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

[Killerfish2012]

Originally Posted by driftflo

i am thinking of gear differentials that provide torque biasing. there are actually no gear "lsd" units.

wavetrac is a gear differential that additionally provides some preload and corresponding a little biasing on coast. i discussed that with dana clark (you should know who he is if you are into wavetrac) in detail. so i dont think i need to do more research on that one!

as i said, these units are quite ok. if you think it is the right one for you and it has enough biasing capacity on coast, get it!

i generally don't like gear type units that much because they are not able to lock up the axle at all. it is no locking but a biasing differential. it feels pretty different. so speaking of "lock" concerning a wavetrac is wrong. just do some research on that or ask dana directly ;-)

a 335i also feels different from a M3. two worlds...

Yes, but we raised on wheel of a 335I with wavetrac entirely off the ground. When gas is applied it not only transferred power, but the vehicle moves. That's the only LSD I've ever seen that can do that. This simulates all zero load conditions as well. Yet to see OS giken, or Drexler replicate the same feat.

[driftflo]

Originally Posted by bigjae1976

Looks like the kit I'm referring to has thinner clutches and lock plates as they advertise zero machine work is necessary. They also say it was designed for zero preload.

sure no preload, as they have no space for a belleville spring without machining! i decide depending on the kind of usage, if i suggest preload or not. i guess they sell you "no preload" as a feature ;-)

Originally Posted by bigjae1976

So what is the design of the E46 M3 diff? Is it a 2 plate clutch like the E39 M5?

the e46 has a gkn viscolok like the e92. completely different technology.

[tikamak]

Originally Posted by bigjae1976

Looks like the kit I'm referring to has thinner clutches and lock plates as they advertise zero machine work is necessary. They also say it was designed for zero preload.

So what is the design of the E46 M3 diff? Is it a 2 plate clutch like the E39 M5?

The E46 M3 one is the GKN viscolok
The E39 M5 is the ZF clutch LSD with 2 clutches.

You can put 4 clutches if the thickness of the clutches and steel plates are less than stock.

[bigjae1976]

Originally Posted by tikamak

The E46 M3 one is the GKN viscolok
The E39 M5 is the ZF clutch LSD with 2 clutches.

You can put 4 clutches if the thickness of the clutches and steel plates are less than stock.

Looking at used E39 M5 diffs...seems much easier and more economical to go with a Diffsonline 3 clutch unit.

Question, by variable ramp angle, do they mean that there are different lock up rates under acceleration and deceleration like a 1.5 way Kaaz? Or does it refer to being able to customize the ramp angle(s)?

A steeper ramp angle means quicker lock up?

[tikamak]

Originally Posted by bigjae1976

Looking at used E39 M5 diffs...seems much easier and more economical to go with a Diffsonline 3 clutch unit.

Question, by variable ramp angle, do they mean that there are different lock up rates under acceleration and deceleration like a 1.5 way Kaaz? Or does it refer to being able to customize the ramp angle(s)?

A steeper ramp angle means quicker lock up?

This depends on the angle is measured.

With OS giken it's the other way around. With ZF LSD (210mm diffsonline for example), an angle of 30/90 means that the LSD is 1 way:
30 angle on acceleration means that it will lock less progressively than 45 degree stock (the end locking is the same though, that depends on the clutches, friction material, etc ...)
90 degrees on deceleration means that the LSD will not do anything when decelerating since the 90 degrees angle will be blocking the spider gears support from engaging the clutches and hence will be an OPEN diff.


Stock is 45/45 which is a very neutral and safe.

Diffsonline can also do a variable ramp 30/60/90 and drexler can do a curve ramp instead of a straight line:
a 30/60/90 is a 1 way diff that on acceleration will initially have a less aggressive lockup pattern and after arriving to a higher degree of slip the diff will still be locking up progressively but at a more aggressive pace as the slip becomes greater. (it could be also the opposite, depending on how the LSD is machined and what the race driver wants the car to behave like)
This kind of variable ramp can only provide a benefit for a very very skilled driver as he will need to be able to pull off special driving tactics like entering a corner and rotating the car with the LSD benefitting from the progressively aggressive lockup.

Now to images:
This is a 55/90 on the right and 45/20
This is the link and image for example:

[driftflo]

the steeper the ramps, the quicker and smoother it reacts!
the more shallow the ramps are, the more lockup is generated. the aim should always be to put in as many clutchpacks as possible and have the ramps as steep as possible.

"variable ramps" are no option in professional lsd building as in a properly configured lsd, the axles will travel no more than 1-2mm at highest! machining a curve that makes sense and can be driven properly without the lsd doing unpredictable things is not possible. don't buy units with "variable" ramps!

drexler doesn't do that either if you look at there pressure plates in detail. the part where the axles travel up and down is a stright line... the curves are done on the upper part of the "V" where the axle will never move to (maybe because it looks better or the plate gets more stable or it simply is easier to machine like that?!)

if diffsonline really means that when talking about "variable ramps" i'd like to see a closeup of their ramps! if they mean you can choose what ramp they should use it would make more sense but then they need to manufacture and harden each set individually what also makes no sense. if they mean it is no 2-way but 1.5 or 1 way they use the wrong words!
so i would really be interested what they want us to say with "variable ramps"?!

[bigjae1976]

Thanks for the spot on answers!!!

[tikamak]

Originally Posted by driftflo

the steeper the ramps, the quicker and smoother it reacts!
the more shallow the ramps are, the more lockup is generated. the aim should always be to put in as many clutchpacks as possible and have the ramps as steep as possible.

"variable ramps" are no option in professional lsd building as in a properly configured lsd, the axles will travel no more than 1-2mm at highest! machining a curve that makes sense and can be driven properly without the lsd doing unpredictable things is not possible. don't buy units with "variable" ramps!

drexler doesn't do that either if you look at there pressure plates in detail. the part where the axles travel up and down is a stright line... the curves are done on the upper part of the "V" where the axle will never move to (maybe because it looks better or the plate gets more stable or it simply is easier to machine like that?!)

if diffsonline really means that when talking about "variable ramps" i'd like to see a closeup of their ramps! if they mean you can choose what ramp they should use it would make more sense but then they need to manufacture and harden each set individually what also makes no sense. if they mean it is no 2-way but 1.5 or 1 way they use the wrong words!
so i would really be interested what they want us to say with "variable ramps"?!

This is a far better explanation than mine. I missed the point where the diff actually pushes on the ramp and this is how the locking mechanism is actuated.

For the variable ramps you can see here as this is my reference:
http://www.zpost.com/forums/showpost...0&postcount=22

This is my friend with a supercharged track weapon 4.10 diff Z4M.

[OM VT3]

Originally Posted by driftflo

the steeper the ramps, the quicker and smoother it reacts!
the more shallow the ramps are, the more lockup is generated. the aim should always be to put in as many clutchpacks as possible and have the ramps as steep as possible.

"variable ramps" are no option in professional lsd building as in a properly configured lsd, the axles will travel no more than 1-2mm at highest! machining a curve that makes sense and can be driven properly without the lsd doing unpredictable things is not possible. don't buy units with "variable" ramps!

drexler doesn't do that either if you look at there pressure plates in detail. the part where the axles travel up and down is a stright line... the curves are done on the upper part of the "V" where the axle will never move to (maybe because it looks better or the plate gets more stable or it simply is easier to machine like that?!)

if diffsonline really means that when talking about "variable ramps" i'd like to see a closeup of their ramps! if they mean you can choose what ramp they should use it would make more sense but then they need to manufacture and harden each set individually what also makes no sense. if they mean it is no 2-way but 1.5 or 1 way they use the wrong words!
so i would really be interested what they want us to say with "variable ramps"?!

What's the best lsd for these cars?

[driftflo]

Originally Posted by tikamak

For the variable ramps you can see here as this is my reference:
http://www.zpost.com/forums/showpost...0&postcount=22

i still don't understand it. "30/60/90 ramps" just makes no physical sense?! ... or is beyond my understanding!

he also speaks about 90% lockup? with 2 clutchpacks?! hmmm... besides the fact, that 90% lockup is way too much in nearly any racing situation i can think of, he has the first lsd out there that does this with just 2 clutchpacks.


anyway: i would still be very interested in more details about the "30/60/90 ramps". a picture of these ramps would be perfect.

[driftflo]

Originally Posted by leigh

What's the best lsd for these cars?

a clutch type lsd with a proper setup that fits your individual needs.

[tikamak]

Originally Posted by driftflo

i still don't understand it. "30/60/90 ramps" just makes no physical sense?! ... or is beyond my understanding!

he also speaks about 90% lockup? with 2 clutchpacks?! hmmm... besides the fact, that 90% lockup is way too much in nearly any racing situation i can think of, he has the first lsd out there that does this with just 2 clutchpacks.


anyway: i would still be very interested in more details about the "30/60/90 ramps". a picture of these ramps would be perfect.

gmd2003
Help us out here
What do you mean by variable ramp angles?
Do you think you can get a picture from dan ?

[bigjae1976]

Originally Posted by driftflo

sure no preload, as they have no space for a belleville spring without machining! i decide depending on the kind of usage, if i suggest preload or not. i guess they sell you "no preload" as a feature ;-)

the e46 has a gkn viscolok like the e92. completely different technology.

So I would understand that preload puts some tension on the clutch packs so they lock up quicker?

Or am I way off???

[bigjae1976]

Originally Posted by tikamak

gmd2003
Help us out here
What do you mean by variable ramp angles?
Do you think you can get a picture from dan ?

I think Diffsonline sells a 2 clutch, 3 clutch, and 3 clutch w/ variable ramps. Not sure if I've seen them advertise a 30/60/90 2 clutch system.

[gmd2003]

Originally Posted by driftflo

i still don't understand it. "30/60/90 ramps" just makes no physical sense?! ... or is beyond my understanding!

he also speaks about 90% lockup? with 2 clutchpacks?! hmmm... besides the fact, that 90% lockup is way too much in nearly any racing situation i can think of, he has the first lsd out there that does this with just 2 clutchpacks.


anyway: i would still be very interested in more details about the "30/60/90 ramps". a picture of these ramps would be perfect.

It's actually 3 clutch packs and a mechanical ramp capable of 30/60/90 lockup , all internals REM polished and a BMW Motorsport 4.10 gear set .I have been racing with it for 3 years and it is fantastic . Laser quick response and makes the car very predictable on hard acceleration , mid corner , and on deceleration .I have never taken it apart so I'm sorry no pics of its internals .

[driftflo]

Originally Posted by gmd2003

It's actually 3 clutch packs and a mechanical ramp capable of 30/60/90 lockup...

an asymmetric clutchpack setup is not the cleanest way to build a lsd but i did some on customer request as well and it seems to work quite ok indeed.

i still do not understand how a "mechanical ramp" is capable of "30/60/90".

you mean 30/60/90 is the percentage of torque difference of the two drivewheels until the lsd opens? or what is this "30/60/90" thing? what does it describe or stand for?
and when does it "30", when "60" und when "90"? i struggle to understand it if we take physical laws as given and even diffsonline is not able to go beyond these laws ;-)

[gmd2003]

Originally Posted by driftflo

an asymmetric clutchpack setup is not the cleanest way to build a lsd but i did some on customer request as well and it seems to work quite ok indeed.

i still do not understand how a "mechanical ramp" is capable of "30/60/90".

you mean 30/60/90 is the percentage of torque difference of the two drivewheels until the lsd opens? or what is this "30/60/90" thing? what does it describe or stand for?
and when does it "30", when "60" und when "90"? i struggle to understand it if we take physical laws as given and even diffsonline is not able to go beyond these laws ;-)

The diff is able to do 30% , 60% or 90% lockup depending on the forces applied to the wheels . So it's similar to the 0-100% lockup behavior of the m diff but much more durable with a faster response . The faster response is from having good pretension with the multiple clutch packs . Again how Diffsonline physically constructs the diff is not known by by me , i just know it works and works well .

[driftflo]

Originally Posted by bigjae1976

So I would understand that preload puts some tension on the clutch packs so they lock up quicker?

Or am I way off???

the main features of the preload spring(s) is that it holds the clutches under some load and therefore the load changes are smoother, especially if a setup without this spring has worn out a little. then there is too much clearance inside and load changes will let the lsd react pretty extreme from loose to fully locked.
so if you have a lsd without a belleville spring that was built with the right assembly clearance this is ok if you will have it rebuilt on a regular basis.

i also like to build racesetups without springs but with some preload generated from the clutches themselves. if you measure this setup you will also have a certain amount of preload but if you open it, there is no belleville spring inside. these are the quickest reaction setups out there.

there are so many tweaks you can tune a clutch type lsd. each one has certain pros and cons.

[driftflo]

Originally Posted by gmd2003

The diff is able to do 30% , 60% or 90% lockup depending on the forces applied to the wheels .

why is a value like this given in percentage and not as an absolute value? right, because it depends on the force applied: apply 100Nm and 30% means another absolute value than if you apply 1.000Nm and take 30% from this.

so it simply makes no sense to give different percentage values. it does make sense to give two values as there is one ramp for power and one for coast. but a third ramp simply is not there!

of course the guys at diffsonline can claim anything they want to. but as long as there is no in depth explanation to this "magical behaviour", i can not believe it.

Originally Posted by gmd2003

So it's similar to the 0-100% lockup behavior of the m diff but much more durable with a faster response .

impossible! look at the way the force is applied to the clutchpacks in the gkn viscolok and look at the activition in a clutch type lsd. if someone has understood both ways of pressure generation it is clear that the above is impossible.

Originally Posted by gmd2003

The faster response is from having good pretension with the multiple clutch packs .

no, the faster response comes from the way, the pressure is applied in a clutch type lsd vs. viscolok.

Originally Posted by gmd2003

Again how Diffsonline physically constructs the diff is not known by by me ,...

so i guess i have to wait until someone will strip down his unit and is posting pics of the internals.

Originally Posted by gmd2003

... i just know it DOES behave like a 30% locking diff when I want it to and a 90% under deceleration . It works and works well .

ok, it locks 30% on power and 90% on decel (this setup would make no sense in a FR car but it could be built, yes!).
...but in what situation will this unit lock 60%?

[gmd2003]

Originally Posted by driftflo

why is a value like this given in percentage and not as an absolute value? right, because it depends on the force applied: apply 100Nm and 30% means another absolute value than if you apply 1.000Nm and take 30% from this.

so it simply makes no sense to give different percentage values. it does make sense to give two values as there is one ramp for power and one for coast. but a third ramp simply is not there!

of course the guys at diffsonline can claim anything they want to. but as long as there is no in depth explanation to this "magical behaviour", i can not believe it.

impossible! look at the way the force is applied to the clutchpacks in the gkn viscolok and look at the activition in a clutch type lsd. if someone has understood both ways of pressure generation it is clear that the above is impossible.

no, the faster response comes from the way, the pressure is applied in a clutch type lsd vs. viscolok.

so i guess i have to wait until someone will strip down his unit and is posting pics of the internals.

ok, it locks 30% on power and 90% on decel (this setup would make no sense in a FR car but it could be built, yes!).
...but in what situation will this unit lock 60%?

I'm running on no sleep this am , I meant 90 % on acceleration and 30 on decel . I understand what you are saying , it's certainly a percentage of tq to the wheels . In every other clutch style diff it was my understanding that you could essential only have it behave x% tq under acceleration and x% tq transfer under deceleration . Most people ( me included prefer less under deceleration as it makes the car more stable under braking . This diff is very stable under braking and does an excellent job of putting down all 525hp in my Z4M on track esp mid corner . So it definitely behaves differently under acceleration vs dec ( has more lockup on acc) . Why don't you email Dan and ask him ? He's very responsive and I'm sure he would answer . I was going to get a standard 30% /60 % lockup 3 clutch setup but bought this carrier bc someone else had backed out and I got a reduced price .

[driftflo]

Originally Posted by gmd2003

Most people ( me included prefer less under deceleration as it makes the car more stable under braking .

no, it is the other way around:
the more lockup, the more the drive wheels tend to be connected together. both wheels spinning with the same speed means the car pushes straight. on decel, more lockup results in a more stable rear end. this is why a (good) 911 setup always has more lockup on decel.

even a 2 way lsd feels to lock harder on power because if you think of the torques applied under hard acceleration vs. hard braking ;-)

so the M3 GT4 (or Z4 GT3 MJ2011) bmw factory lsd units had more lockup (via more shallow ramps) on decel to get a real 2-way "feeling".


anyway: if you like your setup the way it is, this is perfect! how it was actually done or what exact setup you do have is of no importance for you.


my interest in the setup of a certain diff (re-)builder and how he does his setups is not so high that i will contact and ask them.
i discussed with matt monson of guard transmission a while ago. quite interesting in some points but i do not agree with him in all points. but he definitely knows what he is doing!
i also discussed with other such diff-builders a few years ago... which often ended that i explained how it works and how it is done better.
now i simply do it myself. ;-)

i think it is more helpful for me to talk to the technical contacts of the lsd manufacturers, some amg and bmw motorsports engineers as well as companies providing the most important parts like lsd clutch manufacturers, ... (most of them are based in europe, what makes it easier for me).
these companies do a lot of research (some more, some less) and you get first hand information on what works and what doesnt (or works, but not as good as it could).

[gmd2003]

Originally Posted by driftflo

no, it is the other way around:
the more lockup, the more the drive wheels tend to be connected together. both wheels spinning with the same speed means the car pushes straight. on decel, more lockup results in a more stable rear end. this is why a (good) 911 setup always has more lockup on decel.

even a 2 way lsd feels to lock harder on power because if you think of the torques applied under hard acceleration vs. hard braking ;-)

so the M3 GT4 (or Z4 GT3 MJ2011) bmw factory lsd units had more lockup (via more shallow ramps) on decel to get a real 2-way "feeling".


anyway: if you like your setup the way it is, this is perfect! how it was actually done or what exact setup you do have is of no importance for you.


my interest in the setup of a certain diff (re-)builder and how he does his setups is not so high that i will contact and ask them.
i discussed with matt monson of guard transmission a while ago. quite interesting in some points but i do not agree with him in all points. but he definitely knows what he is doing!
i also discussed with other such diff-builders a few years ago... which often ended that i explained how it works and how it is done better.
now i simply do it myself. ;-)

i think it is more helpful for me to talk to the technical contacts of the lsd manufacturers, some amg and bmw motorsports engineers as well as companies providing the most important parts like lsd clutch manufacturers, ... (most of them are based in europe, what makes it easier for me).
these companies do a lot of research (some more, some less) and you get first hand information on what works and what doesnt (or works, but not as good as it could).

Ha , I thought I got lock up reversed based on your previous reply this why I reversed myself . I'll stop answering until I get some sleep lol. At any rate the car puts the power down well mid corner and is very stable under breaking and controlled trail braking which was my goal .