Civil Engineering

As for the “Big Railway”, this topic includes track laying as well as the trackbed (formation) and the bridges and structures and, indeed, the whole of the landscape through which the railway is built. However, baseboard construction also comes under this heading for the indoor modeller.


For a model railway, the construction of the landscape usually follows the construction of the formation and the trackwork. However, it is helpful to know what sort of topography you are aiming at and how you will achieve it before starting work.

For this layout, I decided that I would be using an open top construction for the scenic sections so that the landscape contours would be easier to create. I am using a series of crossbeams and uprights to support the trackbed, which is formed from plywood with a T-rib under it to stop sagging between supports. However, for the storage loops and the station area, I am using a conventional flat-topped construction.

Although the open top method gives greater freedom for creativity in the landscaping, it does require greater discipline and planning for the track layout, since the trackbed forms need to be pre-cut to the intended topography.


The traditional method of track construction is to attach the track to a timber base (with or without a layer of cork) and then use PVA glue to bind a granular material such as grit for the ballast. Unfortunately, the PVA sets as hard as rock and creates a very rigid trackbed that generates the maximum amount of unwanted noise. In his book An Approach to Building Finescale Track in 4mm, Iain Rice likens the track and its immediate support structure to an acoustic loudspeaker system. He points out that, if the system is rigid, then running noise will be amplified. The way to minimise the generation of unwanted sound is to keep the system flexible and this is where foam underlay is so effective.

Iain’s approach is to separate the track from the rigid base by interposing a layer of foam between the ballasted sleepers and the base, and to use a flexible rubber adhesive to hold the ballast instead of PVA.


The track standards I will be able to adopt will be dependent upon the wheel standards I decide to use. I have a substantial fleet of Hornby Dublo wagons and coaches with their original wheelsets and this will define my starting point.

I’m quite attracted to 00-SF which is designed to allow commercial ready-to-run stock with standard wheels to be run straight out of the box with finescale (1.0mm wide) flangeways. The flangeway gaps in S&C is where the improvement in appreances really stands out. However, since the wheels of my original Hornby Dublo stock have a back-to-back dimension of 14.0mm, I fear they will not run on track made to this standard.

Most of the trackwork will represent bullhead on timber sleepers but I will have a length of concrete sleepered track representing an early installation of flat bottom CWR on the main line. I will be setting out all curves with transitions to smooth the entry and exit of my trains and the curves will be canted (banked or superelevated).

The setting out of transitions and the application of cant can be quite complex on the big railway but some considerable simplification can be made for a model. The maximum cant is 6 inches for the big railway which is 2mm at 4mm to the foot. However, cant is measured over the centres of the rails which is 5 feet, or 20mm, so the corresponding cant on OO would be 1.75mm. The normal rules for calculating cant and cant deficiency don’t apply for a model and a cant of less than 1mm will not be noticeable, so a value of 1mm to 1.5mm would be about right for most curves. The important thing to take from the prototype is the rate of change of cant because this translates as twist and creates a potential for derailment. The normal design limit for the big railway is 1 in 400, 1.5mm at 1 in 400 gives a transition length of 600mm which will probably look about right. Interestingly, Templot’s default transition length is 2 chains or 528mm at 4mm to the foot, which gives a cant gradient of 1 in 352. I think some experimentation is called for!

It’s worth remembering that, for a double track railway in the 1960s, all four rails on a curve often remain coplanar. For a four-track railway, pairs of tracks would be coplanar, with a step down in the ten-foot space. This makes canting easier on the model if the roadbed or sub-base is canted, rather than trying to shim the high rail. The railway today economises on ballast by canting each track independently so there will be a step down in the six-foot.

For my turnouts, I am considering different options for outside and inside, they will be adjusted to suit my corrected six-foot dimensions. Whichever I go for, I won’t be using turnouts with hinged switch blades in prime spots indoors because I can’t abide the look of them, totally unprototypical!

Following a recommendation, I bought a couple of books written by Iain Rice: A Pragmatic Guide To Building, Wiring and Laying PCB Track and An Approach to Building Finescale Track in 4mm. In these books, he takes a fresh view of what the modeller is trying to do, what can be done and the implications of the different standards that can be adopted. I reckon I may well be building my own S&C and, to help me with this, I have splashed out on a copy of the Templot software for track planning and plotting construction templates. It is quite unlike any other track planning software (it’s designed for plotting individual templates, not a whole layout) and there’s a steep learning curve. However, after a few hours experimentation over two or three evenings, I suddenly found that I was producing something useful.


Track LayoutIndexSignals and Turnouts