Yet another 6DOF DIY build

Hi,
I started this build way back in 2014 but ran out of time, so the project was more or less on hold for the last two years. However, I recently want back to it and finished it.
Well kind of finished it: It's now more or less functional, but there's still a lot to do and refine.

But let's start from the beginning:

I wanted to build a 6DOF Steward-style motion platform. So I checked out some other DIY builds in the internet.
I decided to go with self built linear actuators, based on 24V scooter motors and bicycle chain drive.

First I tried to research the main geometry and size of the build, so I built a quick mockup model:








Since I only have limited room for the motion platform, I tried to get it's foot print as small as possible, while still building a full size motion platform.



Then, I tried to build the first of the six linear actuators. ... and I failed miserably. Although I tried to drill all holes for the rods as exactly as possible, I couldn't manage to get the rods parallel enough. The linear bearings always got stuck half way.
Also, after building one of the linear actuators half way, it was obvious, that the whole thing would be way to large for the limited room.

Meanwhile, the 6 scooter motors arrived from china:


When I saw the motors with their bike chain sprokets, it hit me:

Why not using bicycle parts to build (much smaller) crank style actuators, instead of the huge linear actuators!

So I build a first prototype of a crank actuator out of one Motor, some cut up parts of an old bicycle, a Saberthooth 2x25A motor driver and a rotary encoder:

The plan was to still use some bungee cords to take some of the static load from the motors:



It seemed to work so far. So I did some load tests:



With the new actuator design, I went back to the "drawing board" and built another mock up model. This time with 3D printed models of my crank actuator design:





This was mid September 2014.

Some other projects came up, so I had to pause the build for 3 or 4 weeks...

To be continued.

In mid October 2014, I built another model, this time with 6 micro servos instead of the 3d printed actuators. That way I was able to do some live testing with the model and different driver software.



Since the foot print of this construction was still too large, I rotated the servos by 180 degrees in the next version. I also 3d printed the upper frame and a (more or less to scale sized) seat:



Well, that's more like it.

The model worked great, so I started to build the real thing.

To keep the whole thing modular (and transportable), I decided to build the lower part with the actuators out of three segments.
Here's a first layout test:


The base plates got a sub-constructions in order to have some room (and anchor points) for the bungee cords:



(The following is an obsolete construction, but it's the only picture I have with some bungees in place...)


And here's an image of the actual actuator construction starting:


For the universal joints, I needed a way to fix them on the cranks. I tried to use the inner bolts from actual bicycle pedals. But I wasn't able to find bolts with the correct thickness. I tried to machine the bolds on my lathe, but since the bolts seem to be out of hardened steel, machining them is really no fun at all!



I (much) later decided to machine my own custom bolts from scratch. But more on this later.

A friend of mine was so kind to weld some new bicycle bearing tubes onto some aluminum base plates. That way I didn't need to cut up more old bicycles for the bearings. And it also looks much nicer:



Another problem to solve was to find a way to connect the rotary encoders to the gear train. Unfortunately the axle of the motors wasn't accessible from the back and there wasn't enough room between the sprocket and the wooden bearing block.
In an early prototype of the actuators, the chain ran in front of the wooden blocks to the chain ring in the front (the silver one on the picture above). So I had some room to set the encoder in front of the sprocket. In order to make the axle longer, I machined a special nut out of brass:


This nut extended the axle enough to install the encoder:




But then it turned out that it was necessary to run the chain on the back of the bearing block to the chain ring on the backside. Otherwise, the two actuators wouldn't fit on the base plate.
And as said before, there wasn't enough room to install the encoder between sprocket and bearing block.

So I had the idea to 3d print custom gear wheels, which connected directly to the front chain ring:


The encoder was installed on the back of the small gear block.

Although this kind of worked, there was a lot of play in this setup and I decided later to drop this idea. But for now, it was good enough.

I continued with building the rest of the actuators:



In order to be able to test the rigs geometry, I machined some temporary, wooden bolds,...



... and built six main rod mockups out of cheap wooden sticks with ring screws as universal joints and a temporary upper frame, to hold it all together:


This wasn't strong enough to put some load on the rig, but it was good enough to see if the whole thing might work in the end:



Well, it looked promising so far!

This was mid November 2014.

(to be continued)

Here's video of the 6DOF model in action:

And here's a short video of the geared encoder:

Finally a video of the first test with the temporary, wooden bolts, rods and upper frame:



You can see how some of the actuators sometimes "pump" due to the play in the geared encoders. Definitely, some changes were needed here...

This is what I call a fine strange contraption. It is full of strange ideas that seem to work. You have to finish it for sure. Keep on.

Interesting concept, @zaggo. My main concern is the behavior under load for the following reason:
You are using nearly 180 degrees of movement of your levers, although the ranges near 0 and near 180 degrees are very ineffective. Within these ranges, only a small part of the available torque goes into the direction needed for lifting or lowering the rig, while most of the torque is lost. This is also the reason why most simulators using the lever approach are only moving within rather small angle range with its center at 90 degrees relative to the moving direction.
And this in turn is probably the reason why most of the 6 DOF rigs use linear actuators.
Anyway, I wish you much success while going on.

PS: Some time ago I did some experiments with the same type of linear actuators that you had in mind, see here and here. But since I only wanted to build a 2 DOF rig, I continued with the common lever concept.

I really like the creativity of your project. I too am thinking to use e-scooter motors as they are cheap and readily available here in the UK and with free delivery. Did you stick with using Sabertooths to drive them and have you experienced any problems using these motors?
I'm interested because I'd like to drive my linear actuator (still in development) with an e-scooter motor.

I'll be following your progress.
Good luck!

Thanks for the feedback.
The 180° of movement in the video were just to test and show the maximum possible movement of the platform. I don't think that I'll really want to use that much movement during actual use.
With the setup in the video, the platform had a vertical range of almost 40 cm! In the meantime I already reduced the length of the cranks (more on this coming soon). And even with the reduced range, I'm pretty sure that I'll use only about 50 to 60 % of the possible movement. I didn't do much real world testing yet, but after the few times, I actually sat on top of the platform during tests, I'm pretty sure that more range would be more to scare people than simulate actual game movement...
But we'll see!

Yes, I still use the Sabertooths (with Kangaroo). They work great with the motors so far, but as said before, although I started the project 2 and 1/2 years ago, there was a pause of almost 2 years until I came back to it 4 weeks ago. Right now it is nearly finished, but I didn't use it much yet, other than some functional tests.
Based on my testing, I'm still pretty confident, that the three 2x25A Sabertooths are ok for my application with the 6 motors.
I'll try to keep you updated!

Ok, back to November 2014

After some messing around with the inner bolts from bicycle pedals (see above), I decided to machine my own custom bolts from scratch. Here's my first prototype:




(The wooden rod was only temporary!)

I turned another bolt and then was Christmas. And then some stuff came up, and then some work and more stuff and then....






... it was 2017

Most of the time in between, the project was packed away under a big table in my workshop.
During the whole time, one thing bothered me most: The unsatisfying mounting of the rotary encoders.
There was a lot of play in the geared solutions I tried.
The one above, where the 3D printed sprocket meshed directly with the chain ring on the front, was the worst in this regard.

I did a second attempt with an also 3D printed gear, bolted on the back of the chain ring in the back (the one with the chain on it) and a fitting sprocket with the encoder on the other end.

At least, this enabled me to remove the crank-with-chain-ring-assembly from the front and replace it with a simple crank.

However, this setup wasn't really much better in regard of play.

So I tried again to squeeze the encoder between the motor sprocket and the wooden bearing block. But after some messing around, I finally decided that the only "clean" way to go would be, to modify the actual motors, and to mount the encoders on the backside of the motor housing, directly on the main axle.

In short, to go from this... :


...to that:



I tried the modification on one of the motors earlier this year and it turned out great. A full success.

The wobbling of the crank (when holding specific positions) was gone and the encoder was finally out of the way, nicely sitting in its place on the back of the motor, where it belongs.

The modification meant to completely disassemble the motor, drill the center hole in the back plate for the extended main axle and also drill and tap the 4 mounting holes for the retainer.

The encoder retainer (the green thing on the image above) was (again) 3D printed.

Finally, I had to drill and tap a hole in the back end of the motor axle to screw on the (also custom machined) brass axle extension.

After that, I only needed to manage to re-assemble the whole thing

I documented the whole process on video and I plan to publish it on YouTube later this year (as soon I find some time to edit the raw material).

This was end of January 2017.

(to be continued)

Any updates on this project ?

It seems to me you could mount lever arms direct to the output of the motor and not use the bike cranks at all.

What is the gear ratio of the motor gearbox ?