Monday, September 14, 2020

Keeping Up with Jason: 3D Printing Beet Racks

If you've been following Jason Hill and his Owl Mountain Models, you might see that some of my personal projects have mirrored his commercial projects. He cut molds for his injection-molded F-50-4 flat cars; I 3d-printed the earlier CS-35 flat cars. He's done steam locomotive parts (3d printed and otherwise), I experimented with a 3d printed boiler for a C-11 Pacific. He's experimented with Harriman passenger car customization, and I've 3d printed some C-60-1 bodies.

Some of our overlap isn't surprising. We've got similar interests; we're interested in the steam era on the SP. We're both likely working from the easily-available plans in some of Tony Thompson's SP books (at least until we get curious enough about details to wander up to Sacramento and the California State Railroad Museum library and archives to see the actual blueprints.) We're interested in making lots of particular models - commercially and injection-molded in Jason's case, and for my own use and 3d printed in my case. We've also just been talking lots and comparing parts produced by each other. Many of my conversations with Jason about interesting models and manufacturing encouraged me to try building various models.

The Tony Thompson freight car book on flat cars had more than plans for early SP steel flat cars to inspire us both. He also included photos and plans of the various temporary sides and sugar beet "racks" that the SP used to make the flat cars useful for occasional traffic. Sugar beets were a big commodity on the SP, often seen up into the 1970s going from the fields to the various sugar beet processing plants located in the Bay Area, Salinas Valley, and Central Valley. SP track diagrams from the 1960's even show a track in Mountain View labeled "sugar beet dump" - about where the Microsoft, Google, and LinkedIn shuttle buses pick up folks at the Mountain View station... or at least where they picked up employees in the days before COVID-19.

Sugar beets were heavy, large, and were shipped in huge volume during the harvest season, so the SP needed a cheap and easy way to ship them. The crop was too seasonal to deserve dedicated cars, too bulky for low-sided gondolas, and too low-cost to deserve anything too nice. So during the first half of the 20th century, the SP would build latticed sides out of two-by- lumber that they could put on any ratty flat car, dump the beets in the top, and open the sides to let them pour out at the sugar refinery. When those cars got too worn in the 1950's, the SP took steel gondolas, then added wooden sides to increase the capacity to haul more of the relatively-light sugar beets. The early cars with the latticed sides are much cooler in my opinion than the later cars - the airy, slatty cars always looked a bit jury-rigged, and battered and worn enough to give a modeler lots of weathering fun.

Close-up of beet rack

Jason and I chatted long ago about the beet racks and how they make interesting cars. Since then, Jason took the effort to cut injection molds for his Blackburn patent beet racks, sized to fit his F-50-4 flat cars. They're beautiful models, with much finer detail than I can get with my 3d printer. All those conversations also encouraged me. I went after similar cars a couple years back, 3d printing a few beet racks based on an earlier, non-patent design also in Tony Thompson's book. If you want a few beet racks for your layout, I'd go buy some of Jason's. I'd still like to tell you about mine because they say a bit about what's easy and hard with 3d printing.

Jason's beet racks are separate plastic parts sized to fit his existing flat car models. Injection molding's good for that; it's a reliable process for high numbers of parts, and parts keep the same dimensions. To keep costs low, making parts flat, thin, and consistent thickness makes the molds easier to cut and run, and minimizes warpage of completed parts. For the beet racks, that means that making the slides as four flat pieces is easiest and the most inexpensive. In contrast, large 3d structures are hard to do with injection molding. Trying to print the flat car and the beet racks simultaneously would require large, deep molds with several pieces that need to slide together to close the mold - a challenge for the major hobby manufacturers, and near-impossible for the garage manufacturer.

3d model arrangement in the printer: bodies printing vertically, with rack ends as a separate part.

With 3d printing, the rules about what's hard and easy are completely turned around. Because of printer miscalibrations, a printer might have slightly different scale in different directions, making it hard to keep parts the same size unless they're printed in the same orientation or axis. Printing thin, flexible things can be hard with the Form One because the part will flop during all the movement as each layer is printed. As I've mentioned before, my Form One's temporary support structure relies on many little sprues to hold up the part and form the surface it begins to print from. Where a part starts printing is often roughest because of these supports; the best detail is usually much better in the middle and top of the parts. I like to think of it as "the 3d printer doesn't like to start new parts". If I can print a new layer that's well-connected to the previous layer and requires no outside support, I'll have better quality and more successful prints.

I'd started trying to do the beet racks as separately printed parts, but found that didn't work at all. The resulting parts were floppy, inaccurately sized, and rough where the supports attached. I ended up redoing the model so that the beet rack sides were part of the flat car model, and the ends were separate parts printed on a separate support structure. I also printed the models vertically, again omitting part of one end so that the support structure wouldn't join to a visible face, and then 3d printed a separate part with two feet of deck and the car end. As a result, I only needed supports along a short edge of the beet racks rather than along one of the longer edges. The attachment to the car body also stiffened the lattice structure.

As always, detail that's close to the plane of major parts is easy to apply and comes out fine. The hinges and door latches are just embossed designs raised up or lowered relative to the rest of the design. The slats printed well as long as the board was well supported, and connected back up to posts frequently. Rather than trying to fiddle to get posts and stake pockets to match, the single-piece body made it easy to have everything look realistic and fitting well.

The beet racks also showed how 3d printing works great for variants. For Jason, cutting a new version of a flat car often means designing and cutting new molds from scratch. With 3d printing, it's much easier to borrow the flat car model, combine it with the beet racks, and print a combined model. It also made for an easier model to assemble, without any need to get the beet rack sides and flat car body aligned correctly, or to figure out how to trim posts to make sure the sides matched the stake pocket locations.

The area I model around San Jose, Campbell, and Los Gatos never had sugar beets as far as I know. All the local sugar beets were grown in the flat lands around Moffett Field, rather than further south in the Valley. For example, Henry Mitarai, a Japanese-American farmer, grew acres of sugar beets on his farm off of Mathilda Ave. in Sunnyvale during the 1930's. Dorothea Lange photographed him in his fields in 1942; shortly after, he and his family were sent off to the Heart Mountain internment camp in Wyoming for the duration World War II. Mitarai didn't return to Sunnyvale after the war; he and his family stayed in Utah and grew sugar beets.

Even if the sugar beet cars aren't appropriate for my orchard layout, they're cool cars. They're also a nice reminder of the history of the Santa Clara Valley: we grew many crops besides fruit orchards, our current urban towns had agrarian beginnings, and we made money in some interesting ways before social networks.

Saturday, September 12, 2020

Rebuilding Glenwood

A string of maintenance-of-way gondolas heads uphill.

Like I mentioned last time, long-suffering Glenwood got some serious rework recently. I had some good reasons to finally return to Glenwood. The ProRail invitational was going to be in San Jose in April (before COVID-19), and the visitors from around the US deserved to see the layout at its best. I needed a distraction from work, and wanted some projects that could fill a weekend.

I also had several years of pent-up frustration waiting to be unleashed. Glenwood’s also an old part of my layout; I laid the track on the upper level about two years into the layout, and roughed in some scenery. The model doesn’t accurately capture the real location. It’s not up to my later standards for prototype scenes. It's not eye-catching enough to be a focal point for the layout. Glenwood’s also in a darker corner of the layout, and in a location that’s not central to model railroad operations, so it’s never gotten a lot of scrutiny. Glenwood deserved be

But being generically dissatisfied is one thing; I need a list of things to fix. Let’s run through the problems at Glenwood.

Glenwood before the rework.

Problems with Glenwood:

  • The road’s unrealistic, just badly-levelled Sculptamold on foam scenery, with sheer drops and no shoulders.
  • The building just above the tracks isn’t prototypical; although there was a small house there, the building on stilts doesn’t match the location, nor does it look realistic for the 1920’s. It also draws attention away from the prototype portions of the scene, hiding that great curve into the cut and tunnel.
  • I'd built a station building, but it’s a coarse plastic model. There’s none of the maintenance of way buildings or outbuildings seen on the maps.
  • The grassy hillside doesn’t quite match reality; prototype photos show more trees. The grassy hillside doesn't hide the unprototypical terrain, and misses the chance for trees as a view block to frame the scene.

So my plans? Tear out the hillside, improve the tunnel entrance, make the scene more realistic overall, and detail the station area.

Redoing Scenery

Step one was ripping out a bunch of bad scenery - taking out the hillside, the cut, and the house-on-stilts.

Before any of that, I took a pass at a bunch of other unfinished business. The turnout in front of the tunnel was a frequent derailment site. I ripped out the track, leveled it out with spackle, and relaid the track. The old Tortoise switch machine was a problem; it stuck out too far below the bottom of the deck, and was difficult to adjust. I swapped it out for one of the tiny MP-5 switch machines. I also took this opportunity to check on the track in the tunnel, pulling up even more track, using spackle to again ensure the roadbed was as level as possible, laid the track better, and sealed in the tunnel so that stray light didn’t ruin the illusion of a tunnel through a mountain.

With the track done, I hit the rough scenery.

The new hillside started out with the focal point: the road climbing over the hill and curving around the top of the tunnel portal. This road’s actually the Glenwood Highway, the first paved road across the Santa Cruz Mountains, and first state highway over the mountains. The Glenwood Highway, built between 1912 and 1921, split off from the current Highway 17 on the ridge between Glenwood and Laurel, dropped down into the Bean Creek canyon, then headed through Glenwood towards Scotts Valley. The concrete road, 15 to 17 feet wide, had banked curves and oiled shoulders. It was the height of modern highway design. When the Glenwood Highway was widened in 1939, the town wasn’t big enough for the highway and the railroad; the SP lost that battle, and the depot was torn down to encourage more space for cars. The current route of Highway 17 later won out, but the jazz-era Glenwood Highway still remains if you drive through Glenwood today.

Road during rework.

Road after rework.

Old photos show the key details of the Glenwood Highway: precise curves and straightaways, the odd slalom around the top of the tunnel portal, and an even descent. I followed an approach I’d used elsewhere. I’d started by roughing out scenery to match the rough slope I wanted, and tore out as much of the old road as I could. As I’ve done elsewhere, I used 1/16” styrene sheet for the roadway, scribed with expansion lines. I cut the styrene at the workbench so I made sure curves were accurate and straightaways were smooth. I glued the sheet to the scenery with Liquid Nails contact cement, and used weights and straight lumber to keep the road flat until it dried. Once the road was glued in place, I used Sculptamold and spackle to finish the fills and shoulders.

Time for detailing the scene.

SP 84 heading out of Glenwood tunnel towards Santa Cruz.

SP 31 coming out of tunnel.

This photo of train 84 coming out of the tunnel shows that the bottom of the canyon had a bunch of pine trees in the 1920's. I'd always intended to capture scenes like this: the conifers in the canyon, oak trees higher up, and the privacy screen of trees between the highway and the railroad tracks. Rearranging the hillside and roadway helped this a bit. Filling in undergrowth and deciduous trees is easy; I've been using either Supertrees or Woodland Scenics Fine-Leaf Foilage. The redwoods and other conifers were more of a problem. I'd covered the hillside around Wrights with Woodland Scenics conifers, but I'd found these slow and tedious to build. I'd started trying to do the same at Glenwood, but eventually figured out that gluing tufts of sponge to the plastic armatures was not how I wanted to spend my hobby hours. Luckily, I'd gone to a model railroad train show right before COVID-19 struck. Grand Central Gems out of San Diego was there with their pre-made trees; I bought a few bags of tall pines, and loved how quickly I managed to get the scene finished. I bought a couple more bags later, quickly filling the hillside. I've always been cheap and unwilling to buy pre-made trees, but spending less than a locomotive to get this scene finished was worth it. With enough trees, it was also easy to give the look of separate areas of fields separated by tree lines. Static grass and a barbed wire fence made of wood posts and fishing line completed the scene.


In the earliest track plans for the Vasona Branch, I’d sketched in a location for the pit for the former South Pacific Coast turntable. Glenwood had been a key spot for narrow gauge lumber traffic; many short trains would carry lumber up to Glenwood; from here, a single engine could pull a longer train through the summit tunnel and down to the Santa Clara Valley. One of Bruce MacGregor’s South Pacific Coast books mentioned the filled-in turntable, so I added it to the track plan as an interesting bit of history. I’d cut a half-moon hole in the homesite for the turntable at a convenient location when I first laid track.

In the last episode, I mentioned finding the valuation maps and spotting the actual location. There were actually two turntables. The one at north end of town that apparently was filled in and tracked over during standard gauge times. The other turntable was at the far end of the siding, hanging over creek edge. Neither matched my guess when I'd first sketched the track plan.

I’d already cut the notch in the layout for the turntable in… oh, 2005, and wasn’t up to moving it. Stories claim the turntable was mostly filled in, but I took the existing hole, added stained balsa wood around to support it, added some debris on the bottom, and called it a day. It'll be a good location to throw whatever clutter I happen to have kicking around.

Glenwood Station:

I scratch built the Glenwood depot using some existing plans, and inference from photos. Gary Cavaglia published plans for the Glenwood depot in the March/April 2003 Narrow Gauge and Short Line Gazette. His drawing laid out the original narrow gauge depot building from the 1870's. That building contained a waiting room and tiny baggage area, and was completely sided in shiplap siding. All the photos from the times I model show a different structure. Apparently, the station was extended some time before standard gauging. The 1920's station extended the office portion of the station, built on a long addition for baggage and freight, and added a large raised freight platform matching SP's standard station designs. The station, however, kept many of the details seen in other SPC depots in Alviso, Agnews, Alma, and Wrights: similar doors, roof supports, and roof peak decorations.

Cavaglia’s drawings of the depot gave me the rough shape: walls were 12 feet tall; the gable peaked at 17 feet, and end walls were 14 feet wide. The drawings also laid out the rough sides of the passenger section. During the reconstruction, a door moved; I assumed the windows stayed in the same location. Using these measurements and various expectations (doors 30” wide, windows three feet off the ground) I could infer other measurements. The waiting room originally had two windows with a door between; later photos show a solid wall and a door to the right. Apparently, the reconstruction kept windows in the same locations but blocked up the original door location.

The later extension to the freight side of the depot was board-and-batten which made guessing at lengths easier - the battens appeared to be spaced 12” apart, so I could make guesses about the overall length of the extension. I sized the freight dock to the space available on the layout, rather than the size of the prototype’s.

SketchUp model of Glenwood Station

I sketched the whole model in SketchUp because I could do so quickly - I already had 3d models for the SP-style windows, so putting together the rough shape was fast. Once I had a rough model, I could compare it to photos and confirm that it looked about right. I could have done the same with pencil sketches or with a cardboard model.

I built the model using sheet styrene, window and door castings from my hobby stash, and vacuum-formed shingle material from Plastruct. I was lucky that all the supplies were already in my hobby stash, for I started the model just as the Covid-19 shelter-in-place started here in Silicon Valley. Like many of the SP stations I built, I used the Grandt Line 5031 windows (12 pane double-hung windows) to match the main windows, and the narrower Grandt Line 5029 windows for the sides of the operator’s bay. (I use those windows a lot, so my box-of-windows-for-projects usually has some on-hand.) I scratchbuilt the freight doors from styrene sheet. It doesn’t take a lot of styrene to be able to knock off one of these models; I usually keep a couple sheets of board-and-batten material, a couple sheets of shiplap, and then strip styrene in 1x4, 6x6, 2x6, and 2x8 dimensions, and that’s all that’s needed for most buildings. I also keep large sheets of 1/16” sheet styrene from TAP Plastics because it’s cheap and useful for bases or backing support. I also had some very beefy .156 x .250 sticks of styrene; these turned out to be really handy for building up a base for the loading dock. I could have done the same with Plexiglas, but that would have required shopping, and also required using power tools in the garage. Building from styrene let me build quickly with just a #11 X-acto knife, a square, and a straightedge.

Overall, scratch building a model like this is quick - probably a week of evenings including design and painting. The worst part is cutting out the window openings. If you haven’t tried scratch building, find some simple building, get $25 in plastic from your favorite hobby store, and start cutting!

Maintenance of Way Buildings

Who was around in Glenwood? Even with the large station, Glenwood never attracted the business one would expect; it didn’t become a wine center, didn’t have a major lumber industry, never attracted farmers. The August 1916 Southern Pacific payroll on showed Campbell station had an agent, warehouseman, and clerk, and apparently had a part time “fruit checker”. Wrights had an agent and warehouseman in their little hamlet. Meanwhile, Glenwood’s large station only had Alfred Feldt, operator, making $80 a month. Feldt eventually moved to San Lucas; in 1920, Edom N. Davis had the agent role. The abandonment proceedings in 1939 declared that only 196 people lived in the Glenwood area.

The section gang was a big chunk of that population during the teens and twenties. 1916 payroll records show six laborers and a foreman in Glenwood, reminding us that the section housing and work sheds deserve to be prominent. The August 1916 records show similar section crews at Campbell, Los Gatos, Wright, and Santa Cruz. The 1916 crew included V. Simoni as foreman, P. Simoni as watchman (perhaps for the tunnel), and G. Simi, M. Mariani, J. Jilla, A. Scarponei, U. Balleroni, and G. Berlacgua on the crew. The 1920 census showed a similar crowd: Benjamin Capp, Vigellio Elli, Toni Gianti, G. Luciano, Joseph Menta, Sam Chientilli, and Angelo Fideli. The railroad apparently was a good gig for the new immigrants. A separate continent of Mexican workers listed their occupation as wood choppers in the same census pages. Valuation maps don't show housing for the workers, but a few 1920's photos show what appear to be bunk cars on the siding next to the tunnel. Twenty years later, a Vernon Sappers photo of the Felton depot in 1935 shows maintenance of way bunk cars on the siding behind the Felton station, suggesting the maintenance of way workers moved closer to the bright lights of civilization.

I'd hoped on hinting at the folks who worked in Glenwood. I've got some bunk car models (care of Jason Hill of Owl Mountain Models fame), but I hoped I could include the section house where the foreman lived. Unfortunately, I'd started figuring out a location too late - I already had the station and an Atlas water tank in their rough locations. Unfortunately, I couldn't figure out an arrangement that didn't appear too crowded. Instead, I added a tool house built from a A&LW Lines laser-cut kit, and an outhouse next to a privacy fence.

Scenery improved? Check. Unrealistic buildings removed? Check. Prototype station in place? Check. Tons of trees? Check. Glenwood was always a place I modeled because I wanted to capture the real look - the curve into the tunnel, the redwoods, and the interaction of the new highway and the old railroad. It’s now got that look I intended.