another spline failure

[billpierce]

I already posted this over at pelican but thought it might benefit some here as well. Thought I would do some maintenance being down for winter. 39,000 miles.

The splines were bone dry and not far from complete failure. I ordered the parts today and will do the work myself. About $850 in parts and the better part of two days should do it.

My son who is a machinist came over and did some measuring. Our conclusion is a defect in the shaft material (too soft) and or too much clearance in the original parts shaft to clutch disc. This would cause a jolting and rapid ware. it's hard to say at this point. Misalignment does not seem to be the issue in this case. I can't speak for others.

My plan is to have the local metallurgic company Rockwell test both the new and old shaft. I may have it nitrated as this has helped in some industrial applications where I have had problems with this kind of failure. Needless to say it will be going back together lubed and checked every 30-40K miles in the future.

http://i218.photobucket.com/albums/cc164/billpierces/IMG_0271.jpg

http://i218.photobucket.com/albums/cc164/billpierces/IMG_0281.jpg

http://i218.photobucket.com/albums/cc164/billpierces/IMG_0272.jpg

[k75s]

Oh great, I can't wait to do mine now :banghead

What year is your R1100S?

What I can not figure out is where exactly is the wear on the shaft. Looking at the male end in your photo, there does not appear to be wear on the very end of the shaft. It looks symmetrical.

But it looks like there is wear in the middle of the shaft ? Not sure I have seen the wear only in the middle of the shaft before. But I have not looked at as many as some folks on this website.

Not sure how Rockwell will be able to test metal hardness on the new shaft without causing some damage to it. Most metal hardness tests require an impact device and then measuring the depth of indentation into the metal.

[awagnon]

Sorry to hear about your splines. They looked about like mine did at 43,000 miles. I did the replacement myself, but didn't check alignment or hardness. I'm going to pull the tranny in a couple of weeks and see what 15,000 miles have done to the new shaft and clutch. I'm anxious to hear what you learn with the hardness testing.

[billpierce]

Oh great, I can't wait to do mine now :banghead

What year is your R1100S?

What I can not figure out is where exactly is the wear on the shaft. Looking at the male end in your photo, there does not appear to be wear on the very end of the shaft. It looks symmetrical.

But it looks like there is wear in the middle of the shaft ? Not sure I have seen the wear only in the middle of the shaft before. But I have not looked at as many as some folks on this website.

Not sure how Rockwell will be able to test metal hardness on the new shaft without causing some damage to it. Most metal hardness tests require an impact device and then measuring the depth of indentation into the metal.

The bike is a 2001 mfg date 7/00. The whole shaft does not protrude into the hub, thus the ware pattern.

I made a trip to Inland Northwest Metallurgical Services with the old and new transmission input shafts for hardness testing. Nice folks, because they knew me that did the testing for free.

We tested on the c scale.

The results are, old shaft 56, new 58. That's hard folks. The guy helping me said that if we went up a scale or two it could even be a 60. We don't know the process they used to harden the shaft so the c scale is a conservative estimate. The most I could get by nitrating would be in the low 60s. Hardly worth the time and money.

Testing puts a very small mark on the shaft. Not an impact just pressure. Obviously you don't want to do it right on the new splines or directly under a seal or where a gear or bushing has to move. But it doesn't damage the new shaft. This is common practice.

Conclusion, it ain't the shaft that's the problem.

Next, I checked the run out on the flywheel, and also made sure that all the holes were in a perfect circle. This test eliminates the possibility of of the assembly running elliptically. or wallowing as it turns. Run out is between +.001 and - .003, not bad at all.

http://i218.photobucket.com/albums/cc164/billpierces/IMG_0299.jpg

Next step is to get my kid to turn a shaft that will fit snugly into the pilot hole in the crankshaft and will extend in to the transmission housing. Then I can put a dial indicator on that shaft and make sure that the bore if the input shaft bearing is lined up with the crank center.

Similar to what this gentleman did.

http://i218.photobucket.com/albums/cc164/billpierces/crankrunout3.jpg

So far the only thing I can see that may have lead to this failure are the badly warped clutch housing and pressure plate. There is also quite a bit of play between the input shaft and clutch splines on the new parts.

More to come.

[lostboy]

This wear pattern looks like an alignment problem to me. This has caused more spline failures on BMW than any other factor. The trick is measuring the misalignment. I don't think anyone has found a way to measure it accuately on R1100/1150, and there's no guarantee that BMW will get it right on the next case.

Several years ago I got involved with a K75S that was going through input shafts every 10,00 miles. These are fairly easy to measure, and it took three new intermediate flanges before I got an acceptable one.

[billpierce]

The trick is measuring the misalignment. I don't think anyone has found a way to measure it accuately on R1100/1150

Can you tell me whats wrong with the method I have described above?:dunno

[lostboy]

The problem is that the input shafts don't seem to be perpendicular to the gear box mounting surface. Measuring off the input shaft has proven difficult to say the least. It woulld be interesting to make a dummy shaft that would accept a dial guage so you could reliably read to the mounting suface.

[billpierce]

The problem is that the input shafts don't seem to be perpendicular to the gear box mounting surface. Measuring off the input shaft has proven difficult to say the least. It woulld be interesting to make a dummy shaft that would accept a dial guage so you could reliably read to the mounting suface.

The input shaft is below the mounting surface. So a good straight edge and a depth gauge should do the trick. I will check that too,after the gearbox is back together.

[nrpetersen]

Can you tell me whats wrong with the method I have described above?:dunno That's the way to do it. You should mount the dial indicator to the crankshaft/flywheel and read the transmission input shaft front bearing housing bore. It will require a special rig to do this because of the tight quarters, but since you are only looking for run out (TIR) it is only necessary that the fixturing holding the indicator body be stiff. I suggest simply making it from wood and temporarily epoxy it to the flywheel. Others have made a steel rig & you could look around to see if you could borrow that. The run out should be less than ~ .003. If it is more, you will have to monkey around with the alignment pins between the engine and the clutch housing.

If you can't borrow one, the dial indicator is available from Harbor Freight or Phase II as I recall for like $25.

It would also be good to have when you have to redo the crown bearing in your FD. ;)

[Boxwrench]

you guys are scaring this newbie owner.
My Rt has only 15k on it and have me wondering and worrying.
what signs are there for this problem?
can't afford too many expensive fixes till I find a job again.
Good thing I have my reliable C50 as back up.

[billpierce]

you guys are scaring this newbie owner.
My Rt has only 15k on it and have me wondering and worrying.
what signs are there for this problem?

The only way you can tell is to dissemble and inspect. Or wait for the failure. I choose option 1. I love this bike so I am going to take the effort to make sure it is right. Thank goodness my boy is a machinist :thumb

It would be helpful (too say the least) is BMW would at least tell is if the defect is with the engine or transmission alignment. But then they would be admitting they made a mistake. :scratch

[billpierce]

The boy Tony, made me a shaft today that we could use to measure alignment of the crank center to trans input shaft center. The material he used is SS rod turned ground and polished, for perfect roundness.

The rod is a press fit into the end of the crankshaft. hole size is .2395 "

http://i218.photobucket.com/albums/cc164/billpierces/IMG_0311.jpg

We then bolted the front case of the trans to the engine with the rod extending past the input shaft. The bore gauge was then installed on the shaft, instrument reads in 10ths.

http://i218.photobucket.com/albums/cc164/billpierces/IMG_0307.jpg

http://i218.photobucket.com/albums/cc164/billpierces/IMG_0309.jpg

http://i218.photobucket.com/albums/cc164/billpierces/IMG_0310.jpg


We checked the alignment by rotating the engine. Results are that the case is offset to the engine by .0145"
We check and triple checked and turned the shaft in the crank bore as well. There is no doubt. It sits high and to the right.

So the next step, Tony will make two offset (by .006-007") alignment dowels to replace the ones in the engine case.

We feel this is the best solution because we don't know if it's the engine case or the transmission case that is wrong. Buying a new trans front half at $350 or buying a good used transmission may net the same results.

Here is a shot of Tony measuring up the dowel pins.

http://i218.photobucket.com/albums/cc164/billpierces/IMG_0312.jpg

Stay tuned more to come

[nrpetersen]

The only way you can tell is to dissemble and inspect. Or wait for the failure. I choose option 1. I love this bike so I am going to take the effort to make sure it is right. Thank goodness my boy is a machinist :thumb
This happens to only a (small?) fraction of these bikes. You can remove the starter and look critically inside the clutch housing to see if there is a lot of wear debris and fretting corrosion around the spline. Look at the pictures by the OP at the beginning of this thread to "calibrate" what you find. My 2000 R1100RT was whistle clean in this area at 20,000 miles so I'm not going to sweat it.

Obviously the spline should be relubricated on any teardown, but I don't think there is any point to arbitrarily disassemble a bike with a clean clutch housing.

.014 is a lot of run out for a spline. It has been dragging the clutch disk around the flywheel face quite a bit then. Too bad there isn't a way to "float" it into concentricity by simply leaving out the alignment pins.

[19991100RT]

What a great bunch of post, info and pics. Thanks Guys. BUT, now you really have me scared. After riding mostly hondas for the past 35 years yet drouling over the BMW's that passed me time and time again mile after mile I went out and just bought myself a 1999 1100RT. Now i'm thinking cripes, here we go ! How worried should I be? She looks brand new, has a 3 month warentee and was "lady driven" if that means anything. But does have 90,000 Km's on it. The Dealer told me that's nothing at all. HAve I bought a "too old" bike for my own good?

[41107]

That was a very fine write-up with pics to really show how it's done.Very nice job.But to those that are concerned about this failure i strongly believe that it has to be a very small percentage.I actually believe that bmw has a problem in pinning it down to production numbers.Certainly a bmw quality control issue that they should address and be held liable for.In particular with low mileage bikes that this happens to.I pulled my tranny 98 rt for input shaft bearings at 55 k but the splines while dry and in need for moly looked perfect.

[canyoncarver]

The absolute best and most explored post I have ever seen! You guys rock, especially your son!

[billpierce]

I finally received the rest of the parts I needed to complete this job. I replaced the rear main seals while I was in there.

If you are interested I posted details and photos here at the pelican parts fourm

http://forums.pelicanparts.com/showthread.php?t=452179

[PGlaves]

I am both amazed and amused every time I hear that these splines really don't need to be lubed, that it is a small number, that BMW doesn't know which bikes, etc.

Every dry clutch BMW built since 1969 has needed periodic clutch spline lubrication - at least up to the Hexheads with which I have no experience.

There is some variability in how soon they get dry and how fast they wear. I've seen dry ones at 30K miles, and dry ones at 70K miles, and most 10K intervals in between. Folks ignore it at their peril.

But lots of folks do ignore it. But they often sell the bike before it fails, and then the next poor schmuck pays the price.

[OUTBACKUFO]

I am both amazed and amused every time I hear that these splines really don't need to be lubed, that it is a small number, that BMW doesn't know which bikes, etc.

Every dry clutch BMW built since 1969 has needed periodic clutch spline lubrication - at least up to the Hexheads with which I have no experience.

There is some variability in how soon they get dry and how fast they wear. I've seen dry ones at 30K miles, and dry ones at 70K miles, and most 10K intervals in between. Folks ignore it at their peril.

But lots of folks do ignore it. But they often sell the bike before it fails, and then the next poor schmuck pays the price.

??? Mr Glaves...

i had mine original go dry at 50K and my 02 1150GS was only two years old... after new clutch assembly and input shaft i have been going ever since... the cycle has 159K on it and been through all weather conditions and riding conditions... though i tend not to clutch/ shift much when off road (normally only in1 and 2nd gears)... i was told by a bmw mechanic since i ride alot more than most it is less likely to dry out... is this a valid point??? also i have in the last 100K have had no shifting problems at all... no sticky or mystery inbetween gears... i feel it is time to lube it or should i wait if i am having no issues.

How much of a job is it to pull off the back of the Trans to lube theinput ?

thanks
OU

[billpierce]

I am both amazed and amused every time I hear that these splines really don't need to be lubed, that it is a small number, that BMW doesn't know which bikes, etc.

Paul, I'm on the 2 year plan. Every other winter it comes apart, lubed with honda moly 60, and reassembled. It's not a big deal, just takes a little time. In my case it was lack of lube and misalignment.

I would love to see you do an article on this (spline lube) in the magazine.

[dbrick]

So the next step, Tony will make two offset (by .006-007") alignment dowels to replace the ones in the engine case.

How will you index the new alignment dowels, to ensure that they are registered correctly and move the transmission case in the right direction?

[billpierce]

How will you index the new alignment dowels, to ensure that they are registered correctly and move the transmission case in the right direction?

By the same method we used the first time. I put the dial bore indicatior back on and checked it. We ended up making several sets of off set dowels before it was finially right. In the end we found that both pins were not off by the same amount.

[Vinton]

I have been a Machinist and Instrument Make for 30 years and your alignment readings could be off. You need to test your indicator setup for sag and compensate for it. You might be OK with what you have done but I would have verified the sag first. Believe me it exist and can give you false readings and you can add to you mis alignment.


Determination of Indicator Sag
http://www.mpta.org/MPTA%20C2C-2002.pdf

When dial bore gages are mounted on extension bars to measure distant objects, the weight of
the gage can cause the mounting beam to bend. This bending causes the gage to give different
readings depending on its orientation in space. To check the amount of sag, mount the gage onto a rigid pipe. Position the dial gage the same distance from the fixed end as will be measured on the equipment. Set the dial gage to the zero
at the 12:00 position. Rotate the pipe and gage assembly to the 6:00 position and read the gage. This is the indicator sag for that position along the beam. If you move the gage closer or farther from the beams mounting point,
the amount of sag will change and will have to be recalculated for the new distance.
If the gage set up is moved to the equipment, any reading taken at the 6:00 position would have to be corrected for the indicator sag that was calculated on the rigid pipe.

[billpierce]

I have been a Machinist and Instrument Make for 30 years and your alignment readings could be off. You need to test your indicator setup for sag and compensate for it.

If you would have read the entire post at pelican you would have seen that we did that.

[rinty]

Interesting thread, Bill.

Would this technique be something that a BMW tech with average skills could do, or would it require bringing in a machinist?

[billpierce]

Interesting thread, Bill.

Would this technique be something that a BMW tech with average skills could do, or would it require bringing in a machinist?

You would need a machinist. I could not have done it without Tony. I do think that anyone who has this kind of failure needs to check alignment.

As Paul said lack of lube is an issue. However, just replacing the bad parts even with lube on a bike with misalignment will cause future problems.

[nrpetersen]

Interesting thread, Bill.

Would this technique be something that a BMW tech with average skills could do, or would it require bringing in a machinist?It doesn't have to be a machinist, but it has to be someone who recognizes the sag possibility of a cantilevered indicator, and jury rigs an appropriate fixture to support that dial indicator rigidly with respect to the flywheel.

Personally I'd cobble a wood plug etc to support the indicator, & temporarily epoxy the whole mess to the flywheel.

It is not a precise measurement, but it is one that requires rigidity. Universal machinist's fixtures won't do the job, as they deflect too much (unless you tip the bike on its end such that the crankshaft is vertical!).;)

[billpierce]

Universal machinist's fixtures won't do the job, as they deflect too much (unless you tip the bike on its end such that the crankshaft is vertical!).;)

Try defecating a 5 inch long piece of 1/4" drill rod, with 40 grams of weight. :stick

[ragtoplvr]

In a previous job I did strain gauge load cells, used to weigh things

I could squeeze a one inch square piece of A2 tool steel with my fingers and it got shorter and fatter. Enough that the gauge glued to it would change resistance.

The amount of defection on a drill rod is large with very small loads.

Rod

[JimMoore]

you guys are scaring this newbie owner.
My Rt has only 15k on it and have me wondering and worrying.
what signs are there for this problem?
can't afford too many expensive fixes till I find a job again.
Good thing I have my reliable C50 as back up.
Hey Boxwrench (and the guy with the '99),

These failures are inherent to the six-speed transmission. Your bike has a five-speed tranny (M97 to be exact). It'll last forever. Don't give it another thought.:dance

[abe456]

I just ripped into my 99 RT @ 20K for the same problem. I've only had the bike 2 months, and here I am with bike parts all over the garage, tranny in the back of my truck for delivery to Rubber Chicken. THe PO hid the problem by filling the trannny with 120 wt gear oil. Now I cant reach the SOB by phone or email, no surprise there.

Splines showed some wear, but not like your pic. Still, tho, you gotta check this out on these RTs or pay even more later. This is my 2nd RT, other one had even worse probs with spline wear.

I'm waiting to hear back from Tom Cutter as to whether or not I should pursue bore alignment. I prolly should anyway.

[PGlaves]

Hey Boxwrench (and the guy with the '99),

These failures are inherent to the six-speed transmission. Your bike has a five-speed tranny (M97 to be exact). It'll last forever. Don't give it another thought.:dance

Go back and re-read my post #18. It contains my experiences with the M93, M94, and M97 five speeds.

[Acejones]

What is pelican and where is it ?

[JimMoore]

Go back and re-read my post #18. It contains my experiences with the M93, M94, and M97 five speeds.
Thanks for the response. You're a fountain of information and a huge asset to the BMW community. Folks, if Paul G says one thing and I say something else, do what Paul G says.

That being said, in my years of scouring the internet for information on these bikes, Ive seen dozens (hundreds?) of cases of clutch spline issues with six speed gearboxes. I may have seen one or two with the five-speeds. Based soley on that, if I had a five-speed tranny I'd be pretty confident my clutch splines were gonna last forever, lubed or not. If I had a six-speed tranny, I'd be very concerned about it, especially if I had less than 40K miles.

[JimMoore]

What is pelican and where is it ?

http://forums.pelicanparts.com/forumdisplay.php?s=&forumid=24