The Great AWD Torque Split Debate: 90/10 vs 60/40

[Chiketkd]

If you do a search, you'll find that this topic has been discussed numerous times before but no conclusive answer was ever arrived at for the USDM SVX with the transfer clutch awd system.

When the SVX was first introduced, several magazines and technical articles talked about a 60/40 torque split during normal driving (with no slippage) - but other USDM Subaru models (Impreza, Legacy) which use a similar awd system have a 90/10 torque split, which has led several people to draw the conclusion that the SVX was no different.

Over the past few months I've been researching this topic and even did some 'field research' in the passenger seat of Tom's '92 LS-L using his Subaru Select Monitor.

My Findings:
The excerpt "Multiplate Transfer Clutch" taken from the 98 page Road & Track Guide courtesy of Earthworm's site www.svx-iw.com backs up my findings in the field:



In the awd multi-plate transfer clutch system, Solenoid C acts as an on/off switch used to fine-tune the torque split front-to-rear. However, torque split to the rear wheels never drops to 0% but closer to 5-10% (this was proven when I had my car up on SVXfiles lift with the fwd fuse engaged, but still had spinning rear wheels).

The above article describes Solenoid C as a "chatter valve" which cycles rapidly between on/off - works at a speed which most hardware store voltmeters will not be able to register. Think of the "chattering" of Solenoid C being like an elevator moving between two floors - if you want to keep the elevator suspended at the mid-point between the floors, using only an on/off switch ('on' the elevator goes up, 'off' it starts falling), someone would be able to accomplish this task by pressing the 'on' the button until the elevator reaches the desired position, followed by a rapid cycling between 'on' & 'off' to keep the elevator suspended at that point.

This 'chattering' of Solenoid C is controlled by the SVX specific TCU. The Subaru Select Monitor is able to register the on/off signals being sent by the TCU to Solenoid C and calculate the effective torque split front-to-rear.

While in the passenger seat of Svxfiles's car, I had him do a few things before we pulled off. He first moved the shifter into the 'D' position - the Subaru Selector Monitor (SSM from here on) registered a split of 60/40. He then put it into the '3' & '2' positions with no change. When put into the '1' position the torque split now registered on the SSM as 50/50. In 'R' the torque split changed back to 60/40.

Svxfiles' then drove off in 'D' while I monitored the SSM. Driving around on level flat road the torque spilt varied between 70/30 to 60/40 depending on whether he was accelerating, at a steady velocity or slowing down for a stop light. At wide open throttle (i.e. power mode), the split went immediately to 50/50 and I even once, for a brief moment, saw it register as 49/51 (which theoretically is possible in the chattering on/off system).

When driving around corners, on entry the torque spilt would initially favor the front wheels at 75/25 but then shift back to 60/40 on exit.

Conclusion: While the SSM isn't infallible, it is the official licensed device authorized by Subaru to correctly monitor/diagnose their cars. So who's right 90/10 or 60/40??? Both. Why, because SVX the transfer clutch awd system works between the extremities of 50/50 to ~90/10 when dividing torque. However, those are the extremities and the torque split never stays at either one unless forced to (i.e. power mode or fwd fuse). But, when cruising along a dry, flat interstate with the torque converter locked up at 65mph, torque split is 60/40 in the SVX.

very informative..I too often wonder as to what the split currently is on my SVX while driving in varying conditions

[Beav]

I didn't realize anyone thought it was a 'one or the other' type arrangement. It definitely is a continuously variable affair.

Monitors are notoriously slow in reporting what is happening. I suppose the ECU/TCU is more concerned with maintaining vehicle operation than reporting it. Real time action can be viewed with a meter that has a duty cycle settting.

I must be missing something, how is it possible for the system to go 49/51? I realize the monitor may very well display that, they aren't totally 100% accurate (believe it or not.) It's just that at 100% application the torque can only be split evenly, right? (applied equally, front & rear)

[michael]

OK so does this mean there is no difference between the usdm and jdm transmissions?

[Chiketkd]

Originally Posted by Beav

I didn't realize anyone thought it was a 'one or the other' type arrangement. It definitely is a continuously variable affair.

Monitors are notoriously slow in reporting what is happening. I suppose the ECU/TCU is more concerned with maintaining vehicle operation than reporting it. Real time action can be viewed with a meter that has a duty cycle settting.

I must be missing something, how is it possible for the system to go 49/51? I realize the monitor may very well display that, they aren't totally 100% accurate (believe it or not.) It's just that at 100% application the torque can only be split evenly, right? (applied equally, front & rear)

Only the most sensitive meters with a duty cycle setting can monitor the rapid on/off of solenoid C. A cheap $10-$20 voltmeter will not be able to accomplish this...

The SSM would have some small % of error, but from my understanding of the 4EAT, 100% of torque is sent to the back of the transmission, where the split front-to-rear occurs. With a 'chatter valve' that works in an on/off fashion, theoretically it is possible to go past the desired position briefly when switching between on & off.



If the torque split can never be more than 50% rear, then it is possible that the SSM has an error of 1%. Either way, while a sensitive duty cycle voltmeter can register the on/off activity of Solenoid C, it would be next to impossible to figure out the effective torque split. The factory made and licensed Subaru Select Monitor was designed to do just this...

-Chike

[Chiketkd]

Originally Posted by michael

OK so does this mean there is no difference between the usdm and jdm transmissions?

There's a huge difference. The JDM and 99+ USDM 4EAT's use a VTD system that electronically varies torque front to rear (no transfer clutch & Solenoid C). On the JDM models, the torque split starts out at 33.33% front/ 66.66% rear - on USDM models it's 50%/50%.

-Chike

[n00b on demand]

60/40 huh? Is that why i could take a sharp corner going 70? ....not that i have

[GreenMarine]

That is AWESOME information!!! I can't believe that everything that I have been reading and repeating to others over the past few years has been wrong (the 90/10 thing).... I can't wait till ECUTune comes out with their Engine monitoring system!!! I want one soo bad!!! Not that it matters all that much anymore since I don't have an Auto Tranny anymore

[IggDawg]

no kidding... this is VERY cool info. and I assume all this is due to the TCU, so the fact that I have (what I assume to be) a legacy transmission, it should still behave the same way? Any idea if this is standard behavior for all MPT 4EATs?

[Mr. Pockets]

Originally Posted by Beav

I didn't realize anyone thought it was a 'one or the other' type arrangement. It definitely is a continuously variable affair.

I wasn't aware that people were confused on the issue, either. I'm glad Chike posted the information for those who, unlike me, haven't had to learn a lot about the tranny because they put it back together wrong.

[svxfiles]

One of the SVXi we ran the tests in was my car which has a Legacy outback transmission (W/4.44 gears ) and another time Diana was driving her stock 92 SVX and I was on the phone to Chi ke, while Di and I ran the same tests.
Since the TCU controls the trans, they both acted the same way. Tom.

[SVXRide]

Originally Posted by svxfiles

One of the SVXi we ran the tests in was my car which has a Legacy outback transmission (W/4.44 gears ) and another time Diana was driving her stock 92 SVX and I was on the phone to Chi ke, while Di and I ran the same tests.
Since the TCU controls the trans, they both acted the same way. Tom.

Tom,
Remember that we did the same thing with the SSM in my car at the "drag meet" at Mason-Dixon. Of course the TCU was sending the 40% into a "black hole" as the Solenoid C was not functioning at the time...
-Bill (enjoying the start of a new season of autoX - especially with a functioning AWD system!!)

[87f383]

awesome info... I don't drive a gloried FWD car anymore... lol

I remember the 90/10 60/40 debates from my Yahoo days. I believed it was 60/40 by default, others disagreed. Eddycat, wherever you are, the joke is finally on you. 60/40!!




Dave

[Beav]

Good grief. Look at it this way: the power goes to the rear of the trans. At the rear of the tranny is a drive gear and a driven gear. The driven gear sends power back to the front, to the front diff. This is how a fwd trans would work. AWD is accomplished by adding a clutch to the rear of the the drive gear. The rear of the clutch is attached to the rear driveshaft. The clutch engages and sends power to the rear wheels. When the clutch is disengaged 100% of the power would go forward. When the clutch is 100% engaged the power is shared by the front and rear, or another way of saying it is .... tada!! 50-50. Physically impossible to be any more than that with the drive/driven gears @ 1:1.

I can't imagine why this would be an issue to haggle over, it is really quite simple.

Regarding the 51%, don't go overboard - it's just a scanner snafu and believe me, every scanner built has plenty of those. Why do you think techs double check the scanner findings with a meter?