Thank you for +1'ing this point/question to the devs. I've redone my tests with correct settings in the WheelCheck program and the results are now as they should be (so please ignore the previous one comparing each % overall strength.) I repeated each test 5 times and these are the average. I looked at each collection of repeats in excel in a graph i found that each set actually had the same amount of tightness in spread so again you can ignore my previous remarks that 100% is more spiky than 60%. What we ideally want to be using is the 60% response curve as it has the highest peak torque range without turning non-linear (unlike 70-100% response curves). The major downside is that for 60% there is a huge 8% initial ffb deadzone. When you add in the fact that in order to avoid clipping the ffb ingame you have to use something less than a "car specific ffb multiplier" of "0.6" (in the megane and it varies for each car but usually needs to be even lower than this for faster and more powerful cars). Using "0.6" also has the effect of increasing the initial ffb deadzone from 8% to 13.3% and also lowers the strength of the dynamic range of forces you feel around the track. If we can get a "Min force" offset, not only could we remove the ffb deadzone completely but we regain some added strength to all the forces (especially felt and benefits on the low end forces...crucial for driving confidence and significantly increases the ffb realism...this is why i use 100% as the deadzone is significantly reduced and the low end is much stronger and responsive and it feels fantastic). As shown here (this is an old graph but is still good to demonstrate the deadzone issue): All it would take is a final step before the final ffb signal is sent to the wheel. Something like: The % has to be represented as a fraction (0.0-1.0) for this particular equation to work though (should be easy to work around though). Here's how the ffb input (at the wheel) vs ffb out (in rfactor 2) response curve (or line in this simplified example/model) would look with and without the offset: Brief explanation....the x-axis is the amount of ffb output (as a percentage ranging from 0-100%) that is calculated by rfactor 2's physics. The y-axis is the amount of ffb (as a percentage of maximum torque producible by the ffb motor) at your gaming wheel. The blue and green shaded regions (at the bottom) represent the range of forces along the y-axis (the amount of torque produced by your wheels ffb motor) that you will not feel because the torque is not enough to overcome the internal resistance of your wheel. And remember, these deadzones regions become even larger when you decrease in game "car specific ffb multiplier" because of the effect it has on increasing the ffb's dynamic range/spread. Can we have an offset feature in rf2 like in iracing pretty please?
Wow, thanks for such a ridiculously detailed thread DrR1pper. So your saying, in a nutshell, that you've proven that the best settings for the T500 are 100% force in driver and around 0.60 in game? I do concur that these settings feel the best but never realised that it is more to do with the wheel deadzone at the bottom of the range rather than max force output. I too wonder what improvement it would make to have a min FFB output offset, maybe more so for lower end wheels with such large deadzones.
100% and 0.55 ingame (you were using it remember!!! lol) for the megane. Might be different for different cars but not sure....haven't had the chance to drive many other cars yet. I did in the 60's f1 at brianza online and it felt so much more realistic and f***ing awesome in my hands! Tim has passed on the ffb base offset onto the devs so fingers crossed!
Strange but for me the T500RS feels best with the value of 40% in the driver. Everything else I personally find not very harmonious in the transitions. As example one of the most impressive vehicles on the steering wheel for me is the Mazda 787B from MaK and the last time I've driven it I have used a overall strength of 40% and in game something around 0.07-0.10 and it was the most awesome feeling in the steering wheel so far, even the self-alignment when load changes was almost perfect and relatively moderate quickly so that I could let go of the steering wheel to align itself to a part, and the steering of the car in general has behaved very dynamic, though I used - Steering torque sensitivity = "2.00000" - instead of 1.00000. Why this car needs such a low multipler in game I do not know but it is still very strong and fast and i like it.
That steering torque value has made your response curve severely non-linear which induces earlier clipping. But because your using such an extremely low in game ffb multiplier (0.07-0.10) you won't be running into that clipped area. If you have the pedal plugin, I would not be surprised if your ffb did not go higher than 30-40% in the corners because you are using such a low in game ffb setting. Using 40% in your t500 control panel is interesting. There is absolutely no difference in the linearity and characteristic shape of the ffb response curve between 40% and 60%. So you must just prefer a lower maximum torque or something. I don't know what made you try a steering torque sensitivity of anything higher than 2 but you really want to be using 1.0 (linear) and the play with the t500 overall strength value and ingame ffb multi to optimal levels.
Yes I've read it would not linearly but it suits me better and I think it fits better in the context of forces transitions for the conditions we have. The very low multipler is just with the Mazda. My steering wheel is mounted on something kinda like a torsions axle ( still experimental ). The effect - dampening and at the same time reinforcing. View attachment 11691 It transmits the resonances and, increasingly, the recoil of the steering ( by preloading ) which is the main issue with this wheels, the accel, speed and thus peak forces.
Please ignore this if you already aware but this is what is happening between ffb signal from when it's calculated by the physics engine to when it's received by your wheel then: Changing the "Steering torque sensitivity" value does the following to the ffb torque calculated by physics engine: This is not an exact representation of how much it changes the linearity, i don't actually know how agressive using "2.0" or "0.0" changes the non-linearity of the curve but this is sufficient representation for the purposes of my explanation. You (speed1) are using the red curve. As you can see, if you using the entire range of forces from rf2's physics (represented along the bottom) then you will be getting heavy ffb clipping for all forces above 60-70% (represented by the plateauing line). However, you said that you use a very low ingame ffb multiplier of something like 0.07 to 0.1. What this does is limit how much of the curve you are using: In this arbitrary (greater than 1.0) non linear curve i made up for this example, you will get no more than 75% of the max ffb force that can be sent to your wheel. Here is how the response curve and maximum forces sent to your wheel will actually look if you imagine that the first 0-29% of physics ffb on the x-axis (which is all you will use with 0.07-0.1 ffb multi) is stretched to represent the new 0-100% of forces you only use from the physics engine: When you factor in that you also use 40% overall strength in the t500 control panel vs the default 60% overall strength, all forces are reduced a futher 30-40%. This means the effective maximum force you can feel at your wheel and for all forces produced by your wheel are limited to 50% of it's full potential. If you're happy with this, then that's fine but i wouldn't advise it as it goes against everthing i've been saying about the problems associated with ffb clipping (which as you can see from your green line is going to cause clipping issues....however i don't know how aggressive the steering torque sensitivity value will make the curve non-linear...it could be a lot more progressive than the green line i'm showing here). But if you want to achieve the same response curve (i.e. the non-linearity you've set with "Steering torque sensitivity" and use up the full potential of forces achievable by your wheel, then what you should do is the following: You would need to use a lower a Steering torque value between 1.0-2.0 that still mimics the response curve you used previously. And then use a higher ingame ffb multiplier of around 0.6-0.7. Your previous response curve will (relatively) un-changed, only this time you will get more power from your wheel for all the forces. So for example:
Oh no now you challenge me I did not really want so deeply That's the problem with you guys, too much theory, I go by emotion and feelings because I've driven many different cars ( again ) to know how it should feel about to produce familiar feelings but still thank you for the instructive information. About the multibler I would have to refresh my memory since it is longer ago than I driven the last time but most cars are around 50-60 and I think rf2 ffb is the best I ever could feel on pc. edit: ah and i'm using the pedal plugin. Thanks to the ingenious creator.
Ah, ok, cool beans speed, as long as your happy. I just don't like to see people at a disadvantage (all be it an avoidable one) or end up chasing their own tails unable to seek more performance from themselves in the cars (like i was).
Years ago I already have been looking for such a good ffb, now for the available conditions it is very impressive. I also think the tire model has a great influence on and I am satisfied so far. Now I just need a rf2 which is more optimized. I'm going to be for a few weeks out of the house, this makes the wait easier for me because I want to use rf2, while verything else could not satisfy my demanding so far.
And.. if we put 100% in driver, and the Steering torque with maybe 0.7. Do we have an almost linear response?
Not for a t500. If you have a t500 and you use 100% instead of 60%, you will need to use 0.53-0.55 in the in game ffb multiplier. The reason being that 100% has a funny effect on the response curve. The first 0-60% of forces sent to the wheel are linear but 60-100% become flattened (clipped). You might then wonder why I'm using 100% and not 60%? Because strangely enough, using 100% has a smaller initial ffb deadzone down from 8% to 2.5%. When you then lower the ingame ffb multiplier from 0.7 to 0.55, it has the effect over using only the first 0-60% of forces at for the response curve of 100% overall effects strength (avoiding the 60-100% of clipped forces). The effective initial deadzone then becomes 4% (2.5% / 0.6 = 4%) As a result, the wheel feels stronger and more response in the lower end forces if you can reduce/remove the initial ffb dead-zone as much as possible. In the ideal (perfect) world, there should be no initial ffb deadzone. If you have some deadzone, it means you lose the first few percent of ffb information and this has a knock on effect on all subsequent forces that follow feeling weaker at the wheel than they should be. If you'll look at this graph, you'll see what i mean: Green is what we get at our t500's (with 60% overall effects strength). Notice the first 0-8% of ffb effects forces from rfactor 2 (x-axis) are producing 0% ffb torque output at the wheel (y-axis). Blue line shows a modified output from rf2 so that we don't get a deadzone.
Exellent research and clear explanations, thank you so much for all the work that should allow users to thrustmaster wheel to get the most,for me about this should be converted into synthesis and placed in a section dedicated to control devices ... Tim ?
Thank you Flex....I appreciate the acknowledgement from others that they find it useful as well and that I'm not just pissing off the everyone with graphs.
I mean, set a low steerint torque (http://i597.photobucket.com/albums/tt54/DrR1pper/1_zpsa2db765c.jpg) to balance the 100% FFB curve (http://i597.photobucket.com/albums/tt54/DrR1pper/t500graph1_zpsd394b45c.jpg).
I stayed silent but I read and learned a lot in this thread about my wheel, so yeah, big thanks a lot for your search!
Cheers DrR1pper, I've a question regarding information loss because of the dead zone. This is true but surely only for the middle range or not ? Which should mean it makes no difference under load such as with some degree of steering angle, especially when driving a corner if I'm not mistaken. edit: why I'm asking, I personally do not find something is missing because of the dead zone, not only that it prevents oscillation, the steering wheel also does not receive each vertical pulse such as height differences what means I can do it through the variable dead zone while increasing the overall strenght, if a car should respond and determine to steer bumps with straight directed steering, because not every vehicle transmits height differences to the steering or only in uneven overrun of the front wheels but not while straight driving.
