Printing the Ultracortex MIII

In the previous story I started to explain how and why I planned on 3D printing the Ultracortex Mark III, which is the headset of the open source OpenBCI EEG system. I performed my print on a Lulzbot Taz 5 printer, a real work-horse in our shared lab with almost constant use for the last several months. Unfortunately, my first attempt did not go so well, but fortunately we can all learn from it.

What you’re looking at is what I consider to be defiance, but more likely it’s stubbornness.The suggested print settings recommended that the headset be printed with PLA plastic and heavy supports, which prevent the plastic from oozing all over the place. Instead I used a free program called Meshmixer by Autodesk to generate custom minimal supports. You can see some of them like tree branches in the image. A broken support is indicated by the red arrow. They were not enough to hold up the overhanging parts of the plastic, and in the end could not even hold themselves up under the high temperatures of the print.

The other mistake I made was to use a high temperature plastic: ABS. I often used ABS for simple parts like electronics enclosures and mechanical parts because of its strength and flexibility. While the mild-tempered PLA requires only 180 to 190 degrees Celsius to print, ABS needs a good 230 to 240 to start squeezing out. Extruded plastic cools down rapidly after being deposited on the previous layer, but often a cooling fan is used to speed it up and harden the layers below so they don’t do weird things like wobble or bend during the print. PLA tolerates cooling very well. ABS does not, resulting in warping and layer detachment due to the rapid cooling. Usually we print ABS without a fan but the harsh overhangs and steep angles of this print demanded it. Unfortunately this also meant that the finished part started falling apart in my hands.

So with lessons learned I decided to use more classic software to generate the supports. I used Slic3r’s built-in support generation with 5 mm honeycomb spacing at an angle threshold of 35 degrees.

It looked like a lot of plastic, but I switched out to black PLA filament and started the print. The front took a full 16 hours to print, the back was similar. Each part took about 150 g of PLA filament, about half of which was support material. I clipped, twisted, and crushed the support plastic off.

Now we’re getting there. For your entertainment here’s a timelapse of one of the 16 hour prints:

And if that was just too fast for you, here it is again in slow motion:

Next I will describe printing the smaller parts.