For the past few months I’ve been working on a project that involves creating a series of plastic miniatures that work with an Augmented Reality app. Creating toys from 3D models built for a game seemed a natural application for 3D printing. The idea being I make a positive out of a game model with a 3D printer, then use that to create a silicone mold for resin copies (a very old school process). To figure out how this all works, I’ve spent a few months immersed in additive manufacturing.
My first stop along this journey was to my local 3D printer store. I visited DeezMaker in Pasadena to see their BukoBot in action. This is a hobbyist printer that uses the same sort of filament printing technology the pioneering MakerBot does. The staff at DeezMaker gave me a good crash course in the different models, usually ranging from $600-1500. The software used to run the machine, Slic3r and Repeiter Host, are free and open-source.
A friend of mine has a BukoBot and was willing to do a test print of a figure. The first step was to ready the in-game model for printing. 3D printers need a model in STL format, which is a simple conversion process from most 3D programs. However, the model itself must be completely solid. You can’t have any holes, floating one-sided polygons, and other non manifold edges. Slic3r (the software used to ‘slice’ the model into vertical chunks used for printing) detected these problems in the STL file and spit out a series unhelpful error messages. The artist that built the model was able to fix these problems in Maya while running it through the slicer to check for errors.
With a cleaned up model file in hand, I visited my friend’s house and we began the print. It took a few hours to slice the model and set up the print area properly. This involved scaling down the model a bit to fit on the printer’s bed and messing around with the machine to get it aligned properly. Soon, the print began, and we waited with giddy excitement.
3D printing a figure. ETA 13 hours. vine.co/v/b6EUgJQO0Xl—
Ralph Barbagallo (@flarb) February 22, 2013
Some 20 hours later, my print was finished and the end results were…..well not that great.
In the BukoBot’s defense, this was a single material print. For complicated objects such as this figure, you need to use a support material. This is a water-soluble substance that is printed underneath overhangs and other floating features. When the print is done, you submerge the object in a bucket of water and wait for the support material to dissolve. Otherwise, it’s going to print ‘fluff’ material to support the model and you’ll have to break it off by hand.
Breaking off support material. vine.co/v/bgPPQBrtxPA—
Ralph Barbagallo (@flarb) February 26, 2013
In reality, these kit printers kind of suck. The engineering tolerances are pretty loose–especially in the BukoBot’s case as the printer is built, in part, by 3D printers. A lot of the pieces such as those that guide the printer head aren’t very precise and result in skewed prints. The printer is also prone to erroneously burping up large blobs of material, throwing off the accuracy of the print. Finally, the object lies on an open bed and cools rapidly to room temperature which warps the material. The end result is kind of a disaster, even on the most simple of objects.
My next step was to see what options were available for high-end 3D printing. I considered services like Shapeways, but their turnaround time was too long and they have draconian limits on file size and triangle count. I talked to a few different places that had professional-quality printers, and eventually printed my model using a uPrint SE by Stratasys at TekPro Group in Woodland Hills.
Sure, the printer costs nearly ten times what the BukoBuot does, but you get what you pay for. The end result is simply amazing. The uPrint uses the same FDM process as the BukoBot but is much more precise. It still took over 20 hours to print, but in this case it’s worth it.
If you look closely you can still see some roughness to the surface. I gave the printed model back to the (elated) artist and he’s working on smoothing it using acetone and sanding. Before we do a resin cast it may be necessary to carve more detail into the figure as well as use putty to fix a few errors we didn’t catch in the modeling phase.
Low-end kit 3D printers may be disruptive to the high-end machines. Certainly the quality and fussiness of printer kits reminds me of personal computers from the Homebrew Computer Club era which disrupted mainframes. At the surface, the war between professional and kit printers seems to be torn from the pages of The Innovator’s Dilemma.
As it stands now, kit printers just aren’t ready for prime time–by far. Newer models have shown dramatic increases in performance, but we are still probably a generation or two away from approaching the quality of high-end machines like uPrint or Objet. Meanwhile, the search is still on for a consumer application for 3D printers that will make an audience beyond hobbyists and manufacturing actually want them.