around all of this, probably because I've yet to tear down the bike. I'm wondering if either the crankshaft or gearbox input shaft (probably the later) could have been damaged because the bike was originally built out of alignment, due to the missing alignment pin. Is it possible that your case/alignment pin adjustment was compensating for damage shaft(s)?

I originally thought that the missing alignment pin was the cause of the input shaft spline wear. But that is not the case. My measurements in 2003 showed that even with the pin in place there was still a significant misalignment. Sure enough, after that diagnosis a dealer replaced the pin, input shaft and some other related parts and there were two more shaft failures. My belief now is that the engine case was machined incorrectly at the factory and the pin was left out to allow assembly to proceed.

We'll know how your situation works out as the bike wears. It would seem that a good way to begin verifying your theory is to put together a kit with instructions and the shaft alignment tool to calculate runout for every bike which is torn down for spline lube, clutch replacement, and so forth. A high percentage of coincidence between damaged and undamaged splines with out-of-alignment and alignment-within-tolerance shafts would gradually prove your theory.

As I said earlier in this thread, I wanted to wait for enough miles on my repair before I start touting this as a "cure." I am totally dedicated to helping whoever needs help on this but I prefer to wait for good proof. When that happens, I can put together a complete package for folks.

Going one step further, it's be nice to refine the case re-engineering so that twiddling with the alignment pins isn't necessary. Ideally, the case would be ready to install, and the pins correctly set up to align the case and shafts to within tolerances, upon return from the machine shop.

That may be possible - it would likely require moving both sockets at the machine shop, which would be preferable to trimming the pins, IMO. My measurements indicated that I could move only one socket but when I got home I still had a problem in the direction normal to the socket movement we made. I haven't put much more thought into how I could have avoided it but that would be part of the "package" you referred to earlier. It would require one more measurement that I was not equipped to make: a precise measurement of the distance between the two alignment pins. Ideally the "package" would include the tool drawing and a spreadsheet/program to take the measurements and produce the movement required in the two sockets.

My bike's ready for a new clutch (and maybe spline lube). How did you make the tool, and could I pay you to have another one made and mailed to me? (also, diagrams and/or detailed pix on how you set up your dial indicator to measure the clutch disk hub wobble would be great too).

I designed the tool and fabricated it on a lathe. With my drawing I believe any competent machine shop could fab it, there is really only one critical cut: the 0.985" dia is what locates the tool concentrically on the crank. I’ll post the drawing a little later today.

[mbohn]

Alignment tool drawing can be found here.

[nrp]

Mark - Wouldn't it be easier to just make a non-precision crankshaft extension that will project thru the unput bearing housing & then use a smaller dial test indicator to measure the runout with its probe on the bearing housing bore?

 

If we could make offset pins with an offset witness mark on them, this problem could be fixed without having to get a machine shop involved once the runnout was measured. The runnout measursment could be made by a skilled mechanic or a dealer.

 

If the crankshaft main bearing clearance is of concern, it could be statically measured with the above setup, and we could assume the crank runs at the average measurment.

[smiller]

My belief now is that the engine case was machined incorrectly at the factory and the pin was left out to allow assembly to proceed.

 

[chrisolson]

Liability issues (if there are any) should be secondary to diagnosing the problem, repairing the problem, preventing it in future bikes, and shoring up the confidence of their existing owner base in their product.

IMHO, liability and the resulting potential legal action is exactly the reason BMW has not acknowledged a problem.

 

As has been said before, the US market is not huge for BMW. To willingly create a feeding frenzy of litigation and expense for an unknown number of bikes (so you have to recall all of them ?) is a tough financial decision.

 

Given that it would be difficult to correlate any recent or significant amount of lost sales of BMW models based on this issue - even as well discussed as it has been - probably only confirms their decision to address this only when pressed.

 

The problem is that the company has not done a better job of understanding, repairing and extending assistance for longer time/mileage periods for specific bikes shown to have an issue.

[mbohn]

Mark - Wouldn't it be easier to just make a non-precision crankshaft extension that will project thru the unput bearing housing & then use a smaller dial test indicator to measure the runout with its probe on the bearing housing bore?

 

You’re saying the dial indicator would be attached to the extension and rotate with the crank? That might work if you could get a true radial runout measure. The trick, I think, would be to get the probe to fit inside the radial gap (between the crank extension and the bearing housing) and still make a radial measure. If it is not purely radial you are going to lose resolution. Also, you could only turn the crank less than one revolution – there is a cutout in the bearing housing that would trip up the probe. You can’t turn that engine backwards, it tends to jam up the timing chain.

 

Maybe another approach, if there is enough interest, is that we have another tool made out of stainless steel (mine is aluminum and wouldn’t take much use/abuse) and somehow spread the cost among users. Maybe you pay a “rental fee” or something like that.

If we could make offset pins with an offset witness mark on them, this problem could be fixed without having to get a machine shop involved once the runnout was measured. The runnout measursment could be made by a skilled mechanic or a dealer.

Are you going to take bike to the machine shop for that measurement with the gearbox off the bike? Huge hassle. And I don’t really see BMW dealers getting involved with this. You’ve got the gearbox cover off anyway so a machining operation on the cover would logically follow. When I discussed “adjustable” options with my machine shop we kept coming back to the question: how rigid would that be? Could your offset pins be kept from rotating out of adjustment? It seems what you are trying to come up with is the easiest way for an affected owner to solve this problem and that is a good thing to think about.

If the crankshaft main bearing clearance is of concern, it could be statically measured with the above setup, and we could assume the crank runs at the average measurment.

I’m not sure why that bearing clearance would be a concern, maybe I don’t follow you.

[nrp]

This is gonna be confusing -

 

M - You’re saying the dial indicator would be attached to the extension and rotate with the crank? That might work if you could get a true radial runout measure. The trick, I think, would be to get the probe to fit inside the radial gap (between the crank extension and the bearing housing) and still make a radial measure. If it is not purely radial you are going to lose resolution. (NRP - You'd loose accuracy yes) M - Also, you could only turn the crank less than one revolution – there is a cutout in the bearing housing that would trip up the probe. You can’t turn that engine backwards, it tends to jam up the timing chain.

 

NRP - If you used a dial TEST indicator (they are comparatively small and have a lot of probe adjustment options and a very short stroke), I think it could be made to be mounted to the crank extension & rotated inside the transmission housing. They cost like $40 (cheap) and if set up properly could bridge across what must be a lube gap. The advantage is that there is no precision involved other than the dial test indicator & that is purchased. The rest is stiffness of the crank extension & since that only has to be like 3 to 4 inches, that should be no sweat.

 

M - Maybe you pay a “rental fee” or something like that.

 

NRP - I was thinking a minor rental plus a healthy deposit. The fixturing would take some design work.

 

M - Are you going to take bike to the machine shop for that measurement with the gearbox off the bike?

 

NRP - No - this could be done by a good mechanic with maybe a video to show people who have not worked with a dial test indicator, how to handle them. Once the runout is measured and the correction vector established, its telephone time to the offset pin maker - which could be anyone here with a lathe. The pins would press into the engine block. BTW - Didn't your old ones press in?

 

M - And I don’t really see BMW dealers getting involved with this. (NRP - depends on how dedicated to customer problems they are) M - You’ve got the gearbox cover off anyway so a machining operation on the cover would logically follow (NRP - But dealers can't nor would they have to do this) . M - When I discussed “adjustable” options with my machine shop we kept coming back to the question: how rigid would that be?

 

NRP - Rigidity is provided by the bolt-up face friction. The pins just hold things in alignment while bolting up.

 

M - Could your offset pins be kept from rotating out of adjustment?

 

NRP - I assume they would press in - or be staked in. M - It seems what you are trying to come up with is the easiest way for an affected owner to solve this problem

 

(NRP - yes. There are some VERY skilled bike mechanics & probably machinists too out there I'm convinced.

 

NRP - If the crankshaft main bearing clearance is of concern, it could be statically measured with the above setup, and we could assume the crank runs at the average measurment.

 

M - I’m not sure why that bearing clearance would be a concern, maybe I don’t follow you.

 

NRP - The crank will be statically at the bottom of the clearance for measurement purposes, yet will normally spin or oprtate with the clearance centered. That difference (which may be not worth worrying about) could be figured in. I would guess it is about .0005 inch difference.

 

NRP - I took apart my R90/6 this weekend & found the splines this time were still greasy and undamaged. It seems like it must have been assembled at the factory (1975) with no lube. I had used Molypaste G.