An early MakerBot operator and electrical engineer talks 3D printing tips, peristaltic pumps, and more!
Derek: Did you end up replacing any of the parts on your MakerBot Cupcake?
Andrew: Well, because Rick was the first one to come out with the heated bed, that was the first thing that I replaced. So I got one of the ceramic tile types of heated bed. You sort of cement on the Nichrome wire and then mount it on magnets. So I had that running for a while and then I finally put on an automated build platform, the standard issue one from MakerBot.
D: Do you still have the belt running on your Cupcake’s automated build platform?
A: I do but I don’t use it that much, and I don’t really use it in an automated fashion. Plus the Cupcake is in the basement and I have the MakerBot Thingomatic up here and I configured it as a static heated build platform.
When I got my kit it was one of the beta test kits and they didn’t have all of the parts for the ABP yet so I went straight to the heated build platform. I like it because I can get my surface layer nice and flat, and no complications. Plus it gives me a little extra height. My max Z-axis is about 120mm.
D: When I put mine together, the belt was.. not quite perfect. First of all it was running backwards, but once I sorted that issue out the belt sort of just came apart and flew out of the machine.
A: Yeah that seems to be a common failure mode.
D: So I just use it as a regular heated build platform. I kind of like it, because when you generate the Gcode for the automated mode it doesn’t move to a position and wait for you to hit OK, it’ll heat up and wipe on the little wiper thing. So even if you’re not doing automated stuff I kind of like how it works.
A: Right, right. And of course on the Thingomatic you have the the centering that happens regardless of which one you’re using.
D: Oh, how does that work on the Thingomatic?
A: They have endstops at the minimum X and the minimum Y, and at the top on the maximum Z point. So it zeroes itself by going to the min X/Y, and then it zeroes by going up to the max Z. So as part of your calibration routine you have to set the maximum Z height, so there’s sort of an initial calibration step where you have to position your stage right in the middle and lower down by manually twisting the little screw so that it’s just kissing the surface of the build platform. Then you send it up to the max Z until it triggers, and once it triggers you open up the control panel and read off the coordinate and that’s what you punch into your start Gcode file.
D: Did the endstop design change in the Thingomatic?
A: You mean from the Cupcake? Yes, in that the initial Cupcakes when they were shipping with endstops, they were all optical. And so in the Thingomatic they’re all mechanical switches. So just the little cantilevered push button type switch. It zeroes itself quite nicely – don’t have to worry about lights doing false triggers, because that was a big one on the early Cupcakes.
D: Did you leave them in on your Cupcake?
A: Just for a little bit. They ended up being more hassle than help so I ended up just taking them out. It was one of those things where you get used to manually adjusting your Z height anyway when you’re laying down the raft, so it wasn’t really an issue. But it is nice once you have the automation taking care of it for you.
D: I only tried very briefly with the endstops to get those going, and the popsicle stick that was connected to the platform would just stab at the wrong spot.
A: Yeah.. It was one of those things that you chalk up as nice in theory but they needed some work.
So yeah, with the Thingomatic the switches are positioned in such a way that it’s actually sort of the main structure. They’re positioned at minimum and maximum trigger points, it’s a nice firm engage or disengage.
D: I was hoping to see one in person on the weekend (at the Toronto Mini Maker Faire) but nobody had one up and running yet. Oh, did you get to work any more on your Star Wars landspeeder point cloud?
A: No, I haven’t had a chance to get back to that yet, it’s been hectic! It’s one of those things where you have to find a good couple of hours in the evening and just take picture after picture.. I also need to set up a circular camera dolly so I can just sit it on the dolly and chug it around. And if possible a servo motor on there to tick it off like clockwork and automatically trigger the camera.
D: That was the My 3D Scanner thing?
A: Yes. It’s really easy to use, you just zip your pictures up into a single file, upload them, and it crunches away and sends you an email when it’s ready. You get a point cloud and then you work your magic in MeshLab. That’s where you really have to check out some of the stuff that Tony Buser has been doing because he’s mastered the art of mesh regeneration and subsequent printing of the results.
D: I was telling you before how I’d scanned my hand with Kinect software, I wonder if I could get some tips from his stuff.
A: Oh yeah, and of course have a look at what he’s posted on Thingiverse. In one of the early tests that he did, he outlines the process. If you go to thing 7930 it’s his version of the Venetian lion and he has some instructions in there. You can see in one of the pictures the support structure that he included there.
Let me see if I can find the one where he details the point cloud. Most of them are Skeinforge settings. You’ll have to ping him and see if you can get some tips.
D: Have you done much work with OpenSCAD?
A: Oh lots, lots. OpenSCAD is my favourite tool. Coming from more of a coding background I find it easier to code it than to fight with a mouse or a trackpad to get my coordinates aligned. I code up my things hierarchically so that I can reuse as many as my objects and parts as I need.
D: And resize portions of them independently.
A: Yeah, resize portions and position things in sort of a relative fashion so that you’re doing a lot of translations and rotations from previous coordinates so it’s that sort of that nesting of things.
D: A friend of mine has a gamepad for his PC, and it’s missing one of the analog sticks. I was thinking that the shape is simple enough that I should be able use OpenSCAD and take a squished sphere and a tube and kind of a divot on the end for the thumb depression…
A: So is it one of those rocking kind of joysticks?
D: Yeah, it pivots. And the stick is just a simple piece that pulls off, so I was thinking to use OpenSCAD to recreate it or use the online 3D scanning option.
A: Do both! It’s one of those things where it’s probably just easier to measure out the critical dimensions, capture that in your OpenSCAD script, and then you can do up some custom ones – maybe make them thumb-shaped. Find a model for a thumb, and attach that onto the end. All sorts of fun with that, like an eyeball shape!
D: Yeah, my friend probably wouldn’t appreciate thumbing an eye to play a video game.
D: We were talking before about peristaltic pumps – do you remember that?
A: Yeah, that’s one thing that I’ve been playing around with. Of course with the MakerBot printers, one of the extruder head options is their Frostruder which is for frosting and clay and other sort of compressible fluid types. Of course the gotcha there is the air compressor portion of it… So I’ve been playing around with some ideas for a peristaltic pump-based compressor. Just been gradually coding that up using some of the gear libraries that are available in OpenSCAD over on Thingiverse and figuring out how to go about coding that up in a printable manner.
One of the nice things with the peristaltic design is that because compresses a rubber tube to do the pumping, it’s completely enclosed. Like the whole sort of material channel is enclosed. You’ll often find those in hospitals being used for pumping medication or blood because the fluid isn’t contaminated in any way by the pumping mechanism. So you can have fun with all the foodstuffs and of course the whole DIY bio sort of community can have all sorts of fun with that!
D: Oh, one thing that I encountered at the Mini Maker Faire – I always have a little container near my MakerBot for the waste plastic, and as it filled up people started to ask about recycling the ABS. The only thing I could think of was that one project some students had put together to do the milk jug plastic?
A: Right, right. Some folks have been playing around with that. It’s one of those things where trying to get it to work.. Well, traditionally those sorts of mechanisms are used in really large injection moulding equipment, so very very high pressures at relatively low temperature and it’s a continuous stream. So in theory you could re-extrude more filament but getting the filament to a consistent diameter is a challenge.
I know what some people are doing is saving up the waste and, as long as you keep all of your ABS plastic separate from other types of plastic, what you can do is dissolve it down in acetone. Some folks will do it in sort of a diluted acetone solution. You add just enough acetone to get it all dissolved into sort of a mush and you have the option of using it as a glue – so you can glue other ABS parts together by brushing on this mixture at the interface and the acetone just evaporates away, so there’s that.
Other people what they’ll do is use it to treat the surface. When you print something off in the fused deposition modeling process you get that visible layered texture to it. So what some people will do is brush this sort of ABS-acetone solution onto the surface so that half fills in the gaps and half dissolves some of the surface away and gives it more of an organic finish to it.
D: Ah. And you normally print with a raft, right?
A: I print with a raft when I’m using the automated build platform but on the Thingomatic stuff no, I just print directly onto the static heated build surface. Because Skeinforge, it’s smart enough (if it’s configured well enough!) to lay down that first layer more slowly than the subsequent layers so it has time to get a nice solid contact with the surface and it fuses itself well with the Kapton so you don’t really have any curling up from the edges.
D: I don’t think I’ve ever had a nice print come out from a raftless attempt.
A: Yeah, you’re doing it on the ABP? The belt has that flex to it so it’s definitely better to have the raft to strengthen the base. I know some folks were playing around with printing bars around the bounding box so that it would keep the outer edges uniform and flat but you wouldn’t get as much waste in the internal area within those bars.
I think there’s some Skeinforge settings that will let you print a perimeter around it, sort of a raised wall for the first few layers and that helps make the belt rigid enough for the main print.
D: Did you have to mess with the leveling of the ABP on your Cupcake?
A: No, because I was always using a raft on that so it was pretty much self-leveling. But definitely with the Thingomatic and the static platform… It’s one of those things where you just set and forget. One trick is that you put a loose piece of paper between the nozzle and platform and set the height so that there’s just a little bit of friction so that you can tug it out with your fingertips but it’s not free to move. And then you manually just gently slide around the XY surface to see where it’s either higher or lower than that midpoint and you adjust the corners accordingly. So it’s a manual process but once you get everything sort of level enough.. Then you forget about it.
D: I’ve had a bit of trouble with mine. There doesn’t seem to be an easy way to adjust one side versus the other on the Cupcake’s ABP. I tried to measure it, and it looked like from one extreme on the left side to the right side there was a 0.9mm variance in height.
A: Which is huge!
D: So some of the techniques in the newer Skeinforge settings where you have it centered on zero and it moves off and comes back across from the side and does the raft in the shape of the model you’re printing – that just does not work at all out for me at all with that difference in height.
A: If you have some really thin silicone washers you could put those underneath the corners on the ABP, so that as you level it can tighten down one corner more than the other because you’re just compressing the silicone a bit more in one area. That may be enough to the levelling.
D: I was looking the other day at the calibration stuff you uploaded to Thingiverse..
A: Oh, yeah, my little comb shaped structures.
D: How much time have you spent calibrating your MakerBots?
A: On the Cupcake it was constant. You’re always tweaking things. And as ABS is exposed to air it gradually absorbs water, so the water content changes the characteristics and you’re always fighting that a little bit. That was the old thermistor-based extruder head that you have to calibrate so when I finally got around to that my extrusion temperature was significantly lower than I was expecting. I think my 220 C was more like 180 or 190, but because my plastic had absorbed so much humidity over time when I tried setting it to 220 the thing was puffing up steam all over the place! So I ended up having to drop the temperature back to 190 and it was fine.
And the fact that I was always printing at a lower temperature could explain why my MK3-turned-MK4 extruder never had a problem with the PTFE insulator tube. I got lucky!
D: One of the MakerBots I saw at the Mini Maker Fair, they were using a stock MK4 nozzle and they had definitely fought with that.
A: Combined with all the drafts coming around there…
D: Yeah, it was pretty windy in the space that we were in. I was thinking maybe that the insulation on the MakerGear nozzle was superior to the one on the stock MK4.
A: Oh it’ll definitely help. Also having your PID controls set properly makes a big difference too, so that it can keep up with all of the incidental drafts going by. One trick that I used to use with the Cupcake is with painters tape I’d hang pieces of paper towel over the exposed sides to cut down on the crosswinds. That was usually enough to keep the warping down to warp factor 1 or 2.
D: There was a model I started printing, the bathtub U-Boat that Michael Curry put together, he suggested standing up some pieces of cardboard around all open sides of your MakerBot while printing the really tall pieces. He said that in his own tests that he had gotten significantly better quality by doing that.
A: Oh yeah, it definitely helps – the taller the build. Even once you have a heated build platform, the upper levels can stress as the plastic cools. If it cools unevenly you’ll still get some warping happening up at the top. Or if there’s a breeze blowing while it’s laying down some layers and not others you’ll get this strange sort of hourglass type of effect on the surface because it’ll be more contracted in some places and less so in others.
D: Did you make any other changes to your Cupcake like a wobble arrester or anything like that?
A: I never did the wobble arrester but one thing I’d uploaded to Thingiverse of course was the bearing brackets. I’d made some printed versions of those just to make it a bit easier to maintain. Because you have those four threaded rods, I made it so that you could drop them down in there and then with little set screws lock the bearings in place from the sides instead of the laser cut kludge where you have it compressing the top. If you’d done and that tightened it down just a little too much you’d get bowing in the rod, or it would be just enough to bind the Z axis. So it was a way of dealing with that in a controlled fashion.
A: All these adventures us early adopters embrace.
D: I think on balance I’ve had fun with it though.
A: Oh absolutely. And it’s one of those things makes you appreciate the nature of the materials that you’re working with and the build process. When you have something like a MakerBot to mess about and make mistakes on when you’re doing a design it gives you that much more confidence when you go to use something like the Personal Factory with your model. You can use the more advanced processes, like higher resolution plastics and even the stainless steel type options. It definitely helps with the confidence that you’re not blowing away $25, $50, $100 on a smallish piece but getting some nice quality results back.
D: I was wondering, do you have any idea how thin of an object you’ve been able to print successfully? Like a post or a cylinder that’s printed vertically.
A: When it comes to sort of long and thin, more often than not I try to think of how I can print it on its side instead. Because then your minimum thickness is the thread thickness, so the deposited layer thickness and width. And then your strength is along the length of your printed object rather then when you do high objects and have that shearing aspect which… If it’s ornamental then it doesn’t matter, but if you’re expecting strength in it then it’ll just kill it. So you either have to print it off slowly so that it’s strong enough between layers, or you can make it hollow and put something like piano wire down the centre of it to strengthen it.
D: Have you ever done that?
A: Not for something as small as piano wire, but what I’ll sometimes do is if it’s a bolt diameter then use a long, thin bolt.
D: When I was making some 4X LEGO minifigs, the model that I was using was originally 2X and designed for using 3mm filament as joints. So when I double-sized that I had to make some 4X pins. So for the waist pin I tried printing that vertically and it was kind of mushing around on the surface and it wasn’t cooling fast enough.
A: One of the tricks I’ve used, with my Automoblox connectors, is when you’re doing tall but sort of narrow parts you really have to do a multiplicity of them. You can’t do one or two, you have to do at least four of them so that each layer has a chance to cool as it’s printing up.
Actually, if you look at the OpenSCAD source for that I use a little trick at the bottom so that there’s a bit of a bevel to it and that helps when you’re printing the first few layers. You’ve probably noticed when you’re printing raftless that the first layer tends to be a bit more squished than the others so you get a flaring out of the bottom. So that’s the trick – to pre-bevel in your design so that when you get the flaring it’s minimal and doesn’t exceed the real diameter that you’re interested in.
D: That’s a good idea.
A: Yeah, all the little things you pick up on the way. You design for the quirks in the manufacturing process. And you really notice that too when you start playing around with the higher-end processes for things like my LEGO-compatible disc buttons – when you print those in plastic it has a bit of elasticity to it versus when you print in one of the stainless steels there’s absolutely no give whatsoever so you have to figure out a way of designing that into the model so that you either have a locking mechanism or use specially-designed ridges.
Derek Quenneville is a 3D printing evangelist who posts weekly on the Ponoko blog. Follow him on Twitter @techknight.