Cousin of OSW (Open Sim Wheel)

Got some reworked ESP_WHEEL from:
https://www.xsimulator.net/community/members/fernando-igor.26352/

that I changed to fit my hardware (Pro micro + IBT-4) and did a wheelcheck. I have just mounted the encoder with a rubber band, so no real "drive" test yet, but it seems promising. Because of my low current power source, I was forced to limit the current to the motor with a 12/100w light bulb. Will do some test just to see if this firmware work with rfactor2. Should be nice to have a working firmware that you could change by yourself.

Have you tried both cards with rfactor2? I have no ffb in rfactor 2 with either ESP_WHEEL_0.3 or FFBWheel.0.16.leonardo.hex... Ok in LFS...

Did some test in LFS. FFB setting in LFS = 50. When stronger forces comes, the power supply trips. I can say that this motor could have been perfect for the job, if it wasn't for the cogging.... If I calibrate so zero comes exactly between two "teeth", than at least it goes straight if I let go of the steering.
https://photos.app.goo.gl/Sk4GoqpcFeXWKCnM2

Hope you guys is just on holiday during Christmas, starting to feel alone here... I have given up to get the leonardo based solution to work with rfactor2. Ordered a STM32. Yesterday i started to think about how it should be possible to solve the cogging issue software wise. And then I search the internet and found this: https://granitedevices.com/wiki/Configuring_cogging_torque_compensation. So now i need the leonardo solution to work with rfactor2 even more. The second option will be to build a PWM manipulator between the STM32 and the H-bridge with the desired function. If I just connect the rotary encoder to that card also it shouldn't be that hard to do.

i have just bought a copy of RF2 - will try it out when i find some time

That will probably be the best buy you ever done!
And fernando have already confirmed that the final version works in rfactor2.
https://www.xsimulator.net/community/members/fernando-igor.26352/

But now my goal is to try to make a cogging torque compensation. Damn, why aren't the final source available....
In my mind I have started to work out the function already, and discovered that the 2 last number on the motor represents the number of coggs/(pols?) My MY1020 has 20 and your MY1016 has 16.

MY1016


MY1020


The cogging isn't that big problem when turning, but going on a long almost straight, it can be disturbing when you jump between 2 "teeth"

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


i have been running RF1 for some time - i now need a more powerful pc to run RF2

my motor windings do not look like the picture you have - i will find the type of motor i have

how many magnets are in your motor - do you have a picture ?

so leonardo board is programmed to work with RF2 - do i need a later version ?

inside it looks more like this

Just picture googled MY1016 That I thought you had....(It's a XYD-6b in the movie right?) . I haven't disassembled my yet, but I found i guy who did:
http://forum.modifiedelectricscooters.com/viewtopic.php?t=2008

I see no reason to take it apart. I will drill the axle from the back without doing that. It seems to be a capsuled bearing, so there will
be more risk to get drill shavings inside if I take it apart.

There are four magnets in my motor.

According to Fernando it should work with leonardo.0.16.hex

My pc is A8-3870k cpu with a GTX1050 2 GB card, and I'm satisfied running RF2. Have all graphics setting on low, but isn't so important for me.

I've tried rf1 a couple of times after I started to use rf2, but when you once used rf2, you will never go back.

It seems that I'm not the first one to go with the MY1020...

you can see the steering wheel is quite worn where he has been holding it tightly !

Some update of my progress... Still haven't received the new board so I can test the wheel in a real drive environment (LFS works, but want to try it in rfactor2 before final conclusion), but have done some more tests.

My IBT-2 cards have arrived, so I tested to use them in parallel, and it works flawless. The current splits 50/50 between them.
I have 2 ATX power supply in parallel which gives me 55 A before they trip.

Another interesting thing for those who want's more current, I found that the FET (LR7843) mounted on the IBT-4 cards actually have a max Current of 161 A!!!. The problem with that card is that the cooler is wrongly mounted, so they fry directly if you go above 40A.
Put them correctly on a CPU cooler and you have real "Monster" h-bridge! I have ordered insulation pads for that, but I probably won't do that when the IBT-2 work so fine in parallel.

I have measured the torque with a scale for luggage, and it was just what I expected. About 1Nm/10A which give me maximum of 5.5 Nm with my current supply. For now I have fitted the wheel from my old Logitech Momo Red that has a diameter of 25 cm. Have tried to do some "one hand" laps in LFS (No pedals connected yet), and I cannot see the need for more torque.

that is some great feedback

i am pleased to hear 2 x ibt2 works - i use them on my sim so i may double up for a test

the ibt-4 is a lot more powerful than i thought - what do you recommend to keep it cool ?

can we see a video of your wheel in action pls

(5.5 nm is a lot of torque for a FFB wheel, i have read the OSW must be turned right down to make it feel realistic)

5,5 Nm its good!!
This motor with 20 brush steps, is it perceptible when it is between 2 steps?

When the simulated car stands still and there are no forces it's annoying, but when you reach some speed, the forces are so strong so than
it isn't a problem. I calibrate it so zero is between 2 "coggs".

One hand driving in LFS
https://photos.app.goo.gl/6cC1b321NQIT0xFP2

To keep the FET cool, you have the solution already with CPU cooler, but it must be the "non-isolated" side to the cooler, not the upside of the component.

would it help if you adjusted the 'dead zone'

i found an article here on using an arduino to simulate a steering wheel

might be a good place to start if somebody wants to code thier own software

https://www.hackster.io/immortalx/arduino-gaming-steering-wheel-f50d67


the original project using a servo drive is here -- https://forum.virtualracing.org/showthread.php/97461-DIY-FFB-Wheel-based-on-Arduino-Leonardo

It's the ESP_WHEEL_0.3 that Fernando have reworked a little that I have been using so far. It still have some errors and doesn't work with rfactor2. After some reading of Bernard Bergers OSW, I realized that rfactor2 is actually the easiest sim to make you own ffb solution to because the ffb output s available via an open API. So I removed the ffb part from the ESP_WHEEL source and added the ffb signal via comport communication from the API and now I have a great ffb wheel for rfactor2. I noticed another thing that I suppose is related to the IBT-2 card. It doesn't produce the same current in both direction with the same pwm output. For now I have just rescaled it linear, but have plan to do a full linearity function so I get the correct Nm proportional to the output from the API. Theoretical, this solution should be very close to the closed loop servo wheels. My solution will not compensate for difference in resistance because of temperature etc.

Have not fully completed the hardware with pedals, but did som test driving in rfactor2 which is coming to a new world when it comes to ffb compared to LFS. I have turned down so I use about 4 Nm at peeks, and with the current wheel (25 cm Logitech Momo) that is more than enough. I can now set the filter to zero without the annoying "rattling" that the G27 gaved. I can feel every little detail on the road, and the car almost compensate oversteer by itself.

thats great news ! can you tell us more about how to use your wheel in RF2

it may be the motor that is the reason you have less power in one direction - if you swap the wires and the motor wires do you get the same problem in th eother direction

a neat thing to do would be fit a torque sensor to the wheel boss so the force can be adjusted in a closed loop - this would compensate for when the wheel gets hot !

is the code available for the ESP_WHEEL ?

i feel this project is getting better all the time and the DC DD wheel is becoming a serious contender