The High Speed Designer Confirmed – R.J. Mitchell’s S.5

Whilst the Schneider Trophy continued to be organised by the clubs, by 1926 the character and costs of the meetings had produced the first confrontation of government subsidised teams with well organised military pilots and support staff. Government assistance for a British entry had already been in evidence in 1925 but thereafter renewed action prompted Flight to comment that ‘Never in the history of British aviation have we tackled an International speed race in so thorough a manner’. Gloster and Shorts were also asked to design machines capable of speeds not less than 265 mph at 1,000 feet and Napier was responding by increasing the Lion engine's compression ratio to 10:1, with a view to approaching the 900 hp mark.

      Faith in this reliable engine was justified when it was now made to deliver 900 hp in the ungeared model and 875 hp in a geared version. Just as metal propellers were found to be superior to wooden ones as tip speeds increased, so it was considered that any extra weight or loss of power because of reduction gearing would be compensated for by greater propeller efficiency. The Air Ministry clearly hedged their bets by supporting three Supermarine entries with two geared and one ungeared engine, a geared Gloster machine and two ungeared ones, as well as a Short seaplane powered by a more standard air-cooled radial engine.

     Also, on the 1st of October, a High Speed Flight was now formed, consisting of military personnel, to test and compete with the new aircraft, rather than relying on company pilots as previously.

 S.5

At Supermarine, Mitchell continued to place his faith in the newer monoplane approach and proceeded to strive for improvements on his S.4 design. Mitchell was now able to profit from the more determined Governmental support and work was soon under way, involving careful appraisal by wind tunnel and tank testing of ¼ scale models. Particular attention was to be paid to floats, flush wing radiators and airscrews. He (and Folland for Glosters) had consulted with Napiers and the new Lion was designed whereby the engine’s frontal area, reduced by shortening the connecting rods of their engine and lowering the cylinder blocks, had repositioned magnetos and the cam covers of the three cylinder engine banks were contoured to mate with the streamlining engine fairings fore and aft. 
     Thus Mitchell was able to reduce the cross section of his fuselage – so drastically in fact that the pilot's cockpit was an extremely tight fit: there was no room for a seat and the pilots sat on the floor of the machine, their legs almost horizontal and their shoulders coming up to and pressing against the underside of the cockpit coaming. Fl. Lt H. M. Schofield, one of the pilots of the newly-formed RAF High Speed Flight described their visit to Supermarine ‘for a fitting’: ‘The method of reaching the seat was to squeeze in sideways and down as far as possible so that the shoulders were below the top fairing, then turn to face the front, and in my case it needed no ordinary effort to get my shoulders home.’

     As there was now insufficient room for the fuel tank in the fuselage, the starboard float was used, which would help towards counteracting the torque of the engine which, during take off, might be expected to cause the opposing float to dig in and swing the aircraft off line before it gained sufficient airspeed to be effectively governed by the control surfaces. Mitchell also offset the fuel-loaded starboard float an extra eight inches from the centre-line as an additional response to this expected problem.

     However, the most telling improvement, apart from the effect of the more powerful engine promised by Napiers, was the estimated increase of about 24 mph by the change from the Lamblin type under-wing radiators of the S.4 to a system akin to that adopted by the previous Curtiss racers and, subsequently, in the Italian Macchi M.39. The new radiators were to be made out of copper sheets, eight and a half inches wide, with their outer surfaces formed and fixed to the contours of the upper and lower wing surfaces; thus the outer sheeting, exposed to the cooling airflow, offered virtually no additional drag. Corrugations on the inner surfaces of the radiators formed channels for the coolant: this was taken along troughs behind the rear wing spar, through the radiators and along the leading edge of the wing and then pumped to a header-tank behind the engine block. The lubricating oil was also cooled in surface radiators, consisting of channels which ran along the outsides of the fuselage and up to a header tank behind the cockpit (not long after flying tests began, it was found necessary to increase the oil capacity).

      Supermarine S.5 on engine test (R.J. Mitchell on left).

In terms of structure, the Supermarine contender continued the move away from the basically all-wood structure of its S.4 predecessor: the new machine now had an all-metal fuselage, of stressed-skin structure (which looked forward to that of the Spitfire), whilst the flying surfaces were still of wooden construction and ply-covered. It was designated S.5 as it represented a complete redesign of the previous monoplane and also incorporated the new information gained from meticulous work at the National Physical Laboratory test facilities, sponsored by the Air Ministry.

     Mitchell had sent down three models for wind tunnel testing: one was a shoulder-wing design with wing roots cranked down and supported by streamlined struts from the floats; a second model had a low wing, similarly braced by struts; and the third configuration was an all-wire-braced proposal with a low wing position to give favourable bracing-wire angles. Biard's problems with forward vision during landing and take-off in the S.4 were no doubt an influence on Mitchell’s considerations and eventually the flat, low wing position was chosen. The new wire bracing between the floats and from them to the bottom of the wing also allowed a wire ‘cage’ to be completed as the wires from the upper fuselage to the top of the wings were fixed immediately above the float bracing attachment points. 
      Mitchell was clearly guarding against any wing flexing which might have contributed to the S.4 crash, as well as providing a further reduction of the weight and drag represented by the very sturdy float struts of the S.4. The balancing out of advantages and disadvantages attendant upon the wish to reduce frontal area and weight against the need to ensure adequate strength and pilot view was set out after the race in the Chief Designer's speech to the Royal Aeronautical Society in 1927:

(a) The primary object in lowering the wing on the fuselage was to improve the view of the pilot, which was never very good on the S.4 The higher position of the wing no doubt gave a lower resistance due to fairing in the outside engine blocks and thus saving a certain mount of frontal area. A loss in speed of about 3 miles per hour is estimated from this alteration. This loss is more than balanced, however, by the importance of the improved view.

(b) The system of wire bracing of the wings to the fuselage and floats was adopted for a number of reasons. The unbraced wings and chassis of the S.4 were very high in structure weight, and it was found very difficult to construct an unbraced wing sufficiently strong and rigid without making it very thick at the root, and thus increasing its resistance. The adoption of bracing was largely responsible for a reduction in structure weight of 45 per cent for the S.4 to 36 per cent for the S.5, with its corresponding reduction in resistance; also for the elimination of the two struts between the floats, and for the reduction in frontal area of the four main chassis struts. Against these must be set the addition of fourteen wires. It is not easy to estimate the final effect of a number of alterations of this nature, but from the analysis of the resistance of the two machines it is given on fairly good grounds that the overall effect was an appreciable saving in resistance, amounting to an increase in speed of approximately five miles an hour.

(c) The cross-sectional area of the fuselage has been reduced by about 35 per cent. This very large reduction was obtained through the redesign of the engine and the very closely fitting fuselage. This almost amounted to a duralumin skin in order to ensure that the very smallest amount of cross-sectional area was added. On several occasions during the construction of the fuselage the pilots were fitted, and much trouble was experienced through their being of varying dimensions … The reduction in body resistance was responsible for an increase in speed of approximately 11 mph

The floats were also reduced in frontal area by about 14 per cent. This was accomplished by using a much lower reserve buoyancy. The reserve buoyancy was 55 per cent for the ‘S.4’ floats and 40 per cent. for the starboard float of the ‘S.5’ [now being used for fuel tankage]. This figure is extremely low and called for very efficient lines.

The estimated increase in speed due to reduction in float resistance is 4 mph. These reductions in resistance of fuselage and floats are due to lower cross-sectional areas and not to improvements in form.

(d) Wing surface radiators were first fitted to the American machines in the 1925 race, and gave these machines a very big advantage in speed. The radiators added a certain amount of resistance to the machine due to their external corrugations increasing the area of exposed surface. As about 70 per cent. of the resistance of a high-speed wing is skin friction, and the corrugations almost double the area of surface, it is reasonable to suppose that an increase of at least 30 per cent. of resistance is added to the wing. It is evident that a saving in resistance would result if radiators could be made with a flat outer surface, and that they would give no direct resistance to the machine. After much experimental work, radiators with a flat outer surface were produced. The chief difficulty experienced was in sufficiently strengthening and supporting the outer skin to enable it to stand the heavy air loads without making the radiators unduly heavy. The estimated increase of speed due to their use in place of Lamblin radiators used on the ‘S.4’ is 24 mph.

Thus, without considering the improvement in performance with a more powerful engine installed, Mitchell had estimated an increase of 41 mph; he could have also mentioned that, with the second S.5 (which was to come first in the forthcoming contest) ‘the hundreds of tiny rivets all over the skin were now flush with the surface instead of projecting like a mass of wee knobs as they had done’.

At the Venice venue of the Schneider competition that year, the Crusader crashed –  the aileron controls having been crossed on re-rigging in Venice. And so the three permitted British entries were finalised as Fl. Lt S. N. Webster in the Supermarine N220, Fl. Lt O. E. Worsley in the Supermarine N219 and Fl. Lt S. M. Kinkead in the better of the two Glosters.  N220, and the Gloster, N223, with the unproven geared engines were to fly flat-out with the expectation that Worsley in the ungeared S.5 was likely to finish if, for any reason, the other two machines failed. 
    After a delay of one day, owing to bad weather, the Italian, Ferrarin, turned off the course on the first lap with two pistons burnt through, to be  followed by de Bernardi on lap two, suffering from a connecting rod failure. The third Macchi, with an older replacement engine, proved no match for the British with their new uprated Lions but then Kinkead retired at the beginning of the sixth lap when violent vibrations made it seem prudent to do so. However, on the penultimate lap and in sight of being placed, the last Italian, Guazzetti, pulled out  when he was blinded by the bursting of a petrol pipe’. Luckily, he managed to get down safely although not before just missing spectators on the roof of the Excelsior Hotel on the Lido.

    Webster led the British S.5 whitewash with an average speed of 281.65 mph, a new record for seaplanes and, indeed, bettering by 3 mph the world speed record for landplanes. Worsley came second at 273.01 mph in the second, ungeared, S.5. Back in England, Mitchell was among those fêted by the Corporation of Southampton and his winning machine was put on display in London.

In the following year, Vickers acquired the Supermarine Company but acknowledged the achievements of Mitchell’s design team by retaining it as an entity at the Woolston factory and by allowing the branch a separate identity under the following title:

THE SUPERMARINE AVIATION WORKS, LTD

(DIVISION OF VICKERS (AVIATION) LTD.)

Mitchell’s name now appeared among the list of Vickers Directors and the 1925 publicity description of him as ‘one of the leading flying-boat and amphibian designers in the country’ was now significantly expanded to:

one of the leading flying-boat, amphibian

and high-speed seaplane designers in the country.

He was still only thirty-four years of age.

* * * * *

For reference sources, see my Blog: “Source Material and References. " An extended bibliography is included in my R.J.Mitchell at Supermarine; Schneider Trophy to Spitfire  which also provides material for wider reading, grouped according to specific areas of interest.