Here's a bit of fun!

I've wondered for a while whether or not the brushless motor driver cards used in the flight control loading systems might be used for a steering wheel force feedback system. So I thought I'd have a shot at developing one....

The experimental FFB wheel is shown right. Mechanically the steering wheel drive is similar to the aileron axis on flight controls. But mathematically the loading calculations are quite a bit different - and the way the calculations are distributed over the PC and motor driver card is also quite a bit different as a result.

A significant difference is the higher rate at which the forces change, and I changed the system design to reflect this. It might be important to say here that I haven't used the standard FFB output from the racing sim - instead I've extracted dynamic data directly from the sim and used this to calculate the slip angles and forces etc required for the FFB system. This has let me play with the force calulations....

The experimental system is configured as shown in the diagram - here are the main elements:

• rFactor - I used this sim to experiment with because of the fairly good data set available via its internals plugin.

• A - A "feeder" application which stands between the racing sim and the main FFB software (B). The feeder app pulls the live dynamics data from the sim, does a bit of processing and sends it to the FFB software. The data is updated each sim frame and principally consists of a steering restoring moment (Mz) .v. tyre slip angle curve and vehicle slip angle data.

• B - The main BFF FFB application which takes the live Mz .v. slip curve data and vehicle slip data and exports it to the brushless motor driver card. Additional force data such for friction, damping and vibrations are also sent to the card.
  The FFB app also receives back from the card steering wheel position data which it passes back to the sim in the form of virtual joystick position movements (E).

• C - The motor driver card uses the vehicle slip angle data together with the wheel position to calculate the instantaneous tyre slip angle. It then uses this to calculate the instantaneous restoring moment using the latest Mz .v. Slip curve data received from the PC.

• D - Brushless motor to produce the wheel feedback torque - I used a single stage belt reduction to bring up the torque to a working level. The motor is the same one as I use in the flight control loading systems and gives smooth torque output.

• E - vJoy virtual joystick used to send the wheel position data back to the sim so that a separate joystick card isn't needed.

The experimental system configuration allows the data comms between the software and cards to take place at approximately the frame rate of the sim. Whereas the actual force calculations onboard the card are carried out much faster to keep the drive system dynamically stable.

By using a separate feeder application I can also write and run different restoring moment .v. slip angle tyre models without affecting the main FFB app or the card programming. For example the tyre model used in the video clips below is a Pacejka "Magic Formula" model with additional processing to handle low speed changes in the steering force response. I thought in the future I might be able to make the feeder app code open source to allow users to experiment with the models and/or write their own to suit specific cars and tyre responses etc.. This is quite challenging, but might suit keener racers who would like to manipulate the forces at a fairly deep level.

The Pacejka Mz curve calcs in the test feeder app use live vehicle wheel downloading data which is pulled from the sim at the sim frame rate - effectively a new Mz .v. slip curve is calculated and exported to the driver card each frame. It changes continually with the continual changes in wheel loading reported by the sim (including impact, pot hole and off-road loading changes). Tyre model settings such as the Pacejka tyre coefficients, camber angles etc are defined within the feeder app coding. Ideally the user would match these to those used in the sim's tyre setup.

NOTE the feeder app could be written with other tyre restoring moment models - it doesn't need to be Pacejka.

MOVIE CLIPS

Here are a couple of quick movie clips to illustrate the system.

This first one just gives a few views of the steering wheel mechanical assembly -

http://youtu.be/H5U3qlRC4yY

The second clip below is a longer one showing the system in action. I've tried to show the on-screen data together with the steering wheel movements. Unfortunately you can't feel the forces but you can watch the oscilloscope data which shows how the parameters change with the car dynamic situation.

http://youtu.be/ZwsFQrLhkL0

Watch the left-most scope screen - you'll probably recognise the shape of the live Mz .v. Slip curve (green trace) - note how it changes as the wheel loading changes. For example braking increases the front wheel download and raises steering torque levels, whereas accelerating has the opposite effect and lightens the steering forces. The pink "pointer" trace on the same scope shows the instantaneous tyre slip angle - so you can see where on the Mz .v. slip curve the steering is operating.

In the right-hand scope the pink trace is also tyre slip angle. The yellow trace is vehicle slip angle - watch how these change as the vehicle corners - when the vehicle starts to let go on a corner you'll see the vehicle slip angle increase and the tyre slip with it....

The green trace on the right-hand scope is the motor torque output.

OVERALL

Overall the FFB wheel seems to work quite well - the torque response to tyre slip is there and the loading responds to live wheel loading changes reported by the sim. The force response is sharp. Additional effects such as friction and damping can be added-in as can vibrations based on engine speed and power etc.

The system configuration with the feeder app allows a lot of flexibility in the tyre response modelling, but this would probably be of interest mainly to enthusiasts given the knowledge levels required.

For the overall system approach to work all the automatic steering help and speed adjustments that the sim can apply need to be disabled so that the steering angles used by the FFB system properly match those used by the sim. This probably adds "realism" to the experience right enough  - it might explain why my driving is so bad (any excuse)!

There might be enough potential in the system to work on it a bit more - I'll post any updates when I have them.        

© This site is copyrighted, If you'd like more information or have any comments please contact me at