Читать книгу Naval Anti-Aircraft Guns and Gunnery - Norman Friedman - Страница 13

CHAPTER 4

THE INTER-WAR ROYAL NAVY

Background

Once it had defeated Germany, the Royal Navy’s outlook on future war changed dramatically. With minor exceptions, from about 1919 on attention focussed on Japan. Although the Japanese had proven helpful during the First World War, they had also shown an appetite for the British Far Eastern empire, as well as for the informal or commercial empire in China. Ideally the British would have stationed a powerful fleet in the East as a deterrent, but the available infrastructure was far too limited. The most ambitious plan was to send the battlecruisers east in 1929, and that had to be cancelled. Instead, the fleet had to be split between the Mediterranean and Home waters. In a crisis the Mediterranean Fleet would steam to Singapore, which was thought to be well beyond the range of Japanese land-based aircraft. Singapore would be the Scapa Flow of a Far Eastern war, the fleet there either forcing the Japanese to remain at home (as the Grand Fleet had bottled up the German High Seas Fleet) or fighting a decisive action. Ideally the fleet would steam to Singapore as tensions rose but before war broke out.

As in the US planning described in the next chapter, the decisive stage in the war would be a close blockade of Japan, which depended almost entirely on imports for survival. To mount the blockade, the British would have to destroy the Japanese fleet in a decisive action. Because the British expected to fight far from any Japanese land bases, they could imagine that air attack during the run-up to the big battle would be extremely limited (the US Navy was in a different position). Any attacks would be mounted by shore- or sea-based long-range aircraft, at ranges which would automatically limit the scale of the attack. Thus the main threats to the fleet prior to the fleet battle would probably be Japanese submarines and Japanese-laid mines. Mass air attacks would be limited to carrier air strikes mounted either just before or during the fleet action.

In 1935 the situation was transformed. British faced a real threat of war against Italy during the crisis over Abyssinia. Massed long-range land-based Italian aircraft were suddenly a credible threat, as was the large number of Italian motor torpedo boats. It was impossible to deal with land-based aircraft by a pre-emptive strike (which might, for example, eliminate a threat further out to sea by sinking the enemy’s carriers). The motor torpedo boats affected British anti-aircraft thinking because light anti-aircraft guns were conceived both to fend off close-range air attacks and also to deal with motor torpedo boats. British understanding of the air threat at the time of the 1935 crisis naturally reflected the recent committee report, and much of what was done on an emergency basis responded to its recommendations. The exception was the decision to convert old light cruisers to specialist anti-aircraft ships, a step the committee had explicitly rejected.

For the inter-war Royal Navy, the key facts of life were limited carrier capacity, in terms of numbers of aircraft, and the limited number of carriers prescribed by the Washington Treaty in 1921. Aircraft carrier capacity was set, the Royal Navy thought, by hangar capacity; aircraft were not to be kept on the flight deck after they landed. There was some interest in the US style of operation about 1931, but by that time it seemed that a new arms limitation treaty might control overall numbers of military aircraft, and the RAF stoutly resisted any increase in rated carrier capacity. This equation explains why Ark Royal was designed in 1933 with two hangars. The Royal Navy adopted deck parks only during the war, and it largely abandoned them after it. The Royal Navy was well aware of what air strikes could achieve: it wanted as many strike aircraft as it could get. In the early 1930s it fastened on a solution: cruisers and battleships would carry strike aircraft armed with torpedoes. The same type of aircraft would fly off carriers. This combination of requirements created the Swordfish. It needed maximum lift to fly from a catapult with a 60kt end speed carrying a torpedo, for the strike role. It needed maximum scouting range because the Royal Navy lacked a long-range shore-based air arm like the flying boats of the US and Imperial Japanese navies. That favoured the relatively low-powered engine available when the aircraft was designed (the 1000hp class engines suited to tactical aircraft appeared about two years later). The Swordfish was designated a TSR, a torpedo-spotter-reconnaissance aircraft. In theory, then, it could find an enemy fleet, slow it enough (using torpedoes) to allow the slower British battle line to catch up, and then support the gunnery action by spotting for the capital ships. The Fleet Air Arm motto was ‘Find, Fix and Strike’, and the British concluded after Jutland that enemy battle fleets might well try to evade their own slower but very powerful battle line. Fixing meant slowing them enough that evasion would become impossible (the new heavy shells issued after Jutland were intended for much the same purpose, to disable any German capital ship machinery they hit badly enough that they could never escape again). The low take-off and landing speeds of the Swordfish turned out to be a great virtue when it operated from small escort and merchant aircraft carriers, and it was one of the few combat aircraft to operate throughout the Second World War. These Swordfish, photographed in 1944, carry underwing rocket rails with which to attack U-boats. (David Hobbs)

Queen Elizabeth was subject to a more complete modernisation than Warspite, the main difference being that she received a uniform secondary battery of 4.5in guns in the new BD mounting which had been under development since the 1920s. The turret-top guns are twin power-driven Oerlikons. Unlike later battleships, she could not accommodate more pom-poms in addition to the four installed on modernisation. She is shown off Hampton Roads after a US refit, 2 June 1943.

Throughout the inter-war period, the Royal Navy tried to maintain a modern fleet by adding weapons and equipment on a step by step basis. The basis of anti-aircraft fire control, the HACS, came first, because once it was in place medium-calibre guns could be added as required. Photographed in 1934, Ramilles shows what was done before the major fleet rearmament begun in 1936. She has a HA director atop her foretop, an octuple pompom (on the platform abeam her funnel, under the two searchlights), and four 4in guns Mk V. These upper-deck guns were relatively easy to replace later with twin mounts, because neither they nor the twins was power-driven or penetrated the upper deck. The major pre-war upgrade to this and similar ships was to replace the single guns with twins.

Warspite was subject to an elaborate modernisation which provided her with the standard automatic battery of a modern capital ship: four octuple pompoms. She also had sided HA directors, so that she could engage attacks from both sides simultaneously (less modern battleships with a single director could not do so). She was also given the new quadruple 0.5in machine guns, which are visible atop her superfiring turrets. The coloured stripes on ‘B’ turret were added during the Spanish Civil War, to indicate to the combatants that she was part of the neutrality patrol. Note that the pom-poms were not shielded.

Once it accepted that war in European waters was possible (in the mid-1930s), the Royal Navy took steps to provide air protection to shipping, in the form both of new escorts (sloops and ‘Hunts’) and conversions of older warships. The air part of a sea war in the Pacific would have been fought, it was imagined, far from shipping routes. European trade routes were all within easy range of land-based aircraft. The potential of air attacks against shipping was demonstrated during the Spanish Civil War, 1936–9. Coventry was the first, armed with 4in Mk V guns taken from ships being rearmed with twin 4in mounts under a major fleet upgrade programme. She is shown in 1940. The shrouded object in ‘B’ position is a multiple pom-pom.

Although war with Italy was averted, European war was increasingly possible, presenting a horrific nightmare in which Britain might have to fight Japan and a European power (probably Germany and possibly Italy as well) at much the same time. Throughout the late 1930s, the Admiralty’s only solution was to fight and defeat the Japanese fleet first, then swing the British fleet back to Europe to face the less powerful fleets there. The opposite sequence did not bear thinking about, because war in Europe would exact attrition and might also occupy so much of the fleet that it would be badly outnumbered in the East – which was exactly what happened.

In 1937–9 the Spanish Civil War demonstrated that in European waters aircraft were an effective anti-shipping weapon, perhaps on a par with (or more dangerous than) submarines, because land-based aircraft would often be within range of shipping and because Asdic (sonar) was thought to have largely solved the submarine problem. From 1937 on considerable effort went into converting warships (and preparing to convert merchant ships) specifically to defend merchant shipping against air attacks. This was apart from the threat, which may not have been appreciated before the outbreak of war, of air-laid mines in harbours. Only the Royal Navy seems to have paid attention to shipping protection against air attack during the run-up to the Second World War. The other major navies concentrated on the air threat to warships.

Anti-aircraft is more than defensive guns. Given limited carrier capacity and a much larger Japanese carrier force, how could the Royal Navy protect itself? The Blackburn Skua was one answer. By dive-bombing it could destroy or at least disable the Japanese carriers and thus gain air superiority. It also had air-to-air capability, but that was a distant second to dive bombing. This Skua is diving with its dive brakes open, in an attitude which would later be called glide bombing. (David Hobbs)

King George V shows the standard anti-aircraft outfit as applied to a first-line battleship at the outbreak of war: a fully dual-purpose secondary battery (in this case of 5.25in guns) and four octuple pom-poms. This was more powerful than that of any other navy: the US standard automatic battery at the time was three or four quadruple 1.1in guns and eight single 0.5in. As in the Queen Elizabeths, all four pom-pom mounts were concentrated in her forward superstructure, each with its own director in a small tub at a higher level. No pom-poms or other anti-aircraft weapons were mounted atop the after superstructure, which at this stage was used for the ship’s boats. The Royal Navy view was that to retain its mobility the fleet had to carry its own ships’ boats. The US Navy set up boat pools at its bases, which enabled it to spread more light anti-aircraft guns over its ships. The Y-shaped structure carried two HA director towers which controlled the twin 5.25in guns on the deck below. Around the base of the main battery director are anti-aircraft lookout sights, used to indicate incoming targets to the Air Defence Officer – a feature unique to the Royal Navy at this time. This air defence platform, open to the sky, was adjacent to the open upper bridge (compass platform) on which the ship’s officers stood, so that the ADO could have immediate access to the ship’s CO. No lighter weapons are evident in the photographs. The design originally called for four quadruple 0.5in guns, but they were apparently eliminated when the turret-top pom-poms were added. By February 1940, plans called for six pom-poms, the other two atop ‘B’ and ‘X’ turrets (rockets were mounted instead when the ship was completed, due to a shortage of pom-poms). Both ships armed with rockets (King George V and Prince of Wales) had them replaced with octuple pom-poms when the latter became available in sufficient numbers later in 1941, and Prince of Wales also had a single hand-operated Bofors on her quarterdeck at the time of her loss (it continued to fire after she lost power). King George V is shown on the voyage to the United States in January 1941 on which Ambassador Lord Halifax was taken to Annapolis, the closest port to Washington. (Naval Institute Collection)

Until about 1936, the Royal Navy expected the major air threat at sea to be against its fleet. Since destroyers were agile, they were unlikely to be singled out for attack, which would be concentrated on capital ships. It followed that a destroyer gun elevating to 40° could usefully support the capital ships. The ‘Tribal’ class was designed specifically for such support, with four twin 4.7in guns capable of 40° elevation. The subsequent ‘K’ and ‘N’ classes had the same gun. An RAN destroyer displays her forward twin 4.7in guns in the Southwest Pacific during the Second World War. (State Library of Victoria)

Caledon, shown here on 1 February 1944, typified the production version of the anti-aircraft (shipping protection) cruiser, armed with standard guns (such as the twin 4in). In 1940 the US naval attaché in London reported that anti-aircraft cruisers were ‘worth their weight in gold’. The US Navy accordingly began to plan a comparable conversion of Omaha class cruisers, to be armed with 5in/25s, but the outbreak of war precluded it.

Initial Approaches to the Fire-Control Problem

The Royal Navy committees to capture wartime lessons included a Naval Anti-Aircraft Committee convened in 1919.¹ The anti-aircraft committee decided that all of its experiments should employ director firing (a new departure for air defence) because this method ‘presents unquestionable advantages over individual firing: greater accuracy, greater rapidity, more continuous firing and simpler control’.² It was not clear whether an anti-aircraft director should incorporate a rangefinder; in 1919 the Royal Navy was integrating its surface directors with rangefinders because wartime experience suggested that otherwise fire-control systems might suffer when rangefinder and director pointed at different targets. For the moment, the Committee was willing to use separate director and rangefinder for experiments. In 1924 arrangements were made to install high-angle directors and separate rangefinders on board capital ships.³ The complexity and awkwardness of existing anti-aircraft fire control made it obvious that calculation should be automated. In addition to aiming guns, the fire-control computer had to calculate fuse settings. Other navies reached the same conclusion.

By this time the RAF had taken over the wartime Royal Naval Air Service. Although not a significant factor in 1918–19, attempts by the new Royal Air Force to shoulder the navy aside became an important factor in Royal Navy views on naval air defence (among many other things). Leaders of the new RAF (and, for that matter, of air services abroad) liked to claim that their inexpensive force made expensive surface navies obsolete. Thus on several occasions between the wars the Royal Navy found itself compelled to explain to the Cabinet why it considered large warships, particularly battleships, viable in the face of air attack. The navy’s need to take air attack into account may explain why the Royal Navy was better equipped for air defence than its contemporaries. In the 1930s the Royal Navy was well on the way to having the heaviest existing anti-aircraft batteries, including the world’s heaviest batteries of light automatic anti-aircraft weapons. War experience was to show that even these batteries were insufficient against the new threat. It was a great wartime disaster that, having provided so much anti-aircraft firepower before the war, British warship designers had not provided much margin for further increases.

Because it believed that several aircraft would have to be devoted to any first-line warship target (using level and torpedo bombing), the Royal Navy assumed that attacks on the fleet would concentrate on high-value units such as battleships and, to a lesser extent, cruisers (slow merchant ships were another story). On this basis the Royal Navy could afford not to develop a dual-purpose destroyer main battery. As the air threat escalated, it seemed to be enough for destroyers to support heavy ships with relatively low-angle fire – the Royal Navy classified the 40° elevation guns aboard ‘Tribal’ class destroyers as fleet anti-aircraft assets, and it developed an anti-aircraft fire-control system for such ships. It is not clear to what extent the inter-war Royal Navy saw carrier-borne fighters as part of fleet air defence. Its carriers worked in close formation with its battleships, and until the late 1930s it was reasonable to imagine that screening destroyers would spot enemy aircraft in time to launch defending fighters (as the contemporary US Navy imagined). That was why the Skua was expected to assist in fleet air defence by destroying the threat at source, rather than (usually) defending the fleet directly. Once a European war seemed to be looming, neither attack at source or direct defence seemed workable. Ironically, that was exactly when radar appeared. It solved the air-warning problem, and thus made direct fighter defence entirely workable.

While it was developing the HACS, the Royal Navy deployed an interim system employing a simple director and calculating instruments. It occupies the upper control level in this early 1927 photograph of the new battleship Nelson. The director is the thimble-shaped structure at the fore end of the platform. The earlier battleships were similarly fitted.

Targets

Like the other major navies, the Royal Navy did not experience air-sea warfare between 1918 and 1939. How well it prepared for the war it fought after 1939 depended in part on how realistically it could test anti-aircraft fire control and guns. Most targets were inexpensive but also unrealistic: puffs of smoke, gliders and towed sleeves (banners). The only targets which came close to realism were radio-controlled drones. The Royal Navy was the first in the world to employ them. In the early 1920s the navy began to launch radio-controlled aircraft, initially intended as anti-ship missiles, from the destroyer Stronghold.⁴ At this stage there was no real prospect of immediately turning the aircraft into realistic targets. It did not help that naval aviation was largely removed from the Royal Navy under the aegis of the RAF, because that eliminated the interaction between aviators and seamen which might have informed both. The navy did provide many of the aircrew of the Fleet Air Arm, but they did not go on to become senior naval officers. That affected the fleet’s expectations in air matters – including targets. Control by the RAF strictly limited the number of naval aircraft, precluding the use of numbers of expendable obsolete airframes as radio-controlled targets.

Work on radio-controlled targets seems not to have resumed until about 1930. In 1932 the RAF modified three standard fleet floatplanes (Fairey IIIFs called Fairey Queens) into radio-controlled drones. The first two were damaged beyond repair while being catapulted by the battleship Valiant, but the third was successfully launched in the autumn. It made a short trial flight, which demonstrated that it could be controlled in the air and landed on its float. Arrangements were made to use it as a Home Fleet target off Gibraltar during the 1933 Spring Cruise, and plans were made to develop a smaller and less expensive drone which could be maintained by RAF and Fleet Air Arm personnel.

The cruiser Sussex shot down the sole Fairey Queen late in May 1933 on the ninth salvo of Run 2. The ship had 48 per cent of her bursts within 100 yds during Run 1 and 59 per cent during Run 2 (100 yds was considered satisfactory). As a result of the Fairey Queen trials, the Air Ministry carried out further trials with radio-controlled De Havilland Moths which were called Queen Bees. Four were to be made available to the Mediterranean Fleet in June 1934, and four more to the Home Fleet in the autumn of that year. Unfortunately these targets could not simulate dive bombing.

Initial experiments showed that anti-aircraft effectiveness was much less than had been imagined. Reporting on Queen Bee firings in the Mediterranean, Vice Admiral 1st Battle Squadron pointed out that the standard of placing rounds within 100 yds of a target was misleading. In one exercise, although 59 per cent were within 100 yds, only 1.56 per cent would have scored actual hits. No more than a quarter of 4in shells within 100 yds would cause damage to an aircraft, and the other three-quarters should be scored as close misses. The fleet commander agreed; probably only th of these shells would damage an aircraft. Experiments by the all-service Ordnance Committee showed that a 4in HE shell had to burst within 70ft to cause decisive damage, and outside 140ft there would probably be no damage at all. The 100 yd standard had been adopted because it was impossible to determine the position of a shell burst any more accurately – the percentage within 100 yds was really no more than a way of assessing accuracy.

The more subtle message, illuminated by somewhat later US experience, was that with so few targets available, there was little interest in shooting them down. The fleet did not learn the lesson the US Navy learned from its own drones, that advertised shell lethality had been badly overstated. The typical test method, bursting a shell near an aircraft near the ground, was not at all good enough.

The navy pressed the Air Ministry to make targets more realistic, both by enabling them to dive and to operate them in pairs. In 1935 it hoped to modify the Queen Bee to dive at angles of 35–40° at 175mph, but it could not operate two of them in formation, to simulate a mass attack.

The navy certainly understood how different drone firings were from other forms of anti-aircraft practice, and it embraced the new technique. During 1935 a Queen Bee base was established at Vigie Creek, St. Lucia, and approval was given to recommission the old carrier Argus as a Queen Bee tender, to carry the Home Fleet Gunnery Co-Operation Unit as well as drones (she would also be the deck landing ship). No similar ship was provided to the Mediterranean Fleet because Malta offered sufficient facilities ashore. However, as aircraft speeds increased through the late 1930s, it was increasingly clear that the Queen Bee was not a particularly realistic target.

Even without drones, it was difficult to arrange realistic practices representing aircraft flying directly towards the firing ship, as would be the case in wartime. Progress in Naval Gunnery 1932 mentioned as an important development increased numbers of exercises in which bombers flew overhead rather than as crossing targets. The Home Fleet wanted 60 per cent of cruiser shoots and 75 per cent of battleship shoots to represent aircraft directly attacking the firing ship, and there was general agreement. Apparently the problem was safety requirements which limited firing close to a ship.

Fire Control

Probably the most important decision of the Naval Anti-Aircraft Gunnery Committee was to abandon the barrage fire adopted during the First World War in favour of aiming directly at the aircraft. Not only did theoretical calculation show that barrage was practically useless, but results in the war were ‘too slight to be counted on’. That required something which measured aircraft motion and predicted aircraft position. The Royal Navy developed a High Angle Control System (HACS) incorporating one or more aloft directors linked with one or more High Angle Control Positions (HACPs) containing the system’s computer, the High Angle Calculating Table (HACT). The HACP was analogous to the below-decks Transmitting Station (TS) used to control main battery guns; the HACT was analogous to the new Admiralty Fire Control Table (AFCT) and its derivatives. A policy for installing HACS in new and existing ships was promulgated in 1929.⁵

In June 1919 the Naval Anti-Aircraft Gunnery Committee formed a sub-committee to test new anti-aircraft fire-control devices.⁶ For trials, the director could be separated from the control (computing) system. The director and the associated rangefinder were the eyes of the system. They provided its inputs, and they would also be the basis of any feedback. It was not sure whether a director should incorporate a rangefinder, but decided that for trials purposes director and rangefinder could be separate. The committee wanted a means of detecting whether a target was flying level – by no means did it assume that it was. However, if the target did fly straight and level the fire-control problem was considerably simplified. Moreover, in 1919 the major air threats were level and torpedo bombing, so it made sense to concentrate on a target flying level.

Professor Sir James Henderson, the Royal Navy’s gyro expert, pointed out that the force required to keep a gyro pointing at a target (moving it from its preferred position) measured the rate at which the direction of the target changed. This measuring device was called an angle gyro. The committee liked Henderson’s design of a two-man director with stabilised layer’s and trainer’s sights. Given Henderson’s angle gyro, the committee favoured a tachymetric computer, which would rely on it to measure directly vertical and horizontal target motion (in terms of angles, not distances). It hoped that such a system could largely dispense with rangefinding, which was difficult and unreliable (given the need to set fuses, and the way in which target range affected gun elevation, it rangefinding could not have been avoided altogether). The Admiralty Research Laboratory (ARL), which developed anti-aircraft fire-control systems for the British army, produced a tachymetric system. So did Vickers.

The rub was that shipboard sights had to be stabilised. Otherwise the ship’s motion would be entangled with the aircraft’s apparent motion. High-angle firing required much better stabilisation than low-angle: the Master Gunnery Gyro being developed for surface fire was not good enough. Unfortunately trials conducted in 1926–7 on board the cruiser Dragon and the battlecruiser Tiger failed. It proved too difficult to maintain constant gyro speed (gyro speed affected the force needed to move the gyro), and existing gyros took too long to get up to speed. The gunnery school HMS Excellent proposed moving the gyros below, but nothing was done, probably because the stepping motors used by the Royal Navy to transmit data could not operate smoothly enough. Several foreign navies, including the US Navy, adopted tachymetric approaches based on gyro-stabilised directors.

During the 1930s the Royal Navy considered long-range fire the dominant means of air defence, so it emphasised sophisticated fire control, just as surface gunnery entailed elaborate means of control. At the very least, heavy anti-aircraft fire could force bombers to keep high and/or jink, which would ruin their aim. It could break up formations, rendering attacking aircraft more vulnerable to fighters and also ruining the cohesion flight commanders needed if they were to make many hits. Aircraft should be attacked before they could get close enough to attack.⁷ The Royal Navy thought that its investment in fire control and guns would enable it to destroy many attacking aircraft.⁸ War experience showed that actual destruction was rare, but that gunfire often accomplished the other goals. Lighter guns, which were more likely to destroy attackers – particularly dive bombers – became far more important, a reversal unexpected before the war.

A ship’s manoeuvres under air attack could complicate fire control. As of 1931 the Royal Navy view was that last-minute manoeuvres to evade bombs were seldom worthwhile. Large turns would ruin gunnery. Fire control (as described below) worked best if a ship followed a more or less straight path.⁹ However, it might be well worth while to manoeuvre upon sighting aircraft, both to bring maximum anti-aircraft fire to bear and to make it difficult for the aircraft to approach on a course giving it the best chance of hitting.

Rangefinding

The committee soon realised that rangefinding such a difficult major problem that it sought a system based entirely on angular rates. The ARL developed just such a tachymetric system for the British army, but as noted that was impossible on board ship. After tests of Barr & Stroud coincidence rangefinders and a German stereo type modified for anti-aircraft use, the committee reported in January 1920 that the army’s UB 2 coincidence unit ‘shows the greatest promise, and meets Naval AA requirements better than any other type’.¹⁰ UB 2 automatically read out target height based on sight angle. Differential gearing connected range and height-reading scales. If the aircraft flew at a constant height, and the rangefinder was kept ‘on’ the target, it read out change of range. UB 2 had a special presentation which apparently made it easier to get a ‘cut’ on any part of an aircraft. To get it onto a target more quickly it had four sets of open sights usable by range-taker and trainer, and also by an officer helping get the rangefinder onto the target. The general arrangement of the navy’s wartime FT 29 precluded open sights. Against an aircraft flying at a steady height of 3000 to 10,000ft, UB 2 could get about eight good cuts (ranges) per minute. This number was roughly halved when the aircraft manoeuvred evasively. Generally it took half a minute to a minute from sighting the aircraft to making the first cut (range measurement). Errors were about 300 to 500ft; cuts were best against higher-altitude targets because they appeared to the range-taker to be moving more slowly. UB 2 became the basis of future Royal Navy anti-aircraft rangefinders, beginning with UB 4, which had been adopted by 1925.

Given the discovery during the First World War that it was usually better to concentrate on target height, the new rangefinders were described as heightfinders. The committee decided that they should be mounted horizontally rather than vertically, to protect its crew. It turned out that if the appropriate rate of change of range was applied to the instrument, the eyepieces were held on a target flying level. Even if the target was moving rapidly, the cut would be less transitory, and an operator could get more frequent cuts. That made the range- or height-finder a potential means of measuring a fire-control solution against reality – a means of feedback.

The main subsequent development was to optically convert slant range (which the device measured) into plan range, the quantity fire-control systems used. Barr & Stroud tested this mechanism in a modified UF 1 rangefinder about 1931. The success of these trials showed that it was possible to produce a dual-purpose (HA/LA) rangefinder, a significant step towards a dual-purpose battery.

Calculation

The 1919–21 sub-committee followed much the same path as surface fire control developers. Initially it hoped that guns could be aimed on the basis of plots of what could be seen, much as in the Dreyer Table used for wartime surface fire control. That proved impossible; it was necessary to predict based on estimated target movement, correcting by comparing estimate with reality. That approach was embodied in the AFCT (for surface fire). It had been pioneered before war by Arthur Hungerford Pollen, and adopted during the First World War by the US Navy. Pollen’s engineer, Henry Isherwood, was helping develop the new surface fire-control system.

The one major success of the simpler approach was a means of estimating time of flight, hence fuse setting. Major A V Hill, who had had considerable wartime anti-aircraft experience, based estimates on current height and sight angle (equivalent to range). Given these data, he could draw a curve, which in turn could be represented by a suitably graduated slide rule. It was not precise, but errors were quite small in most cases (they were worst at low angles of sight and at extreme height, i.e., at long range). Prototypes were made by HMS Excellent. In October 1919, the sub-committee recommended that Hill’s plotting board be issued to the fleet as an interim device, to be made either by HMS Excellent or by the commercial fire control maker, Elliott Bros.

Initial experiments suggested that it was easy to estimate target plan course, but unless an aircraft was near the ground it was virtually impossible to say whether it was climbing or descending. The subcommittee (and others) realised that if an aircraft flew level, the key to prediction was target plan course, the course the aircraft flew over the ground, not taking diving or climbing into account. As in surface gunnery, that meant plan range and also plan inclination (the angle between the observer’s course and the aircraft’s course, in plan form). Speed could be estimated to within about 20 per cent from a plot of plan position. Creating one required time, and speed plotting required a skilled operator.¹¹ Also, range and bearing could not be measured quickly enough or accurately enough. It seemed simpler to guess based on knowledge of the type of aircraft.

By 1920 the sub-committee had concluded that an analog computer like the evolving surface-fire AFCT was needed. The inputs (set-up) would be guessed target speed and course (inclination), and the target would be assumed to fly level. These choices were inescapable unless the director was fully stabilised. The computer would compensate for own-ship manoeuvres. Errors in the set-up would be corrected based on observation. Once the set-up had been confirmed, the computer could use it to calculate deflections (horizontal and vertical) and fuse timing. Given the correct set-up, the computer would continue to predict target position even if the aircraft disappeared in cloud. A correct set-up could be upset only if the target manoeuvred.

The committee enthusiastically supported a proposal for such a computer from Commander D T Graham-Brown.¹² His parameters give an idea of what was expected at the time: own ship speed of up to 40kts, aircraft speed up to 130kts, and wind speed up to 60kts, and for altitudes up to 30,000ft. Graham-Brown’s proposal did not quite get all the way to a computer with feedback, because it did not really model target motion. It was complicated because target motion as seen from a ship was complicated. Even if they could be observed directly, rates were not constant, particularly for a crossing target. The only simple motion was the target’s. Once that motion was known, it could be translated into terms of what would be seen from a ship.¹³

The Committee realised that Graham-Brown’s proposal was the germ of a much better future system, so it referred his design to Henry Isherwood, who had designed the pre-war Pollen fire-control computer.¹⁴ Isherwood explained what was wrong: Graham-Brown solved the problems separately, then connected the solutions. To do that Graham-Brown required far too much mechanism and too many operators. ‘Anti-aircraft fire control must be exceedingly easy and rapid in operation: the fewer the operators, the better the chance of success. Commander Graham-Brown’s system is grafted on to the existing naval fire control, which in my opinion condemns it at once.’

An Interim System: STS and SGU

For the moment, ships received the Standard Temporary System (STS). The Royal Navy later saw it almost as a pre-automated equivalent to the HACS. STS survived alongside HACS on board ships not deemed worth fitting with the full system, such as the First World War-built ‘C’ and ‘D’ class cruisers. This system was in service by 1925, and probably much earlier.¹⁵

Like the later HACS, STS was conceived to support aimed rather than barrage fire. Elements were a UB-series rangefinder (heightfinder), an anti-aircraft Dumaresq, a Hill fuse predictor using a plot of angle of sight, and a deflection calculator. On sighting an enemy aircraft, the control officer estimated its speed and inclination. The anti-aircraft Dumaresq resolved that into virtual inclination and speed along the line of sight. Tangent elevation was applied using a fuse curve or cosine sight. Deflection was obtained from calculators set with height (or range), angle of sight, and apparent relative speed and inclination (obtained using the AA Dumaresq). All of this assumed that the target would continue at constant height, course and speed. The resulting deflections were passed to the guns. Once bursts appeared, the control officer could change inclination and speed to bring the bursts ‘on’. For example, if the bursts seemed slightly high and astern of the target, he could alter the inclination accordingly and increase set target speed. In theory, inclination and speed had to be kept up to date by adjusting the AA Dumaresq, but experience showed that was difficult once guns began to fire. Targets were designated from STS to guns by an Evershed (bearing transmitter) of the same type used for surface fire. Fuse and deflection were apparently communicated by voice.

As a result of sea experience and also of trials on board Tiger in July and August 1926, a standard control procedure was drawn up (issued 30 November 1926). Gun output was deliberately limited to avoid confusion in fuse setting, the rate of fire being governed by dead time. To avoid confusion, for example, the setting for the third round to be fired could not be made until about 2 seconds after the first had been fired. The prediction interval was therefore dead time plus 2 seconds divided by two, giving 8½ seconds between rounds, a maximum rate of 7 rnds/min, well short of what guns could actually do. The only way to do much better was to reduce dead time by adopting an automatic fuse-setter on the gun mounting, which could be kept continuously up to date. That in turn had to be associated with a mechanical predictor. What was done, by 1930, was to set the fuse on the gun mounting rather than some distance away, so as to minimise the delay between setting and firing.

Apparently the combination of AA Dumaresq and deflection calculator was less than effective, because in 1927 there was interest in replacing both with a simple Aldis tube sight (a telescope, rather than the usual simple tube) incorporating deflection rings.¹⁶ Accepting the use of a very simple deflection sight was, in effect, an admission that nothing short of the elaborate measures involved in the new HACS was worthwhile at anything but very short range.

By 1927 all battleships except the Iron Dukes, all battlecruisers, carriers, and cruisers down to and including the Centaur class had the STS. It seems to have survived during the Second World War on board ‘C’ and ‘D’ class cruisers. For a time the STS was associated with a dual-purpose (HA/LA) director developed specifically for the fast minelayer Adventure, which was armed with 4.7in HA guns.¹⁷ Unlike the director associated with the HACS, the Adventure director had no rangefinder. These directors and STS were installed on board battleships (Queen Elizabeth and later classes), the battlecruisers, the carriers and large cruisers (Hawkins, ‘E’ and Kent classes). Installation of HACS in these ships released some for other ships, so in 1926 they were proposed for the Iron Duke class battleships and ‘D’ class cruisers (no installations were proposed for the ‘C’ class). However, the following year it was decided that the Iron Dukes (except Marlborough, which was already fitted) would not be fitted with STS due to their age and the cost involved. Installation of Adventure-type directors was, however, being considered for the ‘D’ class cruisers. This installation was soon rejected due both to topweight and to the need to cut spending (1928).

Roughly parallel with STS was the Single Gun Unit (SGU). In 1924, when SGU was first discussed, it was planned for earlier cruisers (Caroline class and before), for leaders armed with 3in HA guns, and for large gunboats which could have no HA control positions due to limited space and weight. SGU would consist of a fuse plate and a deflection calculator, existing sights being used. By 1925 the fuse plate had been tested successfully and manufacture was proceeding. The fuse plate could also be used with the existing deflection cards. The supply of fuse plates and deflection calculators began in 1927. At this point it was decided that SGU would be fitted to the Iron Dukes (except Marlborough), to Cambrian and Caroline class cruisers, to older cruisers, to gunboats, leaders and new-construction destroyers, and to some auxiliaries. It might also become a backup system for ships with two or more guns on each side but with only a single HACS. The Aldis ring sight was scheduled for trials in 1926, and by 1927 it was expected to replace the deflection calculator pending successful trials. Production of the Aldis sights was badly delayed, although by 1928 supply of fuse plates and deflection calculators had begun.

The prototype SGU was installed on board the cruiser Calypso for comparative trials (against smoke burst targets) against STS in the 3rd Cruiser Squadron during 1930. They were successful enough that it was suggested that SGU was preferable to STS in ships in which one gun could be brought to bear on an air target; to see whether that was true further trials were ordered in Curlew in 1931. Final Aldis ring sight telescope trials were conducted at about the same time, and it was reported to compare very favourably with deflection calculating arrangements in other systems. The first seventeen Aldis telescopes were expected to become available for trials during 1931.

After all of this optimism, 1933 trials showed that SGU worked only under ideal conditions. No more were to be fitted, but ships which already had it were to retain it until something better became available. Two alternative systems were under consideration, either as primary controls for ships without HACS or as alternatives for cruisers with HACS, to engage targets on the opposite side or as fall-backs. Nothing seems to have come of these projects, because it was so much more urgent to improve other kinds of anti-aircraft fire.

The Computer System: HACS

The Committee completed its work before any new computer could be designed, but it appears that Isherwood’s comments led him or others to develop a High Angle Calculating Table (HACT) analogous to the AFCT, using somewhat similar mechanisms. It entered service in 1928, and it remained in use in modified form after the Second World War. The HACT was the computing element of the HACS, the others being the director aloft and the connections to the guns. The HACS was conceived specifically to defeat high-altitude bombers, controlling long-range anti-aircraft fire (defined in 1933 as anything from 3000 yds out).¹⁸ Permissible errors were set by the lethal radii of different shells.¹⁹ Later it was also described as a means of defeating formations before they broke up to deliver dive, torpedo or other close-range attacks.²⁰

The director was the eyes of the system. It incorporated a coincidence rangefinder (heightfinder), which was set horizontally to provide its crew with weather protection. Moreover, setting it vertically would have limited its size (base length), hence its accuracy or its effective range (as the US Navy found with its vertically-mounted ‘altiscopes’ at this time). The director also carried telescopes for layer and trainer, control officer’s glasses and transmitting devices. It was manned by a control officer, layer, trainer, heightfinder operator and communications number. The control officer provided the two key estimates (target course and speed), and he was responsible for corrections based on feedback. When on the target, the layer and trainer had the target in their telescopes. Their view was ‘undisturbed:’ no mechanism threw it off the target so that they could make corrections.²¹ The operators provided the below-decks HACT with angle of sight (layer), bearing (trainer) and height/range (heightfinder). The below-decks computer generated estimated target bearing and elevation, power-driving the director accordingly. There were manual back-ups. Different versions of the director had different means of stabilising it against the ship’s roll, none of them precise enough to make tachymetric operation possible. The Mk III version introduced a cross-roll periscope operated by the communications number.

The HA director tower as shown in the 1930 handbook for HACS I and II.

HACS was conceived as a more automated and integrated evolution of the STS, with the same inputs: range or sight angle from a heightfinder and estimated target course and speed from the control officer. The difference was that the elements were tied together. Range was computed and therefore could be predicted. However, because it seemed the STS worked to some extent, HACS was designed so that it could be de-integrated into something more like the earlier system.

Target course was expressed as angle of presentation, the angle between target plan course and the line across the line of sight.²² It was equivalent to inclination (of a surface target), which was so difficult to estimate that the Royal Navy used special inclinometers for that purpose.

The control officer transmitted it by placing the graticule of his glasses on the target. Angle of presentation cannot have been easy to estimate, but the way it changed could be sensed, and that was related to how close the aircraft came to the ship. Ease of measurement may have been over-estimated before the war because anti-aircraft practice was conducted against aircraft towing sleeve targets. The combination of tower and sleeve was quite long, and that probably made its apparent course relatively obvious. The problem did not become evident until the Royal Navy began experimenting with radio-controlled targets in the late 1930s.

This drawing of the Mk IVG director was used as a wall chart at HMS Excellent, the Royal Navy gunnery establishment. It was dated 31 March 1946. Mk IV was designed to work with 5.25in guns (in King George V class battleships and Dido class cruisers). Compared to earlier directors, it incorporated many more electrical circuits, many of them for Magslip data transmitters.

The trial system was tested on board Tiger in 1926.²³ Instead of the AA Dumaresq and a deflection calculator, it featured a new graphic approach to calculating deflection which became characteristic of later versions of the HACT. Without any precise means of measuring angular rates, speed across (which gave horizontal deflection) could not be derived directly from measurement. Angle of presentation was far too imprecise. The unusual and elegant new approach used the image of a circular tilting ring projected onto a flat screen. The circular ring (deflection circle) represented all the points from which an aircraft flying at constant speed could reach the centre of the ring, the point at which it would meet a shell. The circle measured deflection, because a gun had to aim off by the angular distance between circle and centre to hit an aircraft flying from the circle.

As seen from a ship, the circle was tipped over at the sight angle at which the shell burst. Deflections were measured in a plane perpendicular to the line of sight (the presentation plane, which of course was not parallel to the deflection circle). This projected figure was nearly an ellipse, whose near (upper) side was slightly larger than its far side. The appropriate deflections, vertical and horizontal, were given by the point on the projected circle corresponding to that which the aircraft occupied at the outset, which in turn was set by the aircraft’s course (angle of presentation, the course projected onto the near-ellipse).

A Mk IV director is shown aboard the cruiser Ajax at New York Navy Yard, 16 October 1943. The antenna on top is for its Type 285 range-only radar.

The Mk V DCT was installed on board the later King George V class battleships (Duke of York, Howe and Anson) and the later fleet carriers (Implacable, Indefatigable and Indomitable, of which the first two had RPC for their 4.5in guns). There were two versions, of which Mk V* (for the carriers) had its rangefinder high rather than low.

The radius of the circle was defined by aircraft speed (u) and the time of flight of the shell (t). The circle could be imagined as the upper edge of a cone, the height of which was R, the range to the aircraft when the shell burst. R in turn was time of flight multiplied by the average projectile velocity (APV). The tangent of the half-angle which defined the deflection circle was u/APV. Given enemy speed (u), APV was found by positioning a pointer on an APV drum marked with curves indicating average shell velocity at various ranges. If the combination of u and angle of presentation seemed to be wrong, the table operator could change both until the director seemed to follow the target.

An arrow on the screen was set at the angle of presentation (in production HACTs it was generally controlled by the graticule in the control officer’s binoculars). That gave an estimated present position along the ellipse. Deflection was the angle between present and future positions – between the centre of the ellipse and the present position along the ellipse. It was split into vertical and lateral components using vertical and horizontal cross-wires moved by handwheels which in turn fed vertical and horizontal deflection into the table (the computer). Two operators moved the wires until they intersected at the estimated target position on the ellipse.

Range did not enter directly, although APV certainly depended on it. The elegant graphic technique was not exact. Errors were imposed by the fixed cone angle of the projection unit (about 18½°); by assuming that lines of equal angle on the screen were perpendicular to the axis of the ellipse; and by substituting present angle of presentation for future angle of presentation. Note that there was no direct feedback to correct u, which was an estimate.

HACS I used present sight angle to calculate deflection, meaning that the range-calculating part of the system was divorced from deflection. The appropriate input was future angle of sight, which was given by future range. The I* version (1930) substituted it (as derived from the integrator described below) in both the deflection unit and the own-course and speed gear. This connection made it impossible for the HACS I* to deal with a climbing or diving target.²⁴

The trial installation seems not to have computed future range, but the production HACT incorporated an integrator which calculated it. To do that the computer needed both target speed and inclination, because for a target with any component of speed across (due to inclination) the speed along the line of sight varied with range because inclination (hence the components of speed across and along) varied with range. Although far too crude for calculating deflection, angle of presentation data seem to have been good enough for range calculations. Range in turn was needed both to set gun elevation in combination with vertical deflection and for fuse calculations (using a Hill predictor). For fuse calculation, the operator had to find the intersection between the range plot and a fuse curve. The curve of range against time up to the moment of firing had to be extrapolated to take account of dead time between fuse-setting and firing. To find range at the moment the shell was expected to arrive, the computer had to extrapolate current range to take account of dead time and time of flight (taken from a fuse curve).

The presentation used to calculate deflections graphically, from the 1945 Gunnery Pocket Book. (Photograph by Richard S Pekelney, Historic Naval Ships Association, courtesy of Mr Pekelney)

The computer did not calculate these future ranges. Instead, an operator extrapolated a range plot, using a grid of lines which could be rotated to match the estimated slope of the plot.²⁵ The technique may have been intended to allow for estimates to correct errors of calculation. Unfortunately the range plot was not a straight line, because the range rate along the line of sight varied. Only at very long range could motion across the line of sight be ignored, hence only at long range was the plot more or less straight. The chosen fuse setting was called out by the plot reader. An operator set a transmitter which drove receiving pointers at the guns. Fuses were set manually.

The integrator worked, not in terms of range itself (as in a low-angle system), but instead in terms of the logarithm (log) of range, log R.²⁶ The system had to keep translating back and forth between plan range and plan inclination the slant range (usually angle of sight) measured by the height finder. That required frequent multiplication by trigonometric functions, a complicated process in mechanical terms. Using logarithms simplified it, because the logarithm of a product is a simple sum of the logarithms of the quantities involved – adding is a simple process, mechanically. Working with log R made it possible to combine in a single plot long-range low-angle target motion (for which sight angle was difficult to measure, so the heightfinder worked as a rangefinder) and closer-in motion for which sight angles were used (the Hill fuse predictor was modified to work with range rather than sight angle). It was sometimes also claimed that the log R plot might be easier to extrapolate than a simpler plot of range against time, but even the this plot curved so sharply that the HACS sometimes produced too short a fuse range for an approaching target.

At first there seems not to have been any attempt to use feedback to check and improve the initial set-up. At the outset, the only source of feedback appears to have been spotting. The prisms of the heightfinder were moved at an average rate of sight angle change derived from a plot of sight angle. As the HACS developed, the device was integrated more fully (as in the change from HACS I to I* described above) and the operators learned to use feedback to correct the initial set-up. As experience was gained and the HACS was better integrated internally, range and inclination were used for feedback, as a means of detecting errors in set-up so that they could be corrected (this function was barely, if at all, mentioned in early accounts of the HACS). Range was fed back by moving the prisms in the heightfinder electrically (which is why it was called an electric heightfinder). If the range prediction was correct, the ‘cut’ would stay on the target. Angle of inclination was fed back by moving a graticule in the control officer’s binoculars. The control officer in the director tracked both forms of feedback and communicated observed errors to the HACT team below decks.

The Royal Navy distrusted purely mechanical solutions to gunnery problems. It knew that input data were often riddled with errors: although it did not use the phrase, it understood ‘garbage in, garbage out’. Thus the HACT incorporated feedback and correction mechanisms. In addition to range, angle of presentation was fed back via a graticule in the control officer’s binoculars controlled by the computer below decks. The main data correction mechanism was the ability to choose data, sometimes on the basis of a plot. A human plotter could average a run of data by eye, in effect drawing a line through the scattered points of a plot. For example, one operator was assigned to keep track of the observed log H (the logarithm of height). Initially he set target height based on the control officer’s estimate (from the director). Once rangefinder data began to come in, he used the average reading or followed directions from the plot reader. Together with rangefinder angles, the chosen height gave a series of ranges for insertion into the computer.

Generated range was plotted alongside range from the rangefinder. Even if there was a systematic error, the generated range plot was expected to parallel the plot of rangefinder data, and an operator could use the generated plot to spot errors in the rangefinder data.²⁷ In addition to range, the computer plotted angle of sight (which was equivalent, given fixed target height). The angle of sight plot seems to have been set by the mean observed rate of change of angle of sight, not by any analysis of target motion. It was used to set the angle of sight motion driving the rangefinder prism.

The HACS based its calculations on the motion of the aircraft relative to the ship. Corrections for own-ship motion, drift and (in later versions) convergence were all added to the calculated deflection. There was no wind corrector. It was assumed that wind would affect aircraft and shell more or less equally, but by 1931 it was clear that was not so; that was much of the reason that fictitious target course and speed had to be introduced. Given the estimated course and speed of the aircraft (assumed to be flying level), and estimated height, the computer generated range (and, therefore, sight angle, which was a simple function of range if the aircraft was flying straight and level).

Fleet exercises with HACS I began in 1930. The results seemed promising. Both with HACS I and with the earlier STS, time between first sight of the enemy and first range or height was 31 seconds. Average time to open fire was similar, 53 seconds for HACS I vs 50 seconds for STS. However, the percentage of shots within 100 yds for both range and line was 16.1 per cent for HACS I vs 6.9 per cent for STS. The superiority of HACS I was somewhat exaggerated because it included shots for which data were imprecise. As experience was gained, the gap between the new and the old system widened considerably. Of 148 shots against sleeve targets fired by Nelson in six firings in 1930, 50.7 per cent were within 100 yds for range and line, and 92.5 per cent within 200 yds. In 1931, HACS I got 15 per cent of bursts within 100 yds for range and line, compared to 7 per cent for the STS, and the figures for between 100 and 400 yds were 54 per cent vs 34 per cent.

However, it was already clear that fuse prediction by extrapolation was not working well enough.²⁸ At this time spotting was the major way of correcting for a bad set-up. It was not at all clear whether a control officer could disentangle bursts at long range and at high rates of fire. Was a burst the result of the most recent correction, or a previous one? If a fuse was predicted wrong, meaning that range was in error, then the wrong average projectile velocity would be applied to the deflection unit, and deflection would also be wrong. Delays in reporting spots (bursts) could cause the bursts to appear astern of the target (but expected lags in transmitting spots did not seem large enough to explain the errors).

In 1930 it was reported that HACT teams often had to use entirely false set-ups to bring bursts into line with the target.²⁹ That happened even when range was approximately correct. That spotting was needed to correct for line (direction) pointed to a need for some means of measuring rather than estimating the relative course and speed of the aircraft, rather than guessing both. That was never possible.³⁰

The HACT Mk IV as shown in the 1940 handbook. Fuse number is the fuse setting in fractions of a second. The fuse range crank above it is a typical follow-up. In analogue computers like this one, the result of a calculation was typically displayed on a dial, and it had to be read off and cranked back in for a second calculation. In theory using follow-ups made it possible for operators to smooth data and to correct obvious errors during computation.

HACS I was conceived to control 4in and 4.7in high-angle guns, but in the 1920s the Royal Navy was interested in using heavier-calibre guns. HACS IA and IB were associated with this project. Due to their calibre, hence weight of shell, these guns were considered to possess considerable anti-aircraft potential. That was why the 8in guns of the new cruisers were given 70° elevation, and the 6in secondaries of the Nelsons were given 60° (although the ships had dedicated 4.7in anti-aircraft guns).³¹ Initially there was hope that these heavier guns would have anti-aircraft control on a par with HACS. However, as of 1926 design work had not yet begun, so none of the ships was completed with main battery HA fire control gear. It took several years to abandon this scheme as too complex.³² There was also HACS IC for the carriers Courageous, Glorious and Argus, all of which had dual-purpose batteries (Argus was soon eliminated because she was due to decommission so soon).

Because the director was not integral with the HACT, a director could be connected with any of various HACTs. Because HACS was so expensive, initially ships had only one. It could engage only a single target. By 1931 policy envisaged adding a second HACS for capital ships, typically to be mounted aft with its own calculating position below decks. Guns would have duplicate receivers (for follow-the-pointer instructions), so they could follow whichever director was desired. Installations began in 1931 as ships refitted.

As of 1940, six systems were in service: HACS Mk I, II, III, III*, III** and IV, each with its own type of director and its own type of computer (HACT). HACS I was fitted to all ships up to 1935; as of 1940 it was being replaced by ‘ended’ HACS III (the resulting surplus sets were installed as second HACS in heavy cruisers). Each HACS Mk I director was associated with a particular HACT (ships had several). To be able to switch directors a ship needed duplicate wiring from each HACT to the guns, each gun having a change-over switch from one set of receivers to the other. HACS I** had improved fuse-setting gear. In 1935 arrangements were made for ships’ staffs to modify HACS to accommodate an enemy speed of 250kts. Later they would be modernised with 350kt settings, and errors in the deflection gear of HACS I and II would be corrected. This took time, so as of 1940 it was still planned ultimately to modify all sets for increased target speed (350kts) and for gyro roll correction (HACS I***).³³

HACS II was a slightly improved HACS I* fitted only in Repulse and in the Leander class cruisers (one per ship). By 1939 gyro correction replaced an earlier failed liquid pendulum, whose float did not exert sufficient force. The gyro corrector stabilised the director telescopes in the vertical, and removed the effect of roll from measured director setting. The layer no longer had to follow the target as the ship rolled, but instead concentrated on the angle of sight. It employed a single gyro mounted athwart the line of bearing to the target. It did not provide cross-level (that was incorporated only in HACS IV). Gyro roll correction was exerted by a separate unit in the HACP. The Mk II director had its windscreen shape altered to accommodate an electric junction box inside its screen, and to suit its new rangefinder.

HACS I, II and III all used step-by-step transmission. Elevation was transmitted to the director in 3 minute (of arc) steps, at up to 100 steps/sec; train was in similar steps, but 130/sec. In HACS I and II (including improved versions of I), power follow-ups used M type transmission (motors and hunters) with small DC motors. For HACS I and II, target speed (as reflected in deflection gear) was normally - 45 to 200.4kts; own-ship speed limits for all versions were 0 to 40kts. Steps for speed transmission were 0.25ft/sec per step, with a maximum of 45 steps/sec. APV limits were -870 to 2100ft/sec. Elevation angles in all versions could be 10 to 90° (HACS Mk I* was limited to 9° 30’ to 89° 30’).

By 1938 HACS III was in production at three firms. It equipped some cruisers prior to the Birmingham class and some modernised battleships and carriers. It could provide deflection for barrage fire (see below). A 15ft (rather than 12ft, as in earlier systems) rangefinder was fitted in an anti-vibration cradle. Range conversion (to height) was limited by stops to between 2000 and 20,000 yds. The director had an entirely new roll corrector adjacent to but separate from the HACT proper, employing a gyro (in US terms a stable vertical) driving an oil motor connected to the HACT. This device was trained in the direction of the director, so its correction was applied to the director’s line of sight (a new gyro roll corrector was designed for retrofit to HACS I and II). Angle of roll could be sent (by ‘M’ transmitter) up to a corrector sight in the director, if that was fitted. Users particularly liked it and the anti-vibration rangefinder mounting. HACS III embodied considerable mechanical improvement, including ABC transmission and oil motor follow-up (maximum motor speed of 400 RPM corresponded to 6°/sec motion by the director).³⁴ Target speed limits were 35 to 350kts. Each transmission step was 0.25kts, and the device could handle 100 steps/sec. This version could handle 4in and 4.7in ballistics (APV 870 to 2100ft/sec and 1000 to 2230ft/sec, respectively). The destroyer depot ship Woolwich and the anti-aircraft cruisers all had the C version with provision for low-angle fire.

The associated Mk III director had its plating thickened to 0.08in, presumably to deal with more powerful aircraft guns (later this thickness was described as merely weatherproof). Mk III* had an additional range-taker’s position, and the control officer’s window was enlarged so that an eye-shooting High Angle Director Forward Area Sight (HADFAS) could be installed. The directors in Penelope and the forward one in Malaya did not have this sight.³⁵ Mk III** had a completely round windscreen but no additional range-taker position, and it was reinforced against shock.

By 1938 HACS IV was being installed in all new-construction and also in rebuilt battleships with dual-purpose secondary batteries. It was a modified Mk III with magslip (synchro) transmission and with roll compensation; provision was also made for cross-level. The design was cleaned up to reduce loads on shafts and pinions to reduce maintenance and backlash. Mk IV* was designed specifically to support the longer-range 5.25in gun. It had improved plotting arrangements. When used with 5.25in guns it was fitted for salvo fire. Target speed limits were 0 to 350kts, and the Magslip transmitter could turn at 400 RPM (one revolution was equivalent to 650kts). At least initially, HACS IV was fitted for 4in and 4.5in ballistics (1050 to 2350ft/sec and 1050 to 2250ft/sec). HACS IV* was designed to handle longer-range 5.25in guns on board Dido class cruisers and battleships. It and HACS IV were modified to incorporate drives which could be linked to an Admiralty Fire Control Clock (AFCC) for surface fire. HACS IV could operate both with various versions of the Mk IV and Mk V director and with the later Mk VI (which had been conceived for the Flyplane computer described below). HACS IV and IV* were designed specifically to deal with both air and surface targets, the air-only version being designated AA and the dual-purpose version AASU.³⁶

Ships completed in 1940 had a new Gyro Rate Unit (GRU) mounted on the director, to measure horizontal and vertical angular rates directly. Although this measurement might be described as tachymetric, it could not transform the HACS into something capable of dealing with climbing or diving targets. Nor did it solve the problem of measuring target inclination. That was still fed in by the control officer. The Gyro Rate Unit Box (GRUB) in the calculating position below decks received the two rates and the angle of presentation from the GRU, ultimately to feed data into the HACT nearby. This was still a feedback process. The assumed target ground (plan) speed was set on the GRUB. Given an angle of presentation and a target height and range, this speed implied particular vertical and horizontal rates. It was adjusted until these rates matched the observed rates, the GRU acting in effect as a feedback mechanism. The resulting checked ground speed was designated the GRU ground speed and fed into the HACT. According to the 1939 HACS manual, ‘at long ranges this information should afford a valuable source of information to the Control Officer, but will not be accurate enough to be followed blindly. At short ranges it should be quite accurate enough to be accepted.’ At long ranges and low sight angles the rates would be too small to be reliable, and might even be grossly misleading (according to the GRU handbook).

The layout of the HACT from the 1930 handbook for HACS I and II.

When HACS IV was being designed, director production lagged, so the Mk III director was adapted to Magslip transmission and stabilised in elevation as Mk IV. This director was completely round, larger in diameter, and had the additional range-taker position. By 1945 the typical crew was seven rather than the five of the original HA director: a director officer, a rate officer (GRU operator for AA), a director layer, a director trainer, a phoneman and a local gun direction officer or rating. The director officer still had the angle of presentation binoculars. Note the absence of a range-taker due to the use of radar.

Some ships had HA/LA Mk IV directors, similar to Mk IV but wired also to control the low-angle armament (in 1940 they were in the Birmingham and Liverpool classes and Aurora). This director was modified during the Second World War. Directors had High Angle Director Eyeshooting Sights (HADES) in place of the earlier HADFAS; they were later replaced by GJ 6 reflector sights. HADES was a simple open ring sight which could be moved independently or aligned with the director binoculars. It had rings spaced equivalent to speeds of 100kts and 200kts at a range of 1500 yds. During the war the barrage torches were redesigned for a maximum scope of 12° to either side (instead of 6°), to take advantage of the deflection measurement of the GRU and to take account of higher aircraft speeds. Target speed transmitters were eliminated in favour of a simple dial.

The initial wartime versions of Mk IV were Mk 4 GB and Mk 4 GB AA/SU. Here G indicated the GRU and B indicated scooter control, for quick slueing to acquire a target. Typically one handle was provided for the director officer and a second for the GRU operator; in surface fire mode, the second was used by the rate officer. Because addition of scooter control to the existing oil motors proved lengthy and expensive, a few directors were fitted for electric control (‘E’) and were therefore designated Mk 4 GE. Here scooter control meant that by depressing the ‘scooter’ the control officer could quickly slew the director to a new bearing.

There was considerable interest in duplex rangefinders, but they made the directors heavy; the HACS was designed to work with one rangefinder. Plans called for duplex rangefinders in the 1937 (King George V class) battleships, the second barrel working separately from the HACS table except that the table would feed range to it to assist operators in keeping their cuts on the target.

Meanwhile a new Mk V director (tower) was developed with all possible improvements, such as a sliding roof, duplex rangefinder, and stabilisation for training and laying. With the advent of the Air Defence Officer (ADO), lookouts outside the director could cue it, so a clear overhead view seemed less important; weather protection became practicable. It seemed possible, in 1939, that earlier directors would receive roofs. At this time Coventry was testing gyros intended as the basis for a future tachymetric deflection control system. Mk V equipped the King George V class battleships and wartime aircraft carriers.³⁷ Unlike all the earlier HA directors, it was cab-shaped rather than basically cylindrical. It was completely enclosed for weather protection. Because it controlled dual-purpose guns, Mk V had improved surface firing capability. That required more personnel, because surface firing required a spotter. Thus there was a seat for a rate officer (the GRU operator for AA fire) alongside the layer, with the trainer on the other side of the director cab. Between them was the raised seat for the control officer, with his binoculars and his HADES telescope. At the centre of the cab was the telephone operator, with the range-taker at the rear. The director controlled the ship’s HA/LA armament through an HACT for high-angle fire and through a separate AAFC (fire-control clock) for low-angle fire. Directors were fitted with a Radar Training Sight (RAC) after the war.

At the end of the war the Royal Navy introduced a new cylindrical Mk VI (typically written Mk 6) director designed to work with the Type 275 radar. It was designed to be operated from the ‘Tallboy’ (radar console) in the Control Position. The director was entirely electrically-powered. Mk 6 was designed specifically to provide a better view for the crew and to include arrangements for a local gun direction officer. It could be connected either to an existing computer (such as HACS 4) or to the new Flyplane introduced after the war. It and Flyplane are described in a later chapter. Unfortunately Mk 6 was not designed for tachymetric operation; it was not stiff enough to measure bearing rates as it tracked a target. That caused problems when it was used with the tachymetric Flyplane computer post-war, and it had to be rethought as the post-war Mk 6M.

All of this leaves the question of how effective the HACS/gun combination was, both in reality and in pre-war perception. In January 1938 DTSD tried to summarise the results of anti-aircraft practice in all the fleets.³⁸ Assuming perfect prediction and fuse setting, guns could be expected to burst a third of their shells at the correct range. The standard for success was to burst a shell within 100 yds in front of the target. That was well beyond lethal range, but a pilot seeing such a burst would probably jink and thus ruin his aim. Some of the scores were unreliable because only the Home and Mediterranean Fleets had full facilities to film, and therefore triangulate, bursts. Tables of results showed that ships were much better at getting shells in the right direction (line) than in getting them within the right range. The right combination was achieved about 12 to 17 per cent of the time. In something less than a third of all practices, no bursts at all were placed within 100 yds of the target. In barrage fire, 6in and 4in guns placed their bursts within 100 yds 26.5 and 4.8 per cent of the time, respectively (4.7in guns failed altogether to do so).

The 1938 edition of Progress in Naval Gunnery included the comment that some form of tachymetric deflection measuring device was urgently needed; sleeve and Queen Bee targets, whose speeds were known within narrow limits, and which could not vary their speeds appreciably as they approached, were apt to breed false confidence in deflection control. The 1939 edition of Progress in Naval Gunnery reported that a new fully-stabilised high-angle control system was being developed for future construction (it figured in early plans for Vanguard). At the time (May 1939) it was planned for ships which would complete in 1942–3. The necessary stabilisation system was tested successfully in Coventry in 1938–9. At least initially a tachymetric unit was conceived as an add-on to the HACS. It could not completely replace the course and speed method at medium and long ranges, when vertical and lateral deflections were small (and presumably difficult to measure).

It was also obvious by early 1939 that diving targets would be more and more important.³⁹ If constant height could no longer be the basis for fire control, the next possibility was to assume constant range rate as the basis for feedback. Tests of a duplex rangefinder were underway, one half operating on a constant-height basis and the other on a constant range rate basis. For the moment, since prediction would become impossible once dive bombers had broken formation, barrage fire was the only possibility; it had to begin as soon as the bombers broke formation, regardless of range. Accounts of wartime anti-aircraft action in the Mediterranean suggest that barrage fire was much preferred to aimed HACS fire.

Systems for Smaller Ships

By the 1930s the Royal Navy took the air threat seriously enough that it wanted destroyers to contribute to fleet air defence. Limited gun elevation made sense for a destroyer engaging bombers approaching a ship she was protecting. As the aircraft approached, it would spend very little of its time at high angles of elevation over the destroyer, which therefore would get very few high-elevation shots, however high it could point its guns. An attempt to build a 60° destroyer gun mounting having failed, 40° was selected as the highest for which a satisfactory mounting could be built. The first destroyers affected by the new requirement were the ‘E’ class of the 1931 programme, but they and their successors up to the ‘I’ class did not have special anti-aircraft fire-control arrangements, because the required systems did not yet exist. The 40° mounting was an acceptable compromise for a ship intended primarily to beat off torpedo attacks by enemy destroyers and to deliver its own torpedoes against the enemy’s capital ships.

Going to higher elevation entailed serious sacrifices. The higher the elevation of the gun, the higher its trunnions, which had to lift the gun high enough to allow it to recoil (at maximum elevation) without hitting the deck. Higher-powered guns recoiled further. Above a certain trunnion height, a man standing on the deck would be unable to serve the gun. The only solutions were to provide a pit under the gun mount (reducing deck strength) or to place the crew on a platform above deck, revolving with the gun, as in the contemporary US 5in/38. The Royal Navy rejected such complex arrangements (hoists revolving with the gun mount). DNO reported that the maximum gun which could be hand-loaded at all elevations, and which was really suited to anti-aircraft use from a destroyer, was the 4in (35lb shell), whose twin mounting weighed 14 tons, somewhat more than a 40° elevation single 4.7in (62pdr). Wartime experience showed that these assumptions were badly flawed, and eventually a 55° un-powered mounting was accepted (see a later chapter). Note that even a 40° gun needed an anti-aircraft fire-control system, not least in order to calculate fuse settings. The elevation argument did not apply to the relatively low-powered 4in HA gun, which could elevate to 80°.

A typical destroyer HA director created by converting a three-man rangefinder.

The first ships with an anti-aircraft computer were the ‘Tribal’ class destroyers which were conceived to contribute to fleet air defence. They were given the fuse-keeping clock (FKC), ‘clock’ indicating a simple computer, the rest of the designation emphasising the need to set fuses. The FKC was first tested on board the sloop Fleetwood. This FKC was associated with a bare minimum system. The associated Rangefinder Director (not in the same series as the HA directors of the HACS) was sometimes described as a HA/LA director. The UK series rangefinder was modified to transmit director setting and training as well as range. It carried a control officer’s glasses with the angle of presentation graticule. Typically it was occupied by a control officer, a director layer, a director trainer and a range-taker; in contrast to the HACS director, there was no separate phone man.⁴⁰

The other elements were an optical deflection calculator and a separate fuse-keeping clock which maintained future range so that the fuse could be read without the complication of a plot. Inputs were target course (via angle of presentation) and speed, plus initial range. As in a HACS, calculated range was transmitted back to the rangefinder as feedback. Change of range in time of flight was produced and added to rotate a dial to future range; the dial was marked with fuse curves, from which fuse numbers could be read. The first FKC used an inverted form of the HACS deflection calculator, in which the operator looked through a lens at the ellipse inside. The production version had the same kind of optical deflection calculator as a HACS.

Fuse-Keeping Clock Mk II, as shown in the 1945 Gunnery Pocket Book. (Photograph by Richard S Pekelney, Historic Naval Ships Association, courtesy of Mr Pekelney)

In effect an FKC was an HACT without its plot, with the same kind of feedback (range to the rangefinder, target course to the A/P binoculars used by the control officer). Early FKCs also had power follow-up for sight angle, which was an additional form of feedback: the FKC set the control officer’s binoculars at the generated sight angle, which in effect corresponded to the calculated range (for an aircraft flying level). Angle of sight was needed to convert angle of presentation, as seen by the control officer, into target course. An important difference from the HACS was that the same range operator tuned both rangefinder range and rate, a combination which turned out to be difficult in initial trials on board Fleetwood. Compared to an HACP, an FKC required eight rather than nine personnel.⁴¹

In a destroyer, the FKC worked in conjunction with the standard surface fire-control computer, the AFCC. In a smaller ship, it worked with the less sophisticated Admiralty Fire Control Box (AFCB), which provided the surface fire-control function. At the outbreak of war FKCs were on board large escorts (Bittern class and later sloops, e.g. Black Swan class) controlling 4in HA guns; on board ‘Tribal’ class destroyers; on board ‘J’ and ‘K’ class destroyers, and on board smaller escorts (Hazard class and ‘V&W’ class converted to anti-aircraft escorts). All had gyro roll correctors in their dual-purpose (anti-air and surface) TS (the British term for a fire-control centre). At this time the FKC was planned for the ‘L’ and ‘M’ class destroyers with their higher-elevation (50°) main batteries.

The success of the FKC, compared to the HACT, showed that the new kind of fire control based on computation really worked. No range plot was needed. This success is evident in the wide use of FKCs outside the destroyer series, in ships which might otherwise have had HACTs. The fleet carriers Indefatigable and Implacable had Mk III*, as did the monitors Roberts, Abercrombie and Erebus, many depot and repair ships, fighter direction ships and other major auxiliary ships (such as Prince Robert). The carrier/repair ship Unicorn had Mk II*. The cruiser Ontario had Mk V, and it was planned for the Tiger class. The fast minelayers Manxman and Apollo had Mk II**.

There was yet another series of dual-purpose directors for destroyers, intended to save space on crowded bridges by consolidating the surface director control tower (DCT) and the associated rangefinder director.⁴² Because destroyers were expected to have excellent surface fire capability, these units were described as LA/HA director control towers. The first was installed on the ‘J’ class (and its ‘K’ and ‘N’ class repeat versions). It proved unsuccessful; in 1941 the DCT in these ships was reduced to surface capability and the rangefinder converted into a three-man rangefinder director, a seat being added for the control officer (as well as a windshield to protect personnel). The other personnel were the layer and trainer. The associated computer was FKC Mk II. In these ships the 12ft rangefinder (UK series) had range limits of 2000 to 20,000 yds. The ‘L’ and ‘M’ class had another such director. These destroyers required additional HA capability because their power-operated gun mounts could elevate to 50°, the highest elevation of any British destroyer main battery to date.⁴³ Because this Mk IVTP was unsuccessful against aircraft, a new ‘K’ type DCT (see below) was developed for the later war emergency destroyers. It formed the basis of yet another director series, and was designated ‘K’ Mk I. The K designation was taken from the ‘K’ sight used on the director.

Later destroyers had either the US-supplied Mk 37 or the wartime Mk VI director described in a later chapter. The six Vickers-built Havant class destroyers (taken over from Brazil) had pure surface fire-control systems. Their sole anti-aircraft control system was an auxiliary barrage director.

Barrage Fire

Barrage fire, largely by non-high-angle guns, turned out to be extremely important to the Royal Navy during the Second World War. The idea can be traced back at least as far as the 1919 Naval Anti-Aircraft Committee. In February 1920 its president Captain Usborne suggested using barrage fire by low-angle (secondary battery) guns against torpedo bombers and remote-controlled boats. Experiments having shown that no existing fuse could be relied upon to burst 6in shell making grazing hits at ranges inside 3000 yds, so the ship would rely entirely on the splashes the guns could create. Torpedo bombers were sometimes damaged by the splashes created by their torpedoes. CMB (motor torpedo boat) officers considered their boats (and, by extension, remote-controlled boats) ‘most probably highly vulnerable to a splash falling on board them and swamping them’. Since there was no real knowledge of the size of a splash, or how much of it was water and how much light spray which could not harm an aircraft or boat, the committee asked for experiments. Its report suggested firing shrapnel from heavy low-angle guns. This technique was tested by Tiger in 1926. It seemed that this type of fire would be effective against torpedo bombers.

Trials of 5.5in and 6in secondary guns at long range and low sight angle began with firing by Hood in 1928. A special control system called H.X was installed in Renown, Repulse and Curacoa for trials. At about the same time battleship secondary batteries fired barrage and splash trials against torpedo bombers. It turned out that shells burst too low above the water (at about 8ft altitude). Time-fused HE was likely to be more effective than a splash barrage. The control technique used a fixed fuse setting and a corresponding range set in the director sight. The director was laid on the incoming aircraft rather than on the horizon (as in a splash barrage), and the ship fired rapid salvos of time-fused HE shells. This seems to have been the beginning of a technique that was very effective in the Second World War. The more elaborate H.X system was dropped because it entailed too much complication in ships which already had multiple fire-control systems. Results were not so good as to justify continued ammunition expenditure.

The barrage trials were satisfactory enough that the 1930 issue of Progress in Naval Gunnery announced as policy that the secondary batteries of battleships and the main batteries of cruisers would continue to be used against aircraft, but only at close range. The necessary ammunition and practice facilities would be provided. The possibility of providing efficient high-angle control and anti-aircraft armament for primarily low-angle guns in future ships would be reconsidered. Battleship main batteries would not be involved (this was made explicit in 1933).

In 1932 the Mediterranean Fleet tried barrage fire by anti-aircraft batteries, also an important Second World War tactic (as a way of overcoming the limitations of HACS). There were three alternative techniques: (a) a stationary barrage (at 1500 yds) through which an aircraft had to fly; (b) a falling barrage, one salvo fused for 1500 yds, the rest at 1000 yds; and (c) two-step, fired at a fuse range of 2000 yds and, when the target passed through that range, at 100 yds. CinC Mediterranean Fleet concluded that a form of barrage could be fired against torpedo bombers, and that (a) was preferable against an attack by more than one aircraft. The trials were insufficiently conclusive, but they must have been promising.

Between the wars the Royal Navy became interested in using all of its guns to beat off air attacks. That meant setting up procedures to use heavy guns (normally limited to LA fire) to create barrages through which attacking aircraft would have to fly. The barrage concept explains why early British heavy cruiser guns elevated to 70° and why the 6in secondaries shown on Nelson in 1938 could elevate to 60°. They were by no means anti-aircraft guns: they loaded at a fixed elevation (5°), hence had to depress and elevate between shots. A project to link them with the ship’s HA control system was abandoned. The Japanese followed much the same reasoning in providing some of their 5in destroyer guns with high elevation (up to 75°). Like the British, they did not provide special anti-aircraft fire controls, and the guns fired slowly at high angles because they loaded at low elevation. Nelson’s anti-aircraft guns (single 4.7in and octuple pom-poms) are all shrouded.

As of 1933, policy was for the modern cruisers to use their main batteries against aircraft at short ranges (this did not extend to the older ‘C’, ‘D’ and ‘E’ class cruisers). After further Mediterranean Fleet trials, a provisional method of barrage control was issued to the fleet. By 1935 barrage fire was being considered as an anti-dive bomber measure.

In the spring of 1933 trials on board Nelson showed that her 6in guns could be used against aircraft, but with so much remaining to be done to make the specialised high-angle armament effective (according to Progress in Naval Gunnery for 1936), it seemed unwise to provide elaborate high-angle controls for low-angle guns. In 1935 the question was reopened; geometric data might be provided by a ship’s HACS. CinC Mediterranean Fleet, who was facing the threat of Italian air attacks in the Abyssinian Crisis, suggested that there would be occasions when it would be desirable to augment fleet anti-aircraft fire. That would apply more strongly to destroyers with all-low-angle batteries. He wanted barrage fire using pre-set fuses. When the rangefinder indicated that aircraft were in range, guns would open rapid fire for 30 seconds. Lateral deflection would be estimated in advance, and vertical deflection included in the guns’ elevation. This technique would produce a series of bursts on the line of sight at different heights, rather than a barrage at one fixed point in the sky.

The following year Progress in Naval Gunnery reported increased interest in close-range barrage fire, as theoretical investigation showed that it offered a good chance of inflicting damage. Nelson, Rodney and Achilles (Home Fleet) were ordered to investigate this kind of fire during their 1936 Summer Cruise. It turned out that the most promising means to applying deflection was an HADFAS worked by the HA control officer. Its movements worked pointers on the HACS deflection screen. A prototype was fitted to Coventry, and this device was incorporated in the design of HACS Mk IV. The exercises showed that fire tended to be opened late, so it had to be ordered when even a relatively slow target was about 1000 yds beyond the barrage range setting.

Barrage practice was further formalised in 1937. There were now three types: distant, to harass aircraft when they were first sighted; high, to harass aircraft just before they reached the point of release; and close: against aircraft carrying out close-range attacks. It was reiterated that battleship main batteries were not to be used. The secondary batteries of the Nelsons could be used for all three functions; other battleship secondaries were usable only for close-in; modern cruisers could do all three, but the ‘D’ and ‘E’ classes were limited to close-in. ‘D’ and earlier class destroyers were limited to close-in, but later ones could be used for distant and close barrages. The carriers Eagle and Furious were limited to close-range barrage. Control would be extemporised, with close-in barrage the main priority. Anything else had to give way to the urgent need to modernise fleet anti-aircraft firepower and to equip new ships. Tactics were developed.⁴⁴

In 1938 Nelson reported a barrage technique using her 6in guns against bombers; it was estimated that fifty rounds of 6in HE were equivalent to 250 to 300 rounds of 4.7in high-angle fire. The ship’s technique was reported to the fleet.⁴⁵ The idea was to fire a full broadside at a fixed fuse setting, ‘rippling’ fire to increase the period of time during which shell was bursting in the selected zone. The technique would be used only against approaching targets, shells being fused to burst on the line of sight. The recommended rippling interval for 6in shells was 2 seconds. Four fixed fuse-settings were specified for particular ranges and angles of sight. By this time the use of low-angle guns for air defence was considered so important that a special computer was being developed specifically to use HACS data for low-angle fire control.

Close-range barrage fire was now recognised as an important part of fleet air defence, to be imposed both by long-range anti-aircraft guns and by low-angle guns. Adding heavy anti-aircraft was a new departure. To some extent it was an admission that HACS was limited, and it was also a way of supporting a high rate of fire, which was essential if no gaps were left through which a fast target could pass. It was described as a way of filling the gap between the shortest range at which controlled fire was effective and the range at which a bomb or torpedo would be dropped. Against a torpedo bomber the preferred setting was 1500 yds. The attacker would be forced either to drop the torpedo at excessive range or to pass through the barrage. In either case there was no point in a shorter-range barrage; by this time it was clear that barrage range should not be changed by any step less than 1000 yds. Barrage was also a way to deal with dive bombers, 1000 yds being the most suitable range. The provision of forward area sights was connected with barrage tactics.

Barrage fire by anti-ship guns offered the ability to engage a separate set of targets during a synchronised attack. During the Abyssinian Crisis the idea of a long-range barrage by anti-ship guns became so important that it overshadowed the close-range barrage (which had been far more important). As of 1938 control methods were still being developed.

Detailed instructions for barrage fire were issued in the spring of 1939, with special emphasis on dive bombers. Controlled fire would be continued as long as possible, but once bombers broke formation to dive, the HACS could no longer control fire, and a barrage had to be opened. The problem of torpedo attack also seemed to invite a barrage solution.

Detection

The committee was also interested in the evolving science of aircraft detection by sound. After witnessing an October 1919 demonstration of the army’s system, the committee suggested rigging one aboard a ship. The army system consisted of a large aimable disc to focus sound, tuned to a frequency corresponding to an aircraft exhaust. In May 1920 the Committee laid out specifications for a shipboard sound locator, to be scanned over the sky. It should indicate the presence, direction, and altitude of an aircraft or airship up to 15 miles away, with an accuracy of at least 5°, neglecting distortion due to wind, etc. The device should be insensitive to non-aircraft sounds. Some form of IFF would be desirable. The project soon collapsed; later naval interest in sound detection concentrated on the problem of defending the fleet in harbour.

It was obvious that lookouts were not enough, particularly as aircraft became faster and flew higher. That was one reason radar was so important. The first British naval radars, beginning with Type 79, were air-warning sets. Because they operated at relatively low frequency (the best technology available in quantity), shipboard antennas could not offer very fine beam definition. However, they could produce enough power to detect aircraft at a considerable distance. These sets were classified as aircraft warning (AW) devices. By 1940 they had been refined to the point where they provided good ranges. They could not support director fire control, but they could support barrage fire, because they could indicate that an incoming target was passing within barrage range.

By the outbreak of war work was well advanced on a second generation of radars operating at shorter wavelength (about 50cm compared to 3.5m or 1.5m), offering ranging but not precise direction for fire control: Type 285 for anti-aircraft, and later Type 282 for pom-poms and Type 283 for the wartime barrage director. These sets entered service after the outbreak of war.

Putting the System Together: The ADO

British experience during the First World War showed the importance of designating targets. It was also clear that anti-aircraft fire had to be concentrated on the most threatening air target. By the late 1930s the British therefore became interested in the step before a director and guns were assigned to a target: the choice of target by a designated Air Defence Officer (ADO) supported by dedicated look-outs. The ADO would also be responsible for choosing the appropriate means of resisting air attack. The ADO concept seems to have been considerably in advance of that of any other navy.

In effect the ADO organisation was designed to maintain awareness of the air situation around a ship so that targets could be prioritised and fire assigned to them. The ADO organisation seems to have evolved into the Gun Direction Room (GDR), which was an element of the wartime Action Information Organisation (AIO). The AIO was often considered equivalent to the US Navy’s CIC, but the Royal Navy emphasised the split between ship self-defence (as organised in the GDR) and functions such as fighter control, the latter organised on a much larger scale. The US Navy tended to combine the two functions.

Ramillies shows the air defence position built up from her bridge in this late pre-war photograph. The objects visible atop it are ADO sights, which were used to indicate incoming targets. In effect they were the distant forebears of the later British ‘Eversheds’ and the US Target Bearing Transmitters. (Naval Institute by Ted Silberstein)

The ADO concept seems to have originated in the Home Fleet. During the latter part of 1934 the fleet convened a committee to consider fleet anti-aircraft defence. It concluded that the captain of a modernised battleship could simultaneously engage as many as six air targets, not counting those engaged by Lewis guns. With a modern closed bridge, he had little or no view of the sky from which synchronised air attacks could come at great speed, with little warning. Air defence was essentially different from other kinds of action, in that it would develop suddenly. Moreover, the relative importance of different attacks could shift instantly, a previously innocuous aircraft becoming a priority threat, and vice versa. The committee recommended that the captain be relieved of some of his many activities so that he could concentrate on developing the offensive power of his ship. He would of course retain a veto. However, reaction time was so short that the ship could not afford wasted time.

The alternatives were decentralisation to the weapons or groups of weapons, and centralisation under an ADO. Rigid centralisation might preclude quick reaction in a close-in melee, but there had to be some coordinating authority to collect and disseminate information and to insure that the various weapons were used most effectively. To direct the batteries most effectively there had to be someone not immediately concerned in the actual control of fire, hence able to take a general view of the situation and keep the captain informed. He would control the air lookouts, the most important of which would be located near him, and also in direct contact with the captain. He also needed adequate weather protection. The proposed organisation was immediately tested on board Nelson and Rodney. To accommodate the ADO, ships under refit were being given suitable open bridges. Unfortunately that was impractical in some ships, including Warspite, then under reconstruction. By mid-1936 the ADO idea had been accepted throughout the fleet.

The ‘Ideal System’

In 1931 the Royal Navy convened a new Naval Anti-Aircraft Gunnery Committee to devise an ‘ideal system’ to replace the existing one. In effect it was a critique of the HACS; as in the past, it advocated a tachymetric system.⁴⁶ Also as in the past, the main issue was whether the director sights could be stabilised well enough. The committee pointed out that if all ship motion could be cancelled out, the predictor part of the system would function at sea as on land, so the Royal Navy could use the same system the army used (and the predictor part could be tested on land). As in HACS, the director of the ideal system would have an ‘undisturbed’ line of sight, tracking the target without shifting away to compare actual and calculated target motion. As in the existing HACS, the calculating element should be separate from the director, protected by armour both from enemy fire and from the blast of a ship’s guns. It could also be given collective gas protection (an important consideration in the inter-war period). The director should be capable of handling a target moving at a relative speed of 250kts (high for 1931) at 2000ft (greater altitude would imply slower angular rates) and also with an aircraft passing at 2000 yds on a flank or at a maximum angle of sight of 60° when approaching. The system should handle targets at all elevations up to 70°.

The system would measure rates at a particular time, projecting them ahead based on deduced target motion (including whether the aircraft was flying a straight or steadily curved path, and whether it was flying level or climbing or diving). Hopes for a curved-path predictor (assuming that the aircraft was flying steadily along an arc of a circle) seem to have foundered within a few years. DNO accepted this recommendation, and began work on a system called TS I (Tachymetric System 1). Unfortunately this project was incomplete when war broke out; the Royal Navy fought the Second World War with HACS, which had been considered inadequate nearly a decade earlier. It did try to patch on a tachymetric element, in the form of GRU and GRUB, but that was not nearly enough, and it did not solve the basic limitation of assuming that the target was flying straight and level.

The anti-aircraft battery had to develop the maximum possible rate of fire. In practically all cases (except perhaps destroyers) a ship rolled more slowly than the possible interval between shots. Surface ships timed their shots so that they were always at the same point in the roll, but in this case that was impossible: the guns had to be stabilised so that they could fire any time in the roll. Past practice, in which the gun crew kept elevating and training the gun to match pointers controlled remotely by the fire-control system, was inadequate. The crew just could not move the gun quickly enough, since they would be tracking a fast target while compensating for the ship’s motion.

The only solution was to have the fire-control system remotely control the gun: what the Royal Navy later called Remote Power Control (RPC). In 1931, as in the past, the Royal Navy transmitted dial settings using the ‘M’ transmitter, a stepping motor. Although reliable, it transmitted motion as a series of separate jerks, rather than the required smooth control. The ARL proposed a new rotating-field transmitter (later called ‘Magslip’), broadly equivalent to the US Selsyn. The motion involved would be anything but smooth; the mass of the gun had to accelerate and decelerate rapidly and stop as desired. The available alternative means of powering a gun mount were oil engines (as in car transmissions) and electric motors, the committee favouring the first because they contributed much less inertia to the mechanism as a whole. The Royal Navy tried both alternatives when it began to adopt RPC in 1939. The committee urged development of a new medium-calibre anti-aircraft gun suited to RPC, with a high inherent rate of fire.

Guns

As the first Naval Anti-Air Gunnery Committee began its deliberations, officers of the Grand Fleet asked whether new battleships could have secondary batteries combining anti-air and anti-torpedo (i.e., anti-destroyer) functions. The Naval Staff wanted an anti-destroyer gun of at least 5.5in calibre (as in Hood), and there was some question as to whether so heavy a gun could be loaded rapidly enough. The issue could be resolved only by examining a design, so the committee recommended asking the three mounting suppliers (Vickers, Elswick and Coventry Ordnance Works) for sketch designs of 5.5in dual-purpose mountings (Woolwich was later added). Examination of existing sketch designs of 6in, 5.5in and 4.7in dual-purpose mountings the previous year had shown that the 4.7in gun was the largest whose fixed ammunition (shell and case in one) could be manhandled. On this basis the Committee decided to order six 4.7in guns for trials (the order was soon cut to four). The Committee suspected that the 4.7in would fire so much more slowly than the existing 4in gun that it was already somewhat too large. Heavier guns had to be power-loaded. Separating shell and cartridge would slow the rate of fire, since the loading cycle would involve two separate operations. In any case the trunnions would have to be at the rear of the cradle, near the breech, to limit the height of the mounting (which was set by recoil length when the gun was at maximum elevation, and therefore by how far back the breech was from the trunnions). Since the gun would not be balanced, it would need considerable counterweight, which in turn would make for a heavier gun and mounting. Even then the gun and mounting could not be properly balanced at all elevations.

The 4.7in QF Mk VIII (4.7in/40) was the first inter-war Royal Navy medium-calibre anti-aircraft gun. Its design reflected First World War experience. It fired the largest-calibre fixed-ammunition used by the Royal Navy, although the later 4.5in QF Mk I and III fired heavier rounds. This gun armed the battleships Nelson and Rodney (hence would have armed the more numerous capital ships stopped by the Washington Treaty), the Courageous class carriers, the seaplane carrier Albatross and the minelayer Adventure. Two guns are shown on board the Australian seaplane carrier Albatross. (Alan C Green via State Library of Victoria)

Until the Spanish shipbuilding industry was nationalised in about 1935, Vickers had a majority stake in it, designing ships and weapons. It provided Spain with its own variant of the 4.7in/45, which it designated Mk F. These guns armed the two Canarias class heavy cruisers and also the rearmed cruiser Mendez Nuñez; in LA form they also armed other ships. Mk F is shown on board Mendez Nuñez (as presented in an attaché report dated 6 December 1951).

The Committee reported to DNO in June 1919 that it preferred QF (metal cartridge case rather than bagged charge) guns for rapidity of loading, simplicity of ramming and automatic breech closing. It also preferred semi-automatic operation with fixed ammunition. If the mounting was properly balanced it could be hand- rather than power-trained. A maximum elevation rate of 4°/sec was required. In October the committee pointed out that from a different point of view it was desirable to adopt the largest possible gun. Any fire-control solution reflected the motion of the aircraft up to the point of firing. The pilot might well manoeuvre as the shell rose towards him. The longer the shell took to arrive, the better the chance that it would miss altogether. To minimise time of flight, the shell should have the greatest possible muzzle velocity and also should lose its velocity as slowly as possible. The heavier the shell, the better it retained muzzle velocity. For example, at a range of 4000 yds it took a 3pdr HA shell (3½lbs) 14.5 sec to reach an aircraft at 10,000ft (muzzle velocity 2500ft/sec); a 3in shell (16lbs, 6 CRH shape) with the same muzzle velocity took 9.7 sec. Even though it had a lower muzzle velocity, a 4in HA shell (2350ft/sec, 31lbs, 6 CRH) took 7.2 seconds. The heavier shell was also less affected by wind.

All of this was aside from the fact that the danger sphere (lethal radius) of a shell was proportional to its weight (assuming it carried the largest possible burster). On this basis, in October 1919 the Committee recommended the 5.5in gun: it saw no need to limit gun size by demanding that the gun operate manually. That in turn implied that any future dual-purpose capital ship gun would be mounted in gunhouses, which could incorporate power loading and power elevation and training. Power operation in turn made it natural to link the gun automatically to the emerging computer fire-control system. In March 1920 the Committee compared existing heavy anti-aircraft guns: the standard 3in and 4in, the developmental 4.7in (2400ft/sec muzzle velocity), and the proposed 5.5in (2400ft/sec). Time of flight and remaining velocity were tabulated for guns elevated to 50°, firing at aircraft at various altitudes. Maximum altitude for the 3in gun was 15,000ft, to which time of flight would be 20.9 seconds, with a remaining velocity of 550ft/sec. That compared to 11.85 seconds for the 5.5in (1205ft/sec). All guns but the 3in could reach an aircraft at 20,000ft.

Vickers’ drawing of a 4.7in mounting is taken from its 1923 catalogue. (John A Roberts)

Repulse was fitted with the first of DNO’s BD production mountings during her 1933–6 modernisation. The top of the mounting is visible at the foot of her mainmast, protruding from the awning, its director on a pole just forward of the forward of the two triple 4in LA mountings. The ship retained four single 4in guns, and was fitted with the pair of octuple pom-poms standard in capital ships at the time. Note the single 4in HA Mk V gun visible abeam the fore funnel, and also the apertures for the above-water torpedo tubes. The twin BD mountings seem not to have been satisfactory, as they were replaced by single 4in guns during the ship’s September 1938 – January 1939 refit. Had war not intervened, Repulse would have been rearmed with seven twin BD 4.5in mountings. Instead, in February 1941 it was proposed to remove all the 4in guns and replace them with seven of the standard twin Mk XIX mountings. Repulse was lost before anything was done. (RAN SPC)

Factors of comparison were shell power, rate of fire and weight of equipment. Shell power (in terms of lethal volume or radius) was based on experiments at Shoeburyness in November 1919. On this basis the 3in HE shell had a lethal radius of about 28ft, the 4in about 37ft, the 4.7in about 46ft, and the 5.5in about 55ft. On the other hand, rate of fire favoured a smaller gun: 20 rnds/min for a 3in, 9.5 rnds/min for a 4in, 5.5 rnds/min for a 4.7in and 3.5 rnds/min for a 5.5in. A gun would probably fire in 10-second bursts, the number of rounds per burst being, respectively, 4.3, 2.6, 1.9 and 1.6. Finally, the weight of a gun mounting and 200 rounds of ammunition could be compared: 4 tons 17 cwt for the 3in, compared to (respectively) 11 tons 3 cwt, 20 tons 19 cwt, and 35 tons 18 cwt. About the same number of personnel (including six to supply ammunition) would be needed for each gun (the 4.7in required more than the others). The Committee tabulated lethal volume (cubic ft) per 10-second burst per cwt of gun and ammunition and per square ft of mounting. Weight was by far the most important factor, and the 3in gun did best, because it fired fastest. The 5.5in did best in terms of deck area, and also in terms of lethal volume per man; but the number of men was the least important factor.

DNO’s BD mounting reached widespread service in the form of the 4.5in Mk II, shown here on board Queen Elizabeth, 2 June 1943. The gun cradles were bolted together, so that they elevated together, as shown. These weapons armed the rebuilt capital ships Queen Elizabeth, Valiant, and Renown, and the armoured fleet carriers. Mk III was a simpler upper deck (UD) mounting. When a dual-purpose destroyer gun was needed in 1941, the BD mounting was the obvious choice. Although it was much heavier than the Mk III, it offered RPC and power loading, the latter important in a lively ship. A Mk II was tested on board the destroyer Savage, and a modified Mk IV mounting developed for the ‘Battle’ class. The 4.5in mounting superseded DNO’s twin 4.7in BD mounting, which figured in sketches of small battleships proposed in 1928. In those it was mounted together with the twin 6in secondaries introduced in the Nelsons.

Another factor was the effect of varying atmospheric conditions as the shell climbed. The heavier the shell (the larger the gun), the less it would be affected. Similarly, a heavier shell would be less affected by wind – and it was unlikely that wind at altitude would be known. The Committee compared the effects of these factors with the diameter of the lethal zone for each shell. It factored in errors in fuse burning (mechanical time fuses were not yet available). An error in fuse timing would have the greatest effect on the shell moving fastest, since that shell could travel furthest during the excess time the fuse burned. The error was also worsened by the spin of the projectile, the 3in and 5.5in spinning slightly more slowly than the others.

The Committee concluded that ‘gun for gun’ the heavier gun was superior on grounds of accuracy and economy of personnel. However, weight the lighter gun from a barrage point of view. The 3in was rejected for its limited range. On the basis of numbers, it seemed that the whole secondary armament of battleships and the main armament of cruisers should be adapted for AA fire. That would leave more space for ‘pom-poms, sound indicators, etc’. In theory a battleship might have some single-purpose secondary guns, ‘but there could be no object in having two different types of gun and mounting of the same nature’.

A twin 4.5in BD mounting: front view, and rear view with the cupola removed (the structure visible is all below decks).

Unfortunately no dual-purpose gun design was immediately ready for the projected (but abortive) new capital ships. The secondary battery choice made at the time carried over to the new Nelsons, which were effectively smaller versions of the aborted ships. They therefore had the new 4.7in high-angle gun plus 6in secondary guns intended to deal with attacking destroyers. After that the British ordnance industry was fully occupied developing 8in and then 6in cruiser mountings, so the battleship sketch designs developed in the late 1920s (for the expected new construction when the Washington Treaty ‘holiday’ on battleship building ended) also had a combination of 6in secondaries and 4.7in high-angle guns. However, they did have the latter in new cylindrical between-decks mountings. After successful tests on board the battleship Resolution, a production version was produced for the battlecruiser Repulse, then being modernised.⁴⁷ The 1931 Naval Anti-Aircraft Committee supported this project.

By 1932, three different HA mountings were being designed, all using guns with the same ballistics as existing ones: an improved between-decks (BD) 4in mounting for Repulse; a twin 4.7in BD mounting for Nelson and Rodney; and a twin 4in weather deck (WD) mounting for fifteen battleships, Hood and possibly 8in cruisers. The latter was an admission that the BD mounting required so much below-decks structure that installation required total reconstruction of a ship. Twin mounting offered maximum firepower in given deck space: the Mk 19 twin 4in mounting had much the deck footprint of the existing single 4in gun.⁴⁸ It fitted in well with the coming anti-aircraft rearmament programme, one of whose early goals was to double anti-aircraft firepower at minimum cost in time and material. Power loading had to be given up. That in turn limited any fixed round to about 63lbs, the heaviest which could be hand-loaded and rammed uphill with the gun nearly vertical (anything more would require separate shell and cartridge, and power ramming). However, the resulting mounting had considerable inertia, hence was difficult to manoeuvre quickly enough to match the movement of a fast target. Ultimately it required full power operation (other than loading). In addition to modernised ships, Mk 19 armed new cruisers, some new destroyers, and other units. Sea trials were conducted on board the sloop Fleetwood in 1937. It took about a year to solve the teething problems.⁴⁹

The wartime British naval shipbuilding programme had to cope with delays in gun mounting production. They affected the 5.25in BD mounting used on board the King George V class battleships and the Dido class cruisers. To overcome that delay, two Didos were armed with the Mk III UD (hence much simpler) twin 4.5in mounting. The mountings involved had been manufactured to arm ‘D’ class cruisers, which would have been converted into anti-aircraft cruisers had the war not intervened. Scylla is shown in 1942.

The twin 4.5in Mk III as shown in the manual. (By courtesy of John Lambert)

Was the 4.7in powerful enough? In 1933 tests were conducted, bursting HE shells in flight against air targets to gauge effectiveness. The existing 4.7in (new design) and 4in (new design) were compared to a new 5.1in shell (70lb shell, 2500ft/sec) and to the US 5in/25 (70lb shell). If the new 4.7in was set at 1.00, the rating of the 5.1in was 1.27. Against that the US gun was rated at 0.79, and the new 4in at 0.5 – which made a twin 4in equivalent to a single 4.7in. Against that, the 5.1in could not be fully exploited because its fixed ammunition was too heavy (108lbs); trials in a destroyer had shown that it was awkward for one man to handle. To get the most out of the 5.1in, a still heavier round would be needed (a heavier shell or higher muzzle velocity). It was unlikely to be better than the 4.7in. Fleet experience showed that the rates of fire with existing designs were 13 rnds/min for the 4.7in and 20 for the 4in, not counting delays due to fuse-setting dead time.

On this basis an experimental 4.7in BD mounting was being designed in 1933 for planned trials in Nelson in 1936. If they succeeded, two single 4.7in would be replaced by 4.7in BD twins in both Nelson and Rodney. Meanwhile 4.7in designs with 55lb or 60lb shells were being considered. The 4in BD mounting tested in Resolution had failed to meet requirements, in that its loading rate was only eight rnds/gun/minute. There had been no time to fix the problem before producing two more mountings for Repulse, although they might be modified once installed. Two hand-loaded BD mountings were to be installed in Renown during her long refit (she was completely rebuilt, with different guns, instead). The prototype hand-loaded 4in WD mounting would be fitted in Iron Duke in 1934 for trials. Tests in Iron Duke (September 1934) were successful. A companion single Mk XX was planned for ships which could not accommodate the twin mounting. The prototype twin 4in ran sea trials on board the sloop Fleetwood in 1936.

By this time the British were beginning to mobilise. They had to produce as much as possible within stringent industrial and financial limits – much of the vast industry which had supplied so many ships and guns and shells before and during the First World War was gone. Progress in Naval Gunnery 1934 (describing developments in 1933) pointed out that, due to considerations of weight, space, expense, ammunition stowage and ammunition supply (production), capital ships and cruisers would have to retain the existing 4in guns – four twins in each of fifteen capital ships and two twins in each 8in cruiser, in each case a twin replacing an existing single mounting.

The most powerful of DNO’s BD mountings was the 5.25in, shown here on board the battleship King George V. Note that the two guns elevate independently. (Alan C Green via State Library of Victoria)

Internal arrangement of a twin 5.25in mounting. (Photograph by Richard S Pekelney, Historic Naval Ships Association, courtesy of Mr Pekelney).

In 1934 DNO decided to reduce the 4.7in gun to 4.5in, firing a 55lb shell. It was now known as the QF 4.5in Mk I, and was to be produced in both single and twin BD versions, the latter for certain capital ships. Trials were ordered, and it was decided not to mount the new gun in the Nelsons, presumably in hopes that they would receive more complete reconstruction (as described in their Covers). The single 4.5in gun was abandoned as there was no current requirement for it. The prototype BD twin was installed in Iron Duke in the summer of 1936. Trials showed that it could fire 11 rnds/gun/min.

The 5.25in gun had enough of a punch to make it a viable surface weapon, and it could be fired fast enough at high angles (using power loading) to give it effective anti-aircraft capacity. Using a single calibre made a small general-purpose cruiser viable. The result was the Dido class, which was not considered an anti-aircraft cruiser (the Royal Navy considered the 4.5in a better pure anti-aircraft weapon). This was much the same logic which led the US Navy to adopt a dual-purpose 6in gun for its planned prewar 8000-ton light cruiser (the tonnage of which was limited by treaty). The Royal Navy considered arming its own 8000-tonner with 5.25in guns, but turned back to single-purpose 6in guns and 4in anti-aircraft guns in its Fiji class. Euryalus is shown in 1941.

There was still interest in something more powerful. In 1935 a new 5.25in gun firing an 80lb shell with a separate cartridge was proposed. Preliminary trials were successful, and DNO reported that if further trials confirmed that, this gun would supersede the 4.5in. Trials having confirmed that the gun could fire separate AA ammunition, a twin mounting was designed, and in 1936 it was scheduled for trials on board Iron Duke late in 1937 or early in 1938. Both it and the 4.5in were used as dual-purpose capital ship secondary weapons.

Meanwhile the Admiralty became interested in the army’s new 3.7in anti-aircraft gun as a possible weapon for armed merchant cruisers and patrol vessels which could not accommodate a 4in gun. It was theoretically 50 per cent more effective than the existing 4in Mk V, which was not to be retained in service. Nothing came of this project; the army needed its guns far too badly once they were being produced in quantity.

The 1931 Naval Anti-Aircraft Gunnery Committee proposed a variety of anti-aircraft battery improvements, but money was too short. Even so, much effort went into providing the fleet with the new high-angle control system. By 1933 the British government accepted that war was on the horizon, and efforts to provide sufficient fleet anti-aircraft firepower accelerated with the Mediterranean (Abyssinian Crisis) war scare in 1935–6. By that time the services were submitting proposals to the Defence Requirements Committee.⁵⁰ The navy divided its proposal into two classes, of which Class I was fleet requirements and Class II requirements for trade protection. Class I was counted as making up for known deficiencies in the main fleet, such as those which had been identified by the 1931 Naval Anti-Aircraft Gunnery Committee.⁵¹ There were two stages. First came major fleet units (capital ships, carriers, and cruisers). Second came existing sloops and minesweepers, which were needed at the least to ensure that the fleet’s bases remained usable.

Class II, which was much more expensive, went beyond the deficiencies of the existing fleet to deal with the threat of air attack against trade. Thus it included conversion of existing old cruisers and destroyers for trade protection and building up a reserve of weapons and other equipment which would arm auxiliaries in wartime. The sub-committee on Defence Policy and Requirements approved the Class I programme in April 1936. It approved the Class II proposal at its 40th meeting (24 June 1937) subject to further Treasury approval.⁵²

The twin 4in Mk XIX became the standard Royal Navy medium anti-aircraft gun of the Second World War. This mounting was photographed on board a Canadian ‘Tribal’ class destroyer after the war. It carries the radar of a US-supplied Mk 63 fire-control system. The rear view shows a pair of fuse-setting machines, one on each side. Some wartime mountings lacked them. The mount was simple because it was not powered (except in RPC form) and because it had no integral ammunition hoist: ammunition was passed from a fixed hoist into the back of the mount. (MarComm Museum)

A twin 4in Mk XIX in action, showing the loaders. (MarComm Museum)

It did not help that, unlike the US Navy, the Royal Navy never standardised on its medium-calibre guns or mounts. Thus there were multiple types of 4.5in capital ship mountings; 4.7in (50- and 62pdr) destroyer mountings, single and twin; and 5.25in battleship mountings. Unfortunately this variety persisted as the British began to rearm in earnest in 1937. Not only was production of existing designs slow, but design work was delayed. The problems encountered by the King George V class battleships (when a new twin 14in mounting was suddenly required) are well known, but the destroyer programme encountered serious delays in the supply of twin 4.7in mountings, both that in the ‘Tribals’ and the entirely new between-decks type in the ‘L’ class.

Twin Mk XVI guns on a Mk XIX mounting, from the 1945 Gunnery Pocket Book. The fuse-setting machine on the left has been omitted for clarity, as has the divider between the guns. (Photograph by Richard S Pekelney, Historic Naval Ships Association, courtesy of Mr Pekelney)

Close-Range Weapons

Smaller guns were also needed. In an 11 February 1920 memo, the Naval Anti-Aircraft Committee recommended development of a multiple pom-pom to deal with low-fliers attacking with torpedoes and explosive boats controlled by aircraft.⁵³ The mounting should be director-controlled. The committee proposed that Chatham Dockyard cut down the mounting of the standard 11in anti-submarine howitzer to take six 2pdr pom-poms spaced 3ft apart, the original training gear being retained but modified for director control. Guns should be arranged so that their lines of fire could be made to diverge to fill a desired area with projectiles (the committee thought the natural inaccuracy of the gun, which should be tested at ranges up to 3000 yds, might suffice to give the desired coverage). The Director of Naval Construction calculated that an Inconstant class cruiser could take one mounting in place of one set of torpedo tubes, with little modification. The cruiser’s director could be set up to control the gun mounting.

Trials compared 2pdrs and 3pdrs to decide which was the smallest (for maximum rate of fire) sufficient to defeat a torpedo bomber with one hit. Either was enough if it hit the engine, fuselage, or wings inboard of the outer struts; neither would suffice if it only hit the outer wings. The 3pdr offered a slightly greater chance that a fragment would hit the aircrew even if a direct hit on the outer wing failed to crash the bomber.⁵⁴ That was not enough to disqualify the 2pdr.

Maximum acceptable weight for the multiple mounting was set at that of the 3in HA gun which then armed destroyers: 2 tons 12 cwt.⁵⁵ Rate of fire was to be not less than 60 rounds per minute per barrel.

The initial pom-pom director was a simple dummy gun, the virtue of which was that it removed the aimer from the noise and vibration of the mounting. This is the Mk II version. The straps allowed the single operator to turn the director with his body. The cartwheel sight made it possible to estimate deflection. Mk II can be compared to the slightly later US Mk 44, which was also a very simple non-computing director. Director control would be exerted by the usual follow-the-pointer technique, the mounting being moved by a continuously-running electric motor which the crew could clutch in and out. The mounting should elevate and train at 15°/sec, with a maximum elevation of 45° (as yet there was no vertical-diving threat) and a maximum depression of 15°. As usual, both Vickers and Armstrong (Elswick) were asked to design mountings. Armstrong’s, which was designed for continuous rather than burst fire, was rejected as too complex, and a mock-up of the Vickers mounting was examined at Vickers in July 1923. Rate of fire was better than required (90 rounds per gun per minute) and maximum elevation was 80° rather than 45°. Rate of fire was 90 rounds per barrel per minute, compared to the required 60. The prototype mounting passed its shore trials in 1927 and it was successfully tested on board Tiger in 1928.

The most important British automatic weapon developed before the Second World War was the multiple pom-pom. This octuple one was on board Shropshire. The photograph is supposed to date from 1942–4, but note that the gun mounting is entirely unshielded, and that it is unmanned while it is being loaded. These guns were power-operated, which is why Prince of Wales suffered so badly as soon as she lost power. (State Library of Victoria)

The maximum allowed weight allowed for too few barrels. Ultimately a 15-ton eight-barrel mounting was designed for capital ships. The barrels were mounted so that they could diverge to spread their fire.⁵⁶ Guns had to fired in sequence to avoid jarring the mounting. Initially they were fired in symmetrical pairs, but in later mountings they could be fired singly. The rate of fire was limited to limit the rate of erosion due to bore heating (the 2pdr was considered good for only 3000 rounds). Also, a much higher rate of fire would have required complete redesign of the gun.⁵⁷ Guns were belt-fed from 150-round loading trays, which could be reloaded to maintain a rate of fire of 28 rounds per minute. Thus the gun could fire at 90 rnds/min for 2 minutes 25 seconds or at 108 rnds/min for 1 minute 50 seconds. Any increase in tray capacity would enlarge the mounting.

The quadruple pom-pom was designed specifically for destroyers and cruisers; this one is aboard the destroyer HMAS Napier. This version had separate layer and trainer, and each had his own computing sight, presumably a US-supplied Mk 14. Note the splinter shield in front of the mounting and also the protection for the ammunition belts. (State Library of Victoria)

Like Mk II, it fired existing 2pdr ammunition, large quantities of which remained after the First World War. That limited it to a muzzle velocity of 1900ft/sec, which by the late 1930s was clearly insufficient. The first operational mountings were ordered under the 1930–1 Estimates: twelve Mk M octuple mountings and six directors.⁵⁸ The initial eight-barrel version was designated Mk M and later Mk V.⁵⁹ During the 1930s a modified Mk VI mounting was ordered; it was standard during the Second World War.

By 1927 it had been decided that rounds should all be HE with sensitive-enough fuses to burst when hitting the fabric of an aircraft. Since the shell would not make a visible burst unless it hit, a proportion of ammunition had to be tracers. For initial trials the prototype mounting was connected to an Adventure-type director. Based on earlier trials in the cruiser Dragon, it was thought that a single control officer could lay and fire the mounting and also spot, but that proved to be too much for one man. A target moving almost directly at the mounting was relatively easy to hit, but it was much more difficult to hit a target with a high crossing speed; at any great range tracers gave a misleading impression. Spotting would have been easier if shells had been time-fused, bursting at the expected range. However, that would have required an automatic fuse-setter, an unacceptable complication.

Within a few years the Royal Navy considered several alternative ways to aim light anti-aircraft guns. Eyeshooting, which it eventually much favoured, had the gunner visualise the future position of the target and point at it using either a telescope or a forward area sight (an open sight). He might base aim on an estimate or he might use a telescope supported by some form of rate estimation. Any such method had to allow the gunner to override and select an alternative point of aim if the aircraft manoeuvred. Alternatively, course or speed sights could be set by a control operator for inclination, dive, and speed during an attack.

An alternative was hosepiping. No sights were used, the gunner relying on tracers to mark the trajectory of his rounds. In effect he was moving a hose of tracers. He relied on this visual aid to bring the trajectory onto the target and to keep it there. Lewis gunners used hosepiping because at very close range (500 yds and less) it was as efficient as eye-shooting and required less skill. To support it, the usual proportion of tracers was one to four rounds of ball in a Lewis gun.

The key argument favouring high muzzle velocity was that the target was not brought under fire until the time of flight of the initial rounds elapsed. If it jinked, it could not be brought under fire until another time of flight interval elapsed. In 1936 the DNO wrote that ‘from the control point of view reduction of time of flight is more important than any other single factor’. Time grew shorter and shorter as aircraft performance improved sharply through the 1930s.⁶⁰

By 1928 Elliott Bros. was working on a prototype director. It was a simple sight on which allowance could be made for own speed, for the speed and course of the target, and for tangent elevation (range). The sight was laid and trained by one man, and data would automatically be sent to the mounting. It was understood that the weapon was intended mainly to defeat torpedo bombers. They placed themselves in a vulnerable position by flying low and straight for at least 15 to 20 seconds before releasing their torpedoes. That raised a problem. If the gun opened at excessive range it was unlikely to hit, but it might easily exhaust too much of the ammunition in a belt. Yet tracers had to be expended in order to tell that aim was correct for line, before the bomber settled down into its final run. DNO recommended firing a short burst (to establish line) at 2000 yds, then holding fire until about 1500 yds. By 1933 experience showed that to achieve 70 per cent hits it was necessary to know the range within 200 yds, inclination within 10–20° and speed within 20kts.

By this time there was a Mk I pom-pom director (initially described as a Director Sight). Unlike the director of an HACS, it was not connected to any sort of below-decks computer. It was a simple remote open sight which could transmit elevation and train orders to the Mk V pom-pom mounting. The back sight was set for enemy speed, dive or climb, and inclination. The developers recognised that such settings could not be correct for all ranges because the ratio of time of flight to range would vary with range. To deal with that problem, the foresight could be moved back and forth for range using a cam groove rather than a straight line. By 1934 directors were being modified so that enemy speed could be set more quickly, and also so that it could be released and reset to zero to deal with dive bombers – the emerging threat. The mounting was being altered to give a larger vertical pattern of shots at the expense of the lateral pattern – it was easier to aim for line than for range. Mk I* embodied detail improvements, but it too was basically a dummy sight. In 1935 DNO pointed out that Mk I was not suitable to engage dive bombers. Until the better Mk II was available, and new gun sights had been fitted, ships would have to rely on local control to fend off this kind of attack. Looking back in 1937, DNO wrote that Mk I had been of little use; control officers had to rely on eyeshooting.

A quadruple pom-pom on board the cruiser Bellona after the war. (State Library of Victoria)

A quadruple pom-pom from the 1945 Gunnery Pocket Book. Note that it has been modified for one-man operation, using a joystick. A major wartime lesson was that the usual two-man control (layer and trainer [pointer and trainer in US parlance]) was ill-adapted to fast-manoeuvring targets. That applied to directors as much as to gun mountings. (Photograph by Richard S Pekelney, Historic Naval Ships Association, courtesy of Mr Pekelney)

Mk I lacked a rangefinder, but by 1933 (as above) range was clearly a vital input (by 1937 arrangements were in hand to provide all existing pom-pom directors with 1m rangefinders). In 1934 Mk I** production was stopped in favour of a new Mk II with a 4ft rangefinder and a change of range in time of flight mechanism (range integrator). Just as importantly, Mk II incorporated a forward area sight, the Control Officer’s Forward Area Sight (COFAS). While layer and trainer kept their telescopes on the target, the control officer could aim off using this sight to provide deflection due to enemy movement across the line of sight. COFAS made eyeshooting far more effective, because the control officer was pointing his sight in the direction he envisaged. Wind correction was entered by a sight-setter. That left the control officer free to concentrate on target speed and course. As delivered, Mk II offered synchronous (M-type) rather than step-by-step data transmission to the gun mounting. The ultimate requirement was full power control (RPC). As of 1936 a power-control gear for multiple pom-poms was being tested. Overall director design was kept as simple as possible, first deliveries being expected before April 1936. Given the gap between Mk I and Mk II production, some ships would be delivered without directors in 1935–6. In fact Mk II proved difficult to make, and production was unsatisfactory. Mk II* had Magslip transmission, and Mk IIA had an improved COFAS.

Mk III was a Mk II modified for high- rather than low-velocity ammunition (Mk III* had Magslip transmission). According to the May 1939 edition of Progress in Naval Gunnery, Mk II had not come up to expectation. Production of it and Mk III would continue only until capacity could usefully be switched to the fully tachymetric Mk IV. During the war Mk IIIs were rebuilt with tachymetric inputs as Mk IIIT.

The companion to the pom-pom was the quadruple 0.5in machine gun, seen here before the war aboard Exeter during a visit to the United States. Like its predecessor the single pom-pom and its successor the Oerlikon, this gun was controlled entirely by its gunner; it had no director. It is not clear to what extent this gun was exported just before the outbreak of war. It apparently armed the Argentine Santa Cruz class destroyers built in the UK. It was to have armed the Brazilian destroyers which became the British Havant class as well as the new Turkish destroyers. It may have armed at least two of the rebuilt Greek destroyers. A British official list of warship data (1939) mentions quadruple Vickers machine guns on board several ships, including the Soviet battleship Marat (three quadruple and two single machine guns). Vickers guns were also listed on board several of the old cruisers, though not on Kirov.

While Mk II was entering production, a more elaborate Mk IV, with a tachymetric training unit, was ordered for trials. Its core was the Automatic Control Officer’s Forward Area Sight (AUTOCOFAS), which used tachymetric measurements to set lateral deflection.⁶¹ AUTOCOFAS was fully (locally) gyro-stabilised. The director would automatically feed range, enemy movement, and own ship movement into the AUTOCOFAS calculator, so that the control officer could concentrate on keeping the gun on target. This director had Magslip transmission. Mk IV* was the initial version, a converted Mk II or III. In 1940 the first Mk IVs equipped the battleship King George V and the carrier Courageous.

Mk IV was still rather elaborate, with a four-man crew. Ultimately DNO wanted something simpler, based on a ‘disturbed’ line of sight. Layer and trainer would aim off, following the target. There would be no need for independent operators to obtain and supply deflection. That was how the later US Draper systems (such as Mk 51) worked, and it was how the late-war British Simple Tachymetric Director (see below) worked. DNO hoped to begin this development once the larger Tachymetric Control System Mk I was far enough advanced.

At about the same time that Mk IV was being designed, it was decided that the 2pdr shells should be self-destroying at 2500 yds. That would allow ships to engage aircraft attacking between columns or from the direction of screening ships, but it was much more important as a deterrent. It turned out that an attacking pilot might well ignore tracers climbing towards him, but he would see the self-destroying rounds as they exploded, much as he might see time-fused bursts from heavier guns.

Tracers were associated with hosepipe control. As it emphasised eye-shooting and the associated form of director control, the Royal Navy moved away from tracers; they would only confuse the control officer. Late in 1940 trials were run to see whether that was a good idea, and also to compare local with director control.⁶² It turned out that the combination of tracers and director control was best, except when tracers could not be observed, when fire had to be opened at very short range (so no delay was acceptable), when pointer-following errors were likely to be excessive (due to heavy rolling), and with high-velocity guns using Mk II directors. Tracer was relatively ineffective for locally-controlled guns because the gun layer and trainer needed so much skill to follow the target, estimate and superimpose aim-off (deflections), and at the same time observe tracer and apply corrections. It was also difficult or impossible for the gun crew to see the tracer through smoke.

Quadruple 0.5in machine guns on a Royal Australian Navy cruiser show the splinter shields added in wartime.(State Library of Victoria)

Poor results using director fire were mainly due to errors in following pointers controlled by the director. The solution was RPC, which was being fitted to cruisers and larger ships in conjunction with the new Mk IV director. Both the Metro-Vick-Elswick electro-hydraulic and the ‘Metadyne’ electric systems had recently passed their tests. New director rangefinders (FV 3 instead of FV 2) were being supplied, and a programme to fit Type 282 radars was also in hand. New ship designs would have their pom-pom directors as close as possible to the guns, to eliminate convergence errors.

A lightweight quadruple Mk VII mounting was developed for cruisers and destroyers. In 1932 it was expected to weigh 6¼ tons, still far too much for existing destroyers and surviving First World War cruisers, but acceptable for more modern cruisers. It also appeared on board ‘Tribals’ and some later destroyers, and some carriers. Development continued even though the 1931 Naval Anti-Aircraft Gunnery Committee considered the quadruple pom-pom useless (it wanted a new gun, which emerged as the 0.661in). The committee considered the octuple mounting just good enough to be worth retaining. Once rearmament began, the quadruple pom-pom was easier to produce in quantity than a new mounting, because it had such commonality with the octuple 2pdr already in production. About 1935 tests showed that the quadruple 2pdr had considerable potential against a new threat, the motor torpedo boat (MTB), which became much more prominent during the Ethiopian crisis in the Mediterranean (the Italians had large numbers of MTBs). That was why the ‘Hunt’ class destroyer escorts were ordered modified to accommodate this weapon (with unfortunate, albeit unforeseen, effects on their stability). As completed, Mk VII was overweight: A typical Mk VII* weighed 8.6 tons. By way of comparison, the single 4.7in CP Mk XVII in a ‘G’ class destroyer weighted 8.8 tons. The prototype quadruple 2pdr was tested aboard the cruiser Shropshire in 1934. This prototype was transferred to the destroyer Crusader for 1935–6 trials which confirmed that it was suited to the new ‘Tribal’ class.

The destroyer installation did not include a director, and without it the pom-pom was considered unlikely to get many hits at over 1000 yds range. During discussions leading up to the construction of the ‘Tribal’ class, it was pointed out that the value of the pom-pom might lie more in the enemy’s knowledge that it existed than in its chance of actually shooting down aircraft. Trials on board the ex-battleship Centurion suggested not more than 5 per cent hits, and the pom-pom would have only 10 seconds of firing time before a dive bomber released its bomb. A trial using a fixed position suggested a hitting rate as high as 16 per cent, but that would decline for a lively ship.

The 1919–21 Committee also wanted a shorter-range back-up, a replacement for the wartime 0.303in (rifle calibre) Lewis gun. It had to be more powerful, because future aircraft might well be armoured.⁶³ The result was the quadruple 0.5in Vickers machine gun. Compared to the pom-pom, it had a much higher rate of fire (600 rounds per barrel per minute, or 2400 total, compared to 720 for the eight-barrel pom-pom), but the rounds were far less destructive and the effective range shorter. Effective 0.5in range, as evaluated in 1932, was about 1000 yds (out to 1800 yds time of flight was less than for the 2pdr, but bullets were not dense enough to do much damage). The gun was far inferior to the multiple pom-pom, but also far better than a single 2pdr. Many ships retained twin Lewis gun mounts, which were set up on a temporary basis.

The 1931 Naval Anti-Aircraft Gunnery Committee considered the 0.5in gun inadequate; it wanted a new gun, which would also replace the quadruple 2pdr. Nothing happened initially, but in 1934 DNO announced that ‘to meet improvements in aircraft and new methods of attack’ a new close-range weapon was being developed, capable of disabling an attacker at a mean range of 1200 yds in 6 seconds. He was interested in both an improved Mk M and an improved machine gun firing solid bullets. The latter had to offer reduced time of flight (a heavier bullet and/or higher muzzle velocity). The machine gun requirement led to a 1935 order for a Vickers prototype 0.661in gun (for use in a sextuple mounting). Muzzle velocity was 3125ft/sec. Time of flight to 2000 yds would be 60 per cent that of the 0.5in machine gun. Expected rate of fire was 300 rounds per gun per minute, each shell weighing 3oz.

The 0.661in grew to weigh so much more than the 0.5in machine gun that it was compared to the quadruple pom-pom instead. Thus the 0.661in fired 1800 rounds per minute, compared to about 400 for the quadruple pom-pom; on the other hand, each of its shells weighed less than 10 per cent as much, so the weight of fire was only about 45 per cent that of the pom-pom (though far more than that of the 0.5in). Moreover, the sights planned for the 0.661in were considered much inferior to that of the pom-pom. This mount came too late for the ‘Tribal’ and ‘J’ classes, but in 1936 DNO proposed that the next (‘L’) class have two such mounts in place of their pom-poms and 0.5in. Like other weapons of the period, the 0.661in continued to grow in weight. Originally weight was to have been 2.75 tons, and in various versions of the ‘L’ class design the assumed weight was 3.4 tons. However, at the mock-up stage in 1938 expected weight was 4 tons, nearly half as much as the pom-pom. The decisive argument against the 0.661in may have been its inadequacy against MTBs, given its light bullets and short effective range. In January 1938 the Assistant Chief of the Naval Staff said that he would prefer two pom-poms to one pom-pom and a 0.661in. The sextuple 0.661in was cancelled in 1938.

The improved Mk M was more fruitful. At the outset DNO wanted a larger pattern, which would normally mean more barrels, and/or a higher rate of fire. He also wanted to reduce time of flight, which would reduce the effect of control errors. Weight and loading considerations made it unlikely that the number of barrels could exceed the current eight, and it seemed unlikely that the gun could fire faster. For the 2pdr, that left muzzle velocity. By 1935 DNO had considered and rejected designs with velocities of 2500 to 3000ft/sec because they would entail much more weight, and because they would also entail longer cartridges and therefore slower fire. He became interested in smaller-calibre guns (1½pdr and 1pdrs) with high velocities; the 1pdr seemed to be the smallest for which a satisfactory shell-fuse combination could be produced. Further investigation showed that the 1pdr gun and mounting would be nearly as heavy as the 1½pdr, and that the 1pdr shell probably would not be effective; it was abandoned. In 1935 DNO was designing an octuple 1½pdr for trials, carrying 80 sec worth of ammunition on the mounting (130 shots/gun/minute, 2600ft/sec). It seemed that the 1½pdr would replace the 2pdr, and the 0.661in the 0.5in.

Like the 0.661in, the 1½pdr kept growing to the point where work had to be suspended in 1937. Effort shifted to improving 2pdr ballistics and rate of fire, the goal being a muzzle velocity of 2300ft/sec, which would reduce time of flight to 2500 yds to 5.5 sec. In 1937 DNO announced that he hoped to have an improved-performance 2pdr in production towards the end of 1938, and also hoped in due course to modify the existing gun for higher muzzle velocity and higher rate of fire, using sturdier new gun components.

By 1938 efforts were being made to improve the pom-pom with higher-velocity ammunition and automatic rather than controlled fire (i.e., the trigger did not have to be held down) for the octuple mount. Elevating and training control of the quadruple pom-pom was considered incompatible with fully automatic fire, so that mounting retained controlled fire. The octuple mounting had priority for the new ammunition.⁶⁴ By early 1941 high-velocity ammunition still was not universally available. Trials conducted at that time confirmed that conversion to high velocity offered a considerable improvement in hitting, even though in that case the gun would have to rely on local control, the director being used only to indicate target bearing: the locally-controlled high-velocity gun did better than a fully director-controlled low-velocity gun using RPC. By mid-1942 most 2pdrs in the fleet had been converted to fire high-velocity rounds as Mk VIII* guns.

Trials were run in 1938 to determine characteristics of a light eye-shooting close-range weapon (inside 1000 yds range), to be used particularly by ships too small for multiple pom-poms. At one end of the scale was the single 2pdr; other possibilities were a single or possibly twin 0.8in (solid or explosive round), quad 0.5in and finally multiple 0.303in. It emerged that the 2pdr was most effective, though it would need a new mounting. The 0.303in was completely outclassed. A drum-fed 0.8in offered tactical advantages, in that it could fire for 8 seconds and reload in 12 seconds, compared to 20 seconds for the 0.5in (45-second reload). It turned out that there was no time to develop a new gun.

Commercial Fire Control

In March 1920 Captain Usborne (president of the Naval Anti-Aircraft Gunnery Committee) proposed that DNO seek a fire-control system from Vickers, which was, thus far, the only British firm known to have make a complete study of the AA gunnery problem. Vickers was already working on a director and on fittings on gun mountings to be used for fire control, and it had experience in the design of periscopes. Work did not go entirely smoothly: within a few months Vickers was asked to stop work on the director while it was re-thought.

Vickers proposed its AA Predictor some time in the spring of 1920. The Committee concluded that if suitably modified it could calculate vertical and lateral deflection given measured vertical and lateral angular velocities, which in turn could be measured by Professor Sir James Henderson’s ‘rate of change of bearing and altitude meter’, which used his constrained gyros – it was a tachymetric device. Vickers submitted a modified design in October 1920. It incorporated the desired ability to predict fuse time, and the Committee thought that it might be worth adopting. Usborne asked DNO to pay for production of a prototype.

Vickers apparently soon abandoned work on the Predictor, resuming it in 1924; it produced a prototype in 1928. By this time it had abandoned measuring gyros in favour of comparing generated angles (produced mainly by integrators on the basis of estimates) with what was observed (a process called ‘tuning’). The Predictor also measured the range rate by tuning. As in the later HACS, the range operator adjusted the range rate to keep the target at the observed height as it moved (as the sight angle changed). Once rates were determined by ‘tuning’, they were fed into equations which connected vertical and horizontal deflections to the rates (the equations involved various trigonometric functions of the initial and final sight angles). Time of flight was a function of future sight angle and altitude (i.e., of future range). For any combination of these two parameters it could be obtained from a three-dimensional cam. Assuming constant rates, the deflections were simply the rates multiplied by time of flight. Vickers used two equations, each of which connected the sine of a deflection with time of flight and future sight angle. Instead of solving them directly, which would have been complex (because so many trigonometric functions were involved), it computed each side of each equation. The correct choices of time of flight and future sight angle would cause the two sides of each equation to match. The Predictor displayed the difference between the two sides of each equation, and the operator ‘balanced’ them by turning a wheel until the difference was zero.⁶⁵

It took 3 to 4 seconds for the operators to reach a solution by tracking an aircraft. Once they had measured the horizontal and vertical rates and entered them into the Predictor, it could calculate future angles. Given height and vertical angle (sight angle), the Predictor calculated future range, which gave time of flight of a shell, hence fuse setting. This calculation was done using a three-dimensional cam. Below 10° or 15° angle of sight the system used range rather than height, because at low angles angle of sight did not change quickly enough. None of the rates was really constant, but (as in a Dreyer Table) they could be treated as though they were for a short time. By 1931 the British army was using the Predictor to control medium-calibre anti-aircraft guns. This Naval Predictor, which Vickers exported, was offered to the Naval Anti-Aircraft Committee. It formed the basis of Japanese wartime naval anti-aircraft fire-control systems. Vickers also exported the naval system to other countries, almost certainly including Argentina (for the training cruiser La Argentina) and Spain (for cruisers, presumably beginning with the Canarias class). Vickers also offered the committee a more elaborate device more like the HACS, but it was apparently never exported (and it may never have been built).⁶⁶

More sophisticated devices used current angular velocities and range to estimate target course and speed, and then predicted on that basis. In 1932, when the Naval Anti-Aircraft Committee released its report, the more complex but more accurate technique was used by ARL, by Barr & Stroud, and by Sperry in the United States. Of these, the ARL Predictor was intended for use only ashore. Barr & Stroud’s system was rejected by the Admiralty as too complex.⁶⁷ The Imperial Japanese Navy seems to have bought an earlier, simpler, Barr & Stroud calculator (see below). A majority of foreign armies had tachymetric systems, but in 1931 only Sperry in the United States and Hazemeyer in the Netherlands offered naval tachymetric systems for export.