CHAPTER THREE
The design concept that led to the Boulton Paul Defiant can be traced back to the Bristol F.2B Fighter of the First World War, which combined the performance of a single-seat fighter with the flexibility of a two-seater, with a rear-facing Lewis gun in addition to a forward-firing Vickers. The idea was perpetuated in the Hawker Demon of the inter-war years and reached its ultimate expression in the Defiant.
Having already developed a pneumatic turret for its Overstrand bomber, Boulton Paul were quick to see the advantages of a hydraulic turret developed in France by de Boysson of the Societé d’Applications des Machines Motrices (SAMM) and acquired licence manufacturing rights in 1937. The rapid advance in aircraft structures and performance in the 1930s had led to a reappraisal of aircraft armament, as it was becoming clear that larger gun batteries would be needed to stand any chance of downing the new monoplane bombers. The combination of four 0.303-in Browning machine-guns in the de Boysson turret (instead of the Darne machine-guns used by the French) seemed to be an ideal solution to the problem of delivering sufficient weight of fire onto a target, and J.D. North, Boulton Paul’s chief designer, penned a neat low-wing monoplane to accommodate it. Unlike the Bristol F.2B and Hawker Demon, there was to be no forward-firing armament, as the aircraft was considered to be a ‘bomber destroyer’, which, it was hoped, would not have to contend with escort fighters. The resultant P.82 Defiant, designed to meet Specification F.9/35 was flown for the first time by Cecil Feather on 11 August 1937 at Wolverhampton Airport. The type was eventually chosen ahead of its main rival, the Hawker Hotspur.
In its construction, the Defiant was more advanced than the Hurricane, which was the culmination of the Hart/Fury design stream. At the same time, it was much simpler to produce than the Spitfire. The front fuselage was built up of four L-section longerons, a number of bulkheads and contained the pilot’s cockpit, the floor of which was formed by the upper surface of the wing centre section. The rear fuselage accommodated the turret, which was streamlined fore and aft by fairings made of spruce and three-ply. These were raised or lowered by pneumatic jacks actuated by cams in the turret. Much attention was given to reducing airframe drag, with countersunk rivets being used in the metal skinning to allow a smooth surface finish. The wing was of two-spar construction and consisted of a centre section, two outer panels and detachable wing tips. A semi-elliptical planform was achieved by using a linear taper on the centre section, with a more pronounced taper on the outer sections. Split flaps ran the full length of the wings from the Frise-type ailerons, which (like the elevators and rudder) were fabric-covered. A total of 104 gallons of fuel was accommodated in two self-sealing tanks in the wing centre section. The radiator was mounted ventrally in a similar position to that adopted for the Hurricane.
The turret was hydraulically operated (the pump being electrically driven) and mounted four 0.303-in Brownings in vertical pairs, each gun having 600 rounds in ammunition tanks located under each gun installation. It was able to rotate through 360 degrees and the guns could be depressed to the horizontal, except when facing forwards, when they were restricted to an elevation of 17½ degrees to clear the propeller. The maximum angle the guns could be raised was 72 degrees as the air gunner could not use his sight at higher elevations. Electrical cut-outs were provided to prevent the guns hitting any part of the aircraft.
The prototype Defiant (K8310) was delivered to A&AEE at Martlesham Heath in early December 1937 for an initial assessment. The turret had not yet been fitted, so a metal fairing took its place. Powered by a Rolls-Royce Merlin I, performance testing showed a top speed of 320 mph and the aircraft attained a height of 10,500 ft in 7½ minutes. The handling was acceptable in most respects, the only major complaints being that the ailerons were considered to be too light and the flaps, when fully down, did not produce the desired nose-down attitude on a glide approach. As the cockpit was set further forward than either the Hurricane or Spitfire, access on the ground could be made difficult, particularly if the wing surface was wet. However, the view forwards and downwards was better than on either of the two aforementioned aircraft. The cockpit also proved to be rather draughty with the hood closed.
K8310 was fitted with the four-gun turret in February 1938 and returned to Martlesham Heath for armament trials, with further performance testing taking place the following October. The figures obtained were slightly disappointing as the maximum speed was reduced to 303 mph at 15,000 ft and it took 15.1 minutes to reach 20,000 ft. It had been hoped that the Defiant would have a similar performance to the Hurricane, but its top speed was 17 mph slower and the climb to 20,000 ft took a full 6 minutes longer. It was obvious that more power was needed and the first production aircraft (L6950) was fitted with a Merlin III of 1030 hp and first flown on 30 July 1939. It was delivered to A&AEE (by now at Boscombe Down) on 19 September 1939, together with the second prototype, K8620, for further performance and handling trials.
The provision of a non-slip walkway on the wing root made entry to the cockpit slightly less hazardous, although a spring-loaded step on the fuselage side would have made it easier still. The cockpit was comfortable, with adequate room for a well-built pilot, but in flight a cold draught tended to play on the back of the pilot’s neck. It was also rather noisy, but no more so than contemporary fighters. The gun turret was much more difficult to enter, access being obtained either by sliding back the segments of the transparent cupola, or via a hatch in the floor aft of the turret. The latter method could only be used when the guns were facing forward. The turret was perfectly adequate for an air gunner of small stature, but anyone who was taller than average found the accommodation rather cramped when wearing full flying kit.
The control surfaces were operated by a system of cables, pulleys and levers, which produced little friction in the control circuits and only slight play. The elevator trim tab control was located on the pilot’s left in the forward upper corner of the cockpit and could be moved over its full range in about three seconds. Rudder trim could be applied by moving a quadrant-mounted lever, which was also on the left-hand side of the cockpit immediately aft of the undercarriage and flap quadrants. The full range of movement could be obtained in about one second. No aileron bias gear was fitted.
The throttle control quadrant was conveniently mounted on the pilot’s left and showed no sign of slipping. The propeller pitch could be selected to ‘Coarse’ or ‘Fine’ by a control immediately below the panel on the left-hand side. It was within easy reach, but could be easily mistaken for the boost cut-out control, especially when changing pitch after take-off. The radiator shutter control was located on the right of the cockpit, with the fuel cock on the left. Some difficulty was experienced with the fuel cock sticking, which was modified by Boulton Paul. The flap control was mounted in the throttle control box and could be operated with ease to stop the flaps in any position. An indicator on the left-hand side of the panel was marked in degrees and covered the full flap range.
The control for the undercarriage was located next to the throttle box and was easy to operate. The position of the wheels was indicated to the pilot by red and green lights situated in the centre of the panel–red when the undercarriage was up and green when it was down. A warning horn sounded if the throttle was closed beyond the last third of its movement with the wheels still up. There were two methods of operating the undercarriage in an emergency, consisting of two hand pumps, one operating through a hydraulic system separate to the normal system. Pilots faced with such an emergency were then faced with the unenviable task of applying 200 strokes to the pump before the undercarriage was fully down.
The brakes were pneumatically operated via a lever on the control column and performed well. A standard blind flying panel was fitted and the layout of the instruments was good, with no vibration. For night flying, illumination was provided by two lights on movable arms, one on each side of the cockpit. A dimmer switch was provided for each. They could be set to illuminate any particular instrument or the complete blind flying panel and did not cause any unwanted reflections on the windscreen.
In the prototype Defiant no emergency exit was available from the front cockpit. A sliding hood was fitted, but was found to be almost impossible to open at speeds above 210 mph IAS. Knock-out panels were fitted in the sides of the hood, which were sufficiently large to enable a small pilot to climb out of the aircraft should it come to rest inverted on the ground. Later aircraft were fitted with a hood that could be jettisoned. In a bale-out situation, the air gunner had to leave the turret through the trap door in the floor of the fuselage, but this could only be done with the guns in the forward position. It was considered that the chances of the gunner getting out successfully would depend on his size. A tall man might well have difficulty, particularly if the aircraft was not under control at the time. The gunner’s exit hatch could be opened from the outside should the aircraft go over onto its back on the ground, but it was thought that it would have been extremely difficult for him to extricate himself without assistance.
For a low-wing single-engine monoplane fighter, the view from the cockpit was better than most other aircraft of the period as the pilot’s cockpit was mounted further forward to cater for the turret. Although the view ahead when the tail was on the ground was not good, with the tail up on take-off it was greatly improved, and in flight the view downwards over the leading edge was excellent. No ‘direct vision’ panel was fitted so there was virtually no forward visibility when flying in rain. Instead, the pilot had to open the hood and peer round the windscreen.
Handling trials were carried out at take-off weights of 7220 lb (forward CG), 7390 lb (normal CG) and 7560 lb (extended aft CG). The aircraft was easy to taxi in winds of up to 30 mph and showed no tendency to lift its tail even with CG at the forward limit. The behaviour of the undercarriage on any surface was excellent and the brakes were smooth and progressive in operation. Prior to take-off the flaps were set at 30 degrees. The tail tended to come up slowly during the take-off run and the swing to the left could easily be held with rudder. As soon as the aircraft was airborne the undercarriage could be raised with little change in trim, the whole operation taking around eleven seconds. It was recommended that the flaps should not to be raised until a speed of 120 mph IAS had been reached, as this caused slight tail heaviness. Once in the clean configuration the aircraft could be accelerated to its best climbing speed of 180 mph IAS. Although there was sufficient elevator trim when climbing at this speed, the rudder trimmer did not have enough authority and some right rudder had to be maintained.
The controls were moderately heavy, the elevator being quick in response and effective at all speeds with the engine on, but becoming heavier in the dive. It was also classed as moderately heavy on the glide but in this condition there was a marked deterioration in response and control was not very effective. The force needed to apply the rudder was similar and once again the response was somewhat worse when gliding, although it remained effective. Full stability tests were not carried out, but the aircraft appeared to be longitudinally stable at maximum level speed at normal CG. The level of stability tended to reduce with decrease in speed, until at low speeds the aircraft was slightly unstable. Directionally, the aircraft was stable at all times.
Night flying did not pose any particular problems, although the poor initial rate of climb whilst gaining speed was a little disconcerting at first, especially when there was no horizon. A fair amount of flame was produced by the exhaust, but this did not seriously affect the forward view. The lookout was satisfactory, particularly forwards out of the glass windscreen. In contrast the Perspex hood was prone to scratches, which affected the pilot’s ability to see other aircraft.
During stall tests with the flaps and undercarriage up, the aircraft showed itself to be longitudinally stable when trimmed to fly at 1.2 × stall speed (108 mph IAS). The only warning of the approaching stall was given by the fairly high position of the nose, less effective aileron control and slight tail buffet. The aircraft stalled at 90 mph IAS with the control column almost fully back, but it was not particularly easy to accomplish this as the elevators were not very effective at low speeds, and although the forces involved were not great, the limited amount of elevator travel caused some difficulty. When it did occur, the stall was fairly gentle and generally level, with an occasional tendency to drop either wing quite sharply, but this could be raised by coarse use of rudder. If the wing was allowed to drop, the speed increased and the wing could be raised by aileron. There was no tendency to spin.
With the flaps and undercarriage down the stall occurred at 76 mph IAS, the nose being less high than in the clean configuration. As a result, the control column only needed to be brought back to the three-quarters position. The stall was more marked and a wing, usually the left, was much more likely to drop. If the control column was pulled right back, the aircraft developed a fore-and-aft pitching with variations of speed from 70–85 mph IAS and with considerable tail buffet. Each wing tended to drop alternately, but could be raised by coarse rudder. The aircraft was very difficult to control under these circumstances, but once again there was no sign of a spin developing.
When the flaps were lowered on the approach the aircraft became slightly nose heavy, but this could easily be held. The best speed for the approach was 90 mph IAS. Although the landing was reasonably straightforward, the lack of elevator control at low speed meant that it was rather difficult, but not impossible, to get the tail down into the three-point attitude, even with CG in the forward position. The touchdown speed was about 70 mph IAS. With its wide-track undercarriage the aircraft handled well on the ground; there was no swing after landing, or tendency to nose over at any CG position, although the run was noticeably longer than either the Hurricane or Spitfire.
Both K8620 and L6950 were fitted with a Merlin III with a two-pitch de Havilland propeller of 11 ft 6 in diameter. The take-off run was calculated at 315 yards when corrected for zero wind and International Standard Atmosphere (ISA) conditions, with 560 yards being required to clear 50 ft. The best rate of climb was 1620 ft/min at 10,700 ft, which was attained in 7.1 minutes. The greatest height achieved was 26,000 ft and the estimated service ceiling was 28,100 ft. The full test results were as follows :
With a maximum of 6¼ lb/sq.in boost the highest speed was only 303 mph at 16,600 ft. However, slightly better figures were achieved with an increased boost of 12 lb/sq.in and 100 octane fuel, when a top speed of 312 mph TAS was recorded at full throttle height of 10,000 ft. The maximum speeds at other heights were as follows:
Further performance testing was carried out at Boscombe Down between 22 April and 5 June 1940 using L6954, which was fitted with a constant-speed de Havilland propeller. Despite the fact that the take-off run was slightly longer at 340 yards, the distance to clear 50 ft was reduced to 485 yards. As expected, the change of propeller had little effect on the top speed, although there were slight improvements above and below full throttle height. The major benefit was a substantially better rate of climb, 5000 ft now being reached in five minutes, 10,000 ft in eight minutes and 20,000 ft in twelve minutes.
The Defiant I entered service with No. 264 Squadron at Martlesham Heath in December 1939. However, it was beset by engine and hydraulic problems, which led to a temporary grounding in early 1940 while the snags were rectified by Rolls-Royce and Lockheed. The development of suitable tactics for the Defiant was also causing headaches. Squadron Leader P.A. Hunter, No. 264 Squadron’s CO, reported that the Defiant suffered a serious lack of speed, to the extent that overtaking attacks on a single Fairey Battle flying at 180 mph IAS could only be carried out successfully if the Defiant had a height advantage of between 2–3000 ft. He also noted that because of its weight, it was unable to regain height rapidly after a high-speed dive. With full service equipment and fully armed, the Defiant I weighed in at 8350 lb. This reduced the top speed to between 280–90 mph, depending on engine performance and the state of individual airframes, but quite how it was unable to cope with a Battle flying straight and level at such a slow speed was not explained.
Shortly before the Defiant was issued to No. 264 Squadron, tactical trials were carried out by AFDU (then at Northolt) using L6951, L6952 and L6955. The air gunner’s vision was found to be severely restricted, being limited to a narrow cone directly in front of him, together with the upper and rear hemispheres of the turret. There was very little room in the turret and the gunner’s knees were constantly pressed up against the ammunition boxes. Movement of the turret was effected by a short control lever, with an electrically operated button allowing a high-speed option. The sight was a Mark III reflector, located centrally above and between the two banks of guns.
Practice attacks were made on Blenheims of No. 25 Squadron, also based at Northolt. The easiest form of attack was to overtake the ‘bomber’ on a parallel course, but against any aircraft faster than a Blenheim it would have to be delivered from a dive in order to have sufficient speed. The most comfortable position for the air gunner, when firing from the bomber’s flank, was approximately 100 ft lower than the target at a range of 200–250 yards. Attacks delivered from below were also reasonably successful, although when passing from astern to ahead, the guns could not be used within the vertical 36-degree cone as the sight could not be used. As the Defiant’s guns could not be depressed below the horizon, attacks from above necessitated the aircraft being banked so that fire could be delivered. Skidding was not found to be a practical proposition, owing to loss of speed, so attacks had to be delivered from a slight turn.
AFDU considered that the main advantage of the Defiant was its ability to deliver fire onto a bomber from below, where it would not be as protected by armour. To do this, however, the Defiant needed to be able to carry out its approach without being seen, as the Defiant gunner had great difficulty in firing at the target if it began evasive manoeuvres, particularly those involving diving to lower levels. In these circumstances, any burst of fire was likely to be very short and the weight of fire delivered would, in consequence, be limited. It was also recommended that the smallest tactical unit should consist of at least two Defiants and that a section should consist of four aircraft instead of three.
Unfortunately, the Defiant was never to be tested in the role for which it was designed. By the time German bombers began to appear over Britain in substantial numbers, they were able to operate from forward bases and were protected by fighter escort. Comparative trials with a Hurricane had already shown that the Defiant would be no match for single-seat fighters, as it not only lost out in terms of performance, but also in manoeuvrability. The Defiant weighed approximately 1800 lb more than a Hurricane when fully loaded, but its wing area was slightly less at 250 sq.ft. This resulted in a wing loading of 33.27 lb/sq.ft compared with the Hurricane’s 25.63 lb/sq.ft. When compared with the Bf 109E, there was en even bigger disparity in weight, the gross weight of the Emil being only 5740 lb. The Messerschmitt’s much smaller wing of only 174 sq.ft resulted in a similar wing loading to the Defiant, so that the Bf 109’s biggest benefits were a speed advantage of 50–60 mph and much better climb and dive performance.
Meanwhile trials were taking place to assess the Defiant’s suitability for other roles. Specification F.9/35 had also called for light bomb racks to be fitted under the wings for close-support duties and in early 1940 L6950 was used for dive-bombing trials over the Orfordness ranges. The optimum method of attack was to begin the dive from 5000 ft at an angle of 40–45 degrees, with a pull-out commencing at around 1500 ft. Shallow dive attacks were also tried at angles of 20–25 degrees down to 500 ft. Steep dives at 60 degrees were attempted with speeds reaching 400 mph IAS, but it was found that the height needed to recover was excessive and bombing accuracy was poor. L6968 was used for trials in the Army Co-operation role and although a few aircraft were delivered to No. 2 Squadron, they were quickly withdrawn.
Only two squadrons (141 and 264) were to fly the Defiant as a day fighter, with both units suffering heavy losses during the Battle of Britain. It had always been intended that the Defiant should be capable of fulfilling a night-fighter role and following the severe mauling at the hands of the Luftwaffe by day, the Defiant was assigned to night interception. At first crews operated visually, but from September 1941 the first radar-equipped Defiant IAs began to appear with early AI.IV (Airborne Interception) sets with ‘arrow-head’ aerials on the starboard wing and an ‘H’-type aerial on the fuselage side. The cathode-ray tube for the radar was located in the front cockpit on the pilot’s left, with the control panel on the right.
To improve performance, Boulton Paul looked at more powerful engines and eventually selected the 1280-hp two-speed supercharged Merlin XX for the Defiant II. The aircraft also featured an increased internal fuel capacity of 159 gallons (two auxiliary tanks being fitted in each outer wing section), a deeper radiator and a lengthened engine cowling. Owing to the changes at the front end, a slightly larger rudder was needed to restore directional control. A Defiant I (N1550) was converted as the Mark II prototype and was flown for the first time on 20 July 1940. This was followed by N1551, which was involved in a ground collision with N1550 before being delivered to A&AEE for brief performance and handling trials. It was flown at a weight of 7690 lb but was not fitted with AI equipment, which would have added another 285 lb. The figures for rate of climb showed a considerable improvement, with 2780 ft/min being recorded at full throttle height of 10,700 ft. The estimated absolute ceiling was now 33,100 ft. The Defiant II could reach 26,000 ft in less than half the time of the Mark I as the following table shows.
With its increased fuel capacity the Defiant II had a maximum range of just over 450 miles, which could be attained in MS blower at 10,000 ft with 1900 rpm set. The speed was a relatively sedate 195 mph TAS. The maximum speed was increased only slightly, despite the extra power of the Merlin XX, 313 mph being recorded in FS blower at a full throttle height of 19,400 ft. Other results were as follows:
Handling tests at Boscombe Down were also carried out using AA370, which was the first purpose-built Mark II. The seven preceding aircraft had all been converted from Mark I airframes. It was flown at a weight of 8510 lb at normal CG loading. During early testing the rudder was found to be unsatisfactory, as it appeared to be over-balanced and caused directional instability. Various steps were taken to cure this trouble and finally a rudder was fitted in which the trim tab acted solely as a trimmer, and not as a balance tab as previously. In addition, the tab was mass-balanced and stringing cord was doped on each side of the straight portion of the trailing edge, immediately above the tab. This rudder performed much better and was adopted for production. The aircraft was also flown with a slightly larger propeller of 11 ft 9 in diameter.
The cockpit was very similar to that of a Defiant I, except that it appeared to be rather less draughty in flight. The heating was adequate but tended to produce an oily smell that was rather unpleasant on long flights. This was not particularly surprising as the heat for the cockpit was drawn via a duct from the oil cooler. The hood was now provided with sliding panels on each side and above. These enabled a clear view to be obtained when flying in rain or if the windscreen had become oiled. In addition, the whole hood could be jettisoned should an emergency exit have to be made. The only major difference in the control layout was the introduction of a lever for the constant-speed propeller. This was situated between the throttle lever and the side of the fuselage. The supercharger gear change knob was now located low down on the left-hand side of the panel in the position previously taken by the control for the two-pitch propeller.
Even though full right rudder trim was used for take-off, a strong swing to the left was experienced, but this could be held with rudder without too much effort. Like the Defiant I, there was insufficient rudder trim to fly ‘feet off’ in the climb, but the force required to keep the aircraft straight at its optimum climbing speed of 140 mph IAS was only slight. In normal flight the ailerons and elevator were very similar to the Mark I. However, a difference was noted with the rudder, which was heavier than before and directional stability was now neutral. Pilots also had difficulty in achieving trimmed flight. If the aircraft was yawed, it would regain a steady course without oscillation when the rudder was released, but quite often there was a change in direction. At cruising speed, four divisions of right rudder trim were needed to fly straight with ‘feet off’.
Dives were carried out with the aircraft trimmed for straight and level flight, i.e. four divisions of right rudder trim and seven divisions of nose-down elevator trim. The forward force required on the control column in the dive increased rapidly above 320 mph IAS and 360 mph IAS was about the maximum speed that a normal pilot could hold. If full nose-down trim was used, the aircraft settled into the dive at about 320 mph IAS. Once again, the push force needed increased rapidly above this speed and became excessive at 380 mph IAS, although the maximum achieved was 395 mph IAS.
Up to 360 mph IAS, directional control was satisfactory; the aircraft wanted to yaw to the right with increase in speed, but this could easily be held with rudder. Above this speed, however, the pilot experienced a rather disconcerting change in the control loads on both the rudder and elevator. Quite suddenly, the forward load on the stick and the load on the left rudder were relieved, as if the control surfaces or the tabs themselves had moved. This change was not persistent, but occurred as a sudden jerk, after which the original trim was regained. This was not explained, although it was thought that it may have been caused by a sudden distortion of the trimmers. In view of the excessive stick forces, combined with this ‘snatching’ of the rudder and elevator, it was recommended that the limit for diving be reduced from 400 mph IAS to 360 mph IAS.
When trimmed to fly at 1.2 × stall speed and with the flaps and undercarriage down, the Defiant II was just unstable longitudinally, with any disturbances producing slowly increasing oscillations. The stall itself occurred at 80 mph IAS and was marked by a sharp drop of the starboard wing through about 35–40 degrees, although this could be raised by coarse use of the aileron. After about a second in this condition, the aircraft tended to flick violently over to the right with a significant drop in nose attitude, and if back pressure on the control column had not been released, it was thought that a spin might have developed. With the flaps and undercarriage up, the aircraft was longitudinally stable as speed was reduced, but there was an increasing tendency for the port wing to drop. The stall occurred at 92–4 mph IAS. As a large amount of aileron was needed to hold the port wing, this produced considerable aileron drag, with resultant yaw to the left, but in this case there was a tendency to spin.
The Defiant went on to serve with thirteen squadrons in the night-fighter role until mid 1942. By this time aircraft with increased performance were urgently needed and the type was quickly replaced by the twin-engined Bristol Beaufighter and de Havilland Mosquito. Surplus aircraft were used in several other roles, notably Air-Sea Rescue (ASR). The Defiant flew with five ASR squadrons on high-speed search duties until June 1943, before being replaced by converted Spitfire IIs.
In late 1941 Reid and Sigrist at Desford was given a contract to convert 150 Defiant Is as target tugs with the designation TT.III. The turret was removed to be replaced by a small canopy for an observer, but the take-off weight was still 8227 lb and the drag imposed by the target towing equipment reduced the top speed to around 250 mph. A further 140 target tugs were ordered in July 1941 based on the Defiant II, designated TT.I. The Defiant also served as a training aircraft at various Air Gunners’ Schools and the Central Gunnery School.
The last Defiant to remain in use was DR944, which was issued to Martin-Baker in December 1944 for ejector seat trials. The observer’s position was removed so that an ejector seat could be fitted behind the pilot and a number of ejections were made in 1945 at speeds up to 300 mph using dummies. DR944 was eventually struck off charge on 31 May 1948.