Big LSD Shootout... OS Giken, Drexler, Titan, GKN, ZF...

[okusa]

Originally Posted by italyix

Will be a while before I get it installed, and will post results when I do.

Not acceptable! Get that thing in there and post a full review NOW!!!!

[driftflo]

as i get more and more questions concerning the difference between "my" lsd core and a drexler core i took some pictures and try to explain here...

the lower left is an old ZF OE lsd core.
the lower right is a modified ZF unit to fit E92/E46/E60 M applications.
the top right one is a heavily modified unit with improved oilflow, lightweight and more space inside the casing for additional clutches.
the top left one is a drexler bmw motorsports case.



here are the three e92 units in detail again:




concerning performance it always only depends on the setup. with the modified units you can improve/upgrade the old zf cores to fit even racing needs. i can put any setup you want in any case you prefer. the internals of all four units are compatible and can be exchanged.

i think i wrote it already somewhere: the performance depends on the setup, not on the core used!

some setups can only be built with the two upper units as they offer more inside space/depth.
but the best value for money is the bottom right unit which can still host a very good racing setup that fits all traction needs while still reacting smooth and predictable... and will last one or even more seasons of racing. for non-competition environments it will even last several years before a rebuild is required. i would say this unit can last over 50k miles without feeling a letdown of performance. and it will still lock quite ok beyond that mileage.

so i hope i could clarify this a little...

[okusa]

Originally Posted by driftflo

so i hope i could clarify this a little...

PM'd you a couple of days ago.

[driftflo]

pns are out now, sorry for the delay. i was quite busy the last few days.

[OM VT3]

Originally Posted by driftflo

pns are out now, sorry for the delay. i was quite busy the last few days.

I also emailed you

[kevinlevrone]

Hello,

I am so glad I found this thread, it is probably one of the most informative threads on limited slip differentials that I have encountered until now. With a lot of first-hand information on clutch-based LSDs. Thank you driftflo and the other participants for the useful information.

This being said, I have a 335i 6MT with N54 engine, quite heavily tuned (420+hp, 600Nm). I have M3 front suspension bits which increased the front camber (reduced understeer), also Eibach sway bars that also reduce the factory understeer. After I used a Drexler LSD for 4 years, 50000 km, the rear axle broke inside (not sure what but it made a clicking sound while moving, after a loud "pop" was heard inside). What I know from Drexler is that the diff is fine but the axle pinions are done. Thankfully a second hand axle is not very expensive in Germany.

Ok, so my question now: until now I ran the Drexler LSD with 50/40 ramp angles (the ramps also have the option of 45/30 or 30/45), so more lockup on deceleration than on acceleration. The car was perfectly stable with perfect road manners. But I was never truly happy with the forward traction and traction out of corners (I am using Michelin 265/35/18 A/S 3 rear tires so this may be a factor ).

Driftflo finally answered my email and recommended a switch from 50/40 degrees to 40/50 degrees which can be done easily by flipping the ramps. I did pass this request to Drexler and they will change it when assembling the axle. But I'm having some second thoughts now.

Now my question is, what could be the downside of the 40/50 setup versus the 50/40 setup ? Is 40 degrees too aggressive ? What lockup percentage would this translate to ? I use my car as a daily driver and I woudn't want an unstable car in the rain or snow or some unforseen conditions. But I do want more traction in acceleration.

Thanks for any input on this !

[driftflo]

Originally Posted by kevinlevrone

Now my question is, what could be the downside of the 40/50 setup versus the 50/40 setup ?

this is only a small tweak, not a whole new feeling! you will have a little more lockup on power (giving you better traction out of a corner) and a little less lockup on coast (leading to a less stable rear end while braking... but remember that with the oe m3 lsd or the open 335i rear end you have no stability while braking at all).

Originally Posted by kevinlevrone

Is 40 degrees too aggressive ?

40° is quite ok still (steep enough to give you quick and smooth reaction of the lsd). if you go even more shallow to the 30/45° option it gets aggressive/rough.

Originally Posted by kevinlevrone

What lockup percentage would this translate to ?

depends on the installed/activated clutchpacks. if you have four, 40° will lead to ~50% lockup and the 50° ramp will do ~40% lockup (given the oe drexler clutches are used). if you have 6 clutchpacks it will be 75% for the 40° and 55% for the 50°.

Originally Posted by kevinlevrone

I use my car as a daily driver and I woudn't want an unstable car in the rain or snow or some unforseen conditions. But I do want more traction in acceleration.

the more lockup the more the car will be steerable via the throttle. if you are used to this and have enough driving skills you will love it. if not, you may fail in slippy conditions with dsc off.

[kevinlevrone]

Originally Posted by driftflo

depends on the installed/activated clutchpacks. if you have four, 40° will lead to ~50% lockup and the 50° ramp will do ~40% lockup (given the oe drexler clutches are used). if you have 6 clutchpacks it will be 75% for the 40° and 55% for the 50°.

...

the more lockup the more the car will be steerable via the throttle. if you are used to this and have enough driving skills you will love it. if not, you may fail in slippy conditions with dsc off.

Thanks a lot ! It seems that your recommended setup is exactly what I need. My car wasn't very steerable via the throttle and I wished more of that feeling. I have 4 clutchpacks (and Drexler seems reluctant to change to 6 as you recommended so I think I will leave it at that) and 50% lockup does not seem high at all, in fact I think I need at least this amount of lockup for my engine's torque values. I was somewhat scared because I read back in the thread that according to your estimations for the Drexler diff, 40 degrees meant 85% and 50 meant 70%. But this was probably with a lot more clutch packs (my impression was that more clutch packs only means smoother reaction, not more locking).

[vagmag]

So for an m5 e39 you propose the same zf lsd unit modified by yourself with 4 clutches and 45% on acceleration and 45% on deceleration . Correct me if I am wrong .
I like street drifting but occasionally also driving and drifting on track but for fun ...
I like as mentioned above the car to steerable via the throttle.

Questions :
A) for the above mentioned, what is your recommendation ?
B) what is your favorite setup with zf and why ?
C) maybe your answer on b) will be enough but , why not to use 6 clutches ? Or make it lightweight? Is it the cost , does it change the perormance ?
Thank you

[driftflo]

A) at least 45% with medium preload or even a higher lockup like 65% would be quite good!

B) in general the best would be as much clutchpacks and as steep ramps as possible.
the maximum you can get with the zf is 6 clutchpacks and according to the locking value you want to have you (or i) need to choose the ramp angles required. i have super steep ramps here from 85°, 80°, 75°, 70°, 65°... and down to 30°.

C) it always depends on your focus and budget. most people want a well working lsd that locks well, has not to be rebuilt too quickly and (mostly) is still cost effective.
if you don't mind the costs, i would suggest a drexler "light" core with 8 clutchpacks (and using the drilled z4gt3 clutches with improved oilflow for an even smoother lock/unlock transition) and very steep ramps! this will cost double the money of the 45% zf lsd but is a little lighter, lasts even longer, even smoother transition, ...
on the other side i doubt that you will really feel a difference in street-driving between these two units as the above mentioned advantages are so marginal that you really need to squeeze the car out to its max to feel the difference. i guess a racedriver driving these two units in a z4 gt3 will possibly be able to say which unit is which, but hardly anybody else!

and even the drexler "light" core could be modified to be even lighter (i have a prototype pic of such a unit and know it is working), i could reduce the friction of the gears via special bearings and so on... but no one will pay what it would cost :-(

but as i said, i'd love to build such a "high end street lsd unit". so if you want... lets go ahead ;-)

i'll put the stuff together lateron and post a pic of the items...

[driftflo]

here is a picture of these items.

on the left side you see two drilled (highend) clutchpacks, on the right side two grooved ones (the zf oe clutches had no holes or grooves at all, i don't use them).

in the middle is a 70/70° rampset which would lead to ~40% lockup combined with 8 clutchpacks. a pretty perfect street unit with hardly any downsides at all (except price).

although for a "fun/sideways" setup like you are aiming for i would go a little more shallow to have a higher lockup.

[vagmag]

Well I really would like a high end solution , but to be fair I would not be in position either to understand or pay for it .

To come back to earth ....

A) How much more expensive can a 65% lockup be than a 45% lockup setup ? more lockup % in acceleration = more sideways fun ? ( this is what I am looking for )
B) you have written that the more lockup percentage in deceleration we have , the less is the understeer . Is this right ? So why not have a high amount of % in deceleration ?
C) You write
"the downside of preload is understeering! the more preload, the more understeer!" . You mention that you will make the preload less but not zero ! Is this right ? We need some preload and why ?
Excuse me for the questions , but I still have some things confused in my head .
I certainly need a new lsd , and I try to spend my money the best possible way . .

Ps1 : you will make my unit i just ask to find out what exactly I need and want
Ps2: I use normal street tires Michelin pilot sport

[driftflo]

A) depends on the basis. a setup with 4 clutchpacks can be built very easily in a E39 M5 stock lsd core. if you want 6 clutchpacks, the lsd core needs to be removed (ringgear, bearings) and cnc machined. then everything needs to be installed again.

yes, you can possibly say "more lockup = more sideways fun" although it is a little more complex of course.


B) no! more lockup always means more understeer! if the locking value is higher this means the axle is locked more/longer and if two wheels are connected together and rotate at the same speed this means understeering! no matter of we speak about acceleration or deceleration. pretty easy...

and, to answer your next question in advance already: "more lockup = more sideways fun" is also true, because if you hit the gas (too) hard midway a corner both wheels spin and the car will slide sideways.


C) preload is static lock!
lock = understeering!

dynamic lock (generated via the ramps and applying a force on them) can be reduced or stopped anytime via applying less force or hitting the clutch pedal. the static lockup generated via the preload will always stay there!
but preload helps smoothing the load changes (while hitting and lifting the gas pedal). the lsd reacts more predictable. and the cupspring used to generate this preload can balance the wear of the clutches that are getting thinner after time.
i also build lsd setups without preload for some of my racing customers. if you know how to drive these setups and don't mind the short rebuild intervals this is a pretty perfect setup for experienced race drivers.
i currently also drive a setup without preload in my street-driven e92 because i like the extreme load change behaviours ;-)

[kevinlevrone]

Hello,

I have my Drexler LSD installed in a slightly used rear axle coming out from Drexler and given the fact that my rear axle pinions failed once, I would like to extend its life this time, if possible. So I have some questions:

- Is a break-in procedure necessary, like doing some 8-figures with the car at slow speed ?
- Is an oil change necessary after break-in ?
- What is the optimal oil replacement interval (if any) ?
- Should I reduce the torque at lower RPMs via tuning ? I have pretty good experience in tuning the N54 engine and I can do my own maps. Right now I have the engine tuned to max torque from lower RPMs, but I suspect that this is quite detrimental to the rear axle lifespan.

Thanks !

[driftflo]

Originally Posted by kevinlevrone

- Is a break-in procedure necessary, like doing some 8-figures with the car at slow speed ?

the drexler clutches need no special break-in procedure. driving a twisted road with moderate speed for a few miles will be enough.

Originally Posted by kevinlevrone

- Is an oil change necessary after break-in ?

yes, absolutely. after 500-1.500km you should change the oil!

Originally Posted by kevinlevrone

- What is the optimal oil replacement interval (if any) ?

depends on usage. after 5-10 trackdays or after 2-3 years of street driving you should replace it.

Originally Posted by kevinlevrone

- Should I reduce the torque at lower RPMs via tuning ? I have pretty good experience in tuning the N54 engine and I can do my own maps. Right now I have the engine tuned to max torque from lower RPMs, but I suspect that this is quite detrimental to the rear axle lifespan.

the lsd is able to handle very high torque values. gearset and bearings will fail earlier... as well as halfshafts and so on.

[thekurgan]

Thanks for the information in this thread, @driftflo; I ordered a Jim Blanton 105mm 7 clutch 40% static 100% dynamic, aggressive street diff for my beast and much of the research I completed was due to this thread.

[vagmag]

You have a personal message.

You have written :
B) in general the best would be as much clutchpacks and as steep ramps as possible.
the maximum you can get with the zf is 6 clutchpacks and according to the locking value you want to have you (or i) need to choose the ramp angles required. i have super steep ramps here from 85°, 80°, 75°, 70°, 65°... and down to 30°.

The ramps make the lsd react more urgent ?
Is it the preload ?
What do you propose for me for 45% of lock and 3,62 ratio for m5 e39 ?

[vagmag]

Originally Posted by vagmag

You have a personal message.

You have written :
B) in general the best would be as much clutchpacks and as steep ramps as possible.
the maximum you can get with the zf is 6 clutchpacks and according to the locking value you want to have you (or i) need to choose the ramp angles required. i have super steep ramps here from 85°, 80°, 75°, 70°, 65°... and down to 30°.

The ramps make the lsd react more urgent ?
Is it the preload ?
What do you propose for me for 45% of lock and 3,62 ratio for m5 e39 ?

Probably I didn't get it ...
After I saw this video http://m.youtube.com/watch?v=PEdnH7_7_yc

I realize that:
- preload is not via the ramps achieved. The different angle of the ramps make the diff more urgent or more lazy ... And also through the opposite ramp you can transform the unit to 2way or 1.5 way etc ...

What kind of ramps does the stock m5 e39 have and is considered "lazy " or "slow " ?
To what ramps do you suggest that we could turn into ? What will be the benefits and what will be the negatives ?

[vagmag]

How is "our - yours " configuration is compared to http://performancegearing.com

They speak about static and dynamic lockup .
How is the one and the other achieved ?
Are we going to have on our setup ?

[driftflo]

Originally Posted by vagmag

The ramps make the lsd react more urgent ?
Is it the preload ?

steep ramps: quick and smooth reaction/transition but less lockup!
shallow ramps: slow and rough reaction/transition but high lockup!

so the best is to fit as many clutchpacks as possible and then choose the ramps as shallow as you need to get the required amount of lockup.

i.e. a setup with 4 clutchpacks and 50° ramps (= ~40% lockup) is better, faster, smoother, longer lasting then a 2 clutchpack 30° ramp setup, although this would also be called a 40% lockup.

and no, this has nothing to do with the "preload". preload is static lockup generated via a static force on the clutches i.e. through a cupspring.

[driftflo]

Originally Posted by vagmag

Probably I didn't get it ...
After I saw this video

very good video!!!

Originally Posted by vagmag

I realize that:
- preload is not via the ramps achieved. The different angle of the ramps make the diff more urgent or more lazy ... And also through the opposite ramp you can transform the unit to 2way or 1.5 way etc ...

urgent and lazy is wrong. the rest is right. concerning the ramps see my post above.

Originally Posted by vagmag

What kind of ramps does the stock m5 e39 have and is considered "lazy " or "slow " ?
To what ramps do you suggest that we could turn into ? What will be the benefits and what will be the negatives ?

45/45° 2-way.
suggestions above as well.

[driftflo]

Originally Posted by vagmag

How is "our - yours " configuration is compared to http://performancegearing.com

not comparable as the values given on that website are simply wrong.
a static preload can never have a dynamic percentage value. how can a percentage ever be static??? this is wrong in general!
static lockup (aka "preload") is always given in an absolute value like x Nm of torque.

next mistake is the two values for dynamic lockup! this is simply not possible to provide via a salisbury clutch type lsd. you can realise this with the gkn viscolok (aka variable m lsd).

but maybe their setups will work anyway as they are definitely not like they wrote on their website. but there is a lack of theoretical understanding how clutch type lsd units work.

Originally Posted by vagmag

They speak about static and dynamic lockup .
How is the one and the other achieved ?
Are we going to have on our setup ?

again:
static lockup is an absolute value from 0 to x Nm.
you can also measure the current value of your lsd with a torque wrench. and i never saw a torque wrench that shows its values in % ;-)

dynamic lockup is ONE value from 0 to 100%... hopefully somewhere in between these to values as both extremes (0 and 100) could be built but would make no sense. (yes right, 0 can not be built in real world conditions as there is always a certain amount of friction).
this one value describes the difference in torque between the drive wheels that is required to open the lsd (transition from static to dynamic friction). from this moment on the two drive wheels will rotate at different speeds.