Miscellaneous Notes

• Abstract Design
• Module Construction
• Paint
• Hand Laid Track
• Siding Construction
• Sunshine
• Track Mismatch

Engineering and software coding feature heavily in my work background; so producing track design software was an early and natural activity for me. Non-technical people may find it frightening, but things such as the abstraction of using three two-dimensional vectors to model a module, together with appropriate vector arithmetic, etc., has been highly successful. Apart from the track configuration aspect, the mathematical and computer models have provided a cohesive foundation for everything else, including module structural frame design, full-size paper templates for module manufacture, and module hardware size verification.

Concepts that come out of the abstraction are not always conventional. For example, curves are defined by their number of modules per circle and their arc length, and not by their radius; the latter being an arithmetic result of the former two. Thus, in the Baseline-Plus track diagram, the main curves are 16mpc, i.e. sixteen modules per circle, so they each have a turn of $22\frac{1}{2}$ degrees. And the module arc length is specified to minimize the expense of rail/track wastage. Track and rail is sold in the USA in lengths of 72 inches (six feet, [1.83m]). In the portable railway components, the arc length of double track curved modules is set to 69 inches; resulting in the longest rail on a double track, 16mpc, gauge 3, module being 71.45 inches. The tracks centreline spacing is 10 inches [254mm]; and the outer track, module, and inner track radii are 15.06[4.59], 14.64[4.46], and 14.23[4.34] feet[m]

Part of the "fun thing to do" of the mathematical and computer modelling was to implement transition curves in the form of Euler spirals. The Euler spiral is the modern transition of choice for roller coasters, roads, and railways, and I wanted to understand it. The outcome is that, in the Baseline-Plus track diagram, the four modules adjacent to the two straights (labelled 301-304) are transitions to and from constant curvature modules.

I decided to to use a ladder frame with a single stressed skin for my modules. The sides, being key structural members, are aluminium bar. However, the ladder ribs are wood. Aluminium ribs would be better from some points of view, for example, if a module were destined to be installed outside in the weather. However, module manufacture would be more complex with aluminium, probably requiring welding or pop-rivets, etc., especially when leg interface sockets are brought into the picture. The stressed skin is, of course, the track deck. This, too, is wood, being Baltic Birch plywood. Although, if I had decided on aluminium ribs, probably I should have chosen some form of aluminium sandwich board for the deck. An all-aluminium module is very appealing for permanent outside use.

Modules are glued together using high-quality epoxy. Welding and mechanical fasteners were contenders; but, once it was decided to use wood in the construction, using glue appeared to be the simplest course. I solder, but do not weld, so, to weld, it would have been necessary to buy equipment and learn how to do it: too hard at present. Mechanical fasteners were judged to involve too much accurate drilling and possibly thread tapping, again, too hard.

A second skin on the bottom of the ribs would make a module very stiff. However, it seems that a single stressed skin is perfect for the application. The current modules are strong, light, and dimensionally stable; but, also, are able to twist slightly. This characteristic allows a little track super-elevation simply by using shorter legs on the inside of a curve.

A module assembly jig was made during prototype development. In use, a full-size paper template is installed in the jig. The jig has interface plates that can be moved to accomodate module size and interface plate alignments as shown on the template. Component pieces are prepared using the template; for example, aluminium sides need to be pre-curved if the module requires it. A module is glued together in the jig, and can be removed after a twenty-four hour glue cure.

The image shows a double-track, curved module, ladder frame being constructed. The four separate interface angles are for subsequent modules; they show the wood blocks used to glue the angles to the sides.

Then the ladder frame is glued to the plywood deck.

All the module aluminium was primed with etching primer, this is the green paint on the interface plates; the wood was primed with exterior water-based primer. The finish, on everything except the interface plates, is two coats of water-based porch paint.

As can be seen in various images in the Overview (here is one), the tracks on the railway mainline modules are hand laid with rail screwed to redwood sleepers (I have become really good at using a small screwdriver). The explanation for hand laid track is twofold: my possession of a significant quantity of rail, as mentioned as a design constraint; and, secondly, I do not know of commercially available dual G1/G3 track.

The railway siding modules have the rails screwed directly to the module plywood deck; this is really solid construction, and, by comparison, shows up the problems with the redwood very clearly. However, there is a significant difference in sounding-board drumming; the squishy trackbed on the mainline is well worth having. Shown in the image is a G0/G1 siding junction, using code 250 rail, with fixed point frogs and checkrails, but with moving points to handle the wheel drop problem. One of these points can be seen, opened, on the right of the image.

Erecting a railway outdoors is subject to a greater number of external influences than erecting indoors; one of these influences is the sun. The particular heating effect that has emerged for the portable railway is caused by differential expansion in the modules. As constructed throughout the railway, modules have a wood deck and metal sides; the deck is glued to the top of the sides. As the temperature moves away from [the construction] room temperature a module curves since the metal expands and contracts more than the wood. The module deck becomes concave with heating or convex with cooling.

This "obvious" effect has been a surprising discovery. During initial erection, the heating by the sun, and cooling from a cool breeze, can affect overall railway levelling since two, very carefully installed, levelling bugs on one module will now disagree. After erection, heating by the sun causes a scalloping effect as each module curves so that its ends are higher than its middle. A railway can have different gradients at one end of the day from the other; which steam engines are good at finding.

A potential long-term problem is damage to the epoxy glue that holds the deck to the sides. As the deck flexes between convex to concave the glue must handle stresses not accounted for by the module design. It remains to be seen if these stresses cause degradation and failure of the joints.

These effects of the sun and wind are there, and they can extend outdoor setup time and introduce mild gradients into an established railway. However, I do not want to emphasize them; the overall effect is frustration, but, other than potential stress-caused failure, is a small problem.

Before committing myself to making the portable railway, I made three modules as an experiment, simply to see how manufacturing would work out. In early 2021 I gave these three modules an actual use in the railway; they appear as P1, P2, and P3 in the Exhibition Oval Configuration U.

P2 was designed and constructed as a composite: a transition from straight to left turning 8mpc curve for a 35.0 inch arc followed by a constant curve 8mpc for a 35.0 inch arc; the overall module turn being 33.75 degrees. In re-purposing, this module was equiped with a single transition track. However, the hardware overconstrained the new design, and approximations were necessary; in particular, the straight end of the transition was truncated. Also it appears that there may have been some careless rail bending at the curve end during hardware construction. But the net effect was that neither end mated nicely with its adjacent module; the rails did not provide smooth transition from module to module; there was a kink in the track at each module join. The kinks can be seen in the images below by examining the rail contours. It appeared that the kinks could be removed by subtle, varying, transverse movement of the track. The upper images show the kinks; the lower images show the results of the reworking. The curved end rework result is better than that at the straight end; probably the weakness of the latter is because of the truncation of the transition mentioned above.

The reason for recording this rework is to describe the method, which had been surmised but not tried previously; it seems to have worked well. The key point is that each sleeper ultimately holding the track to the module, is attached to the module before the rails are attached to the sleeper.

1. Anchor the track ends on the module to be modified. In this reported case the module was attached to its mating modules as if it were being used.
2. Mark the positions of all sleepers likely to be affected; blue tape works well. This marking is for reference during track position changes.
3. Slacken all rail-to-sleeper attachment screws likely to be affected; this allows rail movement along the rail.
4. Remove the track-to-module attachment screws.
5. Make the desired position changes and tighten all the rail-to-sleeper attachment screws except on the sleepers that host the track-to-module attachment screws. Use a track gauge as required.
6. At the track-to-module attachment point, remove the rail-to-sleeper attachment screws and the old sleeper, then insert and accurately position a new sleeper. Drill a clearance hole, probably using a #51 drill, in the new sleeper from underneath using the attachment thread in the module as a position guide. Open the clearance hole as required, probably using a #44 drill.
7. Attach the new sleeper to the module with a track-to-module screw.
8. Attach the rails to the new sleeper with rail-to-sleeper screws. Use a track gauge as required.