While in France, the combination of being somewhere near ‘church mouse poor’ and living in a small (32 square metres / 350 sq ft) 1 bedroom apartment, I looked for hobbies that took up little space and so settled on photography (i.e. my Project 365 here) and dabbling in computer programming and to a lesser degree, artificial intelligence (my WmDOT for OpenTTD). It was quite enjoyable, but on my return to North America, I found myself with a lot more space and more geographically stable. I wanted to make things that were a little more physical and so with the encouragement of my Honey, I started into Model Railroading, a childhood dream. One of my shocks was the cost of models! A regular-ish building, measuring maybe 2x3x3 inches would cost $60! There had to be a better way…
Enter 3D Printing
Various technologies exist but the most popular, at least at the hobby level, basically consists of a hot metal nozzle that melts a ‘thread’ of plastic and then uses this melted plastic to paint, layer by layer, your desired object. A kit for a printer like this can be had for \~$1300 which would be about the same price as 20 buildings. And with a 3D printer, most of the time is spent creating the computer model rather than in physical model building, so I started wondering if I could sell some models and let someone else pay for my printer and railway modelling…
Good research starts with a review of those who have already worked on the problem. My Google searches turned up a few sites of note.
The first is a note from 2008 talking about the feasibility of using a 3D printer for model railroading. There were two comments of interest: one, 3D printers are expensive, and two, the resolution isn’t there yet.
Next is a series of posts (1, 2, 3) from about the New Years of 2009 when Shapeways, a 3D manufacturing company, was invited to a model train conference in the Benelux (Belgium, Netherlands, and Luxembourg). Not much was written as follow up, so I don’t think the effort worked as hoped.
Lastly, a came across the blog of Robert Bowdidge, who owns a 3D printer and a model railway. His printer arrived in October of 2009 and he starts writing about trying to create a custom building for his model railway two months later. His concerns are again about resolution, but mostly in regards to getting a flat surface on the finished printed part and warping during cooling. With a little bit of tweaking, he decides that the result is acceptable from a three feet back (which is where most of the observing of the railway will be done anyway). He mentions some of the benefits too: easy to create duplicate parts and easy to create parts with complex geometries.
So the two issues that stand out are cost (of the printer) and resolution.
Professional 3D printers don’t come cheap. Searching online does not also show prices, but from what I can gather, professional printer lines start at $10-15,000 and can run to $60,000 or more and many also require maintenance contracts. The hobby printers are cheaper - $800 for the parts of a RepRap or $1300 for an unassembled Thing-O-Matic. Without someone else paying for it, I don’t think I can afford the professional level machines.
The other cost is that of materials. With many professional printers, they use proprietary powders or epoxies. Prices are hard to pin down, but the few prices I could find suggest $500-1000 per litre or more. In contrast, many of the hobby printers use plastic that runs $50 per kilogram or a couple of cents per cubic centimeter. Although not free, if it defendant within reach (like how we don’t think about paper in our inkjet printers today).
Resolution is a measure of how fine of thing can be printed and still be seen. Ignoring the professional printers (already ruled out on the basis of cost), the basic resolution of the hobby printers is based on the size of the nozzle. For the Thing-O-Matic, I found a quote that the “stock 0.4mm nozzles allows printing at a layer height of 0.32 and 0.27mm right out of the box” (link). That works out to 3 or 4 layers per millimeter or 79 to 94 layers per inch. The record, according to Makerbot, is 10 layers per millimeter, or about 250 layers per inch. That record has stood for over a year, so I have hope that that could be improved by myself or others soon.
My feeling is that if I could get 150 layers per inch (approximately 150 dpi or dots per inch), I would have something workable. I remember when laser printers used to give you 150 dpi and it seemed rough, but very workable. So I think resolution is workable, although I will keep my eyes open for upgrades to that.
Competition is good and bad. On the one hand, if there is some competition, it means the idea isn’t too crazy and someone else may have developed a market for you; if there’s too much competition, it’s hard to stand out.
I found two competitors: Shapeways and MakeMyModel. Shapeways is a print-on-demand service that allows you to upload a model and then they print it and send it to you. So Shapeways itself isn’t a competitor, but their community might serve as competition (like eBay). Their search brings up a couple of things (1, 2), but not a huge selection and nothing in N scale (which is what I am modelling in). As a side, depending on the margin I could generate using their service, this might work as a drop-ship service for my own projects.
MakeMyModel appears to have the same idea I do: create railway models and then produce them on a 3D printer. I wasn’t able to figure out what kind of printer they use, but anything I added to my cart was out of stock, so I’m not sure they’re in business anymore.
3D printing is still a developing technology and as such is not “state-of-the-art” for model making. However, it is available at an accessible price point and allows creation of models at home! As a plus, I don’t think it’s unreasonable to turn it into a small side business. One of the big bonuses of having the manufacturing in-house like this is that the inventory I need to keep on hand is very minimal. And even if the business side fails, I still get to use it to build my own railway!