Showroom FlyPT - 6DOF Brushless DIY Actuators

Yes you are right


EDIT: I corrected the post with the correct value...

I've just been reading about this actuator as a comparison to yours. Double bearings and 1.1kw servo motors ! Surely that's a lot more than is needed? A lot more expensive too but simpler electronics.

https://www.xsimulator.net/communit...earpath-integrated-servo-system-72v-dc.11873/

So, I had to buy a new power supply , and I need some batteries. Waiting for €€€.

But in the mechanical front, it's taking shape.
Copied @Dirty idea (https://www.xsimulator.net/communit...-servo-system-72v-dc.11873/page-2#post-160384):




And mounted the structure.
Thought of mounting the top in a different way, but it was simpler and stronger with a screw:








And here's my new baby:

Looking great. 3D printing is being put to the test here. There seem to be a few components that will be taking plenty of strain that i think have generally been fabricated in metal by others. Can't wait for the testing to begin.

That's a shame about the PSU. Are you going to place batteries between the PSU and the motor drivers?

I have solutions if the prints don't hold.
I'm more afraid of the bottom prints.
In the actuator I can always put an aluminium plate in sandwich.
On the base prints, I can use some simple screws. But let's test it

For the batteries, I was thinking of using two of those (really expensive...) in series:

In series I would get 24V and 230Ah.
They are expensive, but durable, accepting a total discharge.

They would be mounted in parallel with the power supply. Something like this:

Is this OK?
PSU would be adjusted to 24V + something to charge the batteries.

That's quite a serious addition to the system. Is there no way to do this with PSU alone? Was the issue that a surge of current went in reverse back into the PSU and blew it? Can't that surge be dumped elsewhere more cheaply? The lean simplicity of your design (like no potentiometers) is getting offset a little by the electrical complexity.

Yes, in the desire of making a load test, I destroyed the PSU with 75kg helping in a heavy way to lower the actuator...
Right now, my system is like the above but without batteries.
Placed a 10A10 diode (1000V 10A).
Don't know if it's enough or if it will heat to much.
I can also change the PID to handle charge lowering in a different way compared to going up.

The advantage of the batteries would be the possibility to add some solar panels later to charge them and only use them as supply.
Also some more current available.
Here's some photos of the PSU with the controllers:






There's one controller still disconnected and I need to add a plug to connect the remote emergency stop (like in the drawing)

Actuator design:

Told this to @Flyman here: https://www.xsimulator.net/community/threads/building-a-new-6diy.12321/#post-163132

When making the top mounts of the actuator, I thought of a solution where the U-Joints don't rotate.
For that I would have to change the actuator guide to allow rotation on the rod. Seems strange, but fixing the upper and lower joints the rotation would be "dissipated" on the rod, introducing a small error in positioning, but making less noise and friction.
I still might do that, just replacing the printed slider on the actuators and changing the top mount (I already have the printed parts for that).


This is possible because there's no need for a belt to get the position.
So anyone with encoders could go this way.
Certainly the actuator wont work without being connected to the rig or with the ends fixed for rotation.

I think it would be good to get some input here from the many people on this forum who must have encountered this problem of blowing up the PSU with a reverse charge. Your use of brushless motors is innovative but this issue has surely been solved already. @SeatTime or @SilentChill @SixDegreesOfFlight - any suggestions?

For me, the carbon rod should not spin in the top bearing (linear top bearing can not accept this)!!!

New updated Sketchup model with MSPhysics:


Link: https://3dwarehouse.sketchup.com/mo...eec50f0a92/6DOF-Motion-Sim-by-FlyPT-Version-4
(updated plans post at: https://www.xsimulator.net/community/threads/flypt-6dof-brushless-diy-actuators.10799/#post-139681

Rearranged first posts on this thread with:

New code for the ESP32 that now handles 3 actuators on two boards with calibration (https://www.xsimulator.net/community/threads/flypt-6dof-brushless-diy-actuators.10799/#post-139682)
New electronic schematics for the "potless" version (https://www.xsimulator.net/community/threads/flypt-6dof-brushless-diy-actuators.10799/#post-139693)

I just want to say i'm following this project with great interest. I'm in the progress of building a 6-dof myself. And I'm planning to use your ESP32 code with the hall sensor setup.
I've ordered some parts and hope to start testing with a prototype actuator in the next few weeks.

Keep up the good work.

Updated post with current schematics at https://www.xsimulator.net/community/threads/flypt-6dof-brushless-diy-actuators.10799/#post-139693
Go there for full resolution link.



Update also on the materials list.

Very impressive and clean design.
Thanks for posting the sketchup model. Its very useful to get my head around general setout dimensions to adopt for my rig.

I just wanted to check if you think the footrest will clear the front V (between the side pairs of actuators)?
I played around with the sketchup movements and at full forward tilt, I see the modeled footrest is inside the front V.
But, If Im using the measure tool correctly, the footrest appears to be about 800 from bum to heel.
I would think it should be more like 1100. But at 1100, this would clash with the front actuators.

I guess if you are setting it up 'F1' cockpit style, the feet are higher than the bum, so the feet should clear the actuators?

Well, there might be some extreme situations I should avoid with limits because of collisions, but also because some huge angles might damage the actuators or structure.
I think there's no problem, but for starting I think I will avoid some angles.
My idea of pedal support is what you see here in red:

Straight:


Some crazy angle:


The idea is to make the pedal base articulated and use two threaded rods to adjust the height of the pedals.
Using a formula car seated position would solve this easily, but my kidneys are asking for something more natural, and I want the base of the pedals at the same height as the base of the seat.
Design is not final, because I want some new pedals and they might be inverted, so this is going to change.

I have to say that this took me some weeks, I had so many doubts about seated position, and space...

So, I think I don't need another PSU or batteries... Took a ride and it WORKS!!!
I was afraid, because the actuators where moving really slowly in my first test, but it was just a case of PID adjustment
A video I made before adjusting the axis (almost random numbers with axis 5 using axis 2 data... crazy):



First part, a run in LFS.
Second part, some tests with output testing.
Ride was made with a recording of LFS. No video

Some thoughts:
-Violent, it even jumps, but with my weight and missing components, I think it wont move.
-Until now it's solid.
-Some noise, not really from the actuator, but from the top mounts.
The top mounts have to much clearance generating some backlash. Maybe a spring washer is enough.
-Some delay between actuators, so for example heave, a simple movement up generates a small oscillation...
-Adjusted all axis after the video and before the ride, but it needs much more tuning, movements not really matching in game car behaviour.
-I think the easy and more expensive part is made, now comes the less expensive and more hard part, the TUNING! I might need some help here

Video uploaded. Take a look in the first page or in the last post.

I thought of not making this, but here's a current look at the costs for the current version using hall sensors from the motors:


This is only the hexapod, no chair or wheel support.
I spent more with some errors and potentiometers.
If I made this today, this would be the cost.
Parts are from Farnell, Motedis and ebay.

Could have bought a pack of motor and controller for 98,99€:
BLDC 8015A + 57BLF03 (https://www.ebay.fr/itm/ACT-MOTOR-G...552331?hash=item3afed62d4b:g:ZU8AAOSw0hlZMXao)
Motor is slower, but cost would come down to 2636,58€.

I think it's not bad

Post of parts has been updated with some more parts.

Hello ,
Have you see one of your cylinder is off ( the one front camera ) ?
I'm happy for you that everythink goes well !
Your set up is nice … do you want plastic rod to hide the wires in the tube ?

I know. It was an error in the interface. Missed axis3 and repeated axis 2 in the second interface.
If you look, axis 2 and 5 behave the same way.
It was a mess, but now corrected.