3D Printed Cycloidal Actuator
HTML-код
- Опубликовано: 17 фев 2021
- I have designed yet another cycloidal drive! This one is different, though. Unlike most of my projects, it is strictly 3D printed. No CNC required. The cycloidal mechanism is also "inside-out". The cycloidal discs do not rotate, they only wobble, transmitting rotation to the output. This was also designed to maybe be used in a modular robot arm. It's also a 25:1 gear reduction.
Gain access to project files and support my work directly - / levijanssen
Follow me on Instagram - / leviajanssen
NEMA 17 Stepper Motor - amzn.to/3bnO9qn
Large Outer Bearings - amzn.to/3bkebL0
Eccentrically Mounted Bearings - amzn.to/3sh0fbF
Disclaimer: I may earn an affiliate commission when items are bought through this link. 
Наука
I tried to “cut the fat” for this video. More time went into editing, and I actually kinda enjoyed the process! Hopefully it made it more enjoyable to watch. Thoughts on a robotics arm using something like this design?
Would love to see some section analysis or parts coloring in Fusion, to really make clear the mechanics. But you struck a good balance here imo!
Maybe for orienting a small solar panel also using a second one? Or a mirror which shines sunlight indoors ;)
It does make the video better! Your older material about coilgun design were a bit too verbose, but this vide was much faster and more engaging.
I'd be interested in seeing a SCARA plotter with this design!
Absolutely love it~! My first thought was you could add some gas support struts/springs to the side between the two arms to help take up some of those forces while it's extended out? Just a thought~ Keep up the amazing work man!
Very nice quality in content, video, audio and editing!
Bucket of water is a fairly accurate weight. If you don't have a scale, you can measure volume and use density to get a weight.
But wouldn't you have a problem with, that if the arm isn't completely steady the water would 'splash' around in the bucket
I cant imagine doing this in imperial units
Remember to account for temperature.
@@Philip_J Bottles FULL of water taped together.
@@aeiou75 How much does that matter between 1-99 degrees?
Very cool design, and I love how you scrolled through the whole code to explain it as you went. Also, totally interested in seeing a robot arm made of these!
This is awesome! I'm keeping my fingers crossed that this stays exciting enough for you that you keep devoloping it into something even more awesome and useful!
This is just awesome!
I really appreciate your extremely well done videos and your objective way of looking at your own creations! Go on mate! Cheers!
Awesome work , I'm looking to finish a five year project of my own and you have definitely helped me to understand several things that I need to take care of . You gained a subscriber .
This is so cool! Goes to show how much someone with a creative mind and relatively cheap tools can make...
Actually the design was copied fro Paul Gould
ruclips.net/video/puZUKE_Wu8g/видео.html
@@anonymouscoward9459 completely different design. all that it has in common is that they're both cycloidal drives.
dude! awesome work! looking forward, would like to see the next stage!! :)
The design of this thing is great! Amazing work!
Hell yeah! the Monoprice mini V2 my favorite of my 3d printers.
Great project. Thanks for converting to metric units as well 👍
Superb project! I had the same problems with squashed circles before I've tightened elastic belts of the printer.
Just discovered your channel. Informative and well made which makes this an awesome video! Good job and hope seeing more 👍
Nice one!!! 😊
I did an inside out version of a Cycloidal Drive with a BLDC... it reduces noise and weight, while increasing torque and speed...but at the cost of needing an extra encoder and expensive mcu, so using available and cost effective stepper motors is of course a very good choice in this scenario too!!
Great job. Yes make the robotic arm as that would just be fun to see and for you to do. A new printer even a cheep one sounds like a good idea.
I find the cycloidal drives extremely interesting :D
You should look at Paul Goulds cycloidal design , it's what Levi has copied
ruclips.net/video/FplpyZD1i1M/видео.html
Awesome Design Levi!
Thanks for sharing! Cool project!!!
This is incredible, subbed, thank you
Well done Levi!
Awesome! I've gone through a few iterations of a harmonic drive but I have not been happy enough with it to publish yet.
Really cool! I´d definetly like to see a finished arm using those actuators!
Super cool! Robotic arm would be fun.
I have the same problem with my MP Mini. On my MP Ultimate (Wanhauo Duplicator 6) circles are perfect. What a great project, cant wait to see what you do next with that.
I like your experimental & fun approach to designing new things. Your 3D printer is a nice tool to have, allowing you to _"extrude ideas from your brain 🧠"_ 🤔
I'd give a pankake stepper a try... would make the build more streamlined. Brilliant idea by the way, looks very promising!!
There's no proper word for the amount of joy that came up when i saw that this was printed on an MPSM
Cool stuff 😎
wow, the quality of the video is much better :D
Great work
Nice job
That's a great design. I'd like to see you do a whole arm.
Going forward you may want to check Trinamic motor drivers, they drive motors with much more control that h bridges, and you they have evaluation boards that are affordable. Amazing work by the way!
Seconded.
Can't wait to try this on my resin printer
really interested in the refined open source version!
Been a while since I've seen somebody use one of those dual H bridge drivers on a stepper, had almost forgot there was an alternative to the little stepper drive ICs, especially the Trinamic magic ones.
Wow, crazy that it's back drivable! I mean, not super smooth, but not talking worm gears.
The OVAL PROBLEM in a 3D printer occurs because of one of the belts being stretched permanently. I faced the same issue b4, got it right by using a new belt
+1 for this, also the belt tensioners on the MP Select Minis aren't great since they're literally springs and can flex. I'd check out the USWaterRockets 6mm belt tensioners on Thingiverse.
Nicely done. Cycloidal gears are something I’ve always found interesting.
If you wanted more precision and torque, you could use a worm gear system in a similar space.
a worm gear is a non-ideal solution for a robot actuator, because there's a lot of play or friction, and worm gears are necessarily non-backdrivable
I'm happy I just discovered this channel.
Maybe a resin UV printer would make more accurate gears for this?
If you tie a string on the end of your square tube and on the end of that tie a empty milk jug, you can apply a range of test torques by adding measured amounts of water to the jug.
Nice information, thanks :)
very cool
Nice work - I have a similar project underway
Nice project man! Although, don't get mad, the final result looks cooler than it performs: lots of play and it doesn't actually work like a servo (it is not positioning aware)
T'es un génie 👍bravo 👏
Thanks for the metric torque!
Everyone's ragging on me for even thinking to mention imperial, I appreciate the thanks for the metric!
6:45 what I did is buy a 8l water bottle, with that you can just fill it with water and use a kitchen scale to get the exact weight you need, its great for testing torque capacity at any possible weight
I have a mini v1 and was able to fix an oval printing issue by tightening the belts. You should also ensure that the stepping values in the firmware are the factory settings. If they are not your scaling will not be consistent across all sizes of prints.
You can scale the 3D model in the slicer by a very small amount (e.g., 1%), independently for X, Y, and Z.
Absolutely do more
i think you could attach the potent to someone arm and be a sort of exoskeleton , maybe to help those having minimal arm issues
You should get a smaller glass bed for your printer. (Or even better, a PEI flexplate) That extra weight is sure to introduce more ringing than necessary.
Perfect 👏
Wait a second you're not Tom Stanton
Amazing work! However do not confuse a potentiometer with a cheap trimmer. Trimmers (like your blue one) are supposed to be used just for infrequent adjustments.
nice
Hmm, maybe having seperate drive and position detection motors would be a good next step, to minimize the effects of stress and strain on the little plastic bits.
Good job,thank you for sharing ,if you cosider practicality/ease/aesthetic you get good grade for it ,need to work on precision/backlash
Cool project! May I ask how you expanded the build volume of your printer? Did you follow any guides or anything? I also have an mp select mini I’d like to see if I can upgrade.
5:06 [insert me when X -joke here]
Cool.
Neat, looks awesome. Some of the play could very well be due to tolerances between the metal tubing and the screws. Threading on screws should never be used to center a smooth hole, only to clamp stuff together by lock into other threads. Maybe try shoulder bolts?
On this note, another option might be to use layered concentricity by fastening the bolts in question into threaded holes or through threaded inserts then surrounding those inserts or some form of hole posts with the hollow rod you decided to use for its stiffness and durability.
Check different supplyers for printer stepper drivers, sometimes ( in 5 packs ) they can be cheaper than dual H-Bridge and cuts code to pulses ( steps ) and direction outputs making it easy instead of having a step table.
Is there any industrial part/mechanism that uses this reverse (inside out) form?
GREAT!!
Amazing design!
CNC'd metal parts and a stronger stepper and I think you have reasonably compact and robust design.
That's very cool! where did you get those bearings?
Zig-zag a chain of them paired in a scissor formation all receiving the same signal acting a massive linear actuator.
It seems to me those inner cycloidal gears could be rotating slightly and that might be producing most of the play in the mechanism.
2:00 Here you were talking about the pwm signal caused by the stall torque or something. I'm currently doing a project and I am trying to find the best way of detecting the stall in a simple nema17 stepper motor. Do you have any suggestions what the best way of doing this would be?
There is some existing solutions for something like this, if I didn't misunderstand you. It's a small circuit board that attaches to the back of the stepper motor, and senses the motor position with a Hall sensor. Google "closed loop stepper motor" :)
@@OMGWTFBBQSHEEP yes I know those kinds of closed feedback loop stepper motors. Tho I can't seem to find any information stating that you can actually detect the stall instead of compensating for it. Do these boards have pins which allow you to check for stall?
Cycloidal drives are cool, but you have to know it's the reason you can't move the arm anywhere near the max holding torque. By the time you get an actuator on the end of the arm, you'll be close to your 2.5 lb limit. Increasing the teeth per inch will help a little, but a planetary gear will be needed at the shoulder.
Backdriving is actually not a good thing for robot arm. At least not for most applications. The idea is that you should have big gear ratio so you would get hefty torque and no backdriving so if you cut the power the robot will not collapse.
Great video anyway!
That's not great advice actually. In autonomous systems having the compliance and observability of back driveable arms and legs adds a great deal of useful information about how the device is interacting with the world. For instance if you make a walker that is super rigid it struggles to remain stable on uneven ground because it can't adapt to early and late ground strikes. If the system is back driveable you can use torque monitoring to sense contact points and adapt to uneven terrain. The ability to back drive allows for interactive, adaptive robotics
nice project, will CAD files be public?
No he said you have to pay for it on patreon
hello, definitively interested in a full robotic arm. before changing your printer better tuneup your pulse per mm/inch to get real circle. a second printer is always handy also...
Very impressive. Well done.
The robot/motor community is so divisive about which units to use for torque. I believe that SI units are superior to imperial units, but kg cm make the most sense in the usual armature lengths we deal with rather than meters and usually thinking of loads in terms of weight rather than newtons. Anyways, team kg*cm ftw
I am curious to see a robotic arm made from this cycloidal actuator. Can it be used for precision work like motion control, welding etc.
I wonder if you'll consider to show in detail the procedure to mount & connect that L298 have the same one but never managed to get it to work like you.... all or any help you give would be greatly appreciated....
Great job !
Looks like a perfect chicken coop door opener/closer. 🤔
Looks great ,
Have you checked out Paul Gould's videos ?
They looks very similar in design , perhaps a collab ?
8:05 You need your printer tuned in really precise for minimal gaps.
Yes, the ovals are probably from your x and y steps not being calibrated. Search for e step calibration
Yes!
Hello. Could you please explain, what are the advantages of those BLDC actuators in small robot arms, compared to RC servos? Conventional servos seem to have higher power-to-weight ratio, while also being cheaper.
This is not cycloidal or harmonic. This in literature is called the Circular wave drive. There are lots of information in literature and academia about it.
transparent filament would be great :)
Transparent FDM filament is usually more like opaque.. Meaning it lets some light trough but you can't really see trough it :/
I have built a robotic arm with a similar cycloidal drive design, but one of the significant shortcomings of this mechanism is it cannot transfer a higher amount of torque, as shown in the video; when you try to move at higher loads, it starts slipping in the same way. @LEVI Janssen, do you have any solution for overcoming this slipping issue? It cannot be done by using a motor of higher torque as the current motor I am using has high torque, but the gearbox starts skipping at half of the motor's rated torque.
Seems like a great place to start from for a cheap-ish welding positioner??
How about a worm gear? It would make the system a while lot stable when the worm on the motor drives the arm. The printed worm may collapse but a aluminium would be a good option.
nice on
you just need some stuff like end stops , also maybe get a better stepper driver and voila :)
Will you share your design? Thx
Great work
what was the ratio used here , i need to reduce the gear profile for 1:10 ratio with high torque nearly 1kg force
Ewww freedom units :(
Nice job man, very cool. Do you buy your bearings in the US or bulk order them from China or something? I haven't used ball bearings in years.
I buy them off of Amazon. They have a surprisingly wide range of sizes.
Hi guy, its ok? I have a question, how do you make the redution? What calculo did you do? I have a motor nema 17, what the better redution for this? Thank you man.
Good work friend..
Could you please shared the dimention of CAD drawing
Great design Levi ✅✅✅
I like the design a lot, particularly the holding the gears from turning and rotating the case instead.
So I had a couple of thoughts that may or may not be feasible.:
1) to eliminate the bulk of the stepper hanging out the side, is it possible to mount the motor beside the gear box in the fixed arm and belt drive from there to the input of the gearbox?
2) expanding from 1. use a BLDC motor with a small toothed pully on the motor and a larger one closer to the diameter of the gearbox on the gearbox input.
BLDC motors are lighter weight and quite often more powerful than stepper motors, I think that would make it a lot faster and reduce some of the noise a bit. :-)
Great project, looking forward to the next one.
It is pretty compact. what is the equivalent gear ratio?