Designing an Arduino Leonardo Force feedback wheel

So on this Quarantine time i finally got some time to play some racing games Pcars 2, AMS 2 and Fm7 ,Fm4 all on a Xbox 360 controller which i can have some nice results with all assists turned off. But the desire of an force feedback wheel only grew more and more the more i played Pcars and Ams so i decided to look into the options here in Brazil

For perspective a new g29 would cost arround $400 and used g27's would cost me arround $300 and a used DFGT $150 so yeah even extremely used ones are expensive and lets be honest the satisfaction of designing and building something cant be paid

So after a lot of YouTube videos i found out about the DIY route and after some reading of the different builds in the forum i decided to try designing an DIY force feedback wheel using the Arduino Leonardo since i already have some experience on it.

My goals for this project are

-Using Easy to source and budget parts
-Simple construction using sheet material (aluminium,acrylic, carbon fiber...), Threaded rods, Metric hardware and perhaps some 3d printed parts
- Building a simple to build yet powerful Force feedback wheel
- Designing the whole wheel in Cad and open sourcing the design
-IF the build goes successfully i also plan on releasing a full build guide + with CAD files , electric Schematics and a full Bill of Materials

Since i have some experience on cad i wish to first design it in CAD and then building it in real life

Currently i'm on the phase of choosing the main components that will go into the build

Arduino Leonardo R3 Micro controller $4
I've considered the stm 32 but i went for the Arduino which i'm more comfortable with and i think a lot of people also
https://bit.ly/2LkLGQV


BTS 7960 Motor diver $4
easily and cheaply attainable and can handle high power motors
https://bit.ly/360UF35

895 Dual ball bearing DC Motor 12$

The motor i am considering is the 895 option because its the cheaper option with a reasonable torque i could find on aliexpress (i've looked into MY1016 , MY1020, MY 6812... but they were all to expensive for budget minded build but i could look into adding supports for those motors too)

https://bit.ly/35RMKoO

Using a HTD belt drive as a reduction considering the low speed version on 24v and a final 400 rpm speed which would result in a 1:15 reduction and using a 90% efficiency it would result in 13 Nm theoretical maximum torque and a 4.5Nm at 35%

Rotary encoder $12.5

the encoder i considered is this option

https://bit.ly/3cunk2P

But i'm not sure about what resolution would be better 1000 ppr or 600 ppr

Price right Now $32.5


I would like to see what people with experience on custom wheels think of my choices and what considerations they would give

everything you need to know is here - https://www.xsimulator.net/community/threads/cousin-of-osw-open-sim-wheel.10915/

Thanks for the reply i've read that thread and gotta say that you built an amazing wheel But the MY motors Although Not far from the RS-895 I've chosen the shipping is ridiculous ($115) ad sourcing them locally would also end up coasting 50% more so when i add up the arduino+driver+Psu the cost would reach an used g27 that i wouldn't have to fiddle with But i definetly wanna try building a DD like yours on the future

But considering that today i made up my mind on the motor since the next model with shipping would cost double of the 895 and the torque improvement wouldn't be that great.

Also decided that im going with a dual pulley reduction (Similar to Fanatec clubsport v2.5) aiming between a 1:9 or 1:16 ratio.
So here's my question what rpm should i aim at because I've seen some people talking that anything over 200 would be good and i think that's too low and with a 1:16 reduction i would still have 375 rpm with plenty of torque OR going with a 1:25 ratio would be better because i would have more and still enough speed (240rpm)

Forgot to add that tomorrow i'll be choosing the timing pulley setup probally going with either gt3 10mm or HTP 5m 15mm basically wil decide upon whats cheaper since both setups would easily hold the torque with that defined i'll choose the shaft and then start working on the Cad model (which should take anywhere between 5 to 20 days depending on my free time)

You won't get that amount of torque with that motor) It's just calculated figures not correlated with real state.

Rpm is not important

Stall torque is what is important

Well no but actually yes i know that calculated figures don't translate to real life directly but in my calculations i'm considering the efficiency of the pulley system which according to the manufacturer is 98% but in my calculations i'm using 90% squared since it's a dual pulley reduction also from my understanding of DC motors(correct me if i'm wrong ) the most torque they produce is during Stall but it's not very efficient due to the big amperage consumption so this is how i calculated the final torque considering the motor torque is 0,9N instead of the 0,98 that it's listed at.

Stall torque * reduction *(efficiency^number of reductions)= Final torque at the shaft
and
Final torque / steering wheel radius = Force in newtons/10 = force in Kg's
so it would look like
0,9N*20:1*(90%^2)=14.85N
and
14.58/(0,3/2)=97,2 /10 = 9,7 Kg's ​

Yes stall torque is important but if i have too low of a rpm if i use the wheel for drifting it would take ages for it to return naturally so i could countersteer properly and also the wheel wouldn't simulate vibrations correctly but looking at commercial options it seems that as i read here anything over 200 RPM should be good so i'll just keep it in that range


waiting to see what you guys think of my considerations =D

@Sieben

So i was watching this guy's video and he seems to stall his setup with a 1.5L bottle full of water (1.5L) and he says on the description that he's using a rs 775 motor and two sets of 1.5 to 4 inch pulleys so applying his results to my formula it would result in 1.4 kgs with 100% efficiency although his measurements are super imprecise and he says that he's having some slipping it looks like my formula translates well into real life

You just don't know the practical aspect.

And you misunderstood. Rated Torque: 39N.cm/4Kg.cm. equals to: 0.392266 N.m (newton meter)
https://handsontec.com/index.php/product/895-dual-ball-bearing-dc-motor-6000rpm12v/


This motors will be lost in the pulleys. You will get in better way up till some couple of Nm. Even if.

A 1.5 kg bottle, on a 30cm wheel, is not 1.4Nm. If it can handle only this weight on a wheel itself, its 7 times less in NM of force applied how it should be measured. And you wont get a 14Nm of Force, as you presume before)

Using a HTD belt drive as a reduction considering the low speed version on 24v and a final 400 rpm speed which would result in a 1:15 reduction and using a 90% efficiency it would result in 13 Nm theoretical maximum torque and a 4.5Nm at 35%

Your calculations, can be at least fair to way more powerfull motors.

Your set up is somewhere near the G25 values in better case. And don't forget the speed. They are the slowest, with all that gearings for that small motors to give a force.

Logitech DFGT: less than 2.5 Nm

Logitech G25: 2.5 Nm

Logitech G27: 2.3 Nm

Logitech G29: 2.1 Nm - not a retail release version of wheel

Logitech G920 2.2 Nm

Fanatech GT3 v2: 2.6 Nm

Thrustmaster Tx: 3.9 Nm

Thrustmaster T500: 4.4 Nm

Fanatec CSW v1: 4.8 Nm

Fanatec CSW v2: 7.1 Nm

You want the wheel to be direct drive, not geared

Well ideally yes direct drive would way better but the smallest motors for direct drive shipped to me would cost $300 which would be significantly over my budget so the only way for me to achieve a good Forceand still be on budget sadly is going with a geared option

As an example there's still lots of good commercially available wheels that aren't direct drive example for exemple Clubsport V2.5 which i'll be basing the dual pulley design of. And it uses a v belt system which is prone to slippage and i'll be using a timing belt setup that's way tougher and almost zero slip chance.


Picture taken from fanatec's website of the V2.5 belt system

Sorry i missed you response but i gotta reply that first the example i used gave the correct 2.2 N/M of torque since:

Torque=Force * Distance from where the force is applied to the center
so let's get our variables from the example
Force = bottle weight in newtons which converting is (1.5*9.807)
Distance = steering wheel Radius since the force is only applied in one side = 150mm or 0.15M
calculating...
(1.5*9.807)N*0.15M=2.2NM

http://hyperphysics.phy-astr.gsu.edu/hbase/torq.html
this also explains why his motor moves to 90° and stalls
Which turns out exactly the same as my spreadsheet (will post tomorrow on google docs)
​

Now with Proof that my math is correct the motor spec that you sent me the link is exactly the same as the Listing i linked but it's a different motor at the listing i posted you can see that are two specs of the RS-895 motor. One High speed version with a stall torque of 4Kg/cm @12v and 8Kg/cm @24v which is the one you sent the listing and One Low Speed High Torque version with a stall torque of 5.2Kg/cm @12v and 10Kg/cm @24v which is the one i'll be using.

You can easily distinguish the version by looking at the RPMs your motor produces are rated at 6000@12v while the Version I'll be using is rated at 3000@12V and i'll be running at 24V


tomorrow i'll post (since it's messy and in portuguese) the spreadsheet i've been using to calculate the torque so you can look at it and you can show me where you think i'm wrong

Cheers

​

Anyway today i finalized the pulley and shaft decision going with Gt2 at 10mm width which is extremely overkill and the pulleys will go from 16T connected directly to the motor shaft going to a 80t pulley on a 8mm shaft of 1045 steel and on the other end of the same shaft a 20t pulley that will go to another 80T on a 20mm 1045 shaft that will be connected to the wheel at all points where the pulley grub screw sets there will be flat spots for no chances slippage on the shaft

adding all these the final price is sitting at $68 shipped to me so still very budget =D next up are bearings and a PSU which certainly shouldn't raise the price not more than 10 dollars after that i'll start working on the Cad model

​

Mein Gott. Pal, you don't understand that the NM of a base are not measured in the wheel radius. If so, it should be recalculated and it works great. You've got your presume 2.2nm on the wheel, where you hold it. There is a 1meter level, the force should be measured by. If not, recalculated by means of reductions. If that example will hold a bottle in 1 meter level, it will be that figures, but it's not.

And that's only the force applied and that a base can only hold on a 15cm level. The base NM overall is less)

Torque=Force * Distance from where the force is applied to the center
so let's get our variables from the example
Force = bottle weight in newtons which converting is (1.5*9.807)
Distance = steering wheel Radius since the force is only applied in one side = 150mm or 0.15M
calculating...
(1.5*9.807)N*0.15M=2.2NM

Okay first i gotta ask Where did you get this 1 meter lever test Standard because i couldn't find any standard testing procedure for force feedback wheels? because if that's a norm that defines how the wheelbase torque should be rated i'll gladly correct my figure's to it.

But i've read the standards for electrical motors and servos torque ratings Which are also the ones used to rate any device that uses those motors as a power source and both IEEE Std 112-2004
https://engineering.purdue.edu/~dionysis/EE452/Lab12/IEEEstd_112.pdf and SAE J2907 https://www.energy.gov/sites/prod/files/2016/06/f33/vs144_miller_2016_o_web.pdf say that the output should be measured at the Shaft which there's no point in rating something that uses a reduction method to increase torque to be measured at the input instead of the output.

and yes if you test it with a 1 meter lever it wont hold the bottle but thats because you increased the lever thus increasing the torque. but you can still measure it with 1 meter lever and realize that it would use a 224 gram weight to stall the motor but the torque on the shaft would still be 2.2NM and if it hold's a 1.5Kg bottle on a 1 meter lever it would result in a torque of 147.1 N/M so the size of the lever does not change the torque rating.

so yeah all those wheels would react exactly the same with a 1 meter lever since the lever length does not affect the final shaft torque

also for the last time i'll give you an example if you take the Mige 130ST-M10010 which is used on the OSW wheels you can see that it has a stall torque of 10Nm and a maximum speed of 1000rpm and if you get this small dc servo from hobby king that's rated at 1Nm https://hobbyking.com/en_us/hxt-10kg-servo-metal-gear-10kg-0-16sec-55g.html?___store=en_us but is at a fraction of the miges Size an produces 1/10 of its torque why SIMPLE because it's torque is rated after a series of gear reductions and it has only 68 rpm which means it trades it's rpm for torque.



So if you don't present me with a valid Norm that defines the standard for Force feedback wheels i'll keep using the Standard rating for electrical motors an their torque ratings which as i presented is correct and gives accurate measures

You missed one zero in your formula. It should be 0.015 not 0.15 then. And 0.2 nm of torque for that example.

You messed up everything or do it on purpose. Superb, if the lever arm doesn't work))

147 N/m it's a way far a car engine already Good luck with RS motors to get 147 NM)

MY1025 motor.
7.5cm lever.~ 8kg with not a full force and PSU with 12v on a 24 volt motor. 1:3.6 gear.
7.5cm lever - it's 13.3 times less then meter so = 600 g.m = 5.88399 N.m (newton meter)



730 g.m = 7.1588545 N.m (newton meter) on almost full power settings but still with 12 volts but now with 1 meter lever. In EMC Gain Constant Spring 100%. And I don't use a scale bar there to double forces that EMS has also. But whatever.





http://convert-units.info/torque/gram-force-meter/730
https://en.wikipedia.org/wiki/Newton-metre


So good luck with 147 NM with RS motors again)

And compare the rotors at least:



And know something about the loses with RS motors too)

I can't see why the RS-895 dc motor could not be used for at geared FFB wheel, it looks stronger than the Mabuchi RS-550 whicj h is used in the Fanatec GT3 and delivers 2.6 Nm there.

I bet it was not discussion about wheater it can or can'not be used) Don't you find it to be so?)

I could look like it
Judging from this sentence: "This motors will be lost in the pulleys. You will get in better way up till some couple of Nm. Even if."

But I could be misinterpreting it.

I would think that most users would be cool with this motor and let's say 1:10 reduction.

But what numbers did you use?

You should use stall torque right? And that number is not available.

Anyway please share your calculations it's the easiest way for others to learn.

Would anybody happen to know the gearing of the commercial it much be something like 1:20?

Nothing contradicting to 2.2. vs couple of Nm there again. It was meant not to 2.2. nm but for those 14 and then 147 nm the topic starter, calculated over here.