What could possibly go wrong :)

Hi All.
As you can see, i signed up here in 2016 with the intension of making a 2dof to move my diy race rig. Unfortunatley, life gets in the way and they make me go to work every day too so i struggle to get time, but then, dont we all.
Its only been 9 years so i figured i should get on with it, lol.

What i built some time ago was a steel tubed frame. It ran 3 screens, a standard g25 wheel, gears and pedals that must be 20 years old now. The h pattern gearbox finally wore out and failed and the wheel died just about a year ago. The pedals are still going , just. It all got alot of use and abuse over the years so fair play to logitech. Theres alot of real cars that dont last that long, lol.
A little while back i was here for info on a hoverboard dd wheel which now works and has been well tested. Im amazed how good it is really and wished i had made it ages ago. Well done to folks that unknowingly helped me there too.

So the reason for this thread... theres so much 'how to' on this site that its hard not to get stuff working if you look and read abit. But sometimes you just cant find what you are looking for and theres nothing in the search results, or at least, nothing that makes sense to me, lol. I thought i'd show what im doing/ making and ask my stupid questions here while im at it.

I was given 2 hefty wheelschair motors, i believe 300w, with thier gearbox's still on.
I plumped for the MDDS 30 Cytron driver, mainly for the cost.
I have a Keyestudio 'uno' r3, currently loaded with Blame73's modded software.
10k rotary pots
2 x 30 amp 24v psu's

I have a load of 40x20x2 steel box and 35years of experience of welding stuff together.
A friend with an mot garage handed me a propshaft with a UJ on the end that was going in his skip (probably from a Ford Transit)

Something like this is my plan... roughly
https://motionforsimulators.com/2dofs3dofmotionplatforms/

Then i'll mount my whole rig to the top and see what happens, lol.

What could possibly go wrong???

Here's the rig with the old g25 and minus the cables. Excuse the dodgy conservatory, Id' just move in. Its fabbed from steel tube and made with motion in mind so its very strong with no flex that you can tell. Without the seat and screens, it weighs very little and you can shift it on your own easily. But... it is big. It takes up the room of a single bed very easily.

Here's a few of the Hover board wheel.

The odrive in this pic died. I dont know why. It worked well for ages then just went all mad. I figured it was the odrive or the encoder so got new of each. I tried the Odesc v4.2 driver first and that has not gone bad yet so happy days.

So, onto the motion side. What could go wrong?
Well, firstly, the UJ didnt move far enough. I guess it didnt need much angle when driving the Transit but it touched its own 'arms' and locked quite quickly. I didnt have any other options at that time, so i cut the 'muvva' up I welded it to some box and plate and now it has a minimum of 45 degrees in all directions. Im sure i wont need that much and it will hit the motors at the moment, but i can sort that with some motor protection bars later.


Next issue, it became apparent that the rig would be far to high and would almost need a step to get in over the side rails of the frame so i had to lower the UJ height.


I settled on this...

It was at this point i couldnt help clamping the whole rig together. I marked the balance point and two of us lifted the rig on and clamped. One of the lads got in and i acted as the motors and moved him about.
"this will be f*%*^n' awsome" was his verdict, lol. For me, i felt it was quite easy to move, even from full dipped 'braking' back up to level. All good.

Wow, that's some quality work on that frame. Looks like you could throw an engine and axles on it; then take it to the dunes.

Next, i mounted both psu's and fixed the cytron to the ply on the rig. Then it was wireing time and also so me temp mounts for the pots. . The smc3 App was loaded and all my dodgy wireing was fixed, lol. I thought wireing it up was quite easy and it only took about half an hour (allthough it is only temporary wiring).
So far so good and i was only about 8hrs into it, but then i came to a halt.. This happened, lol.

https://www.xsimulator.net/community/threads/unable-to-extract-repo-zip.19199/

Since then, ive changed the old Uno to the rev3 copy, redone the smc3 app part and then still failed to get motion from the interface plugin test button/ slider.
The 500000 bits per second through me as apparently that is not a baud rate? I changed from 115200 to 5000000. Also i tried a few varieties of the Interface Output syntax and hey.... one motor worked with the slider, yay
To check the wireing hadnt fallen out for the 2nd motor, i reloaded the smc3 app and all was good as it should be. . I closed the app and checked it was closed and reopened simtools 3 and..... nothing. Not even the first motor worked without changeing anything so i got fed up and went home.

Heres the settings, does the syntax look right to make two motors move?

Cheers Thick8, yeah weve said the same thing. I do have an old Fireblade engine somewhere, lol.

Nice write up!

I’ll enjoy following this project!

Hey Chris, Long time no chat, allthough i did keep up with your bike rig, nice work man This is my learning curve before i carry on with my bike rig next.

Hi Noorbeast. Sorry its taken a few days but everything is at my works.
So how about this.... i set up the pc and the motion rig to get the SMC3 app images you asked for. I took the pics and then tried Simtools 3 on the offchance. I changed nothing, no settings, no wiring, i hadnt physically moved it, nothing. But now it works as expected ???????? It even moved kinda properly with a quick LFS test. wtf? lol. How can that be? Ghosts in the machine i guess.

Anyways, i'll see what i can get done this week after work and report back but it might be another few days.

Onwards...

Slider.

So today i did a torque and weld test. Unitentionally of coarse.

Theres a lesson here people, lol. I checked everything still ran as it should and my ghosts seem to have been exocised so after fiddling with the limit values i decided to make the link arms. I have some nice sturdy m10 rose joints that screw into what was the wheel bolts of the wheelchair. I realise i might need some brackets to get move movment but hey, im still working it out as i go, lol.
Haveing hand tightened a rose joint into each motor i though i should really turn everythig off first.

Here cometh the lesson..... Do not pull the pc plug out when you meant to pull the motor psu's plug
The screen went black and the motors decided to spin up. Now, with the rose joints attatched, there was not enough clearence at the mounting bracket. It didnt seem to bother the motors though as they ploughed on and bent the 3mm plates that were holding them. Torque test successfull i'd say.
You can see where the rose joint first touched at the edge of the bracket. Im quite impressed, if not a little scared of them now, lol.

I think a little redesign might happen now and also wire up a proper power switch too.
Ya gotta love the fun of failure

Slider.

I havnt had alot of time but i have had a go here and there.....
I cut of the bent brackets, straightened them and glued them back on with the welder. I also made some arms from 5mm plate and bolted them in.
It became apparent that things were going to touch each other so i redrilled the motor mounts to go back a little but it still wasnt enough enough so i cobbled some long bolts and spacers to hold the new arms further back and miss the lower frame rail. When it works properly, I'll cut the rail and redesign it so everything has proper clearence. I also had to knock up a better bracket for the pots so they didnt get knocked off. I hooked it all up and did a lap or two of the Red Bull Ring in the Ktm Xbow, in Assetto Corsa. Everything worked as planned.... Happy days.

The next day, i clamped the sim rig onto the motion rig and pugged in the wheel, pedals etc, with the expectaion of actually driveing the Red Bull Ring. Ummmm, no.

So the motors have mechanical locks on them so you can release them and push the wheelchair. Im guessing the two smaller wires to the motor release these locks when its in 'drive' mode so it can move, and lock so it doesnt roll away and throw some poor disabled person under a bus when not moving under power.

Does anyone make use of those locks via the wires??

Anyway, simrig mounted and level.. Smc3 app to check the pots and open up Simtools and released the locks, aaaaaand..... the front plummeted to the floor, then, it kinda reared up like a crazy horse and as i was still trying to hold it from going forward, it threw me onto the floor as well, lol. Luckily, nobody saw it . It was then i was very thankfull that i had fitted a good, easy access mains switch.

After brushing off the workshop floor dust, i saw it had somehow bent both pots so i replaced them and set up again. I realised that my marked balance point on the simrig included a human to balance out the screens, doh!
As there was no crash test dummies left at work I got fed up and went home, then went out and drank beer.
5 empty glasses told me that at least i now know the motors have plenty of power in them

Tonight, i moved the simrig back to what felt like a better balance. I also noticed massive play on the right motor and found the woodruff key and slot had worn. The key was actually twisting a little allowing about 25mm of travel at the end or the arm. A little tig weld on the key sorted that out so its nice and tight now.
Ive aso got a little issue with my rubber tube pot connectors, the right one keeps loosing the centre every time it goes far forward on the surge. It seems fine in the middle/ level. I'll look into that but im not too concerned at the moment.

The bigger isue is that the whole thing wont sit still. There is a little flex in the frame and the motor mount but not much.
In smc3, if i have the pwm at much more than 40, i get alot of over run. Its like an 80's robot arm. It gets to position and then goes back and forth about 3-4 times before it stops.

What causes this and what pwm settings should i expect to use, roughly?

I see folks using pwm up in the hundreds but if i do that, the motor never settles. Its like it cant stop at the exact point it thinks it should be at. It does this with no simrig intop so no weight enertia. Maybe it is my pot connectors?

I'll sort those next and see from there.... Onwards


Slider.

Hi All, ive been busy with life again but managed abit more on the motion.
Ive repositioned the motors so ive lost all th spacer type bits i had before, yay. Ive removed all the slop from the system, remounted new pot's to be exactly in line and solid.

I still cant get the thing to sit still with any weight on it. With just me leaning on the motion bed, it still rocks backward and forward.

I like to work through these things so i know how it works more intamatley so that said, heres the questions...

1. When the sim car is not moving, i assume the motors hold everything nice and tight? ( i seem to be able to move the bed)
2. What is the most common cause of this rocking oscilation?


Cheers, S.

Those are brakes on the rear of the motor. They are normally applied when the wheel chair is not in motion, then released when in motion. Those are not wormgear type motors (yeah for that). so they may move with enough weight on them because the ratio is pretty low. I have a pair that are 17:1. I can rotate the motor using a Crecent wrench (adjustable spanner) on the output shaft. You could probably hook up the brakes to lock when there is not a motor movement output. Then program a couple of output pins on your Arduino based from the motor input. Basically you'd hook them up like you did your motors. Then just add them in the code and recompile it. Something like if the motor output is greater than zero, then release.

Ok Thick8, nice one. I did consider your plan before but figured the motors would hold everything steady so didnt bother. Now i know differently, i'll look into it.

Cheers. S.

Backlash.... thats the issue im having. I played with the locks but it didnt help but i'll still have them on a switch on the 'dash' i think. I read somewhere here that a deadzone of more than 2 means you have too much play in the mechanisms so i never bothered turning it up past that. Yesterday, I tried a random 7 on the deadzone and it almost fixed my problems enough to do a few laps tonight so its promising now, allthough it was abit like riding a lurching camel, lol.

Sadly i dont think there is any way of stopping the gearbox backlash. What do you folks think? Any ideas?

Cheers, Slider.

Hi Noorbeast, i reckon i have a little more movement than the video shows with 75mm arms.
Ive also noticed there is actually some flex in the frame that hold the motors too. It was only really a basic frame idea so i was expecting to be adding more box to it anyway. I still have to properly mount the rig to the motion yet so i'll make it adjustable so i can play with the balance point.

So i re-read the PID explination and added a heap of Kd to the mix on the back of that. Its way better now and actually useable. Another great help was the 'profile intensity' in simtools was at 100 %. I lowered it to half and found it a little too sluggish so around 70% has made a big difference as its way less violent. I did not really notice the slider before as i was concentrating on the axis assignments. I can drive it now without getting whiplash, lol.

Im much happier now as it seems it just wants more tuning and more rigidity but i ran out of time. I'll see what i can get done