Centre diff lock on AWD auto

[idosubaru]

this variable control has been done before on a 4EAT. unfortunately he is not willing to talk about how he did it, but he essentially built a stand alone TCU for the 4EAT and could control nearly everything. i did talk to him about the duty C thing years ago and he mentioned it would work but wouldn't talk about it any more than that.

[nipper]

this variable control has been done before on a 4EAT. unfortunately he is not willing to talk about how he did it, but he essentially built a stand alone TCU for the 4EAT and could control nearly everything. i did talk to him about the duty C thing years ago and he mentioned it would work but wouldn't talk about it any more than that.

 

doesnt play well with others?

 

Hrmm i never thought of a stand alone tcu.....

 

 

nipper

[CNY_Dave]

OK, gents, I understand this now, don't need to kick off another 'explanation war' jusy for me-

 

What I didn't see, has anyone located the magic wire for an '03 OBW, and is that wire going to be the same for the H6 3.0 4-speed automatic?

 

I'm only looking for 'normal mode' and 'front-rear clutchpack locked', just as a worst-case-scenario fallback.

 

 

Can't believe no one has scoped this signal yet, though!

(Probably should not mention I'm an electrical engineer, heh.)

 

 

To add to the melee, has anyone determined if this wire goes directly to the solenoid coil, or does it go through a power transistor/solid-state relay somewhere down at the solenoid?

Wasn't sure on that from the thread. Sounds like 12V gets sent down this lead (pulses), but could the trans end of this wire have +12V on it, and it gets grounded through a pulsing circuit?

 

 

Dave (ducking)

[johnceggleston]

OK, gents, I understand this now, don't need to kick off another 'explanation war' jusy for me-

 

What I didn't see, has anyone located the magic wire for an '03 OBW, and is that wire going to be the same for the H6 3.0 4-speed automatic?

 

I'm only looking for 'normal mode' and 'front-rear clutchpack locked', just as a worst-case-scenario fallback.

 

 

Can't believe no one has scoped this signal yet, though!

(Probably should not mention I'm an electrical engineer, heh.)

 

 

To add to the melee, has anyone determined if this wire goes directly to the solenoid coil, or does it go through a power transistor/solid-state relay somewhere down at the solenoid?

Wasn't sure on that from the thread. Sounds like 12V gets sent down this lead (pulses), but could the trans end of this wire have +12V on it, and it gets grounded through a pulsing circuit?

 

 

Dave (ducking)

yes, those are the questions.

 

i'm glad we have an electrical engineer attacking the problem. please let us know what you learn.

 

i am not an innovator, but i will follow the crowd.... on occasion...if it feels right.... and it can't hurt.

 

good luck.

[CNY_Dave]

yes, those are the questions.

 

i'm glad we have an electrical engineer attacking the problem. please let us know what you learn.

 

i am not an innovator, but i will follow the crowd.... on occasion...if it feels right.... and it can't hurt.

 

good luck.

 

I looked at the pics on the other thread, and unless there's a control module in the trans, it looks like the wire goes straight to the solenoid.

 

If I had any spare scratch, I'd get a USB-port oscilliscope for my laptop and go record traces under various conditions. Ideal weather coming up for that just about now...

 

Easy way to do it is to put a 0.1-ohm resistor in series in the wire and look at the voltage difference on the 2 sides of the resistor- if you look at the +12V side of a feed going to a solenoid, it can sometimes be tough to see what's going on. Using a resistor and taking Vdiff also lets you measure the current through the wire.

 

 

Dave

[idosubaru]

i have not seen documentation on the wiring for your vehicle to tell you what wire it would be. i would imagine the Subaru dealer should give you a quick glance of the FSM. i've even had them photocopy small portions before.

 

I'm an electrical engineer,

if that was all it took we would have this sorted by now. there are other engineers here, (Aerospace here), and we all have the ability, contacts and some even have the equipment available. and it has been done, but not shared. it's the desire and willingness to share that is lacking!

 

the simple switch method is a fantastic solution and there is really no need to develop a controller (which has already been done) for it. what is needed is substantial indicators that this circuit is in use to avoid leaving it on. that is far more valuable than a high tech controller in my...one who has done this for years...opinion.

[Andyjo]

I've played with it before... i think you just nominated yourself to hookup the oscilliscope

I think it was thhhheee..... 11th pin?

http://www.ultimatesubaru.org/forum/showthread.php?t=50264

that should give you a little more info... as for locating it... i haven't heard of anyone doing it w/ a newer (00+) roo'

I would totally suggest you run it through a relay, and not just a switch, there's alot of current that goes through there, and it can burn out your switch (like it did mine :-\ )

And yeha, if you leave it on, and descide to tow a trailer 265 miles on the highway... yeah.. goodbye solenoid.

[CNY_Dave]

i have not seen documentation on the wiring for your vehicle to tell you what wire it would be. i would imagine the Subaru dealer should give you a quick glance of the FSM. i've even had them photocopy small portions before.

 

if that was all it took we would have this sorted by now. there are other engineers here, (Aerospace here), and we all have the ability, contacts and some even have the equipment available. and it has been done, but not shared. it's the desire and willingness to share that is lacking!

 

the simple switch method is a fantastic solution and there is really no need to develop a controller (which has already been done) for it. what is needed is substantial indicators that this circuit is in use to avoid leaving it on. that is far more valuable than a high tech controller.

 

 

I only mentioned I was an EE because I thought it'd add fuel to the fire, naughty of me.

 

For me, it'd be a switch-of-last-resort, so yeah, a controller is of limited usefulness I think. Something to just blip the sol. enough to prevent serious bind would be nice, as nipper mentioned.

 

There has to be some signal somewhere that varies nicely with speed and a comparator can be put on it to dump a relay when a threshold is reached.

 

How about a nice simple +12V timed relay that you have to keep hitting a button every minute/2-minutes/5-minutes to keep the relay energized?

 

That should be avail. COTS from somewhere.

 

 

Dave

[Andyjo]

Wow... that's some crazy EE talk... damn sparkies....

I believe someone looked up the signals once.

I thing there is an underlaying signal (square wave), which is then pulsed in another set of squares, and controls by the pulse width (does that make any sense to you?)

I started working on a circuit design at once point... i think 2 555 timers and a potentiometer to control the pulsewidth somehow... yes... muhahaha

the frequencies of these things.. i have no clue :-p

[CNY_Dave]

Wow... that's some crazy EE talk... damn sparkies....

I believe someone looked up the signals once.

I thing there is an underlaying signal (square wave), which is then pulsed in another set of squares, and controls by the pulse width (does that make any sense to you?)

I started working on a circuit design at once point... i think 2 555 timers and a potentiometer to control the pulsewidth somehow... yes... muhahaha

the frequencies of these things.. i have no clue :-p

 

I could see how that'd work- the "anti-bind but otherwise full-lock" signal could be sent out at all times, with a "modify the torque split to whatever's relevant" signal overlaid on top of that.

 

Or, it could be 'always send out this square wave' and then 'add this square wave on top but vary the phase to vary the pulse width' (and would at the same time vary the pulse spacing). Either should be something you could tell what it was with a scope.

 

But I agree with the sentiment that building a controller for this is overkill, cats meaow is just something that will give you the 'full lock but just enough release to keep it from binding', combined with a way of not accidently leaving it on.

 

 

Dave

[Andyjo]

I could see how that'd work- the "anti-bind but otherwise full-lock" signal could be sent out at all times, with a "modify the torque split to whatever's relevant" signal overlaid on top of that.

 

Or, it could be 'always send out this square wave' and then 'add this square wave on top but vary the phase to vary the pulse width' (and would at the same time vary the pulse spacing). Either should be something you could tell what it was with a scope.

 

But I agree with the sentiment that building a controller for this is overkill, cats meaow is just something that will give you the 'full lock but just enough release to keep it from binding', combined with a way of not accidently leaving it on.

 

 

Dave

I believe it's the latter of the two methods, but i could be (most likely) be wrong.

how about a switch like this:

i couldn't find the uber secure usb hub on thinkgeek... requires 2 keys, some flip covers, one of those big levers.....

 

and yeah, if you could get it to be 'almost full lock' that would be good, because i think when it's binding, the ATF Pressure in the valves below the solenoid build up enough, and blow out the oring on the bottom of the solenoid. (or that's at least what mine looked like )

[Andyjo]

Here's what we need!

[johnceggleston]

I believe it's the latter of the two methods, but i could be (most likely) be wrong.

how about a switch like this:

i couldn't find the uber secure usb hub on thinkgeek... requires 2 keys, some flip covers, one of those big levers.....

 

and yeah, if you could get it to be 'almost full lock' that would be good, because i think when it's binding, the ATF Pressure in the valves below the solenoid build up enough, and blow out the oring on the bottom of the solenoid. (or that's at least what mine looked like )

 

i like the 2 key idea, so you can only actvate it when you have a passenger, that way there will always be some one else to blame.

[r81gsr95]

so, this modification, when engaged (switch on) would make the car behave exactly like the manual transmissions, which already have a 50/50 split?

 

 

is this the case? This sounds like a very cool modification for anyone with an automatic. I am looking for an auto and dislike the 90/10 split. I love the 4wd system on my GL and when it is locked, it is amazing in the snow. So with the newer outback legacys that have the automatic transmission, can you engage and on the switch allowing the differential to give a more 50/50 split of the power? Thanks 

[idosubaru]

is this the case? This sounds like a very cool modification for anyone with an automatic. I am looking for an auto and dislike the 90/10 split. I love the 4wd system on my GL and when it is locked, it is amazing in the snow. So with the newer outback legacys that have the automatic transmission, can you engage and on the switch allowing the differential to give a more 50/50 split of the power? Thanks 

 

i would ignore the comment you quoted since there are multiple manual transmission set ups and varieties...technicalities aside, this switch simple "locks" the auto's just like your older GL when locked.  you can't drive it on normal streets like this - it'll hop and buck around turns - it's a fantastic mod.

 

I'm unsure what it does or what would be different with newer VDC transmissions so tread carefully if you're talking about one of those - they've been available since 2001.  2001-2004 in H6 OB's only, maybe others in later years.

[johnceggleston]

as gary mentioned there are pros and cons to this mod. pro, 50/50 split. con / 50/50 split.

the same thing that makes it great in mud , snow, or sand, makes it suck on dry pavement.

 

and it is not exactly like the viscous coupling center diff in the manual trans. that still allows for a speed differential front to rear in tight turns. this mod does not.

 

but a few things you may not have read, the computer and the transfer clutch in the auto trans work really really well in most circumstances.

when a front wheel spins, more power is sent to the rear. it does it automatically even if you do not there is a need. and in most cases this is all you need. it also kicks in at WOT. and when the shift selector is in 1 or 2, it stays in that gear, not shifting up or down (except at ultra high rpms to protect the engine.)

 

and so i ask the question, how often at hiway speed do you need locked 4wd. how oftwn in 3rd do you need locked 4wd.

if you want to race run / race your car on a dirt track at fairly high speeds and want locked 4wd, have at it.

or if it snows in your area a lot and you drive on snow covered roads to work most of the winter and faster is better. go for it.

just remember to turn it off when the pavement is dry.

 

imho, the perfect switch is a ''latching'' one like on the cruise control, . you turn it on when ever you want, but when you turn off the car, it turns off the 4wd ''lock''. trans is back to normal.

 

a limited slip rear diff probably costs more but will greatly enhance your ''off road'' or bad weather capability with much less chance of damaging your drive train. adding both a latching switch, and a LSD would be killer.

[r81gsr95]

wow this is very cool. Yeah it would only be for deep snow. I'm moving to wyoming and i've heard the winters can be quite crazy there. I believe a legacy automatic with good appropriate tires would be more than enough to get me around in the winter time, however i really like the idea of being able to flip a switch and "lock" the diff. I'd probably not use it much but if i ever did get into some deep powder and was having trouble getting out, very cool mod to be able to lock it and get myself moving again. 

[idosubaru]

i've installed the switch a bunch of times and love it, i'm a big fan - but even that doesn't help with low grade or average tires or you get high centered.  if you have unmaintained roads/snow to contend with then dedicated snow tires and even studded are the key to safety and reliably and consistently getting where you need to go.

[CNY_Dave]

I have done a lot of snow driving, and briefly this winter I had a switch instead of the TCU controlling the transfer solenoid.

 

I have never been in a situation in the snow, on road or in my yard, where the switch would have done a lick of good.

 

Off-road, mud and rocks, where you can transition quickly from traction to none and back, I think a switch might be useful.

 

I also learned that turning off the AWD when going down slick hill can be a good thing. I had my switch locked, and a snowy downhill with careful braking resulted in the rear end sliding out, because braking force from the front tires was also being transferred to the rear tires.

 

With the TCU in control, throttle off = minimal AWD lockup.

[nipper]

The manual w/o variable control, acts just like an opn diff until it detects slip, then the vscous coupling hooks up to deliver power where it has to go. On dry ground there will be 50/50 split with a plain ol diff. The VLS retursn the power split to 50/50 but there will be some slippage untill it hooks up as this is a mechanical system. An automatic uses the computer to pulse the solenoid on and off. There is also a spool valve to control the amount of pressure to the clutch. the clutches are designed to tolerate some slippage, where as a normal clutch pack can not tolerate any slippage. The combination of all three with inputs from a yaw sensor, throttle, VSS's, and inhibitor switch will instantly manage any slippage. I think the yaw sensor makes it more proactive then other systems. 

 

I will look and see if I can find it, but someplace I saw the output signal for the C, its a square wave. 

 

Ideally, leave it be and its fine. I think the best way to deal with it is like other mfg's do with lockable center diff. They read the pulses for the spedo and when the speed gets to X for the car, the center diff automatically unlocks. This would be ideal to keep one from accidently keeping the clutchpack locked and then making a 500-900.00 mistake.

[torxxx]

I ran 50/50 all winter in my SVX.  Love it that way, its wired in with a switch so I can go back to awd when I'm on pavement.   Lap times on the ice track were considerably faster in 50/50 mode.  I wish I had the e-brake/2wd switch rigged to it, would make amazing driving.  

[nipper]

But 2wd would have made it just a FWD car dragging around a rear LSD diff. SVX also (if unmolested) had a differently calibrated AWD program then the other cars

[eulogious]

Switch makes a world of difference when you are stopped facing up a hill.  Without it front wheels spin and the car starts to slide backwards and to the right, try it again with the switch, all four tires spin at once and I start moving forward.

 

The problem with the AWD is that there is about a 1/4 sec lag between slip and power transfer (for the TCU to calculate the difference and then activate the rear wheels) and if you are stopped on an icy anything, that 1/4 second can make the difference between being stuck and getting moving again.

 

I remember being in a ski resort parking lot years ago when I had my dual range gl-10 and I was stopped on an incline with a early 2000's outback in front of me.  Once the line started to move, the outback was completely stuck.  The front would spin before the back kicked in and they would start sliding backwards.  We all had to back down the incline and the let the outback have a run at it.  Once the outback got up the hill, I had to stop and wait on the hill again for my turn to park.  Once I was signaled to go, I took off with zero issues in the same spot the outback got stuck.  Having all 4 wheels turn at once is crucial to get going again if you are stopped, really on anything icy.

[torxxx]

nipper the idea behind the 2wd ebrake is so you dont have to let off the gas.  You keep your tires spinning as the rear drags around corners (hairpins especially)  As soon as the ebrake light goes off its back into 50/50 (cant ebrake in 50/50, it loads up the clutchpack, and you end up getting 4 wheel braking.   VLSD doesnt spool up as fast as everything thinks, specially with 200k miles on it.  If I had a clutch type, or a torsen I wouldnt even worry about the ebrake switch.  I'd just have to learn how to pre drift a little more.  

 

And the USA SVX's have the same 90/10 AWD system that the Legacy/Impreza/Foresters had.  The Euro/Asian ones used different programming, unless I'm missing something here

 

 

And the tranny in the SVX already has a delayed shift issue if you dont have Quick Change installed.  I run with my shift sol resistor unplugged when I'm on the track.  Then again I'm doing it on ice, where instant tire lock up doesnt really stress the tranny like it would on tarmac.  The 4.44 4EAT is amazing, I've ran that tranny harder than any other auto and have had zero issues with the tranny its self.  Matter of fact, if I ever score another SVX, the 3.54 4EAT is coming out immediately and its getting a 4.44.   Makes it a real car then.  Its just been with the tranny computer programming. between too slow of shifts and AWD delay.   IMO all 4EATs should have came with VTD.  At least then its a fully useable 4 wheel drive system.  

Edited March 22, 2013 by torxxx

[nipper]

Ah ok, i didn't quite understand the parking brake relationship that you were speaking of.

 

The gearing is different, I thought the AWD was different too, my mistake