Jay's Force Feedback Wheel Controller

[jakob]

Hi,

I've finished my X-Sim FFB wheel controller. It allows to assign four independent working X-Sim axis to control the wheel. Single axis like motor RPM for fine vibrations in lower RPM, car speed for a stiffer wheel on higher speed or combined axis with X-Y-Z forces/car movements - or all axis together. Four axis that give a ton of freedom for FFB tweaking.

A G25 with low or medium force feels spongy and soft. The wheel starts to wobble and curbs or crashes are soft as well. This controller will produce fine vibrations without getting spongy and a mild force that wont affect your lap times in online racing. But still strong enough to cut your thumbs on a crash.

The code isn't 100% done. I still have to tweak the left/right movements for an aggressive feel. The controller runs with USO with a 256000 baud rate and 20ms delay (50 position changes per second). The crystal + baud rate needs some work, too. Maybe I can lower the delay a bit more. And a version for SCN5 Users will follow.

It will be very easy to solder - just a microcontroller, some capacitors one resistor and a crystal. Layout, code and parts list will follow.

Next step is getting the h-bridges + motors running and do the profiling job (with help from Nima). The h-bridges will run in locked antiphase mode. Well, I'm not sure if my h-bridges will work proper, because the controller will cause a big PWM-mess on the motor outputs...I hope they won't be gone in a puff of smoke after testing...har har.

I also want to implement a few digital outputs for a shifting light or an 7-segment LED as gear indicator (based on a fith axis) and an encoder input to get the wheel position. But this can be also done by a joystick controller like Mjoy or Leo's controller.

Test run video



-------------------------------------------------------------------------------


EDIT April, 07 2010
Optimized the code and dropped the USO delay from 20ms down to 1ms @ 115200 baud rate. Yay!


-------------------------------------------------------------------------------


EDIT April, 09 2010
Here is the layout in minimal (working, but not final) config. The cables can be soldered directly to the microcontroller or to header pins. VCC and GND can be taken from a joystick controller like Mjoy or a cheap wall wart. The layout is really simple, there is no need of an etched PCB. A cheap prototype board will do the trick.

Jay

[willynovi]

jakob, nice news from you and nice work.
I can´t wait to see that moving wheel in your hands.

I have my work in progress, but little progress.
I decided to use a L298 to build the h bridge, so for a FFB wheel I think 4A is enough.

As I mencioned I'll get values from rfactor plugin, as steering arm force.
If it doesn´t work as I want, I´ll work with X-Sim mixing some efects as you have done.

I´ll keep alert at your progress.

regards, Willy

[jakob]

Nice to hear that other are workin on a wheel, too. Keep me updated please.

[Nima]

hi jakob,
nice work.i don't can wait for test your controller.i can make a profil with slider
for the wheel. with slider is many simpler to make a nice feeling.i have make one for
the sim and i love it. you can use the slider in play in realtime.

[jakob]

I'll solder your test controller on weekend. Hopefully I get the time to finish the code also.

[jakob]

Dropped the USO delay from 20ms down to 1ms @ 115200 baud rate. This means up to 1000 possible wheel position changes per second.

Can't wait to test it with the hardware.

[sirnoname]

This is really a fast communication, USB HID (as G25) has a host hardware polling speed of all 10ms and the FF effects have a delay from about 100ms to 200ms until the reach the user.

[willynovi]

As I know, you can change polling speed for USB HID from windows setup, right now I don´t remember how, but if I find that I´ll post, maybe you can get better feelings from tyour G25.

[jakob]

This is really a fast communication, USB HID (as G25) has a host hardware polling speed of all 10ms and the FF effects have a delay from about 100ms to 200ms until the reach the user.

I've trashed the whole code. The five bytes send by USO (Startbyte + 4xPWM) were written in a string and from that string into the main loop/code. This solution wasn't the real McCoy and sometimes it angered the snort out of me. I had numerous PWM dropouts and byte errors when lowering the delay.

The new code works with an array and the 4xPWM can be processed without detour. I've checked the PWM output with my oscilloscope and its looking clean and without any brakes - even with a 1ms delay.

[njracer]

Hey Jakob this is a really cool project man. I have a couple questions if you don't mind taking the time to answer??

First is your controller going to be for existing off the shelf wheels? ie, you use your board in the g25 instead of the board that comes with the wheel? or is this going to be a project where it'll be scratch built, including motors and all??

if you're going to use your own motors, which i hope you do, how is the wheel going to stop after it reaches its range?? are you going to implement a mechanical stop or software??

are you going to use a gear reduction system? any chance of doing this through software?

very interested in this. i definitely want to build a wheel myself, can't wait till you release more info. with the ability to tweak the ffb through x-sim this is gonna be awesome. this is an awesome project you're doing and can't wait to see what you come up with