Clutch Weights are the most misunderstood part of a continuous varying transmission, CVT, drive system. When you first see them they appear very simple and uncomplicated. But as with many things, looks are deceiving and things that seemed simple are very complicated.
There are many handbooks and reference manuals on the market today that do an excellent job of explaining how CVT drive systems work as well as what to do to tune and make changes to them. But in all of this information there is a void in the area of the clutch weights. In fact, some of these manuals go as far to say that you really don't need to know how clutch weights work, rather only that they do. But it's been my experience that clutch weights are an area where great performance gains can be achieved. Until this year, all clutch weights have looked much the same; so how can they act so differently? The philosophy in the past has been that clutch weights are more "art" than science, but hopefully we will challenge the accepted norm.
In this article we will attempt to show all of the variables that are in a CVT clutch system. So as a tuner you will be aware of all the things you are currently affecting without being aware of it. We will show that clutch weights can be more of a science than an art!
This technical article will start by showing what is right and wrong with the current CVT clutch design. Then we will shift our focus on how the radically new designed "Dual Quadrant" Heel Clicker( clutch weight works, and how it provides us with performance gains.
What Do Clutch Weights Do?
Clutch weights are used for two basic functions. They provide "squeezing forces" and "up-shift force" to the belt. The force generated by clutch weights is called centrifugal force and it is this centrifugal force that gets divided up into both up-shift and squeezing forces at the same time. Both the squeezing and up-shift forces are applied to the belt directly and are related by the sheave angle. The sheave angles are not the same for all CVT clutches. As Fig. #1 shows the Ski-Doo TRA clutch has a 24 degree included angle whereas Polaris, Arctic Cat, and Comet use between a 27-29 degree included angle. What this means to us is that the TRA clutch applies more squeezing force and less up-shift force than the others. It has been my opinion that could be one of the secrets that makes Ski-Doo's very strong drag racing machines.
Sheave angles: TRA vs. Others
It is important that the reader understand that both squeezing and up-shift forces are controlled by the clutch weight. It is only the squeezing force that will be discussed in this article. In the body of this article we will a design a clutch weight to give the optimum squeezing force for a specific amount of torque. By identifying all of the variables that effect squeezing force, the reader can better tune a CVT clutch. When we are finished looking at the numbers, we will than compare the theory and see if it correlates to a real life CVT drive clutch.
There are several key variables that must be understood to achieve optimum squeezing forces in a CVT drive clutch. Several handbooks over the years have been written to give the high performance snowmobiler with a good basic understanding of how to tune CVT transmissions for optimum performance. One of the handbooks that I considered the most informative is the Olav Aaen Clutch Tuning Manual. This manual has been updated several times and has some very good information in it. In this manual is a graph that shows what I refer to as the "thumb nail sketch" on page 16. This graph shows belt squeezing forces versus shift ratio, but it does not provides any actual squeezing force data, it only shows a trend. For the high performance snowmobiler, knowing what the proper squeezing force is gives a starting point for any further discussion on this subject.
Recently technical SAE (Society of Automotive Engineers) papers have been written on this subject. The auto industry has shown great interest in CVT development; mainly because CVT's are inexpensive to build, lightweight, and transfer power efficiently. These papers contain information on proper belt squeezing forces vs. shift ration for maximum power transfer. This information will fill in the blanks so we can understand how and what clutch weights must do to achieve maximum performance. As we will see, our snowmobile CVT clutches have a basic design problem that needs to be fixed.
Figure # 2 is a graph of belt squeezing force vs. shift ratio. The information gained is from one of the above mentioned SAE papers. It is extremely high quality information that fills in the missing data in the thumb nail sketch in Olav Aaen's clutch tuning manual. It is also important to remember that this data is subject to the belt compound and its ability to adhere to the CVT sheave faces. Belt compounds have changed over the years, and this graph should be used as a guide that points in the proper design direction.
Belt squeeze forces vs. shift ratio
Squeezing force, also know as axial force, is the force needed to stop the rubber CVT belt from slipping on the cast aluminum clutch faces. The forces shown in figure #2 above are the forces necessary to achieve a balance between belt stretch and belt slippage. This balance will result in maximum clutch efficiency. To achieve maximum efficiency in a CVT clutch, two variables must be balanced. These two variables are belt slippage and belt stretch. Maximum efficiency in a CVT drive clutch is achieved when just enough squeezing force is applied to the belt to stop it from slipping. This is why understanding the information in Fig. #1 is so important. This means that the high performance snowmobiler must calibrate the CVT clutch to balance belt slippage and belt stretching at all shift ratios to obtain optimum performance from his snowmobile.
Also notice that the upper and lower torque limits in Fig. # 2 are 57 and 100 ft-lbs of torque, respectively. This is in the exact range of today's 440 Sno-cross machines to 800 triples. To use this graph, just interpolate the data between the two lines. Figure #2 is very important; we will come back to this diagram once we calculate the squeezing force generated by the clutch weight minus the clutch spring.
CVT Clutch Design Variables That Effect Belt Squeezing Force
As seen in Fig #2, an engine producing 100 ft-lbs of torque must produce at least 650 lbs. of belt squeezing force at the 1-to-1 shift ratio. Figure #3 is a sectioned view of one of today's CVT clutches. All the important dimensions are shown that will be needed to find out how much squeezing force our clutch weights are generating. The distance from the centerline of the crankshaft to the clutch weight pivot pin is the most critical dimension in the design of a CVT clutch. As we go through the calculations for centrifugal force, it will be apparent that the location of the pivot pin controls the center of mass (COM) of the clutch weight as the clutch shift from low speed to high speed. This COM location is a key point because all forces are calculated at this point.
Stationary Pin Location
Notice that the pivot pin location is the same for three of the four designs shown. This dimension is 81.5mm from the centerline of the crankshaft. The Polaris, Arctic Cat, and Comet design differs from the Yamaha design by about 5 mm. The extra 5mm used by Yamaha allows them to run a lighter clutch weight and still produce sufficient belt squeezing forces.
Figure #3 also shows the location of the roller in relationship to the pivot pin. This relationship is fairly constant between the four clutch designs and very small differences in squeezing force will result from this dimension. This is not to say that roller diameter does not make a difference in belt squeezing force; it does. Roller diameter changes the relationship between the clutch weight rotational position and the belt shift ratio position. Smaller diameter rollers will produce more side loading force/shift ratio than large diameter rollers.
Current Clutch Weight Design
Now that we know that the clutch design fixes the clutch weight in a precise distance from the centerline of the crankshaft, we can start investigating how the weight is designed to give us the proper belt squeezing forces.
As an example, we will use an A-54 Arctic Cat clutch weight as it rotates around the stationary pin. The A-54 weight is the standard weight used in the ZRT 800. On a good cold day a stock ZRT 800 can produce 100 ft-lbs of torque and peak power is at 8500 RPM. After the analysis is done, we can compare directly back to the figure #2 to see if the A-54 weight was properly picked as the weight for that application.
To get to the proper squeezing force to the belt so 100 ft-lbs. of torque can be transferred, several calculations need to be made. The following is a list of the calculations needed to fully understand how the clutch weight system (including the spring) works.
1) Centrifugal Force of three clutch weights
2) Force Balance between the stationary pin and the roller
3) Force Transfer between the clutch weight and roller
4) Subtraction of the force due to the clutch spring
As stated earlier, generating centrifugal force is the primary function of the clutch weight. The clutch weight and its unique design will apply only a small portion of the centrifugal force to the belt. The three 54gr clutch weights will generate more than 2000 Lbf. This is enough force to stop the belt from slipping, but it is the clutch design that allows only a portion of that force to reach the belt.
In the centrifugal force equation the mass always remains constant, but the velocity is dependent on the radius. So as the clutch weight rotates from position 1 through 5, the radius continually changes as the clutch weight rotates around the stationary pin. See figure #4.
Clutch Flyweight Rotational Positions
All of the centrifugal force is calculated through the COM of the weight. This is one of the limiting factors in our current CVT clutch. The COM can only rotate on a fixed radius around the stationary pin. This is why the stationary pin location is very important in clutch design. By changing the stationary pin location and using the same clutch weight the designer can effect the centrifugal force by changing the radial distance to the center of mass (COM).
In our example, the A-54 clutch weight generates more than 2000 pounds of force (Lbf). Now you can understand why a CVT clutch has so much mass and structure! I remember talking to some of the old timers and being told that clutch weights produced a ton of force when they were running at 9000 RPM. At the time I found it hard to believe, but as you can see it is very possible. In fact, if you do the same calculations with 66gr weights you will find that over 3000 Lbf are generated.
Force Balance Between The Stationary Pin And The Roller
When we compare the centrifugal force being generated by the cam weights to Figure #2, you can see that there is sufficient force to stop the belt from slipping. The 54gr clutch weights in this example generate more than enough force. In fact, this much force would surely over stretch and destroy the belt! This means the CVT clutch must go through a force scrubbing (or force loss) operation. This scrubbing operation transfers most of the centrifugal force to the clutch structure and not to the belt. This scrubbing operation occurs in three separate operations. The first is known as force balance.
Many people do not understand why the CVT clutch is designed to generate so much centrifugal force. We only need a small percentage of this, so why generate it in the first place? There are areas where the clutch can not produce enough belt squeezing force to stop the belt from slipping. The clutch is designed to give good average performance between belt slippage and belt stretch.
The first scrubbing operation is called force balance. The force balance is a very simple ratio. It is a ratio used to divide up the centrifugal force between the stationary pin and the roller.
Clutch Flyweight Rotational Positions
In Figure #5, note that it is the stationary pin and the roller that stop the cam weight from simply flying right out of the clutch. As the clutch weight rotates around the stationary pin from position 1 to position 5, the COM of the clutch weight moves closer to the roller and further away from the stationary pin. This movement of the COM of the clutch weight from the stationary pin to the roller shifts the force being generated from the pin to the roller based on distance.
This simple ratio determines what portion of the centrifugal force is going to the roller or to the stationary pin. As the weight rotates, the COM moves closer to the roller thus allowing a greater portion of the force to be transferred into belt squeezing force at the roller. Fifty to seventy five percent of the centrifugal force has already been scrubbed to the clutch after the force balance operation has taken place.
The second of the scrubbing operations is force transfer. Force transfer is the force transferred from the clutch weight to the roller. This is different from the force balance because we now we take into account the curved surface of the clutch weight face itself. In the previous segment on force balance, the discussion centered on balancing the force between two points, the stationary pin and the roller.
Force transfer occurs through the radial curved face on the clutch weight itself. This curved surface has long remained a mystery, but is a very key part of tuning a CVT clutch. This curved surface of the clutch weight is where most aftermarket companies work to create "aggressive" clutch weights. The simplest way of understanding force transfer is to show a couple of examples. Figure # 6 shows the interaction between the clutch weight and the roller.
Curved surface and angle between weight and roller
The force F acts normal (or at 90 degrees) to the surface of the clutch weight. The optimum force transfer angle would be 90 degrees. If the weight could push on the roller at 90 degrees at all positions 1-5, the force balance would be equal to the force transfer.
But as with most things, the optimum can not be achieved in real life. The reason for this is that the clutch weight is fixed at the stationary pin and has to push the roller over 1-3/8 inches as the clutch shifts from position 1 to 5. Figure #6 also shows our five positions as the clutch weight shifts out. See that we have a 7.2 degree contact angle at position 1, but end up with only 29.7 degrees at position 5.
A higher percentage of the force was scrubbed off in positions 1, 2, and 3 during the force balance operation. When we did the force transfer it is positions 3, 4, and 5 that have the higher percentage of force scrubbed. This is a good example of how complex and creative the current CVT clutch is. It uses these different operations to accomplish different things. As I stated earlier, the CVT clutch creates a force and systematically removes it. Knowing how these different operations work will allow us to add or subtract belt squeezing forces as needed.
The final part of the force scrubbing operation is the clutch spring.
Clutch springs are the easiest of the three force scrubbing operations to understand. The clutch spring is required in a CVT clutch to stop the movable sheave from collapsing onto the belt before the engine has made sufficient power to move the snowmobile forward.
Clutch springs are measured in pounds of force at varying load heights. The spring used for our example is a "150-300." This means that at position number 1 the spring force is 150 LBF whereas at position 5, the force is 300 LBF. The rate for this spring is 115.4 lbs/inch. Subtracting the spring force is very straightforward and this form of scrubbing force in the CVT clutch is the easiest to understand.
Total Belt Squeezing Force Results
Figure #7 shows the weakness in the current CVT clutch design. It shows what the A-54 weight does as compared to what is needed. Notice that the A-54 clutch weight applies sufficient load to the belt in positions 4 and 5, but not in positions 1-3. This can easily be verified by simply looking at the clutch when the belt is removed from any snowmobile. You will notice that there are black marks on both sheave faces. This is result of the belt slippage. Not only is this belt slippage is very hard on belts--more importantly, it does not allow the true power of the engine to reach to the ground! If we can fix the belt slippage problem in positions 1-3 and not over stretch the belt in positions 4 and 5, we can create an ideal solution to the problem of all CVT clutches.
Optimum Belt Squeezing Force
The Solution Is Here--"Dual Quadrant" Clutch Weights
Now that the problem has been identified, we now can move forward with a solution. Belt slippage in the lower part of the shift curve is due to several factors. The most important factor is the lack of large contact patches between the belt and the two clutch sheave faces. Figure #8 shows the difference in contact patches at several different positions.
There is very little contact between the belt and sheave faces in the lower part of the shift curve. This relatively small "contact patch" at position 1-3 is a serious problem. But how can the problem be fixed without a major redesign of the CVT clutch? And if we did increase the area of the contact patch, the gear ratio advantage we know enjoy would be gone.
There is another solution to the belt squeezing problem. If more force in the lower part of the shift curve can be applied to the belt, we can improve or totally eliminate the belt slippage problem. Welcome to the new "dual quadrant" Heel Clicker( Clutch Weight.
Belt slippage in CVT clutches has just become a problem in recent years. Fifteen years ago it was hard to find a consumer machine that made more than 60 ft-lbs of torque at any RPM. Today almost all of our high performance snowmobiles produce over 60 ft-lbs of torque. Even our 440 race sleds are at these levels.
The problem is that Clutch technology has not kept pace with engine technology. Fifteen years ago nobody would have realized belt slippage was a problem because of the lower power levels of the day. Today a fix is necessary to improve the efficiency of the CVT clutch. Track dyno testing let's us see than today's high performance snowmobiles only get about 50% of the power from the engine to the ground. In other words, a snowmobile making 100 engine horsepower is only getting 50 horsepower to the track!
A solution has been developed that requires a whole new way of looking at clutch weights. What I have developed is a clutch weight that applies more belt squeezing force in the first part of the shift curve without over stretching the belt thus letting us achieve the best solution.
The technology is known as "Dual Quadrant Flyweight Technology". This technology simply applies a second mass in a different quadrant (see fig. #9) than the current clutch weight. This second mass (know as the "shoulder") is attached or integrally formed with the current clutch weight in the area above the roller. This technology relies on clutch weight rotation to load and unload the belt at the right time.
In previous figures you can see that the clutch weight only has a single center of mass that can only rotate in quadrant #4. The dual quadrant design places the shoulder in quadrant #1. The shoulder has it's own center of mass and only rotates in that quadrant. The forces generated by the two separate masses can simply be added together as if they were independent
Dual Quadrant Flyweight Technology
Notice that the dimension (A) for the shoulder is further away from the stationary pin at position #1. In fact, dimension A is almost greater than dimension L. The shoulder acts as a lever arm and pulls the roller as it rotates around the stationary pin instead of pushing it as the clutch weight does. This shoulder mass and leverage translates into a new force that when combined with the current clutch weight adds tremendous belt squeezing force in the first part of the shift curve.
As the weight rotates from position number 1 to 5, the shoulder portion of the weight has less influence at the roller. If we go back to the force balance part discussed earlier, the A dimension is now decreasing as the length L to the roller increases. This means the roller is seeing less and less influence from the shoulder as the weight rotates. Thus at position 5, virtually no influence from the shoulder is observed at the roller and belt. All this means is that the dual quadrant clutch weight has improved the efficiency of the CVT clutch without over stretching the belt at positions 4 and 5.
Below Figure #10 shows the same positions 1-5 of our earlier example.
Clutch Flyweight Rotational Positions
Using Dual Quadrant Flyweight Technology
click to view the illustration
When calculating the centrifugal force, and the force balance, use just the shoulder portion of the weight. It can be simply added to the existing A-54 clutch weight in our example.
Remember that the distance from the centerline of the crankshaft to the COM (of the shoulder only) has changed significantly. This means that the centrifugal force generated by a 7gr mass is significant. Both the force and the velocity terms have changed so be careful.
The force balance is where the greatest force increase to the roller can be seen. Remember the force balance is a ratio multiplied by the centrifugal force. Since the COM of the shoulder is now situated out further away from the stationary pin and closer to the roller, a lever arm has been formed. If the COM of the shoulder is longer than the (A) dimension it forms a lever arm that acts as a force multiplier of the centrifugal force produced by the shoulder. This mean the dual quadrant design uses geometry and not just weight to add additional force to the belt through the roller. (See fig #10 pos. # 1).
Figure #11 is our final look at belt squeezing forces. The "dual quadrant" clutch weight is so revolutionary! It offers solutions to problems that many never realized they had.
Superior Up-Shifting Forces "Without" Multi-Angled Helixes
The dual quadrant design will also provide superior up-shift force to any CVT clutch. Remember earlier I stated that up-shift and belt squeezing forces were directly related by the sheave angles. When belt squeezing force increases, so does the up-shift force. This simply means that adding the shoulder will stop belt slippage and increase up-shift speed similar to a multi-angled helix. Out testing has shown that when more weight is added to the shoulder location, acceleration also increases just like you installed a multi-angled helix. The rule of thumb is, put as much weight to the shoulder as the engine can handle. You will know the limitation of your motor when the engine has a slight hesitation at clutch engagement. Additionally, our test vehicles have shown excellent belt life when transferring over 150 Hp. This is due to absence of belt slippage.
Optimum Belt Squeezing Force
With Dual Quadrant Flyweight Technology
The Result: Big Gains In Track Horsepower!
Track horsepower is the final result. Some may say the speed of elapse times are the final answer, but I tell customers that traction is now your problem, and as a high performance snowmobiler, it's one of those good problems to have.
Shown below in figure #12 is the result of a track dyno session at Goodwin Racing in Zion, IL. The sled tested was a '98 ZR600 well set up for asphalt racing. The test was a simple before-after comparison with the "best" known set-up and a set of 45-7 Heel Clicker( clutch weights. The track horsepower increased 20-25 percent from 40 mph and up. This same sled was then run for elapse time through the 1/4 mile with excellent results.
Horsepower vs. Speed
With Dual Quadrant Flyweight Technology
In conclusion, the Heel Clicker( clutch weights will work with any clutch kit you may already have. Remember the dual quadrant deign simply fixes a flaw in the CVT clutch design, and will only compliment any other clutch kits on the market. The Heel Clicker( clutch kit comes with two high preload clutch springs (specially designed to utilize the advantages generated by the cam weights), weights, and all the necessary adjustable hardware.