Lee bodnar's why current ffb doesn't work 2011 article

Not sure where to post this so…
First off does anybody know where I can find the original article as it’s no long available on google documents.

2nd does it still applied as it did back in 2011?

I read it back then but can’t remember too much about it apart from the fact that torque based ffb is the answer. Which off course simucube can handle.

Seems to me the hardware isn’t the problem anymore but the sims.

Any thoughts?


1 Like

I found this one, if correct. I remember reading it at the time :slight_smile:

All the current sims do seem to use just torque from the steering axis rather than using sine, square wave or ramp effects. Some are also using damping and friction to alter the feel a bit.

Yeah That’s the same post I found but the document is no longer available.

So is the orginal premise not ture any more?
Edit delete
So are we using torque mode I thought I we were still using positional.

Scroll down for english part:


Thanks for digging the English copy of that article.

He is right, the current consumer-oriented simulators do not measure the steering column torque directly from the user input to the simulator, i. e. the hardware is not in the loop in that. This is because the limitation of the directinput API, which does not allow that kind of flexibility.

There are professional driving simulators, costing much more, that do seem to have force sensors in the steering column. These use CAN or other methods, thus requiring custom interface cards to PC’s, so not something everyone could build at home. They will give more accurate feeling, but I do not know if they are used in racing simulators anywhere. Not at least with ones that use directinput.

1 Like

A future project for you Mika :slight_smile:

Well, we would need to convince one simulator to ditch directinput and develop a good, realtime interface for something else. Not an easy project at all. And not implementable via USB, not very easily at least. CAN would be much better for that type of things…

if i remember correctly, berney from simxperience contacted a bunch of game publishers about switching to something new, but received almost no interest.

publishers want to make money and cheapest way to do that is to stick with msft and directinput

Yep, if there is interest and money then anything could be developed. Not many people would buy five-figure priced steering wheel which would then only support that one sim…

Possibly develop it with a Car manufacturer or racing team using a pro version of AC or RF2 . With the hope it would trickle down to the consumer market in the future.

I can see a point in the not to distant future in the quest to cut the deaths on the roads all cars will be driven by AI by law. Leaving everybody bar the super rich left with the only option to get their driving fix on simulators. If it ever got that way Id buy the biggest electric ai driven car I could afford completely gut it the interior and put a sim rig in it. :slight_smile:

Yep, that would be the way to do it.

But, your original post suggested that torque-based wasn’t used. Well, closing the torque loop in the simulation is not used, but of course the motor control is being given by torque commands and position feedback is used at the drive and the simulation. So, its not torque that is the problem, is how the control loop is built.

True but I could only remember the jist of Lee’s original article that it had something to do with torque.

I had a look at my granity settings and noticed its now set to torque control Im sure it was on positional before but not 100% sure. Did this change with the ioni firmware you did recently? Or is my memory worse than I thought? :slight_smile:

Your memory is worse.

In sim racing application, torque control is used.

1 Like

We do still use positional information as well because the sim HAS to know where the wheel is positioned otherwise you could have torque but you might be going around in circles… It is possible that reading through all of the other posts that having the torque telemetry and a torque reading at the servo that that might compensate for position but I am thinking for accuracy a positional element would always be part of the equation… otherwise you would need some serious computing to have torque information streaming smooth and in real time.

I believe it has already been done for some pro simulator, such as F1 teams, etc, many years ago. Has it trickled down? no.

In my opinion, the main issue with FFB and our racing sims is that the FFB is position based in our games rather than torque-based. I believe this is one of the main reasons why using high torque values of FFB give way too severe and aggressive FFB even though the same magnitude of torque may be detected in real life - the same magnitude of torque may be detected in real life but the behavior of that torque and how it changes with regards to your inputs is completely different to our games which often feel like they want to power through against your will (I’m guessing because the wheel is told to rotate to a certain point in our games VS rea life where the “FFB” is purely reactive to forces).

In real life, FFB is REACTIVE whereas our game’s FFB is ACTIVE.

Also, real life FFB has a huge amount more inertia than our game’s FFB systems.

No matter what sim I play, I have to use lower FFB settings than I would like or else there are many FFB characteristics that are much too aggressive, sharp, powerful, and accelerative compared to real life.

FFB of simracing is still quite a ways away from real life. High powered direct drive wheels have exposed this more than when using consumer wheels with lower power. In some ways, consumer wheels are more realistic than high powered direct drive wheels.

Ffb force is torque based. But feedback back to the PC and game is position feedback.