Evo3 center differential

[asian312]

I've come across a Cusco 'tarmac gear' center diff for an Evo3. Does anyone know if the center diff from an Evo3 will swap into a 2G transmission?

 

[D_Eclipse9916]

how did you get one of those?!!

 

[asian312]

i've met a few people along the way attending driving schools =)

Anyone have a clue if it fits?

 

[D_Eclipse9916]

no it doesnt, send it my way. . I am only speculating from what I have heard, but I beleive we can fit them, but that itll change the ratios. I wish someone would actually come in here with experience.

 

[raped]

I dont see any reason why it would not fit. No, the ratios are not different. It may actually be the same model that is sold for the 92-99 dsm.

 

[jtmcinder]

The Cusco for an Evo3 is what people with DSMs use when they want a planetary center. It's what ACM runs, for example.

- Jtoby

 

[asian312]

A little research has provided a product number for comparison.

510-020-A

From what I've found through searching, the late 91-99 share the same center diff. The 90's and early 91s are a tad different with gear angles. Since the Evo3 launched in 95, I presume that the center diff is the proper year. I'm just waiting for verification from a particular member, before purchasing. *confirmed*

I've read your articles Jtoby about differentials and ran across a thread or two concerning the torque split. What's your take on the 35/65 split? I know that's not the exact torque split you would opt for, but would it be favorable to the stock 50/50?

 

[jtmcinder]

I, personally, would prefer something around 40/60. I thought this a while ago, based on some simple math; it was confirmed when I tried a new and old STi back-to-back. If there was some way to jam the 45/55 out of a 3000GT VR4 into our tranny, I'd do it in a moment.

You have upgraded rear axles and a new rear diff in your budget, right?

Edit: I'm sorry I didn't actually answer your question. If everything else were done on the car and I had the money to try it, I probably would. I'm just not close to being done and don't have the money, anyway. I'm still hoping the 41/59 from the new STi will appear in a form that fits in our cars.

- Jtoby

 

[Terrh]

WOW, I had NO IDEA that anyone made a center diff that would fit our cars with a different torque split!

35/65 has got to be better than the stock 50/50!

I am not the biggest fan of power understeer, and I'm just looking for a fun capable car in my case, not the absolute fastest laptimes, I'd rather have "fun" handling... fun that a bit more torque going to the rear wheels would provide.

Are there any other center diff options that are readily availabe and fit? Is this one available? (and under a grand?)

 

[dsmcrazyspiff]

The only time a dsm has a 50/50 split is with a welded diff or a spool in the center diff

 

[asian312]

You have upgraded rear axles and a new rear diff in your budget, right?

Nope, but then again I make a woping 200-220 HP with my 14B =)

This is the only option hat I know of that has the ability to change the torque spilt. Expect to pay some where around $2000 new.

 

[Terrh]

man, I cant justify spending $2000 for a diff....


Are there any other solutions in the sub $1k range? (even $500ish used would be great )

The only time a dsm has a 50/50 split is with a welded diff or a spool in the center diff

What split do they have then?

and while they'd "technically" have a 50/50 split with a welded diff, in actual driving conditions (and the same applies to the stock setup AFIAK) the torque split would vary depending on actual individual tire grip..... with a welded diff it could theoretically be 100/0 and 0/100 at times.

 

[raped]

i heard awhile ago about draining the fluid inside the viscous coupling and refilling it with a a more viscous fluid. This would do nothing to the torque split but would allow the diff to lock faster. I am still searching for someone who has more information on this.

 

[jtmcinder]

Please, stop saying that welded "diffs" have a 50/50 split, because they don't. They can be anything from 0/100 to 100/0 from one instant to the next. The typical, spider-type diff that comes stock in the center of a DSM produces a 50/50 split.

As to the idea that it's a no-brainer that 35/65 is better than 50/50, many people would beg to differ, including the engineers at Mitsu. And given that Evos usually kick STis' butts....

- Jtoby

 

[jtmcinder]

i heard awhile ago about draining the fluid inside the viscous coupling and refilling it with a a more viscous fluid. This would do nothing to the torque split but would allow the diff to lock faster. I am still searching for someone who has more information on this.

I still can't verify whether the stock fluid is the type that increases or decreases in viscosity as temp rises, so anything you find would be welcome here.

- Jtoby

 

[raped]

I still can't verify whether the stock fluid is the type that increases or decreases in viscosity as temp rises, so anything you find would be welcome here.

- Jtoby

I dont have too much information on that but I did hear a couple years ago that there was a rally car preparation shop located in Lake Zurich called TAD that had access to or did just that (change the fluid in the VC). I have tried to contact TAD but i have not had any luck.

 

[jtmcinder]

What drives me crazy about this is how the claims on various websites just don't line up with the physics. I've seen many claims of increasing viscosity with temp, but that is simply not the norm for fluids. (Gases do this, but not fluids.) In contrast, whenever I read a technical article, the fluid is always assumed to drop in viscosity as the VC is forced to work and generates heat. What causes a VC to transfer more torque as the temp goes up is the expansion of the fluid, not an increase in viscosity. At first, the expansion of the fluid just fills up more of the unit, so a higher percent of the plates are actually shearing through the juice. So, even though the viscosity is going down, the increase in area more than makes up for this. Then, if the fluid gets hot enough, you get the hump phenomenon, where the plates are being pressed together. (This is why half the plates have holes and half don't, by the way.) In other words, the VC starts acting like a clutch-type LSD when the pressure gets high enough.

In any event, I'd love to know if the claims I see on non-technical website are plain BS or whether there really is a VC fluid that increases in viscosity. My experience with the web vs technical journals tells me to ignore the web, but it's so widespread that I still have my doubts. (Of course, more people on the web believe a locked "diff" is 50/50 than not, while I know they are wrong, so I don't know why I take anything very seriously, but I do.)

- Jtoby

 

[raped]

What drives me crazy about this is how the claims on various websites just don't line up with the physics. I've seen many claims of increasing viscosity with temp, but that is simply not the norm for fluids. (Gases do this, but not fluids.) In contrast, whenever I read a technical article, the fluid is always assumed to drop in viscosity as the VC is forced to work and generates heat. What causes a VC to transfer more torque as the temp goes up is the expansion of the fluid, not an increase in viscosity. At first, the expansion of the fluid just fills up more of the unit, so a higher percent of the plates are actually shearing through the juice. So, even though the viscosity is going down, the increase in area more than makes up for this. Then, if the fluid gets hot enough, you get the hump phenomenon, where the plates are being pressed together. (This is why half the plates have holes and half don't, by the way.) In other words, the VC starts acting like a clutch-type LSD when the pressure gets high enough.

In any event, I'd love to know if the claims I see on non-technical website are plain BS or whether there really is a VC fluid that increases in viscosity. My experience with the web vs technical journals tells me to ignore the web, but it's so widespread that I still have my doubts. (Of course, more people on the web believe a locked "diff" is 50/50 than not, while I know they are wrong, so I don't know why I take anything very seriously, but I do.)

- Jtoby

You are right heat does decrease the viscousity of the fluid (Ijust figured that out). I believe it might be the shearing action that causes the fluid to become more viscous.

Edit: and i dont believe the plates ever touch

 

[jtmcinder]

Oh, the plates definitely come in contact. That's the point of having half the plate with holes and half not: to press them together to get the hump phenomenon. And I know our center produces the hump, because it was very clear when I was rallycrossing on snow and ice.

As to shear producing an increase in viscosity ... there's something to this. Several of the patents I've read (yes, I've gone as far as reading patents to try to get to the bottom of this) have mentioned special methyl-silicone and polyorganosiloxane molecules with critical properties to combat the loss of viscosity with temp. At least one these appears to be activated by motion and shearing.

In other words: physics is not enough; we need to know chemistry, as well. And here I am: a guy who has a VCR saying 12:00 ... 12:00 ... 12:00 ... 12:00

- Jtoby

 

[raped]

I opened a vc a couple years back, ill have to open another and get back to you.

As far changing the fluid in the vc i contacted Devo Tuning. Initially they said it was done in New Zealand. I was told to call back for more info. The second time i called I was told it makes the differential more rear biased. I don't see how this is possible through just changing the fluid.

I'll post more information as I get it.

 

[jtmcinder]

Anyone who says changing the fluid in a DSM's VC will change the torque bias of the associated diff should not be called back.

Yes, this attitude has some drawbacks, but (just like Preparation H) it feels good, on the whole.

- Jtoby

 

[raped]

Anyone who says changing the fluid in a DSM's VC will change the torque bias of the associated diff should not be called back.

Yes, this attitude has some drawbacks, but (just like Preparation H) it feels good, on the whole.

- Jtoby

Well I still have to speak with Adam. The person that I spoke with probably made that assumption because he understood that the vc actually connects the center diff basket and the out put shaft (which which ultimately connects to the rear end). In reality we know that the strength or shall I say, limited slip power, has nothing to do with torque bias even though the diff basket is connected only to the rear via the vc.

The only way we could change the torque bias with the use of a vc would be to bias the vc somehow allowing to turn one way easier than the other.

 

[Terrh]

so are there other options?

I won't (can't) spend $2k on a center diff.

It seems like a no-brainer to me that 35/65 would be more fun to drive, whether or not it would be faster on the track, I don't know... I think my 0/100 (all the time) RX-7 is the better tool for me to use on a track, I just want this to be more fun to drive.

I'm glad that I'm not the only person that sees that a welded diff doesn't actually have a 50/50 torque split...

 

[tkrPerformance]

if you get it you will love it.
I have one in my evolution III tranny and the cusco lsd in the rear end also. they work great

 

[GRNDSM]

Please, stop saying that welded "diffs" have a 50/50 split, because they don't. They can be anything from 0/100 to 100/0 from one instant to the next. The typical, spider-type diff that comes stock in the center of a DSM produces a 50/50 split.

As to the idea that it's a no-brainer that 35/65 is better than 50/50, many people would beg to differ, including the engineers at Mitsu. And given that Evos usually kick STis' butts....

If a welded "50/50" center diff goes from 0-100 to 100-0 from one instant to the next (depending on traction). So then what happens to the 35/65 center diff when it looses traction on one set of wheels?