Learning to love difficult prints – Space Pod from 2001
There is an experiment I’d been meaning to try for a while, and my Thing-O-Matic is now printing well enough that I felt I could try it. It is to take files from a site such as the 2001 Model Archive, and see if I could print them with little editing. The result was a qualified success: you can do amazing things with the Thing-O-Matic and support material enabled, but trying to remove delicate parts from the support material can be a real challenge. Some degree of editing of the source files is a good idea, to make them print-ready.
For the first attempt, I chose the Space Pod from “2001: A Space Odyssey”. I loaded it into Meshlab and had a look. Keep reading after the break for a longer description of the process.
The first thing that jumps out is that the arms will be very difficult to print, and that support material will be necessary. Also there are some extreme overhangs on the side. I used Meshlab to clean up the file a bit, simplifying the number of vertices, scaling and exporting to STL. After doing this, ReplicatorG seemed to have no difficulty in reading the file.
I initially tried generating GCode using the default Thing-O-Matic profile with raft and support enabled. However, this did not work too well, and I aborted the print after 45 minutes. There was too much of a gap between the support and the object, so the overhangs around the underside of the pod were full of droopy strings. I’d noticed that overhangs are handled much better when you use a profile with a reduced layer height and a wider thread width, so I used Makerblock’s Profilemaker to make a profile with a layer height of 0.25mm and a width of 0.5mm. I also narrowed the distance between the support material and the object to 0.6 of a thread width.
I set it printing, and came back the following morning to this:
The pod is comprehensively encased in support material, especially the arms. I’d used the excellent Pleasant3D preview tool to check out the GCode, so it came as no surprise. However, the cleanup job would require some effort. The big clumps of material in the forward window, and in the side thrusters, came out pretty easily by just grabbing them with pliers.
The support material around the arms posed a huge challenge. I ended up using an X-acto knife to cut through the ends of the support material, where it folds, and peeling it away one layer at a time. In spite of considerable care, the arms ended up breaking anyway, into tiny tiny pieces. I’d come this far, so I carried on removing the supports and used cyanoacrylate glue (Krazy glue) to repair them.
The results look pretty good. I am particularly pleased with how well the “handgrips” rendered, they go all the way through. There was support material in there, and it came out really cleanly. Overall, edges are sharp and details are crisply defined. The arms mostly printed OK, though they are hollow tubes, and so fragile that they flex if you breathe on them. In the picture, I still haven’t gotten around to re-attaching the claw “fingers”, which printed out correctly. They’re just a couple of millimeters long.
Conclusion of the experiment: the Thing-O-Matic has the resolution and the positioning accuracy to do some amazing things. The arms and details are correctly rendered. The finish of hte model overall is very good, and details are very crisply defined. However, the main constraint in achieving incredible prints is the lack of an easily removable support material. The bulk of the “hands-on” time on this project was spent removing the support material, and repairing parts which were damaged during the process.
Another way to get great results would be to edit the model, so that delicate parts could be printed individually, with little or no support material. However, this would add some more “hands-on” time, and I was interested in seeing if you could from model to print with little manual intervention.
Error - could not find Thing 9036.