(notes to future self)

HTPLA from Proto-Pasta is reallllyyyy good in the kiln. Other PLA (lookin at you, Jessie) kinda sucks.

200F / 2h: no warping, good result

210F / 2h: warped, layer separation

205F / 2h: warped (but not as bad as 210)

200F / 2h: wood PLA + HTPLA, minor warping in wood (barely noticeable), no separation

New idea: permanent christmas lights. I read about this on a company "makers" mailing list, no credit to the author (it's been some time, lazy, etc). The idea is simple: use programmable LED strips, mount them (permanently) outdoors, and update them for the seasons.

The LED strips are the easiest part. Amazon sells them in 5m reels (like this), and they can be daisy-chained for arbitrarily-long strips. The LEDs are spaced out, depending on what you want. For side-of-house application, the least-dense option might work well -- 30 LEDs/m, or ~150 LEDs in a 5m segment. That costs like $20 ish. By math they probably consume about 10A per 5m (60mA per LED * 150 LEDs -> 9A, plus various loss along the way. Each strip can have it's own injected power... The linked ones include a waterproof enclosure. One could also feed them into (say) translucent PEX tubing and skip the silicone jacket.

For an MCU platform, I like the Particle Photon (or its kin). Easy stack over wifi (no USB cable) and the "cloud" interface is nice. With minimal programming I can have a web interface to change features like speed, brightness, pattern, etc.

The physical device will be straightforward: a plastic box, maybe 1-2 buttons, a terminal for the 10A power input, and a plug / dongle to tie into the first LED strip. Probably a CdS sensor (light sensor) to adjust brightness, turn off at dawn, etc.

I prototyped the first setup on a breadboard. The LED strips use 5A power which I can pull directly for the MCU, and the MCU drives 3.3v signal which also works great for the LEDs. FastLED library, of course. A simple button (PULLUP + ground) for "change the pattern" (short press) or "change the speed" (long press). CdS is a 10k resistor + the sensor itself, NBD. I made a generic protoboard and they're on order from OSHPark, who are thankfully still open during CORONAPOCALYPSE.

The generic protoboard will do just fine for such a simple project. It has a terminal input for any power level (5-30v), and onboard solder points tied directly. Solder up the 5v pigtail and use the terminal for power supply input. All data pins are broken out in two positions; use D0 for the pigtail data line. Button is easy (mount to a case, one leg to a data pin, one to a GND pin). CdS is almost-as-easy: 10k resistor in the middle line between GND and an analog pin, CdS between 3v3 and the same analog pin on an outside row. Expose that to the side of the case for light entry.

The two components are a medium-sized cap (like 2-10 uF) and an LP2950 regulator.

TODO: get the boards, design + print an enclosure.

Animations are really the easiest part of all this. Once the hardware is done one could update the animations via WiFi... For the time being I've really got one class of animation: seasonal colors, moving slowly. When I'm sitting right in front of it I can't tell it's moving, but if I look up every 2-5 minutes it will have completely changed. That's driving in a simple function which uses a seasonal pallet of 4 colors: red-white-blue-white for independence day, green-red-green-red for xmas, orange-black-orange-black for tgiving/halloween, etc. I like these because they're both dynamic and subtle.

Also considering "chasers" or some active animation, but I don't want them to be too annoying to my neighbors, most of whom I like.

If you start sintering a bunch of crap, or otherwise start running a kiln a lot, you may wish to know what the *actual* interior temperatures look like.

I rigged up a thermocouple on a Particle Photon, to independently measure and record temperatures. The kiln has an independent digital display based on its own thermocouple, but I can't record those values beyond pen/paper (which is obviously tedious)... and I'd still want some way to independently verify the internal temperatures.

Having an external thermocouple also lets me bury it in refractory to get some sense for heat penetration and transfer delays.

There are a few common ways to age or force "natural" patina onto bronze and copper alloys.

  • Bright green verdigris: copper carbonate
  • Dark green verdigris: copper II acetate
  • Dark brown or black shading: ?? (some sort of copper sulfide?)

The verdigris can form spontaneously just from cleaning, and will "stick" best to very clean (polished) bronze. They can be prevented with light waxing, and can wipe off with a soft cloth.

The darker color, as an "oil-rubbed bronze" finish, is an integral part of the surface material and can only be removed with aggressive wire brushing, sanding, or other surface removal.

skulls with various finishes

Note: this is an addendum to the 3rd step of a print / sinter / polish process.

Solid bronze can be finished and polished to a bright luster, not unlike jewelry. It's a very hard, beautiful metal with some cool properties. It cleans up similar to metal-filled plastics, but there are a few steps specific to solid metals.

tl;dr: pickle, brush, shape, tumble, maybe brush again, polish, buff, wax.

Note: this is step 3 of a print / sinter / polish process.

Now that you've successfully printed with "metal" filament, and possibly sintered a solid metal part, you want to make it look nice. These steps are generally chronological, but you can always go out-of-order or do a step again. If you do plan to sinter, I highly recommend a pass of trimming & shaping before the sinter, then further cleanup after.