Swedish State Railway (SJ)

by Per Olofsson and John Hinson

Semaphores delivered from the end of the 19th century to the Swedish State Railways were normally manufactured in iron with 8, 10, 12 or 14 metre height and equipped with one, two or three arms. T-semaphores were normally 8 metres in height. Semaphores could be locally or centrally operated from a mechanical frame using double-wire or electrical interlocking.

Signals could be mechanically or electrically operated. Mechanical signals could be controlled in several different ways.

  1. with the signal wire controlling both the semaphore and its associated advanced (distant) signal - the latter usually being a disc type
  2. with signal wire run above ground (on pulleys, UK-style)
  3. with signal wire run below ground (in ducting)
  4. locally, by hand-crank

One of the first fixed signals used in Sweden was the T-semaphore. The design of this was influenced by British practice, of lower quadrant type and built in timber. It would be placed on the platform in front of the station building. T-semaphores had two arms, one for each direction.

It was decided in 1915 that all semaphore signals should be of the upper quadrant type.

The last semaphores on Banverket (Swedish National Rail Administration) track were replaced in 1999 but on Inlandsbanan there are still (2002) two or three stations equipped.

T-semaphore signals

T-semaphore at intermediate station
Photograph from the collection of
Per Olofsson

Although the original function of a T-semaphore became redundant as stations became interlocked, some survived at intermediate sidings to protect the siding from movements in either direction when shunting was taking place.

The picture on the left shows a station signal at Jämsjöslätt in the 1930s. One arm has been lowered inside the post to show clear. A wooden T-semaphore protecting a siding is shown to the right. As with early British signals, any driver approaching such a signal showing danger would be expected to bring his train to a stand before any points, rather than stop at the signal itself.

A few T-semaphores survived on private (or former-private) railways at stations into the 1950s. Some protecting intermediate sidings existed as late as the 1970s.

Swedish T-semaphore signal
Photograph from the collection of
Per Olofsson

Drawing of later T-semaphore signal
Drawing of later, iron-built T-semaphore

Single-armed semaphore stop signals

Single-arm semaphore signals
Photograph from the collection of Dr. J W F Scrimgeour

A pair of single-armed semaphore signals at a junction station. The signal on the left is locally operated by crank-handles at the foot of each post - the signal man has a hand on this in the photograph. These rotate the large pulleys mid-way up the post, which in turn raise the signal arm. However, it appears that the other signal is worked remotely - note the ducting. The barriers are wire-operated, controlled from the signal hut.

The location is Ostersund. Here, the station was interlocked, worked from a double-wire frame of the type shown at Stocksund. In 1911 a new railway was built between Uppsala and Enköping and it is probable that the left semaphore protects the line towards Uppsala and that signal was not integrated (at the time for the picture) in the interlocking frame.

The signal arms have similarity to German signals in style (with the "blob" at the end) but are pivoted at the right-hand end. The arms were painted with the top half red and the bottom half yellow. The post, or mast was also painted in these colours.

Semaphores were most commonly installed to the left of the track, and at interlocked stations would be wire-operated from the centralised crank-handle or lever frame. The T semaphores installed on the platform were also operated by hand-crank.

Multiple-armed semaphore stop signals

In the early 1900s Sweden switched to speed signalling and the meaning of the signals was altered. Speed signalling does not inform the driver which route he is taking, only the speed at which he may travel through a junction.

Most Swedish speed signals were provided with two arms, but a third arm was provided in more complex areas. The indications were as follows:

Top arm horizontal
Middle arm vertical
Lower arm vertical
Top arm at 45°
Middle arm vertical
Lower arm vertical
Top arm at 45°
Middle arm at 45°
Lower arm vertical
Clear with caution and 45 km/h speed restriction
Top arm at 45°
Middle arm at 45°
Lower arm at 45°
Clear with caution and 45 km/h speed restriction; route is unusually short or the route is set only to the shorter of two alternative routes

Three-armed semaphore cleared for main line
Photograph from the collection of
Dr. J W F Scrimgeour

In the left view, the signal is cleared for the main, or higher speed, route. The two arms below are barely noticeable as they are flush with the signal post.

The signal on the right is cleared for a lesser route (of lower speed) and displays two arms at 45º.

It is interesting to note that present-day signalling in Sweden reflects the same principles, and the colour-light signals are capable of displaying one, two or three green lights for the same circumstances.

Three-armed semaphore cleared for lower speed route
Photograph from the collection of
Dr. J W F Scrimgeour

Disc advanced signal

Disc distant signal in warning position
Photograph from the collection of
Dr. J W F Scrimgeour

These two photographs show a mechanically-operated disc signal, worked by double-wire. The colour of this disc is green with white surround, but at night the light displays a flashing green aspect when the signal is in the "warning" position.

When the signal is cleared, the disc pivots to the horizontal position out of sight, and the lamps is exposed to show (by night) a flashing white light.

Disc distant signal in clear position
Photograph from the collection of
Dr. J W F Scrimgeour

These signals are called Advanced signals because they give advanced warning of the signal ahead - very similar in function to the Distant signal of the UK.

About the photographs

Additional notes by Lars-Henrik Eriksson and Gunnar Ekeving

Comments about this article should be addressed to John Hinson