CNC Mill upgrades, SLS 3D printing, building 3 Vorons, and exploring some of my favorite projects to make in my garage - 2021 was a content packed year! As I move onto larger projects, I will stay true to my style of dissecting all elements of the machines and products that I am building. I will catch you guys in 2022 🎉
first video I've ever seen on your channel. I'm 300+ lbs. Do you think you could print me some chairs and what would you charge? 2:20 do you experience feedstock binding?
At this scale I'm imagining a granite slab would be more cost effective for a level build surface. This is exactly the type of project I want to do in a year or so. I'm so glad you are documenting your process for us!
Yep at this scale I can't imagine that aluminum plate would be anywhere near flat. Granite is good for that purpose but the thermal properties might be less than ideal (a ton of thermal mass is gonna be very greedy for heat and take ages to heat up cool down)
A honeycomb core panel would be rigid and light, or a torsion box design could work. It really depends on how flat you need it to be for 3d printing, bed levelling works great for surface adhesion, and printing with a raft could be used to create a perfectly flat surface. Personally, I like prusa's new segmented heated bed solution, especially for large format printing.
I'm curious how well that lead screw will work. I'm wondering if all that bouncing is going to be visible as an artifact in the print. Either way, it will be super cool to see a print that large.
Yeah that bouncing is a good indication of why you go for larger diameters on longer lengths (or different systems altogether). That whip is baaad for quality (and everything else)
Larger ball screw would be much better but very costly. At least he can usr the cheap lead screw to get something functional before he remortgages the house, lol.
yes, the speed definetly is wrong - as they sad "motor can run 9kmm in minute". And I thing the belt is better sollution, belts can handle hunderds of kilograms (harley use it intead of chain and other implementation) and also belt finaly produce lower vibrations .
Very interesting project. There's a lot of whip on the lead screws, you could do with a larger diameter of lead screw in order to resolve this. I'm not a fan of the bed levelling mechanism, it will be a lot of work to level and you'll likely find that you need to re-level it both periodically and every time you decided to change your bed temperature. I'd also be concerned about the level changing as the printed part's weight changes throughout the print. Mechanical levelling of a bed of this size is a losing battle. I'd suggest rigidly mounting the bed and relying on the Duet's mesh auto-levelling with auto-probing. You can do a periodic probe of all of the mesh points and the probe can also be used to set your z height with a single-point probe at the start of every print. The mesh auto-levelling will also correct for any tram that you have with the four z-axes.
Couple thoughts on your reasonable comment: Regarding the lead screw whip, the ideal case would be to use a 20mm ball screw at this length. I simply cannot afford to build a system around this component. There is a tensioning nut with thrust bearings on the actuator to remove some of the compression on the lead screw, but this will not eliminate the whip. I'm hoping that the whip is manageable at low enough speeds that I can get decent prints. I have used these same actuators in my CNC plasma table and the cuts were great: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-a-cyTMqul-o.html (just 1 actuator though not 7). Finally, the stand that I quickly threw together for holding up that 1500mm c-beam rocked back and forth when the motor was accelerating. You can really see this on the b-roll shot that zooms into the motor. This may be why the lead screw whip looked worse on video then in reality. The bed leveling mechanism is intended for getting the bed close to level at 60C. The bed is a cut off from a 4x8 sheet so it didn't even come flat. There would have been a significant amount of Z-axis movement when printing a layer if I relied on mesh leveling alone. For higher temperatures, I will let mesh leveling compensate for changes in the build plate because as you pointed out, I am definitely not releveling the bed between PLA and ABS prints. At max compression, the springs can hold 13lbs. If I can get near that compression across all 25 springs, then the bed be able to support around 300lbs (evenly distributed). A 4'x4x'4' block of PLA at 100% infill would weigh thousands of lbs, so I definitely won't be able to use the entire volumetric capabilities of this printer with this configuration. But I do want to talk about this limitation and considerations that need to be made for removing a hundred pound part from the bed. However, the build plate can be rigid mounted as configured.
@@DrDFlo have you considered stabilising the leadscrews with a loosely fitting bearing on a mini carriage that slides around on the screw independent from the extruder. You can make it stay halfway between the extruder and the edge of the frame with some clever cabling trick (I can sketch it out if you want). That said, I'm not sure if it will damage the screw, but the way it's whipping can't be good either right?
I have only watched part of the video so far. Regarding lead screw flex. I am planning a 3 x 6 ft machine which was originally for 3d but has morphed into other tasks. The rotational speed of the shaft is length and diameter dependent. Leaving it to whip will destroy sooner than later everything connected to it. I was looking at 20 diam x 4mm pitch for the required accuracy. Going fatter is a loosing game because the acceleration power needed also increases. I thought of doing as 'csiz' suggested but either the bearing or the screw will wear. Heat gain in the system is also a problem. I am going to try a tooth belt as a track, not as a belt. The other reality is to ignore the 6 mt/ min type speeds. Is that really a necessary requirement? The traverse speed only needs to be something a little faster than the extrusion rate. Particularly with fixed tooling like this, traverse speed is only a small fraction of the total extrusion time, and the joy of saving a few seconds per build becomes a fools mission in my view. Returning to the heat buildup problem. With regular high speeds the screw will lengthen differently to the rest of the machine. Product accuracy becomes thrown out as time progresses. I am separating the drive from the measurement system to overcome this. Notwithstanding this, different coefficients of temperature are not to be ignored when shed temperatures can range over 40 /50 degrees C. (Australia) I will be watching the rest.
@@DrDFlo Why springs actually? How about the bed being supported by locknuts, maybe with washers, would that work? As to the whipping, i'm not terribly concerned? You can always have some sort of a sliding coupler on the moving part.
What an amazing project! I was surprised to scroll down and see only 77k; this is the material and production quality you'd expect from a channel with several million subs, and with content like this, I'm sure you'll be there before too long. Great video, cheers
You might have heard of Extrudinaire, they're gearing up to release an affordable pellet extruder which will retract, and according to their demos, it may retract spectacularly well. It's also super lightweight, so it can be belt driven no problem. However it's only rated up to about 300g per hour. It's intended to be used with an external continuous feed system so it would work with commodity motion systems. I suppose just drilling out the nozzle like you did isn't quite enough, as the slicer expects the bead to be ironed flat, against the swelling that can occur right after the filament leaves the nozzle as the plastic molecules unfurl; but sanding a little flat on the nozzle should do it.
It looks to me like your listed speeds where off by a factor of 20. I'm very curious though, whether creating such a big printer is actually feasible. I'll definitely be keeping an eye out for the next video 🙂
Have you contemplated how this large of a bed will react when heated? I think you will find, that it will bow like nobody's business. Maybe you could use meshed leveling to compensate? Also are you segmenting the bed in several heatzone? If not I think you will have a crazy power consumption even when printing "smaller" objects. Looking much forward to follow you progress in this endeavor.
The plan is to throw the kitchen sink at the bed leveling problem. I have the springs for making large adjustments, the gantry will sit on universal ball joints allowing for quad gantry leveling, and finally mesh leveling. Regarding power consumption, I am only going to use this printer for massive prints or many nested parts so personally, the zoned heating is not worth the extra wiring. But I can definitely see an application where you would want to run this printer with such a setup.
@@RobertSzasz That's a great idea. Use the plasma table to cut some flexures around each of the mounting holes that allow XY movement but not much Z movement.
@@smellycat249 dunno if you realized this, but that's entirely the point behind crowd sourced construction - everyone contributes, builds, and iterates, and at the end of the line you have something awesome
Since you already have CNC of the gantry, a cool hack would be to make an automatic build plate leveler that uses a head with a distance sensor and a driver that can turn the bolts.
Have some minor experience with injection molding. They use a screw also but it move in and out of the barrel slightly to do retraction. The other thing is that the molds have gates in them to shut off the flow of plastic.What about making a gated nozzle. My other million dollar idea was a gated nozzle with a wide thin orifice for printing wide walls. The nozzle would have to rotate to make turns but I think it's something to look at.
I don't think those springs are sufficient. A printer like this is meant to print objects weighing multiple kilograms, and the extra compression resulting from the extra weight of an in-progress print would throw out your vertical tolerances.
Cool project one idea to manage the pellets is to have a small vacuum mounted on the extruder that sucks up pellets and drops them into a small reservoir on the extruder. They used this system on a printer called the part daddy made by see me cnc and is also the system used by many injection molding machines.
I'm not sure if this is a tech that exists, but you can feed additional pellets via air not unlike a venturi sprayer, up and down a wide flexible tube to the head, where the tube is attached directly including a small mesh-covered air escape. Simple optic sensor at the "full" mark in the feed could control the pressurized air source.
Please note the shape of the bottom of the nozzle is crucial for getting a proper line with a flat top when printing. Just drilling a larger hole will reduce that flat area around the whole and it will limit the maximum thickness of the printed line and will make it more difficult to get a nice flat top. Molten filament should be constrained from both the top and the bottom when printed.
To solve the pellet feeding problem you could add a pellet sensor, an optical one since the tube is clear could work, and make a routine to park the extruder to a corner where you would install a big automated hopper. So when the level of pellets goes down, the extruder pauses the print, moves to the 'refill station', reloads and goes back to printing. The hopper would need a set volume to unload or another spiral screw to measure the right amount to refill.
Nice project, if obviously not for everyone. But it does open quite a big range of different possibilities and uses in areas where 3D printing would otherwise not be even considered as a part of the workflow. I can imagine printing thin shells as a foundation foor all kinds of molds, like for laminatng resins etc. Or even as support structures for the final products. The extruder itself is hardly priced with the DIY crowd, but I think it's probably overpriced even for professionals, because I'm pretty sure similar product can be assembled without much technical expertise from the standard parts for injection molding machines and such. And all of them utilize a metal tube, corresponding auger, heater and motor. Nozzle, too, although it's a question if they could be used directly for this particular job, but as you are modifying these also, I see no big difference. I'd also explore the option of adapting, or even using as is, the desktop fillament extruder, like Welzoom B2, which is selling on Amazon for less that $700, and has very similar flow rate, nozzle sizes, and is even lighter than this one. I suppose if an elbow tubing was replaced for its rather small hopper, it might even work without further modifications other thn some means of attaching it to the machine.
I don't think your bed leveling idea is going to work the way you think. You are basically relying on the springs to push the bed flat. More than likely though, the springs are not going to be strong enough and a downward warp in the plate is just going to compress the spring and the screw head will stick up. What you need is a threaded stud rigidly bolted to the bed that then passes through the aluminum block you made with a jam nut on each side. Then you can pull/push the bed flat as needed and it would be locked in place by both jam nuts tightened against the aluminum block.
its hard to imagine how strong ABS parts would be extruded so thick. I can tell a difference when printing the same parts with a .4mm and .8mm nozzles. Something like this opens so many possibilites
2 года назад
You don't even need the nozzle, printing with thicker extrusion works fine to about 200% of the nozzle size.
13:45 at 6000mm/sec (22 Kph!) , the gantry would have made it through the wall on the other other side of the garage... perhaps you meant to say 600mm/s?
That leadscrew has a lot of whip. That is not going to be good. You really should consider using a belt drive system. I dismantled an old pick & place machine (Zevatech 810). It had a 600 watt servo motor for on X axis and 400 watt on Y axis. Here is a video of the kind of speeds it had. ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-NZwLPoxxYI4.html The machine I dismantled was a little bigger than the one in the video. It had about 36"x32" of travel.
There is a tensioner on the linear actuator that can further alleviate this wobble but not eliminate it completely. This is just a limitation of the system, but when I am printing an object that is a meter long a little wiggle in the walls is going to be hardly noticeable
This is the first video I watched from your channel. I am impressed with your teaching skills and that you obviously spent a lot of time editing the speaking part of your videos to make it flow fast and smooth. These are the signs of a great film-maker and editor. Looking forward to seeing how the rest of this project turns out!
We actually build a printer like this at work with this same extruder. Massive Dimensions makes an automatic feeder system for this system. I only have one question, which wire did you use to get the fans to power up? We've been working around it for a while. Also it will move faster than that.... :)
One question. Why did you go through the effort of designing and building such an awesome leveling system. Wouldn't surface mapping by a probe or something be the easiest route? You could have one solid bed and let the software do the leveling Cool printer project really looking forward to seeing some prints and your solution for the pellet feeding
Great to see this kind of build. I work for a company that makes that kind of sized 3D printers (this is a bit bigger than our current largest XYZ offering), so I really understand the issues. It's also fun to see the different design decisions you've made versus what's present in our printers! Really shows you how there are multiple ways of solving things! Good luck with the rest of the assembly, and looking forward to see you tackle things going forward.
how do you keep the overall cube from racking? seems like no matter how rigid the aluminum is, once that weight is moving around, that top square is going to move relative to the bottom one, isn't it?
them some nice toys you have there great use of them keep up the good work How much current are you figuring you will draw and if this works can we do bellows and heat the chamber....how many subs to heat the chamber
The audio levels need some fine tuning. Jumping between the voice sections to music/power tools is very jaring and requires adjusting everytime. Music/power tools are way way louder.
I would have gone for a CoreXY with HTD5 belts and geared stepper motors. You could have used any size NEMA motors this way to sustain huge amounts of speed and acceleration.
Even tho it sounds crazy.. But for the bed you are better off using a solid steel plate with one inch or more. Yes it will be hard to heat but if you consider that you print for a long time the energy expense is alright. Also you would have a way more stable temp. Another benefit is the added weight to the machine lowering resonance. Oh and also the bed will not bow like shot
You should probably opt for rack & pinion for your linear motion on such a large format machine. Or go for a thicker lead screw. There's a lot of whip in that leadscrew when even trying to move the print head slowly. Also, regarding the build plate, did you get cast aluminum? I've heard that that's important to make sure the plate doesn't warp at different temperatures.
For the 5mm nozzle that you drilled out, I feel like you should flatten off the face so that the edges aren't so pointy. Most nozzles have that flat area around the orifice, and I think it helps with print quality.
2 года назад
100% this, I think I tried a bored out nozzle at one point and it was awful. You really need that flat area.
I have never seen your channel before but the quality of this video, content, production, pacing, was incredible, true passion caught on film. Congratulations to you for your hard work and significant effort on the channel and series.
Of course! My goal is to have the workflow for setting up a print on this massive machine to be the same as one of the Voron’s that I built (Automatic bed leveling, nozzle probes, etc). Should hopefully just be a longer wait time for everything to come to temperature.
Loved this video. 3D printer enthusiast here, and I can't help wonder if 10 years from now, all mainstream printing will be from pellets or "cartridges" of pellets, rather than spools. Very cool that dies can also be added inline with the print process.
Triangulate your frame for rigidity might help accuracy. You may also consider a less thermally reactive composite, try filling the extrusion with and epoxy garnet slurry.
D-Flo ...... just an idea (it would maybe make a good video if you could figure it out) - Make a tool head that scans the bed and uses a motor with a head bit on it to automatically adjust the screws one by one by the data from the scanner
2 года назад
Haha this is the exact idea I had about how to make an adjustable bed for my bad back ;)
Hey David, I'm building a 590mm cubed printer, so basically a 2'x2'x1/4" build plate. I got the ATP 5 plate from "Midwest Steel and Aluminum" which was about $120 shipped. I priced it out and a 4'x4'x1/4" tooling plate (what ATP5 is) would cost you ~$500. No idea what the shipping would be. For my build surface I'm using FR-2 (PCB material) from "Professional Plastics" which was about $25 for a 2'x2'x1/16" sheet, shipping was about the same price. I'd include the URLs but RU-vid is being difficult. Looks like your printer is going to be a monster. :)
Bring that water level up. We cut with the bottom of the sheet touching and get zero dross. We do submerged for really long cuts. Your THC, speed and voltage settings play a much bigger part.
If you wanted more threadform through thin material running a rollform tap would have been an option, though don't know if the exit side extrusion would have messed you up. Though I've only seen rollforms run on rigid machines, big bucks to accommodate the size of that sheet.
Your figure on the E3d hotend is about right with a volcano hotend and a 0.6mm nozzle. But it can do a spool in under 3 hrs with a supervolcano nozzle. Cool build though, looking forward to seeing some speed tests! 😎
An internal valve like a ball valve that won’t change in internal volume would be perfect right before the nozzle to prevent leakage. Not sure how that would be done though. For the bed, what about small steppers with screws and a nut to adjust the bed? You could move over the nozzle or sensor over a stepper under the bed and extend until it touches the sensor? The whole bed would have to be allowed to flex a bit. Maybe a 5x5 grid of “bed jacks”?
Cool build! A heated chamber would be a very interesting follow-up once the main assembly is done. As far as pellet feeding goes, I once saw a pellet extruder system that used a vacuum (like a shop vac) to draw pellets from a bin into the hopper through a flexy hose. It ran only at intervals, and I'm not sure how it was controlled -- maybe with a breakbeam across the hopper to tell when the pellets ran low?
Pretty much. In blow molding, you'd usually have an industrial vacuum motor on the hopper on top of the extruder gantry sucking up pellets from below. IR beams were, and likely still are, typical fill sensors. The extruder screws I've seen were able 3-4 meters long with heating bands and thermocouples around the barrel.
13:05 Hey Dr. D-flo. Leadscrews (and ball screws), and any general shaft have a critical speed that decreases by the inverse of the square of the length. And after that, it will start to resonate. Assuming you have a 4 thread 2 mm pitch screw of 1500 mm in length and 8 mm diameter, and that your screw is supported at both ends (instead of fixed) your critical speed might be around 172 and 344 rpm (with a safety factor between 2.5 and 1.25 respectively). In this demonstration you mention that the actuator can go around 8000 mm/min, which translates into 1000 rpm with the given pitch. Therefore, you're above the critical speed, which explains why your leadscrew is vibrating that much
They make fan style through beam probes for vacuuming pulling up pellets in industrial molding applications. Most people use a centrifugal dust extraction hat on a small shop vac to feed the pellets. You should be able to set up a control scheme in either the duet or via a pi for the vacuum on. Try not to step on the bed like that. You definitely put some bow in it after than. The spring loaded beds look like a good idea. But they often minimize a bent sheet that is exaggerated when it heats up. The best way to adjust those screws is with mesh leveling in the duet firmware. You can set your grid to be somewhere close to the screw head and it'll tell you the offset compared to the z zero. I know ABL is a bit silly with all that adjustment. But on my cubic meter build the able with independent z axis motors was pretty awesome. Setup took some time, but the control scheme was incredible. Currently make an idex machine running that reprap firmware and it reminded me how awesome it was for my giant machine too.
For a noisch leveled bed you could go Prusa's new way of putting more single beds together to one big. So the bed is less likely to warp via heat if you put the right engineered base under it. Also at this size it's probably mich easier to have the bed stationary and the XY-gantry moving up and down. The bed then can be pushed upwards by screws from underneath and be leveled this way. So you could use as many screws as you liked. Also your ground base then could be wood or so with the screws in it.
For the lead screw, what about some 3d printed spacers with bronze bushings? The spacers could slide within the channel. Allow the print head to push them side to side, but have them return to the middle of the empty space with something like a constant force spring (tape measure thing) and a stop pin?
You could always use a level sensor on the feeder tube and hook it up to an air venturi loader system. This is very similar to how virgin resin is loaded in plastic injection molding systems.
I greatly admire your talent and knowledge. Greetings from Medellin Colombia. You think that by following your project repository step by step I am capable of making a machine like yours. I want to start. thank you
A 5mm nozzle? No thanks I seldomly need anything more than 1mm personally. It is a cool project though. I just personally would never have a use for it aside from making large containers or something that I could buy at the dollar store.
Wait... 6000mm/s is 6m/s. That motor was running several seconds, and no way in hell is that build more than 6m long. Something is off here, and I'm guessing it's that you ignored the "gear ratio" that the screw adds to the numbers.
What if you took sugar and made it into rock candy and broke it down to granules the same size as the pellets, you think it could print in melted sugar? add a dual extruder and you can print out giant candy canes and ribbon candy
I probably should read more comments to see if the subject has been broached? First, great project. I've had the extruder idea in my head for some time because? As a kid, I worked in both injection molding and extrusion blow molding (about a hundred years ago). Anyhow, one point of focus, for the extruder and feeding it. Pelletized plastic takes on more moisture than filament. You want a dryer on your hopper and you can locate your hopper (think 10lbs, 100lbs) on the other side of your shop if you want. Set up the dryer on the hopper. Run a vaccum tube from the hopper/dryer to the extruder. You can setup your feed system to run automatically to keep your feed tube full. With that in mind I'd look at say a one or two pound? Three? Think volume and how often your feed will have to kick on. A sizeable mini hopper on top of the extruder. This will allow you to keep your plastic dry and feed your extruder without having to constantly be involved. Also it will allow you to do your color blends on your pellets in larger batches which will allow you to get a consistent blend accross prints. That's my two cents. Hopper/Dryer/Feed. Any moisture will become quickly apparent, more so in large formatting extrusion volume. LOOKING FORWARD TO SEEING MORE!
3000-6000mm/s - Are you serious? It is 108-216km/h, not each car can reach it. Seems you made mistake during calculations from "Imperial" to "Metric" system. Your linear motion speed is about 30-60mm/s now.
My covid project was to build a 4 x 5 x6 foot printer. I turns out to be very similar to yours. The heated bed has been a problem. I have used a donated 2x3 foot 240 volt 2160 watt single silicon heating pad on a 8 mm aluminum plate ( Not thick enough ). It heats great but because it is a single pad I am unable ( I think ) to pierce it to mount adjusting bolts and springs near the center. Unsupported it bows in the middle. Mesh compensation helps a little but to date I have only been able to utilize about half the bed. It has been a fascinating learning experience, since I had none, but it has definitely been worthwhile. Particularly the learning g code part (argh) !!! Very cool video and looking forward to your next.
I worked in the plastics industry for about 15 years (Primarily in blow molding and Injection Molding) the best way to do the supply (if you have the headroom is to use a vacuum based hopper system, Noise might be an issue though, I only know industrial applications, so might not be good for you. The biggest issue with pellet based extruders is cleanup. If you need to change colors you'd have to take the screw and barrel apart and clean everything to get all the old color out of it. I dealt with screws and barrels 3-12 ft long and hot (not fun), to clean them and the dies to change color or material. BTW your screw needs to be removed and everything cleaned while hot you you need some major heavy gloves. Found a RU-vid link showing what I mean, these were about the scale I've dealt with in my past, be glad you got a little screw. Not trying to scare you just showing you what might be needed when the time comes for actually making things. ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-67nxFiXOrhc.html&ab_channel=LearnIndustrialMACHINES%26AUTOMATION
