Wednesday, October 11, 2017

Market Street: Construction

A couple months ago, I needed a project that I could really focus on, and preferably one that required hammering lots of 2x4s together. In a feat of poor judgement, I just started building. The result was this Market Street Station layout, a modular layout in the british exhibition layout style.

This new Market Street Station layout is a big change from what I’ve built in the past. It’s modular, so I need to worry about how the pieces connect up. It’s intended for longer passenger cars, so I need to use different standards for curve radius and switches from my garage layout. It needs to be portable, so I need to be able to move the pieces, set it up accurately, and break it down quickly.

From the last layout and planning for this one, I knew that a complex design would just mean I'd spend too much time planning and rethinking. Instead, I decided to get some minimal bench work up as fast as possible so that I could test out module width, track position, and composition of the scenes as fast as possible. With some quick building, I could figure out if I were on the right path.

First Problem: How to Build Fast. In order to get modules built quickly, I decided to try out hollow core doors. They’re available - one short trip to Home Depot got me most of my benchwork - and require minimal other carpentry. Doors often don't seem appropriate for a model railroad because building terrain lower than the tracks is difficult. That's not a problem with Market Street, where most of the area is completely flat. Using doors also cuts some of the work to build a self-supporting structure - they’re pretty stable on their own. By building on a flat surface, I also don’t have to plan every exact curve and track location (as I might with a cookie cutter plywood layout), and can instead start building and worry about exact track locations later.

Underside of door with leg pocket and wiring.

I used 28” doors (Home Depot, $25 each), framed them with 6” wide birch plywood (3/4” for the end plates, 1/2” for the sides). I added pockets to fit 2x2 legs inside, with the pockets glued and nailed to the sides. I placed the layout at a nominal height of 50 inches. Cutting the plywood side pieces accurately was a challenge without a table saw, but I got close enough with a rip guide on my circular saw. The first two modules took two and a half days to build, with most of the time taken up trying to figure out how to build stuff. I’m relying on glue joints a bit too much, but I’m thinking it’ll work. I also learned a couple tricks on the first doors. The door I got was already primed; I had to use a plane on the edges to get down to raw wood for gluing. I also expected that everything would be suitably square if I built it on our patio concrete, but found I still needed to check squareness constantly and adjust. I used 1x2 scraps to reinforce corners. The final modules weighed about 30 pounds each.

Second challenge: the track arrangement. Once I had the modules built, I started laying out track just to get an idea of what would fit.

Experimenting with track arrangements.

As with my Vasona Branch layout, my first goal was to define my average train length so that I could ensure most trains would fit on the sidings, passing tracks, and staging tracks. I ended up choosing four car trains with three coaches and a baggage car or RPO. Although I’m expecting to use shorter Harriman cars for most of my trains, I sized the spurs for eighty foot cars. I've collected modern commute cars over the years; running those Gallery and Baby Bullet trains will help folks understand that the Market Street station is part of San Francisco-San Jose commute history.

With the train length set, I could start placing track on the doors and figuring out track arrangements. I quickly found that 29 inches was an awfully wide space. Even with the Freemo suggestions to keep track six inches from the edge, there was plenty of room for tracks and for surrounding structures and scenery. Door length, however, was an issue. The trackage on the module to the left of Market Street station needed crossovers to allow trains to and from the station to go to either track, the switch for the mainline bypass around the station, and switches at the end of the baggage sidings. This resulted in five switches placed end to end which had to fit in the 6’8” space of the single door.

Track plan near 4th Street Tower

Another essential bit of track was the area east of Market Street station where the tracks diverged to head towards Oakland and Los Angeles. Modeling the Oakland track wasn’t a big priority for me - there weren’t a lot of trains going that way. However, the split was memorable because it was the location of the Fourth Street interlocking tower (controlling track on the east end of the station), the venerable Golden Gate cannery sitting between the tracks (later run by Hunts and Richmond Chase), and the unusual sharp curve of the eastbound track towards Los Angeles.

The curve of the Los Angeles track was a particular bit of reality I wanted to capture. The tracks heading east from Market Street slowly came together and curved towards the south. At Third Street, just as the tracks approached the Golden Gate cannery, the tracks curved sharply, cutting through several backyards until popping out onto Fourth Street and heading defiantly towards the heart of downtown San Jose. Those tracks ran down the middle of Fourth Street, passed homes, offices, and San Jose State University, and finally hit the cannery district south of downtown where Fourth Street ended but the tracks continued. Those tracks are interesting, not only because of the specter of freight trains running past San Jose State, but also because of the political fight between San Jose and the Willow Glen neighborhood about how to get the SP off the city streets.

Because this section of layout couldn't be rectangular, I built it using more traditional construction: plywood roadbed surface and cookie-cutter framing underneath. I found the construction more challenging, time-consuming and tedious, and I ended up getting the curves wrong so that the track to Oakland didn't actually end up in the middle of the end plate. It was a heck of a lot of work just to keep a favorite scene.

Third Challenge: Laying Track The cheap doors use masonite or hardboard for the door surface, so spikes and nails can't be used to hold track down. I decided to try gluing down track using acrylic caulk. (M.C. Fujiwara's video on laying track on Freemo-N modules convinced me to try this out.) I used Walthers code 83 track and #6 switches - matching the Freemo track height, but using sharper switches than Freemo allows on mainline modules. The caulk approach went down quickly, but I’ve already had a couple places come loose, so I’m a bit concerned that it won’t handle abuse as well as spiked track. I did drill holes to put spikes into the masonite door surfaces to hold switches in position; we’ll see how those hold. Next time, I may get doors faced with lauan plywood, or glue thin plywood onto the doors.

Keeping track on the module in place is one challenge, but modules also need track to align at joints accurately. For joining the three key modules, I used steel pegs to ensure all track joints would be in alignment. I turned these pins by hand on a lathe from 3/8" steel rod, making a male and female pin. I clamped the modules together in alignment, drilled a single 3/8" hole, and epoxied both pins in place.)

For the sections leading to staging, I used the Freemo approach of ending tracks 2 inches before the end of the module, and then using bridge rails to cross the joint. One advantage of the bridge rail approach is that there isn’t a rail at the very end of the module that can snag when moving the modules. I've already snagged a rail end once while moving the modules.

But Is It Freemo-Compatible? I'd intended to make the modules fit Freemo specifications, but early-on was forced to break away from the Freemo specs. The modules don’t have the required 26” end plate with tracks centered, though that’s not a big deal; I could always create transition modules to shrink the end plate to the correct size and location.

The bigger problems were curve radiuses and switches. Freemo mainline modules require 42 inch curves and #8 turnouts on the mainline.U I couldn't make either requirement work. The broad curves would have narrowed the Golden Gate cannery scene too much. The #8 turnouts just wouldn't let the track plan fit on the two doors. My specific problem was on the door representing the area west of the station. That track required five consecutive switches - the switch for the baggage tracks, crossover from east to west main, switch to the bypass around the station, crossover from west to east, and far switch for passenger rip track. Making any of these switches #8 would keep the track from fitting on a single door and would stretch the track over a second module. Because I’m also not aiming for modern era, I wasn't sure my modules would be particularly interesting in a standard Freemo setup, and decided to give up on Freemo track standards.

Wiring I did wire the modules in the Freemo style, with an accessory and DCC bus both joined between modules with PowerPole connectors. I didn't bother to wire for signals on the main - the whole scene, in the 1920's, would have been unsignaled (west side), or controlled by the interlocking in the Fourth Street Tower (east side).

I did power all the switches, as I do on my garage layout - it avoids damage to the switches, allows me to power frogs, and helps suspend disbelief by keeping operators' hands out of the scene. Instead of using venerable Tortoises, I tried using the MP-5 switch machines. These new switch machines, imported from the Czech Republic by Richard Brennen and sold by Seth Neumann's Model Railroad Control Systems, are a great improvement. They're much tinier than tortoises, easier to adjust, and so low profile that they hide inside the door. To mount the switch machines, I cut a hole in the underside of the door, then glued a piece of 1.5 inch 1/2 inch plywood to the skin just below the switch. I found that the supplied stiff piano wire didn't work well with the short distance to the switch machine; the 6mm throw pushed the switch points too hard, and occasionally the switch machine wouldn't move far enough over to hit the limit switch to allow throwing the other way. Fixing this required pushing the mechanism a bit further to get it to a s point where it could throw again. One particularly nice feature for a Freemo layout is that these switch machines don't need to be continuously powered. Instead, I can have momentary throw switches on both sides of the layout so operators can work from any side.

So that’s construction. I’ve set up and torn down the layout several times successfully; my biggest challenge is that the combined length - 25’ - is longer than my garage. I’ve set up parts of the layout inside when I’ve been doing track work or installing switch machines. Testing the whole layout, though, requires setting it up in the back yard. Luckily, California weather means I can set the modules up most weekends. I’ve got little chance of rain or humidity. For a weekend of work, I’ve been able to set up the layout on Saturday and keep the benchwork outside through Sunday or Monday. The rest of the time, the modules lean up against the wall of the garage. That’s not a good long-term solution (especially as I do more scenery), but it’ll work for now. Considering the layout is intended for operations and transportation, minimizing the amount of detail on the layout might be a reasonable decision. I’ve got some mock-ups of buildings in foam core which give the impression of rolling through the city.

Coming Up: Electronics, Software, and Operations.

No comments:

Post a Comment