Tuning Evo789 ACD
As we're very much aware, the ACD controls the level of slippage between front/rear. But given the above, the front axle is allowed zero slippage vs the output ratio, making it impossible to transfer power away from it.
Just wanted to point that out.
Newbie
Joined: May 2009
Posts: 29
Likes: 0
From: California
With the design of the driveline, you can't transfer 100% of the power to the rear, the best you can achieve is a 50/50 lock. The driver's front axle is direct-driven from the trans, and the passenger front is controlled by the front diff on the opposite side of the ring gear (which is direct driven on effectively the same shaft).
As we're very much aware, the ACD controls the level of slippage between front/rear. But given the above, the front axle is allowed zero slippage vs the output ratio, making it impossible to transfer power away from it.
Just wanted to point that out.
As we're very much aware, the ACD controls the level of slippage between front/rear. But given the above, the front axle is allowed zero slippage vs the output ratio, making it impossible to transfer power away from it.
Just wanted to point that out.
Example: one side on the pavement, the other side in the mud. In an open diff the wheel in the mud will get the most, if not all the power.
Front power is not direct. It feels direct because: the rear LSD does not lock properly, so its front dominant.
A hollow shaft is connected to one of the gears in the center diff and sends power to the Rear.
Inside the hollow shaft, there is another thinner hollow shaft that is connected to the opposing gear and sends power to the Front (front diff). The front diff then splits the power
And then the ACD restictes the differences.
....so layed out from driver to passenger
Center diff(open), Front diff , ACD (plates)
And the downgraded Rear diff
4 diff's moving around power
What you describe is more of a haldex differential. God forbid ! If it had a haldex I would not have bought an Evo.
I'm typing this from my phone, there are plenty of Evo ACD images/blue prints online
Last edited by Malocas; Jan 26, 2014 at 01:05 PM.
More knowledgeable people can chime in if I got this wrong, but I'm going to try and explain ACD in laymans terms as how I understand it.
First there is no torque vectoring going on. The power is not actively shifted anywhere. You cannot "send" torque rearwards by engaging more ACD lockup. This is not a GTR where those torque vectoring capabilities are present at the center diff.
The ACD is purely meant to lock up an otherwise open center diff. Or in otherwords, an open diff with an electronically controlled variable locker. It can go from 0% locking force (fully open diff) to 100% locking force (fully locked diff). That means if you are coming out of a corner and there is wheel slip at the front or rear axle, with a non-functioning ACD the c.diff will shunt all that power to the slipping wheels and you can't power out. With a functioning ACD the power will be split more evenly (depending on the % lockup) and the axle with traction does not lose torque.
As anybody that has driven a 4x4 with 4Hi / 4Lo mode knows, a locked diff doesn't like to go around corners very well. STi guys can experience this by pressing the DCCD lock button on their c.diff controller. Evo guys can sort of experience some of this by switching from tarmac to snow mode. There is some invisible force thats hindering the car from rotating. Thats why some cars like the lotus elise do not have a locking diff as standard. It doesn't rotate into corners as eagerly and the car is underpowered anyways so theres usually no issue with wheel slip to need a locking diff.
So basically, use the ACD to manage loss of traction. For the given speed, g-forces, steering input, etc. you want to balance maintaining traction with letting the car rotate. And you achieve this balance with managing how much ACD locking force is applied. Generally you would want to have as little locking force as possible without compromising traction too much. There is an optimum balance. Usually lower speeds will favour more lockup (better traction) and higher speeds will favour more open diff.
^that is what I understand of this whole ACD situation. Which is not to say I'm an expert on the topic so people that know more can edit this if they want.
First there is no torque vectoring going on. The power is not actively shifted anywhere. You cannot "send" torque rearwards by engaging more ACD lockup. This is not a GTR where those torque vectoring capabilities are present at the center diff.
The ACD is purely meant to lock up an otherwise open center diff. Or in otherwords, an open diff with an electronically controlled variable locker. It can go from 0% locking force (fully open diff) to 100% locking force (fully locked diff). That means if you are coming out of a corner and there is wheel slip at the front or rear axle, with a non-functioning ACD the c.diff will shunt all that power to the slipping wheels and you can't power out. With a functioning ACD the power will be split more evenly (depending on the % lockup) and the axle with traction does not lose torque.
As anybody that has driven a 4x4 with 4Hi / 4Lo mode knows, a locked diff doesn't like to go around corners very well. STi guys can experience this by pressing the DCCD lock button on their c.diff controller. Evo guys can sort of experience some of this by switching from tarmac to snow mode. There is some invisible force thats hindering the car from rotating. Thats why some cars like the lotus elise do not have a locking diff as standard. It doesn't rotate into corners as eagerly and the car is underpowered anyways so theres usually no issue with wheel slip to need a locking diff.
So basically, use the ACD to manage loss of traction. For the given speed, g-forces, steering input, etc. you want to balance maintaining traction with letting the car rotate. And you achieve this balance with managing how much ACD locking force is applied. Generally you would want to have as little locking force as possible without compromising traction too much. There is an optimum balance. Usually lower speeds will favour more lockup (better traction) and higher speeds will favour more open diff.
^that is what I understand of this whole ACD situation. Which is not to say I'm an expert on the topic so people that know more can edit this if they want.
Last edited by deeman101; Jan 26, 2014 at 03:35 PM.
Newbie
Joined: May 2009
Posts: 29
Likes: 0
From: California
More knowledgeable people can chime in if I got this wrong, but I'm going to try and explain ACD in laymans terms as how I understand it.
First there is no torque vectoring going on. The power is not actively shifted anywhere. You cannot "send" torque rearwards by engaging more ACD lockup. This is not a GTR where those torque vectoring capabilities are present at the center diff.
The ACD is purely meant to lock up an otherwise open center diff. Or in otherwords, an open diff with an electronically controlled variable locker. It can go from 0% locking force (fully open diff) to 100% locking force (fully locked diff). That means if you are coming out of a corner and there is wheel slip at the front or rear axle, with a non-functioning ACD the c.diff will shunt all that power to the slipping wheels and you can't power out. With a functioning ACD the power will be split more evenly (depending on the % lockup) and the axle with traction does not lose torque.
As anybody that has driven a 4x4 with 4Hi / 4Lo mode knows, a locked diff doesn't like to go around corners very well. STi guys can experience this by pressing the DCCD lock button on their c.diff controller. Evo guys can sort of experience some of this by switching from tarmac to snow mode. There is some invisible force thats hindering the car from rotating. Thats why some cars like the lotus elise do not have a locking diff as standard. It doesn't rotate into corners as eagerly and the car is underpowered anyways so theres usually no issue with wheel slip to need a locking diff.
So basically, use the ACD to manage loss of traction. For the given speed, g-forces, steering input, etc. you want to balance maintaining traction with letting the car rotate. And you achieve this balance with managing how much ACD locking force is applied. Generally you would want to have as little locking force as possible without compromising traction too much. There is an optimum balance. Usually lower speeds will favour more lockup (better traction) and higher speeds will favour more open diff.
^that is what I understand of this whole ACD situation. Which is not to say I'm an expert on the topic so people that know more can edit this if they want.
First there is no torque vectoring going on. The power is not actively shifted anywhere. You cannot "send" torque rearwards by engaging more ACD lockup. This is not a GTR where those torque vectoring capabilities are present at the center diff.
The ACD is purely meant to lock up an otherwise open center diff. Or in otherwords, an open diff with an electronically controlled variable locker. It can go from 0% locking force (fully open diff) to 100% locking force (fully locked diff). That means if you are coming out of a corner and there is wheel slip at the front or rear axle, with a non-functioning ACD the c.diff will shunt all that power to the slipping wheels and you can't power out. With a functioning ACD the power will be split more evenly (depending on the % lockup) and the axle with traction does not lose torque.
As anybody that has driven a 4x4 with 4Hi / 4Lo mode knows, a locked diff doesn't like to go around corners very well. STi guys can experience this by pressing the DCCD lock button on their c.diff controller. Evo guys can sort of experience some of this by switching from tarmac to snow mode. There is some invisible force thats hindering the car from rotating. Thats why some cars like the lotus elise do not have a locking diff as standard. It doesn't rotate into corners as eagerly and the car is underpowered anyways so theres usually no issue with wheel slip to need a locking diff.
So basically, use the ACD to manage loss of traction. For the given speed, g-forces, steering input, etc. you want to balance maintaining traction with letting the car rotate. And you achieve this balance with managing how much ACD locking force is applied. Generally you would want to have as little locking force as possible without compromising traction too much. There is an optimum balance. Usually lower speeds will favour more lockup (better traction) and higher speeds will favour more open diff.
^that is what I understand of this whole ACD situation. Which is not to say I'm an expert on the topic so people that know more can edit this if they want.
Last edited by Malocas; Jan 26, 2014 at 04:56 PM.
Thread Starter
Evolved Member
Joined: Jun 2008
Posts: 824
Likes: 24
From: Sydney
So far this latest round of 4WD-ECU explorations have centered on ACD control, with little or no effort put into the AYC side of the controller code.
But, along the way, evidence has been accumulating that AYC control, as we previously thought, is likely to be quite wrong.
The AYC Proportioning Valve output drive location has been identified, there are three peak drive levels and they are different for Tarmac, Gravel and Snow on the C1 ROM.
Also, the left/right valve switching has been found, not that we need to mess with that item.
More importantly, the 3D map at 0x12171, previously thought to be AYC Rear Diff Proportioning % Torque, is not in fact in the AYC routines.
It would appear to be a pre-conditioning map for the wheel slip D*E group.
Lowering the usual value of 100 to 50 in this map (as in Evo7 K2 Rally ROM) was thought to make the AYC behave with more of a 50/50 split side to side and give the car more oversteer.
But, if it reduces the effect of D*E it will reduce lockup when the wheels are slipping, and so also promote a bit more oversteer.
Essentially, lowering the values from 100 to 50 or 65 etc will make the ACD a bit looser with wheel slip. Which on tarmac cornering is probably what we want anyway.
Clearly this AYC stuff needs more work, hopefully we will get a clearer understanding of it soon. In the meantime, I would suggest leaving that particular table alone for now.
But, along the way, evidence has been accumulating that AYC control, as we previously thought, is likely to be quite wrong.
The AYC Proportioning Valve output drive location has been identified, there are three peak drive levels and they are different for Tarmac, Gravel and Snow on the C1 ROM.
Also, the left/right valve switching has been found, not that we need to mess with that item.
More importantly, the 3D map at 0x12171, previously thought to be AYC Rear Diff Proportioning % Torque, is not in fact in the AYC routines.
It would appear to be a pre-conditioning map for the wheel slip D*E group.
Lowering the usual value of 100 to 50 in this map (as in Evo7 K2 Rally ROM) was thought to make the AYC behave with more of a 50/50 split side to side and give the car more oversteer.
But, if it reduces the effect of D*E it will reduce lockup when the wheels are slipping, and so also promote a bit more oversteer.
Essentially, lowering the values from 100 to 50 or 65 etc will make the ACD a bit looser with wheel slip. Which on tarmac cornering is probably what we want anyway.
Clearly this AYC stuff needs more work, hopefully we will get a clearer understanding of it soon. In the meantime, I would suggest leaving that particular table alone for now.
Last edited by merlin.oz; Jan 27, 2014 at 06:01 PM.
I just opened a 401702 ACD rom and noticed that the definitions for the B, C, E, F and G maps are invalid; those tables can't be opened.
This appears to be caused by missing scalings:
SlipFactor for B maps
SteeringAngle for C maps
LockFactor for E, F and G maps
This appears to be caused by missing scalings:
SlipFactor for B maps
SteeringAngle for C maps
LockFactor for E, F and G maps