My 3 DOF with heave

I am in the process of building my own 3 DOF motion simulator (roll, pitch and heave) for flying games in VR. I've never build or even used one before.

I have read tons of threads on this forum and have pretty much completed the mechanical design in CAD. In the upcoming weeks, I want to start ordering parts.

I've read about VR motion rigs not needing a lot of angular displacement, but I'm too stubborn to actually follow that suggestion. I assume that more roll and pitch angle equals more fun. So I'm aiming for +/-20 degrees for pitch and roll. I can tune the angles down via software later, anyways.

I want not only the chair to move, but also the pedals, flight stick and throttle. With the pedals sticking out from the seat quite low and far to the front, a high amount of negative pitch (going into a dive) means the center of rotation of the seat needs to be pretty high above the ground (or else the pedals would crash into the ground). Avoiding collisions between the footrest and the levers connecting the motor arms with the top platform was also tricky.

Check out the picture attached to this posting.

I'll be glad to discuss the mechanical details of my upcoming build with you all later, but for now I have some specific questions regarding PSUs and Motor driver.

I plan on using 500w 24V motors with a rated current of 25 A. Reading a lot of build threads on this forum made me come to the conclusion that it can avoid trouble to have PSUs with twice the power of the motor it's going to drive. So we are talking about about a 1000W PSU for each motor. Also, I've read that modded server PSUs don't like regenerative current from motors, so I'm currently looking at one Meanwell RSP-1000-24 for each motor. I don't want to use batteries in parallel to the PSUs for a more compact and simple setup. For the driver I'm looking at Sabertooth 2x32, mostly because they seem very reliable and well documented. I also like the option to dump regenerative current through resistors.

So I plan on using oversized quality PSUs and dump the regenerative current through resistors... All to avoid shut downs because of power spikes or regenerative current.
Am I over-doing it? Could I drop one or two of these measures?

Second question: I was planning on driving the three motors using two Sabertooth 2x32, so one unit will drive two motors while the second unit drives motor number 3. This means I would need to provide 2000W to one of the Sabertooths. But i don't like the idea of buying one single 2000w PSU. It costs about the same as two 1000w PSU, but if I accidentally fry it, twice the money is wasted on a single incident. On the other hand, wiring two 1000w PSUs in parallel only works good in theory. Of course I could always go with 3 motors, 3 1000w PSUs and 3 Sabertooths... But this seems like a waste. Any recommendations?

Are there any single channel drivers worth recommending? Do they also offer regenerative current dumping? If not, will an oversized quality PSU be enough to handle that current? I want to avoid big batteries or capacitors.

Looking forward to your advice, thoughts and suggestions.

Best regards,
Sebas3

Well, the 2000W PSU would be connected to the Sabertooth that drives two 500W motors. So it's really more like 1000W PSU for 500 W motor.

How do other drivers than Sabertooth 2x32 handle regenerative currents? I have the impression that a lot if not most of them don't have the option to dump those currents through a resistor. Also, I am under the impression that there are lots of rigs without batteries to handle regenerative currents. So how do these work? Are regenerative currents not that critical after all?

After doing some further digging through the threads on this forum came up with the following plan:

One Meanwell RSP-1000-24 (40A) PSU will feed into one Cytron MD30C (30A continuous, 80A max) which will drive one Motor (25A rated).

To handle regenerative currents, I will put one Pololu 26,4V 2,8 Ohm 15W shunt regulator in parallel to the PSU. One forum user fried his shunt after a short while and then successfully used two shunts with lots of heat sinks and airflow. I will try one single shunt with heat sinks and some airflow. We'll see how far that will take me. After all, I don't expect much regenerative current, as the 50:1 worm drive motors can hardly be back driven.

The position of the motor shaft will be measured by 6127V1A360L.5FS hall pots. I use the FS (=flat shaft) variant to make it easier to mount a 3d printed GT2 belt pulley.
I went for the 360 degree continuous rotation variant to avoid breaking the hall pot if something goes wrong.

I came up with a 3d printed belt tensioning mechanism which I will show later.

I am building a 3DOF rig so the described hardware will be there 3 times.

All three motor drivers will be controlled by one Arduino Uno R3 running Blame73's version of SMC3.

What do you guys think about my approach? Any obvious flaws or things I overlooked?

Is was thinking about adding fuses. Its tempting to skip those and rely on the over current protection of the PSUs...
What do you guys think about that? If you explicitly recommend using fuses, what specs (ampere and delay time) should I use?

I will also have to look into the topic of emergency shut off switches.

Most of the non-electrical parts arrived and construction has begun.

General construction is based on 4040 aluminium profiles and associated hardware. I neither have welding skills nor the related hardware so I can't really work with steel. I also like that I'm able to put everything together inside my home. Compared to wood, I expect aluminium to be stiffer and that screws don't get worn out as fast when the whole thing is moving.

While ordering the motors I was able to customize the shafts to a D cut shape without additional cost. This allows for a simpler lever design than what a lot of people use here because no strong clamping mechanism is needed. I simply take a block of aluminium and drill a hole with the diameter of the motor shaft into it. Then I add another hole perpendicular to the first one, tap it, and put the screw in that holds it in place. All this can be done in a small workshop without large or expensive tools.

The diameter of the motor shaft is 20 mm so I need to drill a hole that size into the Aluminium blocks that shall become the levers. I have never drilled a hole so big, so I decided to start with a 12 mm hole, then use a 16 mm and then the 20 mm drill. Each drilling pass has roughly 100 mm^2 of area to remove.

I read about the importance of using the correct rpm for size of the drill and the material to be drilled. Unfortunately I just have a simple drill and drill stand with no way of setting a specific rpm. Or do I? I used an app (Spectroid) on my phone to do a spectral analysis of the sound the drill generates. The drill is supposed to have 1200 rpm on full speed and the spectrum showed a peak at about 545 Hz. For drilling a 12 mm hole into aluminium I am supposed to use 800 rpm. This means I need to change the rpm of the drill so that the spectrum shows its peak at about 363 Hz. Easy peasy.

In fact, I was surprised how easy it was to drill the 12 mm holes. I think widening them to 16 and then 20 mm shouldn't be too difficult either.

Today I pretty much completed the top platform of my rig - see the attached picture.

About the motors (which have not arrived yet): I have ordered what seems to me like the 500 W version of the PGSAW crab pull motors that are often used by members of this forum. I found it on Alibaba:
https://x.alibaba.com/AvP1w3?ck=pdp

It doesn't cost much more than the 200 W version (that's commonly used here), so I went with it. Also, you can customize the shaft length before ordering, which I found very useful during the mechanical design phase. You can even order a shaft with D-cut, so no need for huge clamping forces on the lever arms.

Specifications:
Voltage: 24 V
Gear reduction Ratio: 50:1
No Load speed on output shaft: 72 rpm
Rated speed on output shaft: 60 rpm
Rated Torque: 56 Nm
Rated current: 25A

It's a worm drive gearbox with steel worm and bronce gear.

Three of these Motors, each using a CTC of 100mm, should be able to carry a static load of 171 kg.

Static_Load = Torque_per_Motor * Number_of_Motors / CTC = 56*3/0,100 = 1680 N or 171 kg

I've measured my platform at about 35 kg. My own weight is 85 kg, so I'll have a total platform weight of about 120 kg. So the rig should be able to do vertical acceleration (upwards) in the order of 0,4g (meaning my body will experience 1,4g).

Acceleration = effective_Force / Mass = (Force_from_Motors - Force_from_gravity) / mass = (Force_from_Motors - Mass * 9,81 m/s^2) / mass = (1680-120*9,81)/120 = 4,2 m/s^2 or 0,4g

I've read that it's advisable to only use about +/-40 degree of rotation from the motors. That means a CTC of 100 mm gives a vertical range of motion of +/- 64 mm.

heave_distance = CTC * sin(angle) = 0,100 * sin(40) = 0,064 m or 64 mm

The maximum linear speed will be:

Linear_speed = 2 * Pi * CTC * RPM/60 = 2 * 3,14 * 0,100 * 60 / 60 = 0,628 m/s or 628 mm/s

This is quite a lot, more than needed. The upwards acceleration on the other hand is lower than what would like. What do you think?
Of course, I could use a shorter CTC to get more balanced values, but this would also reduce heave travel.
A higher gear ratio would also help, but I'm already at 50:1, which is the maximum that this company is offering for this motor.

I dedigned for 100 mm CTC but have made provisions for switching to shorter CTCs easily.

The motors will be fixed to the frame by Aluminium sheet metal of 6 mm thickness.

My whole construction was influenced by my distrust in my own ability to completely handle the electric and software side of things. So even if the motors start turning 360 degrees, nothing should break.

While most of the construction is Aluminium, there are some M8 screws and M8 threaded bolts and 8 mm rod end bearings involved in the flow of force. I went with M8 because the holes on the front side of the 4040 aluminium profiles can easily be tapped to this thread size.

The hall sensors are driven by GT2 belts. This means I need some kind of belt tensioning system. I came up with the idea to place the pot on top of the motor, where there is a screw hole in the motor housing that was unused till then. So I designed a 3D printed part that holds the pot and can be screwed on top of the motor. The screw hole on the 3D printed part is a slot hole, so while the screw is not tightened, the pot holder can be moved around. This way, the center distance of the belt pulleys can be varied. See attached pic.
The belt drive will cause some load on the hall sensor shaft. Here's to hoping nothing will break.

The pulleys themselves are also 3d printed. Both the motor shaft as well as the hall sensor shaft have a D cut shape, so it's easy to fix the pulleys on the shafts. The pulley on the hall pot will have 25 teeth, while the one on the motor has 75. So when the motor moves 40 degree, the hall pot will move 120 degrees, which might positively affect precision.

After assembling the top platform, before mounting it to the frame (or rather: the motors, which haven't arrived yet), I tuned the position of the center of gravity. At first I determined the position where the CG should be (at 1/3 of the distance between the two front motor connection points and the one in the back). Then, I placed a wooden bar under the top platform where the CG should be and sat into seat (which had its Flight controls already mounted to it). After making some adjustments to where the seat was the top platform, the CG was at the right place.

This one is about electronics and safety.

The PSUs have a over current protection feature kicking in at somewhere between 105 and 125 % of rated load (according to spec sheet), so at about 42 to 50 A. I think Meanwell is a reputable manufacturer (and I ordered via a reputable shop), so I'm trusting that the PSUs can protect themselves.

To protect the drivers (and also the motors) I added fuses, (one fuse per driver). On this forum I read that it can be good practice to start with a low rated fuse and then work your way up if they blow without a good reason. So I ordered 20, 30 and 40 A fuses and see where I'll end up. Reminder: the motors are rated 25 A. At first I wouldn't have thought about ordering 20 A fuses, but then I took a look at their data sheets. I found that they are quite slow, with 135% nominal current (27 A) blowing the fuse with a delay of 1 to 30 minutes and 200% (40 A) causing a delay of 4 s to 1 minute. I think the motors will take less than their rated current most of the time, so 20 A seem to have a chance to work. We'll see.
Oh, by the way, I'm using automotive fuses. Maybe that's why they are so slow. I had trouble finding other fuses than automotive ones that could take currents in the 20...40 A range. Also, don't forget to get proper fuse holders.

I also added a emergency shutoff switch. I didn't like the idea of running 240V AC up to the seat (very dangerous if the cable gets damaged because it gets cought in moving parts). So I'm gonna use 3 relays rated for 24 V coil voltage and 50 A switching current.
The emergency shutoff switch is normally closed (opens of if I push it) and the relays are normally open. So the relays need current running through their coils (about 70 mA each) in order to keep the rig powered up. If the cable to the emergency shutoff switch gets disconnected, the rig will turn off.

The relays are placed between the PSUs and the Drivers/Shunts (which are wired in parallel). So even if I press the emergency shut off switch, the PSUs will still have power. Also, the Shunts can take care of regenerative currents from the drivers.
Also turning off the PSUs by pressing the emergency shut off switch would require either an additional PSU for the relays or to abandon the "normally closed switch and normally open relays" concept I described above. I didn't find any of those options attractive. Also, I have quite some trust in the manufacturer of the PSUs and their protective features so I don't expect emergencies caused by the PSUs.

The electronics are cooled by a 140 mm fan by Noctua that can be driven using the 24 V I've already got in my system.

The shunts, drivers, relays, fuse holders and PSU are held in place by custom 3D printed holders.

The last pic attached to this post shows the insides of the electronics box. The relays are on the left side. Next to it, to the right, you can see a "voltage distributor" where the shunts (next one to the right) and fuses (also next one to the right, with numbers on top) are connected to. The fuses are also connected to the drivers.
In the lower left corner thete is another small "voltage distributor". Here i am using PSU #1 to power the emergency shutoff switch and relay coils as well as the cooling fan.
So about half of the content of the electronics box is related to safety and has little to do with actually driving the rig. Reminds me of the company i work for.

I chose 6 mm^2 (about AWG9) wires for the high current applications. The medium current applications (emergency shut off switch, relay coils, 240V stuff) are wired using 1,5 mm^2 (about AWG15).

After all i found setting up the wiring to be a lenghty and tiresome process. An hour of work had seemingly little to show for, compared to putting together the aluminium profiles. And those thick wires have a will of their own, so its hard to convince them to to anyhting that remotley resembles something like a tight 90 degree turn. I expected to have plenty of space in my electronics box but it turned out to be barely enough.

However, Tests of the electronics where successful. No smoke anywhere and green lights on the drivers.

I still have to wire up the arduino (connect it to the drivers and hall sensors). I will be using 0,14 mm^2 (about AWG25) wires for that.

I sure hope those motors arrive soon because I'm running out of other stuff to work on.

Nice job mate !!

Waiting for motors to arrive.....

@m50b20: sorry, but i dont quite get what you are telling me

The Motors have finally arrived. Yay! They were much bigger than I expected. Mighty. Impressive. Wooaaah. This is gonna be fun.

Of course, the actual measurements were like the drawing said they would be. Just holding them in my hands hits different compared to seeing the CAD assembly on my screen. Nothing beats seeing stuff with my own eyes.

I mounted the motors, added the hall pots and set up the wiring. I took care to keep at least about 5 cm (2 inch) distance between the motor wires and the hall pot wires. This was done to avoid issues caused by interference.
On the other hand, this made the hall pot wires much longer, which might also cause interferences. We'll see how it works out.
I'm really satisfied about how the wiring looks, at least outside of the electronics box.

First manual test with the motors where successful. Pressing the buttons on the drivers makes the motors turn. Playing around with the on board pot varies the motor speed. Still no smoke when I turn stuff on. I do better than i expected.

I adjusted the position of the hall pots relative to the motor positions. Now the normal operating angles of the motors are covered by the hall pot angles where the hall pot's voltage varies continuously. There is a point where the hall pot voltage jumps from about 4,5 V to about 0,2 V. I guess i dont want this happen during normal operation.

I took a close up picture of the hall pot mount that allows belt tensioning. The CAD screenshot I posted a while ago didn't do it justice.

Next: doing all thar arduino / PC / software stuff that (hopefully) will bring it all to life. You know, the stuff that scares me the most...

I think you should change the ratio between the pinion on the motor shaft and the pinion on the Hall sensor, 75:25 is too much, in the motion sim used for flying there is a greater range of motion which will cause the Hall sensor to go out of its measurement range and emergency stop the motor. In my motion sim I use a gear ratio of 20:20 and it is ok.
Your motion sim is very impressive

I have a question related to the engines you bought, can you say something about what is the gear backlash?

Thanks for your compliment and your advice.

I read that the range the motors are used in should not exceed +/-40 degrees, because it gets harder and harder for them to maintain constant linear speed the larger the motor lever angle is.

So when the motor never moves more than 40 degrees in one direction, the hall pot won't move more than 40*75/25=120 degrees in one direction. That leaves 60 degree margin of error per direction. Or in other words: I have a corridor of 120 degrees to position the hall pot correctly. Sounds manageable to me. I'll give it a try. But if I run into trouble because of this i will follow your advice and reduce the gearing ratio.

The backlash of the motors is about 0,5 mm at 100 mm CTC. I have no personal experience with this, but I think that's OK.

Speaking of backlash, I currently have substantial backlash between motor shaft and the belt pulley mounted to it. That means the motor can turn some before its pulley, the belt and the hall pot starts to turn. This backlash is caused by the motor shaft being about 0,1 mm smaller than expected, or my print turning out 0,1 mm larger than it needs to be. I'll print another set of pulleys to fix that issue.

That's the result of me thinking i was smart, when i was just being lazy. I wanted to fix the belt pulley without screws, just by hitting the tolerances perfectly. I failed doing that in the first try so now I have to print the parts again. Will I learn from that? Probably not.

I keep my fingers crossed for the successful launch of the platform.
The gearbox backlash looks very small , in my 250w motors on a 70mm arm it's about 1.5mm .
Your motors look much better made.

Below is a picture of my 2DOF wooden , it looks not very professional but I treat it as a prototype

Looking good, all in black. And the MFDs are awesome!

I refrained from using wood as I was worried that screwing pieces of wood together would not be stiff enough. Especially if I used softwood.
I have good experience with beech wood, but that would save much cost over aluminium profiles while still not offering connections as rigid.

I replaced the pulleys on the motor shaft with new ones that have no play with the shaft.

Today I connected the Arduino to my pc the first time. I flashed SMC3 and SMC3utils connected successfully and I played a little with the PID and PWM settings. Basically I followed this tutorial up until I completed the "initial setup" part.

https://www.xsimulator.net/communit...3dof-motor-driver-and-windows-utilities.4957/

Turns out I'll have to switch the way the wires on the motors are connected to the drivers.

Won't be long till I need Simtools. I already applied for the free DIY license.