Clutch Torque Capacity

[crogthomas]

By popular request (one person) I’m subjecting you to my random thoughts on clutch torque capacity.
I’m currently putting together a mildly tuned YPVS engine for a track bike and I got to thinking how I could make sure the clutch doesn’t slip. I don’t have the time, money or inclination to test out all the options, but I do know how to operate a calculator. So I started doing some sums, comparing all the different ways of improving the clutch, then making some notes and pretty soon I ended up with a multi page stream of consciousness. Not wanting my efforts to go to waste, I added some long words and decided to share it and perhaps stimulate some discussion. I am most definitely not a clutch engineer, but it’s mostly quite simple physics.

The equation to determine the Torque capacity, T of a clutch is simple:

T=zμFr

[https://x-engineer.org/calculate-clutch-torque/]

T = The Torque capacity of the clutch
The simplest way to not overwhelm the clutch is to leave the engine standard. No fun.

However, clutches are torque limited, not power. If you could maintain the same peak torque output, but simply deliver it at a higher rpm, you would get increased power, without any additional stress on the clutch, (excepting the fact that the clutch needs to be able to stay together at that higher rpm, without disassembling itself around the insides of the crankcases).
It’s difficult to find exact figures, but a quick look online at power curves shows that a standard LC2 makes approximately 46.5 Nm of torque at around 9000rpm, which equals 59bhp. If that torque was not exceeded anywhere, but extended all the way to 12000rpm, the power output would be 78bhp. Or course, most modifications that increase the torque at high rpms also tend to reduce it lower down, so this works in our favour. Fine for racing, but not ideal for a road bike. And of course, limiting how much torque made is also no fun.

The second workable solution is to reduce the amount of crank torque transferred to the clutch via the primary drive. i.e. change the primary drive ratio. There are 26/69 primary gear kits available which would mean that he engine could make 8% more torque before the clutch reaches its existing limit. The downsides are of course the cost, dealing with the overall change in ratio (presumably with chain and sprocket changes) and the fact that the clutch is now spinning much faster, which it might not be entirely happy with.

[crogthomas]

z = Number of friction surfaces
A standard YPVS clutch has 7 plates, thus 14 friction surfaces. Adding an extra friction plate (and steel) is possible by modifying the standard basket, using thinner plates, or with a aftermarket basket (try saying that after a drink or two). Determining the difference is easy, it’s the percentage increase in number of plates. So those two extra friction surfaces will increase the torque capacity by 14% (16/14).

Obviously there is a limit to how many extra friction plates can be added. Not only will you run out of space, but the more you add the more travel is needed to give the required clearance between plates so the clutch can be disengaged. There is only so much travel that is possible with a normal non-freakish human hand, so if more is required the only solution is to change the lever ratio. Changing the lever ratio makes the clutch heavier, for which the solution is to reduce the spring force, which reduces the torque capacity bringing you back to square one. I don’t know what the limit is, but there is only so much space under the clutch cover to work with anyway.

[crogthomas]

u = Friction coefficient generated between the friction and steel plates.
We don’t have much choice of steel plates and there doesn’t seem to be much that could be done to improve them. People rough up the surface, but it seems that actually decreases friction.
www.raybestospowertrain.com/blog/smooth-is-in-rough-is-out

Likewise, there isn't much choice of friction materials that I can see. RD clutches claim to be cork or paper, some sort of organic material. There are carbon plates available, but I’ve seen no evidence that they actually have a higher friction coefficient, just claims about them lasting longer or coping with heat better. So it seems there is not a lot that can be done about this. I can’t measure the friction coefficient easily and everyone will claim theirs are the best, but I suspect that most replacement plates for the RD will be about the same anyway (somewhere around 0.15). As soon as you start messing with the composition of the material you end up making compromises elsewhere. It needs to work from cold and when hot, slip when required, but then not. Most important of all, the end plates run on the aluminium surface of the clutch hub and pressure plate, so nothing too aggressive can be used anyway.

The friction coefficient in a wet clutch obviously depends on not only the friction material but the oil sloshing around it. My Haynes manual says that I should use Type SE motor oil, presumably following guidance from Yamaha back when there was not as much choice of oils as there is today. Type SE oil was a defunct spec even back in 1983 [https://www.thelubricantstore.com/motor-oil-classifications] and I’m guessing the recommendation was made to try to avoid modern car oils which have additives that would not play nicely with a wet clutch. Generally additives (friction modifiers) are bad, unless they are the good additives. This sounds like an advert for some sort of yogurt.

So what are the options?
• Car engine oil. Avoid. Modern oils have the wrong additives for wet clutches.
• ATF, may work okay. Auto gearboxes have wet clutches.
• 4 stroke motorbike oil. Designed to work with a wet clutch, but also lubricate the engine, so possibly a compromise in some way. This is the only one where there appears to be a standard that measures its effectiveness in wet clutches. A JASO standard that gives an acceptable friction index range. The short story is if you were going to use a 4 stroke motorbike oil avoid MB or MA1 and go for MA2.
• 2 stroke gearbox oil. It would seem that this is probably the best option, considering it is designed exactly for the task. Considering the small quantity required this is where my money would go. Pick a brand with a nice picture on the bottle.
• Water. Coefficient of friction increases with water contamination: www.researchgate.net/figure/Mean-COF-vs-number-of-test-cycles_fig1_269635741 This is a good demonstration of the compromises that an oil has to deal with. It may well improve the clutch, but I suspect only very briefly before resulting in a box full of swarf.

Oil additives are absorbed by the clutch material (we know this because we are told to soak the plates before installing them), so replacing the oil with a new better one might not improve the clutch as much as you expect if the plates are already contaminated with an existing oil, or worse something that contained an additive that actively reduced friction.

Lastly, the friction coefficient of a dry clutch plate seems to be nearer to about 0.3, which means a dry clutch of similar mechanical construction to the RD clutch (like a TZ one) can transfer twice the torque (or could use fewer plates or lighter springs of course).

[crogthomas]

F = The Force on the plates
This is provided by the springs. Nearly anything that is done to increase this force will also make the clutch lever harder to pull. Options are to use stiffer or longer springs, shim the existing ones or combinations of different strength springs so as to find a point somewhere in the middle of too light or too heavy.
No one seems to quote actual spring rates for clutch springs, but most ‘Heavy Duty’ springs state that they are something like 10-20% stronger which translates to a 10-20% increase in torque capacity. A mix of 3 standard and 3 HD springs would give you half that increase. Presumably you should only go for 3 of each, evenly distributed to ensure the force on the clutch pressure plate is not lopsided.
Shimming springs to increase the engaged force is possible up to the point that they become coil-bound, but I’ve not investigated how much that would be. I’ve seen people mention adding 1.5 or 2mm shims to the springs. Adding 2mm shims (washers!) to six standard springs, which have a rate of about 13N/mm [thank you Martin Kieltsch], would increase torque capacity by somewhere around 16%.

I’ve not investigated lock up clutches in much detail, but the few I’ve seen seem to use a centrifugal (or centripetal if you prefer) mechanism where the extra force is added over the existing springs as rpms increase. Presumably it makes the clutch lever heavier at high rpms whilst still being light at idle, I dunno. Not a problem if using clutch-less shifts I suppose. A potentially good solution, but you would have to live with the shame of having scooter technology in your motorbike.

A quick fix for a worn clutch is adding an extra steel plate. It makes up for the wear which has reduced the thickness of the friction plates. It’s basically the same as adding shims under the springs, it increases the force on the pack. One plate is 1.1mm, which would increase it by about 10%. It’s probably much easier just to shim the springs.

Fitting new friction and steel plates to a clutch will probably improve the clutch anyway, simply by not being worn. Some replacement plates appear to be thicker than standard, again increasing the force. Most come as part of a kit with stronger springs, thus adding a bit of extra force from both.

[crogthomas]

r = The mean effective radius of the clutch material
Like brake discs, bigger is better. In this case we will refer to it as the mean (average) effective radius. In this case the mid-point of the clutch friction material is a reasonable approximation i.e. halfway between the ID and OD of the friction plate.
Nothing can be done to make the OD bigger (without some major engineering anyway), but interestingly the commonly used FZ/FZR/YZF friction plates actually have a smaller ID. This results in a smaller mean radius and a lower torque carrying capacity for these plates. It seems counter-intuitive, but the narrower band of the RD plates can actually transfer more torque, everything else being equal. It is a very small difference though, 1mm, which changes the torque capacity by approximately 1.6%.

• Standard OD 134mm, ID 116mm, which makes the mean effective radius = 62.5mm
• FZ plates OD 134mm, ID 112mm. Mean effective radius = 61.5mm

If you are right on the very limit, it might make a difference, but I suspect it probably doesn’t matter. The broader FZ/FZR/YZF plates will last longer and dissipate heat better, which might be the reason Yamaha used them.

I said that you can’t increase the OD, but you can sort of. As standard the RD’s have wibbly wobbly steel plates with part of the circumference slightly flattened. This reduces the OD slightly. Some of the aftermarket steel plates available omit this feature, giving a slightly increased OD over part of the plate. Working out how much exactly is tricky due to the odd shape, but we can work out the worst case scenario quite easily to give us an idea of the most we could expect to gain (or lose).
With the outline of a steel plate drawn of a piece of paper and the application of a protractor from my kids maths set I’d guess the flattened part spans about 78degrees of the plates circumference. The smallest radius is 3.4mm smaller than the rest of the plate. Not all of it is that much smaller, but for the purpose of the calculation we will say all 78 degrees are 3.4mm smaller. That’s 22% of the plate. So we do a calculation with 22% of the friction surfaces with the smaller radius and the other 78% of the friction surfaces with the full radius, then add the results together. In the end the difference is only 0.3%, possibly less as the difference in radius isn't 3.4mm around all 78 degrees.

Note that increasing the surface area of the friction plates doesn’t generate more friction. If you increase the surface area, you reduce the pressure per unit area of friction material (since the force supplied by the springs remains constant). You have more area but less pressure. It ends up cancelling out.

[crogthomas]

Finally a bit of validation.

Yamaha themselves used the simple equation above to make what is essentially the same clutch work in a range of different bikes with far more power and torque than even the most highly tuned RD.
The FZR1000 for example, produces 107 Nm of Torque, but has a primary ratio of 41/68 and has 8 friction plates in its clutch, each of them very similar to the RD plates (slightly different in surface area as mentioned above, but not really a big deal).
So despite it having well over twice the torque of the RD, each of its plates only has to cope with 9.8 Nm, whereas in the RD each deals with 9.5 Nm, which as far as I am concerned is basically the same amount. I’ve no idea what springs the FZR uses, but it seems reasonable to believe that it could use the same ones as the RD without any issue, and thus the lever pull still be comfortable.

[Tobyjugs]

Cheers for that. I now need to let it soak in.

[Chewie01]

That's a lot of information to take in.
Just a thought, the FZR has a hydraulic clutch which would suggest to me that the springs a much stiffer?

[crogthomas]

It was a lot to type.
I'd be interested to see any proper controlled tests of various oils, plates, etc. Unfortunately most 'evidence' is anecdotal at best, with more than one thing being changed at once and the improvement being measured against the old worn out clutch that was being replaced.

I don't think the hydraulic clutch will really make much difference. It doesn't add any force, just transfers it differently. Any advantage could just as easily be obtained by changing the lever ratio of a cable clutch. There may be other advantages mind you, better packaging and perhaps a small gain from the lack of cable stretch or compression of the outer.

[midlifecrisisrd]

Only other advantage of hydraulic is no friction from the cable

Steve

[Tobyjugs]

Feb 22, 2023 9:37:57 GMT 1 midlifecrisisrd said:

Only other advantage of hydraulic is no friction from the cable

Steve

And maybe more travel of the clutch plates.(really important for clutch drag)

[Tobyjugs]

I've used carbon frictions and they are from my experience the best.
BUT the thought of all that carbon floating around in the oil puts me off them. If we had oil filters I would be a fan of the carbon frictions.

[Tobyjugs]

When you quote all the small decimal differences as being almost the same I'm not sure if that is good. I really like what you have posted so don't take that remark as derogatory.

The difference in the pull from an R6 lever and an LC lever can be 0.20mm but it makes a big difference in clutch drag.

[Tobyjugs]

I would also like to hear if Robbieben has any thoughts on this subject?

[Robbieben]

The key factor, or limiting factor is the clutch springs and their clamping force where clutch slip is concerned and the ability to pull the clutch lever comfortably via a cable in the case of the LC and many other bikes. The clutch actuating lever fulcrum also affects ability to pull the clutch lever. Short fulcrum = easy pull, longer fulcrum = heavier pull. Number of plates doesn't necessarily give less slip but adding an extra plate and steel does put more preload on the plates as you have effectively shortened the springs and put them under greater tension, similarly adding an extra washer under the spring creates more preload also but limiting travel available to release the clutch and thus creating drag.

Standard springs on a short fulcrum on the actuator create an easily pulled cable clutch, heavier springs add more pressure to the plates but create a more difficult, hard to pull clutch potentially also stressing and bending the clutch actuating rod slightly again this would contribute to clutch drag.

My personal method on my build is going to be standard amount of plates and steels but heavier springs with a hydraulic clutch, 17mm Frando clutch cylinder and a 16 mm master cylinder, that will drastically reduce the effort required to pull the clutch but allow a decent travel on the system to prevent clutch drag. Clamping force is what ultimately prevents clutch slip on a given friction area, heavier springs create more clamping force you just need to be able to overcome that clamping force.