Integrating my Harmonic Drive - 3D Printed 6-Axis Robot Arm

well i asked and you guys answered sort of you want to see my robot arm progress but you also want to see my magnetic levitation platform so obviously i'm going to do both but i decided to go with the base of the robot arm first because i thought this would be the easier project to finish up if only i knew anywho let's jump right into it i'll be working on the mechanical design of the base of the arm first which includes the rotary platform and the first hinge joint which will be powered by that harmonic drive i developed in a previous video also briefly look at some harmonic drive components printed with a marked forage machine and finally i'll be building up this a built from scratch motherboard which will eventually control the entire robot arm it uses trinamic stepper drivers has tons of feature in io support and should definitely be super quick to get up and running but first things first let's start on the rotary joint and before i even start on that i had to make sure i had my priorities in line i still want to use this arm for filming purposes and for general purpose robotics stuff so as such my main priorities for this build are to keep it cheap keep it pretty and make sure it's powerful enough to hold my camera nema 23s were my go-to for low cost power i even crimped on some jst connectors for extra professionalism and less risk of them melting when driving at 3 amps for power transmission i originally wanted to use belt drives and pulleys but i recently discovered this gear profile generator and autodesk inventor so of course i wanted to give this feature a try for the rotary joint i designed an 801 reduction using angled tooth profiles and i'm using an angle because it helps reduce noise and provide smoother motion but does it actually help yeah no it really does the angle of the teeth allows the next tooth to engage while the previous tooth is still fully engaged which just basically makes everything a lot smoother cool so next step is to constrain the joint the simplest way to constrain a rotary joint is this a bearing in a shaft bonus points if you actually clamp it to constrain the shaft axially but is this enough for my design no way bearings like this are fantastic for radial loads but absolute garbage when it comes to axial loads and especially these super cheap bearings that i buy in order to keep costs down so how do we fix this more bearings of course but not more radial bearings instead we have to add thrust bearings these bearings are a totally different beast and perfectly complement radial bearings as they are fantastic at constraining axial compression loads but trash at just about everything else so what i did is clamp two of these thrust bearings together with my rotary platform in the middle i then added a radial bearing in the center and this ensures that the joint is fully constrained by bearing which should help provide super smooth and accurate motion even under high loads unfortunately plastic is pretty well flexible and this assembly has a bit more compliance than i would like but for now this will work so i'd say it's time to test it in a previous video i used this off-the-shelf stepper driver and it worked alright but it's pretty big and for the price it's not very advanced i wanted to build my own motherboard from scratch both to reduce the total cost and to make sure i had all and only the features that i want in the board so after weeks of planning and designing it was ready a huge thank you to pcbway for sponsoring this video and providing me with the boards and stencils i ordered this board with the matte black silk screen which looks absolutely fantastic i then ordered all the components to populate the board with but i had to get a little creative since some all right most of the important ones were out of stock i had to make pre-orders order from multiple sources and even steal some chips from other projects and boards but in the end i just couldn't acquire the mosfets that i wanted to use with these new stepper drivers no worries though i found a similar spec alternative in the same package and ordered them great so next step is to apply the solder paste using the stencil then very carefully place all 200 components and finally reflowed in my diy reflow oven careful it's hot only to discover an issue a big issue i was a little careless when i applied the paste and i used poor technique and applied way too much and even better the paste i was using was expired and the flux mostly separated out lesson learned right i bought some new higher quality paste and did the whole thing again much more carefully and this time it came out perfectly until i tried to test it i discovered that the v inline was shorted to ground that's not good all the solder joints looked pretty good so i didn't think it was an assembly issue and then it hit me and i double checked those mosfets i'd ordered well it must have been really late because uh the pin out of these replacements does not match the original at all so i found these alternate alternatives ordered them reworked the board to add them and great that fixed the shorting issue next up i flashed the bootloader and uploaded a script to blink and led which all worked perfectly but that's where the good news end i then spent the next 20 hours debugging the new stepper drivers at first just enabling the stepper driver would cause the entire board to spaz out turns out i had a minor bug in my circuit design so i fixed that then i was having trouble establishing communication with the drivers on the board spy bus i purchased a logic analyzer to see exactly what was going on here and discovered that the drivers were just not responding to anything so i reworked the whole board to remove all but one driver just to simplify things but no luck yeah well it turns out these 5160 drivers need vm power for spy communication to work so i went and added back some more stepper drivers i'd removed and then things got screwy again after more debugging i realized that my unconfigured stepper drivers were all just trying to talk at the same time i was able to disable them at the right moments and get the 2041 driver working with relative ease but the 5160 still refused to work i could talk to it i could configure it but i couldn't enable it without it immediately faulting so eventually i discovered that my alternative alternate mosfets were a little weird these are dual end channel mosfets and the two sides of this one have different specs which might just be messing with the 5160s built-in fault detection luckily i somehow found the originally spec mosfets they should be on their way soon but i didn't want to hold up this video any further so i'm just going to be trying out using the 2041 driver in dual drive mode which combines the two motor outputs into a single one and provides double the power at 2.2 amps rms not quite the power of my 5160s which will be around 3 amps but it will do for now and after plugging the motor thankfully it's working great next i turned my attention back to the hinge joint luckily this guy already put a ton of work into developing a low cost high performance 3d printed harmonic drive and even more luckily bobby over at maker's acres reached out to me and offered to print some of the harmonic drive components in a more advanced material it's a nylon based filament with the option of adding a continuous carbon fiber reinforcement unlike the chopped fibers in cheap pla you can buy which do increase rigidity but are super weak this special material by markforged increases strength and rigidity bobby sent me the wave generator both with and without the carbon fiber and as you can see there's a huge difference between them this carbon fiber filled version is about as flexible as the pla but it's gonna be way tougher so i'm gonna go ahead and install this now he also sent me some of the flex splines done in the nylon these are a bit more flexible than pla but once again should be way more durable i'll be evaluating the long-term performance of these components versus pla in a future video but for today i need to figure out how i'm going to integrate the harmonic drive into the rest of the design since i wanted the arm to look good sticking a motor off the side was not gonna do instead i wanted to place the motor in the base and i would use a built-in pulleys to transmit power up to the harmonic drive but then i had a potentially brilliant or crazy idea what if i added an eight millimeter shaft that passed straight through the arm length one side of the shaft would have a pulley driven by the nema 23 and the other side would attach to the wave generator in the harmonic drive i know this is really extra but it's worth it for this super clean and symmetrical design i'm also going to be using heat stick inserts for the first time in this project again mostly for aesthetic reasons though the performance is surprisingly good assembly is a little awkward at this point with 8 m5 bolts clamping this whole thing together but it should be incredibly strong i'm still using this test arm from my previous video and that's just because i'm not done with the final design of the arms yet so as to not waste too much plastic i'm just going to pop a hole in this and use it here and wow yeah this thing is ridiculously large it really looked a lot smaller in cad but hey i'm committed now the observant of you may have noticed this tooth profile opposite the harmonic drive that does serve a purpose but i'll get to that in the next one on a totally unrelated note i wanted to quickly mention that i've been working on developing a compact absolute magnetic encoder for each joint but unfortunately due to the ridiculous amount of time i spent debugging the osr prime board i wasn't able to start testing these just yet i am really excited for them though they should add a lot of capability to this arm for almost no cost all right testing time the rotary joint is working fantastic but the harmonic is having issues so it seems like the 3d printed pulley is just slipping on the center drive shaft pretty easy fix so i tore it apart added a flat to the shaft but that's when i discovered that the wave generator also has a few spots where it jams up really bad which turned out to be me using slightly too long screws by accident so with both of these issues fixed i reassembled the arm and now it's working great side note i will be replacing most if not all of the 3d printed shaft clamps and bullies with off-the-shelf aluminum version for the next iteration and yep i know what you're wondering now how fast is it well the rotary is surprisingly fast it can run at 5 radians per second with ease and still have tons of torque and it doesn't max out until well above 10 with low loads of course the hinge is much slower with a max speed around one reading a second but a max usable speed about half of that this isn't too surprising though as the joint has a whopping final reduction of 102 1. as for power it's a little tough to test it right now without bolting the base down but i don't want to do that yes i'm not really sure where i want it to live the rotary joint feels like it's got about three or four new meters of torque at speed and the hinge well it's definitely got way more than that it doesn't struggle at all lifting itself up and the base is not light this thing weighs almost three kilograms already but it can definitely lift more you know what let's let's just do this here seems good [Music] and turns out my desk is actually hollow thanks ikea [Music] so i'm just gonna run through a bunch more tests see how much power this thing really has again keep in mind this is not the final stepper driver it's running at about 33 percent less current so that should be a fairly easy improvement once i get the motherboard up and running overall i'm super happy with the progress i've made on this robot arm it is a surprisingly solid unit i'm really excited to start building up the rest of the arm but for that you're gonna have to wait till the next video thank you all for watching as always a huge thanks to all my patreon supporters if you enjoyed the video please like the video and subscribe if you haven't already it really helps me out and that's all for this one until next time

Channel: 3DprintedLife

Views: 80,762

Rating: 4.9593711 out of 5

Keywords: 3d printed, robotics project, 3d printed project, arduino project, 3d printed robot, Custom pcb design, Trinamic stepper, Custom robot arm, 3d printed robot arm, Designing robot arm, 3d printed harmonic drive, low cost robot, low cost robot arm, 3d printed robotic arm, custom motherboard, custom stepper driver, custom robotics, robot arm, robotic arm, diy robot arm, 3d printer project, 3d printed gears, robot actuator, 3d printed actuator, harmonic drive, diy pcb

Id: 2CdDL6JARkU

Length: 12min 54sec (774 seconds)

Published: Mon Jul 12 2021