The Seafire F.45 was the Fleet Air Arm equivalent of the Spitfire F.21 and was powered by a Griffon 61 driving a five-blade Rotol propeller. As the wings had not been modified in any way from those of the RAF machine, this variant did not feature wing folding. Armament comprised four 20 mm Hispano cannon, the first Seafire to be so equipped.
Admiralty interest in a navalised version of the Spitfire F.21 resulted in the issue of Specification N.7/44, the end product of which was regarded very much as an interim type, pending development of a more advanced variant with laminar flow wing (this was a proposed version of the Spitfire 23 which was subsequently abandoned). The prototype Seafire F.45 was a converted Spitfire 21 (TM379), the main modifications comprising the fitting of a sting-type arrester hook, which increased overall length to 33 ft 4 in, and the incorporation of a guard for the tailwheel. Due to its non-folding wings and the strong swing on take off as a result of the extra power of the Griffon 61, the Seafire 45 tended to be flown from shore based establishments only, although a number were fitted with a Griffon 85 and six-bladed contra-rotating propeller which eliminated the swing on take off due to engine torque.
The handling section in Pilot’s Notes for the Seafire 45 fitted with a Griffon 61 warned that the aircraft was likely to be nose-heavy during taxying, extreme care having to be taken when applying the brakes to avoid nosing over. Concentration was also required on take off as the use of high power levels not only caused a pronounced swing to the right, but the aircraft was also likely to crab in the initial stages of the take off run and tyre wear when operating from runways could be severe. A boost setting of +7 lbs/sq.in was found to be quite adequate for take off, with full power being selected when safely airborne. After take off the wheels could be braked and the undercarriage retracted, however, any failure of the gear to retract was likely to lead to excessive engine temperatures as the airflow to the underwing radiators and oil cooler was likely to be interrupted.
Once in the air and in the climb the recommended speed was 150 kts IAS up to 25,000 ft, reducing speed by 3 knots for every 1,000 ft gained thereafter. Normally during a combat climb the supercharger would change into high gear automatically at around 11,000 ft, however, the maximum rate of climb could be obtained manually by delaying the changeover until boost had dropped to +6¼ lbs/sq.in. Some deterioration in longitudinal stability was noted when climbing at low airspeeds, otherwise stability about all axes was satisfactory. Control forces throughout the speed range were generally light, movements in pitch and yaw being assisted by elevator and rudder trimmers which were very effective in operation. This sensitivity could be a problem during high speed dives as any sudden application of rudder or its associated trimmer was likely to cause the aircraft to skid violently. When diving, the aircraft also tended to become increasingly tail heavy as speed was increased and it was recommended that it should be trimmed into the dive. The elevator trimmer could be used to assist recovery but this had to be handled with care as it was very effective and there was a distinct possibility that excessive accelerations could result.
Stalling speed at a typical service load with engine off and flaps and undercarriage up was 80 kts IAS (85 kts IAS when fully loaded) and 72 kts IAS (76 kts IAS fully loaded) with flaps and undercarriage down. The stall was preceded by airframe buffet which became noticeable 5–10 knots before the stall with aileron snatch in the latter stages. At the stall either wing was liable to go down followed by the nose, but recovery was immediate if the back pressure on the control column was relaxed. Intentional spinning was prohibited but the aircraft usually responded to normal recovery action. All normal aerobatic manoeuvres could be performed except flick rolls and sustained inverted flying. The recommended entry speed for a loop was 320–340 kts IAS and upward rolls could be carried out from an initial speed of 360–400 kts IAS.
When approaching to land the recommended speed with engine on and flaps down was 85 kts IAS increasing to 90 kts IAS with flaps up. With the engine off 10–15 kts IAS had to be added. If the aircraft was at light weight with ammunition expended and minimal fuel, 5 knots could be reduced from the above speeds. For deck landing the recommended final approach speed was 75 kts IAS, a considerably improved view of the deck being obtained if a curved approach was made. During the approach the aircraft was subject to a number of trim changes as its configuration was changed. When the undercarriage was lowered a nose down trim change occurred, the opposite occurring when the flaps were lowered. A nose up trim change also occurred when the radiator shutters were opened. Landing was straightforward, the only possible problem occurring when a missed approach resulted in a go around situation. It was important not to select full power when close to the stall due to a strong tendency to roll and turn to the right. At low airspeeds even full opposite aileron and rudder would be insufficient to counteract this effect and control could well be lost.
Development of the later Spitfires to incorporate a Griffon 87 engine with contra-rotating propellers was potentially of more importance in a naval environment, the close confines of an aircraft carrier leaving little room for error. The absence of torque effects with such an installation was a clear benefit and trials soon commenced at A&AEE to assess the aircraft’s suitability for deck operations. Having already been tested with a Griffon 61 and a five-blade propeller, the prototype Seafire F.45 TM379 was returned to Boscombe Down in June 1945 with a Griffon 87 and a six-blade Rotol contra-prop. Certain aspects of its control surfaces were non-standard and the fabric-covered elevator was not modified to current Spitfire F.21 standards in that the alteration to the horn balance and the trimmer tab gearing had not been made. The fabric-covered rudder had a revised trimmer tab which was split into two parts and splayed 1.6 in at the trailing edge, this also being thickened to about 0.25 in. Above and below the trimmer tab there were fixed split tabs. The ailerons and balance tabs were metal-covered and were fitted with piano-type hinges on the bottom surface.
During the trial TM379 was loaded to a typical service take off weight of 9,385 lbs. Ground handling tests showed that the aircraft could be held against the chocks without the tail lifting up to an engine boost pressure of +1 lb/sq.in. Compared with the single rotating propeller with which the aircraft was fitted on its last visit, there was no tendency to swing or for one wheel to lift during the take off run. There appeared to be improved thrust in the early stages of the run, but otherwise the take off characteristics were normal for the type. To evaluate the stall, the aircraft was trimmed in the glide at 80 mph IAS with flaps and undercarriage down, the speed then being gradually reduced. At speeds below 80 mph IAS a vibration of small magnitude was noticed which had not been apparent when the aircraft had been tested previously. This was not considered serious, however, and the stall occurred at 66 mph IAS, with a normal recovery.
The contra-prop installation on TM379 came into its own on the approach to land and during a baulked landing. Directional changes of trim with speed and power, and the pitching associated with yaw, were very small, which was in marked contrast to the findings of the test when the aircraft was fitted with a normal five-bladed propeller. The rudder appeared to be a little heavier, but this was considered to be an improvement. Some difficulty was experienced in keeping the approach speed down and it was felt that more drag would be advantangeous when in the landing configuration. Baulked landings in particular showed a significant improvement as large power changes merely led to a small directional trim change which was easily controlled by the pilot. As the fitting of contra-rotating propellers had consideraby improved the aircraft’s handling characteristics, especially during take off and landing, A&AEE were able to approve the installation for carrier operation without any restriction. Spinning trials were carried out in September 1945 using TM383, its behaviour being typical of other Spitfire/Seafire aircraft with normal recovery characteristics.
Towards the end of 1945 a more comprehensive handling trial was carried out at Boscombe Down using Seafire F.45 LA446 so that clearance could be given for the service to conduct intensive flying trials. This aircraft was also fitted with a six-blade contra-rotating propeller (Griffon 85) but at the time of test it was limited by the manufacturers to a maximum speed of 403 mph IAS (350 knots) due to concern about directional instability at higher speeds. The results of qualitative testing showed the aircraft to be pleasant and easy to fly and without any serious fault up to the limiting speed.
Although the advantages of the contra-prop installation were mainly apparent when in the air, it was also noted that ground handling was very much better than the Seafire F.45 fitted with the normal five-bladed propeller. As the rudder was more effective, the brakes did not have to be used as much which made taxying much easier. With the throttle open, the rudder centred rapidly when freed. The tendency for the tail to lift on opening the throttle was no different from previous Seafire F.45s. For take off, the rudder and elevator trimmers could be set to neutral and the whole process was extremely simple as there was no swing and the rudder could be held fixed in the central position.
The aircraft was then climbed at 170 mph IAS with 2,600 rpm and +9 lbs/sq.in boost, the elevator trimmer being set to 1½ divisions nose up, with the rudder trimmer an eighth of a turn forward of neutral. When the speed was displaced by +/-10 mph and the control column released, a slow damped oscillation of small amplitude was set up. Rudder centring from small displacements was good, but there was a slight tendency to overbalance at larger displacements when the aircraft was kept laterally level by applying opposite aileron. From full starboard displacement the rudder centred slowly on release, but from full port displacement there was little or no tendency to centralise.
At a fast cruise speed of 270 mph IAS at 25,000 ft with the engine set to 2,400 rpm and +7 lbs/sq.in boost, the trimmer settings were–rudder neutral and elevator ½ division nose up. In this condition static longitudinal stability (stick free) was positive, a speed displacement of +/-10 mph requiring the application of a moderate push or pull force on the control column. On release of the stick after such a displacement the aircraft returned slowly to its trimmed speed, never taking more than two cycles to return. When the rudder was freed after displacement and the wings were held level by ailerons, a fast oscillation ensued which damped out in about six cycles. All the controls were moderately light and responsive inducing powerful roll, but no pitch with yaw. Use of rudder gave accurate turns in both directions with the ailerons held fixed, but turns made with ailerons only resulted in slight slipping in.
From the pilot’s viewpoint the ailerons appeared to float up about a quarter of an inch at the trailing edge in level flight. Rate of roll was extremely high and the ailerons were described as being very light and crisp. On previous tests with the Seafire F.45 pilots had criticised excessively high static friction in the aileron circuits. Although this was considerably reduced on LA446 it was still higher than the recommended value and the ailerons did not centre when displaced and freed. At 35,000 ft, with the engine operating at 2,400 rpm and +2 lbs/sq.in boost, the longitudinal stability was satisfactory and directional stability appeared to be substantially the same as at 25,000 ft.
The aircraft was also dived up to the limiting speed of 403 mph IAS with trim and engine controls set as for fast cruise. There was no change of directional trim of any note, but a push force of around 20 lb was required on the control column. The controls were fairly light and positive, the rudder centring rapidly after being freed following a displacement. At no point did any unpleasant characteristics become apparent and on release of the stick the maximum reading measured on the accelerometer was 3g. A slight swing to starboard was noticed when the throttle was closed at the maximum speed attained, but this was easily counteracted by a very small rudder movement.
At the opposite end of the speed range, with power off and the elevator trimmer wheel fully back (flaps and undercarriage up), the aircraft was put into a glide with the control column free at 106 mph IAS. When displaced +/-10 mph from the trimmed speed, the aircraft returned to this speed with a slow damped oscillation on release of the control column. Directionally the contra-prop F.45 appeared to be no different from normal Griffon-Seafires but when the control column was displaced laterally and released, a lateral oscillation commenced, together with oscillation of the ailerons, which showed no sign of damping out. In circumstances when the control column was held, rather than released, these oscillations were not present. Neither were they present during power off glides at 92 mph IAS with the flaps and undercarriage down. Pilots also reported that longitudinal stability was slightly better in this configuration.
Stalls were carried out which showed that at a take off weight of 9,510 lbs and with a typical service CG loading, LA446 stalled at 88 mph IAS with flaps and undercarriage up and 78 mph IAS with flaps and undercarriage down. The characteristics at the stall were virtually identical to a standard Seafire F.45. To simulate a baulked landing, the aircraft was put into a glide with power off, flaps and undercarriage down, radiator flaps closed, elevator trimmer fully back and the rudder trimmer neutral. The engine was then opened up to full power and a climb was made at 140 mph IAS. No directional change of trim was encountered and the longitudinal change was counteracted by a push force on the control column estimated at 25 lb.
Landings with power off were straightforward. When landing from a powered approach pilots found that this was much easier to perform as any amount of power could be added to adjust the rate of descent without incurring strong changes in directional trim. In view of this it was considered that the contra-prop Seafire F.45 had vastly improved deck landing characteristics, although it still suffered from a poor forward view and a tendency to float. In certain conditions of visibility dark segments were visible in the propeller disc at low engine speeds but this did not worsen the forward view to any extent.
Although it did much to develop the engine/propeller combination of the ultimate Seafire, the F.47, the only F.45’s to serve with the Fleet Air Arm were fitted with a Griffon 61 and five-bladed propeller. Only fifty were produced at the Vickers-Armstrong factory at Castle Bromwich, some of these aircraft seeing service with Nos. 771 and 778 Squadrons at Ford from October 1946. The penultimate Seafire was the F.46 which was of low-back configuration (unlike the F.45) and was essentially similar to the Spitfire F.22. Early aircraft tended to have the original empennage, whereas later machines had the Spiteful-type tail. Most examples of this variant were powered by a Griffon 61 with a five-blade propeller, although a few were fitted with a Griffon 87 and contra-props. Production was once again centred on Castle Bromwich but only twenty-four were to be produced, a few being designated FR.46 following the installation of an F.24 oblique camera. The Seafire F.46 entered service in early 1947 with No.781 Squadron at Lee-on-Solent and the variant also flew with No.1832 Squadron at Culham until retired in 1952.
The last of the Spitfire/Seafire line was the Seafire F.47, the first of which (PS944) was taken into the air for its maiden flight on 25 April 1946. As much of the development work had already been carried out on the contra-prop versions of the Seafire F.45 and F.46, there was no prototype as such and PS944 was followed by another thirteen aircraft in the same serial batch (PS945–PS957). Subsequent batches were VP427–VP465, VP471–VP495 and VR961–VR971. The aircraft in the PS batch were powered by a Griffon 87 but those machines that followed were fitted with a Griffon 88 which had a Rolls-Royce developed fuel injection and transfer pump instead of the Bendix-Stromberg induction-injection carburettor used on previous versions of the Griffon engine. Although this had the effect of adding 70 lbs to all-up weight, a significant advantage was gained in that power was maintained under all conditions of ‘g’. Wing folding was incorporated in the Seafire F.47, consisting of a single break point outboard of the cannon installation.
In December 1946 a deck landing assessment was carried out at Boscombe Down using PS944, the main aim of which was to determine the rate of descent and touchdown speed. Apart from the provision of wing folding, the Seafire F.47 differed from the F.46 seen previously in that it had a redesigned air intake in an elongated duct under the nose, long stroke, anti-rebound oleo struts and flaps of increased chord. All the control surfaces and trimmers were metal-covered. Each aileron had a balance tab and a trimmer tab was fitted to the port and starboard sides of the elevator. The rudder was fitted with a combined trimmer and anti-balance tab. The latter was approximately 18 in long and was split about 7 in from the lower end, the upper part being displaced 15.25 degrees from the lower part of the tab. A 6½ lb inertia weight was fitted in the elevator control circuit which exerted a moment of +66 lb/ins about the elevator hinge with the elevator neutral and fuselage datum horizontal.
In a series of simulated deck landings on a runway, made with and without the aid of a batsman, it was considered that the best approach speed was 81 mph IAS (70 knots). The engine conditions for an approach at this speed were -4½ lbs/sq.in with the propeller set to fully fine, these resulting in a steady rate of descent of around 600 ft/min. On closing the throttle just prior to touchdown (normally on the batsman’s signal) the aircraft sank in a three-point attitude. During testing of the Seafire F.46 a certain amount of nose drop had been apparent when the throttle was closed but this was not experienced with the F.47. In the case of a misjudged, engine off approach, if power had to be fed in to avoid an undershoot developing, a slight nose-down trim change occurred requiring a light pull force. Directional and lateral changes of trim were negligible.
In the case of a baulked landing the engine could be opened up to 2,600 rpm and +9 lbs/sq.in boost with little difficulty following an approach at the above speed. On retracting the undercarriage there was a very slight nose-up change of trim requiring a push force of 1–2 lbs to hold, however, the trim change when raising the flaps at 115 mph IAS was in the opposite sense and needed a pull force of 7–10 lbs to maintain the aircraft’s correct climbing attitude. This nose-down change of trim was most marked when the flaps were moving from the ‘take off’ to the ‘up’ position, the trim change when they were being raised from ‘landing’ to ‘take off’ being small and easily held. It was discovered, however, that a climb out with the flaps in the ‘take off’ position held no problems so that if another circuit had to be made after being given the ‘wave off’, the flaps need not be fully retracted.
Other approach speeds were attempted for a comparative assessment. At a speed of 86 mph IAS (75 knots) the aircraft tended to touch down slightly main wheels first, but with the long stroke oleo undercarriage there was little or no bounce. At 75 mph IAS (65 knots) the approach was rather uncomfortable due to aileron twitching and buffeting as the aircraft was getting close to the stall. At this speed a relatively large amount of power had to be maintained to prevent an excessively large rate of descent.
Rates of descent and (true) touchdown speeds were measured during a series of simulated deck landings controlled by a batsman, the engine being cut on his signal when the aircraft was very near to the ground. From an approach speed of 81 mph IAS the rates of descent on touchdown varied somewhat up to a maximum of 7.2 ft/sec. The average true touchdown speed was 93½ mph thus giving a correction to indicated values of 12½ mph. An investigation was also made on the effect of engine power on the rate of descent by measuring height against time at various boost pressures (flaps and undercarriage down). The results obtained varied from a rate of descent of 8½ ft/sec at -4 lbs/sq.in boost to 29 ft/sec with the engine off.
The recommended approach speed of 81 mph IAS for Aerodrome Dummy Deck Landings (ADDLs) on the Seafire F.47 compared favourably with that of 86–92 mph IAS for the Seafire F.46 and was most likely due to the increase in flap area on the F.47. The latter was also easier to land in a three-point attitude due to the fact that the nose did not drop when the throttle was closed. One aspect of the Seafire F.47’s landing performance that became apparent after the Boscombe Down trial was that although 81 mph IAS was a satisfactory approach speed during ADDLs, for actual deck landings on an aircraft carrier it was best to increase approach speed to 92 mph IAS (80 knots). The reason for this was subject to some conjecture, but Service Trial Unit pilots were adamant that the higher figure produced a more controllable touchdown during actual deck landings.