2DOF Build log - Status: Construction Begun

So this project has not seen any light for a LONG time but it is still in the works. I have been distracted with a heap of other more short-term projects but decided it is time to move forward on this.

One of the short projects was to build the wheel itself. This ended up being the AMG wheel by AMStudio. It turned out really well. The main difference here is that mine is made from real Carbon fibre and I have made a 2.4Ghz wireless version - No cables. Runs off an 18650 cell built in and charges via usb. The battery will last about 40 hours so I don't think I will have too many issues with it going flat mid-race. Just needs the badge put on and some stickers.

I stalled on this also because I was unsure of how to do the pedal/wheel mount. As is, the rails are too high - I need to offset these down (and probably out) and was not too sure on the best way to do it where it is still presentable/symmetrical. Open to ideas....

I picked up an Occulus Rift S too - Not sure how this will go but hoping it will work ok with the sim.

I spent some time over the weekend making up the linkages for the motors to the platform. They are 'ATV' steering links cut down and re-tapped to fit. I also welded the arms to the axle hubs and have it all ready to go bar some spacers to be able to hold it all on securely. I'll probably also use some shaft-lock to assist with keeping it all still.

I'm currently working on a better setup for holding the hall sensors but that should not take too much effort and then I should be ready to move!



Still on the agenda:
Finish the main frame
Mount wheel
Make pedals (started, 20%) Hopefully hydraulic brake, hall sensors for other 2 pedals but tricky with 3.3v from the STM32 or need a second 5v controller
Mod one of my HP power supplies to float the negative terminal so I can run 2 in parallel for 24v (Main motors) 2000w of goodness
Repair or remake my shifter - Low priority, probably going with remake as a smaller/lighter version of the one I made 10+ years ago...
General wiring, tuning, testing etc.

Well the long weekend has given me some more time to put into the sim.

I have now lowered the seat base frame extensions enough to feel comfortable. I have added a wheel mount post and mounted up the DD (MMOS OSW) wheel.

I finally printed the badge mount and put the Merc badge on the wheel too. The wires hanging down are for the pedals to link up to.

To help with progress, I ended up buying some Thrustmaster pedals and should get them this week. I'll see how they go, possibly with a load cell mod they will be ok but i might still make my own later.

I've also raised the base for three reasons. 1 - So that I have room for the power supplies underneath. 2 - To give me more ground clearance at the front for more angle and 3 - To increase the footprint some more for added stability.

The biggest issue I can see myself having at the moment is the fact that there is some play in one of the gears in the motors. The final gear is keyed to the output shaft and there is some slop. To counter this, I will probably have to clean out all the grease and use retaining compound or something like that to prevent the movement. I've had great success with Loctite 680 in the past - That stuff is crazy. I'm just not keen to clean out all the grease and redo it just yet but undoubtedly i will need to I think.

So this weekend bought some elation and some disappointment.

I finalized the wiring for the power supplies, rewired some old DFGT pedals (still waiting on my thrustmasters but couldn't wait) and done some basic testing to ensure the motors were lined up etc.

This was my first ever experience with Simtools - I had never configured it before. I managed to get it configured and done some basic testing with the sim and having the seat move. Lots of tweaking to do but I had a disastrous moment at one stage.

At one stage, I went back into the SMC3 config utility and my sim literally done a backfilip! Scared the absolute crap outta me! Luckily I was not close to it (or on it!) at the time as i could have been seriously hurt.

After restarting my heart and doing some basic sanity checks, I removed the linkages from the platform just in case something was not right still. It all behaved fine??? What the hell. It took me a little time to diagnose and eventually work out what happened.

This could be a combination of me hacking/making my own H Bridge controller or a limitation/error/issue with the SMC3 code but I remembered that when the serial port is opened on the arduino, it does a reset by default. Somehow, when the thing is reset, it activates the motor briefly (at full power of course) so the motors spin for about 0.5 sec which is enough to do a full rotation on my motors. Ouch!

So - I have now removed the capacitor that goes between the DTR line and Reset pin of the arduino - This will prevent this from happening. I'll do some more testing too but I think the best safety here is to ensure the power to the motors is off at all times unless you are needing them on.

Secondly, I've discovered that the play in the motors is just too much. I was hoping it would be ok but there is constant oscillating due to the slack. So the sim is stripped back down and I spent most of today clearing out the copious amounts of grease from each gearbox with degreaser etc. then welded the final drive gear to the bush on the shaft. I then cross drilled the shaft and I have put a hardened steel pin through it with a good dose of shaft lock. This should get rid of the slack.

Not really where i'd hoped to be but its still progress.....

First of all awesome work! I'd love to be able to do something like this, maybe one day...

I have one question about this hub, i'm designing a quick release (similar to the one used in the simucube 2) for my my1020 but i have no idea of its mechanical resistance since i'm new to the 3d printing world. I've seen people using 100% infill and i'm wondering how much torque are you putting through yours, just to have an idea for mine.

The difference with mine is that I have bolts going all the way through to keep the part together, so it's not likely to break. I have no issues with mine but doing it again I probably would have gone with a higher infill. I am using that adapter with the AMStudio quick release (printed 100% infill) and it works fine. I would not recommend putting weight on it as you got into the sim though (you're likely to bend the motor shaft anyway).
https://amstudioprojects.com/product/qr-base-adapter-70mm/
https://amstudioprojects.com/product/quick-release-logitech-adapter-for-amg-diy-steering-wheel/
(No affiliation)

If you're going to do something like the simcube, you would have to do a high, if not 100% infill to maintain strength in the bit that 'wedges' as there is no secondary support.

Very good simple design. What is the thickness of the steel pipe you use? This simple structure requires a thick steel tube to maintain strength?

The upper deck is 25 X 1.5mm steel. It has plenty of strength. The wheel mount is 25X50X2mm - it moves slightly as it has no upper support or gussetts but it is un-noticable when driving.

Very nice project, I want to start too. Do you know the weight of the seat frame? What are the advantages of positioning motors like this?

I'm not sure of the weight of the frame but it's not overly heavy. The big goof that's going to be riding it is much heavier.... As for the motor positioning, there is no advantage. This is just the way I decided to build up my rig. Doing it again, i'd possibly just go for a square frame. This works but it needs to be bolted to the ground as it tips over backwards too easy. TBH, I would not base any of your own design on my rig since almost none of it is 'off the shelf'. The wheelchair motors are VERY strong, about 1.5-2X the size of a 'normal' wheelchair motor. Each weighs about 20kg! They have insane amounts of torque - I can literally stand on the 100mm long 'actuator' and it does not move (I am just shy of 100kg) this is at only about 1/3 the available PWM too.... This is the reason I have not been too conservative with the design and just winged it instead of doing many calculations.

I've done a few more bits and pieces to the rig. The motors were greased back up and reassembled and I managed to get my first ride after some sanity checking of everything. There is still some play in the motors but it is very minimal now. It now shows how much play is in the CV joint though - There is maybe 1mm of vertial play and maybe 1-2 degrees of rotary play but both of these movements are due to the shaft not being solid in the bearing carrier rather than play in the CV joint itself. A quick dab of the welder will sort this out but that means tearing the whole thing down again

I added an eStop button to the rig that immediately cuts power from the server power supplies, it also gives me the ability to isolate the wheel and motion separately which is handy.

The base has had tabs welded to the bottom of the feet for dynabolts and a coat of paint.

I've also acquired a 50" Tv that can be used when not using VR and will double as a viewing screen for others. I made up a portable floor stand for the TV too. Not sure how this will go with motion and no VR but we'll see. I guess I probably don't always want motion when playing anyway.

Lots of tuning to do still - The motors are noisy and tend to vibrate but from what I can tell, this is due to lack of PID tuning.




A vid of my son having a ride on very dialled down settings...

Thank you for your answer, to be honest i'm waiting for a 3d printer and i'm just drawing things in autocad! I draw somethink like what you're using and then tried the simucube way but at 100% infill it will weight around 500g...
Still deciding between a MY1020 800W or 1000W to replace my LY1020 800W which is too weak for some reason, i expect something like 6Nm or more, since you have a 48V i guess you'll get more torque and if this design is working for you it will surely do for me!

I have a 48v motor but I am running it at 12v still but you're going to need a big power supply. These HP ones are good for 80+ Amps at 12v. I am running the feedback at about 40% and after a few mins it hurts my arms. The (virtual) wheel stop is quite solid too. I don't know the actual force but I would estimate it to be over double of the G27 I had.

Yeah, i have a 50A one at the moment but my motor is only absorbing up to 30/32A, i think this is because of its internal resistance. Anyway i got a 106A psu (HP HSTNS-PA01) that i may use for testing but it's way too noisy for domestic use, the 2 fans inside are always spinning at max rpm and it sounds like a plane during take off, so i don't recomend it. Which PSU are you using by the way? Might be useful for other people.

Mine are DPS-800GB which in Australia are rated for 1000W (240v Mains). They are relatively cheap and pretty much silent. I have a couple of the 750s in my 3d printer and these are also almost silent and are only 1U tall, just lower output. I work in IT so these all came free with old HP equipment we were throwing out

I need to grab a clamp meter to test the actual current getting to my motor but I have a couple of good things going - 3 IBT_2 Hbridge modules for the switching and 8GA wire from the power supply to the H Bridges. This reduces voltage drop at high current. I have also reflowed the ICs on the IBT_2 Modules and added heatsink pad to the board to increase thermal efficiancy. The fan draws air over the coolers and even after a hard session they are barely warm.

Another day down with some more work on the sim. I welded the CV joint shaft to the carrier today and this has tightened up the joint well.

Over the past few days I have also been working on the modded T3PA pedals. I hated the brake on this, too sloppy and I am used to a load cell brake. But I aint paying $100+ for a load cell mod

I adapted parts I had on hand from previous projects. I used an INA826 instrument amplifier and a 30kg load cell. Adjustable gain and '0' pots available via holes next to the pedal...

hi,
are you using arduino nano or stm32?

Stm32, specifically the diymore board because it is small.

STM32F103?
MMOS FFB?
I use stm32f4xx with MMOSffb, I also have some STM32F103 that arrived this week, but I haven't tested it yet.
I am looking to test the EMC, but it has not yet migrated to STM32F4xx, and the stm32f103 I have is bluepill, and emc only works on blackpill

I posted the link but it didnt work.

https://www.diymore.cc/collections/development-board/stm32

STM32F407VGT6

EMC will not be ported, or you'll probably die waiting. Its not open source and the dev just abandoned the project. There are differing opinions on what works better between EMC and MMOS, thats just the path i went down. I tried both but only in testing on the motor with EMC, not as a wheel/ingame.

Also, EMC does not run on blackpill either afaik. Only arduino leonardo (or other atmega32u4 board*)

*Not all 32u4 boards have the required pins on a header though. Pro mini is an example, but if you're Really good with a soldering iron, you can still solder to the chip directly.

it works with blackpill, I follow ebolz, even he said he never tested with bluepill. told me that soon he would have ported the emc to stm32f4