There are four plates on each side. One type has teeth on the ID to engage the axle, the other has four ear tabs on the outside to engage the LS carrier. Stack them outer-outer-inner-inner gives you just one rough to smooth surface and is the stock configuration, about 25% power transfer. Alternate them outer-inner-outer-inner gives you three, and about 60% power transfer.
The latter is the "tight" configuration you asked about. Measure the stack height and add shims to get some compression for 70-75% transfer. It's real artsy and you should rig a test set up if you want something repeatable.
Note that anything off the stock configuration is not really useful for street driving. The plates will wear like mad on normal street corners in a high ratio, and are hard to steer because at slow speed of a normal intersection there is little slip at all at the LSD acts a bit like a solid axle.
The only place you will see a notable difference is on a fast freeway on ramp circle. High ratios are too much for autocross, but around 50% will do well. Fast track cars on fast tracks with big turns are the ones that benefit most from tighter ratios.
Alternate them outer-inner-outer-inner gives you three, and about 60% power transfer.
You mean to add one per side between the two (with the ears) ??
Thats the most logical for me. Yes the Diff is for track use. I'd like to play with diff configuration. I think is the only way to learn how it works.
In my LSD, I had four plates on each side as original. There were two of each type,arranged so only one pair of faces were driven separately: AA-BB, only the one A-B interface slipped. I just re-stacked them like shuffling cards to ABAB so there were three slip interfaces : A-B, B-A, A-B.
You can also arrange them for two slip faces: ABBA, so the A-B and the B-A faces slip. You get a 40-45% ratio. Or for cheaper parts, AABA, which slips on both sides of the one moly plate (but it wears faster). This gives you a bit less ratio than ABBA.
The critical thing in making the LSD work as you want is how closely they fit. The outer cover on one side sets the total spacing for all the plates and the central spur gears. Imagine if you put an entire extra plate in there so that the stack was so high that that bolt-on cover didn't fit - as you tighten the bolts, the stack of slip plates gets compressed really tightly. Tight enough and the diff is essentially locked in place and both axles turn as one.
A little less tightness and the plates slip. So you can adjust the real lock ratio by adding a shim of some thickness under the cover or around one of the plates. And the tighter you make it, the more wear you will get. BTW - the LSD actually is meant to works with a bit of looseness in the plate spacing. Gear lube gets in there and affects the stiction because of its shear viscosity.
Of course, the gear lube is important too. The rough looking plates are nodular moly steel just like the syncro rings in the TX. The right gear lube is very important in getting good function in both. Fortunately, the same lube works well in the LSD diff as in the moly syncro TX. I use a single wt 90 EP (the Ep is for moly syncros). Some racers use Synthetics like Mobile 1 or Red-Line products. Ask your fellow racers.
In a race car, the diff is fairly easy to take apart; experiment with one, two and three sliding surfaces on each side, and with various shims . The only real measure is how your lap times are.
I remove two shiny cross (1,8) plates and I put two rough plates. After I reassembe it, I clamp the LSD in a vise with the top cover up. I tight the top 8 bolds with 3 kgms. With another modified stub I tryied to turn with a torque wrench. Start turning when the reading on the wrench was 6 kgm. . You know if is that correct or close to correct ?? I didn't lube it yet. I'll try to do the same test with wet DIFF tomorow to see the readings.
But like you said the only real way to measure is the lap time.
You've got it right. The torque dry is pretty meaningless. The LSD works fine even when it feels like there is no drag at all on the bench when lubed well. Much of the power transfer is influenced by the dynamic stiction of the lube and moving surfaces. BTW - you are best off to lube each part before assembly.
If you want a good baseline, assemble the diff in each of the configurations I mentioned above (including the stock one), well lubed. Then measure the break away torque. It's hard to make this measurement consistently, so be very careful about accurate procedure (i.e. do it just the same every time).
If you have, or can get/make some other shim thicknesses, you can do the same for various pre-loads. Then you'll have something to test on the track, and have a rebuildable configuration for the race season.