Before 1914 navies typically estimated future range by applying an estimated range rate to a clock (the widely-used type was the Vickers Clock).18 The Royal Navy’s Dreyer Table was a sophisticated version of this practice. Ranges provided by rangefinder operators were automatically indicated (by pricking) on a moving paper. The corresponding range rate was the slope of a line drawn through the range plot. It could be set on a range clock, the outputs of which were indicated by an automatically drawn line of estimated ranges, which could be compared to measured ranges indicated on the plot. The comparison made it possible to apply corrections, which were necessary because actual range rates were almost never constant. A Dreyer Table generally also incorporated a bearing plot, but range and bearing data were not combined in any way. The Dreyer Table was conceived as a means of assisting a fire-control officer observing fire, indicating how he might correct range.19 Unfortunately the line on the plot could not really be straight, because range rate varied with range, so a Dreyer Table solution would eventually fail.
The range plot had an unexpected virtue. British capital ships had multiple rangefinders, which gave differing data. When their outputs were all indicated on the same range plot, an operator could visually average rangefinder data and he could see at a glance if one set of data were clearly in error. The plot was valued as a quick indication of the gunnery situation: ranges, estimated firing range and spots.
The Germans seem to have been aware that the British were relying on a plot-based clock; their tactical countermeasure was to zig-zag. That created sudden changes in range rate, which the Dreyer Table could not follow. The Germans’ own clock-based system, which did not employ a plot of any kind, seems to have been better adapted to such manoeuvres. At the end of the First World War the Grand Fleet Dreyer Table Committee concluded that it was pointless to rely on range rates. The most valuable feature of the Table turned out to be its bearing plot, which could detect sudden changes of target course.
While the Dreyer Table was being developed, the Royal Navy pursued a much more sophisticated concept: to create a model (an analogue) of the engagement, separate elements representing shooter and target. Once set with target course, speed and initial range, the model generated (computed) current and future target position, hence range and bearing. Current data could be compared against observed reality and target course and speed adjusted until they matched. The resulting solution remained valid until the enemy manoeuvred. Then it could recover faster than a rate-based device. All major navies used this approach during the Second World War.
It was invented by Anthony H Pollen, who developed his Argo Clock computer under a monopoly agreement with the Admiralty. In 1912 the Admiralty planned competitive trials between the Argo Clock and the Dreyer Table, but abruptly cancelled them. Five ships equipped with Argo Clocks for the trials retained them throughout the First World War: Queen Mary and the four King George V class battleships. First Lord Winston Churchill told the Commons that the navy had found a less expensive and superior alternative to Pollen’s expensive device, but it appears that the Dreyer Table was bought as an inexpensive interim device while an alternative analogue computer was developed by Barr & Stroud.
Money was very tight. Barr & Stroud depended heavily on Admiralty business for its main product, rangefinders, so would have been more amenable than Pollen to cutting the price of its computer. That the British Government did not think that war loomed would have made purchase of the less effective Dreyer Table perfectly acceptable at the time. War intervened before Barr & Stroud could complete development. The monopoly agreement with Pollen made the Admiralty’s manoeuvre embarrassing at the least.
Given war experience, the Admiralty chose to develop its own computer (AFCT) based on Pollen’s ideas and some of his technology.20 Associated with the new AFCT was an aloft DCT slaved to the computer, providing the feedback necessary to obtain a good fire control solution. The Royal Navy adopted massive tower bridge structures after the First World War specifically to accommodate the heavy new DCTs. It also developed means of moving DCTs and turrets under the control of the AFCT, a major achievement.
Because the war had slowed development, Barr & Stroud could not offer the Admiralty an alternative to the AFCT. It had to be content with the export market. Its system became the basis of post-war Italian and Japanese fire control and, it appears, German fire control based on an Italian prototype. In this sense Bismarck sank HMS Hood using a more sophisticated British system against a less sophisticated one – a horrible own-goal.
The new analogue technique was far more automated than the system in which the Dreyer Table had been embedded. That, much more than the technicalities of the new computer, may have been its greatest virtue. It required far less training and it got onto a target much more quickly. As one of the five test ships equipped with Pollen’s computer, HMS Queen Mary reportedly made the best shooting among the battlecruisers at Jutland. Since the battlecruisers’ problem seems to have been very few opportunities to practise, her superiority may reflect the difference between the heavily human element in a Dreyer Table system and the far more automated approach represented by Pollen’s computer.
This automation probably explains why the new battleship HMS Prince of Wales, with a nearly untrained fire-control crew, performed so well when she faced the Bismarck (and, for that matter, why the raw fire-control crew on board Bismarck did so well in her initial battle). At least in theory, an automated system made it possible to hit on the first or second salvo. Using the much less automated Dreyer Table, the gunnery officer of HMS Hood remarked to a US officer that he could imagine not getting onto the target until he had fired several salvoes. Hood’s gunnery officer assumed that he could absorb damage while his higher rate of fire and superior shells destroyed his enemy. Hood was just getting onto her target when she was destroyed. The surprise was not that Bismarck hit Hood so quickly, but that one salvo destroyed her.
Armour
British designers indicated armour and other steel thicknesses not in inches but in pounds, on the basis that a square foot of 1in steel weighs about 40lbs. Thus a 320lb plate was nominally 8in thick. In fact steel is slightly heavier (40.8lbs per square foot of 1in steel), so a 240lb plate is actually 5.88in rather than 6in thick and a 360lb plate 8.82 rather than 9in thick. During the 1930s, when adherence to strict Treaty limits was particularly important, constructors sometimes referred to ‘light rolled’ plates – presumably plates rolled to a specified weight rather than to the thickness implied by the nominal weights.
The British capital ships described in this book used Krupp Cemented (KC) or face-hardened side armour, which had displaced earlier types because it was so much lighter for a given degree of resistance. As an indication of how much difference KC made, the Admiralty Gunnery Manual (Vol I: CB 142) of 1915 equated 5¾in of standard Krupp armour to 12in of all-steel armour and to 15in of wrought iron armour when resisting uncapped projectiles. By 1919 the British called cemented armour simply C armour, their version being superior to the original Krupp type. Because it produced a layered plate, cementing could not be applied below a particular thickness, which was 4in in 1928.21 At that time the standard for 15in C armour was to resist a 16in shell at a striking angle of 30° and a velocity of 1530ft/sec, corresponding to a range of 13,200 yds. The 13in and 14in plates were to resist 15in shell at, respectively, 1480 and 1560ft/sec (striking angle 30°), corresponding to ranges of 17,500 and 15,200 yds, respectively.
The cementing (carburising) process produced an extremely hard surface layer, a relatively deep hardened layer and a tough rear layer which absorbed the shock of impact and protects the plate as a whole from being punched through by the broken face. The hardened face was intended to so damage the attacking shell that it was no longer effective. It was most effective when the shell struck nearly at right angles (‘normal’) to the plate.22 During the run-up to the First World War the British and others improved KC steel by ‘normalising’ it, somewhat reducing surface hardness to toughen the plate. Krupp had concentrated on hardening the face of the armour at a cost in toughness. Normalisation improved resistance to capped shells. Krupp did not