Final Drive
1 - 20 of 21 Posts
Joined
·
637 Posts
There may be some claiming to do so, but I'd expect the cascading damage of a bearing crunching itself to the point of making noise, would ruin any indicators that could be used for failure analysis.
It would maybe be more valuable to do precautionary disassemblies and preload measurement verifications of units that have performed OK for many miles. Not only should the operating bearing preload dimension be determined, but also the shaft and bore diameters too since they might indicate reason(s) for these failures.
The crown bearing has to operate in a preloaded state to give the gearset a precise rotating axis for noise purposes, but the effective contact line in the crown bearing races depends greatly on the supporting shaft diameter. That's why I am suspicious of not only preload errors, but also of bearing installation dimensions.
Time for some bluing checks...........
Just an opinion of a retired MechEngr.
It would maybe be more valuable to do precautionary disassemblies and preload measurement verifications of units that have performed OK for many miles. Not only should the operating bearing preload dimension be determined, but also the shaft and bore diameters too since they might indicate reason(s) for these failures.
The crown bearing has to operate in a preloaded state to give the gearset a precise rotating axis for noise purposes, but the effective contact line in the crown bearing races depends greatly on the supporting shaft diameter. That's why I am suspicious of not only preload errors, but also of bearing installation dimensions.
Time for some bluing checks...........
Just an opinion of a retired MechEngr.
Joined
·
637 Posts
Unfortunately that's all I have myself is conjecture, but that's my latest thinking on how the problem should be attacked.arwadowd said:
The design of a deep groove ball bearing (the crown bearing) being opposed by a tapered roller bearing in a preloaded application is very unusual to say the least. Frankly I have only seen pictures of failures, and they have only been so gross that it is impossible to assign a cause, but that is my current speculation.
What year and model BMW? Obviously I'm interested.
NRPetersen
Joined
·
1,134 Posts
You want to chat with CharlieVT...aka Curtis. He is our in house FD expert. He has torn apart a number of them, rebuilt them, and analyzed them. He has even created a video showing how to measure pre-load.arwadowd said:
Keep an eye out for him!
But, I suspect once he sees your post, he will stop in and say Hello!
Joined
·
1,178 Posts
There have been at least two write ups about this in BMW ON magazine by Paul Glaves & Anton ?. There thoughts pretty much mirror what Niel has stated. The problem is more than likely 99% factory assembly or machining error. The majority are believed to be incorrectly shimmed which then causes the bearing failure. Best guess is the preload is to high and this causes the ball bearings to make small indentations in the race. This may also be escalated when shipping by compressing the rear suspension which puts even more load on the bearing in a static condition.
Other failure was the shaft diameter that the bearing presses on was incorrect. This has only been reported on a few failed drives.
Roy
Other failure was the shaft diameter that the bearing presses on was incorrect. This has only been reported on a few failed drives.
Roy
Joined
·
637 Posts
Let's supposed the factory recommended preload amount is followed. As I recall, they call for about .001 to .002 axial preload. There is no spec for the shaft diameter as it is not a wear item.
This design is certainly unique and without precedent in bearing applications and probably marginal in its load capacity, but the range of life experienced in field service varies more than I would expect for a simple marginal design. Somehow there appears to be a greater preload variation occurring on assembly despite years of failure history and service reports. Of course many ( but not all) field fixes don't follow the preload specs on bearing replacement. There isn't much we can do about these failures if they reoccur.
However, the factory isn't getting it right yet for whatever reason. They should be consistent. I'm not ready to simply say they have bad quality. Germans just don't tolerate poor preload procedures (usually!).
The outer race floats in the housing at near zero clearance between the outer race OD and the housing bore..
My suggestion - The inner race of the crown bearing is shrunk onto the shaft. This makes the diameter of the inner race groove as-installed depend on the shaft diameter due to race dilation. It is just possible that the specified dimensional axial preload could generate substantially greater ball preloads (forces) if the shaft is oversize or maybe even just at the high end of the original factory manufacturing diametral tolerance - and vice vs.
Bearing races and balls are made very accurately. The installation dimensions probably vary more than any other element in this design. I contend that we have got to also start recording the as-manufactured shaft dimensions if we are ever going to get to the bottom of this problem.
Might BMW ever let us peak at their shaft dimension specs? It would certainly help.
This design is certainly unique and without precedent in bearing applications and probably marginal in its load capacity, but the range of life experienced in field service varies more than I would expect for a simple marginal design. Somehow there appears to be a greater preload variation occurring on assembly despite years of failure history and service reports. Of course many ( but not all) field fixes don't follow the preload specs on bearing replacement. There isn't much we can do about these failures if they reoccur.
However, the factory isn't getting it right yet for whatever reason. They should be consistent. I'm not ready to simply say they have bad quality. Germans just don't tolerate poor preload procedures (usually!).
The outer race floats in the housing at near zero clearance between the outer race OD and the housing bore..
My suggestion - The inner race of the crown bearing is shrunk onto the shaft. This makes the diameter of the inner race groove as-installed depend on the shaft diameter due to race dilation. It is just possible that the specified dimensional axial preload could generate substantially greater ball preloads (forces) if the shaft is oversize or maybe even just at the high end of the original factory manufacturing diametral tolerance - and vice vs.
Bearing races and balls are made very accurately. The installation dimensions probably vary more than any other element in this design. I contend that we have got to also start recording the as-manufactured shaft dimensions if we are ever going to get to the bottom of this problem.
Might BMW ever let us peak at their shaft dimension specs? It would certainly help.
Joined
·
12,165 Posts
Several years ago I sent my totally failed bearing along with another members bearing which was replaced before it exhibited any real problem to SKF for failure analysis. I posted the results of their report about three times over the years.
Basically, the bearing engineer at SKF said that the major problem most likely stemmed from improper procedures at assembly, and to a much lesser extent a possible machining tolerance issue in that the inner bore showed fretting, which indicated a slightly TOO LOOSE fit of the inner race on the wheel hub.
The major problem was spalling of the races, ON BALL PITCH, which indicated impact/overpressure at assembly causing Brinelling of the races, which over time progresses to spalling. The balls constantly rolling over the spalled areas causes shock loads to the retainer, which then causes it to fail as a result of, not the cause of the bearing failure.
So, I have been totally convinced that the real issue was BMW's assembly procedures, not a design defect.
Remember, this final drive design is several decades old, and the failures did not start showing up in any noticeable quantity until about the year 2000. Most likely due to the larger weight of the LT exacerbating the problem and bringing the failures to the forefront.
Seemingly, LTs after 2005 show a greatly reduced failure rate.
Now, the NEW DESIGN final drives are failing at a pretty high rate, but a different type of failure.
Basically, the bearing engineer at SKF said that the major problem most likely stemmed from improper procedures at assembly, and to a much lesser extent a possible machining tolerance issue in that the inner bore showed fretting, which indicated a slightly TOO LOOSE fit of the inner race on the wheel hub.
The major problem was spalling of the races, ON BALL PITCH, which indicated impact/overpressure at assembly causing Brinelling of the races, which over time progresses to spalling. The balls constantly rolling over the spalled areas causes shock loads to the retainer, which then causes it to fail as a result of, not the cause of the bearing failure.
So, I have been totally convinced that the real issue was BMW's assembly procedures, not a design defect.
Remember, this final drive design is several decades old, and the failures did not start showing up in any noticeable quantity until about the year 2000. Most likely due to the larger weight of the LT exacerbating the problem and bringing the failures to the forefront.
Seemingly, LTs after 2005 show a greatly reduced failure rate.
Now, the NEW DESIGN final drives are failing at a pretty high rate, but a different type of failure.
The FD is from a 2005 R1200GS. It failed at 45000 miles.
Is there no centralized repository for this data? I understand that NP and David really know their stuff but again, I was hoping one of you wizards had a garage full of these things and reams of data that might someday be placed, thoughtfully and ever so gently, right up a certain blue and white roundel.
I'm not really that pissed. I still got a good deal on the bike.
Jim
Is there no centralized repository for this data? I understand that NP and David really know their stuff but again, I was hoping one of you wizards had a garage full of these things and reams of data that might someday be placed, thoughtfully and ever so gently, right up a certain blue and white roundel.
I'm not really that pissed. I still got a good deal on the bike.
Jim
Joined
·
16,584 Posts
:histerica :histerica :histerica So, this would be your first BMW, then? :histerica :histerica :histericaniel_petersen said:
"There is no problem, there never was a problem, and therefore you must be riding the bike wrong. But if you dump that obsolete bike and upgrade to the latest-greatest new BMW, then this non-problem will be completely eliminated."
Until that one fails as well . . .
Joined
·
16,584 Posts
I wish it were as funny as it sounds. I've had 5 BMWs, four final drive failures, and four failed clutches (would have been five, but one bike was totaled before the slipping clutch got replaced).niel_petersen said:
When it works the bikes are awesome, but when it fails . . .
Joined
·
684 Posts
Shaft dimension specs are classified under a need-to-know basis only. At home, there is a strict non-disclosure agreement between me and my wife.niel_petersen said:
Joined
·
4,404 Posts
If it is a drive off a GS I believe it is an "Evo" drive, not the same as the KLT final drive. Drive design is very different from the KLT final drive and failure mode very different.arwadowd said:
Forensic analysis of a failed drive will tell you what failed (most often crown wheel bearing in the case of KLT final drives) but will not tell you why it failed.
If it were a KLT final drive I'd offer to rebuilt it for you, since it is a GS drive, I know nothing about 'em and have nothing to offer.
Joined
·
461 Posts
Has anyone measured the shaft size ?
Here is a chart of shaft sizes and the bearing fit class .
There seems to be a wide range from falling on to way tight . .0033" shrink fit, that is about double what I have ever seen.
NTN Part Number Bore O.D. Width
6917C3 3.3465 in. (85 4.7244 in. (120 0.7087 in. (18
mm) mm) mm)
(This is a PDF file of the "Bearing Fits" section of the NTN 2200 catalog.
See section 7.4 - "Recommended Fits" - for fit selection guidelines.)
Shaft fits High Low
| Bearing Class 0 3.3465 in. (85 mm) 3.3457 in. (84.98 mm)
|
g5 3.3460 in. (84.988 mm) 3.3454 in. (84.973 mm)
g6 3.3460 in. (84.988 mm) 3.3451 in. (84.966 mm)
h5 3.3465 in. (85 mm) 3.3459 in. (84.985 mm)
h6 3.3465 in. (85 mm) 3.3456 in. (84.978 mm)
h7 3.3465 in. (85 mm) 3.3451 in. (84.965 mm)
j5 3.3467 in. (85.006 mm) 3.3461 in. (84.991 mm)
js5 3.3468 in. (85.0075 mm) 3.3462 in. (84.9925 mm)
j6 3.3470 in. (85.013 mm) 3.3461 in. (84.991 mm)
js6 3.3469 in. (85.011 mm) 3.3460 in. (84.989 mm)
j7 3.3472 in. (85.02 mm) 3.3459 in. (84.985 mm)
k5 3.3472 in. (85.018 mm) 3.3466 in. (85.003 mm)
k6 3.3474 in. (85.025 mm) 3.3466 in. (85.003 mm)
m5 3.3476 in. (85.028 mm) 3.3470 in. (85.013 mm)
m6 3.3478 in. (85.035 mm) 3.3470 in. (85.013 mm)
n6 3.3482 in. (85.045 mm) 3.3474 in. (85.023 mm)
p6 3.3488 in. (85.059 mm) 3.3479 in. (85.037 mm)
r6 3.3493 in. (85.073 mm) 3.3485 in. (85.051 mm)
r7 3.3498 in. (85.086 mm) 3.3485 in. (85.051 mm)
Bob G
Here is a chart of shaft sizes and the bearing fit class .
There seems to be a wide range from falling on to way tight . .0033" shrink fit, that is about double what I have ever seen.
NTN Part Number Bore O.D. Width
6917C3 3.3465 in. (85 4.7244 in. (120 0.7087 in. (18
mm) mm) mm)
(This is a PDF file of the "Bearing Fits" section of the NTN 2200 catalog.
See section 7.4 - "Recommended Fits" - for fit selection guidelines.)
Shaft fits High Low
| Bearing Class 0 3.3465 in. (85 mm) 3.3457 in. (84.98 mm)
|
g5 3.3460 in. (84.988 mm) 3.3454 in. (84.973 mm)
g6 3.3460 in. (84.988 mm) 3.3451 in. (84.966 mm)
h5 3.3465 in. (85 mm) 3.3459 in. (84.985 mm)
h6 3.3465 in. (85 mm) 3.3456 in. (84.978 mm)
h7 3.3465 in. (85 mm) 3.3451 in. (84.965 mm)
j5 3.3467 in. (85.006 mm) 3.3461 in. (84.991 mm)
js5 3.3468 in. (85.0075 mm) 3.3462 in. (84.9925 mm)
j6 3.3470 in. (85.013 mm) 3.3461 in. (84.991 mm)
js6 3.3469 in. (85.011 mm) 3.3460 in. (84.989 mm)
j7 3.3472 in. (85.02 mm) 3.3459 in. (84.985 mm)
k5 3.3472 in. (85.018 mm) 3.3466 in. (85.003 mm)
k6 3.3474 in. (85.025 mm) 3.3466 in. (85.003 mm)
m5 3.3476 in. (85.028 mm) 3.3470 in. (85.013 mm)
m6 3.3478 in. (85.035 mm) 3.3470 in. (85.013 mm)
n6 3.3482 in. (85.045 mm) 3.3474 in. (85.023 mm)
p6 3.3488 in. (85.059 mm) 3.3479 in. (85.037 mm)
r6 3.3493 in. (85.073 mm) 3.3485 in. (85.051 mm)
r7 3.3498 in. (85.086 mm) 3.3485 in. (85.051 mm)
Bob G
Joined
·
1,345 Posts
I've got 9,000+ miles on my rebuild. It's still smooth so far. I believe mine was in the process of failing long before I felt it in my pegs. I remember how smooth and light the LT felt out on the road. It seemed like I lost that gradually. After rebuilding it and riding it again, I KNEW I had lost it. I got it back. It now dances on the road at speed like it used to. I hope I can keep it like that.............................
1 - 20 of 21 Posts
Join the discussion
Join now to ask and comment!