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

lol.

 

So when the software/game tells my wheel to output 50% power ("Force"), my CSW V2 actually outputs around 77% or 78% ("DeltaX"). That's terrible.

 

The wheelcheck step2test measures how much your ffb steering wheel has rotated from a stationary starting position after 300ms of some percentage of maximum torque output is demanded from your ffb wheel. I believe it's performed in increments of 2%. So if on the last increment test (i.e. 100% torque output) your ffb wheel only rotates 235 degrees, then that's all 100% torque output from your ffb wheel motor over 300ms can achieve. But the point of the test is not to see how far it will rotate, as insightful as it may or may not be when compared with other motors. The purpose of the step2test is to check linearity and from your graph, looks strange to see it accelerate up like that towards the end. What wheel do you have?

"it repeats that like 5 or 10 times and then finishes"[/I]. I presume there must be a repeat setting in the wheelcheck.exe app with 5-10 selected. Change it to 1.

 

The repeating part is probably down to what you said about the torque over 300 ms only being able to drive my wheel so far. I believe the step 2 linearty test does 52 steps/outputs/tests (or whatever they're called, lol) for everyone, so it's not repeating from that respective, it's just that the final 5 or 10 tests all rotate to about the same point (around 235 degrees) whereas each output before that gradually made the wheel rotate slightly more. I guess it's like you're saying, by the last 5 or 10 steps, it's not capable of going further.

The wheel is a Fanatec CSW V2 w/ Porsche 918 RSR rim. My linearty looks so wrong according to WheelCheck. I mean it practically looks broken. When WheelCheck sends a 50% output signal, my CSW V2 is outputting 80%. Therefore the 2nd half of all software/game forces are compressed into the final upper 20% of my wheel's output force.

Should I be glad that the lower 50% of forces seem like they're being boosted to 80% of my wheel's physcall output? Or do you think something may be wrong?

 

You can clearly see your wheel beginning to level-out/plateau when going above 50% of max torque output for 300ms. But before i say more, can you repeat the exact same test but with 50ms (if possible), 100ms and 600ms and make the same graph please.

Maybe, maybe not....it's not necessarily that simple. For example, this could be simply due to the CSWv2 having that much less default friction/resistance in the steering mechanism that it's causing the torque output to reach maximum speed before 300ms whilst for wheels with more resistance, even at 100% torque max speed is not reached and so you don't see the non-linearity unless you increase the time duration above 300ms. Just a theory but if you can repeat your test as mentioned above it may confirm it.

There is also the possibility of CSWv2 specific settings to consider....drift mode coming to mind. But let's see the data for the 50ms, 100ms and 600ms if you can please.

Bed for now.

 

But I thought a "perfect" wheel should be perfectly linear? Are you saying that the lower internal resistance a wheel has, the worse lineary it'll have?

Ok, I'll do those other tests and report in the next hour or two.

[HR][/HR]
On a separate note from my personal wheel issues, I thought you'd be really interested in this other cool little wheel linearty test program. It also calculates a chart for you with some other options for editing the chart in accordance to in-game FFB settings (PCars settings). Check this cool little program out https://www.dropbox.com/s/1k0norwiv9ki6ml/FCM.rar?dl=0

Then there's this sweet little page where you can do all sorts of stuff to optimize your FFB based on your wheel's test results. Although it's made for PCars, there may be of benefit for all wheel users as-well. I think a FFB guru like you and many others around here will find the following very interesting https://docs.google.com/spreadsheets/d/12NlbeTytWJQYcXwLLV1g6wIXSUo02cB-y2Zf8Gy-SgM/edit?pli=1#gid=1651868755




[HR][/HR]
Post #956 updated with a better graph.

 

I think something else is causing your non-linear results. I found this CSW v2 graph and apart from a little initial deadzone removal, it's pretty much linear. They haven't needed to use any scoop knee or scoop reduction to get it linear. I also think this was performed with the default 300ms in wheelcheck.exe.



Have you tested with different drift modes (DRI), etc, on your csw v2? Perhaps on of them is the cause of your non-linearity.

 

Is there a way to test "what does my wheel output when the softwarw tells my wheel to output x amount of force"? I thought that's what the linearity step 2 test essentially was, but now I'm unsure because 50ms, 100ms, 300ms, and 600ms all give different results which makes no sense to me ATM. Is there a way to simply get a FFB software input VS physical wheel output graph without 50, 100, 300, etc. ms giving wildly varying results? I just want to know what FFB % my actual wheel outputs VS what it's told to output by the software, regardless of those 50, 100, 300, etc. ms settings.....

I didn't test different drift modes but Jack Spade (the well-known PCars FFB guru) says he has very similar results as me with his CSW V2. He also said, according to his tests, the lower from 3 the drift mode is set to, the more linear the FFB becomes. He plays with drift mode @ 0/off and that may be why.

Drift mode @ 3 is the cleanest, purest FFB setting. The lower you go under 3, the more the wheel slows down, becomes lazy, numb, gets "friction" added to it, and feels like you're driving through thick mud. It totally looses fidelity. Drift mode @ 3 is the only "clean", pure form of FFB without the driver crap interfering and screwing things up. Playing at anything other than Drift mode 3 feels terrible to me. The FFB is literally reducing the performance and FFB quality/fidelity of the wheel with drift mode @ less than 3.

 

Did you mean "Is there a way to test what does my wheel outputs in torque when the software tells my wheel to output x amount of torque?" If so, there isn't any software that can do that without also having some form of torque meter device to measure torque output from the ffb motor. The next best thing we have at our disposal is the measure of resultant changes in steering angle and from that we can infer the amount of torque output from the motor. In the same amount of time, if you double the torque you should get twice the rotation travel. But inferring torque output this was requires making some assumptions. One example is the assumption that ffb motor does not reach it's speed limit under the time duration (e.g. 300ms) for all torque output amounts being tested. Another example is the assumption that the force signal sent to the ffb motor is unfiltered but if you have a DRI setting on your CSWv2 that is influencing the final ffb signal sent to the ffb motor then ofc that could explain your results.

Ok, i thought they might. Can you graph them and post them here please.

If you post the graphs, i might be able to help you figure out what is causing the varying results.

This makes sense it might be drift mode however not in a way that i thought it would based on your results.

If that feels better to you to use drift 3 in game and you drive better and the ffb feels better, then use that. But for testing the linearity of the motor, these effects will affect the results.

 

I think this graph looks very similar to what a linear torque curve would look like with a Steering Torque Saturation(from controller.json) value of 2 applied to it. Maybe if you try a STS of 0.5 and experiment within a range of 0.4 to 0.6 you might find a sweetspot to bring it close to linear. That is, if this graph is actually representative of what your wheel will output in sim.

 

I am sorry for being sooo flippant and conceited ............but I feel for you guys I really do.

If I had to worry about stuff to get a semblance of what I thought felt right I would have given up years ago.

Hard to find enough people can keep up in a F3 and it is the most sedate FFB car in the entire sim.

I reckon I could kick 3/4 of your arses with a default black Momo.

Oh ........I used to ! lol

 

DrR1, drift mode @ 3 apparently uses the least amount of effects. Off, 1, and 2 are artificially slowed down a fidelity-killing while 4 and 5 are the opposite where the wheel tries to react to your inputs by speeding-up your inputs by momentarily accelerating the wheel in the direction you are turning it rather than resisting and wanting to centre (think of what happens when you have your FFB reversed, but only for very short moments/bursts whenever you make a movement of the wheel rather than constantly).

Drift Mode @ 3 should be the purest, cleanest, most unobstructed form of FFB possible with the CSW V2. What's more is that you can easily feel this. You can feel the "fake" FFB killing effects of drift modes off, 1, and 2, and you can feel the odd reversed FFB behaviour of drift modes 4 and 5. But when you try drift mode 3, you don't feel either effects, just simply a pure, clean wheel. Therefore I think it's more than just a "feeling-preference" thing, but a technical thing as well where drift mode @ 3 is technically the purest, least-effects-obstructed, form of FFB with the CSW V2.

Yes, I will post the charts and I have some intetesting results when I experimented with different FFB combinations regarding the "FF" and "FOR" settings. They'll be up in an hour or so.

Good idea. I wonder if I can use those Project Cars tools to apply different STS-like values and then convert that number into STS?

[HR][/HR]
Here you go If using Mozilla FireFox, right-click on any image then click "View Image" for a slightly larger, better quality version.

 

strange. Which drift mode (if any) gives you a linear result with 300ms?

In fact could you repeat the 300ms test with each drift mode and group them into a graph please.

 

Drift mode at any other setting should be irrelevant though as anything under 3 destroys the fidelity/detail, as-well as what seems like reactions and speed, of the FFB due to interference/obstruction from "fake" FFB effects applied by the drift mode. Any drift mode other than 3 butchers the quality/performance of the product's FFB. Having said that, would you still like to see results for all drift modes?

 

yes please.

 

Oh i understand that feelng mate. I have guys spending a fortune trying to get an osum feel of correct physics................with iRacing xD

 

Zoomed-in charts are just a zoomed-in (10% Force by about 10% DeltaX) version of the full-scale charts in order to be able to take a closer look at the minimum force stuff.

The following two charts show a comparison between different Drift Mode settings while force feedback strength is set to the default of FF:100 & FOR: 100






Conclusion

- Linearity suffered the further from "off" Drift Mode was set to
- Damper and Spring @ 100 and 100 had basically no effect on the full-scale chart when I tested it on Drift Mode 3. The results were almost identical hence why I didn't bother charting it's results
- Damper and Spring @ 100 and 100 did make a difference in the minimum force stuff though. Check-out the difference between it (black) and the "standard" Drift Mode 3 setting (green)

[HR][/HR]
The Following two charts show a comparison between different force feedback strength combination settings (FF & FOR) while using Drift Mode 3 (since I'm currently trying to stick with Dri Mode 3 as it seems like the only drift mode which allows the wheel to deliver clean, "pure", un-obtrusive FFB).







Conclusion

In my opinion, I think the following 4 settings would be the most logical to use:

- FF: 50, FOR:150 (rose) - seems almost perfectly linear. It does suffer a bit of max power but it still reaches in the low 90s % of max force. It also suffers in minimum force though
- FF:60, FOR: 150 (magenta) - regains ground in-terms of max force but suffers quite a bit in-terms of linearity and the minimum force is almost identical. It is the most linear of all the combos that reach about max force though.
- FF: 70, FOR: 130 (cyan) - only slightly worse still in-terms of linearity but it does improve on the min force. It also hits about max force.
- FF: 65, FOR: 150 (green) - this has the worst linearity out of my chosen four settings, having said that, it's still an improvement over the default 100/100 (yellow) as-well as the 120/80 (red). Also, it's the only setting that comes close to 100/100's minimum force, in-fact, it even beats it and achieves the best minimum force out of all combos in the test.

It's difficult to choose from those 4. I would lean towards 60/150 (magenta) since it has the best linearity out of all the ones that achieve max force, but in-terms of minimum force, it gets beat by cyan and red, killed by yellow, and destroyed by green.




What do you guys think of my results and especially of my final 4 choices? Opinions/advice please

 

I think this is misleading you spin. There appears to be no clear correlation between increasing drift setting and the output/result. In other words, it looks random. I reckon if you repeated these tests, you'd find them different.

That's because as i've since learnt from Flaux that these wheel profiler settings only enable any artificial spring and damper that is coded in the game. Wheelcheck.exe does not have any so enabling these in the profiler will not do anything.

I address this in my first answer.

Spin, clearly drift 3 makes the torque output non-linear in the same way increasing the rfactor 2 Steering Torque Sensitivity does. You said drift mode 3 feels good/better to you in practice. That means you want the output results from drift mode 3, so you dont want the response of drift mode 0 output by playing with other settings.

If you want to confirm this, try turning drift mode to 0 and increase the rfactor 2 STS value until you find the equivalent value that feels the same as drift mode 3 with STS=1 in rf2.

 

I find this all very interesting. I assume we are looking for best linearity with smallest deadzone.?
Is this test useful with a G25? I ran mine in the FCM wheel test. I got no deadzone, but a nice big spike in the middle of the FCM curve? Is there a place where folks are reviewing test results and establishing guidelines on what the curves mean and how to get your wheel optimized?

Thanks

 

No, no, no.....

I don't want the highly un-linear results of Drift Mode 3 (w/ FF: 100 and FOR: 100). I'd like linear results but, REGARDLESS OF LINEARITY, drift mode at less than 3 reduces FFB quality. Maybe the negative effects of drift mode @ less than 3 aren't reflected/measured in this test. Maybe drift mode FFB interference is caused and activated by the player's input on the wheel rather than what the game/software is inputting to the wheel and therefore may not be reflected/calculated by this test - I don't know. Dri mode at less than 3 causes the wheel to apply a very odd and FFB-fidelity-killing effect where it acts like your wheel is being resisted by super thick mud whenever you apply an input (player/user input) REGARDLESS OF LINEARITY. This does not seem to be reflected in the graph, and neither is the opposite odd effect (drift mode @ 4 and 5).

The effects have nothing to do with simply changing linearity. I know what simply having different FFB linearity curves feels like as I've experimented with different linearity settings with many wheels in the past including the linearity changing effects of things like adjusting the overall strength force of the T500RS, and adjusting the STS in RF2. The drift mode thing has nothing to do with linearity, it just so happens to ALSO affect linearity.

I am NOT trying to achieve a drift mode 0 output. Again, I want an as-linear-as-possible wheel but WITHOUT the artificial FFB effects that drift modes at 0, 1, and 2 add. Drift mode at 0 just so happens to be very linear, but it also induces weird, artificial friction/resistance/thick-mud effects on the steeeing and makes the Thrustmaster T500RS feel like the most free-of-friction/resistance wheel in the history of motors.

Again, the setting to achieve equivalent linearity (e.g. STS) will not achieve equivalent performance/behaviour of the wheel because of the different drift mode (0, 1, 2) doing it's "thing". It'll only be equivalent in-terms of linearity but drift mode at less than 3 adds fake effects from the FFB which seems to get activated whenever the user makes an input and therefore seems to want to artificially resist and make every user's input feel to the user like they're turning the wheel through thick mud. Heck, you can even feel it when simply rotating the wheel on the desktop.



The following is not a direct reply to any of your quotes but just a general question:.....

Theoretically speaking, would the FFB from two absolutely identical lines on the chart feel/behave exactly the same regardless of different FFB strength (FF/FOR) combos used to achieve each line (ignore the whole drift mode thing for simplicity's sake)??...If so, then wouldn't using different combinations of "FF" and "FOR" to improve linearity be just as good as any other method? It sounds to me like you don't advise aiming for linearity improvements through the use of different "FF" and "FOR" combinations and i'm curious about your thoughts behind that....