Mechanical clutch designs meet tough challenges in a multitude of industries: HVAC, bulk handling, packaging, pulp and paper, food and beverage, and especially off-highway and recreational vehicle applications. In these vehicles, manual transmissions use friction disc clutches to completely disconnect engines from their transmissions. On the other hand, automatic transmissions need a way to let the engine turn while the wheels and gears inside a transmission come to a stop. For this task, automatic-transmission vehicles use a torque converter.

A torque converter is a type of fluid coupling that allows the engine to spin independently of the transmission. There are four main components inside its housing: pump, turbine, transmission fluid, and stator. A stator resides in the center of the torque converter to redirect the fluid returning from the turbine before it enters the pump again. Its usually aggressive blade design almost completely reverses the direction of the fluid, which dramatically increases torque converter efficiency.

A subsystem's subsystem

A one-way or mechanical clutch inside the stator connects to a fixed shaft in the transmission. Because of this arrangement, the stator cannot spin with the fluid, which is why the clutch is sometimes called a backstop. In fact, the clutch can spin only in the opposite direction — when it is overrunning. This forces the fluid to change direction as it hits the stator blades. But how does the clutch operate? Mechanical clutches can be thought of as a one-way bearing. They are frictional devices in which cam elements rotate to wedge between raceways to transmit power. When rotation occurs in the opposite direction, the elements “unwedge” and the device overruns, disallowing power transmission.

There are three primary mechanical clutch types: ratchet, roller ramp, and cam.

  • Ratchet clutches are the oldest and most primitive design composed of only two interlocking parts. A toothed piece freewheels in one direction and locks up in the other direction.

  • A bit more sophisticated, roller-ramp clutches have ground rollers that operate in conjunction with an inclined ramp and compression springs for positive engagement. Roller-ramp clutches often experience problems with deflection and stress in the races. Roller ramp designs are also limited by the geometry of the inclined ramp, which limits the number of load bearing elements that can be incorporated into the package.

  • Cam clutches, which we'll focus on here, offer the greatest power density because more load-bearing elements can be packaged into the same space. Adding load-carrying elements distributes load and stress more evenly and significantly reduces the deflection and contact stress at the inner and outer races. In addition, initial engagement is improved over a roller-ramp clutch due to the independent cam actuation.

A cam clutch incorporates a full complement of cam elements (also known as sprags) to carry load when power is transmitted through the clutch. The cams have offset radii so that during rotation, the height of the cam rises and acts as a wedge between the inner and outer races. When power is transmitted, torque is applied to the clutch. This applied torque results in a tangential force at the point where cam and races meet.

The cams must stay wedged to grip by steel-on-steel friction and lock the two races together so they turn as one. For this purpose, the cams use garter springs that energize and help rotate the cams to keep them engaged. The line of compression force through the cam creates a strut of support. The angle of this support (relative to the radial line) is called the grip or strut angle.

Despite the sophisticated nature of cam clutches, design improvements continually enhance their functionality and durability. Today, precision-formed cams deliver greater wear resistance and reduced likelihood of fatigue failure. Ground alloy steel races increase uniform distribution of load for smooth action and longer life, while quality bearings and seals also increase life with minimum maintenance. Note there is a fundamental condition that must be met for a cam clutch to work: the geometric tangent of the strut angle must be less than the coefficient of steel-on-steel sliding friction. As such, extreme-pressure additives should never be used in lubricants without consulting the clutch manufacturer.

The forces involved

Hoop stress is most common in cylindrical and spherical pressure vessels such as pipelines and storage tanks. The stress is an action that tends to pull pipes apart in a circumferential direction, with the pull on pipe walls produced by the internal pressure of the fluid or other drivers inside the pipe. In a mechanical clutch, the inner and outer races are analogous to pipe walls. (The only difference is that pressure isn't caused by a fluid; instead, in clutches it's based on the pressure the cams apply to the inner and outer races.) For this reason, it makes sense to use a similar technical expression to evaluate hoop stress, contact stress, and deflection in these components. If pressure is sufficient to induce a stress in either race greater than the allowable limit of the material, cracking of the races can occur. If the pressure is great enough to induce deflection in either of the races that exceeds the maximum cam rise of the clutch, the cams will roll over. Either scenario leads to catastrophic clutch failure.

To predict when such failures might occur, two techniques were originally published by The Society of Automotive Engineers. Their 1998 Technical Paper Series #981092: Generalized Equations for Sprag One-Way Clutch Analysis and Design, by David R. Chesney and John M. Kremer helps identify variables to address the dilemma of hoop stress, contact stress, and deflection, and include uniform pressure approximation and discrete load approximation. By using these SAE theories and performing the necessary testing and validation, standard cam clutches can be customized. Then once hoop stress, contact stress, and deflection fall within allowable limits, designers can proceed to prototype development with greater assurance of design robustness.

Uniform pressure approximation

In uniform pressure approximation, it is assumed that sprag contact forces act uniformly on the contact surface of the race. Therefore, to apply this method the equivalent pressure of the ‘sprag-way’ must be computed. Equivalent pressure is

Where Q = Distributed load

Fn = Normal sprag load

Na = Number of cams

R1 = Race radius

and W = Race width.

The radial deflection of the sprag-way is computed by using

where 3 is Poisson's ratio.

More specific results

The equations used to measure discrete load approximation are significantly more complex than those used for uniform pressure approximation. Using the discrete load approximation, the expression for the radial deflection of the race at the point of contact with each sprag is given as:

Where ecc is the distance from the mean radius to the neutral axis. The race-deformation correction factors K1 and K2 are defined as

K1 = 1 - acf + bcf

and K2 = 1 - acf

where acf is the hoop deformation factor and bcf is the shear deformation factor.

Tailor-made designs

Multiple design analyses are necessary to identify a custom clutch capable of meeting the various requirements of the application. Initially, the theoretical torque carrying capacity (based on a typical cam and packaging arrangement) is calculated and evaluated.

Let's explore perhaps the most common clutch design challenge — when the thickness and width that can be used for the inner and outer races is constrained. In these situations, a roller-ramp design's primary failure modes are stress and deflection. Cam clutches often address these problems with increased power density by offering more load carrying elements, hence reducing both stress and deflection in the raceways.

Evaluating and charting twelve design alternatives for a system with a 2,600 lb-ft. maximum operation torque, it's clear only seven meet the original torque capacity required by the application. Maximum torque with an applied safety factor must also be tested — in this case, a 1.5 safety factor for 4,000 lb-ft. From this evaluation the seven qualified alternatives can be narrowed down to three feasible designs.

Testing and validation

Stress and deflection analyses show how the different design alternatives all significantly reduce contact stresses and deflection present in the roller ramp design. In addition, each design alternative improves contact stress and deflection as additional cam elements are added to the clutch assembly. However, the calculated race deflection at any set torque varies dramatically between the uniform pressure and discreet loading methods. Theory indicates that the discreet loading method offers greater accuracy. As such, for our example one would expect Design Alternative 11 to outperform Design Alternative 6, and Design Alternative 12 to outperform both in terms of torque-carrying capabilities.

In fact, bench testing validates these conclusions. With test fixtures an engineering team can observe torque versus angular displacement over time for all design alternatives. To continue with our example, these results are shown in tables. The maximum torque calculations have an embedded service factor to allow a 1,000,000-cycle rating. As such, the actual maximum torque is significantly greater. Testing shows that the torque-carrying capability of each clutch increases as contact stress and deflection are reduced. Discrete load approximations also indicate these reductions; this verifies its validity and accuracy in predicting the power density of a mechanical clutch for the most realistic evaluation when compared to actual applications.

Torque calculation
Design option Bearing support Result Cam height, in. Cage width, in. Calculated Torque, lb-ft
1 2 X LS 3.75-in. BB Too weak and narrow 0.5 1.065
1//Revised 2 XLS 3.75-in. BB Includes thrust-plate width 0.5 0.772
2 1 XLS 3.75-in. BB 0.5 1.815
2//Revised 1 XLS 3.75-in. BB Includes thrust-plate width 0.5 1.522
3 2 13-mm wide BB Smaller bearing helps 0.5 1.541
3//Revised 2 13-mm wide BB Includes thrust-plate width 0.5 1.249
4 1 XLS 3.75-in. BB 0.5 1.522 5,160
5 Intermittent rollers On hold 0.5 2.272 7,000
6 2 13-mm wide BB Ideal bearing setup //Standard MR cage 0.375 1.249 4,030
7 1 13-mm wide BB, 1 SB SB width = 0.453 in. 0.375 1.610 5,500
8 2 SB SB width = 0.453 in. 0.5 1.970 7,000
9 1 13-mm wide BB, 1 SB SB width = 0.453 in. 0.5 1.610 5,200
10 2 SB SB width = 0.453 in. 0.375 1.970 6,650
11 2 13-mm wide BB Ideal bearing setup //Custom cage 0.375 1.250 4,100
12 2 13-mm wide BB Ideal bearing setup //Custom cage 0.375 1.250 4,500

The theoretical torque carrying capacity based on cam selection and packaging arrangement can be calculated. Here, BB stands for ball bearing and SB stands for sleeve bearing. Cage type — styles with individual pockets, pockets that group several cams together, and cages with intermittent rollers — also affects torque-carrying abilities. The maximum torque calculations have an embedded service factor to allow a 1,000,000 cycle rating.

Stress and deflection
Torque, lb-ft 2,600 2,600 2,600 2,600 2,600 2,600 2,600 2,600 2,600 4,000 4,000 4,000 4,000 4,000 4,000 4,000 4,000
Design option (Benchmarked design) 5 6 7 8 9 10 11 6×10 5 6 7 8 9 10 11 6×10
Bearing support and sprag height, in. Rollers, 0.5 2 BB, 0.375 BB, SB, 0.375 2 SB, 0.5 BB, SB, 0.5 2 SB, 0.375 2 BB, 0.375 2 BB, 0.328 Rollers, 0.5 2 BB, 0.375 BB, SB, 0.375 2 SB, 0.5 BB, SB, 0.5 2 SB, 0.375 2 BB, 0.375 2 BB, 0.375
Number of cams 35 48 48 40 40 48 54 60 35 48 48 40 40 48 54 60
Cage width, mm 57.7 30.2 40.9 50.1 40.9 50.1 30.2 30.2 57.7 30.2 40.9 50.1 40.9 50.1 30.2 30.2
Sprag width, in. 2.185 1.820 0.865 1.314 1.518 1.157 1.675 .0830 0.830 1.820 0.865 1.314 1.518 1.157 1.675 0.830 0.830
Effective width IR 2.268 1.314 1.763 1.967 1.605 2.124 1.279 1.279 2.268 1.314 1.763 1.967 1.605 2.124 1.279 1.279
Effective width OR 2.633 1.797 2.247 2.331 1.970 2.701 1.762 1.762 2.633 1.797 2.247 2.331 1.970 2.701 1.762 1.762
Hoop stress on outer race, psi 37,264 30,481 38,612 30,892 34,430 40,739 24,343 39,379 39,379 46,894 59,403 47,526 52,969 62,676 37,451 60,583 60,583
Hoop stress on inner race, psi -51,378 -51,489 -74,900 -55,817 -59,396 -72,752 -54,992 -76,950 -76,950 -79,214 -115,231 -85,872 -91,378 -111,926 -84,603 -118,385 -118,385
Contact stress on inner race, psi 471,706 276,275 384,552 311,987 282,975 324,141 280,382 370,124 351,131 342,677 476,978 386,972 350,987 402,048 347,771 459,083 435,524
Safety factor 0.91 2.65 1.37 2.08 2.53 1.93 2.58 1.48 1.64 1.72 0.89 1.35 1.64 1.25 1.67 0.96 1.07
Contact stress on outer race, psi 431.359 245.528 352.088 285.649 251.482 288.067 256.149 338.879 321.489 304.539 436.712 354.305 311.925 357.303 317.713 420.327 398.758
Percent reduction 43.1 18.4 33.8 41.7 33.2 40.6 21.4 25.5
Deflection reported by discrete load, in. 0.0075 0.0048 0.0060 0.0047 0.0052 0.0063 0.0034 0.0060 0.0058 0.0072 0.0092 0.0071 0.0079 0.0095 0.0052 0.0091 0.0089
Deflection reported by uniform pressure, in. 0.0090 0.0069 0.0093 0.0071 0.0077 0.0094 0.0062 0.0094 0.0092 0.0106 0.0142 0.0108 0.0118 0.0143 0.0095 0.0143 0.0141

Analyses on design options 6, 11, and 12 performed at 2,600 and 4,000 lb-ft. show significant differences. The listed safety factor is based on a maximum contact stress of 450,000 psi. Maximum hoop stress is assumed 50,000 to 80,000 psi, while the maximum contact stress is assumed 450,000 to 550,000 using various common guidelines.

THE PROCESS WE'RE DESCRIBING

Design and customization require intense collaboration. Inquiry teams composed of cross-functional groups should meet regularly to evaluate progress. A project management model should manage special inquiries throughout the OEM and contract management process. A clutch specialist might act as a single point-of-contact and interface with an application team made up of an application engineer, a design engineer, a procurement representative, industrial engineers, a product manager, and a plant or operations representative.

After successful collaboration between the clutch manufacturer and designers (including design iterations, analyses, and bench testing) the best design alternative should then be tested on a dynamometer; this dynamometer test should simulate real-life conditions as seen in the field. The results often prove successful, as the selected cam clutch can outlast other clutches by hundreds of hours.

For more information, call Emerson Power Transmission of Emerson Electric Corp. or visit www.emerson-ept.com.