Improving rF2 force feedback?

Joe said:
Thanks for the charts. Lets look into this deep to understand this.

The correlation between FFB and SteeringShaftTorque is 0.90 (90%)
The correlation between FFB and LF_LongitudinalGroundVel -0.02 (2%)
The correlation between SteeringShaftTorque and LF_LongitudinalGroundVel -0.58 (58%)

So, yes there is some correlation between SteeringShaftTorque and LF_LongitudinalGroundVel, but it does NOT transfer to FFB. In other words, yes you see some correlation between SteeringShaftTorque and slip angle (as calculated), but not the same as seen in FFB.
Click to expand...

How can you say it doesn't transfer to the FFB when the FFB trace and the steering shaft torque trace are almost identical?

The data is right there in front of you, it's clear as day.

The FFB curve is a bit smoother than the steering shaft torque, but that's purely down to running the default smoothing level during the test (level 9).
 
Paul Loatman said:
Actually i would disagree with the second one as well, but as you say it's all subjective. In my opinion, the car starts sliding the moment i apply any steering input at all. So when i'm turning into a corner, it's already starting to slide, that torque you feel is from the sliding action of all 4 wheels as they rub against the road surface and try to follow the path of least resistance which is why the steering resists against your inputs.

As i said before, there is no sudden change from "grip" to sliding, it's a range of torque which is constantly changing depending on your throttle input, brake input, steering input, the road surface, and the environment in general. It's like understanding a language, if you don't understand it, of course you won't know what you're feeling.
Click to expand...

Paul:

In order to feel slip angle, one has to define the slip angle. In this case, as TechAde shown, the slip angle may write as: atan2($LF_LateralGroundVel, -$LF_LongitudinalGroundVel))
But, the correlation between FFB and LF_LongitudinalGroundVel -0.02 (2%), so one just can NOT feel how much is LF_LongitudinalGroundVel. Hence, one cannot feel how much the angle is!

Yes, you feel slip, this is because strong correlation between FFB and LF_LateralGroundVel. One cannot tell how much slip angle (5, 8 or 12 degree, etc) is from sole FFB.
Hope this is helpful to all.
 
I think you're stuck in that silly telemetry output values you extracted a week ago, and now you think that everything must relate to it, while probably this is not the case. Actual telemetry like what TechAde posted is much more helpful and telling of what is going on under the hood.
 
I think the problem is the tire patch, no matter what the data shows, it isn't feelable in the ffb, because the bracing of the profile particles on the tire tread not convey the feeling of rubber tenseness, therefore there is no slip angle to feel in rf2 ffb.

It's just a friction surface which seems to let you feel equal in all directions. The main resistors and their difference we feel on the steering wheel are the only elements. These are as example lateral acceleration, alignment and carcasse forces, and due to the geometry a bit defferent with each car and tire.

The most street tires don't let you feel the slip angle as well, due to the tread, rubber, softness and realtiv large slip angle area.
 
speed1 said:
I think the problem is the tire patch, no matter what the data shows, it isn't feelable in the ffb, because the bracing of the profile particles on the tire tread not convey the feeling of rubber tenseness, therefore there is no slip angle to feel in rf2 ffb.
Click to expand...

Look at the data, the FFB output drops from 60% at 2-3% slip angle to 15% at 10°(LF)/13°(RF) (difference presumably due to Ackerman).

If you can't feel that then something is seriously wrong with your FFB configuration.
 
I really can't feel it and i don't know why the data shows that result, but i don't feel something like rubber tenseness from the tire patch.
 
speed1 said:
I really can't feel it and i don't know why the data shows that result, but i don't feel something like rubber tenseness from the tire patch.
Click to expand...

Have you tried with the Skippy? It's obviously very car dependant.

By the way I'm talking purely about slip angle here, I quoted a little to much of your post to make that clear.

The tools are out there (i.e. my telemetry plugin and an app that can graph the data from csv files, Excel will do if you don't have ATLAS) for anybody to try this for themselves and attempt to correlate what they're feeling at the wheel with the data that the sim is outputting.
 
TechAde said:
Have you tried with the Skippy? It's obviously very car dependant.

By the way I'm talking purely about slip angle here, I quoted a little to much of your post to make that clear.

The tools are out there (i.e. my telemetry plugin and an app that can graph the data from csv files, Excel will do if you don't have ATLAS) for anybody to try this for themselves and attempt to correlate what they're feeling at the wheel with the data that the sim is outputting.
Click to expand...

Of course it is car and tire dependet how the overall feeling is, and also the setup makes a big difference, and i know that i once hit the hotspot with the Mazda 787, where i could feel the carcasse working, due to the grip and slip angle, and it was incredible. I liked the power the tires generated and how fast the T500Rs was with that car and ffb setup, and i also could feel something like a slip angle, but to be honest i really don't know if i should feel a slip angel as some explain it should be on the steering wheel, because i never experienced something like the old rf ffb style in a real car, i also forgot how slicks felt, since it is a long time ago that i droved some for test purposes. I think we should feel rubber tenseness and carcasse forces, but i'm not sure about the exact slip angle, that seems a bit unrealistic to me to feel the exact postion, in other words the offset from the longitidunal direction in degrees, in the steering wheel.

edit: ah and it is a long time ago as i've tested the skippy last, maybe i give it another go.
 
Last edited by a moderator:
Led566 said:
You are right.
Theoretically speaking steering torque generated by tyres is at minimum in correspondence of the slip angle where the tyre generates maximum lateral grip.
So, when the tyre exceed that slip angle and the car increase sliding, the steering torque increases.
In FFB terms, when entering a corner one should feel in sequence: torque decrease-max lateral grip-torque increase.
Click to expand...

TechAde said:
Here you go, updated plot attached.

Red trace is slip angle of the LF, orange trace is slip angle of the RF (in degrees).

Slip angle is calculated with the following function:

return (57.2957795 * atan2($LF_LateralGroundVel, -$LF_LongitudinalGroundVel))

Note the 57.2957795 converts the results from radians to degrees.
Click to expand...

Hooray!
rF2 works in accord to the general tyre theory!
(Not that I was doubting...;))
 
Joe said:
Oh, my God. I am out here!
Click to expand...

Look, this is the slip angle formula:
View attachment 14925
Now, if there is a "nearly" linear correlation between Slip angle (alpha) and Vy (as you have demonstrated) you CANNOT have a linear correlation between Slip angle and Vx, because Vx is at the denominator (let's forget the arctan and the modulus functions for a moment...;)) so the correlation coefficient you have calculated maybe right or not but for sure can't be similar to the correlation coefficient between slip angle and Vy.
Hope I've been clear...
 
Led566 said:
Look, this is the slip angle formula:
View attachment 14925
Now, if there is a "nearly" linear correlation between Slip angle (alpha) and Vy (as you have demonstrated) you CANNOT have a linear correlation between Slip angle and Vx, because Vx is at the denominator (let's forget the arctan and the modulus functions for a moment...;)) so the correlation coefficient you have calculated maybe right or not but for sure can't be similar to the correlation coefficient between slip angle and Vy.
Hope I've been clear...
Click to expand...

OK, this is good comment. Lets just do this once for all, plot FFB vs slip angle and get correlation between FFB and Slip angle.
Attached is the chart of FFB vs Slip angle, run about 2 laps time on Silverstone on GT-R.
View attachment 14926
The correlation coff is = 0.325 (32.5%). I am not going to argue if one can feel or not due to 32% correlation. Some do and some do not. I leave there....

@speed1, thanks for cheer;

@TechAda: I don't think your argument by using Torque vs Slip angle to derive your conclusion on FFB is correct. Please directly plot FFB vs Slip angle. I just did that. YOu need large sample data, not a few sec of data. I ran over 4 min over 27000 data points.View attachment 14926
 
Joe said:
YOu need large sample data, not a few sec of data. I ran over 4 min over 27000 data points.
Click to expand...

The length of the data sample is completely irrelevant.

My plots shows FFB v Slip Angle of both front wheels over a number of slip angle inducing manoeuvres, that is plenty of data to draw sensible conclusions from.

Sent from my SM-G900F using Tapatalk
 
TechAde said:
The length of the data sample is completely irrelevant.

My plots shows FFB v Slip Angle of both front wheels over a number of slip angle inducing manoeuvres, that is plenty of data to draw sensible conclusions from.

Sent from my SM-G900F using Tapatalk
Click to expand...

OK, I could not see your label clearly at first. Now I see. You have about 9 sec data (from 00:18 to 00:27). From the chart, I can speculate that very strong correlation, probably > 75%.
My data were including all in two laps: corners, straights, bumps, off side-way, etc, over 4 minutes. Can you give us the correlation results from your data?
 
You must log in or register to reply here.
Back
Top