I searched this forum and could not get info regarding this subject.

Is there anyone that can explain in laymen s terms and in detail the difference, pros and cons regarding a normal LSD (Clutch plates) and a torsion LSD like the Quaife ATB parts?

Ok in very general terms - the ONE advantage the friction type lsds have is they will still provide drive with one wheel off the ground. The problem is they allow slippage, and in general that slippage isn't consistent. The springs on the clutch packs generally set the initial breakaway force but the angles of the spider gears will work against one another and increase the clamping under heavy enough load. What that means is you might get 1-2 revolutions where they provide X amount of slippage, and then after that they grab harder and allow maybe 1/2 or 1/4 as much. On dirt this is of zero consequence really, and it's handy to have drive when one wheel is off the ground (it's also good on boat ramps and stuff like that).

The torsen or ATB type diffs work on a mechanical principle (in its simplest terms a worm gear can drive a normal gear but not vice versa, so they allow 'movement' in one direction but not the other, combine that through further gear/linkups and you end up with the torsen). The one thing they won't do is provide drive if one wheel is totally off the ground. BUt for all other situations, if one wheel starts to slip, they will automatically (to the full extent they are able to which is typically somewhere from 3:1 to 5:1) route as much as 4/5th of the available torque to the wheel with more grip and 1/5th to the slipping wheel, and any ratio smaller than that. As each wheel gains and loses grip, they just automatically route the torque. It's done due to their internal workings, no lever has to move or react, it literally happens instantaneously.

On any road surface and car setup where the drive wheel won't leave the ground, they have by far the best traction off the line and out of corners. They are excellent in snow, and the way they direct the torque is so much quicker than any drivers re-action.

Friction type lsds tend to be 'sticky' when new, then wear to a slightly looser setting, then hover around that for their life (which can be less than a year in many cases). Of the available ones, the clutch plate ones are better - easier setup, longer lasting. The friction cone type generally wear quicker and don't lock as consistently.

You can also get friction type ones that have different lockup characteristics/harshness on deceleration vs acceleration. this is typically a good thing, as backing off on corner entry with a fully locked diff can produce sudden oversteer (not always, but possible) - with a softer/gentler decel lockup, a circuit racer can usually do better corner entry/mid corner which then sets it up for a better corner exit speed. Which is one of the most vital things in circuit racing.

The tor-sen never (appreciably) wears, as its action isn't based on friction. The gears can eventually tear up or even break given long enough or hard enough usage, but generally the 'go' is to make the unit strong enough to survive the application.

IMHO - for sedans and coupes with street suspension setups (i.e. not rocksolid heavy springs) a torsen is hard, very hard, to beat. On a racetrack - with a ute/bakkie, I recall one of the people here having issues with the inside rear lifting on hard corners, and a torsen will 'seem' like an open diff in that situation. On dirt or rallying, I'd go with the clutch plate type, and if I had to re-build it every season, so be it. They are (thankfully) usually pretty easy to rebuild. The problem might of course be parts availability and price. For the bw78 diff (larger than in some 120Ys etc - the one used on fords, larger nissans, holdens, valiants etc in AUS) the friction cones went out of production for a while. They are available again, it seems someone either found a source, or got them made themselves.. They are a simple enough diff to work on (if a little annoying since the centre won't lift out as a complete unit, and carrier spacing (for backlash mostly) is done by shims) but if you need to be swapping ratios, they are a pain, you either need a complete housing setup with the other ratio, or completely pull apart the existing one and fit new gears and re-set pinion depth, preload, and carrier backlash and preload.

Having said all that, there is also a friction type lsd, but it's fluid based. in place of the clutch packs (as such) there is a centre unit filled with a heavy thick(ish) silicone like fluid. Under normal circumstances it is of a certain viscosity, but if the two 'impellers' (not sure the precise term) inside the unit start turning at massively different speeds or directions, it aggravates this fluid and it thickens up like lightning and starts to lock up both axles to one another). It's usually called a 'viscous lsd' and not a bad invention at all, it just isn't particularly cheap. I haven't got major driving experience with them, but based on friends who have, they are a good option for an all rounder, almost having your cake and eating it too. I have no idea what sort of lifespan the centres have, and obviously if they started leaking, it'd be the end of their functionality, and possible that of the rest of the diff internals if it compromises the diff oil and it's ability to protect them.

jmac thanks for the info, going to read it for the 3rd time, maybe it will add more clarity, but from your explanation, it sounds like the Quaife ATB is the one for the Gymkhana boys if you can afford it.

Do you find yourself needing to kick the back end out quickly to get around tight turns as part of the gymhkana comps? If you do, the quaife/torsen type diffs give the 'best' traction and as such it's a little harder to get the back end to cut loose. With a more conventionally 'locked' diff, when you do a tight turn, one back wheel is trying to go slower than it needs to, and one is trying to go faster (since on a turn the outside wheel has to cover more distance and the inside wheel less) - which means even without power applied, there is some hopping or loss of grip, and it makes it easier to apply the power and kick the rear out.

If that is a big necessity, then it's possibly one place where a torsen isn't the go. On the other hand if it's about getting out of a corner as hard as you can (which means getting as much power pushing the car forward with as little as possible lost to what is needed laterally to keep the car turning (so to speak) - then the torsen is the way to go.

IF you 'really' need to throw the rear around quickly and harshly, you could make a _damn_ good argument for a fully locked (welded or solid/spool centre) diff - not the least of which would be because it's a hell of a lot cheaper to weld up a centre than to get _any_ of the adaptable lsd options.

After all that is said and done, you'd also have to face the prospect that each type of lsd would likely require differences in spring and shocker rates to get the most out of them. For instance with a tor-sen on surfaces where the wheel might leave the ground, you'd maybe compromise on rear suspension spring rates a little, and alter shock valving (probably rebound more so than compression) to try and keep the inside wheel down during a corner. A more solidly locked diff (more solidly when one wheel is up specifically) and you could run stiffer rears and still get out of the corner. And Of course, I am _massively_ oversimplifying things.

There's other small/cheap tricks you can sometimes do to help - even with a std diff. One of them is to assess the amount of leaf sag on the rear. Both sides will inevitably droop a little over time, but the right rear more so, since engine torque/rotation will always try and lift the right rear under hard acceleration *(obviously more so in first, just off the line, less so in higher gears). Now over the years, this will end up with less downforce on the right rear than the left. On some cars, the simple trick of swapping springs left to right countered that, and even 'pre-loaded' so that the right rear had more downforce/load whilst stationary, but at launch, as the right lifted, it ended up closer to even stevens, and the potential launch (and coming out of tight turns) was improved noticably. On setups where the springs aren't swappable left to right, then a small added block (like a thinner, longer (for more stability) version of a lowering block on the left rear is 'the go'.

Obviously other options include stuff like relocating the battery to the rear (as far back as possible for drag racing, and usually near or forward of the rear axle centreline for circuit racing, the former needing the most downforce/leverage advantage, the latter needing to avoid weight too far 'south' which negatively affects polar moment of inertia - turning responsiveness and stability/controllability)

We try to be as fluent as possible, that said, no wheel spin or at least controlled and we do not let the rear cut loose to get around a corner.

[QUOTE]On the other hand if it's about getting out of a corner as hard as you can (which means getting as much power pushing the car forward with as little as possible lost to what is needed laterally to keep the car turning (so to speak) - then the torsen is the way to go.[/QUOTE] - That is a 100% accurate statement when it comes to our Gymkhanas, all of which is run on tarmac.

http://datsun1200.com/modules/newbb/v ... hp?topic_id=43079&forum=1 - in that post is a youtube movie, it would give you a very good idea of what our gymkhanas is about.

The other interesting thing I noticed this weekend was that people with LSD and a lot of power; over power there vehicles out of corners, hopping the LSD will keep everything in check, but end up losing a lot of time.

Thanks for the good info, keep it coming.

Based on that video, it appears there is almost enough body roll to unload the inside rear. It'd be a tough call. (and when I say 'enough body roll' I don't mean to suggest it's excessive, or the car isn't well setup - what I mean to say is that it might be at or around the specific threshold of where a torsen would work or not).

Interestingly, a related topic came up on another forum, and this link was posted:

http://www.wavetrac.net/technical.htm

This is basically a combo of a tor-sen with an extra internal mechanism which will thrust/move when an axle is completely unloaded (and in danger of just spinning free, with no drive to the other wheel) and creates pseudo-resistance and gets the diff pushing drive through both wheels - kinda the best of both worlds. There is no doubt that such a device could be designed built and would work extremely well, but I don't know one way or the other whether this design produces the goods. Based on the info on that page, it _should_ work very well.