Salisbury differential - Preload

[Silente]

Hi all.

This year our car will have a salisbury type limited slip differential, with ramp and friction plates.

We know how to manage Locking Effect changing ramp angles and friction faces actually working.

If i am not wrong, what it actually happens is that when wheels are rotating at different speeds, a certain torque is moved from the faster one to the slower one. This torque is proportional to the torque the gearbox is giving to differential case by the Locking Effect.

What i have seen is that, for example, when accelereting out of corner with a godd lateral acceleration, the two driven wheels have nearly the same speed since the differential is locked and they are experiencing different torques, because of the locking effect (anyone can correct this if it is wrong).

What i would like to ask is what is the effect of preload? If i give more preload to tha package, what should happen inside the differential and what changes i should expect on car balance?

[Drew Price]

Increasing preload will require a greater rotational difference in the wheel speeds to allow them to start them rotating against each other.

More preload will increase the total locking 'force' forcing the wheels to spin the same speed, but it will also have more of a tendancy to make the rear wheels feel like they are 'locked' together at neutral throttle or light throttle.

If you have zero pre-load, the car will feel like it has an open diff at neutral throttle.

If you have infinite pre-load if will feel like you have a spool, and the wheels are coupled right to each other.

Adding some pre-load will give you some corner-entry understeer (rear wheels don't want to spin different speeds) but will give you additional lockup for corner-exit acceleration, which may even translate as extra corner-exit oversteer on throttle application.

Increasing ramp angle or clutch mu will give you more lockup during acceleration without the (probably) undesireable corner-entry understeer.


I feel like I explained that kind of badly, but have to run back out to work. Did I get it all right?

Best,
Drew

[Richard Pare]

Preload controls the minimum locking rate - the torque that it takes to spin one wheel reletive to the other - when no engine load is applied to the ramps.

In turn, it will affect the balance of the car when cornering in a neutral throttle situation, and will also affect the "smoothness" of the transition from off-throttle to on-throttle.

For these cars, you probably want only maybe 20 ft.lbs of breakaway torque (cold bench test). Less is obtainable, but the diff needs to be designed to got that low a number without getting the clutches too loose.

By loose, I'm referring to the ability of the clutch and ramp package to accellerate sideways a few thou' before the clutches start to compress against each other. If too much free play is available, it will result in a hammering effect as you get on the throttle and cause a big spike in the locking action, in turn causing push.

The best way to set up the usual design is to use shims to fill up any free space between the clutch/ramp package and the end walls of the diff. Just add shims until the preload hits maybe 10 ft.lb.

Preloading after that should be set using springs, and either coil springs or bellevilles can be used.

Coil springs have the advantage that they will not lose much tension at all when internal clearances open up a few thou', but their disadvantage is that most designs do not have the space for them.

Bellevilles don't take up much space, but will relax a lot with every thou' of wear of the clutches, causing a rapid loss of preload reletive to what will happen with coil springs.

Most manufacturers use one big diameter belleville, but personally I do not like that setup and instead use a circle of (usually) 12 small diameter bellevilles set into machined pockets.

The advantage there is that most small bellevilles come in quite a few thicknesses, as well as 2 material moduluses, so you have a pretty good choice in being able to set the preload that you want. On our small diffs, I use .5" diameter x .018" thick bellevilles (when I don't have the room for coils), and set them in .023" or so deep pockets. That arrangement usually get me about 30-35 ft.lbs of breakaway torque. If I want softer, I can either switch to the stainless versions, or take a few out. If I want more, I add arbor shims under the belleville, and can usually get up around 60-70 breakaway. If I need more than that, I'll use some thicker bellevilles.

In all cases, wear of the clutches will control how fast the preload changes, and most manufacturers do not make their clutches hard enough to resist that wear. If you find that the wear is to fast for your taste, just have them nitrided (radius all sharp corners first!).

With nitriding, the clutches will last just about forever. However, be aware of the fact that there will be an initial "bedding-in" whereas all the microscopic high spots are worn off, causing a relaxation of the springs, and if bellevilles are used, it will show up as a significant loss of preload You can reduce this substantially just by polishing them with a powered scotchbrite pad or wet/dry paper. You could also have them REM'd, but that may be a bit overkill.

Unless your locking rates from the ramps are also very low (80-85 or so degree ramps), the few ft.lbs of preload we are talking about will not affect the car very much since it will be only a small portion of the overall locking rate.

[billywight]

The only diff type that cares about wheel speed is a viscous coupling. All others (salisbury included) react to torque imbalance. Speed is irrelevant.

Take a few minutes to think about preload, then ask yourself what would the optimum differential be? In my mind it is completely free off throttle, then progressively locks as the tyres see a torque imbalance. Why would you want any preload at all? Keep in mind that with a salisbury the torque imbalance will be proportional to throttle imput and lateral acceleration.

[scotty young Taylor Race]

Well put Richard..

We use the small bellville washers to adjust pre-loads as well.

We will not go any thinner than .005 on the shims for final set up though. It will spit em out in a heart beat.

scotty
Taylor Race Engineering

[Richard Pare]

I use .001" shims regularly without any problems - the shims cannot be against rotating surfaces.I've also used sheets of teflon between the clutches when the rules didn't allow different(lighter) ramps but allowed shimming. Never thought that they would last at all, but they did!

The only issue sometimes is that if put against the bellevilles instead of the other end of the diff, is that the edges of the ID hole of the bellevilles can cut thru the shim, but I've only seen that when the edges of the hole were sharp instead of radiused as they should be.

Another tuning tip is that if your ramp gives too much locking both on and off the corners and you don't have a less agressive ramp, you can rearrange the clutches to put like clutches together, thereby decreasing the number of active surfaces and decreasing the locking rate.

Corner entry ramp angles : Lighter is not always better - it depends on what the car needs. A trick in the old Buick-powered Indy Lights was to go heavy on the off-throttle side to help stabilize the car under braking, especially on street courses - those cars were so ass-end heavy that if you didn't do this, the car was very prone to get away from you under heavy braking.