• Hi - We're having intermittent forum issues. Posts, alerts, and anything involving email like signing up or password changes are having problems but not for everyone or everywhere. Josh is working on it. No ETA right now. Thanks Josh - BTW this is a volunteer forum so SLA's are just best effort. EDIT -> I manually batch updated about 25 people stuck on the email notification. Try logging in now with your password. EDIT -> May 8/23. I updated another bunch of users stuck on email notifications. Try to login if you are stuck. If you really can't get in contact us on facebook or if you know a forum member get them to ask us.
  • [Ad-Free Experience]
    Register Today, Craft a Post, and Enjoy an No-Advertising Experience.
    Click Here to Register

Cycloidal drive design & testing

Arbutus

Active Member
Premium Member
A cycloidal drive is a type of high reduction gearing used where compactness, high ratio and zero backlash are needed. These are used extensively in robotics systems such as manipulator arms, indexing devices and precision tables.

I have been working on a design to fit small stepper motors - NEMA 17 and 23, for some time now. The mechanics are deceptively simple, but producing a drive which operates smoothly under all conditions and design loadings is proving to be a real challenge.

3D printed designs are all over the internet. I doubt that any of them will work for more than a few minutes in reality. After many attempts, I have concluded that fused filament techniques and materials have insufficient strength, minimal wear resistance and high deformations. My attempts with Titan-X, ABS/Carbon fibre and nylon all taught some lessons, through various failure mechanisms.

Even though the drives are powered by small stepper motors, the internal torque and contact pressures are extreme and easily deform most plastics. Not all parts need to be hard and solid. The casing is not subject to internal loading and can be designed to use engineering plastics such as Delrin. The cycloidal gear could also be made from Delrin if it is thick enough to resist the in-plane bending forces, but 6061 or even steel would be better. The drive pins however must be rigid and hard, capable of handling the extreme pressures at the contact with the cycloidal disk. And of course the entire system must be rigid and reliable when mounted in service.

This design is for a NEMA 17 stepper.
Cycloidal drive exploded annotated.jpg


Some images of the latest iteration, made from Delrin. The drive pins were made from Teflon to reduce friction. Approximatey 75mmx75mm x 20mm thick.

Cycloidal drive 48.jpg

After an hour of running, the drive pins have been crushed and the system has failed.

Drive plate.jpg


Stay tuned....


Don
 

Attachments

  • Cycloidal drive exploded annotated.jpg
    Cycloidal drive exploded annotated.jpg
    269.1 KB · Views: 17
Last edited:

Xyphota

Super User
Premium Member
Thats really neat! I have been researching cycloidal drives a bit recently. I see some designs have the parts you've labeled as contact pins given the freedom to roll, and using a second cycloidal disc 180 degrees out of phase of the first. What are your thoughts on those?
 

Arbutus

Active Member
Premium Member
Thats really neat! I have been researching cycloidal drives a bit recently. I see some designs have the parts you've labeled as contact pins given the freedom to roll, and using a second cycloidal disc 180 degrees out of phase of the first. What are your thoughts on those?
Yep, the contact pins are steel 3x10mm that sit in a 3mm hole in the casing. They are not fixed in place and could rotate. There is some slop in the hole (drilled 3.00 and pins 3.00) but the slop seems to come from the elasticity of the material when under pressure, not hole tolerance.

I'm working on a second design which allows for a 2-phase arrangement, but until the fundamental design is reliable, I'll avoid 2-phase. The real benefit to the 2-phase design is to smooth out the minor velocity variations due to the 5 drive pins. If there were say 11 pins, the velocity variation is much smaller. But, more pins = more diameter and that's another compromise to consider.

Further, this particular design which started from a posting by White Owl, is based on a 48:1 ratio. Geometry defines the eccentricity which in this case is just 0.5mm. The next iteration will be 25:1 ratio with larger contact pins, a 1mm eccentricity and 7 output pins which may produce a smoother, stronger drive.

Don
 

Xyphota

Super User
Premium Member
From my naive perspective, would a second cycloidal disc not be more reliable? The design looks to be a bit more complicated (perhaps it is way more complicated LOL), but a second disc seems to me like it would roughly cut the contact pressure in half.
 

Arbutus

Active Member
Premium Member
From my naive perspective, would a second cycloidal disc not be more reliable? The design looks to be a bit more complicated (perhaps it is way more complicated LOL), but a second disc seems to me like it would roughly cut the contact pressure in half.
Thats a valid point. The pressures are significant.
 
Top