Getting on the rack

A spinning pinion has to drop into teeth it cannot see.

A pure mountain line starts on the rack at the bottom station and never leaves it. But most rack railways are mixed: the train runs on its wheels along the valley and needs the rack only for the steep part. Somewhere the pinion, turning freely with the wheels, has to drop into the rack while the train is moving.

The trouble is that it arrives at a random phase. If a pinion tooth happens to land on top of a rack tooth, tip on tip, the pinion rides up, and either something breaks or the engine climbs the rack. Roman Abt's answer, first tried on the Harz railway in May 1885, was a short length of rack hinged to the fixed rack and carried on springs, with teeth that grow from almost nothing to full height.

The rack entry

Drive in
Pitch of the entry teeth
Entry piece
Misalignment on arrival
Entry pushed down
Engaged after
Run 200 trains to see how often each setting works.
Each dot is one tooth of the entry piece, plotted by how far it is out of line with the pinion's gaps when the pinion reaches it. With the pitch altered, the dots drift across the green band, so one of them almost always drops in. With the rack's own pitch they stay in a flat row, and if that row is outside the band the pinion never finds a way in. Rack proportions from Snowdon; the entry's tooth growth and capture band are simplified.

Why the altered pitch matters

Try the 120 mm setting with "Run 200 trains". More than half the trains arrive with the teeth nearly tip on tip, and with every entry tooth at the same pitch as the pinion they stay that way all along the entry, until they strike the fixed rack. Give the entry teeth a slightly different pitch and each tooth meets the pinion a few millimetres further out of step than the last, so the misalignment sweeps across the gap and nearly every train finds a way in.

Most Abt entries used a slightly longer pitch, about 124 mm on a 120 mm rack. When the Montreux–Glion line was built in 1909 its engineers chose a shorter one, 116 mm, because on their trains the pinion's pitch circle travelled a little less than the running wheels rolled.

On the Harz trial of 1885 the train ran onto the rack "so gently" that the crew did not notice.

Three generations

Abt's sprung entry (1885). A hinged piece about 2.5 m long, sprung, with teeth rising to full height and an altered pitch. The Swiss rules still describe it, and say it should be shorter than the distance between a vehicle's pinions, so that only one pinion is ever on it. Early entries were taken at walking pace, 2 to 3 km/h.

The synchroniser (Brünig, from 1941). Before the entry, a rubber-faced bar brushes the idle pinions up to the right speed and a row of sprung rollers lines them up, so they arrive nearly in step.

The sliding entry (Marfurt and Tensol, 1996). The entry piece rides on inclined levers. A tooth landing tip on tip pushes it down, and the levers slide it lengthwise until the teeth mesh, so no pitch error is needed. With this work, by Josef Marfurt and Hans Tribolet in the 1990s, entry speeds rose to 15 km/h for powered units and 25 km/h for coaches. Ten kilometres an hour is still the usual limit.

Entries are still being studied. In 2021 a pinion on the Brünig line broke on its way down; the Swiss investigators found its springs had been overloaded, made worse by teeth striking head-on at the rack exits. In 2025 students at OST worked with Tensol Rail on a sliding synchronising rail rated for 20 km/h.

Sources: SBZ (1885), the Harz trial; SBZ (1909), Montreux–Glion; V. von Röll (1923), "Zahnbahnen"; H. G. Wägli, "Carl Roman Abt" (2005); P. Berger, Ferrum 86 (2014); Swiss Transportation Safety Investigation Board, pinion fracture at Giswil (2022); Rossi, OST and Tensol Rail (2025).