Linear actuator using ClearPath Integrated Servo System @ 72V DC

I have downloaded the personal version of Fusion 360 and can export the 3d parts to .STL files.
Select object -> go to file (top left) -> select Export -> select file type.

@Thanos, @Dirty. Could any of you answer my question if Speed or Torque is important (post # 194)

Thank you in advance

hiya @T R Para that was a great video , def cleared up some of my panic mode concerns , so gotta say thanks for that and a bit of relief Phew

I'd say torque is more important when you are talking dynamic loads that are shifted around. The motion platform is not static build to calculate to the max load, you need to "overengineer" is to be able the include the inertia of mass of the load you move around.

Speed is good to have if needed, but in that case you would just have to get larger motors to equally handle load and speed. See this 737 sim for example that has 130ST servos and 160:1 reduction gearboxes to be able to move around 2000lbs realistic cockpit on it:





And no, the 6DOF above is not commercial, Mike build it in his garage...

Nice video, unfortunately I don't have such a large garage.
I think I'll be using the rig mainly with VR. Keep in mind to be able to set up 3 34 "monitors with some instruments from a cessna. At least not a full cockpit.

That does look like my "Slider" indeed

Btw, if some of you native speakers stumble over some strange naming, let me know. I'd be happy to learn the correct technical terms. Some stuff you just don't learn in school. And I think there are still a few German expressions in that assembly

On that slider:
Be sure to print them hot enough, because virtually all load acts across layers! You really want that layer adhesion! Print with 3mm walls and 60% infill. Or better even: solid! The cross-sectional area is just barely large enough to give a safety factor of ~3. So, nothing to brag about. If I had to do them again, I wouldn't strip-away all that material. I did this to save some print time. Also I decided to use PETG instead of PLA, because of its ductile failure characteristics as opposed to PLAs brittle failure mode.

Also, definitely use threaded inserts to screw into the part.

You may think you get a solid connection screwing directly into the material using an undersized hole, and that may actually be the case the first time you do it. But after multiple ins and outs the material will suffer and you loose that tightness. To me, those threaded inserts are definitely worth the few cents they cost.

I use these. There are cheaper ones, but they lack the diagonal knurling. That makes quite a difference on pull-out force, as Stefan from CNCKitchen demonstrated very impressively here:


Also worth watching:

...and...


And something to be aware of in general:
I have made these actuators for a nominal load of 1000N (~100kg or ~220lbs)) axially. And all parts were FEA tested in Fusion360 for that static load case with a safety factor of at least ~3 but the motors can generate A LOT more than that!

Lead 10mm:
4NM --> 2500N
2.4NM --> 1500N

Lead 5mm:
4NM -->5000N
2.4NM --> 3000N

Just be aware of this. A chain is only as strong as its weakest link!

That said, I am quite confident that they hold up well. I put friends and family on the rig without fear.

Cheers,.. Dirty

I'd say that's like asking a pilot "What's more important, airspeed or altitude?"

And the answer is the same: You're in trouble if you run out of either.

If your motors become torque saturated the motor drives will trigger an alert and stop the motion. It's not like they will continue to "do as good as they can". They will shut off.

On the other hand, if they reach their max speed limit, they will still move, but they will be playing catch-up with the commanded position, which will give you "corners" in the motion where smooth motion should be:


In this diagram, graph-height is actuator position, gradient is speed and curvature represents (quasi-) acceleration. These corners (orange graph) have infinite curvature and therefor infinite acceleration, which will give you spikes in the torque demand.

So, insufficient speed can also give you torque problems on top of crappy motion.

Solution: Find a worst case scenario for the torque demand. That is usually when the platform is all the way down and reversing direction upwards. And then make sure that that scenario is covered. Then check how much speed that leaves you with. I have tried to consider this scenario in my Excel sheet. Fill in the orange cells everywhere, get results in the last green sheet. I have also added some "basic diagnostics" but let me know if those advice are nonsense.
I have 420mm/s on my rig, and that's OK. Still, 500mm/s is what I would aim for if I rebuilt the rig today.

Very simple solution: Use a 90ST M02430 on a 1:1 Belt drive and an SFU1610 ballscrew to get this out of my excel sheet:

...and yes, those are the theoretical values. Actual values will be below those, but there is sufficient margin in those numbers.

In my opinion, the position accuracy itself is not that important. You will hardly be able to tell if one actuator is 3mm off or not. In reality it will be well below 0.1mm. But none the less, you WILL be able to tell if an actuator vibrates, stutters, runs rough, etc. So it does make sense to look at the encoder resolution, especially for slooooow and smooooooth changes of direction.

As soon as the speed & torque requirements are met, mechanical smoothness comes next on my priority list. The encoder resolution of these motors is good enough for use in motion cueing. The mechanical side of things is gonna be the bottle neck here, I believe.

Dirty

Thanks Dirty,

Your Excel helps a lot to choose the right servo. It seems that the servo I had in mind is going to do just fine.

Your previous post about the inserts was also very interesting and helped me a lot to better understand the design in detail.
I think it is slowly completing the grocery list to start shopping.

Prost Builder.

As I go deeper and deeper into matter, questions come up again as I am not a mechanical engineer.
Is the rotor inertia even more important when selecting the servo motor. In my understanding, you can only do something with this if you know the ineria of the platform.

Is this correct and do you have any advice? or am I going too far into the matter.

Dizzy Builder.

Too far

...I think. But not 100% sure

I didn't actually calculate any values, but my intuitive understanding goes like this:

1. Without any load, the motor is able to ramp up from standstill to max RPM in under 0,1 seconds. That's way more than I will ever need. --> Motor rotational moment of inertia is negligible.
2. With a lot of load (rig + myself), the motor ramps up from standstill to max RPM in about 0,2 seconds. That's still way more than I will ever need. --> Motor rotational moment of inertia plus rig translational moment of inertia ( a fancy way of saying: "its mass") combined is quasi-negligible.
3. The rigs translational moment of inertia is ~2 orders of magnitude higher than the motor rotational moment of inertia. --> Motor rotational moment of inertia is negligible.

Another way of phrasing it would be:
Imagine how much energy is stored inside that spinning motor axis(0,2Kg) at 2500RPM. And then compare that to the amount on energy stored in a 150Kg mass, moving at 0,5m/s. Like I said, I haven't done the math, but I wouldn't pay too much attention to the motors axis inertia. In fact I didn't pay any attention to it at all.

If someone has a deeper explanation, I am always eager to learn about this.

I don't know if you've seen it already or not but for figuring out the worst case torque demand my Stewart platform force calculator should help.

God! That thing is awesome!!!!


No, I haven't seen it before. I have done some of those calculations by hand for a couple of points on the envelope that I thought were critical, but this lets me do it in seconds! For many more points!

COOOL!!!!!

I just played around with it and one can see very nicely that even though there is only a 150Kg load, the actuators will have to provide a combined 190Kg. Just to hold the rig static:


Then if I want to accelerate upwards from a very low position that grows to 380Kgs.


And then adding some reasonable rotational acceleration will bring me close to my 1000N actuator limit:


That's so cool!!!!

I didn't check the formulae, but the results are perfectly plausible. Question: Does the inertia sphere represent a body that has the same mass, but all distributed on the SURFACE of the sphere, or evenly distributed over the VOLUME of the sphere? I assume it's the surface, so 1000mm is a conservative approach, I think.

I did give your post the "winner" rating, but if I could, I'd hit that a hundred times.

Thank you!

I have been printing a few pieces.
I have done 3 so far.
I know that there is at least one more.
Could you tell me how many 3D printed pieces there are in total ?

Cheers, I'm glad you like it. The inertia sphere is modelled as a solid sphere.

Regarding Fusion 360, the new free version does still have .stl export and they've added back in .step export as well. But I believe it has lost finite element analysis.
https://www.autodesk.com/products/fusion-360/blog/changes-to-fusion-360-for-personal-use/

@Dirty Did your file get removed? I wanted to look over a few more items but Fusion tells me the model is empty..
Thanks

Ups,.... I will check.

I think my (educational) license ran out after 3 years. I haven't extended it yet.

Should work now! Let me know if it doesn't.

Hi @Dirty !

Thanks so much for the thread. I am studying closely and setting up parts to build the rig based on your gorgeous design. Adapting pieces to work with imperial system (aargh! but is easier to find little things like washers and so), and doing some Unreal Engine visualization at the same time to test things.

I have a question on how you hold the bearings in place of the nut holders?



Did you used Loctite? or the force of the system would not let them slide out?

Will keep updated on my build, and thanks again for sharing your files and expierence!

Charlie

Hey @vectorcharlie

Glad you liked it!

I have glued the bearings into place with Loctite 638. Initially I planned on putting end caps on either side of the nut housing to constrain the bearings in place, but when I gave the Loctite solution a half-hearted try, I realised that this is a way cleaner solution.
I had concerns, because I thought those bearing would eventually need to be replaced, but as far as I can tell, after about 9 months of serious (ab-)use, there is no noticeable wear.

And even IF they wore out eventually:
2 bearings: 2$
1 nuthousing: 4$
1 spacer: 50ct
1 drop LT638: 3ct
------------------------
grand total: 6.53$US


Be aware that Loctite cures under two conditions:
1. Presence of metal
2. Absence of oxygen

The gap between bearing and nuthousing is suuuuuuuuper tight, which means that the glue will cure in about 5-10 seconds at room temperature!

Should you screw things up during glueing and the bearings are misaligned, you can heat the part up to about 150° (oven or heatgun) and it will essentially turn back into honey-like liquid. No harm done.

I put a spacer in between the bearings, to make sure that the axial loads are shared evenly between the two bearings. You can see it here:


On the first one I had this plastic "star-thing" in there to keep the spacer in place, but here's what worked a lot better:

1. Lay the bearing onto a flat surface
2. Put a tiny bit of LT638 onto the outer surface of the bearing
3. Put a tiny bit of LT638 onto the inner surface of the nuthousing
4. put the nuthousing over the bearing
5. give the nuthousing a good hit with a wooden block to slip it over the bearing. The table surface will ensure a flush fit.
6. let cure for 5 minutes
7. insert the bolt
8. put LT638 onto the tip of the spacer and slide over the bolt to glue it to the bearing
9. put LT638 onto the inner surface of the nuthousing on the other side
10. put LT638 onto the tip of the spacer
11. slide the 2nd bearing over the bolt in position
12. add LT638 onto the outer surface of the 2nd bearing
13. use a wooden block (with a hole in it) to smash the 2nd bearing into place

This will ensure perfect axial alignment and a good connection between the bearings through the spacer.
And there is virtually no chance for stuff to go wrong.

Use very little Loctite. And always remove excess Loctite quickly, because it will sit there patiently, waiting to creep into any gap it can find to glue together parts you don't want glued!

Cheers,....

Dirty

Dity,

Super insightful!! Thanks for the detailed response! Will give it a try once I get my loctite.

I have another question, more related to the design...

Wondering why you picked such a long "Upper block" (98mm).

I definitely understand and like the dual Igus bearing, but wondering if that distance gives any support benefit. And if so, would 2 pieces even more separated make sense?



The entire design is very thoughtful... So I am sure you ended up with this specific design for a reason.

Thanks!!

CI