Testing Differential for proper lockup- Am I thinking properly?

[phenryiv1]

I am working on helping a buddy prep his 98 M3/4/5 for rallyX use. He just purchased a rebuilt 3-clutch LSD with a 40% lockup. I was checking his current (stock, 25%) LSD for function and I think that the clutches may be failing.

I had 1 rear wheel up in the air while we were test-fitting wheels (rally wheels are typically 15" so we were trying some different 15s on the rear for caliper clearance) and when I went to replace the wheels I was surprised to find that when I was torquing the lugs (again, with the opposite wheel on the ground) the rear wheel turned with <85 ft/lb. on the torque wrench.

On my e30 (with a 3.73S diff) I can easily torque the lugs on one side to 85 with the opposite side on the ground.

Does the fact that the raised wheel spun with less than 85 ft/lb. on the torque wrench indicate that the LSD in his M3 is failing?

[xlDooM]

Was the car in gear?

[phenryiv1]

Crap...never thought about it. PROBABLY yes but I would not swear to it.

Assume it was in gear (I will repeat the test next weekend when we try some additional wheels) - would that be a valid test of LSD function?

[xlDooM]

I don't know, I have no idea how strong the clutches are supposed to be, whether you can overpower them with the leverage offered by the torque wrench (+ the bolt circle radius) or not.

- - - Updated - - -

It's an interesting question. An M52B28 has 280Nm of max torque. This is multiplied by the gear ratio, and first gear has a ratio of 4.2, so that's about 1200Nm. 85lbft is 115Nm, so if you add the 6cm bolt circle radius that's 122Nm. Assuming that "25% locking" means 1.75ish torque bias ratio (I found a curve on the internet relating the two) means that of your total 1200Nm, the diff must be able to output 430ish to one wheel and 770ish to the other? Then the clutches should be way stronger than you can overcome with your wrench?

Another way to think about it would be that if you apply 122Nm to one side, the clutch must be strong enough to output 214Nm to the other wheel, and it should definitely be able to do that considering it takes over a 1000Nm when you give it the beans in first.

I'm not sure if this reasoning is correct, I'm an electronics engineer so my knowledge of mechanics is limited. I think Moron95M3 or what's his name is a mech engi?

[spyderg0d]

Just for giggles and coinciding with your good question. I went out and triedine as you did. I got maybe a inch of wiggle from the wheel and complete lock.. I couldnt get it to slip with well over my weight.. So.. I think your clutches are worn.

[flyfishvt]

Stock LSD comes with 2 clutches. 25% lockup. People add more clutches to increase lockup. It's surprisingly easy. If you have more than stock lockup I bet the person who modified it left the original clutches in place and the original ones are worn.

[phenryiv1]

Was the car in gear?

Now that I think about it, the car being in gear should not matter. The differential should lock the halfshafts independent of the driveshaft movement.

[RockinScott]

Now that I think about it, the car being in gear should not matter. The differential should lock the halfshafts independent of the driveshaft movement.

^^^^ This.

You test the lock up with one wheel fixed (or on the ground) and the other attached to a torque wrench (the pendulum kind). I have always been told that a 25% should break away at no less than 50lb. Have to make sure your torque wrench is centered on the hub which takes a rig of some kind. Doing it off of just one lug nut messes up the math quite a bit.

[spyderg0d]

Y'know what... I think i have like 75% lock on mine. I used a torque wrench centered on a puller on my wheel and it did slip once i got to around 140 ftlbs. And now i remember, once in a while, when i take a sharp turn from a stop like at a stop sign, it feels like a rear wheel fights or almost like a welded diff for a split second then works into the turn. Odd. Maybe the auto i pulled it from was in the yard from a drift accidant.