Cycloidal Speed Reducer in OpenSCAD

by mattmoses, published

Cycloidal Speed Reducer in OpenSCAD by mattmoses May 8, 2011

Featured Thing!


By request of Syvwlch ( http://www.thingiverse.com/syvwlch ) and WilliamAAdams ( http://www.thingiverse.com/WilliamAAdams ), here is a stand-alone public-domain OpenSCAD cycloidal speed reducer. As with the Wankel Engine and Roots Blower I recently posted, this is intended more as an example of an interesting mechanism than as a practical device. If you want a practical printable speed reducer, you might consider one of the other alternatives like

the worm drives on this Tank http://www.thingiverse.com/thing:8080 or differential planetary gears http://www.thingiverse.com/thing:7390 or cascaded spur gears http://www.thingiverse.com/thing:7379 or this planetary gear reducer http://www.thingiverse.com/thing:8460

There are several cycloidal-type mechanisms already on Thingiverse, such as http://www.thingiverse.com/thing:3617 and http://www.thingiverse.com/thing:3736

There are also several interesting external sites like: http://www.zincland.com/hypocycloid/ http://fabricationsofthemind.com/2010/07/09/extruder-design-1-printable-1001-hypocycloidal-gearbox/ https://github.com/triffid/Differential_Hypocycloid http://reprap.org/wiki/Differential_Hypocycloid http://en.wikipedia.org/wiki/Gerotor http://en.wikipedia.org/wiki/Gear_pump

and many many interesting youtube videos such as http://www.youtube.com/watch?v=bRn1K2XeWVE http://www.youtube.com/watch?v=3WvPF6uGCq4 http://www.youtube.com/watch?v=CG2sPuqEXBg http://www.youtube.com/watch?v=AMtyFwMDL7w http://www.youtube.com/watch?v=h236SP86nnQ

This present script is based on a design by M.F. Hill described in his 1928 patent "Internal Rotor", number 1,682,563:


Note that this design is based on an offset hypocycloid, similar to Figure I in Hill's patent. Most of the contemporary designs appear to be based on an offset epicycloid, more closely resembling Figure V in the patent.

The motivated student can modify the code so it generates epicycloidal-based profiles. Hint: start by making a module ``epitrochoidBandFast(n, r, thickness, r_off)". The motivated student could also probably clean up my train-wreck of code and/or figure out how to do arrays in OpenSCAD.

Note also that these rotors can be used for pumps - see the gifs in the comments for an example.

Recent Comments

view all
Does anyone have actual solid part files for a cycloidal speed reducer?
Yep: :)

myPoints = [[0,0], [0, 10], [10, 10], [10, 0], [0, 0]];
myPaths = [[1, 2, 3, 4, 5]];

polygon(points=myPoints, paths=myPaths, convexity=10);
theres a slight optical illusion if you stare at that last one long enough...


Liked By

view all

Give a Shout Out

If you print this Thing and display it in public proudly give attribution by printing and displaying this tag. Print Thing Tag


OpenSCAD for animate!
STLs for print!
DXF for cut!

The minimum clearance around shafts is about 0.6mm. If printed, there might be a little interference at the tips of the internal rotor lobes. The DXFs have the same profile as the STLs, so the beam kerf should provide ample clearance if anyone actually lasercuts it. There are some extra holes in the DXF profiles so pieces can be aligned and bolted/screwed together.

Yet-another-Note: The DXF files were generated directly from OpenSCAD. I am not sure why they preview as having dotted lines in some places, but when viewed with say, Open Office Draw, the lines appear solid.

<p>Does anyone have actual solid part files for a cycloidal speed reducer?</p>

<p>Very good thing! </p>

<p>I guess </p>

<p>cycloid eccentric.stl </p>

<p>and </p>

<p>cycloid driven shaft.stl </p>

<p>are the weakest points in it. </p>

<p>Did you try to either taper these shafts and their holes for added strength </p>

<p>or allow to replace the shafts by bolts? </p>

<p>Looks like it could transfer much more force that way. </p>

<p>(And with 8:1 I guess you get a lot of force.)</p>

<p>Tapering or replacing by bolts is a good idea, but I have not added anything like that. There are probably a number of practical problems that would need to be solved if someone wanted to use this in a real application. It would also probably be necessary to add some type of ball or roller bearing to the eccentric to reduce friction. The pin extrusions on the driven shaft may also have to be replaced with rollers.</p>

<p>Awesome! :)</p>

<p>You can do one dimensional numeric arrays like so:</p>

<p> myArray = [1, 2, 3];<br> for (i = [0:2]) <br> {<br> echo("", myArray[i]);<br> }</p>

<p>Arrays are zero indexed.</p>

<p>Cool, thanks! That will help clean up the code. Do you know if there is a way to pass a generic array of points to polygon()?</p>

<p>Yep: :)</p>

<p>myPoints = [[0,0], [0, 10], [10, 10], [10, 0], [0, 0]];<br>myPaths = [[1, 2, 3, 4, 5]];</p>

<p>linear_extrude(height=10)<br>polygon(points=myPoints, paths=myPaths, convexity=10);</p>

<p>thanks so much for the work. Being the OpenScad guy that I am, there's plenty for me to learn. And these will make for good desktop curiosities.</p>


<p>Thank you, thank you, thank you. :-D</p>

<p>And yes, arrays would go a long towards making OpenSCAD scripts for stuff like this less... uh... idiosyncratic.</p>

<p>I've been playing with it and it is fast!</p>

<p>Great work and great documentation.</p>

<p>Here are three examples of different mechanisms that can be created with the OpenSCAD script.</p>

<p>theres a slight optical illusion if you stare at that last one long enough...</p>