CONTINUOUSLY VARIABLE DIFFERENTIAL BACKGROUND

Although there have been many schemes for varying the torque bias ratio between drive axles connected by a differential, all of them have experienced problems; and the open differential is still the most popular one in use. This has a relatively efficient torque transmission train and a low bias ratio, which works well so long as both wheels have traction. The low bias ratio makes it easy to get stuck if one wheel loses traction, however, because not much torque can be transmitted to the opposite wheel.

The many attempts to overcome this have mostly involved reducing the efficiency of the interaxle torque transmission through the differential. This increases the bias ratio and increases the torque that can be transmitted to one wheel when another one slips. A high and constant bias ratio can cause other problems, however. Since nearly all differentials have an interaxle speed ratio of -1 (meaning that axles rotating relative to each other do so in an opposite direction at the same speed) , and since all differentials have efficiencies of less than 100%, those with a -1 speed ratio always apply a proportionally larger amount of torque to the slower rotating axle on the inside of a curve. Differentials with a high bias ratio exaggerate the greater proportion of torque applied to the inside wheel and thus create an under steer moment, urging the vehicle straight ahead while it is turning.

Many proposals have also been made for differen¬ tials having interaxle drive efficiencies that can be varied during operation. Fluid couplings and friction clutches are among the possibilities. Although achieving some success in spite of complexity and reliability problems, none of these (nor any other differential with a -1 speed ratio) can apply a proportionally larger amount of torque to the faster rotating axle on the outside of a curve. This can be desirable because the vehicle weight shifts to the outside wheel on a curve, and the more heavily weighted wheel has more traction with which to exert the available torque.

It has also been proposed, in Chambers U.S. Patent No. 4,535,651, to vary speed ratios of a differential by means of multiple gear trains that can be engaged alterna¬ tively to increase the torque to a slower moving axle if necessary. This may help tractors and slow speed vehicles get a better traction grip, but it would make a high speed automobile unstable to abruptly change its axle speed ratios. Also, shifting gear trains is cumbersome and expensive and, as proposed by Chambers, is still not capable of providing more torque to a faster rotating axle traveling the outside of a curve.

I have discovered a way of transmitting drive torque through a differential to a pair of drive axles in a freely variable manner that allows more of the torque to be sent to the faster rotating axle on the outside of a curve and generally allows the torque distribution to be controlled in response to vehicle driving conditions. My differential is thus able to achieve torque bias ratios not attainable with previous differentials that use torque transmission efficiencies to vary bias ratios. -In creating new torque distribution possibilities, my differential allows vehicle suspension and steering design to take advantage of torque variation under different driving conditions such as turning, braking, varying weight distributions, and varying traction. Besides being usable between a pair of opposed axles, my differential can also distribute driving torque between front and rear axle pairs. SUMMARY OFTHE INVENTION

My differential varies the speed ratios used in transmitting torque to a pair of rotatable shafts serving as opposed drive axles or as torque distributors between axle pairs, and it accomplishes this by using a continuously variable transmission. Instead of the transmission being fixed in place and used to vary speed ratios between its input and its output, I connect these respectively to the shafts or axles and rotate the transmission with drive torque so that the whole transmission turns between the driven shafts. I then control variation of the transmission to vary the speed ratios used in transmitting the drive torque to the

shafts or axles so that their mechanical advantages also vary. This allows controlled apportionment of the drive torque between axles when they rotate differentially in response to vehicle drive conditions. These can include axle acceleration, axle weight distribution, turning, velocity, engine braking, and possibly others. Information about vehicle drive conditions can be processed to control the variable speed ratios of the transmission to divide the torque between the drive axles according to the vehicle's needs. This can include applying more of the torque to a faster rotating axle on an outside of a curve, as well as more torque to an axle retaining traction when its opposite axle slips. It can also include applying more torque to one pair of axles that have better traction or bear more weight than another pair of axles that are more inclined to slip. DRAWINGS

Figure 1 is a partially schematic, cutaway eleva- tional view of one preferred embodiment of my continuously variable differential; and

Figure 2 is a cross-sectional view of the differ¬ ential of FIG. 1, taken along the line 2-2 thereof. DETAILED DESCRIPTION

To achieve varying interaxle speed ratios within my continuously variable differential (CVD) , I use a continu¬ ously variable transmission (CVT) . These are intended to transmit power from a rotating input shaft to a variable speed output shaft, and they control the speed ratio between the two shafts. My CVD makes use of the speed ratio control of a CVT, but does not transmit power through the variable speed ratio path of the CVT. Instead, I connect opposite ends of the variable speed ratio path respectively to driven output shafts or axles and rotate the entire CVT with drive torque. This directly drives the shafts or axles, which are also free to rotate differentially via the speed ratio path at negative speed ratios that can be varied throughout a continuum by controlling the CVT. My CVD can differentiate between shafts that distribute torque to the front and rear axles of a four wheel drive vehicle, as well as divide the torque between a pair of opposed axles. The best way to

explain my CVD, however, is in the familiar position of differentiating drive torque between opposed axles, and this function is assumed throughout the following explanation, unless otherwise specified.

Many continuously variable transmissions exist and operate in ways that make them suitable for my differential. Most of the available CVT ^l s are traction drives that use balls, rollers, disks, cones, and other friction drive ele¬ ments that can rotate at varying radii. Many of these are potentially suitable for use in my differential. The requirements include compact size, capability of being rotated as a whole, and ability to transmit drive torque to output elements that are oppositely rotatable. If the output elements of a CVT rotate in the same direction, one of them can be changed to negative by gearing.

Not all continuously variable transmissions are traction drives. One CVT under development uses variable duration electric pulses to change the speed ratio coupling between an input shaft and an output shaft. This, and any other CVT that can vary speed ratios and can be rotated between shafts or drive axles to accommodate their opposite rotation, can be selected for use in my differential. Size, ease of control, expense, and durability would be additional considerations.

From among the many available possibilities, I have chosen for illustration a traction drive CVT 50 using rollers 15 rotating between output plates 11 and 12 within continu¬ ously variable differential 10. As true of any CVT selected for differential use according to my invention, CVT 50 is rotated by drive torque, in this case applied to ring gear 13 by hypoid gear 14, although bevel gear, spur gear, worm gear, and other inputs are possible Ring gear 13 forms a rotat¬ able casing carrying rollers 15, which revolve with drive torque and, by their frictional engagement with plates 11 and 12, rotate these as well. Plates 11 and 12, like the side gears of an open differential, are connected respectively with opposite drive axles 17 and 18, which rotate with the drive torque applied to plates 11 and 12 by revolving rollers 15. When my CVD 10 is used to distribute drive torque to

front and rear axle pairs, shafts 17 and 18 become drive shafts to axle differentials, rather than opposed axles.

The angles of rollers 15 relative to plates 11 and 12 are changeable to vary the speed ratio between shafts 17 and 18. As schematically illustrated, arms 16 carrying rollers 15 are connected by a pivot pin 19 that is movable axially between plates 11 and 12. A nut 20, carrying pin 19, is mounted on a threaded rod 21 that is rotatable by motor 25. This moves nut 20 axially back and forth between plates 11 and 12 to change the location of pivot pin 19 and the angles of rollers 15 relative to plates 11 and 12.

As shown in FIG. 1, rollers 15 are angled to engage plate 11 at a minimum radius and plate 12 at a maximum radius. This makes the speed ratio of axle 17 to axle 18 a proportion of 3:2, for example, because three turns of plate 11 at the short radius of engagement of rollers 15 could be required for two revolutions of plate 12 at the longer radius of engagement of rollers 15. Since torque varies inversely with speed ratio, and is proportional to mechanical advantage or moment arm, the torque distribution to axles 17 and 18, when they rotate differentially, has a bias ratio of 2:3. When differential rotation occurs between axles 17 and 18, this bias ratio delivers more torque to axle 18 than to axle 17, for reasons such as axle 18 being on the outside of a curve, bearing more of the vehicle's weight, or maintaining traction while axle 17 slips. By rotating rod 21 with motor 25, the angles of rollers 15 can be reversed from the position of FIG. 1 to reverse the speed ratios and torque distributions and apply more differential torque to axle 17. For normal driving, nut 20 can be centered so that rollers 15 engage plates 11 and 12 at equal radii, making the speed ratio -1 between axles 17 and 18. The 3:2 speed ratio and 2:3 torque bias ratio is not an upper limit, and is merely an example; for speed ratios, and corresponding torque bias ratios, can vary throughout the range available from the particular CVT being used.

Supports 22 for the pivoting of rollers 15 within gear carriage 13 are schematically shown in FIG. 2. Many variations can be made on roller support, roller angle

control mechanisms, and torque drive input; and wnen oτ_ner CVT's are used, other variations become available.

A CVT arranged in my continuously variable differ¬ ential 10 does not transmit power between an input and an output, as is normal for a CVT. Only torque is transmitted from input gear 13 via rollers 15 to output plates 11 and 12, since all these rotate together under drive power. It is only when differential rotation occurs between axles 17 and 18 that there is any counterrotation of plates 11 and 12, to transmit torque between axles. This happens through a rela¬ tively efficient torque train involving rollers 15 between plates 11 and 12; and with differentials having a fixed speed ratio of -1, this would produce a low torque bias ratio. But the speed ratio of CVT 50 is variable throughout a continuum by changing the angles of rollers 15. This inversely changes the radii of engagement of rollers 15 with plates 11 and 12 for varying the moment arms and mechanical advantages of plates 11 and 12, thereby effecting differential torque distribution independently of axle speed. More torque is transmitted to the axle whose output plate is engaged .at the larger radius by rollers 15, and unlike the multitude of differentials having a fixed interaxle speed ratio of -1, more torque can be applied by CVD 10 to the faster rotating axle on the outside of a curve.

Controlling the differential torque distribution with my CVD requires sensing vehicle drive conditions and varying the speed ratio of CVT 50 rotating within differen¬ tial 10 in accordance with the sensed conditions. A few or a multitude of vehicle drive conditions can be sensed, and many sensors for these conditions are already available in the automotive art for automatic braking systems and load leveling systems. I prefer that sensors include axle accel¬ eration (S^) , axle weight distribution (Sw) , vehicle turning (ST) , vehicle velocity (Sy) _/ and engine braking (Sg) , although other conditions can also be sensed. Because a slipping wheel accelerates faster than a wheel retaining traction, acceleration sensors S 31 and S^ 32, applied respectively to axles 17 and 18, can detect loss of traction so that rollers 15 can be angled to supply more torque to the

wheel maintaining traction. When my CVD is used to divide drive torque between front and rear axles, the acceleration sensors can be applied to drive torque distribution shafts 17 and 18 to determine which pair of axles is slipping so as to direct more of the drive torque to the axle pair having better traction. Weight sensors Srø 33 and S-# 34, applied respectively to axles 17 and 18, can detect a shift in vehicle weight to the outside axle when the vehicle rounds a curve, and this can be used to angle rollers 15 to apply more of the differential torque to that axle. A similar effect can be achieved by a combination of turning sensor ST 35, sensing the turning angle of the front wheels, and velocity sensor Sy 36 sensing vehicle velocity, which is especially relevant during turning. Braking sensor Sβ 37 can detect engine braking, to slow the vehicle by use of the engine and drive train, so that if one wheel loses traction and slides during engine braking, rollers 15 can be angled to ensure that more of the torque is exerted by the non-sliding wheel. In four wheel drive vehicles, weight sensors can be used to determine the weight distribution on the front and rear wheels, which can be supplied with proportional drive torque by my CVD.

All of these sensors of vehicle drive conditions, and other sensors that are available or may become desirable, can input to control processor 40, which then has information on the vehicle drive conditions that are relevant to differ¬ ential torque distribution. Processor 40 is preferably a microprocessor programmed to respond appropriately to the inputs from all the sensors and to drive motor 25 via slip ring 24. This changes the interaxle speed ratio and thus varies the torque bias ratio of any axle differentiation so that the result suits the vehicle's operating circumstances.

The capability of my continuously variable differ¬ ential for distributing torque independently of the relative speeds of a pair of drive axles offers automotive engineers new design possibilities that have not been previously avail¬ able. Applying more torque to the faster rotating axle on the outside of a curve, for example, can be done to eliminate under steer moments and possibly improve steering

performance. Suspension systems, which have had ro accommo¬ date the torque bias ratios characteristic of differentials with fixed speed ratios of -1, can be changed to take advan¬ tage of the variable torque distribution that my invention makes possible. Vehicles with automatic braking systems may be able to use sensors that are already in place so that these can contribute to control of torque distribution as well as braking force. The many vehicles with an onboard computer, already performing some functions relative to vehicle drive conditions, may be able to expand this technology to control a continuously variable differential according to my invention. Four wheel drive vehicles can use three of my CVD's—one between each axle pair, and one to distribute torque between the axle pairs—allowing torque to be optimally applied to all four wheels, depending on traction, weight distribution, and vehicle turning.