Guide: Optimal FFB settings for rFactor 2 - The key to being in the "Zone" :D

I have a logitech DFP and my curve is the same too.

 

Thanks for pointing out WheelCheck; that's the only thing iRacing has contributed to the community that I've found useful so far.

My G25 seems to be most linear at about 96% strength or so; but that gives me a 17% to 18% FFB deadband. No big deal, as entering STM appropriately and tweaking for best response works a treat. Hats off for giving us a no-guesswork way to optimize the rF2 FFB values!

 

It's due to my Y scale on the graph, when I remove the curves done with 320°, the 900° curves have the same profile.
I will test 98%, when I test under the the game the wheel with the Kart, I prefer 99% instead of 90%, 91%.

 

So Ive been thinking about the STM function in rf2 and i come to the conclusion that for user who have big deadzone in theirs wheels there is no solution to have the most real response.
How is that? Its because STM set to 0,15000 for example gives us all the time 15% of ffb feeling. So when we block brakes, lose traction etc we will always feel like there is still grip since we cant feel less than 15% of ffb output.

I dont know if it can be done but a function which will scale the ffb output would be a solution. What i mean is when game/software gives u 1% of input it should be 15% of output. Than 2% input should be around 15,xx% output to finally reach 100% input which will match exactly 100% output.

Do u understand what i mean? To be honest i dont know if its a good idea. Just a thought

 

There is a string in the controller.ini file that now I don't remember how it's named but the description say : keep 1 to have linear , 2 high, 0 low , maybe this could help us with wheels that appear to not be linear in response when overall forces are over 50%. Sorry to not be able to test this beforehands, maybe this evening.

 

I am pretty sure it is set to 1 by default

 

Mad_King I don't have a program on my gaming PC to post a picture but my curves look like yours. Only difference is at the top end my curves "max out" the last ~5% if I use anything over 91%, a bit like your 99% curve but completely flat at the top. With 91% mine looks about the same as your 95%, and my 90% looks like your 91%. I guess every wheel is different.

 

Yes Lukas, default is 1, but if our wheels have a curve that is "above" linear, then if I understand what this setting do, the "low" profile in the output curve could bring the overall effect near linearity. On the other hand, if some wheel would perform "under" the ideal line of linearity, then the "high" setting could again make things more even.
Hope it's clear enough.

 

Now I understand what you mean. Any idea how the check it?

 

Just wanna jump in here for a sec. I've realized that I have lost the forest from looking at all the trees.
It was mentioned about running Minimum Force test, (vaguely mentioned). My out put from that test was the suggestion to use 0 Steering torque minimum.
So I did. Now I can once again feel the cars tendency to roll to a greater degree.
I was all excited to try the steering torque minimum thing (running 0.15000) and wound up hiding some essential ffb feeling. (and thought I was in drivers heaven)

I'm probably wrong here, but seems that we need a list of ffb tweaks that give an outlined attack in order to achieve the objective. So that a user doesn't get lost obsessing on only one aspect of ffb and ends up thinking "this doesn't work!"

Another thing I noticed in playing with wheelcheck program. It actually makes a difference leaving the logitech profiler running in the background. (in my opinion)

 

anyone know of a decent free CSV/Graph viewer? (dont have Excel) thanks

 

Leaving the Thrustmaster profiler in the background stops the wheel check.exe program from working or working properly in my experience.

I don't touch other ffb settings in ini files so I can't help you with that.

 

I think you can use Excel Online for free.
And you have other free alternatives like Google Drive (online) and LibreOffice (offline).

 

Hi there

excellent write up, thank you.

I have a question about something else, though. Hope you don't mind as this seems to be the most comprehensive thread on FFB overall.
I don't like all those fake bump and rumbling effects, also that the engine turns off the FFB when off-road resulting in behavior like driving on ice.
I tried to adjust it in the controller.ini but it seems the game always uses it's default settings for those commands - that is:
Off-road multiplier
Rumble strip magnitude
Rumble strip pull factor

Anyone knows something about this?

(Other commands take effect like Steering effects strength,Spring etc.)

 

Ok, yesterday made some tests, and what I found is:
For logitech DFP, and from the graph posted it should apply to other logitech products, the FFB "STRENGTH" in the profiler, does not actually manage the actual maximum force that the wheel can deliver, but it just control it's "erogation curve". For this reason it is quite clear why around 50% strength is the most linear curve in the graphs. The second confirm that I want to give, is that the controller.ini string that dictate if the ffb output will be linear, low or high actually do the same thing. It's much easier to keep the wheel at 50% and output linear than to use other value and a lot of guesswork.
I've tested even another line nearby that roughly say"blend values to allow to feel torque changes even when wheel is at 100% output" , well it works, and make the wheel a bit more "lively" on fast turns, you can test it with different values and see if it improve things.
One last thing: while driving with FFB smoothing = 0 make things more entertaining, I can't be as fast as I can when keeping this value between 5 and 10 , so, before impose to yourself that this value must be set to zero absolutely, do your homeworks.
In short, with my old and weak logitech DFP, what send me in "the zone" (and this feeling is so strong that you immediately recognize it) are:
logitech profiler: strenght : 50, spring 0, damper 0, degrees 900
Controller ini:
Steering torque minimum = 0.02500
FFB set to (default) damper instead of friction
FFB multyplier between 0.7 and 0.8 for the car I've driven yesterday (FR3.5 and howstons 1968)
FFB smoothing = 9

With these settings I'm able to "dance the car around corners".

 

Can u share your graphs? I wonder if setting ffb strength to 50% for DFP reduces maximum output which can be deliver by a wheel.
Since for g27, as u can see from the graphs posted, setting ffb stranght below 96% reduces maximum output

 

Brilliant cheers mate, just gotta figure out how to work it now!!

 

Lukas, I've not made a lot of graphs, just one and it's just like one of those posted with curves from 50/60/70/80/90/100 values, so I thought it was not useful to repeat the same experiment, but I've made hundred laps on the same combination track/car, and the wheel feel just as strong, if it's not,it's just because you before was feeling more output than intended. but more importantly, I can lap faster and safer. One thing you have to consider, is that on the limit of output (being it lower limit or upper limit) most equipments are less precise/reliable/repeatable and so on. I think we are all a bit too concerned with "maximum force" from the wheel, and a bit too less concerned with "how much this force from the wheel match the force that the simulator wish me to feel?", it's for this reason that I see people suggesting to use high strenght values and low smoothing, but both these two factor can induce random differences between input and output. Each wheel should have an optimal range, better the wheel, larger the range should be, the lower value should match with Steering torque minum , while the upper value should be controller with FFB multiplier.
For example, if my wheel is not reliable and output erratic forces over 90% of it's maximum output, I'm sorry, but I should make sure that in normal conditions I never reach that point, the same apply for the lower value.
I hope this helps, I fear we all invested too much effort in this "calibration" and lost a bit sight of our goal.


Edit : I've checked the graphs again Lukas, I see the difference in max output you say, but you see how erratic the lines turn in the upper side of the curves with more strength, you actually can't use them as they are.. if you make the same graph again and again , you will see that the upper portion of the curve will always be different, thus the output will not be predictable. Can you use your wheel al 100% or 150% force ? yes for sure, but you need to tweak the output curve from the simulator, and we have not the means to check if you can flatten the curve again without a tool that could sit between the simulator and the wheel's driver which pass all data through a calibration curve. The only thing you can do to "flatten stuff" maybe is to increase FFB smoothing.

 

http://tof.canardpc.com/preview2/8e853362-8822-42ea-80ce-3a6aa00470fd.jpg

This is the graph about G25? Well, I studied chemistry in high school, and those curve are familiar to me. If this was not a game controller but an expensive lab equipment, beforse using it, I would need to make a "calibration curve", I would need to test my equipment with known data, and obtain these curve. What I would do after that, would be to use only a portion of one of those curves, one that is the most linear. For example, if I were to use the 99% strenght curve, which part I would use? clearly the part from 3500 to 8800. Could I use the part from 0 to 3500 and from 8800 to 10000 ? Yes, but I would need a tweak, because those lines have different inclination from the main one, and being on the extreme of the instrument sensitivity, I must take into account for greater measuring errors.
So, what is the curve I need to choose? This is quite simple: the one that grant me the greatest difference from Xmin/Xmax and Ymin/Ymax and at the same time is the most linear. For example the line of 150% stregnth is flat and long, but the "delta Y" is so little, this will induce the greatest measuring errors.
I hope this explanation make sense to you.

 

So i am now confused. What is our goal is to match input with output so when the software/game gives us 60% of force the wheel gives us 60%. Am i right? I wasnt testing such a low values but according to graphs posted for g27: FFB Strenght set to 80% is giving us about 60% when the input is 100%. Not linear at all.

But for sure will check your foundings and post my feelings.

I may be wrong but from yours explenations u are looking for the specific line linearity, not for the input=output linearity, and this is not the right aproach.