Zen 6DOF Motion Simulator

Beautiful just beautiful all most there

painted all black. now all i need is Thanos controller and servomotors from China but that part will need to wait a bit unfortunately. i'll be back

That is looking awesome. Great idea building the rig while waiting for the motors and drivers.

i have little doubts if 3d printed parts will be strong enough (especially the bottom one) but we'll see...

ok, long story short - those 3d prints started to make strange sounds and i don't trust them. i've decided to change it to laser cut steel and since i have to redesign it a bit i've also decided to make it FOLD BACK so actuators are shorter and it looks much better plus the angles are better. while waiting for parts i've lowered it on those threaded rods and it looks already better.

Parts arrived. I've already painted them and will be exchanging them on actuators tomorrow. Will post more pictures then

I subbed to follow this thread.

quick update:

1. i've assembled laser cut steel parts so no more 3d printed parts. all those parts were roughly 150 euro - i see no value in 3d plastic parts in such case...
2. i had an issue with upper universal mount that was epoxy glued to the pipe - it broke off while i was manually raising the rig so i had to manufacture it from steel with thread (done by local locksmith for 100euro).
3. waiting for the servos but customs mixed currency and instead of PLN (polish zloty) they assumed EUR and made it 4x more so i had to reject the package and it will be there on monday with corrected VAT.
4. already got Thanos controller in the meantime so only waiting for those servos
5. it takes a lof of space in my "playroom" and wife is not happy with it as it's less comfy to watch movies now...

Are you prepared to drill some big holes through your nice floor to fix your machine - YOU WILL NEED IT! believe me
Wickie

He there

Nice build! Plenty of pics! Love watching! Better than Netflix

On 3D printing:
3D prints can be made durable when done right. PETG is OK already, PC (Polycarbonate) is better, but a little more difficult to print. Either way, you will have to WAY over-design the parts. Thick walls, 60% infill, maybe even print solid. Do a FE analysis and make sure to limit the tensile stress to ~4 MPa. That will give you a safety factor of 5-6.

I've been using mine for an accumulated ~50hrs and see no wear at all. The design freedom that 3d Printing gave me made the difference. Metal is of course stronger, but unless you have a 5 axis CNC you will not get those complex shapes and all that highly integrated functionality into the part. All in all I went with 3D printing.

On your geometry:
With this arrangement of the U-joints, you will have to be very careful to avoid gimbal lock. As I see them right now (if I see it correctly on the pictures), you are looking at two potential problems:
1. Your actuator MUST NEVER (!!!) be vertical.
2. Your actuator MUST NEVER (!!!) be perpendicular to the upper platform.

I'm not sure, if there is a safety feature in Mover to ensure that, so you will have to test this very carefully. I'd love to show you a video of an actuator going into gimbal lock that I have here, but it is not mine, so I wouldn't want to post it without permission.

Some background info on mechanical DOFs:
Imagine you're an ant and you want to crawl from the floor up to your 6DOF rig. On your way up, you WILL HAVE to cross 6 Joints allowing for those 6 degrees of freedom.

- The first two joints are in the lower U-joint.
- The third is the extension of the rod from the actuator body (Wanna know why that is?)
- The forth is the rotational bearing that you added on @Erik Middeldorp 's recommendation.
- The 5th and 6th are again in the upper U-joint.

Now, imagine that each of those 5 rotational joints (ignore the extension for now) have a "line" drawn through the rotation axis. Here's what you have to make sure: Those lines may NEVER be co-axial with any other one of those lines! If they are, then you will lose one degree of freedom and you will almost certainly break an actuator.

It might not become a factor! So not all hope is lost Just have an eye on it and notice how on commercial sims, as well as my own, the axis of rotation are arranged especially for this NOT to occur.

Cheers,... Dirty

thanks @Dirty for you feedback!
closest to vertical i can set it while max yaw - 3 actuators fully extended and 3 at their minimums:



when it comes to perpendicular to the upper platform i think it's when 2 (right or left) actuators are fully extended with others at their minimums:


i believe those conditions are highly unlikely however i will need to make sure it won't happen...
from what i see currently without servos i think i will lower the stroke of actuators in the controller anyway as i would move too much i think.

EDIT1:
Additionally, i understand (and please correct me if i'm wrong) that this issue applies to your design as well, right? it is not specific to my joints?

EDIT2:
ok, i think i get it now... so my question about orientation of the upper joints from the previous page actually makes difference - rod that comes out of a bearing should be mounted horizontally not vertically as it is right now, right?
this one: https://www.xsimulator.net/community/threads/zen-6dof-motion-simulator.14617/#post-200175

I could easily change upper mount by adding L shaped steel profile and hence mounting bearings 90deg rotated.
I believe that solves issue on the upper mount.
Theres not much i can do about bottom one - i would need to somehow mount it at angle but i think upper one will effectively eliminate issues.
Agreed?

Hey @pooh

I think you are correct. If you have every other actuator fully extended and every one in between fully retracted, that should give you the maximum possible yaw angle. If THAT doesn't give you a vertical actuator then I couldn't image what would.

I agree. Even if you can create those angles manually, does not mean that you will ever see anything even close to that in actual operation. Like I said, there is still a high chance that gimbal lock will never become an issue for you.

I would do it this way: Leave the actuators at their full stroke. It's OK if they CAN use the full stroke. Rather tune the filters in Mover in a way that they will never NEED those full strokes. You DON'T need those full strokes to reach any ONE insanely high pitch/yaw/roll etc. You DO need those full strokes to be able to produce a COMBINATION of medium pitch/yaw/roll/surge/sway/heave all at the same time.

Be very conscious where you "point" the axis of those bearings. You want to point the axis of those bearings in a direction where the actuator will never be. Ideally, if you had them pointing slightly outwards and slightly upwards they would be safe I think I saw this on @SilentChill 's new rig. It looks somewhat unorthodox, but it does prevent gimbal lock very effectively.

I am trying to illustrate the challenge here:
That central axial bearing that allows the actuator top to rotate, is absolutely essential. And it needs to be exactly where you have it! Nothing you can do about that. But since the axis of this bearing is coaxial with the actuator axis, it points directly AT the upper AND lower U-joints. Your only option is to point the axises of those U-joints away from the actuator axis. Since the actuator is moving constantly I should probably say: "...point towards where the actuator axis is least likely to ever be".

Here's a picture of a professional sim in a neutral pose. Note how all joint axises are perpendicular to each other. The same is on the lower U-joint.


I just wanted to raise awareness for this issue. If you are still early enough in the build, I would try to angle the axis of the first and last bearing in the sequence slightly to get a larger margin.

I will follow this build. Post a video when this thing moves

Thank you so much for your valuable feedback! I will redesign upper mounts. It's great that people like you or @Erik Middeldorp provide this kind of feedback so people can avoid huge risks and other can learn from it as well!
I need now a week or two and should be ready to launch it

Ok, I've managed to get servos out of customs. Had issues with control cable between aasd and Thanos controller like one person in controller thread but after 2h of soldering they are straight now...
This thing is loud and I'm not sure if speed is good enough - those are 2010 ballscrews so in theory it should fully extend (500mm) within a second...
Take a look:

What's your motors RPM? What's the torque? Do you have a reduction in the belt drive?
If it's a 1:1 ratio you should get 500mm/s out of a 3000RPM motor. It seems from the video that that is the case. Does the motor reach its target RPM? You can check that in the controller display.

If you apply thick high viscosity grease (the thicker the better) it will take away a good share of the noise and also you can put rubber pads somewhere between the actuator and the upper platform. Maybe even also down on the ground. Can you tell if the sound comes from the ballscrews or maybe the linear bearings?

Servos are 3000rpm and 750w 2.4Nm. They do reach 3000rpm but i guess it takes time (even though electric motor has instant torque). I don't have reduction. It just feels like it takes more than a second to full extend...

Regarding the sound - it comes more from ballscrew i think. Theres lots of lithium grease put inside the ball nut as well as directly on the screw. I think aluminum profile (the enclosure) resonates a lot as well. We'll see how it behaves after some time.

the motors are pretty poweful at 750w

i reckon you could change the gearing to make them faster

Using the buttons on the main screen for manual movement, does not give the maximum speed. If you want to try the max speed, press the encoder in while on the Bars screen to show the Numbers screen... pressing the buttons there result to high speed jumps to end positions of the actuator!!! Beware, secure the motor or actuator before trying, it can fly on 500mm/s!!!


See what I mean on this video: