New 3DOF rebuild with DIY SFU actuators

Hello to everyone. I'm here again to share my "new" 3DOF motion rig mostly for flight sims. My last thread was the switch from a 2DOF to a 3DOF with heave with 3 motors in star setup, here the link: https://www.xsimulator.net/community/threads/from-2-dof-to-3-dof-flight-sim-motion-rig.18430/

That rig worked fine but the worm gears have a lot of backlash so the small and subtle movements from the telemetry was not transmited to the chair platform because of mostly two things:

1. The mentioned backlash intrinsic to those wormgears.
2. The lack of efficiency intrinsic to those wormgears.

Both of that problems casued lot of energy consumption (and waste of that) and the cogging that let me out of the VR immersion speccially when landing a plane.

Because of that (and because I can't stay still) I started to find a solution. Between them was the install of gasprings/dumpers but it doesn't work as expected. Searching in this forum I discovered the SFU gearbox upgrade from DOF Reality and after explore the design of that GB from pictures I started to design my own GB.

One problem I readed that GB have is the stepping/juddering issue that DOF Reality urges to resolve with a planetary GB (so another layer of complexity, problems and cost). I didnt know if that problem will afect to me because my motors (the same as the previous model) are really powerfull but I didnt want to spend time building something with the fear that it will not work. So the my thougts were use more of the ballscrew instead a planetary GB. In base of my calculations/obersvations the DOF Reallity SFU GB uses around 70 - 80 mm of the ballscrew so if I install a ballscrew longer and I use around 300mm It will multiply the force by almost 4 and the motor will work at more rpm (so more torque).

After some blueprints and prototypes printed in 3D I decided to build this model:

First prototype


Almost final model.

After some touch-ups and reinforcements I builded the final model and started to install them on their emplacement. Here the after and before:


Before

After

After the first tests the result were promising. Now the motors work very very cool and develop a lot of force. After my calculations for this gearbox the motors are working at more or less the 33% of their capacity and it really noticeable after two hours working non stop the motors are still cold (in the previous rig the motors warmed up to a very high temperature relatively fast). And for me the most important is that I can fell all the subtle movements/forces of the plane (specially when I'm gliding to land. Even if you are parked in the middle of the sea with a seaplane you can fell the sea waves and feel as if you are adrift. And I love it how it works now, smooth as silk.

As a mention if the rig is used for simracing the gearboxes are a little bit slow but strong enough as mentioned above but the problem with this design is the long (double) arm with a 5mm platten. Sometimes a violent roll movement can twist a little the rear arm and it needs to be carefull with the "violence" and/or use a 10mm platten instead.

Yes, the footprint isn't compact anymore but I don't care, it was never going to be portable si its ok.

Here are the list of material used for each actuator.

• 750W DC brushed motor (x1)
• 390x30 5mm thik platten for the arms (x2, maybe better use 10mm thick)(attached model).
• 120x81 10mm thick platten CNC machined to interface motor and screwball (x1)(attached model).
• Laser cuted side (x2) (attached model)
• 2005 SFU screwball (x1)
  https://es.aliexpress.com/item/1005..._main.142.2656194dFLBzn4&gatewayAdapt=glo2esp
• Ball screw end bracket FK15 (x1)
  https://es.aliexpress.com/item/4000..._main.157.2656194dFLBzn4&gatewayAdapt=glo2esp
• Zinc alloy ball bearing, block mounted support KFL002 (x2) and KFL08 (x2)
  https://es.aliexpress.com/item/1005..._main.162.2656194dFLBzn4&gatewayAdapt=glo2esp
• Deep groove flange bearing (x4)
  https://es.aliexpress.com/item/1005..._main.102.2656194dFLBzn4&gatewayAdapt=glo2esp
• Aluminum motor separation spacer (x4)
  https://es.aliexpress.com/item/1005...t_main.97.2656194dFLBzn4&gatewayAdapt=glo2esp
• Splint lock nut to attach actuator to wood floor (x4)
  https://es.aliexpress.com/item/1005...t_main.92.2656194dFLBzn4&gatewayAdapt=glo2esp
• Flexible Plum Coupling 12x14
  https://es.aliexpress.com/item/1005..._main.167.2656194dFLBzn4&gatewayAdapt=glo2esp

Here some final tests:

Wow, what a beast! If I only had the room... So simplistic but powerful by the looks of it. I like how subtle it's movements can be.
I had to be a little more conservative than that. I really need to finish my attic room so I can go big.

Are the three struts guided laterally in any way? It looks very wobbly.

No, they are'nt ant yes, thats a problem I noticed too and is the cause of the torsion I mentioned in the original post. With slow movements it is not so much noticeable but the faster are the movements the problem magnifyes.

My next step is how to solve that problem. One option is make shorter arms that will be traduced by less heave lengitude (or height depends of the POV) and overall less range of movements but I think it will be enough too. Other option is install a kind of telescopic pillar with a joint to guide vertically the chair platform.

I'll investigate the best option and I'm listen to your suggestions.

Thanks!

A center mounted telescoping pillar would be the best idea IMHO. Heavy Duty trucks in the US have remote mounted pumps driven by a transmission mounted PTO unit. The shafts that connect the 2 have u-joints and a great deal of sliding length. Steering shafts from the same type of trucks would be good too. Both can be found in truck junkyards.
I'm looking forward to seeing what fix you implement.

I´m still wondering that your Motorunits are still standing, because you are using them out of there middlepoint anyway.

The easiest way would be a fixed axle at the back with an eye at the end. This ensures that your rig no longer moves sideways or lengthways. The eye at the end is then only to compensate for angles. However, I would then recommend a base frame that connects all the motor units and provides a base so that the whole thing is stable. Since your rear motor unit is now absorbing the main forces, I doubt very much that your rig will stay in place.

The actuators are fixed to the ground otherwise will be impossible to keep it stand especially when someone sit on the chair. I'm now making three squared bars to join al three actuators toghether to make the lower structure more rigid but it will not solve the side and length movements.

Yes, I was searching for a kind of PTO or some telescopic column but all of them are or shorter or longer to my needs so one of the option is build it by my own. MEanwhile I'll think other options.

Thank you guys

I understand. Like i said, weld or assemble an angle on the rear motor and mount it directly to the rig with an balljoint. Then you will not have any side or longitudinal movement anymore.

Something like this looks interesting. https://pbclinear.com/product-features/linear-bearings/square-bearings-linear-rails
A DIY mod might be to mount a u-joint to the bottom of the seat and the top of a square tube. That tube sliding in and out of a larger square tube that's mounted to the floor. A little grease between them and you're good to go. You could get fancy and use a bigger outside tube so you could mount a couple of linear rails between them for less friction.
EDIT: Just found this. https://8020.net/linear-motion
This guy's using the sliding yoke of an HD truck 2-piece driveshaft. That's a lot of weight though.

I think it will not be that simple coz the actuator arm draw an arch so it needs to be free to rotate like it is doing now otherwise when it travels to top or to bottom it will pull the whole platform.

Yes, I was finding for that these days but all of the PTOs I found does not have enough travel to satisfy my heave distance.

At this time I'll weld both arms of each acuators together with a platen to make it stronger and also I'll install three bars to join all three actuators and I'll see how it performs then.

You have this with all non linear plattforms but you will not notice it, because it will be only a few mm, less then your travel at the moment.

What you can also do for testing, move the actuators further apart.This means you don't run the risk of singularities in the first place.

Please correct me if I'm wrong. The platform should have absolutely no yaw, sway, or surge movements. The COR should be centered at the centerpoint of the 3 towers, which I'll call position 0,0,0. All X, Y, and Z movements are centered at this point on the XY plane. If this true then fixing that center to the base with a telescoping shaft with a u-joint on top would hold the zero X,Y position while allowing the Z to move up and down. It shouldn't matter if the towers cause uneven pull on the platform because of their arc. The center post (if it is robust enough) will ensure that the platform does not move in the XY plane. Am I missing anything?

You wouldn´t feel the 2mm moving in the Y-Axis at all. And again, at the moment it´s only because the axis are staying to near to each other what will cause the singularity in the Struts. Move all axis few cm apart and you don´t get this movement. So you don´t have to change anything else.

I didn't understand any of that. What is "singularity of the struts"?
let's say the platform performs a roll movement to the right. right strut goes down, left strut goes up, and the rear strut doesn't change position. Because of the connecting rods and ball joints the whole platform will roll, twist, and sway to the right as the center of gravity changes in relation to the center of rotation. The whole platform will be oscillating and swinging around as it moves as there is nothing holding it it place over the center point. Are you saying that the placement of the towers can eliminate all that unintended movement?

Yes, some stupid Comparison would be an Tripod Barbecuegrill.

No, you are totally right. The main problem is that I have almost 390mm of heave and I ATM I didn't find out any solution that convince me completely. Few days ago I welded a platten between the two arms of the rear actuator and the lateral movements are gone. Now I'll do the same with the two front actuators and then I'll joint all the actuators with a steel bar. I think it will do the job.

Well, the displacement is 118mm, is not a small thing. Keep in mind the arms are 237 long and the angle they travel is +-60º

You wouldn´t notice it!

But anyway, try to get the base units further apart from each other, that the vertical struts will never come to 90° vertical and you will not have this problem anymore.