G25MX-R center differential options

[nerd racing]

hmm that is odd... wow... I wonder if you could change the torque bias by changing the rear diff ratio?

[thered56]

^^ That's what I was curious about. What Rocketman said about the center diff being locked as a bad thing with different ratios front to back. I was thinking that this would apply to the tranny without a viscous center diff. With the viscous, as was said, it distributes 57% to the rear, 43% to the front, normally. I'm trying to figure out a formula to see if the difference between front and rear ratios account for this distribution.

I came across this about an automatic AWD viscous system, and I quote: "In an attempt to reduce the coupling's activation time, VW Golf MkII Syncro always transfers 5% of torque to rear wheels (this is achieved by rear driveshaft rotating slower than front driveshaft in normal conditions, causing viscous fluid warm-up and slight solidification)."

http://www.awdwiki.com/en/viscous+coupling/

Also this (although the grammer makes me wonder): "most manufacturer varies the final drive ratio such that introduce a slightly speed difference even in normal condition. As a result, the car actually runs with 95:5 torque split between front and rear. This shorten the delay time."

http://www.autozine.org/technical_school/traction/tech_traction_4wd_2.htm

[Rocketman]

I do know that viscous coupled diffs only work correctly when at the proper temperature, I read something some time ago about it (I think it was referring to the Miata viscous unit) I think it needs to be pretty warm to couple properly. I wonder if it's similar to the case above, where it keeps the diff moving & the fluid warm?

What if you got a rear diff that matched the front? I wonder if you could safely lock the center diff then?

[thered56]

"In normal condition, front and rear axles run at roughly the same speed so the plates and viscous liquid are relatively stable to each other. When tyre slip occurs in one of the axle, that means the alternating plates run at different speed, viscous liquid will try to visco them together. As a result, torque is transferred from the faster driveshaft through the liquid to the slower driveshaft. The greater the speed difference, the larger the torque transfer."

Okay so......With a front diff gears ratio of 4.214:1, and a rear diff gear ratio of 3.909:1, more torque gets transfered to the rear, being the slower driveshaft. Now what makes up the percentage of difference? The front wants to move 7.8% faster than the rear. I suppose that must make the viscous diff grab that extra 7.8% from the from diff and put it to the rear diff. But that's more like a 42/58 split. hmmm....how about this...The rear wants to move 7.2% slower, so the viscous diff grabs that from the front and puts it to the rear? Am I onto something? Well let's see, that would mean that the BF, at 4.388:1/3.909:1, would put 61% of the torque to the rear. Does anyone know what the torque split is on the BF? Is it 39% front, 61% rear?

But, maybe I'm just out to lunch. I know I'm exhausted. lol

[thered56]

What if you got a rear diff that matched the front? I wonder if you could safely lock the center diff then?

Well, the two pages I linked in the second post. They would have to have matched diff ratios. Seems it works for them! I was thinking of that as an option, because I wouldn't wanna go through the trouble of AWD if I couldn't make full use of it. But this viscous sounds better than I thought it to be. I don't have an AWD, so I don't know about this manual locking of the center diff. I'm guessing that it is for the trannies without viscous though. Is it just a switch in the car?

Edit: Another useful link I've come across: http://www.mazdafun.com/product%20in...ve%20train.htm

[DuRWooD]

If you think about it from the perspective of the center diff, for simplicity sake powering a front an rear driveshaft; it sees the front driveshaft spinning faster than the rear. This causes the center plates to move against one another and heat up the fluid. As it heats up, the induced fluid friction on the plates cause the center diff assembly to want to match the driveshaft speeds. Since the rear is a faster ratio, this will cause more torque to be available at the rear; it is pushing off of the front.

The diff itself, which like all other mechanical diffs, is a 50/50 torque split.

[josh95mx]

I know this is like a month old, but I was reading through this and was like WTF WHY would you want full time 4 wheel drive?

I dont want to come off rude but this is an incredibly dumb idea unless your entire purpose is to drive off road in dirt and gravel. 4 Wheel drive 101 NEVER DRIVE YOUR VEHICLE IN 4 WHEEL DRIVE ON THE PAVEMENT. there are reasons for this. It becomes extremely difficult to handle and drive let alone steering when you have all 4 wheels pushing you. That is why Mazda designed this with the front and rear split that they did with an option to lock it into 4 wheel drive mode...

From my understanding of the way that this works, from reading several manuals, and some things online is, it puts the majority of the power to the rear at all times OTHER than at launch. So the only time the front will spin up first is during take off and there is a slight delay where the rear will spin up and engage and this is the Viscus's primary function. from what I read this is due to the possibility that if your front tires were on pavement and the rear on gravel, it would not cause any extra unnecessary wheel spin, which would cause extra stress on the drive system and possible failure and give you the best traction available. This is where the 50/50 lock comes into play and when locked the car now becomes full time 4 wheel drive and both front and rear will spin at the same rate with the same power distribution front and rear, which even in a 4x4 truck you are advised to not drive on pavement locked in 4 wheel drive, and in the Mazda manual you are advised to do the same.

I could be wrong but this has been my understanding of how this system works for many years. What I do know is that I like the split and wouldn't want a full time 50/50 it makes no sense to me unless i was in snow or on a dirt road and needed the extra grip. I purposly did not install the lock as I felt it unnecessary. In my car I can crank the wheel and pop the clutch and to doughnuts all day long on pavement. what more could a guy ask for.

[DuRWooD]

Face palm

[josh95mx]

Face palm

Well set me straight, because this has got me thinking. Am I just completely off track? I am here to learn and is seems that there are a lot of good points in here, perhaps I am just lost and need to read through the thread again.

[jay]

Well, you're not completely off track, just partly. So is most of this thread. Let's try and throw some light on the subject...

"full time 4wd" is just that, the drivetrain sends tq to all 4. Whether that's through 3 open diffs, 3 locked diffs, 3 lsd's, or any combo of them, if it's sending torque to 4 corners all the time then it's full time 4wd. This differentiates from the Haldex stuff VW used, where it's fwd only till the car senses wheelspin, then uses a clutch to start sending tq to the back. This is "part time", or, technically, "crappy" 4wd. Some lambo's did something similar, RWD till wheelspin then start sending power to the fronts. So really it's the fulltime 4wd you want, like the gtr, the newer lambos, mazda, toyota, Soobie, DSM, and most of the rest of the world these days if they're going all wheels driven. In our case, we have an open front diff, an open or viscous and lockable centerdiff, and an open or lsd rear. In a straight line, with 3 open diffs, this setup still sends power to all 4, at all times...right up until one wheel starts losing traction. But that doesn't mean it ain't fulltime 4wd, dig? Now an old Jeep might be more hardcore, with no centerdiff and a totally agricultural setup that makes all 4 wheels spin at the same rate regardless of conditions but there's a reason you only see that setup on old Jeeps. But to say that full time 4wd is only for loose surfaces is also an error, as the 3 diff setup mazda uses works very well indeed on tarmac. It's the same setup Audi uses. After they pwned the rally scene with the Quattro, they went road racing with the same drivetrain and so completely dominated the European road race circuit with it that the sanctioning bodies outlawed Audi's 4wd. Much lesser known is Ferguson's (inventor of this 3diff awd setup) F1 car from the early '60's. It only entered one race, with Stirling Moss driving. It won outright. And 4wd was banned from F1 immediately after. Does anyone think 4wd was banned because 4wd is no good on pavement?

As for torque split.... 57-43 or 50-50, there's no real difference. Since I own one of each in the same chassis, I have a pretty good notion on this. The Alltrac box uses the same ratio front and back and is thus 50-50, and I thought maybe I'd not like how the car cornered sending that much more to the front, but I can't tell the two setups apart there. The tail comes out as willingly as it did before, and the front doesn't push any more than I recall it doing before, so as far as I can tell, Mazda just built in more weakness with their differing-ratios torque split, putting more stress on the centerdiff and making it so that locking that item increases the chances of a kaboom. I could lock the centerdiff on the toyota box as well, but prefer to leave it doing the viscous centerdiff thing. Speaking of which---

I don't know where the notion that the centerdiff fluid has to "warm up" comes from. If anything, the silicone fluid that makes the viscous diff viscous, is stiffer when it's cold and thus makes the diff more limited-slip. When we lost the front CV on the rallycar, we were able to continue for awhile cuz the centerdiff was still sending 50% to the back of the car with the LF disconnected, spinning uselessly what was left of the inner CV and the RF doing nothing cuz well it's an open front diff. But after a mile or less the centerdiff got too hot (with that much more internal spin difference), the fluid got too thin, and it started behaving more and more as a regular open diff, we got less and less power to the back, and we had to pull over. So I can say with certainty that a viscous diff wants not to be warmed up. What it does want is RPM. The limited-slip effect is much more noticeable at speed than at idle. You can see this by putting the car up on a lift and running it at idle in 1st or 2nd. All 4 will spin, but it's pretty easy to stop the rears and the idle speed doesn't drop much. Try the same stunt in 3rd or 4th and it's more difficult to stop the wheels and if you do the idle drops a lot and the engine starts lugging. This BTW is a good way to tell if your VLSCD is in good shape, or if it's tired and worn out. Just make sure it's unlocked, and if you're testing an alltrac, that the selector lever on the gearbox is in the VLSD position and not in the locked or open position.

[DuRWooD]

That's an excellent explaination. When I said the fluid heats up causing increased fluid friction, I meant it the other way around; as the differential speed of the plates increases, the friction increases, thus heating up the fluid. Same funtion however.

A good way to explain the system situationally is to take it to the extreme. Think front wheels on pavement, rear wheels in the air. At a low engine speed, the car would barely creep forward, as there isn't much friction between the plates, floor it, and it will (ideally) spin the front tires on the pavement at whatever the torque split ratio to the front is with respect to the rear. The losses will be directly proportional to heat in the center diff fluid.

Edit: when I say "ideally", I mean infinite friction between the plates, this wouldn't be the case in practice obviously, unless you had a locking diff like the BF trans.

[Rocketman]

I don't know where the notion that the centerdiff fluid has to "warm up" comes from.

Differential motion forces the interleaved discs to move through the fluid against each other. In some viscous couplings when speed is maintained the fluid will accumulate heat due to friction. This heat will cause the fluid to expand, and expand the coupler causing the discs to be pulled together resulting in a non-viscous plate to plate friction and a dramatic drop in speed difference. This is known as the hump phenomenon and it allows the side of the coupler to gently lock.

...

Viscous LSDs are less efficient than mechanical types, that is, they "lose" some power. In particular, any sustained load which overheats the silicone results in sudden permanent loss of the differential effect.

http://en.wikipedia.org/wiki/Limited_slip_differential#Viscous

So perhaps not *every* VLSD works in that manner, but searching around tells me that the silicone fluids in VLSD's becomes more viscous and "effective" as it heats up.

[SgtRauksauff]

Here's another wiki article on viscous couplings:

http://en.wikipedia.org/wiki/Viscous_coupling_unit

dilatant fluid is the key ingredient.

here's a youtube video showing a simplistic operation of a Viscous coupling:

http://youtu.be/ONnhNgUzMjo

[jay]

Well, now we know where that info comes from, but that doesn't say that that info is any damn good. Cuz sorry, it mostly ain't, at least not for mazdas... Anyway, the manufacturers put silicon oil in viscous diffs because it is very much more resistant to changes in viscosity with temperature than other fluids. It is much more resistant to heat expansion than other fluids. They want very much for it to be as thermally stable as possible or they'd be unable to predict what the centerdiff is gonna do next. It's physically impossible for silicone fluid to react that quickly to heat, to expand that quickly, and otherwise behave the way that wiki article describes. Centerdiffs work much much more due to fluid shear forces than temperature induced fluid expansion. That latter effect is more of a byproduct than an effect. Here's a pdf on how silicon oil behaves, published by an outfit that makes the stuff. I'm more inclined to believe them than some anonymous wiki author. http://www.clearcoproducts.com/pdf/l...troduction.pdf
That said, that "hump phenomenon" mentioned in the wiki is a real effect. But it's not something that's supposed to happen in normal driving, and it has far more to do with increased shear resistance with increased plate velocity differences (the remaining mechanical gap closing effects being done more with the amount of air inside the diff unit, not the fluid, which makes much more sense. Air expands and contracts with temp changes right effing now unlike silicone oil.), and it's not something that happened for me when we lost the front CV, otherwise the diff woulda locked up and we'd have finished the stage. But instead we gradually lost drive to the back wheels and parked it. Later, when it had cooled off, I drove it off the trailer. So either Toyota alltrac centerdiffs aren't made to lock up like that, or mine was already worn out and wouldn't have been able to anyhow, and is now even worse for the wear. Since it passed the put-it-in-the-air-and-run-it-in-gear test, I ain't gonna worry about it. I have yet to find whether the alltrac diff or mazda diff was designed to do the hump mode thing. The miata rear viscous diff apparently does not, which is why I say this wiki doesn't apply to us. Seems that later WRX diffs can, but only under unusual circumstances, and you can tell cuz you can hear the poor thing groan as the plates grind together. This is not a noise I've ever heard from any of the mazda or this toyota vlscd's I've ever abused.

[SgtRauksauff]

I actually fully agree with you, from my experience with my GTX transmission during ice time trials. the hotter it got and the more I spun the wheels, the less and less the front would pull at all. I didn't have a broken axle for one front to spin freely in, but I think if heat were to raise the fluid's shear strength, I should've had at least one front and one rear pulling great, instead of only the one back tire.

unless there's something in the fluid that just "wears out" after enough use?

--sarge