Your robots have a much more chaotic energy though. That's something to be... proud of? Yeah, probably :) Great results and inspiring videos in any case!
That's no forgery, that's real art. Part of what we admire in art is the skill of the artist, and the medium which you've used, digital, robotic, technical... I mean, just the fact that you had to engineer a robot and go through part iterations, the fact that you wrote software, and the fact that you made the art digitally anyways... this was insanely impressive. Very well done!
@@atruceforbruce5388 Nothing right now. Roybot had to be disassembled for home repair reasons and for an upgrade. The next Roybot project is very different and looks more like a mathmatical graph. And then I'll work on some new original pieces. Email art@nerdtronic3d.com and ask to be put on a list to get art updates.
It IS a cheap comic book print. All of Roy's stuff is just a cheap copy of work he didn't create. He's a plagarist, not an artist. He wholesale stole the work of over 300 struggling artists.
@@CookieTube Orella isn't talking about Nerdtronic when they mention Roy being a plagiarist. They are talking about Roy Lichtenstein, the painter that Nerdtronic is stylistically reproducing. Lichtenstein was well known for reproducing the work of poorly paid comic book artists, without credit to the originators.
@@CookieTube He DID 1:1 copies though and sold under his own name without giving credit though. It's the same crap like Andy Warhol. Just because a rich person wants to evade taxes, doesn't mean it's art. Roybot just did show in a way how Roy did it, no millions involved. (:
Roy didn’t always use stencils. I don’t know if it is still in Guggenheim, but there was a huge painting there of a living room and you, up close, could see the strokes of a painting brush.
I think, by tackling the problems the original artist may have faced, your own way, you (and Roybot!) created your own art form! And the outcome is spectacular!
I have a (much smaller) pen plotter robot. I've always wanted to fully automate the process such that somebody could request a commission online and immediately see a live stream of their print in progress. Would be cool to see you tackle this challenge!
Next step is writing a software that generates the robot's path. To begin, make it so that you are able to feed a single layer of paint, and choose a stroke direction and type. The softwate will then fill that layer with lines or dots and generate a path. After a few iterations, the final software should be able to accept either single colored layers or any image with a set pallet of colors, and slice it up.
Maybe. I think part of the art part of it is designing the toolpath so that it looks like paint strokes and not all vertical lines. Thanks for the suggestion and thanks for watching.
I'd think a useful first iteration would be to detect outlines, convert those into tool lines and then remove that area from the input. Rinse and repeat. As a second step, remembering the ends of paths and then trying to align the starts of paths of the next iteration with those, so the pen does not need to be lifted by can just move over one stroke width, might be nice. This kind of paint looks like it can take a bit of double-covering without it showing up.
@@HenryLoenwind Seeing the paint strokes is part of the attraction (or you can otherwise just order a big digital screen printing on vinyl). At least, if that is your taste. Both ways can equally be very attractive. Compare it to all the different and literally unlimited ways you can use to convert a photo in photoshop in all kinds of simulated painting styles. Each their own I'd say.
This would be cool in an Art gallery where THE ROBOT IS THE ART PIECE, but it also makes art. If you could figure out a way to have a hopper of pens and make it self load and unload the pens, that would be epic. Great work, and dedication, this project is truly a work of art in itself.
@@Goliath83 I guess my point was that the bigger my channel gets, the more comments, the harder it will be to respond to many of them. But I'll still try.
@@Nerdtronic just an idea i just thought of, but can you not put something at the end of the video or something telling people a word to include in their comment, so if they have a question you can just filter by that word
So, a few comments, observations: A. This is amazing, and you are amazing. I love it. B. You basically built a vertical 3D printer, but without the extruder and heating elements, and with custom motors? As such, I probably would have went the route of trying to make it work with conventional 3D printer software. Imagine if you used Klipper, configured the stops of the new motors, and set up the "pen in/out" mechanism to work with the extruder code. Then, all you'd have to do to create this kind of art is split your image into multiple layers with the shapes to print, then feed the flat image into some kind of CAD software to extrude it into a one-layer shape that you drop into a slicer and print. It would save you a ton of time having to manually create all the paths in illustrator, as well as have saved you the headaches of trying to get the paths to trace right. Still an incredible teaching video and overall process...but if you wanted to supercharge it and make it easier to make more art - this could be how. ;)
Seriously! I think the last 6 months has been working non stop on this project. Except for the 6 weeks I was out with covid. I think I can paint the next piece same size in about 2-3 weeks. After I redesign the print head of course. Thanks for watching.
@@Nerdtronic Dude, this is something you can easily make side living of! Fine tune the software, sprinkle some AI in it to generate images, et voila, big canvas sized AI/robot generated art to sell for big (but reasonable) prices! .... Me jealous
@@CookieTube If he thinks it could take 2-3 weeks to do a piece the same size, now that he has one under his belt, it will take about half a year to finish an entire wall mural. You can't really make money like that, unless it's part of a display or show, like at a gallery or museum, which I do admit would be cool. He would need to rethink his rig in several different ways before it could crank out murals and make some decent money, i'm guessing. It is a cool thing though, not knocking it.
Money aside, I'd rather have your work of art than the original! Amazing commitment to making this work. Wow! I've got a one ton cnc machine in my garage and can appreciate the level of tinkering this requires, having gone through a bit myself. Good job finishing it!
Can I just say that your visualization is greatly appreciated. From those subtle text shadows to those 1 to 1 model renders, your visual language is a perfect matrimony of oral communication and visual storytelling. I’m not even into 3d printing but I follow along because of how meticulously presented your videos are. Know that you are appreciated.
Thanks! I am back. Trying to drag my channel in a new direction and do more stuff like this. It takes me 3 weeks just to edit a video like this. So my goal is to do a video about once every 4 to 6 weeks.
And for the upgrades: You already have stable slides, you can very simply put a linear optical encoder on them for positioning. This would also allow you to use a simpler target-feedback positioning logic. (Calculate the intended path and timing. X times per second, check the actual current position and the intended position 1/(X+A) seconds later and feed the difference into the motor. That way, you can use the motor's automatic mode; the A value ensures the motor never reaches its target and stops as it always gets a new movement command before that.)
What an amazing project! And as always the video production is just perfect. Very entertaining to watch. Your content deserves a lot more viewers to appreciate it. Hope you get them soon!
@@Nerdtronic And on top of a nice painting or graphic, you make a very well recorded and edited vid. I suspect you’ll scrap version 1 and version 2 will address a lot of nagging things, things that annoy only you. Like the pens. And speed and XY registration.
Just found the channel and I'm baffled about the small amount of subscribers you have, it needs more! Your content is amazing keep up the good work love the painting
would love to see an episode covering the electronics and software design in more detail. Ie - what you coded with, what was used for the controller, etc.
I'll think about it. Meanwhile - The hardware is a raspberry pi using the serial pins on the gpio. That is connectd to a serial to can bus module and then to the motors. The motors are 48 volts but the one on the paint head is 24 volts so I have a converter. The software was written in xojo on a mac and it can compile a debug version on to the pi. I never made an actual executable I just always run it in debug mode. The print file format is basically svg but slightly modified.
@@Nerdtronic well, if you decide to make a video, I will certainly be excited to watch it. The amount you shared helps me understand what you did which is what I was after lol
Man this is so cool. From making the robot to software to the art itself my inner child is screaming oh and I’m also a huge IT Nerd and this was probably the best video I’ve seen in years!
Above all I'm amazed you got the funds to do this. I've been doing DIY plotters since 2018 but I never got around to amassing the funds to build anything bigger than A0. When people started copying my artworks via instagram I had to close down all of my social media as those copycats without any talent were better at selling than me. It's been over half a year since I quit making plotter artworks but started working on a new project two weeks ago and I gotta say, the itch is still there. It was nice seeing your explorations, I came across many of the same issues, such as the pen ink feeds, and drilling them out which works really well. I work with high-flow acrylic inks which are a lot harder to control (drilling the caps means they tend to leak all over the place), but they have a nice fast flow if you're into speed. That's the main thing where our machines differ, mine are super fast and have zero backlash, but they can't go bigger than A0 and I didn't have the funds to get decent extrusions, so the frame is quite vulnerable. I'm currently working on a special design to minimize some motion artifacts and I think that's probably the most important step for you as well if you want to go faster. I'm also 100% sure that the backlash is coming from the inside of your motors. I know you got sponsored for the motors, but I really want to suggest using stepper or servo motors. Trust me, 0.9 degree per step motors are fine, you could even use a belt to increase the resolution further. And if you use closed loop motors you can do anything you want without complex movement planning. It looked as if you were using a Duet board, I'm not sure but it looked like that. If you are, that's great because Duet has expansion boards with which you can run closed loop stepper motors without problems. Anyway, a joy to watch your video, and it's good you also show your failures. I tried to get some sponsoring from pen companies etc with my 2k follower instagram back in the days and I managed to get was €100 worth of pens. I am very grateful with this but realistically it's not enough to sustain this on the long term. Do you have any advice for me how I can get better sponsorship deals with companies? Do they care more about exposure or are they also in a way showing goodwill when it comes to helping out artists that are not able to fund their practice even though they make unique quality projects with what little they have?
Sounds like you've had some fun. I'll be doing other types of plotter art with this as well. The other day I bought about 10 more very large canvases. But I'm going to fix the problems before starting the next painting.
Loved the project! Next time use a Duet3D to avoid all the positioning problems and control it like a 3D printer. And to improve the pen mechanism with just 3D printed parts, just hold the pen in two separate points
@@Nerdtronic honestly keep going with these videos, the skill set you have + the want for these types of videos will get you so many views. Exactly what I've been looking for yt wise!
Imagine spending 42 million on a painting. You have more money than 99.9% of any human on the planet and can do so much with your wealth. Then you buy a painting that you could just get a print of. Some people just have too much money.
Well realize also... someone buying a 42 million dollar painting isn't spending every cent they have on it. And it's not an expendible. Not like buying a million dollar pizza. They can keep their wealth in the bank or in property. They can sell it later and get that cash back.
Only 2 or 3 places. It accidentally drew a small line when doing the black dots over the blue. And there was a red dot inside the eye that shouldn't have been there. A couple other mistakes were fixed by the robot where I just made an additional painting file to fix it.
We need more people working hard to pop the ridiculous high-end $$$ Art Bubble. I know you would never do fakes and forgeries but thanks for showing how it could be done.
Some of the high-auctions are definitely ridiculous. But why the animus? If it gets people interested in art and the market interested in finding the next new artists, why do you want to "pop" that? That's a very small insecure perspective.
I’m so happy that UA-cam decided to lead me to your channel. This really is incredible, what you created here. RoyBot has a kingly name and it is fitting! Would this have been easier to do if the robot was oriented in the same direction as gravity, i.e. painting with the canvas laying on the floor?
Probably. But would take up a lot more space since it's 13x9 feet. Originally it was going to be a wall mural painting robot but then I decided to focus on the canvas.
Maybe I missed it but why did you have to do the software and also the toolpathes manually? Any CAM program would be up to the job. I've done linoleum and wood prints in my CNC. You just need to do a vector graphics of your drawing...also automating the points is something cad does.
I did a little thinking some time ago about doing a painting robot but I (of course) wanted the full color range. My idea was to use an image and see what the RGB was and then go to a pallet, scan for the right color, dab said color and then dab the canvas. Then go dab any other pixels of the same color returning to the (already known) location on the pallet to pick up more paint. I got a major headache when I realized that it would have to mix colors on the fly... or something... where's the aspirin? But your bot WORKS GREAT even though there had to be some "interventions" on the first full scale run. And to be honest, the belt was the last thing that I would have expected to become an issue. As a medium I want to use ordinary household PVA and add tint to make the primary colors - should be a lot less costly that artist acrylic paint which is basically the same thing... Then feed the 3 colors through some sort of small peristaltic type tri-pump system to meter the base colors to get the required color. Found some thin clear plastic from one of those oxygen cannular (???) things they stick in your nose in hospitals - so that went into the project box... But I still have to collect a LOT of other stuff before I can begin. Looking forward to see how you go forward on this project and how you solve the issues that ALWAYS crop up.
That painting is amazing on its own right, and it should definitely be displayed in a suitable gallery. Your work is impressive, both the technical and the artistic aspects!
Wow thanks. It’s so custom that I don’t think I will. The software was written in xojo which most people haven’t heard of. It’s pretty hacked together.
@@Nerdtronic It would be great to take a look at some of the code just to see how you handle the motion, and I’m sure some people would be willing to help clean it up a bit! Have you thought about using Bézier curves to generate the trajectories rather than segmenting curves? Or is the segment length just so small that it doesn’t make any difference?
I'm like "hey, this is a Roy Lichtenstein painting I have never seen before" - but then, I'm not an expert. One thing got me though ... all of the Lichtenstein pop-art painting I know are from the 60s and that phone is definitely not a 60s (or earlier) style phone. I am not sure if you did that on purpose but I find this a very nice touch. A great piece of art by a great artist.
I have a CNC plasma cutting table that I’ve also added a sharpie holder to for funsies, but your painting is AWESOME!!! Subbed because of how fricken cool that painting looks! Great job to the both of you!
The machine was off by 2 mm, but hey - at least you had a resolution of 0.18 µm. 😂 Sounds like maybe the actuator was not completely necessary and a simple stepper would have done fine as well. Anyway - great work overall, and yes, I am impressed - thanks for sharing! I would recommend you use linear magnetic encoders for position feedback. That stuff is similar to what is used in digital calipers and is available relatively cheap even for huge machines like yours (well, at least the chinese stuff). Needs some skills to implement, but there's no doubt YOU can. Precision is 5 µm - so still way more than enough for drawings.🙂 But of course this would not prevent the head from sagging...
@@Nerdtronic Thanks for the answer, but sorry - there's a misunderstanding. Don't get me wrong: The Roybot including the software is a master piece and you can be very proud of it. And the actuators are fantastic devices - there is absolutely no doubt! But the built in encoder has absolutely NO clue of the head position. It ONLY knows about the motor itself. But the motor is just one part of the positioning system. The complete belt system including the pulleys plays a much bigger role. A linear magnetic encoder would completely remove issues due to belt and pulley accuracy, belt tension, including change of belt tension over time, plus backlash and even some issues due to temperature changes over such long periods of continous use. An alternative would be a refernce point to check (for example using a hall sensor) every know and then. But relying on a belt + pulley for accuracy is just not working well (as you have seen).
I see. Well assuming that nothing slips the motor returns an absolute position in the form of a multi-rotation angle, which simple math turns it into a position on the wall in mm (execpt for backlash). I said as long as nothing slips. Once when toward the end of the black pass I made a mistake and had the pen mech down when it was off canvas, then put a pen in and told it to draw something without raising the pen first. It was going to crash the pen into the side of the canvas. I went to quickly unscrew the thumbscrew and take the pen out and in trying to prevent it from ripping the paintng off the wall I caused the motor to slip in the gear and then the registration was lost. I was able to recalibrate it close enough to continue. I think the backlash will go away if I tension the belt tighter. But I'm also going to make a gear rack and try that.
@@Nerdtronic Thanks for the detailed insight. Well, accidents like those just happen. What I was trying to say is that even with perfect rotary encoder feedback and perfect calculation there will always be a delta between the computed and the real position. This could be eliminated by linear encoders. But I don't want to bother you with that any longer. One additional idea: Maybe you can compensate at least belt (tension) issues by using two reference points on each axis - one near each end. You can then measure the motor steps in between. The value should in theory always be the same, but over time it will not be the same in practice. I think higher belt tension might lower but never eliminate backlash. But maybe you can get it low enough for the purpose (i.e. to maybe 20% of the pen stroke width).
Being the curious type, I am wondering if you considered canvas deflection when programming the pen travel distance? If not, how would a solid canvas backing affect pen movement accuracy and ink bleedthrough?
I had planned to put styrofoam behind it but it wasn't possible. But no paint/ink bleed through. The back of the canvas is blank. I did consider canvas pushing in. I adjusted the pen depth depending on where it was painting.
13:55 I would just call that art imitating life imitating art imitating life! Roy based his artwork on printed illustrations where slight misalignments like that are common, he hand painted his artwork and now you’re copying his hand painted art with a computer controlled system that has simultaneously flawlessly reproduced the image as well as introduce a flaw akin to the original inspiration.
Haha yeah. I noticed that some of his dots weren't all that great either. Some of mine are more oblong than round near the top. I think it was caused because the canvas is a flexible cloth surface.
@@Nerdtronicit would be interesting to see other ways you could use this robot, beyond replicating existing works. One thought I had would be giant string art- you that thing where people use string as straight lines to create images. I wonder if you could do something similar only using the paint pens or permanent markers instead of string. I mean the whole point of using the string on a small scale is to create a straight line, but using your robot it would be easy to create a straight line with a pen. I’ve seen another YTr try to recreate string are using a CNC machine like yours, but all I could think is “why is he using string? Much easy to use a pen?”.
You pulled it off 100 %. Absolutely perfect. Even the imperfections due to pen holder sagging added a bit of realism and similarity to cartoon inaccuracies.
That's wonderful. I think the artists concepts being perfectly replicated by Roybot is a brilliant idea. Damien Hirst works out concepts and leaves paid employees to carry out the work so Roybot is just doing the same thing.
Wow this is fantastic! And here I wondered why there were so few videos over the past months, then you come out with this absolute gem. Stunning work, both artistically and from an engineering perspective. Great stuff :)
Wow, thank you! It was one of the reasons it's been a while since I released a video. Also some family/house projects that I didn't do videos on, plus being out of commission for about 6 weeks with covid.
This could be fun to combine with something like StableDiffusion. I've lost count of the amount of times I've generated an image only to think "Man this'd look great on my wall". There'd still be a bunch of human work involved (dividing the image into layers, all the pen replacement etc), but it'd be fun to see an almost entirely "machine generated" painting, with an AI generating the original design, and a robot making a real world painting of it.
This is really awesome! The painting looks great! Fantastic job by Roybot and you!! I would love to see Roybot do more Lichtenstein inspired paintings in the future.
To be fair, Roy Lichtenstein stole all of his famous painting from comic artists so even if you had copied one of his paintings it would have probably been totally fair. There's blogs online documenting all of the instances of stolen art. The idea that he somehow "elevated" the art by printing it on a canvas is insulting to the artists who originally created these works. Look up "drowning girl" to see how badly he copies and comes out looking like the less talented artist.
This robot painting the art would be an amazing art piece. Feed it AI images in the style of famous painters and you already have a very solid concept. It says alot about the world we live in and the fear of AI and robots taking over the creative industry. You could sell this to any gallery. People would watch this thing paint for hours in a museum.
I think fair use covers you copying a Lichtenstein painting. Especially since his art is straight up copying comic book panels. People paint reproductions all the time. Might get weird if you try to sell it. Excellent work man! Don't know how I haven't found your channel before.
Thanks! Hopefully it won't be another 12 months for my next video. I had some personal things going on as well. I'm starting my next video on monday, if I can just figure out what to print on this new Galaxy 1 printer.
So impressive. I'm curious how "autonomoous" you think you could get Roybot. Obviously the paint pens slow things down (maybe a tube and pump mechanism instead of the pen cartridge?) but how involved did you have to be on adjusting the paths beyond what you showed in the video? How small would the canvas have too be too do a single color artwork in a day?
I really had to babysit it. I was constantly concered that something would happen to make it ruin the painting. So many close called that I didn't get on camera. That said.. My plan was to use a drink IV bag and fill it with the paint. Drill a hole in the back of a pen and 3d print a rubber adapter so I can attach the IV tube to the pen. Hang it a little above the pen. But swapping out pens was quicker than messing with that and having it leak all over the place.
@@Nerdtronic Thanks for replying, that totally makes sense. I feel like thats inevitable with so many robots, getting it to do the thing is hard enough but getting it to do it without maintenance or intervention is a hundred times more complicated and expensive. This was really amazing. I like your idea, but totally agree I'd be constantly terrified of leakage
This is awesome. I wonder if you could put some kind of notches into the guide rail and use that to accurately gauge the travel instead of using a rubber belt. It'd be difficult to maintain the accuracy to notch them, but it would last longer.
The first paint head had a pot in there to feedback the amount of pressure on the pen. But it didn't work. I must have gotten a crap pot and had glued it in. It never gave accurate readings. The rubber bands held up for the whole print but a spring would have also worked. Shane from Stuff Made Here suggested that I didn't need the double set of rails. I used that big actuator there also and didn't need something that beefy but it was what I had on hand. And actually that actuator can act as a spring all by itself if you set the max torque. I just never experimented with that feature but will when I make robots that walk around.
I remember doing a Lichtenstein painting for art class back in High School by hand. I used a toothpick for the dots. It definitely was not as good an outcome as this but at that time I liked how it came out. A ton of dots though even though it was less than a quarter of the size. When I was playing around with robotics and was working with a sort of robot arm, also not industrial or remotely close to pro level equipment I found gravity would screw up my positioning pretty quickly. Movement going down would be faster than going up. Was that an issue with this set up? I would have thought you might have has less issues if you made it table top instead of wall/vertical mounted like you had. Granted, vertical orientation takes up a much smaller footprint. Very cool project. Love the result.
As a suggestion what about changing out the pen actuator for a solenoid it would be significantly lighter and easier to reinforce than the double slide sprung rack and pinion pen actuator and you could use a motor controller to have variable force so it would double as the relief spring
The pen actuator motor is way overkill. It's what I had on hand and already spoke the same language as the other 2 motors which are on the same can bus.
@@NerdtronicThanks for the reply and fair enough I was trying to think a way to make it more rigid without adding more weight/using metal parts, and to that effect what about swapping out gear system and the double slide for a radial or cylinder cam and a single set of rails with a 'driver' slide and the pen holder be a floating slide, with springs or a soft foam block in between and have the slide held back by elastic bands to tension against the spring and the cam, this would still use the motor as the actuator while simplifying the design and you could 3d print and mount the cam directly to the motor and lay it inline with the pen slide assembly removing the double slide system which seems to be the major mechanical problem in the current design. Importantly I forgot to mention it in my first comment but the art turned out beautifully
The floating slide idea would probably also work with the current rack design which would make fewer parts to remake though if you used the current primary slide it would limit the travel.
I like to think that extremely talented/smart people like nerdtronics and stuffmadehere have always existed but nowadays there's a platform to showcase their talents.
This is an immense amount of work for 1 person, not only designing, printing and assembling the robot, but also making actual art with that whole document recording the whole process. Big thumbs up, that’s fantastic!
This is one of the more insightful comments here. I find that super interesting. I couldn’t believe that he would have free handed his grid of dots! Makes so much sense.
My wife said your robot makes better paintings than mine - and she’s right - very nice result!
Haha tell her thanks. I’ll send you guys an original Roybot painting soon.
Your robots have a much more chaotic energy though. That's something to be... proud of? Yeah, probably :) Great results and inspiring videos in any case!
Cart fard
I would love to see you both collaborate on a build!!
A legend commenting on another legend.
That's no forgery, that's real art. Part of what we admire in art is the skill of the artist, and the medium which you've used, digital, robotic, technical... I mean, just the fact that you had to engineer a robot and go through part iterations, the fact that you wrote software, and the fact that you made the art digitally anyways... this was insanely impressive. Very well done!
Wow thanks. It was a fun and fulfilling project.
@@Nerdtronic how much you wanna sell me that robot made painting for?
Ha this original is not for sale. But other original paintings by Roybot start at about 5 grand.
@@Nerdtronic got a catalog? I'm curiously wondering what "other" robot paintings are avaliable.
@@atruceforbruce5388 Nothing right now. Roybot had to be disassembled for home repair reasons and for an upgrade. The next Roybot project is very different and looks more like a mathmatical graph. And then I'll work on some new original pieces. Email art@nerdtronic3d.com and ask to be put on a list to get art updates.
I kinda liked the white gaps on the black lines, because it made it look more realistic. Like an actual cheap comic book print.
true
It IS a cheap comic book print. All of Roy's stuff is just a cheap copy of work he didn't create. He's a plagarist, not an artist. He wholesale stole the work of over 300 struggling artists.
@@orellaminx3530 L take
@@CookieTube Orella isn't talking about Nerdtronic when they mention Roy being a plagiarist. They are talking about Roy Lichtenstein, the painter that Nerdtronic is stylistically reproducing. Lichtenstein was well known for reproducing the work of poorly paid comic book artists, without credit to the originators.
@@CookieTube He DID 1:1 copies though and sold under his own name without giving credit though. It's the same crap like Andy Warhol. Just because a rich person wants to evade taxes, doesn't mean it's art. Roybot just did show in a way how Roy did it, no millions involved. (:
Lichtenstein just used stencils for the dots. So your version is technically more impressive how it was made. It is a great project.
Thanks!
Roy didn’t always use stencils. I don’t know if it is still in Guggenheim, but there was a huge painting there of a living room and you, up close, could see the strokes of a painting brush.
As a nerd and an art fan I really appreciate this in so many different ways. You should be really proud of that painting.
Wow thanks. I have to say that I am. Will be making more paintings soon.
I think, by tackling the problems the original artist may have faced, your own way, you (and Roybot!) created your own art form! And the outcome is spectacular!
Thanks!
I have a (much smaller) pen plotter robot. I've always wanted to fully automate the process such that somebody could request a commission online and immediately see a live stream of their print in progress. Would be cool to see you tackle this challenge!
hmm... that could be a real neat way to do it if tool changes (pen swaps) were added to the machine.
@@zerumsum1640 yup! i can't find the link rn but somewhere someone made a pen carousel attachment that would rotate to select a pen
Next step is writing a software that generates the robot's path. To begin, make it so that you are able to feed a single layer of paint, and choose a stroke direction and type. The softwate will then fill that layer with lines or dots and generate a path.
After a few iterations, the final software should be able to accept either single colored layers or any image with a set pallet of colors, and slice it up.
Maybe. I think part of the art part of it is designing the toolpath so that it looks like paint strokes and not all vertical lines. Thanks for the suggestion and thanks for watching.
I'd think a useful first iteration would be to detect outlines, convert those into tool lines and then remove that area from the input. Rinse and repeat. As a second step, remembering the ends of paths and then trying to align the starts of paths of the next iteration with those, so the pen does not need to be lifted by can just move over one stroke width, might be nice. This kind of paint looks like it can take a bit of double-covering without it showing up.
@@HenryLoenwind Seeing the paint strokes is part of the attraction (or you can otherwise just order a big digital screen printing on vinyl). At least, if that is your taste. Both ways can equally be very attractive. Compare it to all the different and literally unlimited ways you can use to convert a photo in photoshop in all kinds of simulated painting styles. Each their own I'd say.
This would be cool in an Art gallery where THE ROBOT IS THE ART PIECE, but it also makes art. If you could figure out a way to have a hopper of pens and make it self load and unload the pens, that would be epic. Great work, and dedication, this project is truly a work of art in itself.
A local gallery is intersted in doing this as an art installtion but I want to wait until there is more art to showcase.
jesus, i just looked at the comments and you answered nearly every single one.
now thats some true community interaction ^ ^
I can’t do it forever but I try to respond to some every couple of days
@@Nerdtronic dude dont say that like its nothing, you are responding to more comments than some of the biggest channels on youtube
@@Goliath83 I guess my point was that the bigger my channel gets, the more comments, the harder it will be to respond to many of them. But I'll still try.
@@Nerdtronic just an idea i just thought of, but can you not put something at the end of the video or something telling people a word to include in their comment, so if they have a question you can just filter by that word
Unbelievably high production quality and you conquered every challenge in your way until it was done! I'll show this to my students for inspiration.
Thanks! Hope they enjoy it.
So, a few comments, observations:
A. This is amazing, and you are amazing. I love it.
B. You basically built a vertical 3D printer, but without the extruder and heating elements, and with custom motors?
As such, I probably would have went the route of trying to make it work with conventional 3D printer software. Imagine if you used Klipper, configured the stops of the new motors, and set up the "pen in/out" mechanism to work with the extruder code.
Then, all you'd have to do to create this kind of art is split your image into multiple layers with the shapes to print, then feed the flat image into some kind of CAD software to extrude it into a one-layer shape that you drop into a slicer and print.
It would save you a ton of time having to manually create all the paths in illustrator, as well as have saved you the headaches of trying to get the paths to trace right.
Still an incredible teaching video and overall process...but if you wanted to supercharge it and make it easier to make more art - this could be how. ;)
Thanks. Creating tool paths is part of the art. It's not visible on video but you can see some of the paint strokes.
dude. the effort. i was literally waiting for you to deal with paint flow and you nailed it.
Awesome project! I love things like that and the painting looks great.
Why spend a week painting something when you can spend a year automating it ;)
Seriously! I think the last 6 months has been working non stop on this project. Except for the 6 weeks I was out with covid. I think I can paint the next piece same size in about 2-3 weeks. After I redesign the print head of course. Thanks for watching.
@@Nerdtronic Dude, this is something you can easily make side living of! Fine tune the software, sprinkle some AI in it to generate images, et voila, big canvas sized AI/robot generated art to sell for big (but reasonable) prices! .... Me jealous
@@CookieTube If he thinks it could take 2-3 weeks to do a piece the same size, now that he has one under his belt, it will take about half a year to finish an entire wall mural. You can't really make money like that, unless it's part of a display or show, like at a gallery or museum, which I do admit would be cool. He would need to rethink his rig in several different ways before it could crank out murals and make some decent money, i'm guessing. It is a cool thing though, not knocking it.
@mikemulligan5731 I don't know. One 42 million dollar piece and your can live pretty well for the rest of your life.
@@ThatOpalGuy Art doesn't work like that. You can't copy something and sell it for 42 mil there are a lot of factors that go into that made up price.
You did a really cool work making that robot.
The art looks amazing even with those little gaps.
Money aside, I'd rather have your work of art than the original!
Amazing commitment to making this work. Wow!
I've got a one ton cnc machine in my garage and can appreciate the level of tinkering this requires, having gone through a bit myself. Good job finishing it!
Thanks. Well "Not Over The Phone" is my artwork. I used the robot as a tool to paint it. Thanks again for recognizing the hard work!
This is one of the most amazing things I've ever seen! Truly awesome, as in inspiring genuine awe. Congratulations on a phenomenal success!
Thank you so much 😀
Search for Mike Everman's servo -belt system, which uses a second belt section stuck to the frame as a rack for the drive belt
I'll check it out. Thanks
Can I just say that your visualization is greatly appreciated. From those subtle text shadows to those 1 to 1 model renders, your visual language is a perfect matrimony of oral communication and visual storytelling. I’m not even into 3d printing but I follow along because of how meticulously presented your videos are. Know that you are appreciated.
Wow thanks Jason.
Why didn't you use G-Code with Arc support?
So glad to see you are back on youtube!!! I've missed the content. Hope you keep uploading!
Thanks! I am back. Trying to drag my channel in a new direction and do more stuff like this. It takes me 3 weeks just to edit a video like this. So my goal is to do a video about once every 4 to 6 weeks.
@@Nerdtronic I'm all for this new kind of content! So many more opportunities. 1 vid every 6 weeks is a heck of a lot better than 1 vid a year!!!
@@marcuspagel Ha True. Hopefully I can make that. Would you believe I've lost more subscribers than I've gained so far today on this video?
Not only it goes to the point but the way it also showcases the challenges and problem silving along tge way is fantastic, thats a sub.
Wow thanks!
This was absolutely insane. My god.
Massive props for putting the insane amount of effort in to make Roybot!!
Thanks. It was fun.
And for the upgrades: You already have stable slides, you can very simply put a linear optical encoder on them for positioning. This would also allow you to use a simpler target-feedback positioning logic. (Calculate the intended path and timing. X times per second, check the actual current position and the intended position 1/(X+A) seconds later and feed the difference into the motor. That way, you can use the motor's automatic mode; the A value ensures the motor never reaches its target and stops as it always gets a new movement command before that.)
What an amazing project! And as always the video production is just perfect. Very entertaining to watch. Your content deserves a lot more viewers to appreciate it. Hope you get them soon!
Thank you very much! I hope I get them soon too. View count is pretty depressing at the moment.
@@Nerdtronic
And on top of a nice painting or graphic, you make a very well recorded and edited vid.
I suspect you’ll scrap version 1 and version 2 will address a lot of nagging things, things that annoy only you. Like the pens. And speed and XY registration.
You're actually a genius. Such a beautiful painting!
as a fan of both weird art AND cool robots, this is awesome! sometimes being lazy is more impressive than doing things manually
Just found the channel and I'm baffled about the small amount of subscribers you have, it needs more! Your content is amazing keep up the good work love the painting
Thanks.
would love to see an episode covering the electronics and software design in more detail. Ie - what you coded with, what was used for the controller, etc.
I'll think about it. Meanwhile - The hardware is a raspberry pi using the serial pins on the gpio. That is connectd to a serial to can bus module and then to the motors. The motors are 48 volts but the one on the paint head is 24 volts so I have a converter. The software was written in xojo on a mac and it can compile a debug version on to the pi. I never made an actual executable I just always run it in debug mode. The print file format is basically svg but slightly modified.
@@Nerdtronic well, if you decide to make a video, I will certainly be excited to watch it. The amount you shared helps me understand what you did which is what I was after lol
Wow!
You're a real renaissance man. That being said, you're expert level at anything you apply yourself to.
Wow thanks.
Wow, this is wild! Can't wait to see what you do next, hope to see more of RoyBot.
Thanks. I have another art robot project coming. It's much smaller. And you'll see more of roybot.
Man this is so cool. From making the robot to software to the art itself my inner child is screaming oh and I’m also a huge IT Nerd and this was probably the best video I’ve seen in years!
Wow thanks!
Above all I'm amazed you got the funds to do this. I've been doing DIY plotters since 2018 but I never got around to amassing the funds to build anything bigger than A0. When people started copying my artworks via instagram I had to close down all of my social media as those copycats without any talent were better at selling than me. It's been over half a year since I quit making plotter artworks but started working on a new project two weeks ago and I gotta say, the itch is still there.
It was nice seeing your explorations, I came across many of the same issues, such as the pen ink feeds, and drilling them out which works really well. I work with high-flow acrylic inks which are a lot harder to control (drilling the caps means they tend to leak all over the place), but they have a nice fast flow if you're into speed. That's the main thing where our machines differ, mine are super fast and have zero backlash, but they can't go bigger than A0 and I didn't have the funds to get decent extrusions, so the frame is quite vulnerable.
I'm currently working on a special design to minimize some motion artifacts and I think that's probably the most important step for you as well if you want to go faster. I'm also 100% sure that the backlash is coming from the inside of your motors. I know you got sponsored for the motors, but I really want to suggest using stepper or servo motors. Trust me, 0.9 degree per step motors are fine, you could even use a belt to increase the resolution further. And if you use closed loop motors you can do anything you want without complex movement planning. It looked as if you were using a Duet board, I'm not sure but it looked like that. If you are, that's great because Duet has expansion boards with which you can run closed loop stepper motors without problems.
Anyway, a joy to watch your video, and it's good you also show your failures. I tried to get some sponsoring from pen companies etc with my 2k follower instagram back in the days and I managed to get was €100 worth of pens. I am very grateful with this but realistically it's not enough to sustain this on the long term. Do you have any advice for me how I can get better sponsorship deals with companies? Do they care more about exposure or are they also in a way showing goodwill when it comes to helping out artists that are not able to fund their practice even though they make unique quality projects with what little they have?
Sounds like you've had some fun. I'll be doing other types of plotter art with this as well. The other day I bought about 10 more very large canvases. But I'm going to fix the problems before starting the next painting.
Loved the project! Next time use a Duet3D to avoid all the positioning problems and control it like a 3D printer. And to improve the pen mechanism with just 3D printed parts, just hold the pen in two separate points
The quality of this video is crazy good, from the animations to commentary, geez man
Wow thanks!
@@Nerdtronic honestly keep going with these videos, the skill set you have + the want for these types of videos will get you so many views. Exactly what I've been looking for yt wise!
Lichtenstein is one of my favorites! This looks so similar to what I've seen in the met, so incredibly precise - it's almost TOO good!
Right? 😁 Thanks!
Didn't he plagiarized lots of his work?
Imagine spending 42 million on a painting. You have more money than 99.9% of any human on the planet and can do so much with your wealth. Then you buy a painting that you could just get a print of. Some people just have too much money.
Well realize also... someone buying a 42 million dollar painting isn't spending every cent they have on it. And it's not an expendible. Not like buying a million dollar pizza. They can keep their wealth in the bank or in property. They can sell it later and get that cash back.
Welcome back! That is an incredible project. I love seeing engineering used for stuff like this
Thanks!
The painting is genuinely beautiful! Did you have to do touch ups by hand in the end?
Only 2 or 3 places. It accidentally drew a small line when doing the black dots over the blue. And there was a red dot inside the eye that shouldn't have been there. A couple other mistakes were fixed by the robot where I just made an additional painting file to fix it.
@@Nerdtronic very nice! This seems like a good business idea but I'm not sure about the licensing of selling artwork replicas
We need more people working hard to pop the ridiculous high-end $$$ Art Bubble. I know you would never do fakes and forgeries but thanks for showing how it could be done.
Thanks - yeah I wouldn't recreate a real Lichtenstein. But strangely enough I think my Roybot art prints cost more than an actual Lichtenstein print.
Some of the high-auctions are definitely ridiculous. But why the animus? If it gets people interested in art and the market interested in finding the next new artists, why do you want to "pop" that? That's a very small insecure perspective.
i assumed the whole thing was just a money laundering front
In 1996, my high-school computer lab had a plotter. This seems like a fancier version of that. Still very excellent project and outcome.
Yeah it’s a large plotter. But originally I designed it to paint on a wall. Thanks for watching!
@Nerdtronic kudos on your hard work, problem solving, and for building on the conversations of fine reproductions of pop culture.
I’m so happy that UA-cam decided to lead me to your channel. This really is incredible, what you created here. RoyBot has a kingly name and it is fitting!
Would this have been easier to do if the robot was oriented in the same direction as gravity, i.e. painting with the canvas laying on the floor?
Probably. But would take up a lot more space since it's 13x9 feet. Originally it was going to be a wall mural painting robot but then I decided to focus on the canvas.
@@Nerdtronic Good call, it's amazing what you did and what your RoyBot can produce!
Maybe I missed it but why did you have to do the software and also the toolpathes manually?
Any CAM program would be up to the job. I've done linoleum and wood prints in my CNC. You just need to do a vector graphics of your drawing...also automating the points is something cad does.
Art, my friend, art.
Some premium content right there👍
Thanks. Hope to make more!
I did a little thinking some time ago about doing a painting robot but I (of course) wanted the full color range.
My idea was to use an image and see what the RGB was and then go to a pallet, scan for the right color, dab said color and then dab the canvas.
Then go dab any other pixels of the same color returning to the (already known) location on the pallet to pick up more paint.
I got a major headache when I realized that it would have to mix colors on the fly... or something... where's the aspirin?
But your bot WORKS GREAT even though there had to be some "interventions" on the first full scale run.
And to be honest, the belt was the last thing that I would have expected to become an issue.
As a medium I want to use ordinary household PVA and add tint to make the primary colors - should be a lot less costly that artist acrylic paint which is basically the same thing...
Then feed the 3 colors through some sort of small peristaltic type tri-pump system to meter the base colors to get the required color.
Found some thin clear plastic from one of those oxygen cannular (???) things they stick in your nose in hospitals - so that went into the project box... But I still have to collect a LOT of other stuff before I can begin.
Looking forward to see how you go forward on this project and how you solve the issues that ALWAYS crop up.
I could be wrong but I think even Stuff Made Here's wall painting robot only does one color at a time. He did CMYK but not all 3 at once.
Whoaaa now that's a comeback video amazing content
Wow thanks. I'm starting on my next video on monday. So it won't be another year. 🤣
You and Roybot defiitely hit the mark, even given the drift in registration when painting the black.
Thanks
That painting is amazing on its own right, and it should definitely be displayed in a suitable gallery. Your work is impressive, both the technical and the artistic aspects!
Wow thanks!
What a killer result. Amazing accuracy, the finished painting was definitely worth it.
Thanks. I think so too.
Awesome Michael! What development tool did you use to create your painting software? :)
It was written in Xojo running on a mac. It can compile a debug version for Raspberry Pi. Super easy way to quickly create a GUI on Pi.
The fact that this has less than 20k views is criminal. Get off your ass UA-cam algorithm! This is gold.
Yeah youtube.. what's up? 🤣 Thanks for watching.
Such a cool project! Can't wait to see what it paints next! Are you going to open source any of the design or software?
Wow thanks. It’s so custom that I don’t think I will. The software was written in xojo which most people haven’t heard of. It’s pretty hacked together.
@@Nerdtronic It would be great to take a look at some of the code just to see how you handle the motion, and I’m sure some people would be willing to help clean it up a bit!
Have you thought about using Bézier curves to generate the trajectories rather than segmenting curves? Or is the segment length just so small that it doesn’t make any difference?
I'm like "hey, this is a Roy Lichtenstein painting I have never seen before" - but then, I'm not an expert.
One thing got me though ... all of the Lichtenstein pop-art painting I know are from the 60s and that phone is definitely not a 60s (or earlier) style phone.
I am not sure if you did that on purpose but I find this a very nice touch. A great piece of art by a great artist.
Great observations. Thanks.
I have a CNC plasma cutting table that I’ve also added a sharpie holder to for funsies, but your painting is AWESOME!!! Subbed because of how fricken cool that painting looks! Great job to the both of you!
Awesome! Thank you!
What a beautiful channel! The design, construction, music, edition, absolutely brilliant
The machine was off by 2 mm, but hey - at least you had a resolution of 0.18 µm. 😂 Sounds like maybe the actuator was not completely necessary and a simple stepper would have done fine as well. Anyway - great work overall, and yes, I am impressed - thanks for sharing! I would recommend you use linear magnetic encoders for position feedback. That stuff is similar to what is used in digital calipers and is available relatively cheap even for huge machines like yours (well, at least the chinese stuff). Needs some skills to implement, but there's no doubt YOU can. Precision is 5 µm - so still way more than enough for drawings.🙂 But of course this would not prevent the head from sagging...
Thanks for the comments. The motors have the encoders built in. So I can just ask them what position they're in and they send that info back.
@@Nerdtronic Thanks for the answer, but sorry - there's a misunderstanding. Don't get me wrong: The Roybot including the software is a master piece and you can be very proud of it. And the actuators are fantastic devices - there is absolutely no doubt! But the built in encoder has absolutely NO clue of the head position. It ONLY knows about the motor itself. But the motor is just one part of the positioning system. The complete belt system including the pulleys plays a much bigger role. A linear magnetic encoder would completely remove issues due to belt and pulley accuracy, belt tension, including change of belt tension over time, plus backlash and even some issues due to temperature changes over such long periods of continous use. An alternative would be a refernce point to check (for example using a hall sensor) every know and then. But relying on a belt + pulley for accuracy is just not working well (as you have seen).
I see. Well assuming that nothing slips the motor returns an absolute position in the form of a multi-rotation angle, which simple math turns it into a position on the wall in mm (execpt for backlash). I said as long as nothing slips. Once when toward the end of the black pass I made a mistake and had the pen mech down when it was off canvas, then put a pen in and told it to draw something without raising the pen first. It was going to crash the pen into the side of the canvas. I went to quickly unscrew the thumbscrew and take the pen out and in trying to prevent it from ripping the paintng off the wall I caused the motor to slip in the gear and then the registration was lost. I was able to recalibrate it close enough to continue.
I think the backlash will go away if I tension the belt tighter. But I'm also going to make a gear rack and try that.
@@Nerdtronic Thanks for the detailed insight. Well, accidents like those just happen.
What I was trying to say is that even with perfect rotary encoder feedback and perfect calculation there will always be a delta between the computed and the real position. This could be eliminated by linear encoders. But I don't want to bother you with that any longer.
One additional idea: Maybe you can compensate at least belt (tension) issues by using two reference points on each axis - one near each end. You can then measure the motor steps in between. The value should in theory always be the same, but over time it will not be the same in practice.
I think higher belt tension might lower but never eliminate backlash. But maybe you can get it low enough for the purpose (i.e. to maybe 20% of the pen stroke width).
Being the curious type, I am wondering if you considered canvas deflection when programming the pen travel distance? If not, how would a solid canvas backing affect pen movement accuracy and ink bleedthrough?
I had planned to put styrofoam behind it but it wasn't possible. But no paint/ink bleed through. The back of the canvas is blank. I did consider canvas pushing in. I adjusted the pen depth depending on where it was painting.
13:55 I would just call that art imitating life imitating art imitating life! Roy based his artwork on printed illustrations where slight misalignments like that are common, he hand painted his artwork and now you’re copying his hand painted art with a computer controlled system that has simultaneously flawlessly reproduced the image as well as introduce a flaw akin to the original inspiration.
Haha yeah. I noticed that some of his dots weren't all that great either. Some of mine are more oblong than round near the top. I think it was caused because the canvas is a flexible cloth surface.
@@Nerdtronicit would be interesting to see other ways you could use this robot, beyond replicating existing works. One thought I had would be giant string art- you that thing where people use string as straight lines to create images. I wonder if you could do something similar only using the paint pens or permanent markers instead of string. I mean the whole point of using the string on a small scale is to create a straight line, but using your robot it would be easy to create a straight line with a pen. I’ve seen another YTr try to recreate string are using a CNC machine like yours, but all I could think is “why is he using string? Much easy to use a pen?”.
You pulled it off 100 %. Absolutely perfect. Even the imperfections due to pen holder sagging added a bit of realism and similarity to cartoon inaccuracies.
That's wonderful. I think the artists concepts being perfectly replicated by Roybot is a brilliant idea. Damien Hirst works out concepts and leaves paid employees to carry out the work so Roybot is just doing the same thing.
Thanks. Yes I think if Roy was sitll around he'd enjoy this project. He used a stencil to get his dots aligned. I used motors and software.
What an amazing art piece, and the machine being part of it makes it even better. Amazing.
Thanks
Bot Ross
Hahaha! That would have much more complex tool paths. No happy little accidents.
Wow this is fantastic! And here I wondered why there were so few videos over the past months, then you come out with this absolute gem. Stunning work, both artistically and from an engineering perspective. Great stuff :)
Wow, thank you! It was one of the reasons it's been a while since I released a video. Also some family/house projects that I didn't do videos on, plus being out of commission for about 6 weeks with covid.
Looks like lichtenstein to me
Right? Thanks!
When you find out how simple the original was done, you'll kick yourself
I've owned Lichtenstein prints. This was such a treat. You took 7 weeks. Old Roy probably took the same time. Amazing.
Haha. Thanks.
I cannot imagine how he got his paintings to be so consistent. He must have invented some tools to help him step the dots properly.
@@dgillies5420 you would be surprised at what a steady hand can accomplish
@@dgillies5420 For some of them he used a stencil and a toothbrush to push the paint into the canvas.
I think my prints might be more expensive than a Lichtenstein print. 😂🤣
The way I just kept saying "nice" as I watched this video 😭😭😭 so goooooood!!!!!!
Wow thanks!
This could be fun to combine with something like StableDiffusion. I've lost count of the amount of times I've generated an image only to think "Man this'd look great on my wall". There'd still be a bunch of human work involved (dividing the image into layers, all the pen replacement etc), but it'd be fun to see an almost entirely "machine generated" painting, with an AI generating the original design, and a robot making a real world painting of it.
This is really awesome! The painting looks great! Fantastic job by Roybot and you!! I would love to see Roybot do more Lichtenstein inspired paintings in the future.
More artwork is coming. Some will be in this style, others not.
To be fair, Roy Lichtenstein stole all of his famous painting from comic artists so even if you had copied one of his paintings it would have probably been totally fair.
There's blogs online documenting all of the instances of stolen art. The idea that he somehow "elevated" the art by printing it on a canvas is insulting to the artists who originally created these works. Look up "drowning girl" to see how badly he copies and comes out looking like the less talented artist.
You killed it, that is very impressive. That robot can compete with the other youtubers with millions of subs. Nice job.
42 million? Money laundering.
This robot painting the art would be an amazing art piece. Feed it AI images in the style of famous painters and you already have a very solid concept. It says alot about the world we live in and the fear of AI and robots taking over the creative industry. You could sell this to any gallery. People would watch this thing paint for hours in a museum.
Copying a Lichtenstein is perfect considering all of his art is plagiarized.
🤣
The gap is cool, looked like a missed print common in comics. It's a features !
I think fair use covers you copying a Lichtenstein painting. Especially since his art is straight up copying comic book panels. People paint reproductions all the time. Might get weird if you try to sell it. Excellent work man! Don't know how I haven't found your channel before.
Ha me too. Hope so. Glad you found it.
Amazing work! Loved this project
Thanks!
Now you need to connect it to AI and make it generate its own artwork!
I would love to see a part 2 of this! Maybe even integrate it with stable diffusion to create completely original artworks!
The painting came out fantastic but what is truly amazing here is your knowledge and skill in all the tech! Amazing work!
Unreal. I can't believe how good it looks at the end! You're a genius!
Great work. Can you elaborate on why you choose a robot/painting solution instead of a stencil/Litho solution?
Yeah.. I wanted to build a robot.
I strugle to find free time to take a bath, you got time to build a RoyBot! Amazing!!!
This has to be a valuable “work of art” as it is one of a kind using this technology. Absolutely astounding engineering…
Thanks!
I especially admire how you overcame the obstacles that occurred on the way... brillant!
Thanks!
Haven't seen you for a while, but it was WELL WORTH the wait!
Thanks! Hopefully it won't be another 12 months for my next video. I had some personal things going on as well. I'm starting my next video on monday, if I can just figure out what to print on this new Galaxy 1 printer.
Just having Roybot set up so the public can watch it work would be a gallery piece all of its own!
Yeah that would be fun.
Could yo upload a full description to the project and steps to build such a masterpiece? Would love to have a painting robot.
This is extremely impressive. Well done! Please keep the videos coming!
I'm finally back working on videos again and an update to Roybot.
So impressive. I'm curious how "autonomoous" you think you could get Roybot. Obviously the paint pens slow things down (maybe a tube and pump mechanism instead of the pen cartridge?) but how involved did you have to be on adjusting the paths beyond what you showed in the video? How small would the canvas have too be too do a single color artwork in a day?
I really had to babysit it. I was constantly concered that something would happen to make it ruin the painting. So many close called that I didn't get on camera.
That said.. My plan was to use a drink IV bag and fill it with the paint. Drill a hole in the back of a pen and 3d print a rubber adapter so I can attach the IV tube to the pen. Hang it a little above the pen. But swapping out pens was quicker than messing with that and having it leak all over the place.
@@Nerdtronic Thanks for replying, that totally makes sense. I feel like thats inevitable with so many robots, getting it to do the thing is hard enough but getting it to do it without maintenance or intervention is a hundred times more complicated and expensive. This was really amazing.
I like your idea, but totally agree I'd be constantly terrified of leakage
This is awesome. I wonder if you could put some kind of notches into the guide rail and use that to accurately gauge the travel instead of using a rubber belt. It'd be difficult to maintain the accuracy to notch them, but it would last longer.
The first paint head had a pot in there to feedback the amount of pressure on the pen. But it didn't work. I must have gotten a crap pot and had glued it in. It never gave accurate readings. The rubber bands held up for the whole print but a spring would have also worked. Shane from Stuff Made Here suggested that I didn't need the double set of rails. I used that big actuator there also and didn't need something that beefy but it was what I had on hand. And actually that actuator can act as a spring all by itself if you set the max torque. I just never experimented with that feature but will when I make robots that walk around.
I remember doing a Lichtenstein painting for art class back in High School by hand. I used a toothpick for the dots. It definitely was not as good an outcome as this but at that time I liked how it came out. A ton of dots though even though it was less than a quarter of the size.
When I was playing around with robotics and was working with a sort of robot arm, also not industrial or remotely close to pro level equipment I found gravity would screw up my positioning pretty quickly. Movement going down would be faster than going up. Was that an issue with this set up? I would have thought you might have has less issues if you made it table top instead of wall/vertical mounted like you had. Granted, vertical orientation takes up a much smaller footprint.
Very cool project. Love the result.
As a suggestion what about changing out the pen actuator for a solenoid it would be significantly lighter and easier to reinforce than the double slide sprung rack and pinion pen actuator and you could use a motor controller to have variable force so it would double as the relief spring
The pen actuator motor is way overkill. It's what I had on hand and already spoke the same language as the other 2 motors which are on the same can bus.
@@NerdtronicThanks for the reply and fair enough I was trying to think a way to make it more rigid without adding more weight/using metal parts, and to that effect what about swapping out gear system and the double slide for a radial or cylinder cam and a single set of rails with a 'driver' slide and the pen holder be a floating slide, with springs or a soft foam block in between and have the slide held back by elastic bands to tension against the spring and the cam, this would still use the motor as the actuator while simplifying the design and you could 3d print and mount the cam directly to the motor and lay it inline with the pen slide assembly removing the double slide system which seems to be the major mechanical problem in the current design.
Importantly I forgot to mention it in my first comment but the art turned out beautifully
The floating slide idea would probably also work with the current rack design which would make fewer parts to remake though if you used the current primary slide it would limit the travel.
I like to think that extremely talented/smart people like nerdtronics and stuffmadehere have always existed but nowadays there's a platform to showcase their talents.
You nailed it! It is 100% convincing. Amazing job on the design and software. You have super-human patience.
Thanks!
would love to see more of this robot!
This is an immense amount of work for 1 person, not only designing, printing and assembling the robot, but also making actual art with that whole document recording the whole process. Big thumbs up, that’s fantastic!
Wow thanks. More to come.
Absolutely amazing! Glad UA-cam recommended this video. Great job!
Wow thanks. Hope to do more stuff like this in the future.
Roy used a metal screen.
This is one of the more insightful comments here. I find that super interesting. I couldn’t believe that he would have free handed his grid of dots! Makes so much sense.