ROB-80E

Thanks Dallas, I haven't uploaded any XML because I haven't finished yet. Although for the ACD tuning side of things, it's probably as best as it is going to get now. There is still aspects I'd like to figure out which would make understanding what's going on better. But I haven't had any help in this so it's a slow process and I doubt I'll be uploading what I have for the silent. The amount of hours I've spent on this is unmeasurable now.

I'm just about up to trying to make a start on the AYC side of things....it's programmed totally different to the ACD in terms of structure. I'm guessing a hang over of how it had to be programmed for the CP9A AYC ECU. Easy to see why it hasn't been explored in great deal yet.

Floppyz

I can help you to test and log on italian streets, on different evo model (RS or GSR, evo 8 or 9)... Just tell me what you need to log and in what conditions, if you need...

bye

Dallas J

iTrader: (1)

This to me is whats most important and why I'm following along here. I've been on the fence of just getting the fully integrated Motec system to run the ACD and Engine ECU along with all the datalogging capabilities just so I can actually program things directly. Right now with the ACD there's just too much lack of control for different scenarios. I do plan on logging pump pressure with my current data system to see what actions seem to cause pump pressure changes but its still a lot of guess work to know if changing one parameter effects others.

ROB-80E

I'd imagine the data logging with the motec would be much better, but as far as capabilities go, the stock ecu is very capable and would be far better to customise specific parameters. So to say there is a lack of control is far from the truth...there is WAY TOO MUCH control, and we just have to narrow down how to harness that. I believe I've made massive inroads into that.

Pump pressure is not what you want to be logging. Pump pressure only shows the pressure at the accumulator. Once the accumulator falls below a certain pressure, the pump kicks in to charge it back up again. It's the valve drive values that you want to be logging...because that is what is actually creating lockup or not.


So as a rundown again this is the "control" of the ACD that you have with the standard ECU:

P table (the big 9x36 table) - this one is REALLY hard for me to explain and understand, but I'll give it a go in a very simplified way (in which i'll probably miss some aspects off it). I believe it's essentially a Front to Rear Wheel Speed Difference target/correction map. So basically, this sets up where the ecu will know when a front to rear wheel speed difference is going to occur (Steering angle V's Speed) and the amount of difference that will occur (the table value). The higher or lower the value in the table (around hex 80), the more "expected" wheel speed difference there will be. A hex value of 80 is expecting no wheel speed difference. If there is a difference, you'll essentially get a correction value of lets say 1 (this is probably the wrong way of explaining "technically", but for understanding it works). On the low end of the scale, hex 0 is expecting lots of wheel speed difference, but the correction value will also be 0. On the high end, FF is also expecting a lot of wheel speed difference, but it will give you a correction value of lets say 2. However, with a value of FF, the way i think this works is that because the wheel speed difference needs to exceed the expected difference to get a correction value, with an FF value, i don't think you will ever get a correction value output because the wheel speed difference will never exceed the expectation (if that makes sense?). In practice, this is why we get so much better handling characteristics when all the FF's are reduced to something around C0. It gives the wheel speed difference a chance to exceed the expected value and we get a ACD lockup result from it. Lighter steering feel can also be achieved buy reducing the 80's at zero degrees to 70's or 65's, as there will be a lesser correction value when the wheel speed difference increases as you begin to turn. BUT, don't just focus on this table...it's one of many...but this is the INITIAL table. Side note - I'm still deciding how best to scale this table.

J tables (evo 8/9 only) - these are multiplied with the P map. I'm trying to understand which one of these two are a coefficient table...or maybe neither. But, if you use these J tables as a coefficient, it makes sense. A value of 1 has no effect on the P table result. A value of 0 will nullify it.

----------------------------------------------------------------------------------------------------

D tables - it's axis data is one of 2 values. (JxP) - Front to rear wheel speed delta. OR Front to rear wheel speed delta. It's a bit confusing to me how or why it chooses which result to use...there's an XOR command in there. But, it's one of those two is what you get as the axis input to achieve an ACD lockup value. A big value = 100% max lock.

E tables - Speed coefficient for D table. A value of 1 has no effect on the D table result. A value of 0 will nullify it. (ie DxE = lockup value).

When you take into account that Vehicle speed, steering angle and wheel speed difference to get a result from the P, J D and E tables is all taken into account to get a lockup value, this is a highly customizable (and very confusing) result.

----------------------------------------------------------------------------------------------------

F tables - Acceleration G-force lockup adder (self explanatory). More G-force, the more lock you get. Although, these are reasonably small lock values compared to E table. A thing to remember...the map chooses to use the DxE or F table lockup value based on what one is higher.

----------------------------------------------------------------------------------------------------

A tables - Throttle lockup adder (self explanatory). Can be specifically tuned around Throttle % and Speed.

B Tables - Steering angle coefficient. A value of 1 has no effect on the A table result. A value of 0 will nullify it. Can be specifically tuned around Steering Angle and Speed. (ie AxB = lockup value)

----------------------------------------------------------------------------------------------------

G tables - Braking G-force adder (self explanatory).

C tables - Steering angle coefficient. A value of 1 has no effect on the G table result. A value of 0 will nullify it. Can be specifically tuned around Steering Angle and Speed. (ie GxC = lockup value)

H tables - Steering angle velocity coefficient. A value of 1 has no effect on the CxG result. A value of 0 will nullify it. (ie GxCxH = lockup value)

Like the DxE and F lockup value, the rom will choose whether to use the GxCxH lockup value or not. The AxB result is always in the equation, however GxCxH lockup value is also added if it's larger.

----------------------------------------------------------------------------------------------------

End calculation looks like (GxCxH + AxB) or just AxB + DxE or F

Time constants of all these tables also comes into play. Whether they play a role in just attack and decay times, or also provide additional weight (multiplication) to the lockup value i'm still not sure.

But that's pretty much it for how the ACD gets an output. Take all of that and times it x3 for each mode (TARMAC GRAVEL and SNOW)...except P table, that effects all 3 modes. Oh, and x3 again (the pin 50 configuration switch...pin 50 grounded = No AYC and no ABS).



So Dallas, still think that "there's just too much lack of control for different scenarios"? hehehe

Dallas J

iTrader: (1)

Great post. I should qualify my statements just a little I suppose.

By lack of control, I meant more lack of knowing what effect a table may have beyond its stated effect. Or the P-table and how that can be used. Im still trying to understand how I can sort out something like mid corner grip vs a near spin where I may have similar G force but opposite steering angles.

For pressure logging, I dont mean the at the ECU or intercepting pump sensor voltage. I have a sensor to put inline with the line too the ACD. My pump is relocated so its easy for me to add a fitting and sensor inline which I have but just need to run. Seems it should give me a bit of direct info on the ACD condition that I can directly related to whats happening on course.

Your post is very comprehensive to your finding, going to need to spend some time reading and digesting . For now though, I need to go pack for racing this weekend.

ROB-80E

The AYC control block is officially doing my head in. Will be a while before we see anything on this I'd say. Definitely need help with the programming path flow side of things. But not only that, trying to follow what gets multiplied/added/subtracted/compared with what is a nightmare! Made good progress of the groupings of tables and not too bad progress with what inputs are what...still a few unknowns though. But yeah, no clear understanding of what table specifically does what.

Floppyz

And thinking that AYC is not only lockage system but most important a right/left torque transfer. That cause the rear weels can "steer" the car from the back, but doing this in different way if we are going in the inside the corner, breaking, steerin, slideing or accelerating....
I think japans ingineers brain melted like ***ushima some years ago.... but always a great stuff at least!

I drive lot of GSR and RS evos and the first feeling you have with AYC is like, when you reach car limit in cornering asking little bit more, an invisible "big hand" hidden in the sky, gently touch the back of you car limiting the slideing and makeing everything easyest (end sometime safer)... but that dosen't means that car is cornering faster...

Even if I'm an electronic ingineer and love automations and its fine tuneing, personally I prefer RS mecchanical diff. just because the car do exactly what your instinct expect what the car should do in that particolar moment even if is good or bad thing, so no hidden invisible hand that decide car behavior annoying driver's sensations... Finally i think car is more performing due the better driver feeling improvenig the pace in corner drivenig, the result is that the car cornering faster!

I hope I explained myself...


bye

kyoo

iTrader: (29)

bumping this for the tuners - can anyone set up for me an acd tune that's open in all scenarios except on power? with various degrees of lock once on power?

if toying with stuff, hit me with a PM so I can test!

GTA.RS

Are you after light steering just before half throttle and then ramp the A maps reaching wot ?

kyoo

iTrader: (29)

something alone those lines - just less resistance on turn in and through the corner itself, and a gradual lockup as you power out of the corner

GTA.RS

Merlin is a mate of mine down here and I have been making several bins and testing the results. Rob has been getting right into the ACD lately. Going grey I bet
You can try my P map to test for the steering and powering out of corners. Are you running RS rear ?

kyoo

iTrader: (29)

I've got a 1.5 way clutch diff in the rear, helical in front, and currently have a modified acd.

what's your P map? shoot me a PM to talk more please!

GTA.RS

Kyoo you can try both of these P map numbers and test how they feel. I've tracked both.

ROB-80E

There's 2 ways you can get gradual lockup on power out.
A map (Thottle lockup table) combined with B map. As my posts say above, B map is a steering angle lockup reduction, so use this to reduce the lockup amount of the throttle when the steering wheel is turned.

You can also reduce the values in the F map (Accel g-force). Doing this will just provide less lockup in general when accelerating.

Better turn in...as Aldo said, reducing the first column values of the P map works well (9x36 table thought to be AYC). But you should also tune the C map which is the deceleration lockup reducer table. To me it sounds like B and C tables could almost be identical for you with values only in the first to columns.

kyoo

iTrader: (29)

sounds simply, but complicated in that i don't have the ability/software/hardware to make these changes. @Dallas J , thoughts on the comments above compared to what you've been doing?