U-Joint and Drive Shaft Discussion

I got to looking at the deal with BMW drive shaft issues we've all seen and did some research.

Now we all know (or should know) that the types of u-joints used in most MC drive shafts are variations on the 100+ year-old creation known as a Cardan Joint (for one name). They are, of course, meant to allow transmission of rotation from an input to an output shaft when the two are at some angle from each other. But they have a well-known weakness, in that, whenever the angle between the input and output shafts is not zero degrees, the two shafts will wind up only transiently be operating at the same RPM. This happens at four positions per rotation. The entire remainder of the rotation one is under varying rotational speed with respect to the other. This is under constant variation, producing constantly varying rotational acceleration (positive or negative) with respect to the other. The forces to produce these variations add stress to the u-joint

For simplification, if we assume that the input shaft from the transmission is at a constant rotational speed, then the driveshaft is constantly varying in speed pretty much always. And both the peak acceleration force and the peak speed differential of the drive shaft are a function of how far off from zero degrees the axes of the two shafts are.

From what I can tell, at 60 MPH, the FJR transmission turns about 3000 RPM. This is 50 RPS, so at that speed the drive shaft would be experiencing a 100 Hz oscillation in its rotational speed, with the amplitude driven by how misaligned the two shafts are at any given instant.

So at what point exactly, are the shafts perfectly aligned? Good question, I suppose, but I think the best answer is pretty much never. But one thing is a pretty safe bet: that the more heavily loaded the bike is, the more likely the peak angle is to be higher. Tweaking preloads would allow you to maybe find a sweet spot where the suspension spends most of its time at or near zero degrees, but I don't see an easy way to measure just where the exact alignment is.

So now the skeptics will ask: How can the drive shaft have varying speeds? Well, remember those rubber dampers in the rear wheel? They act as a damper between the pumpkin output and the wheel itself. And the rear tire has flex in it as well.

So what's the deal with some BMWs? Mostly a shorter driven shaft. A lot (all?) BMWs have the same type of u-joint at the front and rear of the drive shaft. In theory this allows the input shaft of the pumpkin to have the same (constant) velocity as the transmission output shaft, as it cancels out the varying RPM of the drive shaft. But there ain't no free lunch. The drive shaft itself still experiences those oscillations. And here's the deal: The BMW drive shafts on some of these bikes are extremely shorter than, say, the FJRs, as the distance between the two u-joints is much less (half or less?) than the distance between an FJRs rear axle and that front u-joint. So the BMWs likely have much greater variance in that angle, with the result that these additional torsional stresses are far greater than the usual Japanese bike drive shaft configuration.

The deal with BMW is the paralever rear suspension.

It requires not one, but two U-Joints, one on each end of the drive shaft. To accommodate the movement the shaft is splined and in two pieces ... effectively the shaft is free to change its length while moving as one part slides in the other.

If those splines are not lubricated the shaft can seize putting enormous strain on the UJs, at some point one of them will cry "foul".

If it is the driven end, the bike stops with little harm done, if it is the rear end the drive shaft, and what is left of the UJ, spins furiously until the rider stops or the housing breaks ... at which point, the rider stops.

RichDoyle said:

Now we all know (or should know) that the types of u-joints used in most MC drive shafts are variations on the 100+ year-old creation known as a Cardan Joint (for one name).

Click to expand...

Didn't know that. I'm just happy I don't have to lube a chain, and I basically just have to put gas in the thing.

AS I understand the first two posts, the Beemer requires two u-joints. As a benefit of that, the two u-joints help cancel out the oscillation that is caused by a single u-joint when it is not perfectly straight. I get that. I also get that the Paralever requires a varying length drive shaft.

But, I don't really get why a variable length drive shaft requires two U-joints. And, based on how reliable the FJR is, with no noticeable effect of the oscillation caused by the single u-joint, why BMW would design the mechanism with two. Is the shaft shorter because it has a second u-joint? Or is the second u-joint needed because the shaft is short? Or, is the drive shaft at a more severe angle under normal use than the one in the FJR?

Can't answer any of those questions.

It is likely that the BMW system is supremely reliable IF those shaft splines are kept lubed. This is not a problem that is associated with the other well-known final drive failures. It appears they finally fixed that.

Personally, with the potential for the swinging arm casting to beak so catastrophically, I think the failure of the rear UJ is a serious safety matter

don't be like Beemer riders and ANALIZE everything !!!!!!!!!!!

radiummadman said:

don't be like Beemer riders and ANALIZE everything !!!!!!!!!!!

Click to expand...

If it helps .... I changed my coolant by letting the old stuff run out, then filling up with new ... No "blow jobs", promise

We were just discussing this in class.

The students are asked to calculate the front driveshaft yoke speed with an input speed of 1000 rpm and a 30 degree angle.

The answer is the shaft will spin from 700 to 1300 rpm at different points in its rotation.

Geezer said:

AS I understand the first two posts, the Beemer requires two u-joints. As a benefit of that, the two u-joints help cancel out the oscillation that is caused by a single u-joint when it is not perfectly straight. I get that. I also get that the Paralever requires a varying length drive shaft.
But, I don't really get why a variable length drive shaft requires two U-joints. And, based on how reliable the FJR is, with no noticeable effect of the oscillation caused by the single u-joint, why BMW would design the mechanism with two. Is the shaft shorter because it has a second u-joint? Or is the second u-joint needed because the shaft is short? Or, is the drive shaft at a more severe angle under normal use than the one in the FJR?

Click to expand...

The idea of "canceling out" is misleading. It really still exists in the drive shaft itself, though not in the pumpkin input shaft. And this ONLY works if the transmission output and the pumpkin input shaft are kept in parallel. This means the driveshaft length between them must shorten to a minimum when the two shafts are aligned to a maximum when displaced to the max. And there is no doubt the shorter shaft is at a more severe angle at the maximum than the much longer shaft of the FJR.

I don't see where variable length drive shaft needs two u-joints (gotta think about that one) but the two u-joints, if they're to be kept in parallel, suggests the use of a variable-length drive shaft, unless some very strange linkage were employed.

I agree that the repeated in/out movement of the ends of the shaft adds yet more stress on the u-joints, especially if lubrication gets compromised. For whatever reason, the several pictures I've seen of BMW failures manifested at the rear u-joint.

For whatever reason, the several pictures I've seen of BMW failures manifested at the rear u-joint.

Either can go. I've heard of examples of failures at the driven end too. I think we see more pictures of the rear failure because the pictures are awesome

Geezer said:

... And, based on how reliable the FJR is, with no noticeable effect of the oscillation caused by the single u-joint, why BMW would design the mechanism with two. Is the shaft shorter because it has a second u-joint? Or is the second u-joint needed because the shaft is short? Or, is the drive shaft at a more severe angle under normal use than the one in the FJR?

Click to expand...

Ha! That's a knee-slapper! You're really asking why the Germans "over engineer" their products?

Denver_FJR said:

Ha! That's a knee-slapper! You're really asking why the Germans "over engineer" their products?

Click to expand...

An interesting term. Means under tested, under supervised engineers, and too many parts. The entire rear of the bike is crap. Single sided swing arm? Probably a marketing decision to make the bike look different. Never mind there is no valid reason to have it. Para-lever rear end? Para-broken is more like it. Too many parts, too high maintenance, too flawed in design, under supported bearing housing, and on and on. Before defend BMW, go look up the scheduled maintenance interval for wheel bearings. That one exists at all is an admission of the crap quality they use.

BMW has told their dealers that they are focused on buyers with excess disposable income that will upgrade to the next year bike every year and only ride 1000 miles a year, so dealers should be happy to make profits on all the low mileage used bikes. The guy that rides it for hundreds of thousands of miles "doesn't help the BMW bottom line".

Why is this thread on a FJR forum anyway. I realize some FJR owners are also paperweight owners, but we don't discus paperweights, do we?

Trying to keep the joint straight is probably just a thought exercise. Has anyone heard of the FJR u-joint failing?

The yamaha engineers put splines on both ends of the u-joint. I'm guessing they are there are for ease of assembly and being tolerant of what little potential sliding action there is in there.

Eh. Got better things to do with my time.

HYCLE said:

We were just discussing this in class.The students are asked to calculate the front driveshaft yoke speed with an input speed of 1000 rpm and a 30 degree angle.

The answer is the shaft will spin from 700 to 1300 rpm at different points in its rotation.

Click to expand...

I've seen that much, or close to it, on some machinery that turns slowly and it seems OK in that application. This variation in RPM is probably why they started using CV joints in front wheel drive vehicles.

I stuck the discussion here because we do have drive shafts with u-joints and I wanted to point out that the single u-joint is a more reliable implementation. But I also want to emphasize something I said initially that does apply to the FJR. There is vibration and rotational variation induced in the rear end. I've roughly calculated that the possible variation in the angle the u-joint has to handle is around 13 degrees, with probably the max deflection happening with the bike more heavily loaded or driven hard in the twisties.

Here's something else implied by all this movement: We talk about "lubing the u-joint", but there really is no practical way to clean/lubricate the bearing parts within the u-joint. Autos used to routinely have zerk fittings to allow for this, but we don't have that. So when you have the shaft out to lube the splines, it's worth checking that the u-joint itself is in good condition - rotates freely around the spider axes, no significant play in it, no "gritty" feeling when operating.

I have had these types of u-joints fail in cars - one failed catastrophically, thrashing around underneath the car smashing stuff. The other was benign, where the needles disintegrated, but the spider remained trapped in the races.

There have been a very few u-joint failures within the FJR community. But it's very, very rare. at 98k, I swapped out mine for one with 7k, just because I had the opportunity while doing an engine swap. (long story of woe posted long ago) The 98k u-joint showed some surface darkening of dirt and mild corrosion from the elements, but moved freely and showed no indications of binding. And I was running in rally trim, or more heavily loaded, all the time, than most riders do. Pushing 950 lbs with rider.

Not sure where you've been. "We" don't talk about lubing the u-joint. Because, as you noted, you can't. And that's not uncommon among modern cars either. The old Lube, Oil, Filter is long gone. Modern cars don't have very many lube points and today's LOF is just an oil/filter change and eyeball inspection.

Have a stimulating conversation. Just don't see much point in discussing something that doesn't need to be changed or stressed about.

Geezer said:

AS I understand the first two posts, the Beemer requires two u-joints. As a benefit of that, the two u-joints help cancel out the oscillation that is caused by a single u-joint when it is not perfectly straight. I get that. I also get that the Paralever requires a varying length drive shaft.
But, I don't really get why a variable length drive shaft requires two U-joints. And, based on how reliable the FJR is, with no noticeable effect of the oscillation caused by the single u-joint, why BMW would design the mechanism with two. Is the shaft shorter because it has a second u-joint? Or is the second u-joint needed because the shaft is short? Or, is the drive shaft at a more severe angle under normal use than the one in the FJR?

Click to expand...

The BMW uses 2 links to locate the rear diff. These 2 links create a virtual swing arm whose length and center can be ascertained by drawing a center line through each link as seen in the side view. Where these 2 lines cross is the center of the virtual swing arm. By playing with the relative angles of the 2 links you can create a long virtual swing arm that is resistant to jacking and you can move the virtual center where ever to mitigate lifting under braking and other unwanted movements caused by various dynamic forces. The trade off is the rear diff is not pointed at the output end and therefore needs a second u-joint to allow for this. It also means that the front ujoint plunges back and forth on the front splines.

64Y80 said:

Geezer said:

AS I understand the first two posts, the Beemer requires two u-joints. As a benefit of that, the two u-joints help cancel out the oscillation that is caused by a single u-joint when it is not perfectly straight. I get that. I also get that the Paralever requires a varying length drive shaft.
But, I don't really get why a variable length drive shaft requires two U-joints. And, based on how reliable the FJR is, with no noticeable effect of the oscillation caused by the single u-joint, why BMW would design the mechanism with two. Is the shaft shorter because it has a second u-joint? Or is the second u-joint needed because the shaft is short? Or, is the drive shaft at a more severe angle under normal use than the one in the FJR?

Click to expand...

The BMW uses 2 links to locate the rear diff. These 2 links create a virtual swing arm whose length and center can be ascertained by drawing a center line through each link as seen in the side view. Where these 2 lines cross is the center of the virtual swing arm. By playing with the relative angles of the 2 links you can create a long virtual swing arm that is resistant to jacking and you can move the virtual center where ever to mitigate lifting under braking and other unwanted movements caused by various dynamic forces. The trade off is the rear diff is not pointed at the output end and therefore needs a second u-joint to allow for this. It also means that the front ujoint plunges back and forth on the front splines.

Click to expand...

Thanks, that answers my question. It sounds like the angle of the shaft is too great for a single u-joint. I think I prefer the way the FJR does it; simpler, less expensive, and more reliable.

Geezer said:

Thanks, that answers my question. It sounds like the angle of the shaft is too great for a single u-joint. I think I prefer the way the FJR does it; simpler, less expensive, and more reliable.

Click to expand...

BMW don't do it that way for fun

They argue that the paralever rear suspension brings real benefits for the rider. Their aim was to eliminate torque reaction from the shaft drive. This "feature" of all shaft driven bikes is a real thing, and motorcycle journalists have bitched about it since shaft drive was invented.

Back in the real world I surely can't be the only rider that never found the slight rise and fall of the back end an issue not worth obsessing over.

If you are riding on the limits, then I guess keeping the whole plot parallel at all times has some benefits ... but who actually rides an FJR "on the limits" outside of a track day?

I applaud BMW for their innovative design, but preventing a bike from squirming a bit under hard acceleration, while good, comes a distant second to making sure it can get me home.