Galvanic etching

• Power Supply

One of the reasons I'm tinkering with embedded control is that I'd like to make mechanical devices move, and move with a certain intelligence (i.e. informed by sensors, or following a prescribed program, etc). This is at least a two-part goal, with an embedded control component, and a mechanical fabrication component.

With respect to the fabrication side - I'd like to make these sculptures / objects / mechanisms out of metal, and I'd like them to look pretty interesting, too. Thus, I've searched around a bit for some DIY metal fabrication techniques, and found a great resource at steampunkworkshop.com. I followed the advice of Jake von Slatt over there, and last night built a little galvanic etching set-up.

Briefly, this involves creating a digital image, printing it on inkjet glossy photo paper with a laser printer, ironing the image onto a copper plate, peeling off the paper backing so the toner remains as a "mask", dipping the plate into a solution of copper sulfate, and then running a current through it to etch out the exposed part of the plate not protected by the toner mask.

I tried this a few weeks ago, but I didn't have a laser printer, so I instead experimented with masks made from black rustoleum paint, a sharpie marker, and color t-shirt transfer paper which I printed upon with my inkjet printer. Although I could definitely see that portions of the plate etched, none of those methods proved a satisfactory mask. I really needed a laser printer.

It turns out Kinko's won't print on inkjet glossy paper using a laser printer; they say the photo paper leaves a residue on the roller. Luckily, at a flea market in Columbus last week, I found a booth where the helpful proprietress was selling laser printers for dirt cheap - $25 each - and I happily snagged one. (I may go back and get a few more - they're cheaper than buying replacement toner!). When I told her what they were for, she was interested in the galvanic etching process, and I promised her that I'd show her some of my results. Maybe I'll etch her business name into a copper plate and see if she'll trade it for another printer. :-)

Anyway... Two nights ago I printed up an image, (the word "ALCHEMIST"), but foolishly didn't reverse the image, and equally foolishly used a low DPI setting, so it came out quite pixelated. Furthermore, (devil's in the details), I cut my metal part with tin-snips prior to ironing the image to the metal. Since the metal edges warped and curled with the force of cutting, these sharp edges scratched the face of my iron (luckily for me all my clothes are burlap ;-), but worse, the edges prevented good contact with the flat iron face. Thus, only portions of the toner melted satisfactorily to the copper. Next time I'll iron it on the as-received flat metal, and then cut the piece away.

When I first tried this a few weeks back, I used a lawn mower battery as my power supply. Nice amperage, but it wore down pretty quickly. My bud and co-conspirator Marshall thus modified a computer power supply for me to use, and after probing the dozen or so wires to determine which would give me ~12 volts, I hooked it up. Below are some images of the set up. (Click for closer view).

Above is the quick-and-dirty set-up I used for the my first real galvanic etching. I told myself that if it worked, I'd make a more permanent/convenient set up.

I measured the current flowing through the circuit, and it was a depressingly low 1.3 amps. The electrical wiring in the Milo Arts building isn't necessarily beefy enough for industrial use, so I made sure I plugged it into the outlet I haven't blown yet. Additionally, my studio neighbor is in a Rockabilly band, and she and her bandmates were practicing at full volume. I could feel their amplifiers hum before they started, so I knew our side of the building was pulling a good amount of juice even before I plunged my circuits into water. Maybe it wasn't such a bad thing that I was only drawing a minimal amount of current (at least compared to my lawn mower battery), but I was concerned this would take an impractically long time to etch.

I played with the position of the plates in the solution, and found that the closer the plates, the higher the current flowing. (Naturally). It kind of felt like a game of "Operation" - get the metal parts as close together without touching. I could get about 2 amps this way. (Click for close-up).

Below is a close-up of the plates in the bath. Useless side-note: the mystery toothbrush is so-called because it appeared out of nowhere a week ago in the company van.

One website linked to from steampunkworkshop.com, called GreenArt, provided a great reference for calculating the times necessary to etch to different depths, based on the voltage of your source and the surface area being etched. For my surface area, at 12 V, I calculated that it would take about 2 hours to etch to a depth of 0.5 mm. After about 90 minutes the toner began to flake off, and so I removed it from the bath, and although I really have no convenient way of measuring the depth of the etch, it's significant (enough to catch my fingernail on). Thus, I think the calculations on that site seemed about accurate. Below is a pic of the workpiece, as well as a part I tried a few weeks ago, masked with a t-shirt iron-on transfer, which didn't etch, but which did retain the image from the transfer fairly nicely. (A tribute to Durer and his printing process! :-) I bought the raw metal from Blick Art. They sell copper, tin, aluminum, and more.

As I said, I hadn't reversed the image prior to laser printing it, and thus the word is backwards. (One of my friends mentioned that I could use it for lithography. Perhaps next time I'll try printing an old hundred-dollar bill. :-). Additionally, due to the poor contact with the iron at the edges, the portion with the "T" flaked off, as did parts of the "M" and "C". Furthermore, every 15 minutes or so, I used the mystery toothbrush to scrape off some of the residue depositing on the workpiece. I think this also scraped off a bit of the toner mask towards the end of the process, which left streaks in the portions above and below the word. Next time I'll just let that powder residue deposit on the workpiece, and see if it still etches nicely, and not bother with the occasional cleaning during etching. Below is a photo of the finished part, reversed, so the word reads in the right order.

Note that the rough edges of the lettering is probably due to the pixelated nature of the image I printed for the toner transfer; I only used 72dpi, instead of 200 or 300. The next one I do will probably come out much more smoothly, now that I've worked out a few of the kinks. Fun stuff!

Sign in