A Portable Railway Overview

Producing this railway has been a mixture of a mechanical engineering project and a minor personal adventure, executed in the context of an ordinary person's life progression. In retrospect, it is clear that my years' investment in toy trains weighed on me between leaving my house and garden railway in June 2013, and re-settling into a condominium five months later.

I wrote some objectives early on, in September 2013. Here they are; the first five were important to me; the latter two were nice-to-have preferences. The comments are current, briefly describing interpretation and implementation of each objective, and its outcome.

• Manageable by One Person. This means that the railway must be an assembly of pieces that are not too heavy, not cumbersome, etc.
  The railway components are modules, and each module is nominally six feet long, nineteen inches wide, and weighs less than twenty pounds. Most people, even old geezers, can handle modules like that; although it must be noted that, during setup and takedown of a railway, a person handling these modules can discover personal klutziness very easily. A set of modules, forming a railway, can be erected by one person; but that is hard work, two people is much faster and easier.

• Permit Multiple Railway Designs. Railways of different sizes and shapes must be accomodated. This means following some of the ideas implemented in toy train sets, such as standard straight and curved sections that can be assembled in arbitrary order. Also, it means that it must be fairly straightforward, relatively fast, and relatively easy, to make a couple of new modules to accomodate a new railway configuration. A corollary of this objective is that expansion of a small railway must be simple.
  Think H0 train sets. The modules are numerically coded by shape, and all modules of the same shape are interchangable. There are not enough modules existing to enable much railway reconfiguration, but arbitrary assembly order is a simple given. At this point, making a batch of new modules is a feasible, non-dread-inducing, task; although it is not fast and easy. But storing and transporting the existing modules is a difficult issue; and that difficulty is inhibiting further construction and expansion.

• Support Experimental Projects. This means support for both gauge 1 and gauge 3 at a minimum; also it means a large minimum curve radius, to support long-wheelbase rolling stock.
  It was very natural to follow the common inside and outside track configuration seen on a lot of steamup portable railways. And, by making each track dual gauge, gauges 0, 1 and 3 are supported. Gauge 1 exists on both tracks since it is the most popular gauge. The nominal radius of the mainline curves is fifteen feet. The minimum radius is eight feet, at the turn-in to the G0/G1 inner track steamup siding. The minimum radius on the G1/G3 outer track is eleven feet, on the steamup siding curves.

• Glitch Free. The assembled railway must be smooth running. Derailments must be very infrequent and due to operator error. Any junctions must operate well and be flagged to show point position.
  The tracks are smooth running, and derailments due to the track do not happen. Making the dual gauge junctions work well required a lot of careful design and manufacture. Optional flags indicate, at a distance, which track is selected.

• Robust. Damage is the issue addressed here. And damage usually occurs during packing and unpacking, to and from storage, rather than during use. The railway must be easy to store and be self-protecting with built-in stand-offs, or something functionally similar.
  Robustness quickly segued into addressing storage and transportation. Two storage racks have been made; and my 1987 Chevy Astro van holds these racks and the other stuff which make up the whole main oval, without steamup sidings.

• Small First. This is rather obvious: the first railway should be as small as possible without compromising any later desire to expand.
  The first railway configuration was a simple oval that I named Baseline; it was never made. I decided instead to make Baseline-Plus, which is the same oval with four junctions intended to support steamup sidings, The sidings themselves, however, were not in the construction plan - the storage and transportation bogey raising its ugly head again.

  The module numbering on the module map ties-up with the module numbering in the storage racks shown above.

• Quiet I want the railway not to make noise by vibration of the baseboard, usually manifesting itself as a hollow "empty box" kind of noise. The railway modules should not be drums. This is not about rail hiss, nor clickity-click, nor noises generated by the locomotives.
  The mainline is laid on a squishy roadbed which works well to deaden sounding-board noises.

When I first started to look into the design of a portable railway, and considering that I had constraints on workshop space at the condominium, I did not expect actually to make anything. My view was "If I were to make one, this is how I would do it." In other words, I thought that paperwork would be the final product, and I did the design solely for personal satisfaction.

My design approach includes significant theoretical components and mathematical modelling. Thus, computer code is used for design, and this code contains, for example, implementation of two-dimensional vectors to describe module and track geometry, and numerical integration to produce Euler spiral transition curves. This may be overkill for toy trains, but it was a lot of fun for an engineer. However, once I had decided to go ahead and make the thing, there were real payoffs from the paper study (engineers would say "Of course there were."). One example is the very real assistance in construction provided by full-size, accurate, paper templates for module hardware construction. Another example is the very satisfying smooth transitions of trains running between straight and curved track; just like the real thing.

The design approach is quite abstract inasmuch as it separates module features to the extent possible, with the objective of maximizing manufacturing and usage flexibility. So, a key design characteristic is emphasis on various independent interfaces. Concentrating on an interface, and not what uses the interface, allows subsequent useful changes depending on circumstances. For example, each double-track, nineteen inch wide, module can be connected to another similar module; or it can be connected to two single-track, nine inch wide, modules. There are two key interfaces involved here; one specifies how two single track modules can be joined; the other specifies part of the railway track configuration, in this case two dual gauge tracks at ten inch centres. In the Baseline-Plus configuration, this connection flexibility allows the siding junctions, and hence the sidings, to be placed in different positions on the main oval, yielding, minor but useful, variations of the configuration.

As a second interface example: at some point in the paper study, I stopped trying to incorporate legs into module design; instead I created the module-to-leg interface. Using this interface specification, all that the modules have is a few sockets for legs to be plugged into; the legs can be whatever suits the railway environment. The legs I have made to date are telescopic and give a module deck height range of about two to four feet. The decision to make these adjustable legs came about partly in order to accomodate outside railway use on undulating grass. But, if the modules had been intended only for erection on a level concrete floor, then fixed-length broom handles would be perfectly acceptable supports. Similarly, permanent installation in a garden would offer other support possibilities; but the interface in the modules themselves would not change.

Another consideration affecting design was existing hardware. I had a lot of code 215 and code 250 Llagas Creek rail salvaged from my garden railway. I had perfectly functional junctions, also from the garden railway, that I could not bring myself to rip apart. I was given a large amount of code 250 brass rail, probably Sunset Valley, from a dismantled indoor railway. I had G1 (not dual gauge) track and junctions, both owned and donated, with different specifications. In the new railway design, I wanted to enable use of all of this hardware to the extent possible. This consideration led to a track interface specification; this interface is a helpful starting point, although details, such as differing rail sections at module interfaces, clearly need joint-specific planning. The railway currently uses all three of the rail types mentioned above; and one module hosts a salvaged junction.

In what follows, it is worth bearing in mind the kind of things described above, and which I had in mind when I created the railway design. In summary, I wanted design flexibility with the objective of allowing desired features, together with configuration expansion potential; think H0 train sets. For the most part, I believe that I have achieved these things; what exists is successful. Actual further expansion, however, is stymied by that pesky storage and transportation bogey.

Here is an early image, no track yet, which shows half-a-dozen modules connected together in an arbitrary way, supported by the telescoping, adjustable, legs, plugged into the sockets in the bottoms of the modules.

At the bottom left of the image can be seen a single track interface; just in view, hanging down, is a purchased over-centre latch, which is the mechanism for holding together connected modules. The connecting module will have a "keeper" (manufacturer's term), which the latch hooks onto. The deck is 6mm Baltic Birch plywood, also known as furniture plywood, painted with latex porch paint. The green bugs are spirit levels used during railway erection.

The spirit levels are turning out to be quite successful in guiding railway erection; although I note that they do require both accurate installation and accurate use. With regard to installation, a key requirement, that escaped me for a long time, is that the levels should monitor the module interfaces, not the rails. The rails follow the module decks very closely, of course; but maintaining the module levelness as defined by the interfaces is what results in both good vertical closure of a continuous railway at erection time, and minimal dips and rises along the tracks of any railway.

In the image below are two modules latched together, lying upside down; one module is single track, the other double track. Also to be seen, in the green painted surface, are the pin and pin-locating hole for the second track of the double-track module. The locating hole is the middle hole, the other two are service holes that are not significant for the current description. The holes on the brown tops of the interfaces are for, absent, latches and keepers. As a general rule, latches are installed only where needed, since latches are a little expensive. It is true that, as more railway configurations become possible, the need increases for sometimes-unused latches.

Other features of module construction can be seen in the image: the module interfaces are implemented in aluminium angle, which is epoxy-glued to aluminium side bars, both being glued to the Baltic Birch deck. The blocks seen under the single track module aluminium angle provide gluing area to ensure a good joint. These blocks are wood (oak); and this choice was made because the deck also is wood. I considered an all-metal, welded, construction, which, for example, would be desirable for long-term outside use. But I elected not to do it that way because it seemed unnecessary for the target, temporary use, railway. Unnecessary in this construction context means that I did not need to buy an arc-welder and learn how use it. However, all-metal construction is not restrained by any railway design consideration, and could be used.

The views below show how the interface is used on single and double track modules. The top of the rails is always two inches above the horizontal centreline of the locating pins. Whatever comes between the pins and the rail tops, here being aluminium angle, plywood deck, rubber track-base, redwood sleepers, and rail section, must honour the two inch dimension. There is no specification for joining different rail sections; normally, that is handled within the module itself, and it is not difficult to do that.

As indicated earlier, legs can be of any design that supports the railway modules. To date, only one style of leg has been made; these legs are made from thin wall aluminium telescopic tubing. Two of the legs are shown in the image in their current configuration. The first leg design, featuring collet clamps and plain rubber feet, is shown in the image in the Module description.

The collet clamps are the items mentioned in the introduction as a problem requiring rework. These nylon clamps did not live up to their manufacturer's claim, they slipped under an axial load of 20-25 lbf, not the 70 lbf claimed. As may be obvious, 20 lbf is not adequate; I think that the clamps crept under ordinary use as well. Finding a replacement clamp involved a quite arduous search, but ultimately was successful. Installation of the new clamps on the sixty existing legs involved rather a lot of re-work. In addition, a new batch of sixty legs has been made using the new clamps from the outset. The new clamps grip the tubing with an adjustable over-centre latch, and they function well inasmuch as they do not slip. However, the new clamps are not as convenient as the collet clamps: in use, they are somewhat fiddly to position and tighten.

For railway erection, a new module is latched to the existing module assembly, then is held level, as indicated by the built-in green bugs; legs are plugged into the sockets in the bottom of the module, then adjusted for length, and the clamp tightened to set the lengths. This is where having two people is highly desirable: one holds the module rigidly and level, the other, usually the author on his knees, does the adjustment and clamping. The clamps provide the coarse adjustment for leg length. Fairly early on, the plain rubber feet of the first design were replaced with screwed feet, which provide about two inches of fine adjustment for leg length. Provided that the module weight is taken off the leg, fine adjustment can be effected by rotating the leg in its socket.

At the top of each leg, unseen in the image, is a magnet. The sockets in the modules are fitted with steel washers, originally intended simply to spread the load in the wood module ribs. After leg insertion, the magnet-to-washer attraction works very well to hold the leg in its socket during module placement, so that the leg does not fall out of the socket. Probably it is not really necessary to state that the need for the magnets was an experimental discovery, not a brilliant conceptual design idea.

I like to refer to the leg length range as going from kid's viewing height, at just over two feet, to old geezer's running height, at nearly four feet. But otherwise, the adjustment capability is very convenient for sloping or otherwise undulating ground.

I consider these legs to be successful. They are a little manufacturing-time and money expensive; but you get what you pay for; they work well and are flexible in use. Their weaknesses are that they must not be thrown around and damaged, which would prevent telescoping; and they must be kept clean, it is bad for telescoping tubing to get gritty dust between the tubes.

This image shows a junction module interface, module ribs and other detail, and use of the leg interface. The track had not been fitted at this point.

If the erection site is indoors, a simple oval railway takes two people about two hours to setup; miscellaneous tasks, such as checking the overall integrity, tidying-up, perhaps track cleaning, and so on, adds another hour or so. Takedown takes an hour, and has been considerably less with several people. However, I am very wary of quick, that is when damage occurs. If the erection site is outdoors, the process can take a lot longer; much depends on the weather.

During erection, individual modules on a few legs are wobbly; single track modules on extended legs can be quite unstable. However, once a few modules are latched together they all support each other and the railway takes on considerable heft. The fully assembled Baseline-Plus oval has plenty of stability.

Closing an oval, in three dimensions, when the last module is inserted, is something that erectors of all portable railways have to deal with. I have been pleased, and relieved, with how well this goes with these railway components. I credit computer design, accurate manufacture, and careful module-to-module latching during erection. The two horizontal dimensions usually are within an inch or two for the final connection on a level floor. Careful use of the built-in levels gets the vertical connection under an inch as well. Erecting on lumpy grass makes the horizontal connection more difficult, and probably will cause vertical difficulties as the legs settle into the grass over time.

This image shows the Baseline-Plus configuration at the 2018 Denver Maker Faire, just after erection. Everything except the tables fits into the red and white van.

This image was taken at the 2018 Denver Hobby-Expo, in a different location to the Maker Faire. There is a lot to be seen:

• The world's largest sunglasses.
• A double-track G0/G1 & G1/G3, mainline curved module.
• A single-track G1/G3 mainline curved module.
• A single-track G0/G1 mainline junction.
• A single-track G0/G1 siding junction (ISJ).
• All three of the rail sections mentioned earlier are in this view.
• The presence and absence of squishy trackbed on the mainline and siding.
• The fearsome crowd control, comprising clip-on plastic stand-offs and orange tape.

I believe that it is important to describe negative outcomes of projects; there are always design aspects that are not welcome for one reason or another. My view is that the portable railway design and construction has worked-out well, and that the weaknesses are necessarily inherent rather than correctable faults. I have already indicated weak characteristics of the telescoping legs; below I describe a few other important weaknesses.

Fragility is a clear problem with anything portable. Earlier, I mentioned the personal klutziness that leads to things such as bashing a module corner into a storage rack when loading the rack with the module. Usually the result is dinging the paint or something similar, which is annoying, but not seriously damaging; and the cause is carelessness, not a weakness of the design.

It turns out that the module assembly latches are the primary source of inherent fragility, mainly because of their location. For example, on the storage rack, the latches hang down and can catch on a track of the module below; moving the upper module then can interfere in a damaging way with the track below. Also, if a module is placed on the ground, an unlatched latch can stand up in an unstable way and then collapse later, with the possibility of damage to either the latch or the module. Little damage has occurred on the modules; once, a latch was bent and had to be unbent; but the damage possibility is always there, as foreseen by the robustness objective. Being aware and being careful is adequate to avoid problems.

Another fragility weakness of the modules is the rail joiners, which stick out at the ends of the rails; and, although I do not recall an incident, I am always very aware of the possibility of bending them into uselessness. Also, since they stick out, the joiners can catch in things such as transportation blankets or someone's clothes, and then an insensitive yank can do damage. Once, in the workshop, fortunately not at an event, a joiner caught in a latch, resulting in a rail being ripped from its sleepers.

Another weakness is that considerable care is necessary when latching three modules together. Three modules are connected at the same time when two single track modules are latched to one double track module; this happens when junctions are put in the mainline, for example. When making this connection there are four module pins and six rail joiners to engage, all at the same time as supporting two, maybe all three, of the modules with one hand, leaving free the other hand to throw the latches. The trick is to concentrate on the task, get comfortable, not rush and force things (damage alert!), and then the process becomes quite smooth.

I have settled on a target railway configuration that provides me with a construction stopping point. This configuration is more complex than I had intended originally, mainly because it has a steamup siding for each of the inner and outer tracks. It has become clear, as railways have been erected at events, that steamup sidings are close to essential. Indeed, on one occasion I was asked if a steamup siding for the inner track could be provided.

As I write, component modules for the configuration shown below are complete. It has been necessary to make one more storage rack, basically to hold the modules for the steamup sidings. Of course, needing three storage racks means that everything no longer can be moved by one trip of the transportation van. This becomes a problem at the end of public events, where, normally, the whole thing has to be taken down and moved out before they close the doors and go home; there just is not time to make more than one trip. This problem quite possibly will be handled by purchasing a transportation trailer, which then needs to be stored ....

The steamup sidings are single track with no storage capability; this is intentional. After consideration of observed track congestion and its fallout at other railways, including more than one damaging accident, I decided to separate physically actual steamup from train storage and staging. This decision led to the junction-to-nowhere in each of the steamup sidings. The idea is that a storage yard can be attached to a steamup siding, allowing people to fuss over the makeup of their trains without getting in the way of the real action. Part of the junction-to-nowhere module in the inner steamup siding can be seen in the track configuration image, labelled ISJ.

In the Exhibition Oval configuration the sidings are long enough to hold two normal trains end-to-end. So there is room for one train to be staged whilst another is being steamed; thus, at the time of writing, the provision of storage yards is not an issue. The image below shows a train at one end of the inner siding; there is plenty of room for a train behind this one, out of camera range.

Arbitrary configurability is a capability of the portable railway about which I am pleased; it was an original objective, although at that time really I did not know where it would take things. This objective was a reflection of a desire not to be constrained to one railway configuration rather than anything else. Of course, the real constraints on multiple configurations are construction effort and cost, together with storage and transportation.

A little imagination easily produces new railway configurations requiring more modules. But even the Exhibition Oval can be modified to suit a particular erection site; for example, both steamup sidings can be on one side of the oval if that is desirable. The next image shows the Exhibition Oval and some modifications to it. The top left configuration is the existent oval as introduced already, and as seen in some of the images. The centre left configuration adds some, non-existent, storage and staging modules in an arbitrary layout. The bottom left configuration is possible currently by using the existent modules (and omitting some) to provide shorter steamup sidings and a slightly skinnier railway. The ticks are feet and ten feet markers.

Reporting the cost of a hobby is difficult. Clearly, labour dollar costs are zero. I have accurate purchase costs for the railway materials in the form of a log backed up with receipts. But, at present, I have little inclination to spend the time and effort to analyse my expenditures; although I did so in the early days of production.

I think that it is important to consider the whole cost of running trains in this not-the-cheapest-of-hobbies, rather than a large emphasis on locomotive prices, ignoring other expenditures. I think this emphasis happens rather a lot. The cost of the track used by the locomotives and their trains is important and significant. In real life railways, the cost of the right-of-way simply dwarfs the cost of the rolling stock. Within the hobby, if you fly or drive to spend a few days at Diamondhead, or the Summer Steamup, or Staver Locomotive, how much are you spending on running trains?

So, my estimate of the cost of this set of portable railway construction modules is not in dollars. The railway has cost the same as two or three medium priced G1 mainline locomotives. A quarter to a third of the cost has gone into production tooling, which still exists; and the rest has gone into production consumables. I have some inventory, which is dead money until it is used; for example, I have a surfeit of latches, I bought too many because I got a good unit price. If I buy a trailer, that will be another six thousand dollars; do I count that as part of the railway cost?

I am content with the cost of this configurable portable railway; job satisfaction is good, making a railway is a non-trivial accomplishment. And I did not need to deal with a stupid manager, job satisfaction is really good.

The Denver Garden Railway Society (DGRS) has a live steam track at the Colorado Railway Museum which is used by the members of the DGRS live steam special interest group. Other than that fixed location, we do not have a thriving live steam group in Denver; it seems that live steam interest exists mainly on the West and East coasts. A few of us are trying to stir up more activity in this town of three million people; progress is slow, but we hope positive. The portable railway described here is showing potential as a useful tool. And so this section displays a few images of erected railways, including recent activities of the Denver Area Small Scale Steamers (DASSS).

This outside shot, with the Rockies in the background, shows the railway deck at its highest elevation, i.e., some of the legs in sight are extended to their fullest. The railway deck height is about four feet above the undulating and sloping ground. On the other side of the oval, not in sight, the legs are close to their shortest length.

This foreshortened view was taken at the 2017 Maker Faire, before the inner steamup siding was made. The track transitions show well if you have the eye for them. Transitions do not stand out; of course, that is the whole point. The clip-on orange tape crowd control is quite effective, on adults especially, on kids mostly.

This image was taken at the Colorado Railroad Museum, the event was the DGRS 2018 annual steamup. A canopy has been erected over the inner track steamup siding.

A development of note is that a junction from my Garden Railway has been fully refurbished as part of the portable railway component set. The track interfaces enable the junction to be used as part of a storage yard for the outer steamup siding. The image below is a comparison between module 903, which is part of the Exhibition Oval, and the salvaged junction's module, labelled BS1. This module also is different in that its deck is an aluminium and plastic sandwich board (Dibond?), also salvaged, from a discarded real estate sign. The module joining the two junctions is the junction-to-nowhere that is part of the outer steamup siding for the Exhibition Oval.

The full Exhibition Oval was completed in late June 2019 when the outer steamup siding construction was finished. At the time of writing, mid October 2019, a railway has been erected twice. The delay is mainly because of the temperature, intense sunshine, and windy thunderstorms that have imposed themselves on Colorado for the past few weeks. [Acknowledgements to Global Warming]. The first, outdoor, setup came with intense sunshine, and wind that went from almost calm on Friday morning to hot, dry, and 20+mph by Sunday lunchtime. It was an enjoyable weekend, albeit wearing because of the weather. The second, indoor, setup was at the 2019 Denver Hobby Expo, where the weather was not an issue.

Different railway configurations were used on these two occasions, to suit the site requirements. The outdoor setup was configured with both steamup sidings on one side of the Exhibition Oval, to be close to the clubhouse. The indoor setup had the outer siding on one end to give more aisle space at the sides.

Here is the outdoor setup erected on sloping grass; there is about eighteen inches slope from right to left in this view. Eighteen inches variation is about the maximum possible with the standard telescoping legs. The railway was tied down with white straps and auger-style anchors as a precaution against the wind. [Click on the image, or any image that follows, to see a high resolution version. Return here with your browser back-button.]

These views show the sidings turning-out from the main oval; although all the flags show that the points are set for running around the oval. The inner track siding has no staging extension. The outer track siding has an elementary staging extension connected to its junction-to-nowhere. This staging line also contains a draw, which moves the narrow gauge rail from left side to right side. The draw, in turn, enables incorporation of a salvaged junction. Junction 901, in the left image, has its rails on poplar sleepers, in contrast to the redwood sleepers of the rest of the tracks.

What follows are a few views mainly showing the track configuration in use in the two configurations. The locomotive in the first image is an Accucraft 7/8 scale model of a 2 feet gauge Kerr Stuart Wren.

This shows the Rockies and a blue sky - as well as the railway. This was Saturday, before the wind picked-up. There are several locomotives in this view, including a Shay looking for its train out in the country.

The first train of the day on Sunday, raising steam on the inner track siding. The locomotive is a forty year old Aster Schools; the Southern Railway train is by Pete Comley.

A little later; other crews have arrived with their trains. The blue train is parked on the outer steamup siding staging extension; this train placement does not interfere with access to the main oval from the steamup siding. The locomotive is a Roundhouse Lady Anne disguised as a Colorado mountain engine; the coal train comprises kits by Bob Hartford, terminated with a scratch-built caboose.

A mixed-traffic train, hauled by two Roundhouse Berties, heading out to the country part of the line.

Here is the indoor setup, just complete on the day before the Hobby Expo; the module storage racks can be seen inside the oval. The lighting is a bit dim, definitely not up to the outdoors standard set by the Sun. This configuration, with 50% more modules than a simple oval, because of the sidings, takes an extra hour to set up.

This view shows the outer siding on the end of the oval; it is twelve feet shorter than it would be on the side of the oval, but is quite adequate for most purposes.

Another similar view. The green train and the Shay are parked on the staging track; the blue train and the red engine are on the steamup track.

The inner siding, which is on the side of the oval.

The flags that show point positions have been well-received by operators. However, that does not mean that misteaks do not occur. This flag was well-and-truly attacked. Since the flag is made of rubber, no damage was done to the locomotive that left these wheel tracks. Importantly, the locomotive driver was alerted to the mis-set points, and the train did not go boldly where it should not go.

This is the end of an overview of the history, design, construction, and deployment of a configurable portable railway. The project has been successful; the objectives have been fulfilled; and the railway has supported running trains on three miniature track gauges, with more operations to come.

More technical aspects of design, decision making, and construction, of the railway are available via the links on the Portable Railway page.

Finally, the author hopes that reading this description has been interesting and worthwhile; he can be contacted at info@ngdr.net