Это видео недоступно.
Сожалеем об этом.
3D Printed Robot Actuator Finals HaD Prize 2018
HTML-код
- Опубликовано: 20 окт 2018
- / gouldpa00
This is my entry video to the Finals of the 2018 Hackaday Prize.
Details at hackaday.io/pr...
CAD at www.thingivers...
You deserve to be more known! Amazing work
Thanks fabcol. I have only just started sharing my work.
Paul Gould lol, there are RUclips vids of this exact design from 3yrs ago and there is a startup from a few yrs offering modular robotic design.
@@justtestingonce I have seen videos on designs using outer pins. I have not seen designs using having the dual housing and dual output put shafts. I have also not seen designs with magnetic encoders for both motor position and joint position like mine. Can you please add some links, I would like to learn more.
I love this so much! Paul this is incredible!
I just want to say your design is probably the coolest in the open source maker community :)
Thanks. I am working on making it easier to make.
I believe you are right! Great implementation of Solid Mechanics principles.
Awesome design and work! I've also used the AS5161 absolute position rotary encoder from AMS on top of a commercial servo. It works great!
I like the AMS range. They are a bit expensive but really the only good/easy solution.
amazing work! please keep us updated, this is fantastic
Really good build ! Thanks for sharing this video !
Amazing work! Thank you to show us your designs, they are very inspiring!
Maybe one day my cycloidal designs will be as good as your. 3D printing does allow for rapid development and testing of new ideas.
Great work! Awesome to see lots of random talent in Perth :D
one word: SUPERB!
Thank you! Hoping to mill some of these out on my CNC in slightly larger scale for some fun robotics.
What are you making?
this would be perfect for making brushless servos for a CNC as the brushless servos implementation with encoder has no holding torque. I mean any gearbox with repeatability should work
Loving the credits too! Your work is amazing. I recognise the filament box! ;)
Are you from Perth?
@@PaulGouldRobotics Yep! There's actually a local 3D printing community on Facebook: 3D Printing Enthusiasts - Perth & Western Australia. The folks there would love to see your project!
@@whatthej I've just joined the group.
Waw, I'm impressed ! Excellent video !
Thanks, more robotic quad stuff soon.
@@PaulGouldRobotics Hmmm, I love it !
Ｉlove this project! Thank you!
Really enjoyed this, please post more tests and that. Thanks. Regards
so this guy can basically summon a robot army from his basement and destroy you irl.
Maybe. I started a Terminator arm but all it ended up doing was playing catch with my kids. ruclips.net/video/_P160S7XQl8/видео.html
Wow... The gear system has my attention.
Can you show and explain it in absolute detail?
Please see some of my other videos, go to my hackaday page, or see my Thingiverse page. Everything is open source, including mechanical, electronics and software.
Thank u for sharing great ideas!
Very nice! Great work.
I really really want to see an update video on the quad!! :D
I'm working on it.
@@PaulGouldRobotics im super excited! Was hoping this wasnt a project that was posted about and then stopped. Good luck with it!
hi, Why don't you use the sensors of the motor instead of the as5147
Most High pole count motor that are cheap and used the drones/quadcopters do not have inbuilt hall effect sensors. Hall sensors only give 6 counts per electrical phase. This makes holding torque very jumpy at low/no speed. AS5147 gives +100 counts which make for a very smooth holding torque and also works nicely with Sinusoidal commutation. (I'm working on FOC).
Damn, I was planning to do almost the same. The difference is that I was going to put the motor inside the assembly attaching the eccentric disc to its body, not the shaft.
I would like to do this also but I could not find a motor small enough. Do you have any motors in mind?
Looks like the response was deleted by spam-filter...
thank u verymuch bro.
This is the same mechanism used for Exo skeletons right?
I'm not sure. I haven't seen the inside of a an exoskeletons joints.
thats like a harmonic drive gearbox like we use it in our ABB robots. quite impressive that you pulled off 3d printing one, whats the backlash on these?
I have made about 8 of these and it varies between none and 0.5mm @200mm depending on how my cheap 3D printed behaves that day. "Wear-In" is also a problem. This is fine for a Quadruped but would not be good for some requiring precision. So I came up with ruclips.net/video/2fDbMWHQbC8/видео.html The torque loading is lower but the backlash is always none.
Is it possible for you make these actuators in a small size for a smaller robot ?.
The smallest I've made them is about 45mm dia and 40mm high. Smaller then that and servos might be better. Also going smaller means that precision needed to be higher and 3D printed parts are not that good.
Nice gear. Grate job on the reposition to the last state after force it to be bended , if it has a force feeback would be perfect!
Sorry for my bad english
The Series elastic Actuator (leg) version has spring deflection measurement. See the middle ribbon cable (and hidden sensor) at 2:23 of the video
Amazing work thanks for sharing !
After i saw your work i instently subscribbed :D
Thanks, What robotics are you into?
@@PaulGouldRobotics I have started learning Arduinos
Great actuator, ingenious way to get positional data back from a BLDC motor. Is there an upper bound value in terms of motor KV or is the system’s design proportional to the KV? Any results as to the longevity of this system? I’m assuming those 3D printed parts, specifically the gears, take some serious wear and tear.
Yeah, 3d printed gears do really well in low load technique, but can see terrible wear once friction is involved. I'm going to try making this with both a hardened aluminum rotor and a tpu cartilage sheath. I think this is a brilliant idea.
this video was enough to subscribe
Awesome to see some progress in this project.
And I wish you good luck with that entry, it is really good project.
I hope you don't mind if I adapt those for a robot arm, I always wanted to build one just for the lols.
With your gearbox combo that thing could actually be useful, sadly it has to wait rl is busy right now.
Hi Tim, This is an open source project. Please share, make and improve. I have CADed up an arm. hackaday.io/project/157812-3d-printed-robot-actuator/log/147660-robot-arm-design Also check out some of the robots arms at "gear down for what"
awesome work. Very pleasing to see your different gearboxes laid out as I also have an assortment of abandon gearboxes and motors. Curious did you have any success with a dual stage cycloid gearbox? Thanks so much for posting this. Not all heros wear capes
I have the dual stage cycloidal gearbox working well now. I will create a video soon.
which software you are using on the laptop for controlling the actuator?
It is a custom .Net app Check out github.com/gouldpa/Inverse-Kinematics
Amazing design!
Продолжайте пожалуйста, очень интересно!
That's an absolutely brilliant execution of an underappreciated technique. Now we just need to make it quieter. Mind if I borrow the concept for a project I'm working on?
Thanks, it's open source. It is meant to be copied, modified and improved. Replacing the outer pins with bearings will make it quieter but the size and cost will go up. What are you making?
@@PaulGouldRobotics nothing at the moment, but I was thinking about attempting a mini industrial scale arm bot. I've designed and built a few 3d printers, and some other bits, so I feel confident that I can master the control software. I paint cars regularly, so I thought it would be cool if I could automate a custom paint job. I think training the bot would probably be the hardest part.
@@PaulGouldRobotics couldn't the pins be mounted in mini scale bearings for the same effect? If I'm not mistaken, the noise is from irregularities in the print sliding over the smooth surface. Thereby creating vibration.
Do you have any info on the previous gearboxes you tried out? Im particularly interested in the magnetic and harmonic ones.
I will post some assembly videos and the CAD files in the next few weeks of several magnetic gearboxes and one failed harmonic drive. Check out magneticgearing.blogspot.com/ for external research papers. Search Thingiverse.com for harmonic drives. They are way better than mine.
Could you tell more about your experience with other drives, planetary harmonic and magnetic? What are the differences and why did you settle with cycloid one? I guess planetary has backlash and magnetic are expensive, but what about harmonic?
Hi osmanpasha@home
That is a difficult and complex question to answer. I'll add something to my Hackaday page in the next few days.
@@PaulGouldRobotics Thanks!
that is really good. approx what price are the components for an actuator?
Great project! What is the reduction ratio of you gearbox?
Hello, I want to build one of these amazing actuators just for learning purposes, I’m very amazed on how this so simple gearbox works, so, do you have a list of parts (the hardware I can not print) needed to build one of these actuators ?
Check out the built instructions at hackaday.io/project/157812-3d-printed-robot-actuator
There are simpler versions on my thingiverse and hackaday pages
Awesome job!
Very, very cool!
Any text reference in order to determine the reduction ratio of BLDC.
like Those using Big Turnigy motors usually go for 6:1
but for anti gravity type its 9:1
I want to know actual relationship to this
It really depends on you application. If you tell about your project, I can suggest options for you. ODrive have some good docs on BLDC motors.
@@PaulGouldRobotics idk why comments are not updating here please share your contact
@@shashankshekhar6549 instagram.com/gouldpa00/ or hackaday.io/paulgould
Very nice work! Have you measured the backlash of the cycloidal gearbox? Also what material did you print the gearbox parts in?
The test actuator has 0.5mm of backlash at 70mm from the pivot = ~0.4degs (it has half the outer pins installed). The Hip Actuator has about 1mm of backlash @ 70mm or 0.8deg (with only half the outer pins and half the inner pin installed). The knee actuator has less than 0.1mm of backlash @ 70mm or 0.1deg (half the outer pins). It mostly depends on the precision and settings of my cheap 3D printer. All of the 3D printed parts are cheap ABS. The pins are brass on Silver steel rods. The order of development went hip actuator, test actuator then knee actuator. This means that I am getting improving as I go. I will have to try Nylon next (after a hot-end upgrade). Once I get one leg "right", I will make the rest.
Thank you, Paul. I think a lot of people can use this in their project. You are getting quite low backlash considering the manufacturing method. I agree with nylon. What gear ratio are you showing the video?
They are working way better than i thought they would. The ratio is 26:1 for the big actuator (test, hip and legs) and 21:1 for the small one
hope you win!!
Where is this project at now? I'd love to be a part of your project.
Hi Alan, I'm working on different types of cycloidal gearboxes and a simpler/cheaper FOC brushless motor controller. This design is all open source. What would you like to on this project? There are already about half a dozen people (that I know off) making variations of this gearbox design. Many have made major improvements.
So did you eventually get the quad to walk and jump?
I'm still working on it, while making improvements as I go.
Great Job! What's the gear reduction ratio for the one shown?
25:1
Wow dude, nice work, u give me the inspiration for back to my quad bot, are you think psoc 5 can do all? With no driver?
Please make a quad too. It would be nice in a years time we can have robot quad races and challenges. The FET driver creates high current/fast gate signals for the FETs as well as a charge pump for driving the high side N-FETs. An all-in-one solution could be using the STSPIN32. I have started a design for this but the CPU is limited. I'm looking at the SAMD21 (Arduino Zero) with a FET driver. Check out hackaday.io/project/165217-arduino-brushless-actuator-foc-controller
Достойная Работа !!!
Inspiring work, This is precisely what I needed thanks a lot...
What are you making?
Is there any way to make the torque output higher? I’m working on a project that needs to lift a arm similar to your weighted arm test, showing the rough torque output.
I have limited the Motor power to about 60W peak for the tests. The motor "should" be able to go 10x that for short bursts. I will push the actuator until it breaks in the next few weeks, after I have made a few more. If you need more torque and less speed you could always increase the reduction ratio.
@@PaulGouldRobotics Really looking forward to a followup on this. It really looks like the best low weight, low backlash, small, high reduction transmission for robotics. What is wear and tear like, how much power is lost on friction, at what kind of forces does it break? Amazing work!
Wowwww super cool!
Great work man!
What sort of efficiency did you get, haven't been able to exceed 80% with my design
I don't know how to do an efficiency test. If I am above 60% (with my cheap 3D printer), I would be happy. Please go to hackaday and send me some images of your gearbox.
@@PaulGouldRobotics you measure the torque output and divide it by the theoretical torque output given by your motor torque multiplied by the transmission ratio
@@jamesquinn6662 That would give me static torque. I would be more interested in dynamic torque. Maybe I need to make a dynamo? At this point the actuator creates more torque than I need and uses less power than I though it would. So, I'm happy with it. If I am aiming at high efficiency I would not use 3D printed ABS parts and futher develop the SEA (Serial Elastic Actuator).
Dude. Amazing.
Nice video, didn't understand the mechanics, think I need the lay mans description...
It is a 1 to 25 reduction cycloidal drive (gearbox) modified to suit 3D printing couple with a large drone brushless motor.
@@PaulGouldRobotics Thanks, gives me something to google with :-) Does yours work like the animation in wikipedia - en.wikipedia.org/wiki/Cycloidal_drive, quite ingenious.
you are the best!!
Oh my, it will precisely wiggle its tail!
Check out my tail video, it has a gripper.
you need a pid in there so descend is the same speed as ascend, but otherwise awesome.
I have the current/power limit set very low for these tests, only 60W from a 600W motor. The system has some PID position control but also need velocity control too.
Paul Gould sure you have seen it already, but just in case ruclips.net/video/WJkf264yxTU/видео.html
I like James Bruton's robots. He makes some good robots fairly quickly.
Paul Gould generally amazing what all of you are doing, good luck in the competition.
@@PaulGouldRobotics James should definitely use your actuators for his "big dog" project instead of using lead screws. Why not connecting with him ?
Very cool!
Good job my colleague!
I like it very much!!
Great video.
Anyone else having volume issues? Volume maxed out on my phone and can't hear much.
It's not your phone. RUclips shows that it is -15.1dB so it is no surprise that you can't hear it. Hate to be negative but he seems to be whispering and I had to pause and rewind to see the pictures that just flashed by. Interesting project but poor presentation. IMHO.
Sorry about the sound. I know how to make robots and struggle a with the videos. Will try harder next time.
@@PaulGouldRobotics No worries. I thought maybe it was something on my side, so was just asking if it was the same for everyone. I was able to hear it later at my desktop.
I want to know what Mike had to say!
He suggested that it was possible to put the motor shaft, input housing and output housing all on the same side as a series of concentric bearings. Very clever. This means that the input housing and output housing could be dual sided. The motor shaft could go all the way though and the eccentric shaft be supported on both ends.
@@PaulGouldRobotics I see. This makes it possible to create durable joint by 3D printing! Nice, thanks!
It would be an idea, use hard disk motors
Hard Disk motors run at high speed but have very low torque. Low KV drone motors are cheap and powerful.
And also "damn you like aluminium profiles so much!"
I draw lots of them as part of my day job. They are clean, strong, light, cheap and available in different profiles.
Shhhhh. Dont wake up the wife... Haha that's what I'm imagining. Great work !
Sorry for the poor audio. I've got a new mic now, hopefully the next video will be better. And remember I make robots (when the kids are asleep), not RUclips videos :-)
i love it
Bravo !
Hi, have you followed up on this?
I have been working on a all 3D printed version with no pins. I am in the process of designing the electronics to be cheaper. One based on Arduino and one on ST-Spin. My plan is for makers to create each actuator for under $US100, including motor, gearbox, electronics and position encoders/sensors. I have been working on a tail. hackaday.io/project/163766-low-cost-3d-printed-robot-tail
@@PaulGouldRobotics Hi Paul, Thanks for your response. The progress is looking great! If you need with anything, give me a shout.
Wow that is soper thing
boston dynamics, you are here?
I would like to see some quadruped races or obstacle courses in the near future. There are a number of people making them.
@@PaulGouldRobotics good luck, bro. that work very impressive!
Heh, I wonder how durable exoskeleton it would be possible to make with a plastic printer
I'm not sure, but it is low quality plastic (ABS), same as Lego. The printed parts are not that consistent. I would say that you would want to trust an exoskeleton with your life. Also the power to size/weight is not that great with 3D printed plastic. Metal is much better.
@@PaulGouldRobotics Sure, but it still would be interesting to see how far these can be pushed, especially if stronger filament is used.
Even if it wouldn't be as useful as metallic version, it would still be awesome as a home project!
@@normoloid Ok, so for a hip joint you could make the diameter about 150-200mm, thickness about 45mm, place the motor inside and make it out of nylon with a higher reduction ratio. I will CAD it up and see what bearings I need.
@@PaulGouldRobotics Also, with a little designing of the "bones" it should be possible to house batteries in them, altough then I would probably put the batteries inside a metal tube fitted inside the plastic, which would make the whole thing a little bit stiffer than with just plastic.
Why put a huge battery pack on the back when you can just as well integrate atleast part of them inside the structure itself, making the whole thing much more balanced package. :)
how could u dislike this lol
It feels like the bear motor
Can you please tell me what is a bear motor?
@@PaulGouldRobotics ruclips.net/video/JgeoR5utVUw/видео.html
Btw i don't really know anything i just sayed that because it just looked similar '-'
Прикольно
ᵂʰʸ ᵃʳᵉ ʷᵉ ʷʰᶦˢᵖᵉʳᶦⁿᵍˀ
I make robot stuff, not RUclips videos.
WwwwoooOOOww
+
aaa
Wow... If you want a business partner, then partner hit me up! Like seriously let me know.
I hadn't planned on making money out of this. It is just nice to share ideas.
@@PaulGouldRobotics For real, no point in doin anything if isn't fun but for what its worth if you ever find yourself sharing ideas in Tokyo or in a few months Oregon let me know.
@@petrich123 OK