Underground Radio Tests on the Nene Valley Railway

This report has been compiled by John Rabson, with additional contributions from Rob Gill, Chris Trayner and Roger Goodchild.


Tests were conducted in a railway tunnel to establish the best frequencies for radio communications. Best results were obtained at UHF, next best being an LF induction/earth current system. VHF was of somewhat marginal performance, and mobile telephones did not work any useful distance into the tunnel.

Caves are not the only holes we find in the ground. There are railway tunnels in many parts of the British Isles. We do not usually venture into these but in an emergency RAYNET or similar organisations might be asked to provide communications in and from them. It is a good idea to do our homework beforehand. Enquiries about previous work revealed that one or two things had been tried but not written up. Some years ago tests were done in a mile-long tunnel on BR's West Coast Main Line. UHF was effective but VHF was not. Some tunnel tests were also done in Essex a few years ago with similar findings, but no report seemed to have been published. We felt is was time to stop theorising, find a tunnel and do some experiments.

Unfortunately, it is not easy to get permission to use railway tunnels since many of them are regularly used: the light at the end of the tunnel tends to indicate an approaching train rather than a breakthrough! I sought advice from railway enthusiasts and ended up arranging to use Wansford Tunnel on the privately owned and run Nene Valley Railway in Cambridgeshire, on Monday 9th February 1998 with Chris Trayner, Rob Gill and Roger Goodchild.

Wansford Tunnel is 550m long and straight, 7m high and 7.5m wide. It runs through oolitic limestone and is lined with Victorian brick. It contains two standard gauge railway tracks, one of which runs only about 150m into the tunnel and is used in winter to store rolling stock in the dry. The notional markings on the sleepers, which were supposedly at intervals of one chain (22 yards or 20m) were not always accurate. We therefore took our own measurements of distance into the tunnel.

The Tests

We tried four different frequency bands: LF, VHF (145MHz) and UHF (433MHz amateur band) and 934MHz CB2.

On LF we used John Hey's cave radios. These worked well from one end of the tunnel to the other, using loop aerials. It made no difference whether the loops were placed between the rails or at the side of the tunnel, or in the 6 foot space between the two tracks.

We then moved one station to the surface above the point where the tunnel was at its deepest , about 20m below the surface, and replaced its loop aerial with an earth current system. This used a 40m wire lying on the ground and running at right angles to the tunnel, with a step-up transformer between the transmitter and the earth wire. (It was not possible to run the wire along the line of the tunnel because of the layout of the roads). This configuration worked satisfactorily for a distance of 250m along the tunnel, centred on the point vertically below the surface station. Reception was noticeably better with the loop vertical than with it horizontal.

A pair of 433MHz hand-helds with rubber duck aerials worked perfectly satisfactorily from one end of the tunnel to the other, as did the 934MHz mobile equipment which used a rubber duck aerial at one end and a 10 element beam at the other. We also checked how far into the tunnel it was possible to get service from the local GSM900 mobile telephone cell. The answer was about 10m from the eastern portal and 4m from the western portal: GSM1800 was not tested.

VHF hand-helds were much less effective than UHF ones. We got rather less than half way along the tunnel before we lost our signal altogether. Replacing the rubber duck aerial at the tunnel mouth with a 4 element Yagi gave us only a few metres increase in range. Replacing the rubber duck in the tunnel with a mobile whip on a rucksack gave a large improvement. We repeated the tests using a field strength measuring receiver with the beam horizontal (vertical was not so good). The results are shown on the graph.

The received signal strength varied very little for the first 100 metres into the tunnel probably an artefact due to receiver saturation. Thereafter the signal fell off steadily until, at the mid point of the tunnel, it was more than 55dB weaker than at the entrance. At about this point it fell below the FM threshold of the receiver and therefore was unusable further along the tunnel. We thought at first that the presence in the tunnel of rails with rolling stock on them might have enhanced the propagation at VHF; but the signal strength, as we said, started falling off about 100 metres into the tunnel, while the rolling stock and rails it was on extended to nearly 150 metres into the tunnel. This suggests that the presence of all this metal had only a minor effect on the propagation. We would welcome comments from experts in wave-guide matters as to whether we are on the right lines in our reasoning.

We then tried an earth current system on the surface as before, but on 145MHz. The earth spikes were 2 metres apart and fed by a pair of wires about 2 metres long, connected to the transmitter via a 4:1 balun. The transmitter output power was 1W. This station could just be received in the tunnel, provided the tunnel set was below the surface station and had its rubber duck pressed firmly against the wall of the tunnel.


From these results we conclude that the best choice of band for working in tunnels is likely to be UHF. Second choice is an LF induction/earth current system. VHF seems to be of somewhat marginal performance, and mobile telephones do not seem to be much use in tunnels. If anyone else is thinking of doing such tests, they must obtain permission from the railway company concerned. We found the Nene Valley Railway very helpful and even brought in two cheerful volunteers to provide refreshments. The exercise did, however, cost us a 100 donation to their funds, which was met in part by contributions from some of the participants and in part by sponsorship from The Word Factory, Eyke, Limited.

If anyone would like to know more about these tests, please email me.