Rack Railways For Rail3d

There are quite a few items of stock, scenery, track, etc. available for those who want to model rack railways in Rail3D, especially Swiss ones. On this page, we aim to bring together a few notes on how rack railways are constructed and operated, with some ideas for implementing them in Rail3D.

This page is still very much under construction, but please feel free to contribute

1 General notes

Conventional railways rely on adhesion between smooth steel wheels and rails. They are limited to a maximum gradient of about 5%. With specially adapted-railcars that have all axles driven and are fitted with extra brake systems, it is possible to go to 7 or 8%. Beyond that, you need some means of positive engagement between train and track, usually a rack.

There are, broadly-speaking, two categories of railways that use a rack:

1.1 Pure rack railways

These are normally short tourist lines, linking a base station to a mountain summit, and having a rising gradient throughout. They use the rack for all traction and braking, the conventional wheels simply providing support and guidance for the train1. Maximum gradient is typically 20 or 25%.

Trains are very short — no more than one or two passenger vehicles in most cases. Locos and rolling stock are designed with “uphill” and “downhill” ends, and can’t operate facing the other way. The Wengeneralpbahn, which operates two pure rack railways “back-to-back”, has to provide a reversing triangle at its summit station so that stock can be moved from one line to the other.

Other examples include Snowdon, Mount Washington, the two Rigi ralways, the Jungfraubahn, etc.

1.2 Mixed rack and adhesion

These are otherwise conventional railways that include rack sections to allow steeper gradients on certain parts of the line, thus reducing construction costs by avoiding the need for long tunnels and other complicated civil engineering. Compare the Rh B’s adhesion-worked Albula line with its many tunnels (built in the affluent 1890s) to the low-budget rack-assisted fo (completed 1915–1926).

These lines use loco-hauled trains or railcars. They have to provide both rack and adhesion traction and braking systems on all their vehicles. They may have rack sections with gradients in both senses, so vehicles usually don’t have specific “uphill” and “downhill” ends.

Lines that use loco-hauled trains (e.g. fo/bvz, Brünig) are usually limited to a maximum gradient of 12 or 13% because of limitations on drawbar forces. There are some much steeper mixed rack-and-adhesion lines (e.g. Martigny-Châtelard, 20%), but these only use railcars.

Apart from the fo, other well-known examples incude the bvz and the Brünig line in Switzerland.

2 The Rack per se

All rack railways have some kind of toothed element, usually in the middle between the rails. On mixed rack-and-adhesion lines the rack must be a few cm higher than the running rails so that traction and brake pinions do not foul pointwork on adhesion sections.

There are three systems in common use:

2.1 Riggenbach

The Riggenbach system uses a “ladder” structure, where the teeth are formed by rods joining two flanged plates. The idea is that the cogwheel cannot jump sideways out of the rack. The usual pitch of the teeth is 100mm.

Riggenbach patented his system in 1863, but essentially the same idea was developed independently by an American engineer called Silvester Marsh, and Marsh was able to put it into practice on the Mount Washington railway in 1869, two years before Riggenbach’s Rigi railway opened in Switzerland. A curious feature of the Rigi and Mount Washington lines is that, unlike most other rack railways, they are built to standard gauge.

Apart from the Rigi, the Riggenbach system is used on quite a few older lines. One well-known example is the Brünig railway.

2.2 Abt

The Abt system uses two or (occasionally) three plates set with their teeth out of phase. The pitch of the teeth is usually 120mm on each rack. Like the Riggenbach system, it is impossible for the cogwheel to jump out of the rack sideways. An additional advantage is that the lower effective pitch of the teeth gives a smoother action.

This is by far the most common system. Well-known examples include the Furka-Oberalp/bvz system.

2.3 Strub

The Strub system is effectively an Abt rack with only one, thicker plate. It is more recent and thus also less common than the Abt and Riggenbach systems. Apart from cost, one advantage is that it is less susceptible than the Abt and Riggenbach systems to problems with ice and snow. It was first used on the Jungfraubahn — other examples include Martigny-Châtelard.

Strub rack is usually compatible with Riggenbach, thus Riggenbach lines sometimes use the simpler Strub rack through pointwork or in less steeply-graded sections like stations.

A variant is the Von Roll system — it also has a single rack, but the cross-sectional shape is different. For Rail3D purposes, the two can probably be considered as interchangable.

2.4 Other

The Riggenbach, Abt and Strub systems can be used on gradients up to about 25%. The Locher “fishbone” system has a pair of opposed horizontal pinions engaging teeth cut into the sides of the rack, so that it does not depend on gravity for the engagement of the teeth, and can be used on extreme gradients. The only example is the Pilatus railway, 1885 - maximum gradient 50%.

For completeness, a quick mention of the Blenkinsop system — the very first rack railway. Blenkinsop didn’t believe that a steam loco with smooth wheels could provide enough traction to pull a train, so proposed forming teeth on the side of the running rail to engage with teeth on the loco’s driving wheel. His system was tried out on the Middleton Railway in 1812, but it was soon found that adhesion provided enough traction and was cheaper.2 3

The Fell system is not a rack — it uses a smooth central rail with drive or brake wheels pushed against it laterally. The theory was that it could allow higher speeds on gradients too steep for adhesion but not steep enough to need a rack. It was never much of a success in its intended use on main lines, but a simplified version is used for additional braking on the Snaefell railway in the Isle of Man.

2.5 Rack entry sections

On mixed rack and adhesion lines, special arrangements are necessary to provide for proper engagement of the pinions with the rack where a rack section starts. Rack entry sections have sprung teeth designed to lessen the wear and tear on rack and train and the shock of the impact as the rack has to overcome the rotational inertia of the brake pinions. As a passenger on a train entering a rack section, you will hear a series of bangs advancing along the train as the brake pinion on each coach engages with the rack.

The maximum speed for entering the rack is very low — not more than 10km/hr. Once a train is fully engaged with the rack, it can accelerate to the rack speed limit — typically around 30km/hr. Trains don’t have to reduce speed for leaving the rack.

In Switzerland, the start of the rack is indicated by a white disc with letter “A” (Anfang) or “C” (Commencement); on the back of the sign is “E” or “T” (work it out for yourself…).4

2.6 Switches

Pointwork on rack lines is difficult and expensive. For one thing, you have to arrange for movable bits of rack where the rack crosses the running rails, and for another you have to make sure that all the teeth of all the movable bits of rack match up exactly in both positions (the tooth pitch can’t vary by more than ± 1mm). Consequently, rack railways tend to avoid pointwork as far as possible. 5 6

On mixed lines, it’s usually quite easy to reduce the gradient through a station so that you can interrupt the rack. The problem is that trains then have to slow right down again to re-enter the rack as they leave the station.

On pure rack lines, you have to have at least a few points, but you often see ingenious solutions to keep the number down. One trick is to use the trainshed of the base station as a depot or carriage shed when trains are not running, so that the points for the platform tracks are used twice. Another common dodge is to use a traverser or sector plate for access to the depot tracks.

Some systems don’t allow conventional points at all — with the Locher system, for example, the only way to build a switch is to have a big steel plate with a length of straight track bolted to one side and a curved section on the other side — to change the route you flip the plate over 7. Sliding track sections are also used sometimes.

At the moment, there’s not much you can do in Rail3D to model any of these things. It might be possible to do something by (ab-)using the Point Indicator feature, but it’s probably not worth it.

One thing you might be able to model without too much trouble is the new flexible track point developed for the Rigi railway8. This just pushes a length of track across from one position to the other. Sounds simple, but there looks to be an awful lot of engineering under the floorplate. You could (presumably) do this by using a script to modify the track types of the two links forming the switch so that the one in use is visible and the other isn’t.

3 Traction

Most rack railways use electric traction — if you’re on a mountain, there’s usually a source of cheap hydroelectric power nearby. The few older lines that didn’t manage to raise the cash to electrify either closed down or carried on with their existing steam engines until the point where steam became an attraction in itself. Diesel traction is rather uncommon, except as a backup on steam or electric lines.

3.1 Steam

Steam-powered rack railways use small tank engines — even on mixed lines, there was not much point in having big engines on the rack sections, because the total weight of the train was limited by drawbar forces rather than loco power. On the Brünig line in steam days, it was common for heavy passenger trains working over the rack section to be split into two portions, each with a loco at both ends, to keep within the drawbar limits.9

On pure rack lines, the usual arrangement is for the loco to push one or two trailers up the hill. Most lines do not use couplings between vehicles — as long as the whole line is on an up-gradient they aren’t needed. Originally this was also a safety precaution — if the loco derailed or ran away, the trailer would not risk being dragged off the track, but be brought safely to a stop by its automatic brakes.

There is a well-known problem with steam engines on steep hills — boilers are designed on the assumption that the water-level will be constant from one end to the other. If the loco is tilted forwards, there is a danger that the water level will fall below the top of the firebox, causing the boiler to overheat. A rule that applies to all rack lines is thus that steam locos run with the chimney pointing uphill. On lines that go “up one side and down the other” (e.g. fo, Brünig), turning facilities had to be provided at the summit stations.

On pure rack lines, where there is a relatively uniform gradient, the boiler is often mounted on the loco at an angle, usually set about halfway between the maximum and minimum gradients on the line, to keep the water level as close as possible to being parallel to the boiler axis.

Locos for mixed working usually have an extra set of cylinders for driving the rack pinion. On adhesion sections the steam to the rack cylinders is cut off; on rack sections both sets of cylinders work together, often in compound. Locos for pure rack railways of course only need one set of cylinders. There is usually one drive pinion, and one or more separate brake pinions, although the Gaisbergbahn in Salzburg for some reason decided that it needed tandem drive pinions (when the locos had been tried out for a year or so, common sense prevailed and the extra pinion was removed).

Mixed rack and adhesion loco, Brünig railway

3.2 Electric

As rack railways tend to be small and self-contained, you can find practically every imaginable electrification system somewhere, from third-rail dc (Martigny-Châtelard) to three-phase ac (Jungfrau, Gornergrat). I expect that there’s one operating on the Märklin stud-contact system somewhere…

In the early days of electrification, the usual idea was to use small electric locos, which simply replaced steam locos, and operated in the same way, pushing passenger cars up the hill. Later on, pure rack railways tended to move over to railcar operation. The Schynige Platte line in the Berner Oberland is one of the few pure rack railways that still use electric locomotives.

Mixed rack and adhesion electric loco, Schöllenen railway (now the Göschenen-Andermatt secion of the mgb)

While it is relatively straightforward to arrange an electric motor to drive a rack pinion instead of adhesion wheels, it is quite an engineering challenge to provide both types of drive in the same loco. Especially with the big traction motors needed in the early days of electrification, it was often a struggle to fit everything in between the frames of a narrow-gauge bogie. There are a number of different approaches:

  • Two independent drivetrains, which is mechanically simple, but is heavy and leads to complications when the two have to be synchronised somehow when entering a rack section (e.g. the Deh4/6 of the Brünig, which have two conventional adhesion bogies with a third bogie between them carrying the drive to the rack pinion).
  • Rack drive permanently coupled to adhesion drive — the drive pinions simply rotate freely on adhesion sections, doing no harm. The problem is that the ratio between the effective speeds of the rack and adhesion drives varies with wear of the driving wheel tyres, so that in practice the wheels will be trying to go slightly faster or slower than the rack drive, leading to frictional losses and increased wear.
  • Rack drive permanently engaged, but adhesion drive can be disengaged with a clutch on rack sections (e.g. the bvz/fo crocodiles). The disadvantage is that the rack does all the work on rack sections, and thus wears more quickly.
  • Rack and adhesion drives coupled through a differential — complicated, but seems to be the preferred solution nowadays.

Furka Oberalp hge 4/4i of 1940 — note the very long bogies. The length is due to the traction motors being mounted at the outer ends, rather than between the axles

4 Braking

All rack trains have at least two braking systems:

  • The service brake, which acts on at least one rack pinion on each vehicle in the train, and is capable of stopping and holding the train on the steepest gradient on the line.
  • The dynamic brake, fitted to locos and railcars. This brake is able to dissipate enough power to hold the train at a constant speed during descent. In other words, since trains usually climb and descend at the same speed, it must have the same continuous power rating as the traction system (about 8kW per tonne of train mass is needed for typical speeds of 15km/hr on 20% grades).
Ideally, electric trains use regenerative braking, feeding power back into the contact wire, but of course this only works as long as there are ascending trains to take up the power. If there is not enough load, the train has to resort to rheostatic braking, dumping power into a set of resistors on the car roof.
Steam locos are usually fitted with an arrangement that allows them to use the cylinders to pump air through a restrictor valve when descending. The cylinders have to be cooled by injecting water into them — this is why you will still see puffs of steam coming from the chimney of a rack loco even when it is going downhill.

Trains on pure rack systems usually also have:

  • A roll-back brake — this is a simple ratchet device engaged when ascending, to prevent the train getting out of control in the event of a traction failure.
  • An overspeed brake — this engages automatically and brings the train to a stop if it exceeds the line speed limit.

5 Signalling and operation

Because of the low speeds and short braking distances, signalling is not really needed to prevent collisions on rack railways (with the possible exception of tunnels). It is mainly a question of regulating traffic on single-line sections to prevent impasses or undue delays. Consequently, signalling practices are sometimes rather different from other railways.

A common feature of most lines, imposed by the limited length of trains, is that it is usual for two or three trains to ascend or descend together as a batch, the drivers of the following trains keeping a safe distance behind the one in front on sight. The last train of the batch is the one that has the formal authorisation to be in the single-line section — the trains ahead of it usually carry a coloured disc to indicate “another train following” to the crews of opposing trains waiting at passing stations.

In the simplest case, there are no fixed signals. Trains proceed from one passing loop to the next on radio or telephone instructions from the line controller, and train crews change the points by hand at intermediate stations (example: Brienz-Rothorn railway).

In more sophisticated systems, stations are staffed and have entry and exit signals. Points and signals are controlled by the stationmaster under instructions from the line controller.

6 Resources

6.1 Footnotes

1 Unlike most conventional trains, stock for rack-only lines often has wheels that are left free to rotate on the axles, to reduce wheel wear. (↑)

2 Spartacus on Blenkinsop’s rack railway (↑)

3 The Pictorial History of the Locomotive on Blenkinsop (↑)

4 Roland Smiderkal’s Rack signals page (↑)

5 Photo of a switch for the Riggenbach system (www.luwi.ch) (↑)

6 Photo of a switch for the Abt system (Roland Zumbühl) (↑)

7 Flippable track on the Pilatus railway (Wikipedia) (↑)

8 Bendable track on the Rigi railway (pdf) (↑)

9 cf. Jeanmaire, Brünigbahn — Improvements in rolling stock construction and better couplings mean that trains can be rather longer nowadays (↑)