OSW, 1:1 FFB, Power-Steering

[Marek Lesniak]

Your is 8.8Nm. Clio is set to 13Nm. 13/8.8 = 1.4773... ~1.48. Set FFB Multiplier to that value and see how it drives then. Keep in mind though, stock CSR-E rim has a diameter of 300mm and real car's probably has 320mm. If you want to also take an account for that, you should...
300/320 = 0.9375
0.9375 * 1.4773 = ~1.385
...set FFB Multiplier to 1.38.

 

[Spinelli]

Your is 8.8Nm. Clio is set to 13Nm. 13/8.8 = 1.4773... ~1.48. Set FFB Multiplier to that value and see how it drives then. Keep in mind though, stock CSR-E rim has a diameter of 300mm and real car's probably has 320mm. If you want to also take an account for that, you should...
300/320 = 0.9375
0.9375 * 1.4773 = ~1.385
...set FFB Multiplier to 1.38.

If I'm understanding Vittorio correctly, he's saying the FFB seems too strong so I don't think upping the FFB strength more will help but, in-fact, make his problem worse. Of course I could misunderstanding him. I think he's saying his wheel is often outputting close to it's max force which therefore means the in-game car's physics are often generating close to 13 Nm, and that he's surprised that the Clio is so constantly/often reaching those high 13 NM / 100% forces (I'm assuming he's running 1.00 FFB multi and 1.0 STS).


Vittorio, if the Nominal Max Steering Torque is set to 13 Nm then that means the game will tell your wheel to output 100% force when the in-game car's physics hit 13 Nm. Sorry, I think you know this and I don't mean to be condescending, I just want to be 100% sure


Also, from the game's 13 Nm and downwards, the 100% output signal to your wheel will scale down linearly with the in-game physics (eg. 13 Nm = 100% FFB output, 6.5 Nm = 50% FFB output, etc.) unless you adjust STS (Steering Torque Sensitivity) from 1.0. Then that will apply a curve to the FFB output signal in relation to the in-game car's torque. For eg. 13 Nm = 100% FFB output, but 6.5 Nm = 70% FFB output (if you're STS is more than 1.0), or, 13 Nm = 100 FFB output, but 6.5 Nm = 30% FFB output (STS lower than 1.0).

NOTE: The above regarding STS is what I personally think it does from playing/feeling as-well as the name "sensitivity" (just like steering/pedal sensitivity which also adjusts curves but those of user-input rather than FFB output). Some don't agree with me that STS works like that so please don't take what I said about STS as 100% set in stone.

 

[DrR1pper]

NOTE: The above regarding STS is what I personally think it does from playing/feeling as-well as the name "sensitivity" (just like steering/pedal sensitivity which also adjusts curves but those of user-input rather than FFB output). Some don't agree with me that STS works like that so please don't take what I said about STS as 100% set in stone.

That is how it works. This was proven to be true by TechAde some time ago with motec plots of recorded ffb output vs virtual steering wheel torque in rf2 which conclusively showed that STS works just as described.

 

[DrR1pper]

Your is 8.8Nm. Clio is set to 13Nm. 13/8.8 = 1.4773... ~1.48. Set FFB Multiplier to that value and see how it drives then. Keep in mind though, stock CSR-E rim has a diameter of 300mm and real car's probably has 320mm. If you want to also take an account for that, you should...
300/320 = 0.9375
0.9375 * 1.4773 = ~1.385
...set FFB Multiplier to 1.38.

This will make the wheel ffb/virtual torque output 1 to 1 but will cause massive ffb clipping if the optimal ffb multi for the clio is at or below the default 1.0 value (which i'm pretty certain it is).

Vittorio, correct me if i'm wrong but are you saying that your ffb wheels peak output of 8.8Nm already feels so/too strong (or maybe more accurately described as feeling too/over reactive) which makes you question whether the real clio should have/experience peak torques of up to 13Nm? If so, i can completely understand. The reality is that torque for torque, in real life the real steering wheel should not feel as reactive as our ffb wheels. The reason for this is that the real car has very large inertia (in it's steering system mass) which the forces/torques generated by the tyre-road interaction must travel through first (which dulls how quickly it causes the steering wheel to change speed/direction) before it's affect reaches the drivers hands. This inertia component is completely missing in ffb wheels which means we get the forces/torques generated by the tyre-road interaction directly. Unfortunately, there is no way to resolve this problem with consumer wheels. You cannot fix this problem by software alone as the problem is occurring at a physical hardware level and therefore requires both a hardware and software solution.

 

[vittorio]

Exactly Spinelli and DrR1pper, this is what I was trying to say. If 8.8Nm already feels so strong (STS and FFB multi at default 1.0), are the 13Nm set as Nominal Max Steering Torque correct for the Clio? (which should have power steering if I'm correct). Only real steering wheel torque data I have found are (quouting http://http://isiforums.net/f/showthread.php/19205-CSW-or-T500RS?p=268857&viewfull=1#post268857):

One of the Design Judges (Steve Lyman aka ‘Ergo Man’) was kind enough to send along the
following data. This is static measured torque data he has acquired over the years using a Sensor
Development hand wheel torque sensor:

Autokraft Midget = 6.0 Nm (55 in-lbs) No power steering, 380mm (15.0”) Ø steering wheel.
Lola T332 = 5.5-8.5 Nm (50-75 in-lbs) No power steering, 318mm (12.5") Ø steering wheel.
Dodge Viper SRT-10 = 4.5 Nm (40 in-lbs) Power steering, 350mm (13.7" Ø steering wheel.
Lotus 20 = 4.0-4.5 Nm (35-40 in-lbs) No power steering, 330mm (13.0") Ø steering wheel.
Lola T190 = 6.0-7.5 Nm (55-65 in-lbs) No power steering, 318mm (12.5") Ø steering wheel.

So I wondered is Nominal Max Steering Torque set that high for the Clio because of the missing power steering simulation? IIRC the Clio has lowest Nominal Max Steering Torque of all ISI cars. But looking at the data quoted I wonder even more why Nominal Max Steering Torque is set that high for all cars (assuming these data are correct)... All I want is realistic simulation and realistic forces.

(Thanks for the explanations how STS works and missing inertia component simulation, interesting reading. Though I don't understand why it's not possible to simulate inertia components, guess I have to read Bodnar's whitepaper )

EDIT: the data are static measured torque, should find the exact source what exactly got measured but I guess probably only static (not in race) forces, which would explain these low values.

 

[Comante]

Well, the answer maybe in the question: the data you posted name a "hand wheel torque sensor", the name suggest something that measure the forces that the hand feel or need to exercise. The ISI data should be referred to the wheel shaft, their reading is naturally higher not taking into account the lever that the wheel diameter create.

 

[DrR1pper]

Well, the answer maybe in the question: the data you posted name a "hand wheel torque sensor", the name suggest something that measure the forces that the hand feel or need to exercise. The ISI data should be referred to the wheel shaft, their reading is naturally higher not taking into account the lever that the wheel diameter create.

+1

Hand wheel torque sensor is measuring the amount of torque the driver is applying to the steering wheel system which is also the amount of torque he "feels" from the steering wheel. But how much torque he will "feel" is determined by how much torque the driver wishes to apply to the steering wheel.

Virtual steering wheel torques from the sim however are the torques produced from the tyre road interaction that are then outputted to our ffb wheels, not the amount of torque the driver is applying to the steering wheel.

 

[Lazza]

So I wondered is Nominal Max Steering Torque set that high for the Clio because of the missing power steering simulation? IIRC the Clio has lowest Nominal Max Steering Torque of all ISI cars. But looking at the data quoted I wonder even more why Nominal Max Steering Torque is set that high for all cars (assuming these data are correct)... All I want is realistic simulation and realistic forces.

This is further complicated by the fact that a high NMST for a vehicle has the effect of generally reducing the FFB experienced on the majority of consumer wheels. But as Spinelli reminded me earlier, once you have a powerful wheel and set its STC (steering torque capability) to a value that exceeds the NMST of a vehicle, the game effectively then ignores the NMST for force calculations. So setting a high NMST in an effort to reduce the calculated steering force to a level more befitting a car with power steering will stop working when you configure a more powerful wheel. (is this what ISI was trying to do with the high NMST value? I don't know)

I haven't yet tested the options for a modder with the new steering system, and whether you can 'trick' it into producing lower forces. I've got a feeling you could, because of the parameters you can set which appear to let you violate some basic leverage principles.

*Edit:

This was proven to be true by TechAde some time ago with motec plots of recorded ffb output vs virtual steering wheel torque in rf2

Nitpicking, but Motec wasn't involved. At the time my plugin also didn't do the FFB output. It does now so you could easily do the same test

 

[DrR1pper]

Nitpicking, but Motec wasn't involved. At the time my plugin also didn't do the FFB output. It does now so you could easily do the same test

Sorry if i incorrectly recalled which tool he used to graphically show STS's affect on ffb output. It was quite a while ago.

 

[Marek Lesniak]

Exactly Spinelli and DrR1pper, this is what I was trying to say. If 8.8Nm already feels so strong (STS and FFB multi at default 1.0), are the 13Nm set as Nominal Max Steering Torque correct for the Clio? (which should have power steering if I'm correct).

You can check in telemetry, how much steering torque the car generates in corners but doing as I proposed, will also tell you, if you were already near max torque with FFB Mult at 1.0.

The reality is that torque for torque, in real life the real steering wheel should not feel as reactive as our ffb wheels. The reason for this is that the real car has very large inertia (in it's steering system mass) which the forces/torques generated by the tyre-road interaction must travel through first (which dulls how quickly it causes the steering wheel to change speed/direction) before it's affect reaches the drivers hands. This inertia component is completely missing in ffb wheels which means we get the forces/torques generated by the tyre-road interaction directly. Unfortunately, there is no way to resolve this problem with consumer wheels. You cannot fix this problem by software alone as the problem is occurring at a physical hardware level and therefore requires both a hardware and software solution.

I wouldn't say it has very large inertia. Whole steering rack assembly weights about 6-7kg, so a bare rack might be something like 2-3kg. There is some friction from seals and on steering gear but not that much. Then, you have several kg's of steering shaft in few parts. Here, you might get a bit more inertia. Of course, it all counts but don't overestimate its effects.

 

[DrR1pper]

I wouldn't say it has very large inertia.

Yes, you're right. I meant to say it has a very large inertia compared/relative to ffb wheels.

 

[Spinelli]

That is how it works. This was proven to be true by TechAde some time ago with motec plots of recorded ffb output vs virtual steering wheel torque in rf2 which conclusively showed that STS works just as described.

Wow! Thank you! I wish you chimed in during my "battle" with Paul Loatman about this a few months back.

---

You know what would be cool? If ISI posted the NMST of every car. I'd say right near the FFB multi. I wouldn't call it "nominal max steering torque" though as most people probably won't know what the heck that means, lol. I don't know if many people will care about it but it would be really cool to know how much virtual torque equals 100% FFB output. Maybe some people with high but not very super high powered wheels (eg. Vittorio's 8.8 Nm wheel which even has the capability to go up to 10 Nm or 11 Nm , the Accuforce at 13 Nm, etc.) wil be interested in this as maybe they would want to scale all the cars' FFB strength relative to each other. Knowing the NMST of every/any car you drive would be essential to achieving FFB relative consistency. You can then adjust this by using the FFB multi (no need to adjust NMST because adjusting the FFB multi essentially does the same).

I think it would be very cool. Just have a number like "13.25 Nm" beside the FFB multi or something. Or it could say "13.25 Nm @ 100% FFB multi". Or, ISI can just get rid of the FFB multi all together and replace it with the NMST setting which is basically what iRacing does if you enable it. Or, ISI can keep the FFB multi but just replace the numbers and percent symbol with the "xx Nm" so basically the game does the math between the particular car's NMST and your FFB multi and outputs that number as you adjust the setting rather than saying "xx %".

 

[JmbJr]

STS effect on FFB output

Sorry if i incorrectly recalled which tool he used to graphically show STS's affect on ffb output. It was quite a while ago.

I made a thread back in Jan that provided more in-depth analysis on how the STS affects the output FFB.

http://isiforums.net/f/showthread.php/27988-Comments-on-STS-with-some-data-analysis

Cheers

 

[DrR1pper]

I made a thread back in Jan that provided more in-depth analysis on how the STS affects the output FFB.

http://isiforums.net/f/showthread.php/27988-Comments-on-STS-with-some-data-analysis

Cheers

I know JmbJr, however, STS's affect on ffb output was also proven a year ago by TechAde here: http://isiforums.net/f/showthread.p...B-MPs-each-car?p=347924&viewfull=1#post347924. I referred his proof because it was my first experience of a proof for how STS works. I should have mentioned yours as well. Sorry.

 

[JmbJr]

I know JmbJr, however, STS's affect on ffb output was also proven a year ago by TechAde here: http://isiforums.net/f/showthread.p...B-MPs-each-car?p=347924&viewfull=1#post347924. I referred his proof because it was my first experience of a proof for how STS works. I should have mentioned yours as well. Sorry.

No need to mention me, because I based my analysis on the amazing discussion between you and TechAde. My curiosity was mostly if I could figure out the model ISI is using on the STS to see if I could "linearize" my T300. But that is away from the topic of this thread

 

[DrR1pper]

Btw, good job figuring out the ffb equation with STS. It's clearly a perfect match when you trace a graph of the equation over your plots.

 

[DrR1pper]

Jmblr, I was reading the old thread and this bit from one of TechAde's posts caught my eye (here) where he said...

...we see an STS of 0 produces ~85% FFB, an STS of 1 produces ~92% FFB and an STS of 2.0 produces ~99% FFB.

...and i couldn't help myself but want to check it against the equation you gave. I find it to be a thing of beauty to see such a perfect match.

 

[JmbJr]

Jmblr, I was reading the old thread and this bit from one of TechAde's posts caught my eye (here) where he said...



...and i couldn't help myself but want to check it against the equation you gave. I find it to be a thing of beauty to see such a perfect match.

That is awesome, DrR1pper! I am glad our approach gave the same answer!

Have you tried any STS?, or do you use any STS value? I did some tests a while ago with STS = 0.3 and MinForce ~ 10%, which gives a "linear" response output according to my spreasheet. The FFB felt a little "loose" on corner entry, but I could feel more force variation near saturation (FFB ~ 1).
Right now, I am using STS = 1 and MinForce ~ 4%, mainly because I really don't want to spend much time right now tweaking the STS instead of racing (did that for PCars and that was the whole reason I bought a PC and start racing in rF2 )

 

[DrR1pper]

I've always kept my STS default.

 

[Jokeri]

This is further complicated by the fact that a high NMST for a vehicle has the effect of generally reducing the FFB experienced on the majority of consumer wheels. But as Spinelli reminded me earlier, once you have a powerful wheel and set its STC (steering torque capability) to a value that exceeds the NMST of a vehicle, the game effectively then ignores the NMST for force calculations. So setting a high NMST in an effort to reduce the calculated steering force to a level more befitting a car with power steering will stop working when you configure a more powerful wheel. (is this what ISI was trying to do with the high NMST value? I don't know)

I haven't yet tested the options for a modder with the new steering system, and whether you can 'trick' it into producing lower forces. I've got a feeling you could, because of the parameters you can set which appear to let you violate some basic leverage principles.

*Edit:



Nitpicking, but Motec wasn't involved. At the time my plugin also didn't do the FFB output. It does now so you could easily do the same test

You can alter the NMST via upgrades.ini or the whole steering if you like.

Also:
SteeringShaftBaseLeft=(0.0,0.1775,-0.372)
SteeringShaftBaseRight=(0.0,0.1775,-0.372)
SteeringShaftAxis=(0.0,0.0,-1.0)
SteeringInnerTable=(0.24144,0.17,-0.372)-0.15244,0.17,-0.372)
SteeringInnerTable=(0.15244,0.17,-0.372)-0.24144,0.17,-0.372)

if the steeringshaftbaseleft/right middle value is same as the middle value in steeringinnertable you have zero FFB
make the value bigger and you have more ffb