Showroom 2 DoF using DIY SFU 1605 ball screws

Hello everyone, here is another iteration of my 2 DoF simulator.
Some backstory: A bit has changed over time including switching to worm gear drives, however they proved to be not adequate for this type of sim rig. Honestly they had more backlash than the wiper motors I previously used. This proved to be quite annoying especially with smooth surface transistions. Here is a video in which you can see that the upper platform jumps around quite alot.

I could have modified it a bit, maybe I could have used gas struts to dampen the motion or machine a new worm gear so it would have minimal backlash. However I chose to go a different route. ​

Recently I saw a few threads pop up using this type of screw lever mechanism. Each one of these posts used their own design . I summed up that they all share a few design principles:

• Three pivot points. One for the motor/screw coupler, one for the lever arm and the last one for the ball screw nut.
• The first pivot point needs to be at the center of ball screw mount.
• The second pivot is at around half the length of the ball screw.
• The third pivot point is in the center of the bearing housing.
• In order for there to not be any crashes, the lever arm needs to be about twice as long as the bearing mount arm.
• The resting position is when the bearing mount arm is at 90 degrees to the ball screw. The angle at which the lever arm is in relation to the bearing mount arm doesn't matter.

Another thing that I was sort of limited by and didn't want to do is custom machined parts. I know I could have made it so it looks more elegant, however frustrated with my previous version I wanted this now .

For this build I used the following components:

• 1605 SFU ball screw 250mm
• KF001 12mm housed bearings
• F688 flanged bearings (for the ball screw nut pivot)
• 8mm rod ends
• Electric steering assist motor
• 16x16 square rod (for the frame)
• 25x5 flat bar (for the lever arms)
• 80x5 flat bar (for the motor/ball screw mount)
• 12x2 precision pipe (for the pivot joints and spacers)

Here is a build video for anyone intrested:

​

Overall im quite satisfied with the result. It's a massive improvement over a worm gear drive. However there are some distinct pros and cons.

Pros:

• No backlash
• Very smooth movement
• Large range of motion from a 250mm ball screw
• Relatively low price
• Easily repairable
• Silence

Cons:

• Ball screws are not self braking
• Due to the above a higher power PSU was required and the motors run a bit hotter.
• The pivots and arm lengths need to be calculated or better yet simulated so that no superposistions or crashed occur.

Here is a video on how it rides:

​

I'll be happy to answer any questions .

add silence to the Pros

Nice. A much simpler design than mine. https://www.xsimulator.net/community/threads/sort-of-diy.19294/#post-257578
I suspect that you have a good deal of cogging during slow movements without a planetary gear between the motor and screw.

Yes, silence is a factor I initially didn't take into account.

Very nice project. Yes, mine is simpler but doesn't look as professional as yours. I actually don't have any cogging or atleast don't notice it. I can spin the ball screws by hand and they smoothly move the platform. I had some binding issues when the self alligning bearings were tilted. In the build video you can see that I tapped them with a hammer and punch to get them in line with one another. The other thing that caused cogging was when the motor was at a very slight angle in relation to the ball screw. On the other hand I use large torquey motors from electric power assist steering racks, so maybe they can overcome any cogging during slow movements.

Often cogging is caused by the control algorithm. Most that I have seen use a move to position algorithm. That implies that the position you are moving to is where you want to stop. So the control lurches towards the target point and starts to decelerate m only to be given a new set point to lurch to.

If you are moving along a path which is composed of via points that’s a different algorithm. In that case you want to move at a speed which is dictated by the target velocity.

I have the impression that a lot of DIY builders just copy a design they think that works without understanding how it really works. That mechanism shown in post #1 is IMHO ridiculously complex. It uses a ball screw to convert rotary motion of the motor to linear movement, then a lever to convert that back to rotary motion and a pushrod to finally move the rig in a mostly linear direction. AFAIK this solution was originally invented by a company to fix their problems with backlash of worm gears. They replaced the gear with the ballscrew+lever mechanism which makes sense if you need to fit the fix into the existing space.

However, if you build a new rig and are free to use anything you like it makes no sense at all. Instead of the 1-to-2 lever ratio you could simply use a ball screw with 10mm pitch to directly push upward like a normal linear actuator. Or you could spent just a little more $ on a gear with less backlash. I just bought a 1:20 precision planetary gear with <15arcmin backlash for $35. Rotary actuators are much easier to build.

Well that's exactly the problem I faced. I've built this simulator first using wiper motors, then I used worm gears (albeit cheap and one had significantly more backlash than the other) and now I'm using this design. I didn't want to scrap this build or make some more serious modifications to it.

If I were to do a completely new build I would do it like you said. Some sort of linear actuator on the corners of an aluminum frame.

All things considered it was a fun experience to build and I'm happy with the result.