FFB in race sims is stupid

[redi]

...and Leo Bodnar tells why (https://docs.google.com/viewer?url=http%3A%2F%2Fleobodnar.com%2Ffiles%2FWh y%2520Force%2520Feedback%2520In%2520Computer%2520S imulators%2520Does%2520Not%2520Work.pdf).

Excellent article, food for thought. Especially for race sim developers.

[Viper]

YOU'RE stupid!

Oh... Sorry, pre-programmed response that I should have gotten rid of many, many years ago. :mrgreen:

Thanks, I'll have a read, should be interesting. :)

[phil23]

Most importantly to me Race Series, GTR2 and all good rFactor mods like V8 Supercars Unleashed are much better for the FFB we have than not having any!!!

Don't get me wrong, if one day people like Leo can create something better then I'll be more than appreciative, but for now it makes my sim racing more enjoyable and involving:thumbup:.

[redi]

YOU'RE stupid!

Oh... Sorry, pre-programmed response that I should have gotten rid of many, many years ago. :mrgreen:

Thanks, I'll have a read, should be interesting. :)

AM NOT!

I'm wondering if game devs will be open to improve this standard way of implementing FFB.

Still, I'm not sure whether the current way of implementing FFB is completely off. Leo argues that we steer a car by force, not by steering wheel position, but that is only partially true I think. The driver has to excert force onto the steering wheel in order to get the wheel to a certain position, but the position of the steering wheel (and that of the front wheels) does determine how a car turns. We do not push the car into a certain direction with the steering wheel, the tyres do.

He also argues that if you would have a car that had no forces act on the steering wheel, i.e. you would control the car only by the steering wheel's position, you would not be able to drive it. I think that that is not true. Mercedes has made a car that was controlled by a joystick (http://autofix.com.au/blog/mercedes-benz-scl600-joystick-controlled), i.e. there are no forces of wheels and steering rack on the joystick, and yet this car was perfectly driveable. Hence, I think that you don't need this real-life force feedback to be able to drive.

Still, he is right about the nature of steering, which is applying a force through the steering wheel, which moves the wheels (and therefore the steering wheel) until the wheels point into the direction that makes the car go into the right direction. And while turning the steering wheel, you'll get counterforces depending on factors like car speed, tyre condition, road surface etc. Hence, it's a two-way interaction that needs to be modeled.

[Monsoon]

YOU'RE stupid!

Yellow card for you! :whip:

Haha jokes. :whistling:

Very interesting read, a bit technical but still easy to undrstand what he's going on about. I feel almost cheated with the way sims and FFB has been working over the years! :angry:

[Monsoon]

I agree with redi, surely steering position is still relevant. Wouldn't the steering position determine the amount of force needed to get the wheels there? Or something :?:

God I need to read everything again:laugh:

[lesthegringo]

Bloody hell, I'm pretty certain I'm going to leave it to the experts....

[redi]

I agree with redi, surely steering position is still relevant. Wouldn't the steering position determine the amount of force needed to get the wheels there? Or something :?:

God I need to read everything again:laugh:
Leo is right that steering a car is all about forces. Turning a car can be divided into a number of steps:

1. Basically in order to turn the car, the front wheels have to be in a certain position in order to be able to deliver the appropriate force to turn the car.

2. Getting the front wheels into that position requires turning the steering wheel to a certain position (the position of front wheels and steering wheels is directly correlated).

3. Turning the steering wheel to that position requires a certain amount of force, and that force depends on a number of factors like car speed, tyre grip (which includes road surface properties and tyre properties), front wheel orientation and steering rack properties.

Hence, you have a closed loop system with components exchanging forces while turning the steering wheel, finally resulting in a position of steering wheel and front wheels that makes the car turn as you want it.
This basically means that the force that is required to turn the steering wheel must be dynamically calculated and modified.

[freejrs]

But surely the force required to move the steering wheel can be directly linked to the amount of rotation required for the same vehicle maneuver, depending on circumstances?

Obviously it depends on many other factors like vehicle speed, road surface camber etc. etc. but the way Leo explains it makes it seem like the rotation value is redundant and is always second to the force input required to move the front wheels.

I understand the differences between the two ways of calculating and relaying FFB information but what advantages would this new/better/more accurate method of calcualting FFB give us?? Higher resolution of forces, more accurate forces coming through the wheel??

[-=Prodrive=-]

I think that a lot can be done with current wheels with better FFB commands. The difference between RACE07 and GTR2 is massive in FFB terms. Whilst I might bag SimBin at times for various things, the one thing they got really right was the FFB in RACE07, its so much better than GTR2, and actually helps me when trying to work out what the car is doing. One FFB trait I wish was there was if you were understeering, that there would be a build up of resitance in the wheel followed by a light rumbling feeling when you have actually gone totally over the edge with understeer.

[redi]

Well there is a difference between knowing what a car does by feeling the FFB and making a FFB wheel feel as if you're driving a real car. The former can help your sim racing a lot but still make the steering feel unrealistic, while the latter is what Leo wants to 'cure' by a proper FFB algorithm.

I agree with freejrs that Leo's approach does not mean that the current approach is useless. As I tried to explain in my previous postings, position of and force on the steering wheel are inherently linked. To make it feel realistic, however, would require a constant calculation and update of the forces on the wheel in a much more thorough way than that is done now. In case of very fast steering wheel movements, you could get very inaccurate FFB if the sampling/FFB algorithm isn't fast enough.

[freejrs]

The difference between RACE07 and GTR2 is massive in FFB terms. Whilst I might bag SimBin at times for various things, the one thing they got really right was the FFB in RACE07, its so much better than GTR2, and actually helps me when trying to work out what the car is doing.

The way FFB is calculated in the two games, GTR2 and RACE 07, is exactly the same as far as I can tell. Indeed they use the exact same lines of FFB code.

The difference may be that the default settings used in both games are significantly different and also the fact that RACE 07 has a much improved tyre model over GTR2 (imo) which conveys the forces coming through the wheel in a more pleasing manner.

While I do prefer the FFB from RACE 07 to GTR2 I think the difference is very small and if you use the same FFB lines in each title you will experience very similar effects coming through the wheel.

Well there is a difference between knowing what a car does by feeling the FFB and making a FFB wheel feel as if you're driving a real car. The former can help your sim racing a lot but still make the steering feel unrealistic, while the latter is what Leo wants to 'cure' by a proper FFB algorithm.



The thing is I still think we need some of the "fake" information coming through the wheel in order to drive a simulation accurately. If we had absolutely 100% just realistic forces coming through our FFB wheels we would be missing some information that is vital to driving in a virtual world. Obviously we get other information through visual and audio cues but it's not enough sometimes. Without everyone having a motion sim rigged up I still think we need some extra information coming through the steering wheel, however subtle.

[gears]

The thing is I still think we need some of the "fake" information coming through the wheel in order to drive a simulation accurately. If we had absolutely 100% just realistic forces coming through our FFB wheels we would be missing some information that is vital to driving in a virtual world. Obviously we get other information through visual and audio cues but it's not enough sometimes. Without everyone having a motion sim rigged up I still think we need some extra information coming through the steering wheel, however subtle.

Agreed ^

While that article was an interesting read, I don't give it much credence.

Leo's plugin for rF was awful. So was real feel for that matter.

I know i'm the minority on this, but I thought both of those plugins were rubbish.

[Headsoup]

The thing is I still think we need some of the "fake" information coming through the wheel in order to drive a simulation accurately. If we had absolutely 100% just realistic forces coming through our FFB wheels we would be missing some information that is vital to driving in a virtual world. Obviously we get other information through visual and audio cues but it's not enough sometimes. Without everyone having a motion sim rigged up I still think we need some extra information coming through the steering wheel, however subtle.

And there, I think, is the dilemma. You can have the added forces to know what the car is doing, but it doesn't feel quite real, or it can feel real but you don't get the extra forces.

I don't think we could ever get everyone to agree on which one is better or preferred... Maybe make the added forces 'obvious' and toggleable?

[redi]

The way I understand the article, is:

Current situation: wheel gives position, software gives force as a result.

Leo: wheel must give force, software gives position resulting from that force.

While I think it should be: software gives resistance of/force on wheel, wheel gives position, software gives new resistance/force on new position, etc.
The steering wheel determines by its position which way the front wheels point, which in turn determines what the car does. Getting the steering wheel turned into the right direction requires force, and that force depends on the parameters mentioned earlier.

Hence, looking at the above I would say that the current implementation isn't so far from 'the truth'.

[redi]

And there, I think, is the dilemma. You can have the added forces to know what the car is doing, but it doesn't feel quite real, or it can feel real but you don't get the extra forces.

I don't think we could ever get everyone to agree on which one is better or preferred... Maybe make the added forces 'obvious' and toggleable?
I think that these extra forces that make up for the lack of the seat-of-the-pants feeling is something different than what Leo is getting at. Leo is saying that the way that wheel and software are working together is fundamentally wrong. On top of that you have the issue of conveying forces to the wheel that aren't there in a real car.

[freejrs]

Agreed ^

While that article was an interesting read, I don't give it much credence.

Leo's plugin for rF was awful. So was real feel for that matter.

I know i'm the minority on this, but I thought both of those plugins were rubbish.

I agree with you for the most part.

Leo's plugin seemed interesting but when I found out you had to change the settigns for every car you drive before launching the game I realised it was not a workable solution or even one that was worth trying out.

Realfeel can give a very good feeling through the wheel but only with a handful of mods. What's the point in having something which only works for some vehicles?? With certain mods it is really worth using (CART Factor and AE86 are two) but if you are looking for something that just works it ain't it.

That's why I made my own FFB settigns using the Controller.ini in rFactor to match my FFB settigns in RACE 07. Amazingly it works with all mods and feels quite acceptable as well.

The way I understand the article, is:

Current situation: wheel gives position, software gives force as a result.

Leo: wheel must give force, software gives position resulting from that force.



Leo is saying that the way that wheel and software are working together is fundamentally wrong.

As you demonstrate it is more or a mirror image, a reversal of the actual process that is taking place. Surely it gives a very similar result?

[gears]

That's why I made my own FFB settigns using the Controller.ini in rFactor to match my FFB settigns in RACE 07. Amazingly it works with all mods and feels quite acceptable as well.

Yeah the Race Series has the best FFB out-of-the-box of any sim yet.

And they went the extra mile to give you specific FFB settings depending on what wheel you're using. Not one size fits all like other sims.

[freejrs]

Yeah the Race Series has the best FFB out-of-the-box of any sim yet.



You havn't tried NetKar Pro then??:-D

[gears]

You havn't tried NetKar Pro then??:-D

lol yes I have :-D

But there's no AI and the netcode is garbage, so.....

My point still stands :-P

[redi]

As you demonstrate it is more or a mirror image, a reversal of the actual process that is taking place. Surely it gives a very similar result?
IMHO it must give a similar result, personally I think this whole thing is just a chicken-egg problem. As I tried to describe in my own interpretation, the process is a constant interaction of forces and positions. Depending on when or where you start looking in this process, you'll first see a force or a position, but eventually you need both.

It could be that the current FFB implementation simplifies things, for example it does not feed the force calculations with updated steering wheel positions and/or vice versa. However, I don't see how Leo's proposal is going to solve this as there, too, you'll be able to take shortcuts in order to keep the algorithm simple. IMHO in the end, the FFB algorithm needs to be a closed loop system with constant input of both steering wheel position and calculated forces on the steering wheel, with an output of momentary forces on the wheel.

[freejrs]

lol yes I have :-D

But there's no AI and the netcode is garbage, so.....

My point still stands :-P

Point taken......:laugh:

IMHO it must give a similar result, personally I think this whole thing is just a chicken-egg problem. As I tried to describe in my own interpretation, the process is a constant interaction of forces and positions. Depending on when or where you start looking in this process, you'll first see a force or a position, but eventually you need both.



In the article Leo states.......

You are turing a corner and the car goes a bit wide. What do you do? Do you add extra 10 degrees to the steering? No, you add a little force and the wheel turns a bit more.

I don't see any logic in this.

I believe you need to add both the force and the degrees of rotation simultaneously and both are of equal importance. Without one the other does nothing. You cannot apply a force to the wheel without having it also move through a certain amount of rotation.

I just don't see what benefit this new method would bring to the table. Would people even be able to tell the difference between the current method of calculation and the new one proposed by Leo?? :?:

[redi]

In the article Leo states.......

You are turing a corner and the car goes a bit wide. What do you do? Do you add extra 10 degrees to the steering? No, you add a little force and the wheel turns a bit more.

I don't see any logic in this.

I believe you need to add both the force and the degrees of rotation simultaneously and both are of equal importance. Without one the other does nothing. You cannot apply a force to the wheel without having it also move through a certain amount of rotation.

I just don't see what benefit this new method would bring to the table. Would people even be able to tell the difference between the current method of calculation and the new one proposed by Leo?? :?:
Indeed. What I marked in bold in your comment is exactly what I have been saying in my previous postings. You cannot start with one and then get the other and that's it, end of the line. It's a constant interaction of two effects that are interdependent.
Leo just seemingly turns things around while the actual subject is circular anyway as the things under discussion depend on each other.

Changing the direction of the car requires a force by the tyres onto the road. Changing the force of the tyres on the road requires turning the steering wheel, and turning the steering wheel requires force because the road exerts a reaction force on the tyres (Netwon's Third Law).

[Headsoup]

Indeed. What I marked in bold in your comment is exactly what I have been saying in my previous postings. You cannot start with one and then get the other and that's it, end of the line. It's a constant interaction of two effects that are interdependent.
Leo just seemingly turns things around while the actual subject is circular anyway as the things under discussion depend on each other.

Changing the direction of the car requires a force by the tyres onto the road. Changing the force of the tyres on the road requires turning the steering wheel, and turning the steering wheel requires force because the road exerts a reaction force on the tyres (Netwon's Third Law).

I think what Leo is saying, is that from the car's point of view, it doesn't matter that the wheel has been turned 15º, it matters that X amount of force was applied. The position of the wheel is only important to the driver, the car only reacts to which direction forces are acting in through the steering rack and how much force. So 0 force means the wheel is centered, that's the only reference point needed. Makes sense really.

[redi]

I think what Leo is saying, is that from the car's point of view, it doesn't matter that the wheel has been turned 15º, it matters that X amount of force was applied. The position of the wheel is only important to the driver, the car only reacts to which direction forces are acting in through the steering rack and how much force. So 0 force means the wheel is centered, that's the only reference point needed. Makes sense really.

Not completely true in case of road undulations of course, or when you drive with one front wheel on the grass ;) In such cases a force can be exerted on the tyres/wheels and in order to keep the steering wheel centered you have to apply a counterforce to the wheel.

It is true that when describing a car that is turning, you only need forces to describe what is going on. From the forces follows the steering wheel rotation. However, in order to reach the proper force that the tyres exert onto the road surface, you'll have to turn the wheel a certain amount. Again, it's chicken-egg :)

[JumpStart]

Whatever... turn it off and play, all games suck w/o FFB!

[gears]

Whatever... turn it off and play, all games suck w/o FFB!

Well there you have it. Case closed :thumbup:

Take that Leo! :laugh: :mrgreen:

[Headsoup]

Not completely true in case of road undulations of course, or when you drive with one front wheel on the grass ;) In such cases a force can be exerted on the tyres/wheels and in order to keep the steering wheel centered you have to apply a counterforce to the wheel.

It is true that when describing a car that is turning, you only need forces to describe what is going on. From the forces follows the steering wheel rotation. However, in order to reach the proper force that the tyres exert onto the road surface, you'll have to turn the wheel a certain amount. Again, it's chicken-egg :)

I was wrong about 0 force means centred wheel, as of course the net force can be 0 at any angle..

But I don't think it's chicken and egg. The car provides the wheel with all the forces that are acting on it, you apply forces in return to change direction, the net force is worked out and applied as either a stonger feeling of resistance on the wheel (if there is greater force from the car), or the car starts changing direction (if the greater force is from the driver).
If FFB can continuously calculate the net force shouldn't that give accurate FFB on what the front wheels are doing? (It could be how it already works, I'm no expert on the matter :unsure:)

[edubz123]

ok, noob question: If I raise the 'centering spring/effect' strength in my Logi G27 driver then I have to apply more pressure/force to turn the wheel/car.

Doesn't this satisfy Bodnar's argument?

[Headsoup]

^ I think (not expert opinion!) not really, because that pays no respect to what forces are acting on the car, it just pulls the wheel back to center, all the time. I think that's exactly the kind of thing Leo is saying is wrong. The car should be applying any centering force from the component forces created between the contact patch and steering wheel, not 3rd party software.

[redi]

But I don't think it's chicken and egg. The car provides the wheel with all the forces that are acting on it, you apply forces in return to change direction, the net force is worked out and applied as either a stonger feeling of resistance on the wheel (if there is greater force from the car), or the car starts changing direction (if the greater force is from the driver).
If FFB can continuously calculate the net force shouldn't that give accurate FFB on what the front wheels are doing? (It could be how it already works, I'm no expert on the matter :unsure:)
That is exactly what I've been saying in my past postings :) The only missing thing is that in order to make the tyres exert force on the road surface to make the car turn, you have to change the direction that the wheels are pointing. This you do by turning the steering wheel, i.e. change the steering wheel's position. Hence, steering wheel position and tyre forces to change the car's direction are directly linked, and the link between those two is that you have to apply force to turn the steering wheel because of the counterforces coming from the tyres and steering rack. That makes it chicken-egg, because tyre force requires steering wheel position which requires force to change (and hold).

ok, noob question: If I raise the 'centering spring/effect' strength in my Logi G27 driver then I have to apply more pressure/force to turn the wheel/car.

Doesn't this satisfy Bodnar's argument?

^ I think (not expert opinion!) not really, because that pays no respect to what forces are acting on the car, it just pulls the wheel back to center, all the time. I think that's exactly the kind of thing Leo is saying is wrong. The car should be applying any centering force from the component forces created between the contact patch and steering wheel, not 3rd party software.
Indeed, the force required to turn the wheel should be constantly calculated and fed back to the wheel.

[stagman]

Aw shucks, I've been happily simming for years without FFB and was thinking of getting a FFB wheel. Should I not do it??? :)

[redi]

Aw shucks, I've been happily simming for years without FFB and was thinking of getting a FFB wheel. Should I not do it??? :)

As others have said, it's not really relevant whether FFB is totally realistic (the whole sim racing stuff isn't realistic anyway ;)). Fact is that FFB gives you more information on what your simulated car is doing, which gives you more control and more immersion (and IMHO more fun) :)

[edubz123]

Aw shucks, I've been happily simming for years without FFB and was thinking of getting a FFB wheel. Should I not do it??? :)

^if you see a new Logi G27 for less than USD$250, jump on it.

[Stigasawus]

Hey, dont beat yourselves up..SIMULATE. Theres lots of forces you NEED FFB for. Real world car..slow speed = heavier steering (feel) so that when it gets lighter as the speed goes up your sensory inputs dont fade = no judgement, just reaction even on a straight due to track surface variations/camber then theres steering/suspension geometry (toe-in gives self centering but less of it gives less scrub & rolling resistance, then theres castor/camber angles..roll steer,bump steer, even tyre pressures & brake bias. The existing software aint too bad ( I'm sure I can feel road camber on Conrod at Mt Panorama esp near the top, or is my brain simulating that?) Dont forget that you turn in with the st. wheel but then your right foot gets involved & if your Cooper S is understeering or your Mustang is oversteering, the point where you havent got REAL FFB is where you become a passenger (at least you might spin the latter & "dissipate energy in all directions" but the only FFB left in a Mini is clenching at seat level).
I couldnt drive a bunch of pixels until I discovered FFB, then it started to SIMULATE !!

[AndreiRS]

Is that guy the same who did that FFB plugin for rFactor?

[saguenay500]

euuuuuuuuuu :?: about position or force.. i still not see why FFB are stupid...

:thumbdown:

eric

[Hitman_M3]

This all is quite interesting, and seems to confirm why I ended turning off the FF on my MOMO and just raising the spring effect almost to a 100% while also increasing steering ration in the sims ... it felt way more natural to use force applied to turn the wheel than to have it spin effortlessly to an angle. When you drive hard a car, on the road or on the track, you strive for a steering wheel angle that points imaginarily through the corner's apex to the exit line and that is done more with force than with simply rotation. Though I disagree with Leo in that in my opinion it's not so much dampening but probably the tyre slip angle what exaggerates that effect.

Now I'm driving with a TSW720 and harder springs, and it feels more natural to me than any FF wheel I have ever tried, where the experience you have when acting with force against it is at least ... weird. You either feel the clicking teeth, or the sudden let go of the motor that was countering your force and that is not nice.

I'd be interested to see how the hydraulic system of the ECCI wheels works, but I decided on a TSW because it had optical encoders and 720 degrees.

Thanks for the article :thumbup:

[Leper Messiah]

interesting to read, can't profess to understand all (much) of it. ;)

But, it would be interesting to see how the professional simulators do it. Like the redbull simulator. Would the methods employed in those coincide with Leo's opinion?

[AndreiRS]

Well my steering wheel, with the spring effect set to 0%, still has the "center spring feeling". I don't know how people config their wheels, but mine is almost like the real thing. And I don't really agree with 100% of what was said there, first because its not only "force" what is needed to steer, you have to turn some degrees too... Otherwise they would just put a metal bar there, and the more force you apply, the more the car will steer. LoL Like those things to measure force.

[mrpowcz]

...
I couldnt drive a bunch of pixels until I discovered FFB!LOL This is a genius sentence :lol::rofl:
To add something into discussion, I don't really care if the concept is false in this case, I simply like sims with FFB much more. Must depend on wheel heavily though, with my old momo I didn't like it at all but g25 is another story (how many times has this been said??)!

[Hitman_M3]

Interestingly, in the airplanes world the F16 has a joystick that responds to force and not just to travel, and there are mods to change the Thrustmaster F16 HOTAS into that type of input for combat simulators. So it's not like you couldn't do a steering wheel with rack and pinions that presses against load cells when turning left of right (You can use reduction gears to convert 2-3 turns of wheel into some milimeters of travel, as needed by a load cell).

[redi]

Interestingly, in the airplanes world the F16 has a joystick that responds to force and not just to travel, and there are mods to change the Thrustmaster F16 HOTAS into that type of input for combat simulators. So it's not like you couldn't do a steering wheel with rack and pinions that presses against load cells when turning left of right (You can use reduction gears to convert 2-3 turns of wheel into some milimeters of travel, as needed by a load cell).
But that is something different than what Leo is arguing. The force-control that Leo is writing about is the actual force required to physically move the tyres, wheels and steering rack. The steering wheel does move and there is a direct relation between force applied, steering wheel movement and wheel movement. There is also a physical relation between the force needed/applied and the movement of steering wheel, wheels and steering rack.

In case of the F16 joystick, there is no relation between the force on the joystick and the force required to move the flaps. The joystick hardly moves, so there is no relation between joystick movement and flaps movement, and the force required to move the joystick has no relation to physically moving the flaps.

[freejrs]

But, it would be interesting to see how the professional simulators do it. Like the redbull simulator. Would the methods employed in those coincide with Leo's opinion?

Well the Red Bull simulator, as far as I understand, is just rFactor Pro running on a motion rig. Whether or not they developed their own FFB methods for it is anyone's guess.

[Hitman_M3]

But that is something different than what Leo is arguing. The force-control that Leo is writing about is the actual force required to physically move the tyres, wheels and steering rack. The steering wheel does move and there is a direct relation between force applied, steering wheel movement and wheel movement. There is also a physical relation between the force needed/applied and the movement of steering wheel, wheels and steering rack.

In case of the F16 joystick, there is no relation between the force on the joystick and the force required to move the flaps. The joystick hardly moves, so there is no relation between joystick movement and flaps movement, and the force required to move the joystick has no relation to physically moving the flaps.

What the article is about, as I understood it, is enabling the part of the equation that is missing, i.e. the force you apply and not just the feedback you receive with the opposite or counterforce. In a real car you apply force to get an angle, but the amount you apply is according to the article decisive for some things & calculations which in the current state of FF isn't. The only way I can think of to add that force to the current FF steering wheels is a load cell that receives input on the force applied by the user over the wheel, same as the current brake load cells (Which actually do simply turn it into travel), hence the comment.

In any case, with modern assisted steering even in F1, there is also no longer much relation between your force at the steering wheel and at the wheels themselves.

[guybo]

His article is excellent. I'd like to meet Leo Bodnar and buy him lunch! But I do think FFB is useful and not totally stupid. I think it's stupid how it's implemented and when drivers turn the knob to 11 and get crazy FFB. Lately I've been playing with FFB and making it subtle, not as strong as I can handle. I use it to literally feel when I am hitting a curb or just into the grass or if I missed the curb on the apex or if I hit a certain bump.

I've found that the graphics, the sound and FFB can all come together with a certain synergy that helps me be more precise.

[redi]

What the article is about, as I understood it, is enabling the part of the equation that is missing, i.e. the force you apply and not just the feedback you receive with the opposite or counterforce. In a real car you apply force to get an angle, but the amount you apply is according to the article decisive for some things & calculations which in the current state of FF isn't. The only way I can think of to add that force to the current FF steering wheels is a load cell that receives input on the force applied by the user over the wheel, same as the current brake load cells (Which actually do simply turn it into travel), hence the comment.

In any case, with modern assisted steering even in F1, there is also no longer much relation between your force at the steering wheel and at the wheels themselves.
A load cell would be quite a challenge to get to work with a wheel that can rotate 900 degrees as a load cell measures force with a very limited movement range. I think that a more novel way is to measure the current needed in the FFB motor to counteract the steering wheel movement. I haven't thought it through at all, but it seems a natural system to use.

[MickeyMouse]

It seems the article is indeed arguing that front wheel reaction is based on steeing wheel force and not rotation. I hope I'm misunderstanding him, since that is totally wrong. The force to turn the steering wheel is incidental to get the wheel to the proper rotation you want. The front wheels only generate turning grip based on their slip angle with respect to the surface, not how much force is being applied to the steering column. IMO, the principle behind FFB wheels is correct, but a new type of hardware is required to fully replicate the feeling from a real car.

Interestingly, in the airplanes world the F16 has a joystick that responds to force and not just to travel, and there are mods to change the Thrustmaster F16 HOTAS into that type of input for combat simulators. So it's not like you couldn't do a steering wheel with rack and pinions that presses against load cells when turning left of right (You can use reduction gears to convert 2-3 turns of wheel into some milimeters of travel, as needed by a load cell).
The Saitek X-65 actually does have a loadcell in it and doesn't move at all. As with the first version of the F-16, a lot of people find it very difficult to control. I'm not sure why newer real fighter planes (F-16, F-22 and F-35 that I know of) use the force based joystick. I assume it has to do with the auto-trim capabilities in the fly by wire system. In modern planes you basically select the pitch you want and the FBW holds it with the stick released. You kind of nudge the plane around rather than control it directly.

[redi]

It seems the article is indeed arguing that front wheel reaction is based on steeing wheel force and not rotation. I hope I'm misunderstanding him, since that is totally wrong. The force to turn the steering wheel is incidental to get the wheel to the proper rotation you want. The front wheels only generate turning grip based on their slip angle with respect to the surface, not how much force is being applied to the steering column. IMO, the principle behind FFB wheels is correct, but a new type of hardware is required to fully replicate the feeling from a real car.

What Leo is saying is that the true basic physical process in steering a car is interacting forces. He's not saying that steering wheel rotation is not happening, but he's saying that in order to get to the proper (steering) wheel rotation, you'll need to apply a force to the steering wheel. Getting to the proper steering wheel rotation and tyre slip angle is a result of forces.

If you know all the forces, you know what is going on with the car. However, if you really want to take that approach, as Leo is suggesting, then FFB wheels need to have a very different design.

What I am arguing, however, is that because force and position are directly linked, if you implement FFB correctly then a different FFB wheel design is not necessary.

[MickeyMouse]

What Leo is saying is that the true basic physical process in steering a car is interacting forces. He's not saying that steering wheel rotation is not happening, but he's saying that in order to get to the proper (steering) wheel rotation, you'll need to apply a force to the steering wheel. Getting to the proper steering wheel rotation and tyre slip angle is a result of forces.
...which is the way it is now. The physics engine calculates the resistance force from the angle of the front wheels (aligning moment) and transmits that to the steering wheel. To maintain the current steering wheel rotation or to get a new one, a force must be maintained, applied or released by the user.

[redi]

...which is the way it is now. The physics engine calculates the resistance force from the angle of the front wheels (aligning moment) and transmits that to the steering wheel. To maintain the current steering wheel rotation or to get a new one, a force must be maintained, applied or released by the user.
Exactly. If the update of the force feedback is quick enough, you won't be able to tell the difference between the current implementation and Leo's proposal. Sure, on a fundamental level the current approach is not physically correct, but in practice the difference is/can be negligibly small, with the advantage of a much simpler implementation. Geez, that sounds like the basic definition of 'simulation' ;)

[paddomuncher]

All thing aside, i wonder how the car designers are going to tackle this problem.
It cant be long before we get a production drive by wire car that will need ffb or a system or have i missed something..

[MickeyMouse]

All thing aside, i wonder how the car designers are going to tackle this problem.
It cant be long before we get a production drive by wire car that will need ffb or a system or have i missed something..
I read a couple of years ago that they are working on that. Really small Smart-like compact cars have such a steeply angled steering column that it gets in the way of the engine and makes the steering wheel too tilted. Apparently they are/will use an FFB wheel and servos. For a simple passenger car, I think speed sensitive FFB is all that would be required (the faster you go, the heavier the wheel gets).

[DaPion]

Originally Posted by Stigasawus
I couldnt drive a bunch of pixels until I discovered FFB

... it’s a very nice and sharable sentence.

However I’d like that some FFB algorithms were more realistic.

For example, the steering wheel little movements that contribute to give the perception of speed in a real car, as far as I know, are not obtained considering the suspension oscillations caused by the small bumps on the road surface, but introducing an artificial vibration, that by some parameters can be modified in terms of frequency and amplitude to suit one’s own preferences. A similar method is used to simulate the passage of the wheels over the curbs.

If this is true, that is if I understand the meaning of the different parameters in the control file, in some situations FFB is more a trick than a simulation.