Broadly speaking, a need existed for two types of switch – those which operated instantaneously for use, for example, in booby traps and those whose action was delayed for a predetermined time. In view of the likely conditions of SOE use, the following principles were laid down for the design of explosives initiating devices.
1) They should incorporate a safety pin of the interlocking type such that it could not be removed if the device was in a dangerous state. In other words, if the device had been accidentally tripped and the striker was prevented from initiating the explosion solely by the safety pin, that pin could not be withdrawn.
2) Where appropriate they should incorporate a suitable arming delay.
3) They should be weatherproof and in some instances waterproof to a specified hydrostatic pressure.
4) They should withstand storage and operation under climatic changes in temperature, pressure and humidity from Arctic to Tropical zones.
5) They should be able to withstand rough handling resulting from air dropping or during operational use.
6) They should be based upon mechanical and not electrical firing.2
‘BOOBY TRAP’ SWITCHES
The main instantaneous switches which were available, or shortly to be available to SOE at the outset, were: (i) the pressure switch, which responded to a weight being placed on its hinged metal plate by releasing a striker to ignite a detonator, (ii) the pull switch intended to be inserted in a trip wire, and (iii) the pressure release switch (or anti-disturbance switch) which operated when the load on the case was released. Minor improvements to these switches were made by Stations IX and XII to increase their reliability and safety, and in some cases they were modified so that they could be incorporated into some of the camouflaged devices produced by Station XV.
Methods of disrupting road transport were high on SOE’s agenda. A simple device for disrupting light motor traffic was an adaptation of the medieval ‘Calthrop’ which would be scattered on roadways or aircraft runways. In this, three sharpened triangular knife edges are welded together to form a tetrahedral unit which, when dropped on a flat surface, always presents at least one sharp point upwards. In addition, the SOE Double-Bladed Dagger Jackknife incorporated a hooked tyre-slashing blade.
For attacking heavier traffic an explosive tyre burster was developed. In this, a ring of explosive surrounded a pressure switch, all within a 2 in diameter container made of two telescopic parts ¾ in high. A force of 150 lb (68 kg) caused the mechanism to operate. Station XV had great fun incorporating these items in simulated animal droppings appropriate to the country concerned.
The tyre burster also formed the basis for two improvised anti-tank mines described in the Technical Review No. 1. The first consisted of a cardboard box filled with PE or 808. The lid pressed on the end of a vertical dowel rod, the end of which rested on a tyre burster on the bottom of the box. This operated at a load of 150 lb (68 kg) and was intended for use against lightly armoured vehicles. A more powerful device containing a much larger charge of PE was contained in a square box, at the four corners of which were placed tyre bursters supporting the lid. This operated under a force of 600 lb (272 kg) – a man could even jump on it safely. After preliminary trials at Station IX on plasticine filled mock-ups, a full-scale trial at Farnborough in the summer of 1943 demonstrated its potential.
RAILWAY SWITCHES
Other devices designed to operate immediately on activation included a modification of the railway ‘fog signal’ used extensively in civil life in foggy conditions to warn train drivers of a red signal ahead. They operated when crushed by the leading wheels of the locomotive. Since in SOE they were intended to be used to blow up a section of track ahead of the train, the method of use was to employ the fog signal to detonate a length of Cordtex which in turn detonated the main charge some metres in front of the train. For SOE use some modifications were needed to improve the clips which held the device to the rail to avoid the signal being pushed along the track and pulling out the link with the detonating fuse. Various layouts were devised by individual Country Sections to suit local situations and sometimes involved the use of several signals linked together to overcome the possible misfire of a single device.
Alternatively, a rail charge could be set off using a pressure switch under the rail. This had several advantages over the fog signal because it was relatively inconspicuous and easily concealed. However, a solid base had to be established by clearing away ballast and the gap between the pressure plate and the underside of the rail had to be bridged by an adjustable extension rod. This was a somewhat fiddly job not easily performed in a hurry in the dark. Account had to be taken of the variable deflection of the rail depending on the solidity of the base and the weight of the train. Various modified versions of this switch for railway use were developed and tested. Despite their disadvantages these switches were used extensively in France, especially in the period after D-Day to delay the re-supply of forward German troops. It is recorded that the Resistance achieved over 950 interruptions to the railway system during the single night of 5/6 June 1944.
Other devices which operated instantly on activation included the altimeter switch which could be set to fire when an aircraft reached a predetermined height. There is, however, no reliable record of this having been used operationally. It is argued that the attempt on Hitler’s life, when a bomb disguised as two bottles of Cognac was put aboard his plane but failed to detonate, might have been more successful if an altimeter switch had been available to the dissident German officers. A photoelectric switch which responded to the switching on of a light was also developed and tested but was not put into production. The OSS is reputed to have used a similar device to blow up trains as they entered tunnels, thus adding to the effectiveness of the simple charge.3 A number of anti-disturbance switches were incorporated in camouflaged devices produced at Station XV.
THE IMBER RAILWAY SWITCH
One of the variations of the Railway Switch was the Imber Switch, an ingenious, mechanical device originally conceived by Imber Research and further developed at Station IX, which conducted trials on it at Longmoor in July 1943. Because the Germans were in the habit of searching the track ahead of ammunition trains and other likely targets, there was not much time for saboteurs to lay their traps between the inspection and the arrival of the train. So Imber came up with a pressure switch which could be activated by up to the eighth train to pass over it, which allowed much more time to set it in place. It would not have been too difficult to arrange the explosion on the eighth time the pressure rod was depressed by the deflection of the rail, but this might have been by the eighth axle of the first train. This switch was therefore designed to distinguish between axles and whole trains.
The Imber Switch measured 3 in × 2¼ in × 6 in high when the telescopic actuating rod was extended. Like the basic Railway Switch, it was buried in the ballast and the rod length adjusted until it touched the underside of the rail. Because this switch could be set for any subsequent train up to the eighth, reliable information on the order and contents of trains was essential. When the first axle of a train passed over it the rail movement pushed down the rod, which indexed a ratchet wheel, which, if this was not the target train, would initiate a delay in allowing the rod to return to its former position. This was sufficient to ensure the entire train had passed before the switch reset itself. In early versions this delay was achieved by a dash-pot arrangement. Experimental work revealed that at the high temperatures to be expected in the Far East the dash-pots became unreliable and so a mechanical ‘clutter mechanism’ was designed for it. This was similar to the escapement in a clock and had the effect of retarding the return of the actuating rod. Linked to the ratchet wheel on a cross-shaft was a disc which had a cutout in its edge. When this cutout came round to the relevant position on the arrival of the target train, a sprung hammer was released to set off a percussion cap on the end of the Cordtex fuse.
An example of this switch is in the Natio
nal Railway Museum in York.
TIME DELAY SWITCHES
Of major operational importance were switches which could be set to operate after a predetermined time delay. The most accurate were clockwork delays, but these were more expensive to produce and generally not robust enough to withstand rough treatment. Having the disadvantage that their tick could reveal their presence, they were used only on special operations where accuracy of delay was paramount and appropriate means of camouflaging them could be taken. There are frequent references in trials reports and diaries to the ‘Eureka’ clock. Its exact nature is not known but tests of its robustness were carried out successfully. It may have been simply a carefully designed time delay fuse. Alternatively, it may have been intended for use in conjunction with the Eureka/Rebecca navigational device (see Chapter 11) to arrange for the ground station of this equipment to be switched on at previously arranged times.
Development of this device was probably in the hands of Lt Col H.H. King, a retired officer and keen amateur watchmaker, who worked away unobtrusively in a small office in Station IX. It is not known whether any of his models were ever used in the field but after the war he was given the credit for the ISRB Allways fuse, details of which have not been found.
The need for a simple, inexpensive delay which could be readily produced in thousands was met by two rival models: the Time Pencil and the lead delay. The latter was a product of MD1 and was never widely adopted by SOE, who employed almost exclusively the Time Pencil.
THE TIME PENCIL (SWITCH NO. 10)
The basic concept of the Time Pencil originated in a German device invented in 1916. It was later developed by the Poles and Gubbins brought back a sample from Poland in 1939. It was taken up by Cdr Langley at Section D and further developed and improved through the years. Its final form is shown on p. 8 of the plate section. The striker was held back by a steel wire kept under tension by a strong spring. Its operation depended on the thinning and failure of the wire by a corroding solution contained in a thin glass ampoule which was brought into contact with the wire by crushing the bulb. When the wire broke the striker was released and operated a percussion cap attached either to a short length of Bickford fuse or to a detonating fuse. The rate of corrosion and hence the time delay was controlled by changing the composition of the solution.
In most descriptions of the Time Pencil the solution in the ampoule is said to be ‘an acid’: but there is evidence that work commissioned by Langley and carried out by Bailey at the British Scientific Instrument Research Association (BSIRA) led him in 1939 to choose an aqueous solution of copper chloride as the most suitable corroding fluid. This acts electrochemically, and is easily demonstrated by dipping a piece of iron in a copper solution when copper is deposited on the surface of the iron and a corresponding amount of iron is dissolved. The time delay was controlled by the concentration of the copper solution or, for longer delays, by the addition of glycerol which, because of its higher viscosity, reduces the rate of corrosion. The glycerol had the advantage of acting as an anti-freeze; but on the other hand the increase in viscosity at lower temperatures led to a considerable increase in the delay time. This feature probably led to claims from the field that Time Pencils did not work under freezing conditions.
Five compositions coded by the colour of the safety strip became standard, covering the nominal ranges of: 10 min – black; 30 min – red; 5½ hr – green; 12 hr – yellow and 24 hr – blue. These timings were dependent on the temperature and the nominal values were stated to refer to 15°C, but even at the same temperature they were found to exhibit undesirable variations in the delay times. Unreliability in the case of the shorter nominal delay times could have devastating effects. For example, in July 1943 an agent complained bitterly that a red Time Pencil with a nominal 30 minutes delay had operated in just three minutes. Such incidents tended to give the Time Pencil a bad reputation among agents. No reliable statistics were available to enable the seriousness of the problem in the field to be quantified but the situation was regarded with great concern and a major programme of research was started urgently at Station IX in March 1943 to try to identify the origin of this variability. To make the experiments statistically significant at least 15 replications were made in each set of experiments. As it was impractical to carry out these tests on actual Time Pencils, they were done on laboratory replica rigs. At first very elementary methods were used to obtain the timings. For shorter times, kitchen alarm clocks were lined up on a shelf and the experimenter had to keep watch and record the time at which each operated. For longer times where it was not very practical to have the experiments kept under observation overnight, the breaking of the wire was arranged to drop a pin into the works of the clock and so record the time at which the wire failed and the clock stopped.
It was typical of the practical problems encountered that, because for some time the laboratory did not possess a refrigerator, it was not possible to test the effect of temperature on these timings! Later, a well-equipped testing laboratory fitted with an array of electric clocks and including constant temperature chambers controlled between – 40 and +60°C was set up by van Riemsdijk (see plate section, p. 8). This was used for the routine testing of batches of production pencils.
In all, over 700 experiments were carried out during the summer months of 1943. The statistical analysis was made more tedious by the fact that Station IX did not possess a single calculating machine: use was made from time to time at weekends of a hand-operated Brunsviga machine at Oxford. Among the many factors affecting the delay time was the composition of the corroding solution. The picture which emerged was rather complex and not easily understood. For example, it was found that, contrary to expectations, an increase of concentration of copper chloride led to an increase of the delay time from about 20 minutes to 80 minutes. On the other hand, when copper bromide replaced the chloride, an increase in concentration decreased the delay. Copper nitrate and copper sulphate proved to be much less corrosive, but again, increase in concentration (as with the chloride) lengthened the delay. A wide range of timings could be obtained by using a mixture of copper chloride and copper acetate. To avoid the use of glycerol a series of experiments using n-propyl alcohol instead enabled a wide range of timings from 30 minutes to 15 hours to be achieved.
At this stage the work was beginning to take on the character of a more academic study of corrosion science and further work was carried out on an extramural contract with the Physical Chemistry Laboratory at Oxford under Dr J.G. Davoud, who later joined Station IX. Although it might have been possible at this stage to change the solution composition to take account of the work carried out, production of the pencils would have been interrupted. This would not have been justified as the reproducibility of the results was only marginally better than that of production pencils. The standard deviation of a single result throughout all the laboratory work was about ± 20 per cent, comparable with that for production pencils, i.e. there was a 95 per cent chance of an individual pencil operating within 20 per cent of its nominal value. Statistical analysis gives no information on the likelihood of rogue results arising from one-off factors. In the experimental programme only 5 out of 700 led to a rogue result. An investigation was undertaken to try to identify other factors which might affect the performance. One factor which was identified was the cleanliness of the wire, since traces of grease on the wire as supplied lengthened the time delay. There were also occasions when the wire supplied by the contractor was already starting to rust! Difficulties also arose when a shortage of supplies meant that different sources of wire had to be used. In particular, wire imported from the USA exhibited characteristics different from those of the British material and the solutions had to be varied to take account of this. It was not clear whether in the production process sufficient care was taken to ensure that the tension on the wire was constant, nor whether attention was paid to ensuring that the screw holding the wire was properly tightened. Major failure of the switch to operate
correctly could have been caused by one or more of these factors. All of this pointed to the vital importance of rigorous Quality Control (QC) by the manufacturers and by Station XII, and to the need to tighten up the specification. This was done in collaboration with Maj Bedford and Capt Ault at Station XII. In all, some twelve million pencils were produced under the control of Station XII.
SOE was not the only organisation experiencing problems with devices equipped with Time Pencil-type delays, although in one case it was not the delay itself but the firing mechanism which was at fault. The failure of a German version of the device (or a captured SOE one) probably prevented a change in the course of the war. On 13 March 1943 a group of disaffected German officers made an unsuccessful attempt on Hitler’s life when he visited the Army Group Centre at Smolensk. Over lunch, Col Henning von Tresckow asked one of Hitler’s party if he minded taking two bottles of Cointreau back to headquarters on the return flight, explaining they were part of a bet he had made with another officer. After Hitler had boarded the aircraft the 30-minute delay was crushed to initiate it and the package handed to the aide. The bomb failed to explode, the reason never being fully explained. With astonishing coolness, Tresckow telephoned headquarters to tell them to hold on to the package as there had been a mix-up. Another officer on the daily courier flight retrieved it the next day and substituted a genuine package. On examination it was found that the capsule had broken, the solution had eaten through the wire retaining the firing pin, and the firing pin had struck the percussion cap which seemed to have ignited. But the charge had not exploded. One theory was that the heater in the aircraft’s hold had malfunctioned, a not unusual event, and as a result the explosive, which was sensitive to cold, failed to react.4
SOE Page 8