DIY 3 DOF pegleg first motion rig

Finally pulled the trigger and started my first motion rig build, i've settled on a 4 actuator pegleg design. Will probably add traction loss later.

The actuators are made of:
Aluminium plates
4040 extrusion
HDPE sliders(high density polyethylene)
1204 ballscrew
ZY1020 24V DC motor

I cut the aluminium plates on my small 6040 CNC router, this is old machine guarding i was given 20y ago.



The plates were cleaned up in a tumbler.


Trial build of the first actuator with 3d printed sliding parts, these were later replaced for HDPE parts

Cutting the HDPE


A set of parts for 1 actuator


All 4 actuators assembled


Test setup for a single actuator

Testing with my old rig as a load. I tested with upto 130KG (me sat on the rig), The IBT2 started to struggle and cut out after around 10 senconds of running the SMC3 sinewave with me on the rig.
I don't expect this to be an issue in the final build as the load will be shared between actuators and i may use 2x IBT2 per actuator to make sure.



Next i need to finish the motor mounting of the other 3 actuators and start on the control box.

Steve

nice project broo...

how u attach the sensor/potentiometer..??

u can give the ibt2 little bit of thermal paste, but dont use liquid metal based, just use cheap one, that will help a lot

and set the voltage little bit down, i usually use around 22-23v for safe use

Good point with the thermal paste.

Position sensor is a 100mm stroke slide(Fader) potentiometer, 100mm stroke is the longest i,ve managed to find so far which will limit usable stroke to 85-90mm. (actuators would have 125mm stroke)
Pot is mounted to front of actuator with 3d printed brackets, the bracket that connects to the moving part is going to be re-designed as it wobbles


4 actuators now complete except for mounting of pots on 3 of them and i need to design and print end travel bumpers and feet.


I still need to order parts for the control box and figure that out, the test in my first post was at 12V, powered by a HP server power supply. Hoping i can get away with 1x 12V power supply per actuator.
Steve

Lots of progress and issues since my last update.

I printed some TPU bumpers 3 times as i modelled them without measuring first(for some reason no photo of them). These are so when the rig sinks under gravity without the actuators powered it will rest on the bumpers and not destroy the potentiometers.


I also reprinted the potentiometer brackets twice as thick as they sere distorting and printed the arms that join the potentiometer to the sliders out of TPU so if i do get an overtravel situation they should flex.


Added aluminium box section between the motors and the top plate to make the motor mounting rigid, the box section is not sttached in any way. Just clamped between the motor face and top plate.

From the start i've been unsure if the server power supplies would be enough without a battery, to assist with this i installed a 10000uF for each motor with a simple precharge circuit which gets bypassed by a relay when the capacitor reached 7V (ish).


Many issues with the IBT-2's
It turns out the one i used for initial testing is faulty and heats up even with virtually no load on it. This one may also be responsible for destroying the one it was in parallel with or that one may of just been a dud.
Another has a faulty screw terminal so i need to desolder a terminal from one of the duds to try and save it.


I finished wiring all these and with 4 working IBT-2's was able to test my 2 channel controller, the plan is to make 2 of these, one front and one rear each with a 750W power supply for each motor.

Testing didn't go well, the 3 core wire i had used for the pots and run in the same socking as the motor cables was picking up a lot of interference from the motor wires, every time a motor changed direction i was seeing big spikes in the pot reading within SMC3.
Also had connection issues with the 3.5mm headphone connectors i had used for the pot wiring.
I found some USB A and micro USB breakout boards leftover from another project so decided to use these and standard usb cables for the pot wiring. these solved the interference issues.

Now running the various tests within SMC3 the IBT-2's remain cool, even when using 1 per motor.
This seemed a little odd so i fitted a power meter between the power supply and controller. A single actuator running the square wave test with most of my weight on it was generally drawing 3A with a peak of 7A when changing direction. These motors are rated at over 26A (@24V) so might be excessivly oversized.

I'm now thinking of 2 power supplies in series to give 24V and am undecided on 1 or 2 IBT-2 per motor.

Actuators now fitted to my rig, it's a little nose heavy with only the weight of the rear actuators holding the tail down. Hopefully this won't be an issue in use otherwire i'll nee to move them a long way forward to to between the pedals and wheelbase support.

Still need to remove the existing legs and fit the cross braces between the actuators.
The rubber feet are a bit soft, these may need replacing with something harder and something to spread the load on the floor.
Built the second controller, this time with a single 24V input. Exactly the same as the first except for the common input and capacitor precharge circuit modified for 24v.


Another Faulty H-Bridge on this controller, i'll try running 1 per actuator for now.
I need to modify the first controller for 24V and hopefully get it all powered up tonight.

Both controllers now in a working state and running off a pair of 750W HP server power supplies to gige me a total of 62.5A @ 24V.


Installed and configured simtools, first actuator test was fine. Then the front left actuator seemed to get higher and higher and not really go down again. I need to look into this and see what's going on.
The rubber feet are way too soft and are what's causing all the bouncing, the actuators and bouncing are just the same with me say on the rig.
Video shows testing with simtools, heave first then pitch. difficult to see the pitch with the one actuator not responding properly.

Will see what's wrong with that actuator tomorrow night and try printing some harder feet out of TPU.

I think the actuator not going down is another faulty IBT-2 H-Bridge. Will check tonight.

Better balanced when loaded but still have a feeling it may topple with a fast pitch forward.

My issue is the the front actuators would need to move a long way forward to clear the steering gantry and then I wouldn't be able to run a cross beam between the two fronts as the heal plate and bass shaker are in the way. It would also take a lot of actuator stroke to get much pitch.

More issues off the list for now.
The squishy bouncy feet are gone, replaced with TPU printed feet.

The front left actuator not lowering was another faulty H-Bridge and i later had the same issue with the rear left.
I have some suspicions the way SMC 3 is controlling IBT-2 H-bridges (Mode 2) is responsible the these failures.
If i'm right, it's probably fine with small motors but bigger motors will damage the H-Bridges.

Anyway onto the good stuff, i got it working with simtools. Manual testing anyway.

Unladen testing:
Still nose heavy without a driver.

Testing with me in the rig:

I feel that your four motors are out of sync,. Did you find anything?

nice progress, how much power draw when it stand still with you on it..??

maybe you can monitor the power draw when you sit on it, measure each corner, or at least measure front and back. with that you can assume the balance is right...

I've done very little output tuning, in reality these motors are far from ideal. They don't move until PWM gets to between 50 and 60 then move quite quickly with a lot of inertia. P gain is set to 110 with 0 for I and D , much higher on any of the gains results in in the motor surging and stopping when following a waveform. Lower P gain leaves a huge lag between target and position.

Any power measurement when standing still will be a random number.
The ballscrews are 4MM pitch and with the cogging of these motors the rig doesn't always drop when powered down. If it does then it's the rear that drops and i have to lean forward for the front to drop.

Does anybody know if in the SMC3 GUI i set motors 1 and 2 to sinewave, will they be syncronised?

Below image is how the motors are connected as per SMC 3 thread running Mode 2.


Pins 1 and 2 on each H-bridge are to drive the motor in either direction, 3 and 4 put the H-bridge into standby/sleep with the 7 and 8 for +5 and gnd.

Notice how pin 2 on each H-bridge is connected to a PWM capable pin (9 or 10) on the arduino but pin 1 on the H-bridges is not.
PWM for that direction is presumably controlled by PWM firing pins 3 and 4 ( on the H-bridge) using arduino pin 3 or 5 which are PWM capable. Pin 1 wounl be held high whilst driving in this direction.

The reason i think this is bad is that the data sheet it mentions "PWM capability of up to 25 kHz combined with active freewheeling", i may be completly wrong with this but active freeweeling sugests the device needs to be enabled for this to work and not in sleep mode?

I've had 3 (or maybe 4, lost count) IBT-2's fail in the same way, they will lift the rig but not lower it, the first failed whilst the actuator was sat unloaded on the bench so it's nothing to do with the mass or inertia of the rig
I think it's that in 1 direction the the H-bridge is being drived correctly and in the other it is being put to sleep between each pulse and thereby has no active flywheeling in that direction.


Datasheet here: https://www.digikey.co.uk/htmldatasheets/production/70497/0/0/1/BTS7960.pdf

Been doing some testing of different ideas to drive the H-bridges how i understand they should be driven.
Reflashed the uno to mode1 and interfaced to the H-bridges using optocouplers as AND gates, this worked in that it drove the H-bridges but for some reason the uno froze after powering the motors for 20 seconds or so with either the gui or simtools.
For the short time it worked it was much quieter and smoother than running mode 2. restarting the uno made it work for another 20 seconds or so, no damage done just something upsetting the uno.

I did a similar test using mode1 and a ULN2004 stepper driver to convert the signals(all bits i had leftover) same results, the motor moved in a quiet controlled manor then the uno froze after maybe 10 seconds.

So whilst it feels like i made no progres, i'm more convinced than before that smc3 mode2 is an issue with big motors at least.

I have some more optocouplers ariving tomorrow and think with a bit more thought i can make it work.

More testing with SMC3 in Mode1 and using optocouplers to interface with the IBT2.
I extended the wires to the H-bridges so the arduino could be mounted away from all the high power wiring as the arduino was only freezing with power on. Still no good, then I tried powering the arduino from a sub c port and it worked!

Completely stable in SMC3 and simtools, smooth motion and the noise from the motors is gone. Only done this to the front two actuators so far so plenty more soldering to do.

Power with me sat on rig stationary is too low to use for balancing the rig.
1.1A 24.7V 27W total for all the actuators.


Actuator positions are not finalised so balance will change anyway but for now there is more weight on the rear with me on the rig.

The pile of dead IBT-2's is getting bigger!


I have now got everything working with SMC3 running Mode1 and converting the signals with optocouplers. Need to sketch out what i've done and post it here.
2 optocouplers per motor output.


SMC3 settings It's hard to tune a 4 actuator rig because as far as i can tell SMC3 only lets you move 1 axis at a time so that 1 actuator lifts 2 others off the ground when in motion.


Simtools config i used to test with Live for speed.


Working in Live for speed and yes i know my driving is shocking in the video.
Excuses: This is my 4rd ever lap and second with wheel on this sim.
I didn't enable the force feedback until a few laps in as the wheel went nuts ripping out a usb cable on my first attempt.
Don't know the track, is it real or fictional?
Don't know the car.
Usually dim the lights when driving so i can see the screen.
Steering was not centered.
Had auto shift enabled and was shifting manually.



The above 2 runs were done back to back, a minuite or so after i checked th etemperature of the H-bridges, 3 were cool to the touch. the rear right was hotter to the point where i couldn't keep my finger on for long so probably 65C. Either there is more weight on this actuator or this H-bridge is failing.
This is with no fan cooling them.

Peak power during these 2 runs was 454W (18.5A) total for all actuators.

Big thanks to SimTools for my DIY License.
I got the rig working in iRacing last night, felt more like a roller coaster than a GT3 car.

Having read some of the manual and changed the assignments from pitch and roll to surge and sway. Manually tuned the min/max in tuning center.
It feels remotely like a car, a 90's car with soft wallowing suspension but a big improvement.

Had one weird issue last night, exited iRacing to change some settings and the rear 2 actuators started dropping then going back up every 10 seconds. Re launching iRacing didn't fix it but a reboot of the pc did.

Still lots of tuning to do and some actuator repairs/modifications.

Thanks noorbeast, will take a look.


Modified the SMC3 sketch to show how i wired the IBT-2 h-bridges to work in Mode 1