Challenge: Constant Velocity Joint (M. Gesik)

[Zolko]

I do not know yet if ASM4 can articulate the CV joint so may need ASM3 sooner than later.

To be honest, I wouldn't know how to assemble that using Assembly4. I did a classical cardan joint, and you can see that it's not a constant velocity transmission. In this case, it's mathematically exact, so you can even set the deflection angle to 89° and it'll still work. Now, how to do the same with that complex setup .... ? Generally speaking I have experienced that you need to be able to describe mathematically the model for Asm4 to be able to represent it. In this CV joint case, I don't understand how it works, which parts are fixed, which move ...

Now, as with the hexapod, if you accept to "cheat", meaning that you suppose that it is a constant velocity joint, and you impose the exit arm to have the same rotation angle as the input arm, and the middle part to have a deflection angle 1/2 of that of the exit arm (which is given by the gears), it might be easier. But then, you won't be able to check that it is indeed a constant velocity joint.

Cardan_Asm4.png (368.57 KiB) Viewed 2669 times

[jruiz]

phpBB [video]

source: https://www.reddit.com/r/mechanical_gif ... ity_joint/
Thingiverse: https://www.thingiverse.com/thing:4191753

After watching the video I googled a little and found the following article (date of publication: August–October 2009) that could be of interest to some of you:
A new kind of spherical gear and its application in a robot's wrist joint

which can be downloaded for free here

[ppemawm]

To be honest, I wouldn't know how to assemble that using Assembly4.

The CV joint assembly has been completed as far as I can go with Assembly4:

This is a view of all the interconnected parts using the local coordinate systems (LCS) provided by Assembly4. The LCS were attached to model edges so it is prone to topological problems if any features are changed.

Assembly4 provides a very convenient means for attaching hole LCS's, as well as selecting and adding fasteners to the model. Not all of the fasteners, circlips, or bearings are shown in this image.
.

Picture5.jpg (107.48 KiB) Viewed 2572 times

The last three (3) joints (see arrow) of the top 3-legged ring could not be assembled with Assembly4 since it has no solver. AFAIK it is not possible to "close the loop" of all the interconnected parts using Assembly4 without a circular dependence.

The geared links as shown were approximately aligned using the Part > Link Attachment Offset properties.
.

Picture6.jpg (117.78 KiB) Viewed 2572 times

.
Perhaps a master sketch approach similar to that used by OficineRobotica's hexapod may be a solution:
https://forum.freecadweb.org/viewtopic. ... it=hexapod

That is what I will research next unless Zolko can offer any other suggestions....not quite ready to give up after spending so much time on creating the parts.

[Zolko]

To be honest, I wouldn't know how to assemble that using Assembly4.

The CV joint assembly has been completed as far as I can go with Assembly4:

Nice. I tried my luck also, but without success either. I used the STL files as they came, and reverse-engineered their geometrical attachment points. Then, I tried to calculate the relevant attachment points in the assembly, but I'm stuck: I would need the result of the intersection of a surface (a cylinder) and a curve. Unfortunately, I cant attach the files themselves because they're too big. I could provide the skeleton, but since it doesn't work it's not of much use.

CV_joint.png (545.4 KiB) Viewed 2559 times

[Zolko]

I would need the result of the intersection of a surface (a cylinder) and a curve.

gotcha ! I got the help I needed, and so could build a working skeleton and assembly:

CV_joint.png (527.11 KiB) Viewed 2507 times

This is now an exact solution, exact to the point that I don't really know the real dimensions of the parts, so I had to guesstimate these dimensions. But it behaves correctly from -40° to 40° deflection, and full 360° rotation.

The CV joint assembly has been completed as far as I can go with Assembly4:

If you want to assemble your parts (which are much nicer) with my method, I attach the skeleton, which is the assembly where I have simply removed the parts. Parts are attached to :

CVJoint_parts.png (840.15 KiB) Viewed 2507 times

CVJoint_skeleton.png (360.19 KiB) Viewed 2507 times

The pictures show where the attachment LCS are in the respective parts:

• bielette_A to LCS_arm_A_1
• bielette_B to LCS_arm_B_1
• liaison_10_3 to LCS_common
• corps_lisse to LCS_input and LCS_output, BUT: these must be offset in the AttachmentOffset in the Z direction by 45mm and -45mm

And you have to play around with the RotX-RotY-RotZ buttons, obviously

[ppemawm]

...I attach the skeleton...

Thank you for that. I have been thinking too much in two dimensions. There is much to study in your file.
I will try to apply it to my model parts over the next week or so.

Genius you are, once again.

[aapo]

Hi,

This looks very very interesting, and seems to be quite an achievement! Congratulations Zolko and ppemawm! Would it be possible to include this in the FreeCAD examples, please? Section "Beyond Advanced"

[jpg87]

I do not know yet if ASM4 can articulate the CV joint so may need ASM3 sooner than later.

Can you please elaborate on why you use ASM4? Never used it until now but I would have used ASM3 without thoughts.

gotcha ! I got the help I needed, and so could build a working skeleton and assembly:

Hi
this weekend I couldn't resist the urge to try this mechanism with Asm3 WB.
Here's what I get.

CardanHomocinJpg87.png (80.18 KiB) Viewed 2342 times

Animation video:

http://help-freecad-jpg87.fr/02assembla ... layer.html

[chrisb]

Great video! Perhaps it would be even more convincing if shown at a motion angle of 45°.

[jpg87]

Perhaps it would be even more convincing if shown at a motion angle of 45°.

I remade the video by simultaneously varying the angle between the entry and the exit from 25 ° to 41.8 °: we then see that the arms with toothed sector collide (see more?).