Sunday, May 24, 2015

Tabula Rasa

For some time, Stephanie's been wanting something like a coffee table for the living room to use instead of a series of tiny folding tables which get moved into and out of the room as necessary. After, at long last, looking at the design of some mid-century modern tables, I realized that building such a thing would be remarkably easy. And it was.

Table, with dog for scale.

I started with a couple of sheets of plywood (a thin birch piece for the surface, a thicker pine one for the underside). I scribed a triangle on the surface, used a couple of old paint cans to draw curves into the corners, clamped them together, and cut out the shape with the usual jig saw.

Using a little simple geometry, I found the center of the triangle constituting the lower piece, drew some lines out from there towards the corners, and put the brackets for the legs about two thirds of the way out from the center towards each corner. I used angled brackets for the legs, and cut the ends of the legs at a matching ten-degree angle. Doing the top and bottom as separate pieces conveniently allowed me to drill through the bottom piece without worrying about marring the top.

Once that was done, I glued the top and bottom parts together (clamped together for about a day), put iron-on birch trim around the edges and trimmed it back, stained it a reddish-brown (sorta mirroring the living room's largely red color scheme), and put on lots and lots of coats of polyurethane finish, occasionally sanding between coats.

Helpful scale dog provides scale.

The result is, it appears, a nicely stylish table, sized and colored to suit the room it's in, which cost maybe $75, with the manufactured legs and brackets being the most expensive components. Once I get the CNC mill going, I may experiment with smaller side tables in the same style, but with some Atomic Age inlay.

Thursday, April 2, 2015

Brick. Or stone. Or chalk. Or something.

OK, one more exotic material. Laybrick isn't, despite the name, brick-based. Rather, it's a composite of plastic and chalk dust, and is marketed as providing a sort of stone-like appearance. And it kind of does. The surface doesn't have the gloss of plastic or the metallic sheen of the copper and bronze filaments I've played with. Rather, it's got a faintly porous, matte surface like unpolished white limestone or perhaps a ceramic. However, what I thought it most resembled was bone. So:

It prints very smoothly, certainly. It's very difficult to see any kind of grain in this, even close up. There were some issues with the top layer and the protruding spikes, but I suspect that's up to my printer settings, since that sometimes happens with other filaments I use as well.

One of the advertised properties of this material is that it provides different textures depending on the temperature at which it prints. Cooler is smoother, hotter is rougher. I attempted printing an object with the same temperature-varying plug-in I used to print the wood pieces. The results weren't as impressive. There were striations, but they weren't particularly visible. I may try again with a greater temperature range.

Sunday, March 15, 2015


One last experimental filament: wood. This filament is a mixture of plastic and fine-grained wood fibers, alleged to produce a wood-like look and feel. Sounded fun, so I ordered a small roll. Long story short: not bad.

The first bit of difficulty with this is that it comes loosely coiled in a bag, secured with a zip-tie, rather than on a spool like other filaments. I think it's because the filament is relatively brittle, so it can't be wound as tightly. OK, but how do you put it on the printer? You could cut off the zip-tie and put it on a spool holder, but you run the risk of it uncoiling all over the place. What I did in the short term was to loosely coil some of it around an empty spool and print from there, putting the main coil back in the bag until it's needed. So far, seems to work all right.

The second potential problem is one I'd been forewarned about, so I was able to get out ahead of it. Word on the street is that the fibers in the filament cause clogging. I replaced the usual 0.4mm nozzle on my printer with an 0.75mm one, and I had no clogging issues at all.

Printing was a little iffy because adhesion seems inconsistent. On my first print, the tall, skinny item I was printing broke off about a quarter of the way up when I pulled it off the plate; the bottom of the piece was stuck to the printing surface more strongly than some of the layers farther up were to each other. On a later print, some of the support material didn't stick at all well to the plate, so the piece came loose and didn't finish successfully. Printing seems to be best on blockier pieces which have a lot of surface area touching the plate.

The filament itself doesn't look all that woody once printing is complete. It's a matte brown which certainly is a wood-like color, and the texture is, indeed somewhere between plastic and wood, but it's not, in itself, all that convincingly woody. It's most like a sort of really thick, dense cardstock. However, here's where another of the wood filament's properties comes in to play. The precise color is sensitive to printing temperature. Higher temps darken the wood fibers. And as it happens, there's a plugin for Cura which takes advantage of this very property. The temperature of the hot end changes during printing, creating light and dark bands, emulating wood grain, and I think that works pretty well. For another piece, I'm considering using a bit of stain and shellac.

So now that I've got that printed out, I thought I'd take stock of where our Castle Panic  set is now. At this point, the walls and most of the towers are stone-painted pieces. The wizard's tower is copper with heavy natural patina. I've also got custom pieces for tar, some painted flame tokens, a monster-themed cup we keep the dead monsters in, and a one-piece dice tower which has nothing to do with Castle Panic, but we use it anyway. At the rate we make new pieces, and at the rate the original set is wearing out from heavy use, I may end up rebuilding the whole game.

Thursday, March 5, 2015


I recently became aware of some exotic new materials for 3d printing. Notably, I've come across some materials by colorFabb. They're playing with a number of unusual things, but the ones I'm dealing with right now are their copper and bronze filaments. These aren't just colored. They're PLA filament with a substantial addition of copper and bronze powders. The ideas is that with appropriate treatment, they look for-real metallic. So how do they work?

They print quite well in my quick tests, sticking to the usual blue-tape-and-hairspray surface I usually use as well as any PLA with no noticeable warping. The marketing material indicates that the metal content means that they cool faster than regular filament, so they're particularly good for pieces with overhangs. I haven't done anything with exaggerated overhanging material yet, but it at least sounds plausible.

The printed material seems a bit fragile. It doesn't fall apart if you breathe on it wrong, but it does break more readily than 100% PLA. But the big difference I noticed was the weight. Not surprisingly, the metal-bearing filament is a lot heavier. It's not as heavy as an all-metal item would be, but there's definitely some heft. It also feels a bit cool to the touch, a bit like something metallic would.

However, they don't come out looking like metal. The bronze comes out as a sort of sandy light yellowish brown while the copper comes out a red which is definitely in the copper color range, but neither appears metallic.

That's where the post-processing comes in. The first step is to sand the surface of the print smooth, optionally hit it lightly with a little black paint to create the illusion of a dark patina, and finally rub with a brass polish.

(That bit on the right is a late Warring States-era dagger axe, the first in a series of historically accurate polearm hair sticks I'm making for Stephanie.)

So far, it's coming out OK, but not great. I'm getting a definite metallic sheen, which is definitely understated in the photo but still not all that impressive. Clearly, I need a better sanding and buffing technique. I've seen a Dremel used with some buffing heads, but in my tests, I've been gouging the plastic pretty badly with it. However, the steel wool I'm using isn't getting outstanding results. Finely detailed pieces are also a problem. I'm having a difficult time with all the corners and little protrusions. I gave up on the retro spaceship from the first picture and printed a little sword in the copper to see if that would work better. There is one interesting thing, though: within a few minutes, the bronze head on the dagger axe was already developing a pale green patina. So, yeah, actual metal.

That's nifty, but there's one other notable thing about this filament. It's expensive. It's really expensive. It costs significantly more by weight than regular PLA, and it's much denser, so the same weight gets you a fraction as much length. All told, it's more than an order of magnitude costlier than plain PLA on a meter-by-meter basis. So it's neat for specific artistic effects, but run test prints with cheaper stuff first to make sure they're absolutely perfect. Mistakes with the metal-bearing filament are costly.

Wednesday, February 18, 2015

A Tale of Two Hexes

As a long-time gamer, I've played and continue to be involved in games which use hex maps, going back to Squad Leader and other Avalon Hill products. One of the earliest such games I played with the Metagaming/Steve Jackson classic Ogre, for which I've already done a bit of printyness. Having 3d cybertanks ready to roll is great, but what about the other units? The internet being what it is, somebody's already way ahead of me. The units are great, but can I add anything to that? Nifty as they look, they're missing something important for actual game play: stats. But that's easily dealt with. Whip up a hexagon of the appropriate size, extrude it into a suitable thickness, add a little 3d text around the edges, and drop the unit miniature in the middle.

One of the other games I'm into is the likewise Steve Jacksony RPG GURPS, which likewise uses a hex grid to regulate movement. Here, again, the internet's got my back, with these lovely floor tiles which could be used to build nice dungeons. All I have to do is to add some hexes, as in this test piece.

The problem here, though, is one which has plagued gaming map-makers for decades. People tend to build in rectangles, but maps are in hexes. The source tiles I'm working with are naturally in various sizes of squares and rectangles, and while I can lay down a hex grid of whatever kind I want on top of a batch of tiles, I have to decide between making either a geomorphic hex grid which doesn't make sensibly shaped rooms or a geomorphic square grid which ends up with a bunch of not-quite-hex-shapes around edges. Oh, well.

Monday, February 9, 2015

The Factory States

So what is my robot printing?

Well, a guy over on the SJ Games forums had an idea about printing out the 3d Ogre standup sheets, created as part of the Ogre Kickstarter campaign, on clear plastic, producing clear Ogres where you only see the design, not the material it's printed on. I thought that was rather clever, and then realized that if I could extract the shapes from the graphics in the PDF, I could print out the shapes and assemble my own plastic Ogres. I got to the point of having .stl files ready to print:


 And then I thought, "Hang on. I'm dealing with a 3d printer. Why am I messing around with 2d objects?" The goal became designing and printing a full 3d Ogre.

Design was largely an exercise in figuring out how to use Sketchup and what plugins I wanted to add in. Starting with some of the pieces in the "2.5d" design, there was a lot of use of the extrusion and scaling tools, and it all became much easier when I found a plugin that made geometric solids.

So far, so good. But printing was a problem. When I converted the design to gcode, Slic3r dropped significant parts. The conning tower, for example, simply vanished:

Why was it doing this? I had no idea. But on a hunch, I switched to different software for slicing and printing. Cura seemed to treat the design more as it was intended:

 That looked more promising, so I sent it to the printer. And this is what I got:

This Mk. IV is about three inches long. I note that features smaller than about a half-millimeter may as well not be there. I may play with this a bit and add a few more surface features like some tread texture. I also want the conning tower to be a bit different in various ways, and the front part should have steeper angles. And at some point, I should build a Mk. V.

Monday, February 2, 2015

Instant Stencil

The 3d printer continues to subtly transform the way we do things around here. Here's the latest example: certain people I'm married to needed to paint a superhero logo (Ms. Marvel's, to be precise) on a dozen or so small confections (kaju katli; yum). We already had the can of (edible) gold spray paint and a container of (equally edible) gold dust, but we also needed a stencil. It's certainly possible to do that with, say, paper or cardboard, but there are problems. Regular paper? Easily soaked through after a use or two, and at any rate floppy. Cardboard? Still a limited number of uses, one has to find the right kind of material (which means digging around for...well, you're not really quite sure what), and there's an inverse relationship between ease of working and durability, so even with some reasonably good material I'd probably have to cut two or three.

Enter the printer. The PLA is non-toxic, and it's easy to make the design. I fired up Sketchup, drew a square about the size of one of the pieces of kaju katli, used the pencil tool to draw the symbol (a stylized lightning bolt) on the square, deleted the lightning bolt space from the middle of the square, used the rectangle tool to add on a small handle, extruded it to a thickness of about 2mm, and exported the STL for printing.

Maybe two minutes to build the model, about that long to print, and we're ready to roll.

In the event, it didn't work out all that well. The rough surface of the kaju katli meant there were leaks around the edges. But it wasn't any worse than if we'd used a paper one!