 Hello, everybody, all of the many, many people. Okay, there's nobody in here, but. So I'm going to talk about 3D printing in Linux, and I'm going to give you an overview about what 3D printing is, and then I'm going to talk about how you can go from an IDEA to a physical object with free open source software, and then talk about some of the hardware that's out there for 3D printing and the various support of it from Linux, and some other options that are out there if you want to do that sort of printing on Linux. To start with, 3D printing is the process of making a three-dimensional solid object of virtually any shape from a digital model. The most common method of 3D printing is when you're doing something that's additive, so basically you're taking your material and you're adding it in layers on top of a material. This is different from personal shop tools which rely on subtractive processes where they're either cutting or drilling or lasering out something from a material. It's not a new technology. 3D printing is something that's been around since the 1970s and it's only been in the last five years that the cost of 3D printing has dropped significantly because of patents expiring in the space, and so it was possible for people to build their own implementations without having to worry about having to pay the greater their patent costs, and as the patents started to expire, people felt brave enough to do this, and so while there were 3D printing previous to five years ago, there were certainly people that were doing it in violation of patents and building things up, like the early people doing the first generation reprap stuff, it's only been recently that it's been sort of consumable by people that are not already really well-versed in shop tools. So there's lots of different ways to do 3D printing. The most common method is what's called Fused Filament Fabrication, or FFF. Now, you might also see, if you look on the internet, this is called Fused Deposition Modeling, or FDM, but that term is trademarked by Strabasis, which is one of the big 3D printer manufacturers of proprietary printers, and so I don't like to use it, so I'm gonna keep using FFF. Now, FFF is when something, there's a lot of different ways of doing this, something is driving the source material, which is usually a plastic filament, into a heated nozzle, and so the heated nozzle melts that material and extrudes it, is either a small bead or a small line, and the width of extrusion is controlled by the diameter of the hole in the nozzle. A lot of designs also have a small amount of tanking where the nozzle is melting the material, and it's being kept in sort of a tank reservoir, and then as pressure is applied down in, it's forcing the material out of the tank that's already being held in the liquid state to keep a constant flow going through the printer. Sometimes that's done by simply having a nozzle be shaped like this, as opposed to being shaped like this with a timing point at the end. Sometimes there's actually more of a tank in the design of the extruder. The nozzle then moves horizontally vertically, making a layer, and then moving up, which is what's on the 3D printing, is called Z-axis, and printing another layer until the object is complete, and so you can imagine that it's printing a layer, moving up, printing another layer, moving up, and so forth, and so on. Now, there's a couple of different types of materials that are commonly used for this style of printing. Acrylonitriputidine styrene, which is commonly known as ABS, this is most familiar to people because it is the plastic they use to make Legos out of. Now, they do not 3D print Legos, though you can, if you wish. You can't call them Legos because in the Uri-Laking Material Market, but you can call them stackable bricks. ABS is a strong plastic, and that's what people use it. When they make Legos, they do a molded injection model, where they inject the plastic into the mold and out from the Lego bricks. It requires a high nozzle temperature about 230 degrees Celsius for most ABS mixes to get a reliable print out of it, and it usually requires a heated bed to get a good adhesion for the first two layers. The downside of ABS is that it's not terribly environmentally friendly. It smells like burning cancer when you're printing with it, and you're not gonna fool anybody if they see you printing with ABS, and they're like, oh, man, that doesn't smell. So there's another option, which is PLA. PLA is polylactic acid. PLA is biodegradable, it's bio-friendly. It smells like pancakes when it's printing. Some people say it smells like syrup or pancakes or a mix of pancakes and syrup. It is derived from renewable resources and not oil like ABS. Sometimes we see people mixing it with PHA nowadays, which is, I'm gonna have to look at this, polyhydro-alpinoate. PHA is also biodegradable, but when you mix it with PLA, it improves the strength of PLA and gets it closer to the strength of ABS. So if you're buying PLA from a Thunderlight ColorFab, they're gonna mix PHA in there to make that PLA stronger, and it still has the same rough melting point. PLA's melting point is slightly less than ABS. It also can be printed on printers that don't have a heated bed. Though if you do have a heated bed, if you turn it on below, you'll get a better result out of PLA print in general. Now, when the manufacturers make these plastics, they produce them in pellet form. It's the standard form for plastics. You can't really print a pellet. Although I did see a printer once that was 3D printing with pellets, it was just 30 feet tall. So what you want to do is you want to be able to take those pellets and then turn them into a string of filament, and you want that filament to be of a consistent width. Now there's two standard widths in the 3D printer industry. One of them is 1.75 millimeters, the other is 3 millimeters. My printers use 3 millimeter stuff. A lot of printers that I don't own use 1.75, so there's still both are being widely produced to help the market out. Cheap filaments out of China are going to vary widely in quality. Sometimes they're going to have air bubbles in the filaments. Sometimes they're not going to be the consistent width and you're going to get bad prints as a result. Sometimes they may not actually even be the material they sold you as it being. I had bought a couple of schools on discount on Amazon from China that said they were ABS. But when I got them back, they printed awful. And I said, why is this printing so poorly? And a friend of mine says, probably not ABS. And he said, have you tried dissolving it in Salmon? ABS dissolves in acetone, which is nail polishing. And so I took a sample of it, I put it into acetone and three days later it was just sitting there sticking out of acetone. It's not ABS, and that's why it printed like that. So buying good quality materials definitely has an improvement on the quality of your prints. If you're not spending $40 to $50 on your ABS, you're not getting good ABS. You can get PLA, good quality PLA for $30 to $40 US. And that's not to say that you won't find a deal on good quality, you know, sometimes. But in general, if you're paying less than that, you're not getting good result. Now there are some less common materials that also work with FFF. High-impact polystyrene, when it's not colored, it will print translucent. It's also pretty good because it prints it roughly the same temperatures as ABS. And it's a little easier and a little smoother. It has a bit of a sheen on it, so people like the look. The downside to high-impact polystyrene is that if you print something large on it, the bottom will literally burn. And you can see the char marks on the bottom layer. So for small things, hips is great for introductory printing, but for big things, you end up having to paint the bottom because it's gotta burn. Polyvinyl alcohol PVA is nice because it dissolves in water. PVA is basically just glue stick. And it's very tricky in finicky to actually print with, but a lot of people like printing PVA on a dual nozzle printer with PLA as the main material because the heating points are roughly the same. And then you can have the supports in the model, which I'll talk about later, be PVA. And then you can throw the whole thing in a bucket of water and leave it overnight. And the PVA will soon dissolve and you'll be left with a plastic one. Tea glass, prints almost clear and has the right capacity to act as a light pipe. So a lot of people use it for printing specifically light pipes for electronics so that if you want your LED on your Raspberry Pi to start on the board and then come out the other end, you print this little pipe and you put it in your case and then when it's on it, it comes out the other side. And then there's a new material called Eastman Amphora. And Eastman Amphora is light ABS, but a lot stronger. And so it doesn't shatter or dilaminate like ABS can on your tension. If you have got one of the bags, the green bottle opener that Lulzbot printed that has the Fedora logo on it and the Flock logo, that's actually printed in Eastman Amphora. And they did that specifically so that it wouldn't dilaminate or shatter when you went to open a tough bottle. And then it's there. Yeah, it totally works. Yeah. Now there's even some more unique materials that you can use for FFF. There's nylon, so it's a number of different nylon. Nylon's are very finicky to deal with. You have to usually be very experienced in doing these to get them right. You wanna put down a lot of stuff on the bed. A lot of people who are doing free printing use tape. There's a number of different kinds of tapes you can use to help that first layer stick. Nylon stick to tape, it won't stick to glass. The tricks for nylon are usually to actually spread glue stick on the bed and then bring it to heat. And then you might have a better chance of having the nylon stick and it still might not. There's a nylon called bridge that was released in March of 2014. And that tends to print better than a lot of the other nylons. Since then, the company that makes most of the nylons has made a couple more that are really reliable in printing with varying degrees of why you wanna print that. Some people would like to do crazy things like try to run an edge trimmer line through a 3D printer because functionally that's just a plastic line. The biggest reasons why you wouldn't wanna do that are it's not set up to be good for 3D printing. It's inconsistent with materials that are in it or whatever was cheapest that one for the plastic factory. And so some people have reported success with it. I don't recommend that path. Labric is actually a stone filament. The company that makes Labric basically takes stone, pulverizes it to dust, mixes it with a plastic solvent and it all gets stuck together then they string it out as a filament. And when you print it, most of the plastic is burned off and you're left with majority stone dust. So you have an object that has to feel and texture a stone. It does not have the same strength as pure stone. But for printing little models or toys or things that want to have that stone look and feel you can use Labric. The same is true for lay wood which is just wood pulp which is for the same plastic. When you print it, most of the plastic burns off and you're left with the wood. You can actually do fun tricks when you're printing the lay wood where you adjust the heat during the print, print grain into the wood model. NinjaFlex, which is a thermoplastic elastomer is a new material that came out about a year ago. There's also SemiFlex, which is another material. They're flexible plastics. They have the same basic qualities as a rubber. And if you have a printer that is able to drive forcefully enough the filament in there without clogging you can actually print out these flexible items. The Lowell Spot TAS printer has a special add-on that the open source community specifically designed to do nothing but this flexible material printing. And so it can do very reliable prints with this. If you try to print it on other printers there's a whole internet sort of sub-community of people that are desperately trying to grab how to hack their printers to reliably print their flexible stuff. They can be done. It's just a little harder. Can they do that for the Mini too? The Mini doesn't currently support the flexible materials out of the box because it doesn't have the interchangeable head unit. But they're looking at figuring out if there's an easy way to add that in the future. Is this a filestruder thing or? The Flexi screwdriver, yes. Now this is my Lowell Spot TAS. This is actually the TAS 3. It's a few generations back. Printing at Red Hat Summit 2014. Sadly, it's the only video I have of my 3D printer actually printing something. Most of the time I just take pictures of the fails that it finishes at the end. But it's printing some Fedora logos on the bed there so it gives you an idea of what it looks like when it's in action. Now, how do we get to that point? 3D printers are not magic. You just can't go to the printer and be like, print me a puppy. It's not gonna do that. It doesn't know what you're talking about. And in general, 3D printers are really stupid devices. They don't have a lot of intelligence. They basically are expecting someone to tell them, move my motors like this at all times. And I don't know what I'm doing. I'm just doing what you tell me. And so when you talk to a 3D printer, you have to send an explicit list of instructions for it to do with its motors and its thermocouplings and bring it to this temperature. Turn the fan on at this speed. Move my motors this much. So what we have to do is we wanna start with an idea. Okay, I wanna make a puppy. Great. So the first thing you're gonna have to do is generate a 3D model of your puppy. Now, there's a couple different ways you can make a 3D model. You can use visual CAD tools. In the proprietary world, this would be things like solidworks, AutoCAD, anything that has the capability of doing more than 2D CAD. In the free-noving source world, we have things like FreeCAD and LibreCAD. They're not awesome. I'll be honest, but they are showing progress. Both of them are in active development. They both are adding different features. And so what I have found is that people who have experience in CAD look at one and go, nope, not doing that. They look at the other one and they're like, well, that's almost just good. That's the right sort of tool. So it really boils down to what type of tool you've learned on, what you're comfortable with. So try one, if it's not good, try another. There are some other lesser-known CAD tools out there in the open source community, but most of them have been abandoned and they don't have active development going on. There's also 3D animation tools, so things like Blender, which are designed specifically to let you create 3D models for the purposes of rendering and for animation. And they work great for 3D printing as well, as long as you remember that you can create 3D models that your printer can't actually print. And it doesn't matter in the Shrek universe if Shrek has air gaps and weird holes inside of the model that the printer then tries to interpret and doesn't know what to do with it. So in general, you're using a tool like Blender to make a 3D model. You want to look for the number of tutorials that are out there on the necessary steps to take your model, which looks good in Blender and renders great on your computer, and then making them 3D printing. The key word you're looking for there is usually watertight. You want to make sure that the model in Blender is actually watertight. All right, now let's say clearly great series of video tutorials from Edir I don't remember his name, sorry. It's on Blender website. It's using Blender specifically for 3D printing stuff. You have actually to buy it, but it's created commons and it's really very good. So if you like Blender and if you have used Blender before, also if you haven't used Blender at any time, you can learn how to design a 3D printable model. Now the last tool is the tool that I tend to use most of all when I'm creating models which is OpenSCAD. OpenSCAD is a programmers interface to CAD where you literally write a program and then you hit the compile and render button and it shows you the 3D object that gets built as a result. So you can call functions for specific operations, addition, subtraction, merges, unions, modifications, move, layer, shrink, grow, scale. All of these are functions that exist in this programming language that's sort of a pseudo C language. And then you write your program and it generates the result. So most of the models that I've created which is not a time to be fair, I've done in OpenSCAD. Or you can use an existing model. There's a large community of people on the internet that have been uploading models for things they've built and putting them out there under permissive licenses, some more permissive than others. The two big sites that are out there for open source friendly models are Thingiverse which is kind of open source friendly and BLD3R, BLD3R is sort of a people that were angry at Thingiverse when they created one and they said, well, we're gonna do ours and believe the open source everything here is gonna be either open source or greater commons. It just doesn't have as many models as Thingiverse does. Thingiverse is a much larger universe of models. There's also a lot of other sites that have models that aren't open source but are free to download. Things like Mine Manufactory has a lot of really high quality models that people have uploaded that are free for download. Some of them are actually created commons with a vast majority of them are. Now, some 3D models are not printable using an FFF printer without having to do all sorts of special games. Now, FFF printers print from bottom to top. This means that when you have what's called an overhang without a gradual change of degree you're effectively trying to print thin air. So to visualize this, let me stick my arm out. If I was attempting to print myself in this representation, my arm is overhanging thin air. It is not a gradual overhang like this but rather a very sharp overhang. And so when the printer gets to about here the printer doesn't realize that it can't print in thin air the nozzle will move out and start extruding plastic in thin air. This results in spaghetti, any number of terms for utter print fail when you have just the sea of extruded filament in thin air. It will not result in the model you expect. So what you can either do is design a model that has a gradual change. It doesn't have these explicit overhangs which generally is when you can do a smooth or a gentle curve, you can make the arm sort of curve out or whatever it is that's overhanging or you can print with support. And so most of the tools will take your 3D model detect your overhangs if you tell it to and then add support material underneath it. So in my case when my arm is out like this it would generate a mesh of support that the printer then prints up to that point and then when it gets to the point of printing the arm there's a layer of plastic meshing that would hold up the actual solid piece that I want. Then when you're done you either break off or cut off the supports and are left with the final product. So that results the problem but there's a lot of post work you have to do on the piece and if the details are very intricate that require the overhang it's very easy to snap off or cut off the thing you were trying to protect with the mesh on the overhang. So you have to be careful with the model. Now we've got a model we want to be able to turn this into layers. When the FFF printer is printing it needs to be able to print a layer. One layer up, next layer up, next layer out. So we need to slice the model. Now when you slice you're going to be slicing to your specific printer. You can't slice to the universe. It depends on the type of printer you're using the size of the nozzle you're using the size of the bed the G code specifics for that printer. So you want to go ahead and have your slicing tool ready and configure it for your printer before you start slicing models. This is why people don't generally share G code for their models and instead share the actual 3D object before that. Now the standard file format for 3D printable models is STL. The 3D printer doesn't know how to speak STL 99% of the time. They speak G code. G code is the rough equivalent in 3D printing to post-script. It's a set of machine instructions that tell the printer, do this, do that. If you have any shop equipment your shop equipment talks G code as well. Now slicing programs take the STL they allow you to set the options you care about shrink it, grow it, duplicate it do all these things and then say okay now slice it to G code and then you can slice all of that code into something that your printer will understand and then you can give to the printer. The most common open source is Slicer with a 3 instead of an E. Slicer is a giant pile of pearl. It is somewhat notorious for being slow but working pretty well. It just wrapped a major version recently and the new major version has a lot of interesting new features, it is still under active development. A lot of really smart people who understand math way better than I ever will are adding a lot of really interesting things to the slicer environment to try to both improve performance and add interesting features. One of the interesting things that went into the new Slicer is when you're printing a 3D object you don't need to print solid plastic in the middle because no one's ever gonna see that. You just want it to be strong and that's generally it. And you can adjust what's called infill the plastic in the middle and say well I'm gonna print less of that and you can turn the percentage of infill down. If you put it down to zero it will try to print a hollow object. Some models will take a hollow object some models you'll end up printing in thin air if you try to print a hollow object. So the trick is always for most people is to minimize the plastic that has to go into the infill while maximizing the strength of the models so the model doesn't collapse on itself or fail to print. So one of the things that went into the new Slicer was a new patterning for infill that minimizes plastic but geo but the geometry of it is such that it has a structure that is super strong. It looks weird the layout when you look at it but the infill inside of it uses a lot less plastic at a great strength. So you can choose that type. I tend to not be so fancy with my infill I tend to just tell it to do cross and then the other way and then this way and then the other way. There are some people who do honeycombing that's been around for a while and again you can tweak that for your model and your preferences. There's another slicer that's built into Cura I'll talk about Cura a little bit later. You can see most of the slicing part now is C++ and it's generally much faster than it was a year ago so there is no more steep problems I guess. The interface is still, each bench was very bad parallel and I've sanded some benches and it's very bad parallel even if it's parallel it's usually very bad parallel. Yes. But you're right and I think I timed even the new slicer versus Cura and I think Cura is still being slicer in every test I've run against it on speed so. Probably, but it's still so little time that it doesn't matter. It depends on when you're slicing giant models which is what I'm usually throwing at it. When you're slicing really big models, it's evident. Slicer still takes about 45 minutes if it doesn't die and Cura will still do it in about five. I remember any slice that was like this on the OEM or Scameport, it took like three days. Oh, Scameport? Yeah. It ended up with 60 gigabytes of SSD as a swap to actually do. Yeah, the slicer, the Scameport slice tool that no one should ever use ever again but existed before Slicer did, yeah. Absolutely, ugly code. Everyone had looked at it and I was like, oh I'm gonna improve it, they ran away screaming. So now we've got our G code, now we need to print. So there's a couple different ways you can do this. If you wanna control the print from your computer and your printer is connected directly to the computer you can use printer control software. Print run is one that's out there that's been around for a while. Cura is another one. Cura actually has its own slicer code built into it. It was originally built for the Ultimaker printers but it is open source and so Lulspot worked it for their printers and a number of other people with specific rep-rap fan and printers have been using it for their printers as well. There's also Repeteer host which is out there but only version 0.9 and older are open source. The upstream for Repeteer host after 0.9 decided he wanted to make some money off of this and proprietorized all the features that he added after that. So it basically, most of his community said, well we're not gonna use your stuff anymore and got away from it and stopped using it. So there is still Repeteer host out there for running but it's not open source and we don't ship anything newer than 0.9. Now this control software basically just establishes the connection to the printer 99% of the printers I've seen show up as a USB serial device. Sometimes they detect incorrectly as a modem and network manager tries to build a connection to them. It's pretty easy to tell network manager not to do that. You just have to tell it to know that's not a modem. You don't need to try and dial it out. Do they make a 300 bottom model squeal when they do it? The printer may make an awful noise but I don't think it actually does make a thing. So and then when that connection is established then this software sends the G-code across the line to the printer and the printer buffers it sometimes depending on whether it has a flash or not some of the cheaper printers literally are taking G-code line by line across the wire and printing that away and then if anything happens that print's gonna fail sometimes they buffer it onto an SD sometimes they buffer it into flash. The smart printers at least are doing some buffering of the G-code. Usually they're sending it across once holding it and then printing the entire time. That still doesn't mean you can disconnect your computer and walk away because sometimes that will cause the printer to lose its mind and stop working. A lot of printers can print without being connected to a computer. They'll read an SD card that's plugged into it and read the G-code off of that and print it. So the NISO printers will have an SD card slot and you can put your G-code on that SD card and they're directly from it and then you can take your computer and walk away and your computer locks up and you don't have it. You can also do manual motor and temperature control by these connected programs by sending the specific G-code instructions across. One of my favorite tricks that I have done in the past is to write a series of G-codes that makes the stepper motors move but not extrude any plastic or heat up. This causes the printer, which makes a lovely wine for each, that is slightly different for each of the stepper motors to play a little song. So I can make my printer play the Tetris music, which is kind of special. Not probably. No, when the print fails, the printer actually does my artwork shake. It's this dip. Now, the print goes and sends the instructions off and again, the printer is stupid. It is literally just, I moved here, I extrude this much plastic at this temperature, I move over here and do it some more, do it some more. It really has no concept of what it's done before, it has no concept of state. If you pause the print and hit the extruder or knock it over or shift it slightly and then unpause it, it'll be like, oh, it's so good, and it just happily just keeps doing its thing. And so it's really easy for printers not calibrated well to get out of alignment during the course of a print and fail as a process of that. And so a lot of people who build their own printers spend a lot of time struggling with this because they don't have their belts tight or they don't have their calibration. And the printer doesn't know it's doing anything bad, it's just like, I'm doing what you told me. Now after the print is done, you have to remove the print from the bed. Now some materials need to cool first or when you try to get them off, you'll warp them and tear them and shatter them. The most common mistake that people make when they start through your printing with ABS for the first time is that suck at that print goes, they go at it with their little razor blade or they're nice to get it off the bed and it bends off the bed and you end up having this well. One of my legs has bent off like this because that's how I put it off the bed. You really wanna wait for ABS prints to cool a little bit before you start attacking them and getting them off the bed. Small objects usually don't show warp in but even non-lodal fedora logos that we printed, when you get under them when they're just done, they peel off. PLA does not need that much of a time. It doesn't need to cool too much work on it off the bed. PLA will usually come right off when it's done. If it's still enlarged like that, like a little flat surface and it's hot, you literally just peel it off like that. You're gonna be hurt. But if it's like a normal sized object, it's fine. That's true. Now, all FFF printing, even at the highest resolution I've seen results in visible layer. Even when you're maxed out, you can still see the layers on the print. Now, if that bothers you, and most people who are 3D printing are not terribly bothered by the layering, but some people really want a smooth look. They wanna know how to smooth the print. So, there's a couple different ways of smoothing a model. One thing you can do is you can usually brush on a thin layer of the solvent for the material and smooth out the layers a little bit. You're actually melting the plastic when you do this, but if you carefully you can do it. There's a couple different techniques to do to do this for ABS specifically. ABS is nice because the solvent is cheap, easily accessible, and you can literally go to any drug store to get it. It's nail polishing. And it's very good for your health. Yeah, it's good to smell at the same time. It's also slightly flammable, but you wanna be careful. Don't throw it on an open flame or anything. So, the two tricks are, there's two methods that are out there right now for doing acetone, which is nail polishing. We're smoothing of ABS. One is heated smoothing. We're basically, the method that I, the way I describe this is you take a glass bell jar, you take some aluminum foil and you make a little stand and you put the stand inside the jar and then you put your ABS object on top of the aluminum foil stand and then you pour ABS into the bell jar, not out of the object, but underneath the stand and then you put the whole contraption open on the top onto your heated bed or your printer and you turn the heat on the heated bed. The acetone heats. It turns from liquid to vapor and the vapor smooths out the top of the jar, comes out the top and it passes the model in a relatively even manner and then it smooths the model. Now, if you leave it in there too long, you won't have any model anymore, it'll be gone. If you leave it in a little less than that, we'll have a drippy model that looks like, you know, Salvador Dali paint today. But if you leave it in there for a short window of time and then you take it off the heat, which will end up with something that looks very smooth, it looks almost extrusion total. The other mechanism of doing it if you're not fast is to do cold paper smoothing where you take nail polish remover, you soak some paper towels in it, you line the inside of a top of work container with the paper towels, you make your little stand, you put your object on the stand inside this container, you put the lid on it and you come back six hours later and it will have eventually, in that way it will have turned into gas and passed across. And this is, people tend to prefer this one if you're willing to wait because you get much cleaner results and there's much less likelihood that it will turn into a melty drippy, missing mess. And you're less likely to back out. It's also not involving a heat source and a flammable liquid. So. The top of her jar is made for ABS. Yeah, you would want to make sure that you're not using a plastic that has acetone as it's all. Top of work tends not to use ABS as a plastic. So. The other thing is there's a new technology that the vendor called Smoothon has put on the market. Smoothon is pretty famous in the crafting scene for making a lot of materials for mold making and things like that. They put out a material called XTC 3D and this is literally where you take two chemicals, you mix them together, you brush it on with a paint brush, it fills gaps and it fills the gaps between layers and it leaves you something that is sandable, smoothable. It looks the same way as a net result on a finished vapor smooth print except you're actually applying it with a brush. And it's not that expensive and it goes a really long way. I think it's like $40 for a box about this big and I think I could probably stretch it across 100 prints about this big. I have never tried it, do you already want to buy it? I think it's worth trying at least once if you want to smooth models. It's a good technology and it works pretty well. Especially if you want to paint afterwards, it's good to be able to do that without having to worry about the paint really bringing out the layers on the model. Because even if you sand and smooth the model, you can usually after you paint it, you still see the layering because it comes through with the XTC 3D. It really results that. I'm working on a costume piece that I'm going to try and go big with the XTC 3D on. So we'll see if it actually works as nice on the large scales it did on the small for me. All right, so let's talk about the 3D printers that are out there. We'll talk specifically with an ear towards the open hardware ones. Now, this is the sort of types of hardware that are out there right now. Stratasys and 3D systems are the two big players in the 3D printing market. They mostly focus on the high-end scale of the market with large industrial fabrication of it. They've been making printers since the 1970s. They're both generally pretty unfriendly to free and open source. They hold the majority of the patents in the space which generally limits what other people can do because we have ideas on how to do things because these people have figured them out because they've been spending the last 40 years trying to figure out ways to do things. And then anytime anybody implements one of their patents, they go out and try to play for old days and if they can't, it's even a new blueprint. Both vendors have support for virtually all types of 3D printing technologies. They are known for adding support for things like food, sugar, chocolate printing. So if anyone is doing revolutionary on a major scale, they're usually funded by or directly generated by one of those two companies. Now let's tell the story of MakerBot. MakerBot originally sold kits and pre-built open hardware printers that were well supported by free and open source software. The printers themselves would have been considered open hardware had the terminology existed of time. It didn't really. They encouraged people to hack, modify, improve their printers. They posted all the designs and schematics under open licenses. And were generally considered heroes in the 3D printing open community in the rep-rath community until 2012. And in 2012, the people behind the company and MakerBot stopped making their hardware open. They claimed they were losing too much money to cheap knockoff Chinese versions of their printer. And they were frustrated at seeing these knockoff versions out there and decided that this would not stand and that they could not have these people making their cheap copies anymore. And so they went to make actual, real, solid, proprietary printers from then on out. Now the real story is slightly more nuanced because the real story is that the people making MakerBot were not able to find a good model to make money off of their open printers. And the fact that they were charging way more than they needed to for the printers and that the other companies were able to make printers that were not cheap knockoffs, but rather just as reliable as the MakerBots at a much lower price point meant that they shouldn't have taken a harder look at how they were making money and the price points. But they went to go talk to venture capitalists. The venture capitalists said, no, no, no, stop doing that open hardware stuff. You need to start to pride-rising your printers so you can make all the money and keep all the electrical property in yourself. So MakerBot announced at the open hardware summit in 2012 that they weren't going to be open hardware anymore and that we could all suck it. They didn't actually use the phrase, we can all suck it, but it was clear for me. They stopped the writing open source software for their printers. They do still support Linux with their MakerBot printer family, but they use it with a proprietary software tool that is known to the side. And their printers also have knocked off in quality in later years. The most recent revision of their printer. There's a class action lawsuit pending against MakerBot with the knowledge that they may have knowingly released a printer to the market that was fundamentally flawed that would never work right. Cubify. Cubify is the consumer printer that three systems makes. Cubify uses proprietary filament cartridges. So if you want to buy your filament, you have to buy it from them. It literally checks the DRM signature at the cartridge to ensure that you have put it in and authorized cartridges to the printer because that's how the big printers work, so why not make their small printers work that way too? The controlling software is only available for Windows and Mac. I have seen a couple people who have bought Cubifies on sale on discount at places and then have spent months hacking all of this nonsense out of them and the machine literally does not resemble anything like it started with. So many of the parts have to be changed out and modified in half. They probably could have been better off by just building a reprap, but they're crazy. And this is the printer that sold at Staples. So if your parents bought you a 3D printer at Staples and you really desperately wanted to be able to run any kind of filament into it, I guess you could endeavor on such a project. But you do see it. So sometimes we'll see this printer out and about at shows and things. Reprap printers, I keep saying that term. Reprap is the project that was created to develop 3D printers that can print most of their own components. The idea being if we can make printers, if we can have a community of people who can print other people printer parts to make their own printers and can be a self-sustaining community. And they've been pretty successful. They've released four designs, which are the Darwin, the Mendel, the Prusa Mendel, and the Huxley. They keep promising they're going to put out a fifth design, but they haven't done one since 2010. So... Who are you referring as to them in this case? The official reprap site, the one that's... Is it written where and stuff like that? Yeah, I know that Joseph Prusa is effectively the reprap community these days. And he has put out models. And he has put out models, but he has made it clear that this is him, not the reprap community. So he really is trying, he was trying to put out the models when I spoke to him last to sort of encourage the official reprap community to either say these are now official reprap models or to be inspired by his models. So there are newer models that he's put out there, but those are all Prusa models that are reprap derived versus official reprap models. Okay. It's a semantic difference. Yeah. But there are new designs that are coming out aggressively in the open hardware space. And I'll talk about a couple of the other ones a little bit later. But the reprap project was always intended to be free software and free hardware. And they are very well supported in Linux. There's no proprietary things in the mix. The firmware that drives them is all very well known. And there's a number of different three-year firmware's you can choose from. If you decide to design your own firmware. The Lowest Lot Taz is open hardware. This is one of the two printers that I have. It in fact is the first hardware device to receive the FSF certified Respects Your Freedom certification. They run about $2,000, a little more than $2,000. They have the widest range of supportive materials of any 3D printer on the consumer market. They like to use the term prosumer. They have an active open source community that improves and modifies the printer design. I told the story about flexible material. The open source community wanted to print into Flex when the material came on the market. But every time you run it through the standard nozzle for the Taz, it clogs. Because you've got a nozzle like this. The flexible material would get into the top of the nozzle. It likes to expand when it's heated. And so it would fill that gap and then it wouldn't push anymore. It couldn't push past it, but it was too strong. So they came up with a different design stronger motors, a direct channel drive. So there is no possibility for pooling this just constantly forcing the filament down now the extruder end that results in high quality. This was sort of sketched on a napkin style design by the open source community who didn't really understand how all of the motors work together in tandem. The upstream for Lowell's Bot started working with that community to refine that design. And then people started trying to build hacked versions of it. And then Lowell's Bot generated a version that they actually sell. And they released all the schematics for an open source. So if you want to divide the steppers and think of them yourself, you can do it. Current models has five. It is rep wrap based. Although if you put it next to other rep wrap printers it doesn't quite look the same. They have gone in different directions. It's sort of a tree off of that. It acts the same way. It just looks different. It has a 12 by 11 by 10 inch print area which is pretty freaking huge for a printer in the rep wrap family. Although there are definitely bigger ones. It prints at 200 millimeters per second which is its top speed. If you print at that speed it is practically just lying around the bed and plastic is flying everywhere. So I don't usually run it that fast but you can if you want. It has an all metal hot end which you can get up to 300 degrees Celsius which allows it to print a lot of the materials that earlier printers would have had trouble with doing melting floor. Didn't you use it? Yeah, the task five has the all metal hot end. It also has a PEI treated bed. PEI is polyetherimide. And basically this is a plastic sheet which you can attach to the glass and almost all 3D printed FFF materials love to stick to it. In fact, some materials love to stick to it a little too well like Ninja Flex which if you don't treat the PEI bed you're never getting that Ninja Flex thing off. It's gonna stick there forever. Does it work with nylon? Yes. Nylon all you gotta do is put glue stick on it just to get that initial stick and it works. It's really nice. And then little spot came out with the mini which is also open hardware and also has the FSS certification which works extremely well. Yes. Now the next thing about the mini compared to the task because it is smaller. It is slightly faster. It is 275 millimeters per second. It's self levels and it's self cleans which is amazing. I cannot tell you how much amazing that is. It's self levels the bed. Basically it has the all metal hot end. It runs a slight current through the nozzle and then it has four PSO electric disks on each corner and then it moves the nozzle and lowers it slowly until it just touches to determine the role. It's not the right way. I know, right? It's perfect. And they're making a next generation printer in the Taz family. It's probably not gonna be called the Taz anymore. They're gonna re-brand it. That's gonna be the mini self leveling and self cleaning with the Taz's flexibility and form factor. It's gonna be like the perfect rep rep printer ever. So I bought this one now and I wanted to tease Tricia inside to be able to wrap it in hardware and software because in the small footprint which... Yeah. It has a built in handle and literally just pick it up with one hand and you're like, I'm taking my printer away. As opposed to the Taz where you kind of are. And the self cleaning part? Self cleaning part is on the back of the bed it has a slot where you insert like a, it's like a gritty foam almost like a pad and you put the pad in there and before it prints it brings the nozzle up to 75% heat for the job and then runs the nozzle back and forth into this pad and scrapes off all the extruding filament and it comes out clean. Okay, so it's not self removing of the printed object? It does not self remove the object. That is, no one has quite gotten that. Just like hack a day is sort of hacky pushing it off with a giant. Nobody's got to self removing from the bed yet. That's a much higher end attack than nobody's using. It is also got the PEI treated bed so it's got the good stick. This is, this printer runs 1350 US which is, they can charge three times for the quality this printer does. It was used with, so we use four or five different printers for this and the other. It's a nice printer and it's open source and it works great in the nice. All right, the Orion Delta style printers, sometimes they're called Orion, sometimes they're called Delta's. Delta seems to be the term that's running out now. This in the reprap family. This is where the printer head moves in three directions. The printer head's actually attached to big rubber bands and the motors are running on rails off the side and so it's moving the position like this and so that it looks like a marionette where a puppet is in the middle. The bed doesn't move, the base is fixed. This is great for printing really tall objects where moving the bed will cause it to shift just slightly and you get print fails as a result. So printing things like prosthetic arms that are very tall, tall bases, the deltas do a fantastic job of printing that. The biggest downside to the delta is they tend to have very small bases and so if you have something that wants to be wider and you want a different style. Although I've seen some pretty crazy big deltas if you really want to get into that. I have seen a video where Delta actually like pulled this thing. Yeah, the 35th doll delta that was taking the pellets. It wasn't taking filament. That's pretty crazy. I saw that in person, that was insane. This one specifically is from a vendor called CME CNC. It is also open source, it runs about $1,500. The Form Labs Form One is a different kind of printer. It's not an FFF printer. This is a stereolithography printer. It prints using UV sensitive resin. So basically what it does is it directs a precise laser across the tray of resin which causes the resin to harden where the laser intersects the resin. The build platform then rises up and then another layer is generated and it repeats until the model is completed. So this is different from an FFF printer because FFF printers print bottom to top and the SL printer with the resin prints top to bottom. So on some of the designs for the SL printer it actually appears like the object is emerging from an ooze when it prints. Which is very terminator too. SL technology was very patent-concumbered but the general consensus is that most of the patents in this space have either recently expired or will expire any day now. And it's very difficult to be sure of the patents because so many of them are so structured in a way that you can't tell what they're actually drawing the patent. You'll only find out when they come after you. So when the Form Labs printer was created they were Kickstarter and the Kickstarter was going fantastically well until 3D systems came along and was like, by the way, we actually hold multiple patents which you are infringing upon for this technology. They had a core case. They settled on the grounds that by the time the core case would be done the patents would be expired. Form Labs agreed to not sell any printers until the patent expired and to only honor the Kickstarter and to pay a fee to 3D systems that was equivalent to the number of printers that were successfully Kickstartered. And so it was only in the last nine months that Form Labs has actually been able to sell 3D printers again because they could only honor the Kickstarter. They had a store up and you could click buy and they'd take your money but they couldn't send it to you until the legal terms had ended. So... What's going to happen? It's on, it's on. It's all done. They've just started selling printers again. They're actually blurring the printers. They actually have the Form One Plus out now. So the biggest downside with the Form One is it's a very cool printer. It prints with a very high resolution. The big downsides are that most of the resins result in a pretty fragile print. It's not gonna be as solid or strong as an FFF plastic print. The other thing is that this printer is only supported on Windows or Mac. There is no Linux control software for the Form Labs Form One. I have offered repeatedly to take their QT-based software and ported to Linux. I've no charge to them under NDA where I would literally just say I don't even have to open source it. I would just put the binary down. You can put it on your website and offer to people and they have an answer. So the good news is there's a couple of... Now that all the patents have expired there's a couple of people that are building open source 3D resins printers and there's a couple of them that are actually coming to market. There's one that's on the market called I think it's the Muse printer and it actually looks really reliable. It's roughly in the same price point. It actually comes out about half the cost of the Form Labs with a much bigger resin tray so you can print bigger objects. And for resin, the people who make the Form Labs printer also make resins for it and they've made a resin that is very strong that is almost as strong as ABS and good prints. And they're actually using it in all sorts of crazy tests where they're like holding weights on a crane with a printed chain and so it's actually pretty impressive. It's very expensive resin. Standard resin cost is about $149 US for a leader. The tough resin is $600 for a leader, so. So it's like 20 pounds more in the ABS. But a leader goes a long way on a crane. So on the low end, it's not the tough stuff, the leader is roughly equivalent in cost to ABS for the printer. It's not more expensive for the materials and the amount you get out of it because the resin that doesn't get get by the laser can be run through the machine. There's a lot of different ways you can do 3D printing, but almost none of them work with Linux and they cost way, way more and so I don't talk about them. If you wanna talk about them later, I can have you explain them to you. But it is possible to create 3D objects from Linux. You can do it today. Your best bet is to go with something in the rep-wrap family, but don't just buy a 3D printer and staples and expect it to work with Linux. It's not gonna work. Are there any questions? I'm like way over time, but I'll take some questions if you want. Yeah, I have one about the laser, but is it from colors at the same time? It's a good question. So the best way to do different colors at the same time is to have multiple nozzles. So the TAS printer, which is their bigger printer, actually has an add-on that is two nozzles. And so you can run two different colors. It is a Royal Paint Calibrate, like I have this kit and I hate using it because calibrating is so immensely difficult. And that's because you have two nozzles where when you heat one, the heat envelope affects the other nozzle. And so keeping them at a constant heat with reliable flow is almost impossible. And then the nozzles have independent height settings. And so you've got to configure it to get the heights right. And then relative to the bed, which isn't always level, it is, the last time I got it calibrated, it took me and the Lowell's bot people who worked for the company three hours to calibrate it for a single print. And it went out of calibration after the print finished. So it's hard to do dual nozzle printing in a model that's not built for it. There's a lot of people who have built custom hacks to do multi-nozzle printing in rep-wrap and in the Lowell's bot universe. And there's even people who have built what's called a Hydra head for the Delta printers, which is eight nozzles running on a Delta. I love it. But for the most part, if you want multiple colors, get some paint. Yeah, but I think this has more use for PDAs. Yeah, what I really wanted to do with the dual nozzle is to be able to print support material that would dissolve. High impact polystyrene, which we talked about briefly, is nice because it will dissolve in a chemical called limonene, which is possible to get. It's not CVS drugstore easy, but you can get it. And then the downside to that is when it dissolves, then your model smells like lemon because limonene is the chemical they used to add to cleaners to make it smell lemon fresh. Any other questions? Great, well, thank you so much for coming. I appreciate it.