LSD's...again

[AndyW]

Yes, I have pined for an LSD and waivered over which one to get for about 3 years now. So, let's just call it my annual winter LSD post and move forward.

Trying to decide just how much I want to spend and how much it is worth it to get an LSD. I have read numerous threads on the topic and I think I have it settled to between a Wavetrack and a true clutch type. The Wavetrack is tempting because it has the low maintenance benefits of a helical but doesn't have the "wheel lift" concern of the M-Factory. However, it seems a lot of people "move through" a helical to a true clutch type in short order on a track car. Don't really want to spend twice. Would appreciate any and all input. In fact, I would encourage a heated debate, with name-calling, if necessary...

[iminhell1]

For you, clutch. It's tunable. Don't need to say any more than that.

If you're lazy, torsen style. Personally I'm lazy.

[spidertri]

Went through the same thing a couple months ago. Was trying to convince myself that a torsen would be enough, but I knew that if I got the torsen I'd always wonder "what if I got the clutch type?"

So I ordered a 4 clutch unit from limitedslip.de and installed it myself. It only took one drive after the install to know it was there and working. Loving it so far at autox, track day in two weeks to really try it out.

[mad_hatter]

Here's everything I know about LSDs and factors to consider, from my own perspective. About my experience, I sold these for the last decade or so - albeit for muscle cars and pickups.

Helical LSD
This diff is the ideal unit for an application that sees mostly street and daily driving, and that's the reason so many manufacturers are using this as an OEM Fitment. It's completely maintenance free, 100% silent, and has excellent driving characteristics.

It's the friend who puts their arm around you after you've had a few too many at the pub and walks you home... It gets you there safely without you really knowing that it's helping you.

Both Quaife and Wavetrac use preload springs to mitigate the zero-load issue (wheel lift), and I'd be surprised if M-factory didn't, I think it's necessary to make the diff work - maybe theirs just isn't strong enough? The zero-load issue is a downside to the helical diff. Another downside is it will never fully lock the wheels; by design they're automatic and infinitely adjustable. That means it's never going to be 100% predictable, which is important if you're driving at the limit.

Clutch Type
This is the best choice for an application that won't be used as a daily, but rather as an occasional car, and in my opinion the only LSD for a RWD competition car.

The pros:
-Predicatable
-Tunable. *some units may not have variable ramps, but always with different clutch stacks.
-Will fully lock the rear axle

The cons:
-Requires maintenance to operate properly
-Can cause adverse handling, if you get the tuning wrong
-Can be chattery or noisy (however, this is a sign that something is wrong)

One thing to note: please, do not under any circumstance buy one of the retrofit kits that installs in your open diff with a couple preload springs. In the industry, those are known as "lunchbox lockers" and I've never heard of anyone having a good experience with it. It's not a question of "if" it will fail - but "when"... and then you need a whole new diff anyway.

As for my own 128 that I'm using for STX autocross: I've been in touch with MFactory about getting one of their clutch-type LSDs.

[bbnks2]

Originally Posted by mad_hatter

Here's everything I know about LSDs and factors to consider, from my own perspective. About my experience, I sold these for the last decade or so - albeit for muscle cars and pickups.

Helical LSD
This diff is the ideal unit for an application that sees mostly street and daily driving, and that's the reason so many manufacturers are using this as an OEM Fitment. It's completely maintenance free, 100% silent, and has excellent driving characteristics.

It's the friend who puts their arm around you after you've had a few too many at the pub and walks you home... It gets you there safely without you really knowing that it's helping you.

Both Quaife and Wavetrac use preload springs to mitigate the zero-load issue (wheel lift), and I'd be surprised if M-factory didn't, I think it's necessary to make the diff work - maybe theirs just isn't strong enough? The zero-load issue is a downside to the helical diff. Another downside is it will never fully lock the wheels; by design they're automatic and infinitely adjustable. That means it's never going to be 100% predictable, which is important if you're driving at the limit.

Clutch Type
This is the best choice for an application that won't be used as a daily, but rather as an occasional car, and in my opinion the only LSD for a RWD competition car.

The pros:
-Predicatable
-Tunable. *some units may not have variable ramps, but always with different clutch stacks.
-Will fully lock the rear axle

The cons:
-Requires maintenance to operate properly
-Can cause adverse handling, if you get the tuning wrong
-Can be chattery or noisy (however, this is a sign that something is wrong)

One thing to note: please, do not under any circumstance buy one of the retrofit kits that installs in your open diff with a couple preload springs. In the industry, those are known as "lunchbox lockers" and I've never heard of anyone having a good experience with it. It's not a question of "if" it will fail - but "when"... and then you need a whole new diff anyway.

As for my own 128 that I'm using for STX autocross: I've been in touch with MFactory about getting one of their clutch-type LSDs.

Wavetrac doesn't use "springs" to mitigate the no load situation. The diff relies on INERTIA of the spinning wheel. As the wheel spins up the internal "waves" climb each-other spreading apart a friction disc creating a locking affect. I have my doubts the system does much if anything... and not without some serious wheel spin first to provide the inertia. Not exactly a performance driving benefit. At low speeds and tight turns where the inside wheel might lift it seems slower to respond than the M3 viscolok and there are tons of videos out there showing how at low speeds any preload generated seems to be easily overcome and you get 1 wheel peel. Like going up a driveway or one wheel on ice.

I vote clutch type all day. I installed a wavetrac after having the E92 m3 diff and the wavetrac is definitely better for people afraid of breaking/burning out a clutch type diff. Drag/big power cars maybe. A clutch type, even the M3 diff, feels better for autox/roadcourse driving.

Maybe look at the OS Geiken diff?

[AndyW]

Originally Posted by iminhell1

For you, clutch. It's tunable. Don't need to say any more than that.

If you're lazy, torsen style. Personally I'm lazy.

Yeah, see, your second point is very valid. There's a reason I have a DCT...

[AndyW]

Originally Posted by bbnks2

I vote clutch type all day. I installed a wavetrac after having the E92 m3 diff and the wavetrac is definitely better for people afraid of breaking/burning out a clutch type diff. Drag/big power cars maybe. A clutch type, even the M3 diff, feels better for autox/roadcourse driving.

Maybe look at the OS Geiken diff?

I have, but my installer is not a fan of OS Giken due to some bad experiences previously. Plus, cost is a factor. I mean, is an LSD really worth the cost of a PS2 turbo in bang for the buck?

Edit: Just discovered that OS Giken has zero warranty apparently. At least at Diffsonline.

Edit 2: According to Diffsonline, OS Giken no longer makes a model for the DCT trans. So at least I know it's not an option now...

[bbnks2]

Originally Posted by AndyW

I have, but my installer is not a fan of OS Giken due to some bad experiences previously. Plus, cost is a factor. I mean, is an LSD really worth the cost of a PS2 turbo in bang for the buck?

Edit: Just discovered that OS Giken has zero warranty apparently. At least at Diffsonline.

Edit 2: According to Diffsonline, OS Giken no longer makes a model for the DCT trans. So at least I know it's not an option now...

Yeah I had read people break them on all platforms. In terms of being an affordable Clutch type diff that's still your best bet though.

Most other clutch type options will be 4k+ usd. Mfactory is a good option and a bit cheaper than that.

Don't get me wrong, a Wavetrac is still better than stock... but it's the cheapest option by far for a reason.

[TimCSquared]

I've driven both, and prefer the Helical. A Helical/Torsen/Quaife is called a ATB (Automatic Torque Biasing) Differential. Meaning it automatically biases torque to the driven wheels based on wheel transfer. How does this work? It has to do with the helix angle of the gears. Beyond a certain angle, helical gears cannot be reverse driven because the frictional forces from the axial force vector overcomes the radial force vector.

Essentially, from 50-50 Weight Transfer (sitting still), to 80-20 weight transfer, the diff will automatically bias the correct torque to both wheels. Fantastic! Once the weight transfer goes above this, the torque bias remains constant.(however, you can run a shear thickening fluid which "locks" the diff when excess speed differential occurs, which is neat)

A clutch type is less polite. It will have a set point where the torque begins to bias, a ramp up, and a maximum bias. Yes, this is tunable, and you can get it just right. However, when your alternative is a differential that automatically always has it right, why bother?

I will admit the clutch-type has the advantage at biases greater than a 80-20 split. Personally, I have never ran into an issue of needing a bias above this in the 1er. By the time you have that much weight transfer, mashing the gas is not advisable.

Ive heard many people complain that the torsen isn't adjustable and thus they didn't want it. Except that is the whole point: you don't need to adjust it, its already correct.

I am something of a suspension/handling purist. The torsen will prevent any uncomfortable handling discrepancies compared to an improperly tuned clutch type. That being said, if the clutch type is configured properly, no worries. But how do you validate that it is? What sort of test procedure is applicable? I'd love to hear how y'all clutch-diff-lovers tuned your diffs.

I installed a Blackline (Chinese knockoff) Helical LSD in my 128. Read more here:
https://www.1addicts.com/forums/show....php?t=1835048

[bbnks2]

Originally Posted by TimCSquared

I've driven both, and prefer the Helical. A Helical/Torsen/Quaife is called a ATB (Automatic Torque Biasing) Differential. Meaning it automatically biases torque to the driven wheels based on wheel transfer. How does this work? It has to do with the helix angle of the gears. Beyond a certain angle, helical gears cannot be reverse driven because the frictional forces from the axial force vector overcomes the radial force vector.

Essentially, from 50-50 Weight Transfer (sitting still), to 80-20 weight transfer, the diff will automatically bias the correct torque to both wheels. Fantastic! Once the weight transfer goes above this, the torque bias remains constant.(however, you can run a shear thickening fluid which "locks" the diff when excess speed differential occurs, which is neat)

A clutch type is less polite. It will have a set point where the torque begins to bias, a ramp up, and a maximum bias. Yes, this is tunable, and you can get it just right. However, when your alternative is a differential that automatically always has it right, why bother?

I will admit the clutch-type has the advantage at biases greater than a 80-20 split. Personally, I have never ran into an issue of needing a bias above this in the 1er. By the time you have that much weight transfer, mashing the gas is not advisable.

Ive heard many people complain that the torsen isn't adjustable and thus they didn't want it. Except that is the whole point: you don't need to adjust it, its already correct.

I am something of a suspension/handling purist. The torsen will prevent any uncomfortable handling discrepancies compared to an improperly tuned clutch type. That being said, if the clutch type is configured properly, no worries. But how do you validate that it is? What sort of test procedure is applicable? I'd love to hear how y'all clutch-diff-lovers tuned your diffs.

I installed a Blackline (Chinese knockoff) Helical LSD in my 128. Read more here:
https://www.1addicts.com/forums/show....php?t=1835048

Pretty sure the only type of diff that fluid thickness significantly impacts would be a viscous differential not a gear type ATB.

Not even the e92 m3 visco-lok is a viscous differential. It's a hydropump unit and the fluid you fill the diff with only really matters for the engagement of the clutches. More friction modifier can smooth out clutch engagement. A few stock cars do get viscous diffs from the factory but not BMWs.

A wavetrac and other ATB's don't have a "lock %" but the torque biasing of a wavetrac is said to be 2.5:1 which is equivalent to I think a 30% or 40% (at least that's what I have read) lock on acceleration. There is no decel lock. 30-40% lock on accel not being all that aggressive and most aftermarket clutch type diffs lock higher on accel. I think 30-40% is the range of most street type diffs. So yeah there is a lot different with a decent built clutch type diff especially when you consider that you can run a 1.5way and add a little decel lock.

[iminhell1]

... and the wheel off ground is easily fixed by not turning the stock EDiff off; with the torsen style.

I'll also mention that there is some tuning you can do via fluid viscosity with a torsen. But it's trial and error and heat dependent.

[AndyW]

You guys are the best.

I think I just figured out the correct metaphor for this (are we still allowed to have sense of humor in 2021?).

Clutch-type LSD is the crazy high maintenance super model.

Helical is the sane but pretty girl next door.

[joeo]

I've been using the wavetrac for a little over a year now; car is a track only car. In hindsight, I would get a clutch style.

I'm not sure if something is wrong with the wavetrac, and I can't be 100% sure that I'm doing a 'one wheel peel' since I'm not logging individual wheel speed. Folks behind me say that I'm spinning the inside wheel...and well - it's annoying not being able to put the power down.

Was there a difference over stock - absolutely. But then some dumbass went and upped the boost with MHD and now....


-Joe

[TimCSquared]

Originally Posted by bbnks2

Pretty sure the only type of diff that fluid thickness significantly impacts would be a viscous differential not a gear type ATB.

Not even the e92 m3 visco-lok is a viscous differential. It's a hydropump unit and the fluid you fill the diff with only really matters for the engagement of the clutches. More friction modifier can smooth out clutch engagement. A few stock cars do get viscous diffs from the factory but not BMWs.

A wavetrac and other ATB's don't have a "lock %" but the torque biasing of a wavetrac is said to be 2.5:1 which is equivalent to I think a 30% or 40% (at least that's what I have read) lock on acceleration. There is no decel lock. 30-40% lock on accel not being all that aggressive and most aftermarket clutch type diffs lock higher on accel. I think 30-40% is the range of most street type diffs. So yeah there is a lot different with a decent built clutch type diff especially when you consider that you can run a 1.5way and add a little decel lock.

Yeah, what I meant about the shear thickening fluid is something I am doing research on now. Essentially, running a non-newtonian fluid which shear-thickens can help in a torsen style differential when you get wheel lift. Consider: when a wheel is unloaded in a torsen (and you are on power) it beings to spin like crazy, which thickens the fluid. If the fluid can thicken enough, it will prevent the helical gears from rotating and lock the internals together. Interesting idea, right? In practice, the fluid never fully "hardens", and the shear-thickening ends up acting like a feedback loop for relative wheel speed differentials.

The diff I am used to is a 20%-80% split. I think most race-spec differentials are. This number can be adjusted in the design of the diff by changing the helix angle, so it is not surprising at all that different diffs would have various maximum biases. A more aggressive angle has other consequences in the design of the diff housing, I have yet to measure the the torque bias split of the Blackline I installed in my 128, but it is in all likelyhood less than a 20-80 split, being an "economy option".

What do you mean an ATB/Torsen doesn't have a decel lock? The helical gears work both directions. Many successful race-cars have applied their rear brakes through their differential in this manner, which will bias the torque as necessary.

To hopefully clear this up, I've attached a plot of how a helical style differential will apply bias between the two rear wheels. It is largely simplified, but you get the idea.

[1and1]

Hmmmm, I'm so close to clutch, but I dunno.

The car is reasonably aggressive, and likely to get more so with cdp's on the horizon (already here, waiting...) and their associated ability to play with stage 2 mhd ots tunes, and then likely to play with a bit of the corn.

And yet I just don't see myself in need of the same 60-90% that ya'll clutch enthusiasts do- My aggressive is the drive to work stupid early on empty roads... erm... In Mexico... The rest of the time it's driving home in traffic and weekend errands

Looks like a Wavetrac for me I think

yep

[lowside67]

Originally Posted by AndyW

Clutch-type LSD is the crazy high maintenance super model

I know it's just an analogy but I really do not think maintenance is a genuine concern with a clutch type. I had a Diffsonline ramped clutch style in my E36 which did 5 years of 4-driver autocross abuse (think 24-30 launches per day and more) plus many track and hillclimb events, including after I converted to a V8 with a dogbox with no issue whatsoever other than an annual fluid change.

Likewise, 2 years with my 128i with the Diffsonline build (which got a lot less use admittedly), still like new.

I believe strongly that the BMW works best with a ramped clutch style (ramps are important - otherwise with how much rear end traction you have, the car is going to push).

-Mark

[AndyW]

I love this group. Thank you all for your input...and for posting quickly. Diffsonline had a person back out of a build for my installer and had a "slot" open next week for a build so I had to make a decision today if I wanted to get the thing installed next month.

I have elected to go with the wavetrac. I agree the clutch-type is superior but I chose the wavetrac for the following reasons:

1. it is a "good enough" solution for my current skill level and a step up over stock.

2. It is $1700 less than the next clutch type option.

3. It has a lifetime warranty.

4. It is easier to resell if I end up upgrading to clutch type later.

5. Local race car builder (and my installer) likes them and has installed them on several PRO3 cars with success.

6. Less maintenance and not having to mess with tuning it to be perfect for my car. I admit it...I'm lazy when i can be.

I'll post experiences when the build gets done. I'll be having them do the M3 flywheel, berk resonated mids, wavetrac LSD, new guibo, new stiffer trans and LSD mounts, and all new associated hardware. Wish me luck!

[rac]

For me torsion style if the car is more important to you on daily driving and clutch type if more important on track.

I will probably never use a torsion style diff for track duty again if I can avoid it. Too much push understeer under power, with a fixed design im stuck with behaviour i dont like.

[iminhell1]

TBH I think you'll be happy no matter what you'd of gone with. In the situations you want a diff, you'll notice a big change for the better.


And like we've mentioned, fluids can change the way the diff acts slightly. My limited experience is with my MFactory unit and Redline shockproof fluids. I personally like the super light weight. It gives a little faster reaction than a thicker fluid. For sliding around and transitions on tight courses it's what I needed. Plus the lift off unlock isn't as aggressive/abrupt. But keep in mind that my main use is drifting on frozen lakes on studded tires. I'm no where near the norm.

[bbnks2]

Originally Posted by TimCSquared

Yeah, what I meant about the shear thickening fluid is something I am doing research on now. Essentially, running a non-newtonian fluid which shear-thickens can help in a torsen style differential when you get wheel lift. Consider: when a wheel is unloaded in a torsen (and you are on power) it beings to spin like crazy, which thickens the fluid. If the fluid can thicken enough, it will prevent the helical gears from rotating and lock the internals together. Interesting idea, right? In practice, the fluid never fully "hardens", and the shear-thickening ends up acting like a feedback loop for relative wheel speed differentials.

Seems like you are speaking conceptually on quite a few things.

A fluid that thickens with wheelspeed (like how the hydropump in the M3 diff locks up) would not work the same way with a helical, would it? I highly doubt the fluid would be able to keep the torque of a spinning wheel in check... unless it pretty much solidified... and even if it could it would be creating resistance inside the entire housing... basically like applying the brakes. So yeah finding a fluid like that and applying that to a helical? Never seen or heard of that being available or being done before and I don't see how it would even achieve the goal of locking up a single wheel vs slowing down the car. The housing of the diff would have to be divided at the very least.... basically, it seems this concept is best be applied in the form of the GKN visco-lok lol... not just dumping this type of fluid into a regular helical diff.

Originally Posted by TimCSquared

What do you mean an ATB/Torsen doesn't have a decel lock? The helical gears work both directions. Many successful race-cars have applied their rear brakes through their differential in this manner, which will bias the torque as necessary.

A wavetrac does not lock on decel. I don't think any of the other brand do either. The wavetrac again uses its "waveform" technology to provide a little bit of preload to cause a slight bit of lock on decel but like already mentioned this "waveform" tech does little to nothing. It's a ridiculously low amount of preload that is generated and basically non-existent at low speeds where the waves don't have enough velocity to ride up on each-other.

Is it something? sure. Is it as much decel lock as a true 1.5 way LSD or a 2way? No. Do you even NEED decel lock? That depends on how you're driving the care really. Just debating the differences between a helical and a clutch type and why you might want one or the other.

An Mfactory helical is cheap and gets the job done. Helical definitely feels like it's still leaving something on the table though but it is a big step up to move to a clutch type at 3-4x the cost. Hard to justify for most including myself.

[wootloops]

I see mentions of the MFactory helical diff.. but has anyone tried the MFactory "Plate" type LSD? I assume by "plate" this means it's a clutch-type LSD?

Price is not far off from their helical unit.

http://www.teammfactory.com/catalog/...te-lsd/BMW/E8X

[spidertri]

Originally Posted by wootloops

I see mentions of the MFactory helical diff.. but has anyone tried the MFactory "Plate" type LSD? I assume by "plate" this means it's a clutch-type LSD?

Price is not far off from their helical unit.

http://www.teammfactory.com/catalog/...te-lsd/BMW/E8X

I looked for reviews on the new version of that diff for months. Could never find anything definitive. Seems like some Honda guys are using it but not much from BMW applications.

I even emailed MF and they just told me that it has multiple ramp profiles machined into the pressure rings and that it's driveable on the street. I'm not sure what the 50% reduction in NVH claim meant because it's 50% of what? Videos I found of the V1 were incredibly noisy sounding so it really didn't give me a good feeling that the V2 would be quiet.

I can say that my 4 clutch diff is silent on the street. I didn't go for super aggressive ramp angles because I wanted it to be somewhat civilized and have a not as abrupt lockup. But it still locks very quickly when there is any unloaded wheel slip.