HO-Scale Model Layout Test Track #4


I have a small "layout" at home, which is actually just a glorified test track.

This describes how I designed and built the fourth iteration of it.

I have a page detailing the construction of Layout #2, and this one focuses on the latest version, “Layout #4”.


As in many projects, success starts with identifying the scope and goals of the project, and often most importantly what are the non-goals. In this case the main non-goal is that I am not building a “real” model train layout. What I really need is a test track to fine tune engines and a “lab” area to experiment with projects which I often apply later to the Randall Museum Model Railroad. As such, what I need is:

  • At least one loop to tune engines.
  • A few garage tracks to store a few engines & cars.
  • No scenery. I am not even pretending this to be a “real” model train layout.
  • Experiment with a modular construction.
  • Experiment with track changing over time.
  • A fairly limited footprint.

A lot of my model train activity focuses on maintaining the Randall Museum Model Railroad, as can be seen in the blog. As such, pretty much every new electronics module or thing that I prepare or design for the museum is tried at home on the test track first. Thus the “layout” focus is not really on the layout itself; it is an experimentation bed.

Design History & Evolution

Layout #2 was mostly based on the “4x8” principle, even though technically it ended up as a 4x9.5 ft size footprint. Two double loops, a little bridge, etc.

What I learned from that:

  • The “4x8” form factor was limiting. Four feet is way too large. It takes too much space if placed in the middle of the room, and I cannot reach from one side to another if placed against a corner.
  • I tried to mitigate that by using an access hole in the middle, but that means I could not use the place underneath for storage.
  • The layout was too low for using when standing up, and I did not like to sit down to operate it as I constantly needed to get up then sit down to do anything.

The bottom line is that I realized I wanted a layout at the same height as my standing workbench, and with a form factor where I could be “in the middle” and reach everywhere.

Eventually Layout #2 was replaced by a Layout #3 which was more or less a L-shaped test track on a workbench structure. It was interesting due to its limited corner footprint, but lacked that essential running loop, which I often need to fine tune engines.

Then in 2018, I saw this really amazing module in the online MRH magazine forum:


The CSX Ridgecrest Module.

This module was first presented in 2017 in Model Railroader: http://mrr.trains.com/issues/2017/build-a-small-model-railroad

The original plan can be found here: MRR HO scale CSX Ridgecrest.pdf

I really liked that module. I thus decided my design goal would be to have such a module on one side of a layout, then a storage yard on the other side, and connect the two using two track loops, like this:

The yellow track module represents the CSX Ridgecrest track plan, and the blue track module is a storage track.

One idea here was that the right side of the loop was going to be a movable bridge, and the structure would thus form an horizontal U (e.g. ⊃ ).

I made a lot of iterations on the layout, and I mean a lot. For example at some point I had this design with the two modules with loop track (that somehow changed from double to triple track), but that was a hefty 4x10 ft footprint:

And eventually that got simplified to this, where the blue part at the bottom was going to be an entire curved removal bridge (don’t ask how that even exactly works… where does it go?):

Scenery wise, the design goal at that point was something like this:

Note that at that point I was doing a detailed design of the switching module to decide which turnout and track brand to select… but wait I have all that EZ-Track sitting in a box, so why not use that to do a mock test version? That would require simplifying the switching module a lot since the EZ-Track turnouts are way space inefficient…

However I quickly realized I did not have the space for it, and honestly the movable bridge seemed like a hassle. The CSX Ridgecrest module is 2x6 ft but I don’t have that kind of space. I’m allocating 3x8 ft, and the loops take 2 feet on each side.

Thus I decided to use that typical design adage -- keep it simple:

  • Drop the double track loop and use a single one.
  • Drop the storage yard.
  • Reduce the yellow track module to a minimum (2 turnouts and 2 tracks).
  • Compress the oval loop into a “dogbone” shape.

The end result is very different from the starting point. The result is what I call not the “dogbone” shape but the “ski mask” shape:

The track is mostly built using EZ-Track straight segments and 18” radius curves. There’s a bit of cheating where in a couple places I drop it to 15” radius curves. The inner turnouts are the EZ-Track #6 turnouts, and the outer ones are the #4 turnouts.

Note that the tight radius means I won’t run long engines on this, or at least not with cars. For example the Rapido RDC SP #10 can run, or an SD70 engine can run, but their couplers overhang prevent any car from being pulled by these. That’s not really an issue as the only times I expect to run these engines they would run alone, for demonstration or for testing purposes.

I do have small base engines and old era short cars -- my usual D&RGW line up, and these can run fine on these tight radius without any issues.

I originally was going to scenic this, and I wanted to have a little bridge, a little river, and add buildings like on the CSX Ridgecrest module. I however quickly dropped these requirements as honestly these are not so useful for my use at home. I prefer the flexible aspect of the EZ-Track and I have already altered the track in minor ways.

The final “schematized” track plan is the following:


My previous layout was built using a rigid plywood board on top of a rigid frame surface. Here I wanted something different -- as I pointed out, a whole part of this is experimenting.

One experiment was to build the layout using a modular frame with foam on top instead of plywood. For storage reasons I also don’t want to have any feet. The modules should hold using a cantilever design or some kind of wall brackets.

Going back to the original switching module, the frame was going to work like this:

I made two boxes of 2x4 ft size out of wood:

  • The side and inner supports of the boxes is 3” x ¾ boards, material is #2 common whitewood from Lowe’s.
  • L corner braces are used to hold them together in a square frame open frame.
  • L metal brackets are used to support the frames at the desired high on the wall with some plywood in between to provide better load support.
  • Covered with 2x2 ft pink Foamular 150 boards.

For the surface, the material is “Closed Cell Extruded Polystyrene” or XPS. EPS may or may not be the same thing (EPS = Expanded, XPS = Extruded… same difference?). The typical choices are Foamular 150 vs 250 where the number represents the compressive strength (15 psi vs 25 psi at 10% deflection according to the Owens Corning documents. In the colder Eastern US parts they can get the stuff in 4x8 ft sheets of 2 inch thickness. However in our warner CA climate, the only available are 2x2 ft sheets of 1 inch thickness for the Foamular 150.

It is generally recommended to paint these “pink foam” boards on all sides with a latex paint to avoid moisture absorption (not really a problem here). Oh by the way that stuff is highly flammable and releases toxic chlorine gases when burning. Just saying… something to be aware of.

The two boxes are lightly secured on the L wall brackets. The original design was to apply scenery and track and then cut the track at the end, allowing the modules to be removed independently. But then after I was done with the modules and had them in place, I decided to drop the scenery and go for EZ-Track, which cannot be cut. OTOH that one is trivial to remove / disassemble and thus requires no cutting, nor gluing.

After I dropped the idea of a removable “front bridge” or whatever it would be called, and instead opted for a “dog bone” (or rather “ski mask”) design, I added one more square module on the left to hold the loop. For that 3rd module, I used 1x4 furring string boards [Lowe’s link] instead of the same wood as before as I wanted to compare the characteristics -- I had never used “furring string boards” before. This was interesting: Fairly light / low density; longitudinally, the boards are fairly straight; transversely, the boards are slightly curved; one side is smooth, one side is very very rough, e.g. unfinished, unsanded. It was however easy to work with and I ended up making the module not entirely square -- the outer lower right section is cut in a slanted manner to avoid having a protruding edge in the middle of the room.

No L-brackets support that 3rd module. It has an actual simple 2x2 leg on one side, and is screwed to the workbench on the other side.

I also decided to create two small bridges for the inner loops, which are the parts in yellow below:

What’s Next (and What’s Not Next)

Ideally, what I could or should do next, but won’t actually do:

  • Paint the module’s wood (in and out, to prevent the wood from warping).
  • Paint the pink foam (to hide the obvious pinkiness).
  • Add a foam core base.
  • Add the final track (e.g. replace the EZ-Track by Atlas / ME / Peco, etc).
  • Add a foundation of dirt/grass.
  • Ballast the track.
  • Add temporary paper structures.
  • Add final structures.

To be honest I’m not saying this will “never happen”. That’s not really how these things work. It depends on what interests me at a given point, and right now I have a ton of other ideas to explore, and having an “unfinished” track makes it so much easier to alter and change it. Because there’s no scenery, no ballast, no glued track, everything can be modified in a pinch.

Instead what I have done:

  • Power the 4 turnouts by an NCE Q-Snap. Add support for that in my Android DCC control software.
  • Connect the NCE Command Station to the JMRI computer, to have full home control for trains running, trains programming, and turnout control.
  • Try out the NCE EB-1 Circuit Breaker. That’s a bit overkill at home, but I use these at Randall and it’s convenient to have a way to replicate testing at home.
  • Experiment modifying DCC turnouts with LEDs to directly create direction indicators. I applied that at home before reusing that technique at Randall.
  • Temporarily modify one of the loop curves with a surelevation to replicate one derailing condition from Randall.

The next project is to modify an EZ-Track segment to work as a double switch with a LED indicator, so that I can isolate one siding as a programming track and make sure the rest of the layout is not connected to the programming track.


[Back to main page]