The SAE paper is pretty good and you can access for free via JSTOR btw. It explains principles but not values and in some scenarios the "gain" is very small e.g. the engine braking/abs effect, is barely detectable even on track. Due to the date it probably relates to the R32 system and things had evolved by the R34, so it may not be complete for the R34.
You may be aware but a popular modification is to allow the driver to attenuate the lateral g sensor output to give the impression of running on a lower mu surface and increase the transfer case clutch response speed and magnitude. Looking at logged data it is surprising how little 4wd is used without this modification. It makes it much more forgiving.
Something else not covered in the SAE paper is that the R32 ran at 0 clutch pressure until called on and this resulted in up to a 500ms delay before torque transferred to the front wheels. This is part of the reason for the characteristic big slip angles you see in video of R32s. The 33 and 34 addressed this by running a constant base pressure on the clutch pack, this speeds up the response but also means they were never fully rwd like a 32.
The "Fast start" feature is mentioned in the paper but not explained. Essentially it just looks for >20 throttle opening when stationary and pre-pressurises the clutch ready for a launch.
I have some some logged R32 GTR data from Silverstone if that would help (17mb csv). It has 4wd clutch line pressure and long/lat g and throttle position (no wheel speeds sadly, which according to the sae paper is at least half the picture).
Hey,
I didn't know lateral G modification was also actively done, but I did suspect it because some kits allow a gain control on it. LonG doesn't seem to cut it when trying to match something like the Mine's car. That's good info, thanks.
R32 case didn't run at 0 pressure IIRC, just less. 33 and 34 internal pressure was increased for faster response speed, however it did not mean they were always transferring torque. That seems to be a bit of a myth perpetuated for a few decades now and anyone who tunes the cases told me that it's just to allow for faster response. The extra pressure in the case isn't even enough to transfer any torque IIRC. If you want you can push it even more than the 33 and 34 did.
You can get near-instant response even out of 32 cases, but it doesn't mean there's a "constant 5% transfer" or whatever else people make up. But that's just what I was told.
It's a bit difficult to simulate the response time accurately because there's no actual lag value
, I'd have to use heavier filtering which is...not the same thing, so I've opted to just take what I can get. R32 drives very good on-video out of corners so I'm not super concerned.
If you would like to send me that file, drop me a PM, I'd gladly take a look at it.
At this point I have most of the curves already and it's more of an implementation method limitation thing, but data is always nice to have.
PS: Would you happen to know how much power R34 or R33 RB26's make? I have a very good suspicion that the current power output is much too low, perhaps even 20, 30ps. It's closer to an R32 than anything.