Let's talk about limited slip differentials
 

What's the truth behind all the different setups that are available and how can a person go about choosing the right one for their needs? Long ago I put together a clutch type lsd that had a higher than normal preload for drag racing. It performed exactly as I had hoped.

Now that I'm focusing on road course driving I'd like to know which direction to go. I'm sure others have questions about his as well since I've yet to get a straight answer when talking with people and searching gives me tons of answers but leaves me with even more questions. How does the factory torsen compare to a cheap RX7 clutch type lsd, compared to an aftermarket OS Giken, etc...and what are the merits to each one. Are there limits to them or is it going to be best regardless of the power level, etc...

FWIW when I was researching differentials for my LS1 Miata, most of the road racers (not drag racers) agreed that a Torsen type differential was worth the extra money over a clutch type differential.

After reading up on it a lot on several different boards, and talking to racers, the general order in preference (and in increasing cost, and in increasing torque ratio) seems to be:
Torsen T2
Allison TrueTrac
Torsen T2R

I went with an Allison TrueTrac (for a Ford 8.8 housing) which is very much like a Torsen. A lot of others used the Torsen T2 or the T2R. I could have saved about $200 if I had bought a clutch diff but I wanted to have good corner exit drive stability which as you can imagine is pretty key in an LS1 Miata. The corner exit behavior with the truetrac is great, totally controllable.

Now I know you are probably wondering about what can go in a stock Miata diff housing. What I am getting at above is the stock Miata TII Torsen should be great (at least over a clutch type) unless you are making too much power for one. People say the OSG differential is supermodel tits awesome on a road course, except for the price.

The question more is, what's your budget?

If you can afford it, get the OS Geiken and be happy forever.

If you can't well, just leave the Torsen in there. I've seen it used in a 350whp car with ok response. It makes the car a bit difficult to drive and much more tiring. The engagement is not as smooth and the lock/unlock points are not always quite what you expect.

That being said, the 250whp cars seem to have a blast with them.

That is interesting... opposite of what I'd lean towards purely from tracks have berms... get a wheel airborne and a Torsen looses its function there AND can cause some issues when getting traction again (impact on spider gears and boom?).

I'm going to clutch type rx7 (rebuilt) unit as my next LSD.

I ran a Tomei Traxx diff in my AE86. They are the same as a KAAZ unit, but seem to run cheaper in the market. They are setup out of the box with less aggressive lock up (more streetable), but if you switch the plates around you are right there with the KAAZ

changed from the torsen to a osgiken unit .. one of the best upgrades i have done for track use. now i am able to power out of corners much harder.

Torsens do not have spider gears so there is a lot more tooth engagement area than with a spider gear setup.

Another advantage of the torsen type is the bias ratio. That is, it can provide a greater torque ratio between the inside and outside tire. So in tighter corners it is less likely to cause understeer (push) when exiting the corner under throttle.

Also the transition from turning torque split (say 3:1) to straight torque split (1:1) is reportedly smoother with a torsen than a clutch type. That is it has better manners transitioning the torque split when turning into or out of a turn. Less 'chatter'.

And there is much less change in performance over time. A clutch diff will wear from day one and change its behavior, like a shock absorber. A torsen degradation curve is not nearly as steep.

I would rather have the three above characteristics than whatever the benefits are with a clutch type when curb jumping. That should be a relatively short period of time anyway if the suspension and driving is decent. I've done it and it helps get the car rotated, but I've used it more on really tight tracks personally. From what I have heard/read most going from clutch type (not OSG) to torsens for road racing are happy with the results.

Pros and cons to anything. Pick your poison. Keep in mind the comparison I am making is to a 'normal' type clutch diff, with spider gears and clutches/springs in a standard type carrier. The OSG (which also uses clutches) is much different than this. The OSG is really nice, but 3-4 times more expensive than a true-trac, and 6-7 times more expensive than a LSD.

Again I am not saying you can get a truetrac in your Miata housing. You have to do a diff swap, which for parts and hassle factor gets you up into OSG territory.

The torsen spins the inside wheel when there's little to no traction on the inside wheel. Like I state on my sig, I hate my torsen with a passion.

This was my mindset like two years ago. There's a reason I have a Torsen back in my car. Rx7 clutch LSDs are absolute trash, regardless of the preload.

When considering bang for the buck... still the same opinion? Did you ever run with one that was rebuilt?

I've never had a torsen in a miata but I have had them in 3 camaros that I've had and have also had the trutrak diff as well, comparing them to my RX-7 diff I would say that I find the RX-7 diff more stable, I always would find the torsen type diffs getting confused as it where when ever you have more than a moderate slip angle, it seemed like the power would always jump from one tire to the other making a slide more unpredictable than any other clutch type, locker or spool that I have ever driven. But maybe its just me, but I'd rather have a dinosour that just always does the same thing.

Forgot to say that my RX-7 unit is rebuilt, w/ oversized discs (tighter than stock). I think it all comes down to driving style.

I don't remember a Miata with that model name.

Interesting opinions. Does anyone feel there's a power level at which one setup would be preferred over another or even a certain type of track configuration that one particular setup would shine? Or do all these thoughts rule the world in every configuration?

Didn’t you have an OS Geiken in your car at one time?

Haven driven a very tight time attack on a go cart track with curbing on it I really feel that having more lock up in the diff with one wheel very lightly loaded should make me faster.

When making a hard turn on smooth pavement without the bump on the inside wheel however the torsen hooks up just fine.

Bob

I'm kind of curious, who is running a rear swaybar and having problems with the torsen?

At cal speed last year. He also had the Xidas on and a Fastback. It spoiled him!

Sav is.

I mention the sway bar because there is a really tight banked hairpin at my local track that was causing me to lift the inside rear tire. I would just spin that tire on corner exit. Pulling the rear sway bar solved it for me. I assume it helped increase droop travel and let the inside tire get some grip. Worked for me on that particular corner. I think some of the spec drivers found the same thing. May have zero relation to what Sav is seeing. YMMV

FWIW I have my rear bar off right now.

Sav is getting the same thing. That's the nature of Torsen, trutrac, etc.

I don't run a rear bar on the Volvo. I hate my car with one on. Apparently a lot people here shake their head at a lack of rear bar though.

Same opinion. Mine was preloaded to 110ft.lbs initially, no idea what it dropped off to after it "broke in". The car would NOT turn in, even with the RSB attached, and then it would spin the inside rear HORRIBLY on corner exit. This was all at 200whp. Less/more preload would have exasperated one problem or the other - having both problems at the same time quickly convinced me that the Rx7 diff is the wrong way to go. Kotomile has my Rx7 diff now, even despite my warnings about how shitty it was. He said he wanted something stronger than the Torsen, and the Rx7 diff is definitely that. I swapped back to a Torsen and all of those problems went away - nice predictable turn-in, nice smooth power application on exit.

At 270, the Torsen starts to get a little crappy, and it's really holding me back now that I'm at 350. I have the power to spin the inside rear even if it's not completely unloaded, and once it begins to spin I have to pedal it until everything hooks back up again. It's really bad at some tracks and not too bad at others. As Bob said, it's really bad when you are coping with bumps or pavement undulations, like exiting the Omega at WSIR or turn 11 at Thunderhill. For instance, Laguna Seca doesn't cause an issue, except in turn 11, since the track falls away from the apex to the trackout point. If I get into the throttle when I want to, the IR spins and I have to pedal it.

I borrowed an OSG from Emilio for RTA last year and it's absolutely magical. The car turns in properly, and it just hooks up and goes on exit. Even if you have a little slip angle on the way out, it will just haze both rear tires without even putting up a fight.

As far as RSB vs. no RSB, it helps a little on exit but my car needs the RSB to be balanced everywhere else - if I can't get the car rotated correctly there's no hope of getting away from the apex with any sort of drive.

Is a preloaded torsen possible? Is it possible to squeeze a small self-contained viscous or clutch unit inside the torsen and attach it at the center of the output gears? How much room is in there? It wouldn't have to provide for the full strength of limited slip, it would only be preloading the torsen unit. If it put 50 ft-lbs of preload on the torsen, then the torsen would treat a wheel in the air as having (0+wheel inertia+50 ft-lbs) of torque, and would use that to apply torque to the grounded wheel.

I wish I could tell ya how amazing it was to swap from a Torsen to an OSG.... but I went straight from a viscous to OSG - now THAT was a change. I'm only at 225 too, but I'll take Sav's word that I'll be glad I've got the OSG once I put in my built 1.8 at the end of the season. It's awesome fun coming out of #5 and #8 at Barber, very predictable.

I was searching for a diff after driving Johnwag's car with a Cusco. However since putting the AST dampers in I don't feel the need to buy a diff. I was getting on the gas just as soon as I did in John's 130whp car with no-more inside wheel spin when sucking-up the curbs.

It's my understanding that when you unload the inside wheel on a Torsen, it turns-into an open-diff. If the high-speed valving is right, the inside wheel stays on the pavement and never unloads the inside wheel, it won't spin.

I bet a real diff in my car would be epic.

That is what a Torsen R is. I believe they come stock on S2000's?
I blew up my stock '99 Torsen autocrossing it. Lifted the inside rear, got on the throttle, it spun up the wheel, and ate some teeth off of the worm gears when it landed. Switched to a Cusco Type RS. Turn-in suffered compared to the torsen, but exiting the turns is much better.

Quick note because I didn't know: Per Torsen's website, the Type A (Torsen T-1)differential is superior to the Type B (Torsen T-2) differential for getting power to the wheel with traction. IIRC, the Type A Torsens were found on 94-95 cars, and the Type B Torsens were equipped on 96+ Cars.

The Racemaster (Torsen T-2R) was designed to provide the higher bias found in the type A Invex gears while using the newer T-2 Equvex gear style.

Food for thought.

at 300ish whp, the torsen started acting up, at 400whp, it is useless, feels like an open diff. The osg is without doubt the best diff i have tried ever :) no chatter og clunking when driving on the streets, cant really tell there is a racing type diff in the car as it feels more like a stock mazda diff. Compared to lets say kaaz diffs, well.. it cant really be compared(kaaz is klunky and terrible all the time).
when accelerating, the car would wander from side to side as the torsen struggled to find grip all the way past 80-90mph. with the osg, it just feels planted, much more stable..
Turn in is also great. we are building another 400whp miata now. and were putting a osg unit in it from the start. no point even running the car with the torsen. Better of selling it before it breaks(like mine did)

I've been thinking about the Quaife LSD. Does anyone have experiences on the Quaife?
http://www.horsepowerfreaks.com/part...ip_A.T.B/16679
The price of the Quaife LSD is about 800 euros in Europe and the Os Giken LSD is about 1500 euros.

Was this on a stock Torsen(R) unit, or were you using an aftermarket gear bias differential (such as the Kaaz)?

It's strange to think that you would be breaking individual tires loose - I would expect that the coefficient of static friction between a tire/asphault is less than 2 1/2 times the coefficient of kinetic friction - especially a sticky race tire which tends to have a significantly less frightening transistion from "stick" to "slide"

I could see it rolling out of a corner

kaaz is not a gear bias diff, it's all clutch.

and it's not all that terrible on a street car. I've had mine a long time and it's only clunky if you're trying to zip through U turns. not anywhere else.

the biggest issue I have with it is how it straightens the car out when you get on the gas in turns.

On diffs…note that it can be misleading to just swap and choose the "best". For the best results given very different diffs, you should also tune your car setup and driving strategy a bit. Some that "promote" understeer will perform better if other things are also changed to accommodate that tendency. They can also affect brake setup as the nature of any rear lockup characteristics can become more or less pronounced. Which can be a big thing for those that like a lot of rear brake (retaining stability) as I do. On some cars I've setup, a diff swap could allow a few hundred more PSI at the prop valve's knee point. This stuff is all interrelated...

coming out of sharp, high grip 3rd gear corners on a track, using 225 R888 tires, it would spin the inside tire like posessed.. it simply was not able to transfer enough torque to the outside wheel to break it loose before the inside started spinning, atleast thats what it feels like. And then i had to get of the throttle a bit, find grip, and easy back on the throttle again, just like running a open diff :) This was when using the stock mazda torsen diff. When opening the diff to install the OSG unit, there was massive wear on the spider gears inside the diff, on the sides trying to transfer power to the outside wheel. This had happend in just a few months use (it did display the same issue from day one when running 400whp), and the unit had no problems at all when running lower whp settings. seems that there is a point where the stock torsen just is not up to the massive task it is being dealt.

i must sound like a OSG sales rep . but it absolutly transformed the car. :bigtu:

Seems like EVERYONE that's tried/owns an OSG diff sounds like a sales rep. Speaks volumes for how badass that diff is.

In case some missed the article…while the test used an S2000, the details are still interesting:

http://grassrootsmotorsports.com/articles/whats-diff/

Quaifes work similarly to a Torsen, get the inside tire light and it acts like an open diff.

Although…at least some Quaifes have a preload adjuster set via belleville washers…you have to open up the unit to get at them. I got a few Qs cheap after I helped them out with a lawsuit. Put one in an E30 M3 and another in a boosted Lotus Europa with nitrous antilag. I haven't compared a Q to an Giken but they were better than the usual BMW diff setup for racing. Those smear some understeer over the turn-in process whereas the Qs feel open at that time.

You'd think that, but two of the top CSP cars run a stock Torsen and RX7 diff. I know its not track, but diff problems seem to exaggerated in auto-x. I've driven two different cars with OS Gikens (a CSP Miata and X Prepared S2000) and I couldn't tell you there was difference between it and the Tomei diff I used in my old Corolla (AE86) in terms of performance. Maybe its just because I'm cheap, but I think I'll stick with a Tomei/Kaaz diff when the time comes.

Maybe those two CSP cars had the "whole package" tuned to work well together…quite often, LSD swaps need setup changes to max out potential.

Mine's only clunky when cold and quiets down after a couple of minutes driving. I do change the oil religiously though. Importing those cans cost me a fortune every time.

Speaking of Quaife, AWR and Flyin no longer carry the gearsets for the 5spds... new US distributor, bad sales, Quaife stop making them? I know it was crazy expensive - but the wideratio was temping with the 5spd. - not trying to change the topic, but I'll be needing something to handle 400rwhp (I'm serious) - been researching a various swaps - finding something with proper shifter location being a problem.

My OSG should not be a problem at that power level.

My KAAZ is soooo loud, even with the KAAZ diff fluid. Could have something to do with the solid diff mounts, idk.

Read my posts here about disconnecting the rear sway bar.

http://forum.miata.net/vb/showthread...=362012&page=3

Maybe you have an advantage you haven't fully explored yet. This discussion is about Lotus'. Perhaps it's applicable, perhaps not?:

"The Elise was always designed from the outset to work without an LSD. We have recently started to offer an LSD as an option on the Toyota engined cars, primarily in response to market demand from the Autocross enthusists in the USA, who need one to be competitive when accelerating away at full throttle from very slow, tight corners in first or second gear.

In this type of competition they do not tend to run high speed (100mph +) corners and therefore the increase in understeer on this type of corner which you get with an LSD is of little negative consequence to them and they therefore are better off with an LSD.

In our experience an Elise or Exige equipped with an LSD is at a disadvantage to one without an LSD on a typical European race track. On top of that the LSD bluntens the steering feel and respsonse of the car which we don't like.

If you want an LSD then by all means fit one, but please understand that there are negative as well as positive effects. In the instance you describe, instead of spinning the inside wheel as you accelerate away at full throttle (which can be easily fixed by modulating the throttle! Very Happy ) a car with a 2:1 LSD like the obne we supply will provide twice the torque to the outside wheel as it does to the inside one. This will increase the slip angle of the outside wheel and the car will tend to oversteer heavily on corner exit, requiring a reduction in throttle if you are not to spin.... It isn't much quicker, honest.

The optional diff we supply is a Torsen unit, the aftermarket unit supplied by Motorsport is a plate diff, with a similar 2:1 bias and no preload so in terms of action it mimics the Torsen closely."

By the way, I datalogged the bejesus out of many Elises and my own…with all sorts of drivers from myself, to noobs, to Aliens, half a dozen National Champs, etc. An interesting thing about the open diff on that car…although it is slightly slower than the LSD version, it feels better as Lotus notes, and this can be measured in the steering data too, you can see it in the various driver inputs. When you spin up an inside rear on an Elise (mostly RR), you get some "extra" acceleration back as it spins back down, lol. Another thing about autocross is that you tend to use / need larger front sway bars than other uses require. In racing it does not matter if a part is 1 second or 0.001 quicker since it is competition. The LSD does not do much to help the Elise, nothing like what it does for pony cars for example.

No, there is no room inside a torsen. We use one on the FSAE car because of the very tight turns we have.

Also to note, you won't have to service a Torsen, unlike a clutch pack. (No clutch plates to wear).

I would say Torsen style for the street, clutch for the track.

The Torsen T2R is preloaded
http://www.torsen.com/products/T-2R.htm

By combining the patented Equvex gearing from the Torsen T-2 with a coil spring pre-load and/or friction enhancing plates, the Torsen T-2R can remain locked, providing torque to both wheels, under low torque inputs. The addition of the friction plates allows the Torque Bias Ratio (TBR) to be tailored to differential design needs.

bullshit the torsen doesnt need service. it wears as much as anything else, even more, with the proper amount of power applied to it :D
OSG for street and track : its better at everything, period

True the torsen does have wear parts, they look something like the conical springs in a clutch type but are much smaller and there is one for each of the helical gears.

Yeah sure everything wears. But if you have a properly set up clutch pack, it will need servicing far before a torsen.

P.S - the T2R is slick.

Has anyone here tried the S2K diff swap in there miata? I thought about it but couldn't find the stub shafts that you need out of the S2K cheap enough so I just went the RX7 route. Those Honda dont let parts go cheap.

my torsen lasted less than 3months at 400whp and track use. The helical gears were the parts that wore out.. the osgiken has lasted twice that already :) I would say that the giken unit is properly set up too.

no point in arguing, osgiken diff= best diff for our miatas :P Please find someone that has tested one, that disagrees .

hf-mx5t tracks his car a bit so I would def tend to believe what this guy is spouting.

Has anyone every used a Carbonetic clutch type diff? I know they make one for the S2000 so you could use it in an MSM. Might be an interesting comparison to the OSG.
I don't think Wavetrac makes a miata LSD do they?