Best differential on a wet track?

All else being equal (i.e. predictable, rain tires, optimum suspension setup, correct racing line)...

which diff is most suitable for wet conditions? Let's assume the whole track is wet and grip level is fairly consistent everywhere.

I've read Torsen diffs are smooth but i've also read that in some situations, these diffs may keep varying traction between two driven wheels too much back and forth and predictability may suffer.

I've also read somewhere that a quality clutch-type diff like the OS Giken is better in any situation, dry or wet.

Assuming we're putting a diff in a well sorted rwd car with good rain tires and a nice linear NA powerband (<300hp), which diff is ideal for wet track days?

I'm a big fan of a good clutch plate diff myself. They feel a bit more predictable to me, even though a torque biasing differential may be good for better lap times with a driver that knows how to use it. My cars don't have enough power for it to make much difference, especially in the dry, though.

That is a lot of assumptions. So why not assume you also have the perfect diff and be done with creating the perfect mythical car?

You could always get a clutched helical......yes, they do exist strangely enough...

Clutch type. Not a fan of the way Torsens lock up.

I really like worm-gear diffs, like the Eaton Truetrac or Torsen T-2.

An electronically controlled differential is actually best in that situation.

That said, a clutch/helical hybrid like OSGiken's is the best of both worlds.

The OS Giken is not a hybrid, it's a clutch pack with a LOT of clutches.

Another vote for clutch packs. Torsens and other non-loaded helicals have trouble in very low traction situations and can essentially turn into open diffs if one wheel is below a certain threshold. Think of their torque transfer as a multiplication, and if one wheel can put down 0 torque then there's no transfer. The idea behind a loaded helical is to get around this by providing enough internal friction to load up that inner wheel. it's the same idea as putting on the brakes a bit when you're offroading.

In reply to Keith Tanner:

Bingo, the only "loaded helical" as you put it that I know of is the AutoTech Wavetrac. Wavetrac

Several guys in our hillclimb group run wavetracs and really like them.

My RX7 has a clutch type, which is bagged out and pretty much useless.

In reply to Keith Tanner:

Nailed it.

When our suspension was setup poorly, the inner wheel would lift, and the Torsen would do nothing.

I understand dragging the handbrake 1 click is a good way to cheat your way through that.

I think a 1-way clutch-type LSD with a fairly mild setting performs best on track in any situation. A helical/wavetrac is a good compromise if you want to avoid the added maintenance of a clutch-type LSD.

GURU in Australia used to make a loaded helical - I've got one. It helps a bit, doesn't help a lot. Going from a GURU to an OS Giken on our little kart track was worth either one or two seconds a lap on my 175 rwhp Miata because I could get on the power so much earlier. The sharp corners on the track are enough to really unload the inner wheel.

I've found that there's a power limit on Torsens as well. Once you get past a certain power/weight ratio, you can spin the light wheel hard enough to get the Torsen into "open mode".

GameboyRMH wrote: a 1-way clutch-type LSD with a fairly mild setting performs best on track in any situation. A helical/wavetrac is a good compromise if you want to avoid the added maintenance of a clutch-type LSD.

I suppose a clutch-type diff performs better and is more predictable because on throttle during lock, equal torque goes to both driven wheels?

Can a "loaded" diff, like the wavetrac, have a bias ratio allowing 50/50 torque split during lock?

Keith Tanner wrote: GURU...loaded helical...It helps a bit, doesn't help a lot.

If power-to-weight ratio of a car is well within the capacity of a loaded diff, would an OS Giken clutch-type still have some advantage over a wavetrac unit because of the way it locks during off and on throttle? I'm basically interested in finding out which type of a diff would be the most predictable in wet conditions.

DaveEstey wrote: Several guys in our hillclimb group run wavetracs and really like them. My RX7 has a clutch type, which is bagged out and pretty much useless.

Is it useless because it's worn out?

Junkyard_Dog wrote: Not a fan of the way Torsens lock up.

What don't you like about the way a Torsen locks up?

Keep in mind the diff does other things than just keep both wheels accelerating. When clutch packs are 'sprung' for the dry, they will be too tight for the lower level of overall traction in the wet. This will tend to keep both wheels going the same speed on deceleration, and during turning.

If they are severely over-sprung, they will severely hurt your turn in.

Torsen will probably adjust itself to the situation better, while your clutch pack should probably be adjusted differently between the two scenarios for truly optimal performance.

In my case, I only go to the track when it rains :-) -- to conserve tires and to be competitive with more powerful 2wd cars. I've been focusing on increasing mechanical grip and now I'm trying to shop for a diff upgrade.

SSATB wrote:

GameboyRMH wrote: a 1-way clutch-type LSD with a fairly mild setting performs best on track in any situation. A helical/wavetrac is a good compromise if you want to avoid the added maintenance of a clutch-type LSD.

I suppose a clutch-type diff performs better and is more predictable because on throttle during lock, equal torque goes to both driven wheels? Can a "loaded" diff, like the wavetrac, have a bias ratio allowing 50/50 torque split during lock?

Keith Tanner wrote: GURU...loaded helical...It helps a bit, doesn't help a lot.

If power-to-weight ratio of a car is well within the capacity of a loaded diff, would an OS Giken clutch-type still have some advantage over a wavetrac unit because of the way it locks during off and on throttle? I'm basically interested in finding out which type of a diff would be the most predictable in wet conditions.

DaveEstey wrote: Several guys in our hillclimb group run wavetracs and really like them. My RX7 has a clutch type, which is bagged out and pretty much useless.

Is it useless because it's worn out?

Junkyard_Dog wrote: Not a fan of the way Torsens lock up.

What don't you like about the way a Torsen locks up?

Sorry, I don't have any experience with the Wavetrac. I've never found my Giken or the similar feeling Positraction in my V8 car to be unpredictable in the wet, but it's a fairly rare condition around here.

SSATB wrote:

GameboyRMH wrote: a 1-way clutch-type LSD with a fairly mild setting performs best on track in any situation. A helical/wavetrac is a good compromise if you want to avoid the added maintenance of a clutch-type LSD.

I suppose a clutch-type diff performs better and is more predictable because on throttle during lock, equal torque goes to both driven wheels. Can a "loaded" diff, like the wavetrac, have a bias ratio allowing 50/50 torque split during lock, consistently?

The clutch-type performs better because it can't go into "open mode" like a helical gear type can. And yes it can happen more easily than you might think, as Keith pointed out. The Wavetrac is far less likely to do this but it's not completely impossible.

Also a clutch-type LSD won't put equal torque to both driven wheels - only a locker or spool does that, or a clutch-type set extremely tight so that it acts like one. The "tightness" of a clutch-type LSD (set with springs and/or shims) is what determines how close you get to having spool-like performance on lockup.

A helical diff of any kind also won't put equal torque to both driven wheels - there are torque ratios involved...it's complicated...but they act very much like a 1-way clutch-type LSD with a mild setting, except that they can unlock if one wheel has very little traction (a "loaded" type like the Wavetrac is much harder to unlock this way).

This might have to do with my driving technique and/or how I set up cars, but I like to have the clutch pack set just "tight" enough to prevent one-wheel-peeling. I also think that unless you're on the power, open diff behavior is most desirable (lockup without power will only reduce the car's ability to rotate - I don't want that!) - thus the one-way diff.

Thanks everyone for the discussion. Very informative for me.

I've tried to set up my car for wet track days. I've removed both of my sways now. My dampers are valved with softened high speed rebound with just enough low-mid speed rebound and good mid-high speed compression. At this point I would want a diff that doesn't lock up suddenly to transfer power to the wheel with grip and has a low "blocking factor."

In reply to SSATB:

You're gonna swap your diff when it rains??

I don't get it- Do you live in Papua New Guinea?

The guy who is best prepared for the conditions is going to go faster!

In reply to GameboyRMH:

The even more screwy thing with helicals is that they sometimes have different ratios between manufacturers.......quaife is something like a 25-75 split and torsens are a 33-66 split iirc.

yamaha wrote: You could always get a clutched helical......yes, they do exist strangely enough...

S2000, no? And Wavetrac... and technically you could take any Torsen style diff and make the bellville washer stronger for more initial binding.

yamaha wrote: In reply to GameboyRMH: The even more screwy thing with helicals is that they sometimes have different ratios between manufacturers.......quaife is something like a 25-75 split and torsens are a 33-66 split iirc.

That's not quite how they get rated... it's a maximum torque bias.

Audi had high bias center diffs in the original 01As. After the B3 chassis they went to a lower bias diff, but apparently the difference wasn't in the gear angles, it was in the amount of preload in the Bellvilles.

SSATB wrote:

GameboyRMH wrote: a 1-way clutch-type LSD with a fairly mild setting performs best on track in any situation. A helical/wavetrac is a good compromise if you want to avoid the added maintenance of a clutch-type LSD.

I suppose a clutch-type diff performs better and is more predictable because on throttle during lock, equal torque goes to both driven wheels?

Opposite. if the diff is "locked up" then equal SPEED goes to both rear wheels. Torque split can be up to 100/0. Actually it can go negative in a tight enough corner, the outside wheel will be dragging (negative torque) and accelerating the inside wheel (additional torque over and above what the ring gear is giving)