[CharlieVT]

I recently completed another failed final drive rebuild. Info posted here for those who may be interested.
Two basic points: First, the preload measurement technique originally suggested by Dman gives results very close but not exactly the same as the technique in the BMW service manual. (Since I used different measuring tools in these two techniques, the differences could be measurement error).
Secondly, and this is more speculative, is that failed drives that I have seen where the crown bearing failed, where I have been able to glean some information, had too much preload.
This final drive had the classic crown wheel bearing failure. Failure was of the retainer, attached pictures show the broken retainer. Both the innner and outer races were spalled, and it seems to me that the spalling of the races must have preceeded the failure of the retainer.
There was evidence that this drive had been rebuilt before. The bearing seat in the cover was scored, evidence of a spun bearing race, which happens when the drive is ridden for some distance after the drive begins to show signs of failure.
The current bearing did not show signs of having spun, and the history given by the rider indicated that the bike was not ridden far after evidence of failure. These findings made me think this was a twice failed drive. One failure with the current owner, one with a prior owner which the current owner had no knowledge of.

I measured for preload of a new bearing using two techniques. The BMW method requires a BMW special tool, and a slide micrometer. I developed a technique using the final drive cover in place of the special tool. Many repeated measurements give me confidence in the accuracy of the measurements.
The other method I used, originally introduced by Dman uses a dial indicator; this method is documented in my final drive rebuild video.
The two methods, both repeated several times on the current rebuild, did not give exactly the same results but quite close. (Something I have been curious about, and a reason why I was interested in doing another drive.)
Using both techniques, I arrived at a preload shim thickness of 0.20mm.

The original shimming in the drive was 0.65mm. Much thicker than my calculations for the new bearing. That doesn't tell us much since it is not possible to measure needed preload for the old, damaged bearing. But if you imagine that the bearings are manufactured pretty close to the same specs, and the the machining differences are in the final drive components: cover, housing, crownwheel assembly, tapered roller bearing and it's shim, then it would suggest that the drive may have been excessively preloaded by about 0.65mm-0.20mm= 0.45mm. (I have previously rebuild a drive that had a bearing replacement following failure without preload check, and the new bearing which was shimed using the original shims was over-shimed by 0.50mm by my calculations). This seems to me to be quite a lot of excess preload. This is fairly speculative, but both the old failed bearing and the new one are 19 ball, *** bearings made in Germany, and no other components in the final drive have been changed.

 

[meese]

I really wish you had some experience with the latest generation final drives. I have a destroyed GT FD that I'd love to send you for analysis . . .

 

[Voyager]

I measured for preload of a new bearing using two techniques. The BMW method requires a BMW special tool, and a slide micrometer. I developed a technique using the final drive cover in place of the special tool. Many repeated measurements give me confidence in the accuracy of the measurements.

Is there a possibility that you will make a new video showing your new technique?

 

[andoni]

Thank you very much for sharing this useful information, Charlie. At this stage it is a bit beyondmy abilities, but if my FD fails, at least I will know these important issues so that I can share with a mechanic.
Regards, Andoni

 

[CharlieVT]

I really wish you had some experience with the latest generation final drives. I have a destroyed GT FD that I'd love to send you for analysis . . .

Well, I guess I'd better not open that door. Too many projects and no personal motivation since I don't have a bike with an Evo drive. I do find it interesting that BMW has had a run of drive failures that span over time and different models. It furthers my suspicion that the problems are not related to a bad batch of bearings, or a poor design. It makes me think that they have QA problems with the workers in the FD assembly line.

 

[CharlieVT]

Is there a possibility that you will make a new video showing your new technique?

I might get around to that, or maybe just a photo sequence. Not sure when though. Certainly not until the next time I rebuild a drive, whenever that might be.

Here's a brief description of the technique: I recommend reviewing the BMW Service Manual technique to help put this method into perspective.

The BMW special tool used in the Service Manual techinque for measuring crownwheel preload is used to stabilize the crownwheel assembly with the bearing so that the relative height of the outer race of the crown wheel bearing can be measured with respect to the mating surface of the drive housing. Stabilizing the crown wheel assembly in this way is critical because if not stabilized it will wobble during measurement and no accurate measurements can be obtained.

I got the idea of making a special tool to use in place of the BMW one from a rebuild techinque of Airhead transmissions where an old tranny cover is modified to make a tool. (The Airhead tranny uses shims to limit end play of shafts very similiar to the KLT final drive in that the shim is captured behind the cover.)

I had a spare cover from a trashed final drive that I was ready to modify, but after looking at things while thinking about what cuts I was going to make in the old cover, I realized no modifications to the cover were needed. Simply place the cover (preheating the cover is recommended) onto the crownwheel assembly with no seal or shims installed. Rotate the cover so that the bolt holes are misaligned. The bolt holes are not evenly distributed, and a couple of bolt holes can be made to align while others are not aligned. Using a couple of aligned bolt holes opposite each other, bolt the cover down to make sure it is fully seated. The other bolt holes which are not aligned now have the mating surface of the housing at the bottom of the hole; measurements are made through these bolt holes in the cover to the housing mating surface. Now using an accurate reference, the height of the crownwheel bearing outer race can be calculated with respect to the height of the drive housing mating surface. (I used two veneer calipers, you'll sort of need pictures to visualize how I did it, but there are several ways the measurement could be made. The key is stabilizing the bearing.) This is measurement "b" in the Service Manual. Measurement "a" is much simpler to get, just follow the procedure in the service manual. This gives you the two measurements needed to calculated shim thickness according to the service manual.

 

[simoncharles]

Thanks Charlie. It´s an honour to have someone like you posting on this forum.

 

[CharlieVT]

Thanks Charlie. It´s an honour to have someone like you posting on this forum.

You are kind, but don't be confused. I'm just a hacker with a hobby.



"Last year I couldn't even spell "professional". Now I are one."

 

[AlaskaFish]

You may call yourself just a hacker if you wish Curtis, but what you are hacking into is one of this community's weakest links! Sort of our Achilles heal, if you will. And I, like many others on this list, do appreciate all your efforts along with the time it takes you to share them with the rest of us folks.

John

You are kind, but don't be confused. I'm just a hacker with a hobby.



"Last year I couldn't even spell "professional". Now I are one."

 

[meese]

Well, I guess I'd better not open that door.

I understand, Charlie. BTW, do you have a brother that's as curious as you are?

 

[LAF]

You may call yourself just a hacker if you wish Curtis, but what you are hacking into is one of this community's weakest links! Sort of our Achilles heal, if you will. And I, like many others on this list, do appreciate all your efforts along with the time it takes you to share them with the rest of us folks.

John

Here Here!!!!!!

At least you have provide information on getting one shimmed correctly.

And I think your assumption on Q&A is correct also. You and others have shown pics of mode of failures and to me it looks like you are correct. How else could it be so sporadic? And I do believe it is sporadic, too many LT's running around that don't have any problems, and others that succumb to every one of it quirks.

You are far too modest on your achievements.

You, in my eyes anyway, have given hope that we can get one rebuilt, or rebuild our own, and run it in good faith.

I know your life's ambition is to NOT rebuild final drives, but you have shed light on one of this machines biggest quirks.

At least your information gives us more knowledge and a fighting chance.

 

[Alpac]

Charlie,
just wanted to say thanks again for sharing your findings and assumptions on the causes of final drive failures. Since mine failed last fall I became very interested on that subject. Although identifying the reason why a FD failed can only be speculation your arguments always make perfect sense to me. I still have to rebuild my old one using your very well documented video which I will do soon hopefully. I have learnt quite a bit just by reading your posts :thumbup:

 

[c00k1e]

This is VERY interesting (having had 6 failed FDs and bought a brand new one from BMW). I hope to do a tech session at the 1st UK LT do,on the old FD based on your vid. Is this new technique going to be more accurate?
Have you ever done a rebuild on a failed FD from your old technique?

 

[CharlieVT]

This is VERY interesting (having had 6 failed FDs and bought a brand new one from BMW). I hope to do a tech session at the 1st UK LT do,on the old FD based on your vid. Is this new technique going to be more accurate?
Have you ever done a rebuild on a failed FD from your old technique?

Good questions.
Is the "new" technique going to be more accurate? I think that this technique, which basically follows the BMW service manual, is more prone to measurement error due to the number of measurements, how they are taken, and the multiple calculations that must be made. However, I have no basis to say that one technique is going to be "better" than the other as far as being able to properly calculate preload shim thickness. Having done it both ways, I prefer the technique that uses the dial indicator that was introduced by Dman and is shown in my video. But in any future rebuilds, I will use both methods because I am curious and want to continue to compare the results.

Have I ever done a rebuild on a failed FD from my old technique? I don't know of any of my rebuilds that have failed, if that is what you are asking. But I don't have good follow up on them. I hope if one of my rebuilds fails, the owner will notifiy me (as opposed to sending me a letter bomb) . I would really like to look at such a drive. Many (most?) of the drives I have rebuilt are sitting on shelves as "spares" where the owner replaced their drive and sent me the failed one for rebuild to keep as a spare. I have +40K miles on my own rebuilt final drive. That was a preemptive rebuild, not a failed drive, but it was shimmed using the technique in the video. There are a couple (few?) drives I have done which are being ridden and have not failed as far as I know; I have no idea what miles have been put on these drives.

Regarding reports from people like yourself who have experienced multiple failures, I have often queried as to whether the new bearing was just installed using the old shims or if measurements were made to check for proper preload. I am very suspicious that many of these rebuilds were done without checking preload, and if done by a BMW service department, if queried they'll say they did measure for preload, but does anyone really know if they did? [Now ranting here] I frankly doubt them even when it was reported that they claimed to have down the measurements. My guess is they just slapped in a new bearing. In the early days they didn't know it was going to fail again. It wasn't until there were enough failures and repeat failures here in the US that BMW of NA started telling dealerships to replace the whole drive rather than rebuild it when doing warranty work. (I am just guessing that BMW made that decision based on reports that folks with failed drives were getting whole new drives). But even then, no one knew if the "new" drive was assembled any better than the one that failed, and I believe there we several people who got a factory new drive that went on to fail.

Good data is pretty hard to come by with respect to reasons for failure and success rates of rebuilds, that's why there is so much speculation.

 

[c00k1e]

........Regarding reports from people like yourself who have experienced multiple failures, I have often queried as to whether the new bearing was just installed using the old shims or if measurements were made to check for proper preload. I am very suspicious that many of these rebuilds were done without checking preload, and if done by a BMW service department, if queried they'll say they did measure for preload, but does anyone really know if they did?.........

I know for a fact that one of my rebuilds was done with old shims. It was done in less than 24hrs at a non-BMW shop. He ordered the bearings and whacked them in (literally!) - lasted less than 1000 miles (I knew a lot less then than I do now).
I am still of the opinion that pivot bearings can and do have a devastating effect on the FD, even if the FD is perfect. I have seen 2 bikes here that failed the wheel wobble test and the movement was coming from the pivot bearing - would the FD go next?



.

 

[CharlieVT]

I know for a fact that one of my rebuilds was done with old shims. It was done in less than 24hrs at a non-BMW shop. He ordered the bearings and whacked them in (literally!) - lasted less than 1000 miles (I knew a lot less then than I do now).
I am still of the opinion that pivot bearings can and do have a devastating effect on the FD, even if the FD is perfect. I have seen 2 bikes here that failed the wheel wobble test and the movement was coming from the pivot bearing - would the FD go next?

You raise some good points. Some of the professionally trained engineers on this board (I'm not one of 'em) have suggested that the bearing can be damaged by forceful pressing onto the crownwheel hub; hammering a bearing onto the hub would be a bad thing. Who knows if on the BMW assembly line, some worker didn't figure out how to save time by forceful pressing of the bearing onto the hub, setting the stage for future ball and race spalling. This maybe another factor, in addition to improper shimming that has contributed to drive failures.

Many reports of loose pivot bearings have been posted here. All the final drives I have looked at had some visible wear or notching of the pivot bearings. The pivot bearings are clearly a wear point. Having seen a few of 'em with significant notching of the pivot bearing races by the roller pins, I have suggested that when these pivots are found to be loose the bearings should be removed and inspected rather that just re-torqued. Re-torquing is a good quick fix, but I suspect that inspecting the bearing will probably find that the bearing is degrading.

As to whether loose pivot bearings would contribute to final drive failure.... well, I dunno, but it sure can't help. One engineer friend who is knowledgeable about stress vibrations even suggested that a worn tire (think of the howling Metz) could cause damaging stresses over time. These are things I know little about, but wouldn't dismiss completely. I do think however, that considering the normal road and riding conditions that the bike is designed to withstand, that a properly setup final drive would tolerate worn tires, and loose pivot bearings for a long time. My intuition tells me that these factors would not be a primary cause of FD failure.

One poster on this board suggested that I like to "harp" about bearing preload; well I guess I do. That said, I still consider that FD failures probably have multiple contributing factors (we know that not all FD failures are crownwheel bearing failures, but that is apparently the most common failure for the early drives). Since it is the crown wheel bearing that is the failure point, it stands to reason that damage to the bearing during assembly, and over stressing of the bearing by improper preload are two likely causes of failure.

 

[Alpac]

Charlie,
How many FDs have you rebuilt and have you had one of your rebuilt FDs failed again?
Thanks

 

[CharlieVT]

Charlie,
How many FDs have you rebuilt and have you had one of your rebuilt FDs failed again?
Thanks

Somewhere between 5 and 10 I reckon, sort'a lost track.
None that I know of.

 

[11862]

Charlie - What is a "slide micrometer"? Is it a traditional depth mic? If I understand the BMW service manual technique to use different depth mic readings, (and I don't have a copy), your dial indicator scheme is much more likely to get a consistent preload.

I agree the bearing has to be stabilized to get a repeatible axial preload, and that any brinneling from hammering etc could initiate a premature failure.

It sounds like you are seeing some examples of radically over preloaded bearings. Do you have any readings of shaft diameters?

FYI - Some years ago the auto manufacturers had shipping damage where the wheel bearings don't rotate but are radially fatigued. I don't know how they got out of that problem though.

 

[CharlieVT]

Charlie - What is a "slide micrometer"? Is it a traditional depth mic? If I understand the BMW service manual technique to use different depth mic readings, (and I don't have a copy), your dial indicator scheme is much more likely to get a consistent preload.

I agree the bearing has to be stabilized to get a repeatible axial preload, and that any brinneling from hammering etc could initiate a premature failure.

It sounds like you are seeing some examples of radically over preloaded bearings. Do you have any readings of shaft diameters?

FYI - Some years ago the auto manufacturers had shipping damage where the wheel bearings don't rotate but are radially fatigued. I don't know how they got out of that problem though.

Sorry, my terminology is kind of casual and no doubt incorrect. I used the term "slide" when I should have said "venier" which I probably would have spelled "veneer" And I probably should have called the instrument a "caliper" as opposed to "micrometer". However wrong, I have always thought of a caliper (inside or outside) as a device that could record a dimension, but did not have a scale to measue the dimension.

So, whatever you call 'em the attached pic of the instruments I used for FD measurements. The two "slide" micrometers have a rod that comes out the end which I used for depth measurements in the BMW technique. One of them is metric, the other inches.
The dial indicator (I have called it a dial micrometer) I used for Dman's technique.

I have not measued shaft size of the crownwheel bearing seat on the crownwheel assembly hub. I have read suggestions that this might have been oversized, contributing to bearing stress and early failure. All I can tell you is that in every drive I have done, the bearing would just drop onto the hub by gravity when the bearing was heated to approx. 200 degrees F and the hub had been placed in the freezer for 20-30 minutes.