3D-printed clay-extruder die

In UCSC’s soon to be fab lab (and an existing one), I talked about two of the fab labs on campus and mentioned that I was planning to try using the DSI studio in the basement of the “Science and Engineering Library” (which no longer has any science or engineering books or journals, so should be called the “Study Space on Science Hill” instead). On Monday, I got certified to use the Ultimaker 3 printers and attempted printing the clay-extruder die I had designed:

The Scott Creek extruder is designed to work with ⅛” aluminum dies, with no allowance for different thickness of dies. So plastic ones have to be made with an ⅛” flange, with the body of the die as a plug that goes into the extruder tube.

I printed the design in PLA with a 0.4mm nozzle and 0.2mm layers, with 4 walls, 20% gyroid infill, and 20% infill overlap.  I chose gyroid infill, because the stresses on the die would be both compressive and bending. I was worried that the 3mm flange might not be strong enough, especially as I expected some stress raising around the inner edge of the ring that holds the die in place.

The die printed fine, with no visible blobbing and only tiny amounts of stringing, easily brushed away.

On Tuesday, I took the die to my ceramics class and tried extruding some clay with it (I had a scrape on one of my fingers and did not want to get my hands wet practicing throwing pots). The die worked just fine and showed no damage from being used—the photo above was taken after the die was used and cleaned. So it seems that the 3mm flange is thick enough not to cause problems (at least for a few uses—I’ve no idea how long the die would last in production use, and I don’t really care). Next week I’ll trim the soap dishes, stamp them on the bottom, and leave them to dry on the greenware shelf.  Trimming should not take too long, so I should have time to try throwing again.

If I design other clay-extruder dies, I think I’ll make the bevel on entrance to the die bigger, so that there is more “wedging” of the clay as it is pushed through the die.  I had intended a wider bevel, but I had a bug in my OpenSCAD code, where I had used a constant “3” instead of the parameter “extend” in one function, and I had not noticed that the bevel was narrower than intended in the final STL file.  Of course, there wasn’t really room on the die for more bevel (unless I made the soap dish even smaller), so perhaps the bug was not such a bad thing.  I’ll have to think about how I could modify the OpenSCAD program to make the bevel large in the middle of the die, but reduce it near the edges, so that the die entrance does not get too close to the sides.

UCSC’s soon to be fab lab (and an existing one)

Yesterday, I took a tour of the space in the basement of Baskin Engineering that is being converted to an “Experiential Learning Facility” (a name only a bureaucrat could love—but at least they are looking for a new name!). The space used to be two machine shops: a professional one that only the lone machinist was ever allowed into, and a student machine shop that was never used, because you had to get someone with a recharge number to pay several hundred dollars for the once-a-year training then $12 an hour to step into the room.

Most of the old machine tools were sold as surplus, because they were too difficult to use for a student space. Some of the more basic metal working tools were kept (a bending brake, a large shear, an arbor press, a Rotex punch).  There is also a small CNC mill and a lathe.  There are some basic woodworking tools and a laser cutter, and some 3D printers will be moved in once they get the next room over (before March, I think).  I believe they will also get some sewing machines, though what sort is far from clear.

The space looks like it could become a fabulous maker space for students, but I suspect that it will not, because of stupid policies by the administration.  They outlined 4 uses for the space: a first-year design course, 1-year group projects (like senior capstones), thematic areas (like the S-lab on sustainability that has moved into much of the space and provided equipment for that portion of the space), and engineering clubs.  The engineering clubs will get permanently (or at least annually) reserved spaces for the club to meet, work on their projects, and store their projects and tools.  The lion’s share of the space is reserved for 4 clubs, which will have to apply for the space, but it is expected that FormulaSlug (car building), SlugBotics (robot building), and Rocket Club (rocket building) will get 3 of the 4 spaces.

Note: what is missing?  There is no provision for someone to come in and use any of the tools unless they are in one of the classes or clubs! You take the freshman design course (which has not been created yet) and learn to use some tools—great!  Now you never get to use those tools again until your senior design course—too bad!  Unless you are interested in one of the 4 things that the 4 clubs are doing, in which case you are set up in style!  Why the administration omitted such a critical part of the facility is a mystery to me.

I urged the few students who were there (interns working on helping create and advertise the space) to get students to start a new Maker Club and apply for one of the club bays.  That way anyone could join the Maker Club and (after getting suitable safety training) use the space and tools! The Maker Club could also provide “super users”, who could help maintain the facility, providing training and safety supervision. It would also be good to open the tool use up to faculty, staff, and emeriti (again, with appropriate training), so that a community can be built that has some continuity from year to year. With the current administration, though, I’m not holding my breath waiting for a sensible outcome.

If the administration gets their heads out of their behinds and figures out how they are going to make the space actually useful to individual students (rather than just instructional space and a clubhouse), then I’d be glad to help out with things like short workshops, 1-unit courses, and even some scheduled supervision of the space.  But I fear that the administration is going to silo it up so that very few students and no faculty or staff have access—repeating the mistake that PBSci made with the old machine shop (though not quite as badly, since a few hundred students will have access, instead of 1 or 2).

This bad management is not an essential part of creating a makerspace at UCSC.  The Digital Scholarship Innovation Studio (DSI) on the lower level of the Science and Engineering Library has a much more open policy—anyone with a ucsc.edu address can take their training (online), do a simple in-person assessment, and use the equipment. Actually, I’m not 100% certain about alumni—they can take the training and sign up for the assessment, but they may need a current ID to actually use the equipment. The answer in the FAQ for “Who is allowed to use DSI equipment?” is “Any UCSC affiliate (students, faculty, and staff) who has completed Library training.”  That is a little ambiguous about affiliates who have retired or graduated.

DSI has 4 Ultimaker 3 (2-nozzle) printers and plan to get 2 Mini Prusa printers with a Mosaic Palette accessory (for 8 colors) set up during Winter quarter. They also have 4 small GlowForge laser cutters (11″×19.5″), but lack the ventilation to cut acrylic—so just cardboard, MDF, and thin plywood (or engraving on some harder materials). Currently you need to use materials provided by DSI, but they don’t have a recharge mechanism: they provide a small amount free to students. They provide PLA filament (the 3mm Ultimaker filament, not the more common 1.75mm filament) and sheets of birch plywood. I imagine that they’ll have to set up a way to charge for materials if they get really popular.

I took the online training for the 3D printer this evening—it is fairly simple, though there are a few bugs in the Canvas course that I pointed out in the feedback form at the end.  I suspect that they’ll get those bugs cleared up soon. I’ve signed up the in-person assessment Monday morning, and I’ll see if they let me use the 3D printer.  If they do, I’ll take the online training for the GlowForge laser cutter and get checked out on that also.  I’ll also find out if there is any role for me to play in their space—I’d be glad to help students learn to use Cura or OpenSCAD—perhaps even FreeCAD, once I’ve learned how to use that myself.

I have something to print that is too big for my little Monoprice Delta Mini: a die for a clay extruder, to fit a 4″ round Scott Creek extruder. Unfortunately, the extruder die design I had to match my broken soap dish will not fit—the biggest dimension is more than the 4″ diameter.  I decided to try scaling the design down, making a soap dish only 80% of the size of the existing one.

The Scott Creek extruder is designed to work with ⅛” aluminum dies, with no allowance for different thickness of dies. So plastic ones have to be made with an ⅛” flange, with the body of the die as a plug that goes into the extruder tube. I think I like the NorthStar extruder design better, but I’m not about to spend $400–500 to get one.

If they let me, I’ll try printing this on Monday (it uses 50g of PLA, for a cost of about $3.33, or about 13 minutes of minimum-wage staff time).

Ceramics class?

I’m hoping to be able to get into the beginning ceramics course taught through adult education at Santa Cruz High.  It is one of those courses where the registration opens at a specific time, and if you aren’t on the computer obsessively refreshing, then you might not get in. (I’m not telling when the window opens—I don’t want more people competing for the slot!)

Although I have long had a fondness for pottery as objects, I’ve never learned to shape clay myself. I don’t expect to get good at it, but it should be fun to learn first-hand how much skill is involved.

I do have one specific project I’m interested in doing—replacing a broken (and badly mended) soap dish.

I had two of these soap dishes, but at various times they have been knocked off the counter in the bathroom and broken. This one was mended with superglue, but I’d really like to replace it.

I can’t remember who we bought the soap dishes from originally (I showed one to a few potters that I thought might have made it, but they did not recognize it as anything they or anyone they knew made), so now I’m thinking of making one myself. I think it should be a doable project for the beginning ceramics course.

One of the potters I showed it to suggested that the soap dish had been made with a clay extruder, which seems extremely likely to me. I sent email to the instructor of the ceramics course, asking about whether there was a clay extruder available in the studio. There is, but she did not know what model, and she suggested that I wait to see whether I get into the course before looking into it further.

I decided to start designing a clay-extruder die to roughly match the cross-section of the soap dish. I used SVG (manually edited as text, but viewed with Inkscape) to create the outline. I then printed a thin piece of plastic with half the outline cut out, to see how close I had come. It took me 3 iterations to get an SVG file that satisfied me as being close enough.

Here is a PNG rendition (from Inkscape) of the 3D shape. The image is upside down (because Inkscape has positive Y down the page).

I looked around for ways to create custom extruder dies, and found that there are various materials used for the dies: ⅛” aluminum,  ½” HDPE, marine-grade plywood, and 3D-printed PLA. I have no way to do precision cutting of aluminum, HDPE, or plywood (though I could probably get a membership at Idea Fab Labs and learn to use their CNC mill), so I decided to try designing for 3D printing, even though my 3D printer is not big enough for the approximately 112mm square dies used by the North Star extruders. I can find other 3D printers around town or on campus and use them, if needed.

I used OpenSCAD to create the model. It was very easy to do a simple linear extrusion of the SVG file, but the clay extruder dies I’ve found online all have heavily beveled edges on the top side of the die, to compress the clay as it enters the die. That was a bit harder to manage in OpenSCAD.  I ended up using the openSCADbezier.py plugin for Inkscape to convert the SVG file into an OpenSCAD file that could generate points along the curve, then I used the BOSL2 library for OpenSCAD to generate an offset outline for the entrance shape on the top surface of the die and to “skin” between the exit shape and the entrance shape. I had to fix one bug in openSCADbezier.py, as it had a check for strings that only worked in Python 2, not Python 3.

Here is an image of the STL file showing the beveled edge on the top of the die.

I don’t think that I’ll end up printing this STL file, as it would take 40 hours to print on an Ultimaker 3 (at 0.1mm resolution and 100% fill). I could make a cruder version in about 5 hours (at 0.2 mm resolution and 30% fill), but I don’t know whether it would be strong enough to withstand the pressure of the clay. Even if I never 3D print this die, it was fun figuring out how to create it in OpenSCAD.

I looked for other ways to convert my design to an actual die and found a few:

1. I could use a commercial 3D printing service, but ones like Shapeways would charge around $140.
2. I could get North Star (who make extruders) to use their CNC mill to make one for about $40 out of HDPE. I’d have to get them the SVG file somehow, as their ordering procedure for custom dies involves making a hand sketch on a paper form! Update 2023 Jan 3: North Star no long offers this service. “Unfortunately, we have discontinued our custom die service until further notice. North Star is focusing primarily on production of our main product line. Equipment, manpower, and materials are dedicated solely to this effort. We will review our ability to offer custom die services sometime in 2023.”
3. I could order a 3D-printed custom die from a shop on Etsy (https://www.etsy.com/listing/1372088315/custom-die-for-clay-extruders) for about $12—the procedure for submitting the SVG file there is pretty clear.  I need to find out how solid the dies are and how well they hold up—I’ll need to decide between their cheap dies and the more expensive ones from North Star (which are almost certainly stronger and smoother). Update: I contacted the seller, and the dies are printed with 20% fill and 0.2 mm height, which Cura estimates would take about 4:25 to print. The seller claims that they work as well as the North Star dies, but hasn’t used them long enough to know how durable they are. Prices are likely to go up as they get more experience with selling them.
4. There is another Etsy shop selling custom dies for 4″ clay extruders, but they are a bit more expensive, and their images don’t show the beveling—I would certainly be better off with a North-Star-made die.
5. I could join Idea Fab Labs and learn to use their CNC mill, which would cost me at least $150 (for a one-month “pro” membership).
6. I could join Idea Fab Labs and 3D print there (I suspect that they’d allow a long print, especially if it was on one of their Ender 5 printers, rather than the more popular Prusas). But that would still cost me $110 for a one-month membership, and the 3D-printed version would probably not be as good as one milled by North Star.

Right now, I’m leaning towards getting a 3D printed die from the Etsy shop, since the price is low and I can live with wasting the money if the print fails. The North Star die is a safe backup.

But first, I have to get into the class and then find out what model extruder the classroom has.

Update 2023 Jan 6: I registered for the class successfully.

Dragon key holder

In 3D printed names, I showed a key holder for my home and office keys, which finally failed after 3 years.  My family has been bugging me to replace it anyway—pointing out that it is stupid to have one’s name on a key holder, as it makes it easy for anyone to figure out what doors the keys are for if they get lost.

The key holder cracked at one of the bolt holes. It was still sort of usable (the keys were still constrained by the bolt and the cover plate was not yet coming loose), but it looked bad and would sometimes snag when I was trying to get it out.

I decided to make a new key holder with a dragon motif, based on a sweater from Past Times that has gotten too moth-eaten to wear, but that has a pattern I like.

I took only the dragon part of the pattern, which is 18 pixels by 109 pixels (20 by 111, if you include a 1-pixel white border).

My first job was to transcribe the knitted pattern onto squared paper, then type it into the computer.  I had hoped that OpenSCAD’s “surface” function (which I used successfully for the quantum dot pendant) would let me convert this image easily to a 3D relief, but it did a terrible job, as each black dot became a sharp peak.

I ended up using the Pillow fork of the PIL Python package to manipulate the image and export it in PNG format.

    im_5=im.resize([5*(num_rows+2), 5*(num_cols+2)], Image.NEAREST)
    im_max=im_5.filter(ImageFilter.MaxFilter(3))
    im_big=im_max.filter(ImageFilter.SMOOTH)

I resized it by a factor of 5, ran it through a “max” filter to spread out the black, then smoothed it.
The PNG file is the negative of original image. Unfortunately, the checkerboard grid does not print well at the size I needed to make the image.
I edited the pixels manually to get a somewhat more printable shape, then did the same sort of spreading and smoothing.
I initially printed just the dragon on a little oval, to see how it would come out, before adding the rest of the key-holder design from the original key holder.  Because OpenSCAD produces huge STL files when the “surface” function is used, I simplified the STL files with https://myminifactory.github.io/Fast-Quadric-Mesh-Simplification/
Here are the 3 samples. The top one tried just expanding the pixels to 5×5, the second one did the smoothing, and the third one got rid of the checkerboard patterning.
The first two prints were failures. The top print was done face-down on the glass bad (as I had done the previous key holder), but there was too much spreading on the first layer. The second print failed because the glass plate rotated during the print—I had to add hot glue to it again, to keep it in place.
The final key holder is ok, but it would really look better if I had a smaller nozzle (the 0.4mm nozzle limits the horizontal resolution).
Here is the finished key holder with keys—I expect that it will last another 3 years (though I did use 50% fill this time, to make it a little stronger).
I’ve uploaded the design to https://www.thingiverse.com/thing:4939976.
	

Final 3D-printed “quantum dot”

In 3D-printed “quantum dot” and 3D-printed “quantum dot” revisited, I wrote about my attempts to 3D-print the image from https://scitechdaily.com/direct-visualization-of-quantum–dots-reveals-shape-of-quantum-wave-function-of-the-trapped-electrons/.

I finally got good prints from the resin printer at work (they had to clean the optics on the printer) and a decent print of the “stage jewelry” version on my Monoprice Delta Mini printer. I gave away all the prints (including the failed ones) to the physicists who provided the data, except for the one best print in each size.

The STL files from OpenSCAD are ridiculously large (17.8MB and 19.1MB), but they can be reduced using https://myminifactory.github.io/Fast-Quadric-Mesh-Simplification/ without much loss of detail to under 1MB.

The OpenSCAD program, scaled data file, and two STL files are available at https://www.thingiverse.com/thing:4939939

Here is the resin print, which is 50mm in diameter. The peaks come out clean and sharp, but my only color choices were black and clear (the only two resins BELS had).

The back of the print has the scaling information, but even with sanding the spots from the supports are annoyingly visible.

The stage jewelry version is twice as big, with a diameter of 100mm (I measured it at 104mm—I think my printer calibration may be a bit off).

Again, the back has the scaling information. Using “concentric” for the bottom layers made for some interesting patterning.

Here are the two quantum-dot pendants side-by-side, to show the relative sizes.