Friday, June 26, 2015

Don't Trust Robert's Advice on 3D Printed Buildings

In the past, I've declared that 3d printing isn't great for buildings. Most buildings are easy to make using sheet styrene or sheet wood - I can probably build them faster than I could draw them on the computer. I only need one copy of a particular building, so the ability to print multiple copies isn't so interesting. Buildings tend to be large; they're either slow to print, or won't fit at all in the 3d printer. I'll admit I've used the 3d printer to make details (such as the architectural details for the Market Street station, or the freight doors I borrowed from the Guggenhime packing house. But buildings? Not worth the trouble.

Well, actually, sometimes the 3d printer can help with buildings, and two recent projects highlight when the printer might come in handy.

J. S. Roberts and clerestory windows Early 20th century buildings often used clerestory windows - raised sections of roof with small windows - as a way to bring light into the center of a building. We usually think of this with sawtooth roof factories, but you'd also see this in frame structures. The J. S. Roberts packing house off San Carlos Street got light into the fruit sorting area on the third floor with a raised roof and some small windows, all above the barn-like building's normal roof line.

J.S. Roberts clerestory windows

Clerestory windows like this are a pain to scratchbuild. There's not a lot of good, tiny window castings if you're trying to model a strip of two foot high windows. Because the clerestory roof is often only raised a couple feet, building the walls around the window can be difficult. If the clerestory is on a peaked roof, then I've also got to cut the ends to match the roof slope precisely. The last time I did a model like this, I used some 3d printed blocks to form the clerestory section, but still had to cut the lapped siding sheet to fit. For that model (the Earl Fruit Company packing house on Ryland St.), I wasn't completely happy with the overall clerestory look (too few windows, clerestory section too tall), but I decided it would do.

When I completed the main section of the J.S. Roberts packing house, I knew I needed to make another clerestory roof to finish the project. However, the 1930's photos I'd seen showed a very short space between the two roofs. I couldn't think of any commercial windows that would work, and I knew the whole assembly would be tedious to build.

Clerestory 3d model

Roof and access hatch

Instead, I cheated - I sketched the whole clerestory roof - siding, windows and all - as a single piece. I didn't have to worry about cutting the angles against the roof (as long as I measured the angle correctly.) I could size the windows to exactly match the space. I did have to draw out the location of *every single board* for the siding - tedious, but doable on this relatively small model. The result was too big for the printer, so I split it in two, printed two copies, and glued the pieces together to make a completed clerestory assembly. After painting and attaching corrugated siding for a roof, I was done. The windows, although solid, got some black gloss paint to make it look like there really was glass in those panes. The result was just right - a roofline that matches the prototype and I didnt have to cut any fussy angles.

In fact, I was happy enough with the result that I did another 3d print for the access hatch on the roof. Again, by 3d printing it, I didn't have to worry about cutting the correct angle into plastic siding, and didn't have to worry about fabricating a tiny box perfectly square.

Campbell passenger shelter for the Peninsular Railway

The Sourisseau Academy, one of the local history organizations in San Jose, posted a picture recently of downtown Campbell in the 1920's. The photo, showing the area around the railroad tracks and Campbell Ave., reminded me that the former Peninsular Railway, an interurban line, used to go from downtown San Jose to downtown Los Gatos via Campbell. The tracks cut through the new suburbs of Willow Glen and turned here and there through orchards. Just before the tracks reached downtown Cambpbell, they curved off Campbell Ave. and onto Railway Ave to pass next to the old S.P. depot. The Sourisseau's photo shows Campbell Ave. at Railway Ave. - the curve of the interurban tracks, the overgrown mess of the SP's former garden at the intersection, and a small passenger shelter at the intersection which must have been handy both in the rain and on really hot days in summer. (The photo is visible in the Sourisseau February 2015 video, showing the interurban tracks turning south onto Railway Ave.)

Now, the passenger shelter wasn't much - a bench for a couple people, simple siding, and some minor architectural details to please the town burghers. It's also exactly the sort of model that's easy to make in styrene with a couple hours work. I was still curious whether I could make one in 3d, and did some quick sketches.

Applying the siding was the annoying part; rather than just pulling some lapped siding sheet out of a drawer, I had to draw each board: correct spacing, correct overlap at the bottom, etc. The rest was pretty trivial, with the hardest part being the choice of how to print the model. (I chose to print it upside down, with the roof having the bad surface where the model attached to the 3d printer.) Some filing, and some Campbell shingles completed the model. It's not much, but it showed me that I could create small structures quickly. More importantly, I got to play around a bit with the design and experiment with the decorative eave ends. For my layout, 3d printing the shelter is a bit of overkill, but if I needed three or four, it might have been a suitable project for 3d printing.

I still think 3d printing isn't particularly useful for 3d structure models, but these two projects reminded me that 3d printing could still be handy either for work that would be challenging to do by hand, or where I wanted to experiment with details or shapes.

Showing off the 3d Models

I’ve obviously been writing lots about the 3d printed freight cars over the last several months. Apart from showing them to the local model railroad lunch group, I’ve had few chances to show them off in person, show off what 3d printing can do, and share models of otherwise unavailable freight cars.

Last weekend, I went off to the Bay Area Prototype Modelers meet. BAPM is a yearly get-together for sharing models and discussing techniques. BAPM is arranged as show-and-tell; they provide tables, you put down models. Unlike NMRA meets, there’s no contests or judging, no explicit demands for lengthy write-ups describing the prototype - just modelers coming together to share what they’ve been doing. BAPM isn’t unique; there are prototype meets like this elsewhere in the U.S., giving modelers a chance to share. The meets also tend to attract folks interested in specific prototypes, and in modern (1980 and later) models. (See BAPM’s photos from the meet, or read Tony Thompson’s review of the event.)

This sounded like a great crowd to see the freight cars - even if my models aren’t perfect, I suspected folks would be interested in using 3d printing for making specific prototypes. So I hauled a bunch of my freight cars off to Richmond, spread the cars out with a bit of information on how they were made, and had some good conversations.

I went in with a bit of an agenda - the same one I’ve shared here often enough. I wanted to show that a home 3d printer (ok, a pricey home 3d printer) could make models that were approaching the quality of injection molded kits. I wanted to highlight that the technology worked really well for making lots of cars, both because printing a new car was easy, and because I’d end up with lots of “good enough” test prints as I was working on my design. I wanted to show that designing the 3d model was tedious, but possible for folks who were novices at making CAD drawings. I wanted to stress the race between having enough models, and finally getting the design with all the detail I want.

Folks heard that message loud and clear. Harry Wong, one of the organizers, walked up early in the day, and asked “which of the cars was 3d printed”? I waved at all twenty cars laid out on the table. “All of them.” (Pro tip: when showing off 3d printed models, always bring a whole bunch of models to stress that you can make these at will.) Tony Thompson, who wrote the SP freight cars books I used for source material, loved seeing SP prototypes that weren’t available commercially.

I also shared how the 3d printer made it easy to make all the little parts for set dressing - lug boxes, drying flats, piles of sacked prunes. Those parts might not be cheaper than the commercial ones, but it’s nice to be able to make new ones whenever I’m in a set dressing mood.

I also broke one of the prime directives of BAPM. In violation of all the “don’t touch the models” signs, I brought a handful of 3d printed parts for people to touch and examine so folks could get a real feel for 3d printing. (I explicitly had an area on my display to hold the “ok to touch models”, and marked them all so it was obvious what was touchable.) A couple folks noticed the slight warp I’ll get in some of the flat cars (sometimes fixable with careful drying in the sun, sometimes not). Others could feel that the 3d printed cars could be robust enough for operations, or get a sense for the level of details on the models.

Several others asked about what it took to make a 3d model, and whether their particular model could be printed easily, either at Shapeways or on the Form One. One modeler scratchbuilding wind turbine blade loads for flatcars asked about 3d printing some of the odd-shaped mounting hardware that the real railroads use to hold the windmill blades. His particular pieces were perfect for 3d printing - small, slightly odd shapes, with enough bolts and other details to make for an interesting model. Some folks were asking about boxcars; I mentioned my one try at a boxcar and that the 3d printer didn’t do perfectly flat surfaces as well as styrene sheet.

Meanwhile, I also got to see some great models. Along with lots of traditional modeling, I saw three or four other folks building with Shapeways. Jason Hill of Owl Mountain Models had a Shapeways master for the Southern Pacific F-50-4 flat car he intends to make as an injection molded kit. (I spent a good ten minutes comparing my model and his to see what detail I left off; Jason's crawled all over the similar Union Pacific flat car at the Orange Empire Railroad Museum, so he's matched the prototype much better than me.) Other folks were using Shapeways to make specific boxcar doors that weren’t available on production models. There were also a ton of great, traditional models; Jesus Pena showed off the huge fleet of UPS trailers he’s been building for his Free-mo Richmond intermodal yard - at least forty trailers, with plans to double that number for realism.

BAPM is definitely a great meet, and worth attending; it also got me fired up to do the next set of refinements on my models. Next year, BAPM will again be in Richmond on Saturday, June 18, 2016. I’m hoping I’ll have some new models to show off there!