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

Hey guys,

I intended to write this consolidated post for the optimal ffb (force feedback) settings for rfactor 2 users sometime ago but I had been reluctant to do so considering how much time it took writing the countless posts in different threads. However, it'll save me time and effort from repeating myself in the future (not that i really minded) if i do make this dedicated thread/post, so here goes.

The below may look daunting but it'll be worth it....i promise.


[HR][/HR]

Defining the problems

There are two problem areas that affect the ffb performance of your wheel in rfactor 2 (but this also applies to any and all racing sim titles). These two problems occur in the low end forces (known as an "initial ffb deadzone" issue which causes a complete looseness in the ffb initially which affects your performance mostly in the slow speed corners) and in the high end forces (known as an "ffb clipping" issue which causes the wheel to max out at 100% force too early which affects your performance mostly in the fast speed corners).

The reason we should address these two issues is that they are the key issues to unlocking our fullest driving potential, allowing us to drive more subconciously (a.k.a. being in the "zone"/"flow") which is where we operate with the least amount of conscious effort and yet perform at our best and more consistently so. In my own personal experience having found the perfect settings for my own wheel, i have seen a giant and easy 1-2 second lap time and consistency improvement in all of my personal bests (where my performance had plateaued for many months leading up to it). The improvement did not occur the moment i made these ffb adjustments but in my case it was less than a week later (so maybe I could argue that it was pretty instant after all).....hopefully that gives you enough reason to read on.



[HR][/HR]

Setting up our wheel profiler correctly before we start

Before we get started, we need to setup our wheel profilers correctly. Here i have screenshots of the Logitech g25/27 and Thrustmaster t500 profilers with the correct settings. You should have spring and damper and any spring-centering/auto-centering disabled. These effects are not realistic and should only come from the dynamic ffb calculated by rfactor 2's physics engine. Adding it like this from the profiler is not at all realistic as it simply adds a layer of static (i.e. fake) effect ontop of the dynamically calculated real-time ffb in rfactor 2 (which you don't want to be degrading in any way). CORRECTION: Thanks to Flaux for correcting me on this. Turns out enabling spring/damper in the profiler seems to then allow rf2 to apply it's own spring/damper affects onto the final ffb output. Whilst i still believe it is "artificial" in nature, it may very well be desirable for some specific cars since it can make the ffb feel more realistic due to some inherent problems with how conventional ffb hardware works. However, having said all this, i've not been able to test this for myself so you will need to test for yourself. Please read edit 11 at the bottom of the post for further details. Lastly, you can play with the "overall effects strength" a little but try to leave it as close to default as shown below (and if you want to increase them, do not go higher than 110-112% for the logitech g25/27 wheels and no higher than 70% for the thrustmaster t500...and i use 70% on my t500 which feels best imo, so don't worry with upping it a little on g25 wheels too. Going any higher will not increase the maximum force of your wheel but will instead start to cause the very issue we are trying to fix in "Part 1" below). This generally applies to all other wheel users as well.






[HR][/HR]

With that out of the way, let's begin with the high end force issue (a.k.a. ffb clipping):

Part 1 - the high end forces (a.k.a FFB Clipping - the most important issue): What is it?

To answer that, i must first explain how ffb works. The physics engine in rfactor 2 calculates the forces generated at the virtual steering wheel and this correlates to some amount of force generated at your ffb wheel. The relationship between how much force calculated ingame to how much force is generated by your ffb wheel is decided by rfactor 2. Please note however that when i say "the amount of force generated by your ffb wheel" i mean as a percentage of the maximum force producible by your specific ffb wheel....as the same 40% ffb signal sent to a leo bodnar wheel, t500 and a g27 will each generate a different amount of force felt by the user. Each wheel is producing a force that is 40% of the maximum force that is producible by the different ffb motor's in each wheel, just emphasizing the fact we're not dealing with absolute values here as rfactor 2 doesn't care what wheel you are using.

Fortunately, we are able to control the relationship between the forces calculated at the virtual steering wheel and our ffb steering wheel. This control is called the "car-specific FFB multiplier" ingame:



(side-note: Also set your ffb smoothing to 0, pretty much works on all wheels but some older wheels like g25/27 may experience rattling which is unfortunate and you can reduce/remove the rattling by adding smoothing but it is strongly advised you set it to 0 as this will give you the purest/raw ffb detail)

Generally speaking, increasing this value will increase the strength of all the forces on your wheel. Vice versa, decreasing this value will decrease the strength of all the forces on your wheel. At first this may seem like a great idea as you can just increase the strength of your g25 by increasing this value. But it comes at a detrimental (yet not immediately apparent) cost to your driving performance. At a certain point/value, you will hit what is know as the "ffb clipping" margin. What this means is that whilst on track, when you are driving in situations that produce the highest force at the virtual steering wheel (such as high speed corners), the game is asking for more than 100% of your wheels maximum force. Of course this is not possible so instead your ffb wheel remains at a constant 100% force, meaning you miss out on that all important detailed information regarding the cars balance and traction when you drive in that ffb clipped region. In fact it's worse that just "missing out" on the information, what's actually happening is that your getting mis-information from your wheel. Think about it for a second....your ffb wheel is telling you the force is constant yet the car could be increasing in forces at the virtual steering wheel which would otherwise be telling you that the car could be about to oversteer (or worse yet snap oversteer) or understeer. Instead your mind is being told that the car is in a stable 4 wheel slide when on the screen you are being told a different story. At this point, the ffb is completely useless to you and you have to rely on guess work from visual cues on screen alone whilst also manage the confusing sensory mis-information you are receiving subconsciously from your ffb wheel.

To better illustrate this, here is a pictorial representation with a simple sine wave that you can imagine as being the change in ffb force strength and direction as it is sent to your ffb wheel as you navigate a high speed left-right s-curve corner:

(colour coded words represent areas on the diagrams)



Now, if your like me and want to have as strong an ffb wheel as possible without running into the ffb clipping issue then you need to know what is the right "car specific ffb multiplier" value (which as explained earlier is not dependent on what wheel you use....unless you have a bodnar wheel where perhaps you don't want to use the full force of your wheel because it is simply too strong). By default, the "car specific ffb multiplier" value is set to "1.0". For all the cars i have tested (maybe only 70% of all the cars available for rfactor 2 at present except the newly released kart i tested where "1.0" is fine) this is simply too high and will give you the ffb clipping issue. (A summarised theory as to why ISI and other sims set the default value so high which unintentional causes an ffb clipping issue is that most users are still using older wheels such the g25/27 which are pretty weak in ffb. So in order to boost their general ffb weakness, especially noticeable in the low end forces and not helped by the huge initial ffb deadzones on these older wheels, they must set this value higher by default in order to make the wheel feel like it has any strength at all....fortunately "Part 2" below explains the new tool we have at our disposal to remedy this problem). The most ideal "car-specific ffb mult" value is (almost) always lower than "1.0" but you don't want to go too low and end up not using (i.e. wasting) the full range of ffb force output at your wheel's disposal. (As of recently, i've find some cars requiring a little more than 1.0...you really need to test with the tools explained a little later. It's very easy). Here are the 3 possible outcomes:




So, how do you find the ideal/optimal value you ask?

The quickest, easiest and most simplest way (short of asking others who have found them already) is to use TechAde's excellent "rF2 Pedal & FFB Overlay Plugin" for rfactor 2 (Download Link). Simply download the latest version, drop the "rF2PedalOverlayPlugin.dll" file into your rfactor 2/plugins folder. That's it. Then when you go into the game and on track, wait about 10 seconds and in the top left hand corner you will see this (only smaller):



The blue bar represents your clutch position, red:brake, green:accelerator and lastly the key point of interest, the yellow bar represents the ffb force sent to your wheel. When this yellow bar reaches 100% (a full bar) it turns red to show you that it is maxed out at 100%. You do NOT want this bar to ever turn red when you drive around a track unless you are driving at the limit of traction around a corner, over kerbs, gravel or crash/bump the car into a wall or another car. If you see this bar turn red in a normal race situation around a track (also not including spinning as this is not "normal" and falls outside the normal driving envelope)...you are getting ffb clipping which as explained is a big no-no. To find the optimal "car specific ffb multiplier" value simply involves you testing different values around a track. Doesn't take a lot of time and once you find the right value for a car, it's usually the same for all tracks. If other tracks do clip however, just set it a tiny bit lower and leave it as that for all tracks.

Additional explanation by TechAde:

For the few cars i routinely drive, i have found for example that the Renault Megane value is around "0.75" and for the Panoz it is around "0.85". You will have to test this for yourself and to find values for other cars. It really doesn't take a lot of time and it's completely worth it. It can get a little tricky trying to perfect this value to make sure the ffb is not clipping when your going around high speed corners and can't focus on the bar graph in the corner and i sometimes end up deciding to notch it down/up 1% every so often to check if it's still clipping a little bit or not.

When i first tried this out i wrote a piece here describing my experience and emotions of my discovery (though i cannot and do not claim to have discovered this as it is nothing new that others have not already discovered and shared elsewhere on the internet). You can skip it and read on if you want of course, i just go into more detail how my performance improved exponentially overnight and why.

http://isiforums.net/f/showthread.php/17796-Take-a-Bow-ISI?p=247786&viewfull=1#post247786


[HR][/HR]

(side-note: please complete part 1 before working on part 2 as they are somewhat interrelated. Working on part 2 first and then working on part 1 after will more than likely result in you needing to repeat part 2 again.)

Part 2 - the low end forces (a.k.a initial ffb deadzone issue): What is it?

The majority of ffb wheels suffer from an initial ffb deadzone issue simply because the initial low end forces are not able to overcome the internal resistance of the wheel due to friction (with the exception of the best wheels such as the leo bodnar wheel which are direct drive and what little internal resistance exists is miniscule compared to the full range of forces it is capable of....some 200-300x stronger than a g25/27).

Here is a graph of the ffb response curve of my T500 wheel in the green line (to demonstrate this):



Notice how the first 8% of forces sent to the wheel produces 0% force output at the wheel. This demonstrates the static friction of the wheel. As a result, we also get a delay in any ffb force output at the wheel. Take for example, a 20% force sent to the wheel actually results in a 10% force output at the wheel. As well as the initial deadzone region, this is not ideal and the result is a loose feeling of the wheel around the center when driving in a straight line, initial turn ins and exit of corners and much lower ffb forces than should be in the slow speed corners where the forces are generally no greater than 10-30%. Much like the ffb clipping issue, the wheel is giving you mis-information and negatively affects your performance for the very same reason. The wheel is telling you the car is in a stable 4 wheel slide when the car balance could be changing and in fact oversteering or understeering. (I must mention however that although this graph shows an 8% deadzone, the reality is that this is for static friction and not rolling friction which in my experience is around 3-4% for the t500. Even though it sounds like a tiny amount, in practice it still makes a significant difference when you feel any and all previous deadzone removed).

The blue line represents a modification to the response curve that cuts away that unwanted deadzone. rFactor 2 developers have recently added a new line to the controller.ini file (but as of the recent updates to rf2 i believe it's now the controller.json file) in the ":\rFactor2\UserData\player" folder. The line is called "Steering Torque Minimum". To find the ideal value, you need to test for yourself and this doesn't take long either. My advice would be to start with a value of 4% (which equals a value of "0.04000") and then jump into the game with any car (make sure you have fine tuned your ffb multiplier first however as this can affect the correct deadzone value) and see if your wheel is oscillating when stationary (Thanks to Comante for correcting me in saying that you should do this with the car at very slow speed for the best/most-accurate results....explanation below in "edit 3"). If it is, go back and lower this value but if it is not oscillating then increase it until you find the wheel starts to oscillate or vibrate. Rinse and repeat till you find a value that only just starts to oscillate/vibrate the wheel a tiny bit and the value is accurate to within 0.1% (e.g. 4.1%, no need for 4.1275% accuracy lol). Then back it down a tiny bit, maybe by about 0.1% (so imagine you find it just starts to oscillate/vibrate at 5.4% then take it down to 5.3% or 5.2% instead) to avoid the small oscillation/vibration all together. (Thanks to Luke Dykstra for noting that you must have the "steering torque minimum" value represented to 5 decimal places for it to work. So that's "0.03000" for 3% deadzone and NOT "0.03" or anything else. - UPDATE: no longer needs to be 5 decimal places with the new json file format for the controller.json)

This is what happens when you set the "steering torque minimum" too high causing that unwanted oscillation:

You will now have ironed out any and all initial ffb deadzone on your wheel in rfactor 2. Enjoy!

Update to part 2: You no longer need to mess around in the json file to change the steering torque minimum value . The function is now accessible ingame and the UI setting is called "FFB minumum torque". Here's is pic of it by Justy:




[HR][/HR]

Finished

Ok, that's all. Boy that took a while to write up.

I will be updating/correcting as i go along re-reading it and i will post below with any edits or new editions I make.

Hope you find it useful in unlocking your own full potential too.


[HR][/HR]

Updates/Edits

edit 1: Thanks to Luke Dykstra for noting that you must have the "steering torque minimum" value represented to 5 decimal places for it to work. So that's "0.03000" for 3% deadzone.

- edit 1 update: no longer needs to be 5 decimal places - see edit 8 for details.

edit 2: Those who read my posts in the past and who also followed my settings of 100% overall strength in the t500 control panel to reduce the initial ffb deadzone, I now use 60 or 70% overall (60% being the default value) as it seems to feel better and with the new steering torque minimum (of 3% with 0.75 ffb multi for the megane) the ffb feels sublime!

edit 3: Thanks to Comante for pointing out that it is much better/accurate to test the oscillation/vibration when trying to find the perfect "steering torque minimum" value when the car is moving even if only 1km/h instead of being stationary when you test this. The reason being that with the car moving very slowly, it is when the wheels are actually rolling that there is the least amount of resistance on the movement of the steering wheel (imagine trying to turn the steering wheel to park when your stationary vs with the car moving ever so slightly which requires less steering effort by the driver). You want to remove the oscillation/vibration in all driving situations and if any oscillation/vibration were still to remain it would show itself in the car moving slowly situation.

edit 4: If you do share settings for the same wheel, note that any difference in your Logitech (or whatever wheel you use) profiler settings will affect what the "steering torque minimum" should be for each user. For example, increasing the "overall effects strength" in the Logitech/Thrustmaster profiler will reduce the deadzone a bit which means the amount of deadzone that remains to be removed with the "steering torque minimum" amount could be different for different users. This is why i urge you to not only ask others for what values they are using but to couple it with your own testings to confirm they are right for you too. Perhaps share what your profiler settings with the specific rfactor 2 values you use on your specific wheel please.

edit 5: Wajdi raised a good point regarding what "ffb smoothing" value ingame to use. To help others find the right value for themselves, here is an explanation for what ffb smoothing is and why you may or may not wish to use some amount of it:

For those who don't know what "ffb smoothing" is ingame, it is a way to smooth out the high rates of change in the ffb signal. Why would you want to do this you might ask? If your wheel suffers from annoying rattling in rfactor 2 you can increase the "ffb smoothing" value ingame to help alleviate the problem. But if you plan on using anything higher than the default value given to you by rfactor 2 (after you hit the "detect" button in the controller settings ingame which will configure a recommended "ffb smoothing" value for your particular wheel) then only raise it as much as need order to reduce the rattling issue of your wheel to acceptable levels. If the value is set too high however, you are at risk of losing crucial ffb data.

Here is a simply demonstration of how smoothing affects the ffb signal (which i should add in this example could be considered with excessive smoothing if it were an ffb signal):



If the rattling (common for g25/27 users) is not an issue or you have non, feel free to set "ffb smoothing" to "0" for the purest ffb straight from rfactor 2's physics engine. Setting "ffb smoothing" to "0" gives you the best possible ffb feel and detail. rf2 defaults this to "4" which i would always advise setting it to "0", especially for CSW/T500 users, there is absolutely no reason to use anything but "0".

Here's another graph showing the possible outcomes to different smoothing values:



edit 6: As a case and point that you need to do your own testing of the car specific ffb multiplier, the newly released Karts are perfect with the default "1.0".

edit 7: If you want to generate your own graphs for the response curve of your own wheel, here are the instructions on how to do so...

edit 8: Thanks to Mee for reminding me with the new json file format for the controller.ini file, now called controller.json, you do not need to input values to 5 decimal places for the "steering torque mininum" value to work. Just put in a value to as many decimal places as you want/need.

edit 9: In case you have trouble downloading the WheelCheck.exe file from the host site...

edit 10: Steering torque minimum now accessible ingame (i.e. on the fly) via the UI setting in the controller settings page/menu called "FFB minimum torque".

edit 11: Flaux made some pretty interesting observations/discovery regarding spring and damper settings in the profiler and how they relate to rf2. Seems i was wrong and that enabling spring/damper effects in the profiler does not produce a layer of artificial spring/damper effects onto the ffb wheel. Instead enabling them in the profiler will then allows the game to apply any spring/damper affects it may have coded for the ffb. I still believe these affects are artificial because if they were a natural product of the dynamic ffb simulation you would/should not be able to disable them

Having said that, since writing this guide i've come to understand the possible value of these settings to the ffb realism since they can/possibly help remedy some issues with consistent steering system inertia and damping simulation that is inherently impossible to achieve accurately with the way conventional ffb hardware+software works. I would advise that you give it and try and see for yourself. It could well be worth it.

Here is Flaux's post that covers the details...

and a new update from Flaux...

So it sounds like you may need to test and try different combinations to see what works best for you.

edit 12: Thanks to Hazi for suggesting i include a very concise explanation of the rF2 Pedal Overlay Plugin by TechAde:

 

Thanks for taking the time to share this with all of us--and excellent advice.

 

Great write up my friend thanks for taking the time!

Edit - I should add what I learned about the minimum torque setting today. It seems you have to make sure there are always 5 decimal points in order for it to be effective. For example 10% requires 0.10000, if you have 0.1000 it will not recognize that.

Also - for the record I have had my new G27 up to 0.21500 (21.5%) to be on the edge of oscillation.

 

Thanks for mentioning this. I'll add it to the op.

I'm surprised you are using such a massive deadzone value of 21.5%. I knew the deadzone was big on the g25/27 but not that big. Out of interest, what ingame ffb multiplier are you using with that?

 

Actually just been playing with it right now. I have found Malaysia GP to be a good testing sight with all the corner variations. I have now added the plugin you mentioned which has been a big help. So far the Megane I have found ideal setting at 88% and the Formula Renault is down at 79%. What a difference in ffb definition in the turns. Overall strength really hasen't changed for the most part, as higher settings are over the g27's capabilities anyway. Once again thanks for the great write up I'm sure this will improve many peoples rF2 experience.

 

Music to my ears of your experience with it and very good idea using Malaysia GP to test on as you rightly said it covers pretty much every type of corner in one track!

I'm still shocked by the 21.5% value you need to remove the deadzone on your g27. Hot damn! Oh and are you now able to use the default 100% overall effects strength in your logitech profiler? Increasing this value just adds non-linearity which had the nice effect of reducing the initial deadzone amount somewhat but at the cost of inducing something similar to ffb clipping at the profiler level (no longer needed of course because we now have "steering torque minimum" to offset the deadzone completely).

Here's an arbitrary example graph i made some time ago (which was made to demonstrate what the "steering torque sensitivity" value does) and in this case higher sensitivity value represents a similar non-linearity effect that a higher overall effects strength in the Logitech profiler produces:

 

Yes the high speed corners, so far in both cars I have tested are what seem to really push the FFB to its capable limits, besides curbs, grass, and oversteer conditions..

 

Thanks

 

I'm surprised with the high ideal ffb multi value you found for Panoz. I'm using 0.55 as I told before and it feels really good with 2-2,5% of minimum torque. I would imagine raising the multi by 0.3 would make the steering feel really heavy :s (which would effect going to slow corners negatively because in those situations you need to turn the wheel fast) Maybe I just don't have muscle for the job

But you have helped me understand the importance of ffb setting even better so I thank you very much for that

 

This is an insanely interesting thread, and I think I'm finally ready to get rid of the ffb clipping on my G27 wheel.

I found some "nice" settings last summer, which I have refined over the last year. However, these settings just gave me a really heavy wheel (which I really like) but also lots of clipping of course.

But my question is, can I still have a nice heavy wheel, but avoid clipping? Where would I start on a G27, with both the Logitech Profiler, ingame settings and .ini.

My fear is that I'll just get a limp wheel that I cannot get used to, after driving for a year with a heavy but clipping wheel.

 

Thank you for this - very easy to understand

 

Excellent post DrR1pper.

 

Yesterday, I tried your setting with overall strength at 100%. my driving feeling has improved especially at low speeds. This evening I will try to change as you wrote in the first post of this thread.
thank you for sharing with us your experiences. thanks

 

Thanks for this post! Finally i understand the FFB in my wheel in relation to racegames.

 

Should be sticky, I think you should expand it to show all the other control settings not just ffb. Exaggerated yaw etc. (If you had time!) I think this post here is what makes the difference between rf2 feeling good, or amazing.

Well done sir.

 

Yeah it does make the wheel very heavy in the panoz in slow corners! I've been driving it at old spa online and the last (hairpin) corner makes the wheel crazy heavy. But to my memory, I don't recall it clipping at all so it was the ideal value for the panoz. I will keep my eye out on the ffb bar though to double check the next time I'm in it but I think it isn't clipping.

edit: Don't forget that i now use 70% overall strength in the t500 control panel now and not 100% which requires much lower ingame ffb multiplier where i was using 0.46 with the Megane before (because of the non-linearity and ffb clipping from the wheel's profiler side of things that 100% induced). But i don't want to confuse new comers to the thread by the difference in the profiler settings and why i was using 100% instead of 70% (because before steering torque minimum was introduced, using 100% was the only way to try and reduce the initial deadzone amount).

 

Firstly, thanks for the nice comments guys.

Yeah I think you can still live with using a compromise between the old and new settings to maintain a strong/stiff presence in your wheel. But if you haven't yet already tried using the new steering torque minimum with the default Logitech profiler wheel strength and non clipped ffb value, I would urge you to do this first because it might surprise you how much stronger your wheel will generally become after removing the huge deadzone on the g27. If its still not strong enough though then go ahead and increase the Logitech profiler overall effects strength a little (but not more than 110-115% to avoid going into severe non linearity).

However if this doesn't feel acceptably as strong as your old settings to you (due to perhaps excessive ffb clipping with your old settings which made it feel very stiff/strong) then the only other option is to get a new wheel. In my own experience a t500 was a great improvement over a csr-elite which was a good improvement over a g25/27. I really enjoy my t500.

 

Thanks I was not sure about the relation between the new minimum force value and the oscillation.
Now it is perfectly clear

 

Hehe . I wasn't either for a time after it was introduced into the recent builds but I finally realised it was overly aggressive initial values that I tried (thinking the 8% from my graph) and thinking a small 3% would not change anything meaningful to the ffb experience. How wrong I was. It is incredibly sensitive to this value.

 

Funny you should mention this because maybe 6-12 months ago i turned down all the virtual head movement strengths by perhaps 50-75%. Simple reason being that although it is a very nice effect to simulate head movements in a virtual car, if it's too aggressive (not that i'm saying it is or it isn't in rf2 as i really don't have a well defined opinion on it atm) you can end up reducing the effectiveness of your visual reference/cues of the car's pitch, roll and yaw rates. However i think some time ago it was reset and i forgot about it and haven't gone back to check if it's reverted back and/or needs lowering again.

I will check it out again in the near future.