CNC Machined vs 3D Printed Cycloidal Drive

Parts list (check website article for full list, affiliate links):

AliExpress:

In this video, we will take a look at my new updated version of the cycloidal drive that I made in the previous video, with 19:1 reduction ratio and see how it performs made with CNC machined parts versus made with 3D printed parts.

Like my page on Facebook:

[RandomCrap]

Those parametric equations would've been a HUGE help to me and my design team a couple months ago XD. We were using a lengthier method to create some 3-D printed cycloidal disks for a robotics project. Glad to see this method though! Saving this video for future reference :)

[m.sierra]

As both motors were skipping, I was really curious about the internal damage caused. Sad that you didn't show that. I'd love it even more if you would maybe replace the top with acrylic to see what's going on inside

[chadjensenster]

I made a cycloidal for the gearboxes of a robot arm for my capstone mechanical engineering project. They are nice because it's easy to have the reducer on the robotic arm and it looks good too. Great video. Thanks for taking the time and presenting it.

[Tom_videa]

I think it would also help to use sturdier material for 3D printed parts. I personally use polycarbonate for these high stress parts, as PLA is softer and it can wear quite easily.

[HappyMathDad]

I have been trying to make an almost 100% 3d printed planetary gear box with good efficiency. And I've already noticed a huge difference between gear generator softwares. I think that's how youtuber Paul gold got 30 Newton meters of torque out of a smaller stepper with a 3d printed cyclodial gear. Your measurements of a 1 to 3 decrease in performance Is significant. I would love if Paul measured the efficiency too. But beggars can't be choosers. Love the video!!!

[aldobruno]

Many congratulations!

how much was the cost of the CNC parts?

[chrisnatale]

Going through the CAD design for a cycloidal drive was very helpful. Thanks!

[christianedelmann]

I used to work at a company that made those kinds of planar polymer bearing/bushings and they are almost exclusively made to be pressfit into a tight fit bore. Maybe a design iteration would be to get a ground thin walled tube and pressfit the bushes into them. Also those kinds of contact surfaces should be hardened and ground to a fine surface finish. Obviously not super easy to do but better gearbox life

[dekutree]

I tried SF1 bushings at one point too (and HK needle bearings before that), but nowadays I press PTFE tube into brass tube to make low-friction bushings whatever length I want and without any seam. Although it is fairly time consuming cutting and deburring a large number of tubes to precise length, even with a tubing cutter. And not all PTFE tubes are concentric, so it can be trouble finding a good supplier.

[Niels_tenDam]

U can use 3MF file format instead of STL. It can store more info with smaller size.

[eDRoaCH]

this is a really interesting comparison. I wonder what the best of both worlds would be. If you just put the cnc cyclodial disks or maybe other parts in the 3d printed version I wonder if its power/weight ratio would beat out the other 2 alone

[matthewmeans]

after seeing the failure of the threaded inserts at 18:17, I had to look at my own design and question whether the inserts would fail for a 13 Nm application on the output shaft, however in looking back I realized that you only had the torque distributed on two pins for that whole load rather than the whole 6 pins available. This leaves me more hopeful? I'll comment back on this whether I get significant plastic deformation for my version as well.

[wrgates]

I noticed a few times you had issues with the powder coating. You can specify that it should be to drawing dimensions after the coating is applied to avoid oversizing the part and keeping it to your tolerances in the drawing.

[soscilogical]

This is really good because it tells me that it's a really nice design and it's far to many components to be a low cost solution for a budget product, which is good info.

[wenbodong]

"it's a precision drive" then "let's grind it with a dremel" JK lol. Great work man!

[empatikokumalar]

What is the name of the device you used to measure at 0:29?

[Atomy]

Wow that's some crazy 3d printing on those dowel pins and screws

[thomasmackay]

Regarding modeling technique; is there a reason why not a 0 to 2pi/(number lobes) and a radial symmetry by number of lobes? is that to avoid surface modeling lines? coming from a Catia V5 modeling experience. gap with a spline just irks me and probably is creating some odd tribology down stream.

[sceiluug]

i am wondering of there is also a reduction in noise in between them

[stevebabiak]

The housings on these should have some sort of mounting flanges on them; clamping them as was done in this video shows that the housings themselves move up and down, implying that all forces aren’t being transferred to the load.