Variable Torque Distribution in a Gen 1 Liberty

Liberty Valance · Post by **Liberty Valance** » Fri May 25, 2012 4:39 pm

Hoping to continue a previous thread with a fresh topic of discussion
4EAT MPT Clutch Assemby (<== See this link for some pictures and history)

To summarise the story. I have purchased a 1990 Gen 1 Liberty AWD Sedan and discovered that the auto transmission is not original.

Instead of a Multi Plate Transfer (MTP) Clutch it has Variable Torque Distribution. VTD was available in Australian delivered SVX in the early 1990's and it is likely some would have entered Australia in front cuts of imported RS turbos.

The car has an original TCU, Torque Converter and Front Final drive (diff 4.11), the 4EAT Auto transmission and VTD transfer housing have been bolted on behind those components.

I understand the VTD system uses planetary gearing to bias torque delivery 64:36 towards the rear wheels.

With Duty C Solenoid disabled (0 volts) the torque transfer is locked at 50/50.

With the FWD fuse installed (12 volts on duty C output) I get an unmistakable bias of torque to the rear wheels.

Everything until this point makes sense but here is the strange thing...

At idle I get 6 volts on the duty C output of the TCU but the second I touch the accelerator for light acceleration, normal driving or cruising I am getting 1-3 volts on Duty C and the vehicle is very much in AWD mode (close to lock up).

Has anyone got a similar setup?

What is normal range for the duty C output under normal driving?

I would have expected a range of say 9-12 volts to keep the transfer clutch disengaged unless there is a difference between Vehicle Speed Sensors or the Throttle Position Sensor indicates hard acceleration.

thunder039 · Post by **thunder039** » Fri May 25, 2012 5:15 pm

would it be the system making it closer to a 50/50 split for acceleration for a reason? iam aware that the subaru system does measure throttle and causes a change in torque distribution that way aswell a speed differences

Gannon · Post by **Gannon** » Fri May 25, 2012 5:44 pm

Before we get any further into this, you must first understand that the voltage to any of the solenoids is always +12v. Its the duty cycle that varies. Duty cycle is the ratio of 'on' and 'off' time that the voltage is at +12v. So say for example, if the voltage switches on for 5ms (milliseconds)and off for 5ms, thats a 50% duty cycle. If its on for 2ms and off for 8ms, its a 20% duty cycle. These examples are of a 100Hz cycle.

This is a video taken of a duty cycle meter (not a volt meter) that shows the signal to the duty solenoid C on my 2001 Outback. It varies from 30% to 99% to illustrate the amount of lockup to the MPT. This is actually set up to read the opposite signal just to make it easier to understand, it actually should show 70% and then drop to 1% on acceleration if it were truly showing the duty C signal. Get yourself a multimeter with duty cycle measurement, it will make more sense than trying to read volts.

Anyway, in this video I leisurely drive to an intersection (you can hear my indicator) and then I floor it till I get to 100kmh. So in 1st at low speed and full throttle, it jumps to 99% and peters off to about 70% as speed increases, soon after it changes into 2nd gear, it drops to 50 something and then decreases to 40 odd until I lift the accelerator.

http://www.youtube.com/watch?v=x4k9KIH55fo

Liberty Valance · Post by **Liberty Valance** » Fri May 25, 2012 6:16 pm

Hi Gannon, thanks for taking an interest in my dilemma.
I had considered wiring in an Oscilloscope to measure the duty cycle but considering the subaru service manual is content to perform diagnostics based on DC volts I figure a good fluke multimeter will give a reasonable approximation of the DC voltage switching frequency (a bit like RMS volts on AC). By that I mean if the TCU outputs 100% duty cycle the volts will read close to 12 and if the TCU output is 0% duty cycle then volts will be close to zero. Like wise a 50% cycle will give 'around' 6 volts and so forth. Do you think this assumption is too simplistic?

Liberty Valance · Post by **Liberty Valance** » Fri May 25, 2012 6:22 pm

thunder039 wrote:would it be the system making it closer to a 50/50 split for acceleration for a reason? iam aware that the subaru system does measure throttle and causes a change in torque distribution that way aswell a speed differences

Perhaps you are right but I can't see why the vehicle would think it needs more grip when given just the smallest amount of throttle on a hard dry surface. Or cruising along at 60 in a straight line on a flat and level road.

Gannon · Post by **Gannon** » Fri May 25, 2012 6:36 pm

The volt meter will do as on the DC volts setting, it calculates an average voltage which will do the trick, but wont be exact. To get a reading reversed like mine, measure the duty cycle between the dutyC output and constant +12v rather than ground.

Im betting the difference is in the toothed wheels as this will cause lockup the moment you start to move.

I dont have a video, but in my testing with the duty cycle meter, I found that while driving the reading was around the 30% and varied with throttle. I also did some testing on dirt and found that if I was coasting with my foot off the throttle and pulled on the handbrake, the rear wheels would lock up and the reading didnt change. If I was driving with just a small amount of throttle and pulled the handbrake, the wheels would skid for maybe about 6 inches and the duty cycle would jump up and prevent them from locking up. Pretty quick really.

Liberty Valance · Post by **Liberty Valance** » Fri May 25, 2012 7:41 pm

Gannon wrote:Im betting the difference is in the toothed wheels as this will cause lockup the moment you start to move.

I'd have to pull the transfer housing off again to be sure but in the mean time I have just tried to count the teeth on the first picture I posted.
There is certainly 30 teeth on the MPT clutch basket and while I can't see the whole VTD wheel I have deduced that there are 50 inner teeth on the clutch plates and 6 outer teeth for every 10 inner teeth which also totals to 30.

Gannon · Post by **Gannon** » Fri May 25, 2012 8:20 pm

These are the tone rings you need to compare

The VTD seems to have more teeth

Liberty Valance · Post by **Liberty Valance** » Fri May 25, 2012 9:49 pm

What you are saying makes sense but I though the Vehicle Speed sensor reads the clutch drum. At least that's what it says in the service manual and when you look at the sensor it sits a few mm away from the drum. (see pic attached)

Gannon · Post by **Gannon** » Fri May 25, 2012 10:09 pm

Sorry my mistake, the toothed ring I pointed to is the parking gear. I was looking at it upside down.

vincentvega · Post by **vincentvega** » Fri May 25, 2012 10:13 pm

the 4EAT Auto transmission and VTD transfer housing have been bolted on behind those components.

How do you know this? You might just have the VTD rear end bolted onto the original box.

I would have to check my notes to be sure which is which, but when comparing the MPT and VTD clutch setup, 12V on the solenoid opens the clutches on the MPT box, but locks the clutches on the VTD box. The problem you are seeing is probably because you have an MPT TCU driving the VTD rear end backwards. You have a few options:

Run an early WRX TCU. This will suit the throttle sensor polarity of your car and is meant to drive a VTD rear end. Added bonus is power mode on this TCU is nice and agressive.

Run the standard TCU and just leave the dutyC solenoid disconnected for road use (essentially an open center diff with rear biased torque split). When you feel you need to offroad, switch 12V to the solenoid to lock it. If this causes your TCU is spitting error codes after doing this I can help you get rid of them (or just ignore it)

The speed sensor difference is also a problem. The sensor counts the outside drum of the clutch pack. 30 teeth in a MPT, and 22 in a VTD. When I ran a MPT TCU with a VTD rear end it always shifted a bit funny, and I blamed it on this. The WRX TCU is set up for this speed sensor, and shifts properly.

Post up the TCU part number and I should be able to tell you what it is out of. I have a heap of them here.

The early 4EAT is just a Jatco auto box (also used in the isuzu Mu and some skylines) with a subaru front and rear end. You can mix and match up til 99 when the box design changed.

There is tons of info around on this stuff. Search this board and http://www.rslibertyclub.org for "VTD"

vincentvega · Post by **vincentvega** » Fri May 25, 2012 10:16 pm

one other thing I think might work but have never tested.

The DutyC solenoids operate differently in these boxes but look the same. I reckon you might be able to fit the MPT solenoid to the VTD box and have your MPT TCU at least drive the clutches in the right direction.

You then might be able to use a Jaycar speedo adjuster to solve the speed sensor ratio problem (if you find its necessary)

I would love to hear the results if you do try this.

Gannon · Post by **Gannon** » Fri May 25, 2012 10:19 pm

vincentvega wrote: I would have to check my notes to be sure which is which, but when comparing the MPT and VTD clutch setup, 12V on the solenoid opens the clutches on the MPT box, but locks the clutches on the VTD box. The problem you are seeing is probably because you have an MPT TCU driving the VTD rear end backwards. \

Now that is what I originally thought cos that is what I have read in the Gen3 FSM, but I just figured I had read it wrong when reading this thread. I also followed a thread on NASIOC and a guy put a VTD rear in his XT Foz and left his TCU as it was because he believed the output between the MPT and VTD TCU's was the same. Slightly related, he also had VSS issues

Liberty Valance · Post by **Liberty Valance** » Sat May 26, 2012 6:28 am

vincentvega wrote: The DutyC solenoids operate differently in these boxes but look the same. I reckon you might be able to fit the MPT solenoid to the VTD box and have your MPT TCU at least drive the clutches in the right direction.

I possibly should have mentioned earlier that I have already put the MPT solenoid into the VTD box.

So are we saying in electrical speak that an MPT Duty C Solenoid is 'normally closed' and a VTD Duty C Solenoid is 'normally closed'?

vincentvega wrote: You then might be able to use a Jaycar speedo adjuster to solve the speed sensor ratio problem (if you find its necessary)

I would love to hear the results if you do try this.

Sounds interesting, could you elaborate on this idea a little more?

Liberty Valance · Post by **Liberty Valance** » Sat May 26, 2012 6:43 am

vincentvega wrote:How do you know this? You might just have the VTD rear end bolted onto the original box.

Just a hunch and some detective work. I have since located a receipt for an exchange transmission tucked away in the back of the owners manual so have good reason to believe the auto was swapped 6 years ago. The part number on the my bell housing still matches the VIN plate and the front diff ratio is correct. There is obvious new blue silicone gasket sealant around the joint between the front section and the auto transmission section.

vincentvega wrote: The speed sensor difference is also a problem. The sensor counts the outside drum of the clutch pack. 30 teeth in a MPT, and 22 in a VTD.

Are you 100% positive about this? I think both have 30 but that is only from a photograph. Will have to pull of the transfer housing cover again to be sure.

vincentvega wrote: There is tons of info around on this stuff. Search this board and http://www.rslibertyclub.org for "VTD"

Will do, thanks.

Liberty Valance · Post by **Liberty Valance** » Sat May 26, 2012 12:36 pm

This VTD discussion has been very interesting but I will have to sort it out another day.

Bigger problems to solve as the transmission just lost drive. Now only has 1st and reverse.

I have a donor Gen 1 Liberty available with a good 4EAT and MPT clutch.
Will have to install it next week.

Thanks, Scott