"Wall Mountain" N-scale Model Train Layout

This layout proved to require considerably more effort than I anticipated, especially for a layout consisting of a single 50" run of almost-straight track. There was many a twist in that "almost".

The History

In the foreground you see the old benchtop from the Physics lab that was renovated in Christmas 2006. When Bruce and I saw these marvellous pieces of wood in the dumpster we grabbed them. New Zealand Rimu, at once I knew the stuff should be recycled in some beautiful and exciting way. Back in California I had an idea for a wall-hangable layout using the Fleischmann N-scale cog track and locomotive. Here was just the material for this job: Beautiful wood that was thicker than an N-scale layout needs to be wide.

This is the woodwork part of the first version of the layout. The route up the mountain is the edge of the frontmost piece. The two routed recesses at either end will become stations. The idea is that the grainy Rimu wood represents rock, the rock of a mountain, up the side of which a cog railway will snake.

The layout was partly inspired by discovering the rack railway system from my favourite N-scale maker, Fleischmann, partly by the sight of the thick lab-bench wood, and partly by a postcard that Dominique sent to us years ago. This post card showed a number of European trains, including the Pilatusbahn. There are other cog railways such as one climbing Snowdonia in Wales (http://www.snowdonrailway.co.uk/), and Mount Washington in New Hampshire (http://www.thecog.com/), but the Pilatusbahn seems to be the epitomy. There is a good potted history of cog rail at http://www.cograilway.com/history.htm.

Note in the picture above the two electronic signalling boxes, one at the entrance to each station, are already built into the "rock". These are GP2A25 Electro-optic proximity sensors for the automatic controller, but they are about the size of a trackside signal when viewed in N-scale. The circular hole leads to a small control panel to activate the layout.

The layout is intended to be hung on a wall, just like a picture. The train will periodically leave one station and climb or descend to the other.

The Problems

Around February 2009, Fleischmann went into receivership and was bought by the company that previously bought Roco and turned that company profitable. The blogs intimate that it should have been no surprise, as their technology had not matured as did other manufacturers, for example my brand new loco pictured here is still a 3-pole design, and I gather all Fleischmann locos remain so.

The Fleischmann cog system is supposed to be able to climb up to a 30% grade grade. I dutifully designed the route with a maximum grade of 30% grade.

This is the Fleischmann 9119 Rack Rail. Unlike the regular Fleischmann N-scale track, the rack rail is not made with ballast formed into the plastic base of the track, but resembles the common Peco flextrack and settrack arrangement with individual sleepers (ties). It is the higher profile type of track, and the rails do not like to deform when bent, making it hard to manage.

Note also the frequent holes placed to accomodate nails to hold the track in place.

The sleepers are held together by the toothed rack in the middle of the track. This protrudes above the rails and is engaged by a cog on the axle of one pair of drive wheels in the locomotive.

The first problem is that the rails have a tendency to buckle. They bend around corners left and right just fine, but not in the other direction. When trying to make them conform to the curvature of hills, they tend to spring out of the plastic retainers on the sleepers/ties, damaging them.

The second problem is that the rail needs to be nailed in a lot of places, for which it has numerous holes. However, when you have the nails in deep enough, the rack is no longer flat, and it dips compared to the rails where the nails are located. The dips in the rack and the irregularity of the rails are not a problem, except when you approach 30%, before which grade the slightest imperfection leads to the cog letting go and the locomotive slithering and juddering to a gruesome crash at the bottom of the hill.

The third problem concerns not the grade, but the rate at which the track can transition from one grade to another. This is an underside view of the Fleischmann Edelweiss Cog Locomotive, an 87305. Note that it is a six-wheel arrangement, and that the cog that engages the rack is on the middle of the three axles. This is, I believe, a design flaw. When the track curves in a convex fashion, there is a tendency to lift one or both of the end axles off the rail, leaving the loco see-sawing. Worse, a concave curvature lifts the axle with the cog, and the loco is vastly more likely to let go and slide down. It would function much better as a four-wheel design (perhaps like those 0-6-0 steam locomotives whose middle axle is largely cosmetic).

The Second Version

This is the track half, viewed from the mountain side, after a rework to reduce the radii of curvature. The most visible change is the use of double-sided tape to mount the track. This proved to give much less undulation in the rails.

Just to give some idea of the steepness of the railway, this is a picture taken with the camera horizontal. You are looking at the whole train from within the cutaway mountain, and it is climbing out of the tunnel towards the top station. The black faces mate with the other half of the structure to complete the tunnel.

This closeup shows a section of track with a serious "bump" in the near-side rail. The pins have been added to press the bump down into the correct arc. This may not seem like much, but it is enough to lift the wheels on one side and stop the train. A concave, rather than convex, bump, makes it let go and slide down.

If you can see no error in the arc, compare the height of the near-side rail with the height of the rack. In the vicinity of the nail, note that the apparent height of the teeth of the rack above the line of the top of the rail is somewhat reduced. This undulation can barely be felt by running your finger along the track quickly, while wearing surgical gloves to reduce the sensation of roughness, but it causes a perceptible roll in the motion of the loco.

Bending the rails absolutely demands a rail bender. The use of a rail bender at such scales is virtually unheard-of, since modern track flexes relatively easily in the X-Y plane, but in the Y-Z plane it is another story altogether.

Here is a rail bender fashioned from the bearings of an old disk drive and some aluminium and steel plates and bolts. Thanks to Brian Clark, workshop master. The main plate is 45mm by 35mm.

Experience from Elsewhere

September 2009 saw me complete work on Sprocket BahN, another layout based on the same Fleischmann rack rail system. That layout is reasonably reliable, but will not run for days and days without intervention, as one might hope for a layout that is fully automated and expected to keep a timetable.

In the course of Sprocket BahN I hoped to sidestep all the problems described above. I encountered yet another one: The cog meshing with the rack converts force along the line of the track into a lifting force that almost overpowers the weight from the steel masses in the locomotive, resulting in the loco having greatly increased sensitivity to small particles or discontinuities on the track. It regularly loses electrical contact with one or other rail and "sticks".

On any given track there will be a speed setting that gives the loco enough momentum to trundle over the discontinuities. The upshot is that you have to run faster than you might like to make it through the difficult parts, especially the concave regions where the second derivative of height (the rate of change of gradient) is largest.

--- there may be even more to come ---

The Layout Itself

Here are the Top and Bottom stations before the track is laid. The buildings are from the Hornby Lyddle End range, but they have been sawn to yield "relief" fronts (like something from Portmeirion).