In October 1917, Sir William Weir, a Scottish industrialist who was then Director of Aircraft Production at the Ministry of Munitions, began to place orders for a new engine which, although unproven, promised a quantum leap in power output and thus in aircraft performance. The engine had been designed in response to a request for aero-engines producing at least 300 hp, weighing no more than 22 lbs per hp and with an overall diameter of under 42 inches so that they would fit in existing airframes.

An original Dragonfly cylinder designed with the unusual valve springs used by Bradshaw to keep the overall diameter down. One expert later described the cylinder head layout as “Probably the worst example of air-cooling ever used on a production engine.”

Four designs were submitted, three of which had 14 cylinders arranged in two rows and one, from ABC motors, for a 9-cylinder engine, the brainchild of the company’s gifted designer, Granville Bradshaw, which promised to deliver 340 hp and yet weigh less than 600 lbs. These figures, together with the fact that a single-row design was obviously easier to build than one with two rows, were enough to persuade Weir to adopt Bradshaw’s design and a prototype was ordered. Since ABC Motors lacked the facilities to build it, the job was entrusted to Guy Motors, who hand-build it in just 28 days, a feat for which Weir promptly sent a congratulatory telegram to the company.

The Nieuport Nighthawk intended to replace the S.E.5a as one of the RAF’s front-line fighters. Granville Bradshaw, the designer of its Dragonfly engine, is the small man with glasses standing at the right of the picture

The prototype appears to have run sufficiently satisfactorily to convince Weir that the design was viable. In January 1918, he decided to phase out production of those engine designs he considered obsolescent and to concentrate production on just four types, the V12 Rolls-Royce, The six-cylinder inline Siddeley ‘Puma’, the 230 hp Bentley BR2 rotary, and ABC Dragonfly, which between them were intended to power all front line aircraft in 1919.  In addition to the batch of 1000 Dragonflies already on order with Vickers, orders were placed with 16 other contractors for almost 11,000 engines, of which at least 4135 were to be delivered by 1 June 1919.

With the promise of significantly increased performance that this new engine offered, many new aircraft were designed around it. Henry Folland, designer of the S.E5a, who was now chief designer at the Nieuport and General Aircraft Company, planned to use it to power the Nighthawk, a fighter which he hoped would replace his own S.E.5a in front-line service. The Siddeley Siskin was another Dragonfly-powered fighter design, and the Sopwith Company produced a whole range of designs around the new radial, including the Dragon, a Dragonfly powered version of the Snipe, which was already on order. The Armstrong-Whitworth Ara, the Austin Greyhound, The Avro Manchester, the Bristol Badger, and numerous others were intended, if selected for production, to be powered by the Dragonfly too.

The first flight of the Sopwith Dragon, a Dragonfly-powered version of the successful Sopwith Snipe, which the Dragon would have superseded had its engine proved reliable.

However, as the first production examples became available for bench testing, it became obvious that all was far from well; the engines overheated so badly that carburation was adversely affected, causing irregular running. At first, Weir accepted the problems as no more than teething troubles and remained confident that they could be resolved, but things got worse rather than better. Power output was well down on Bradshaw’s promise, producing only 276 hp at the design speed of 1660 rev/min. although 295 hp was possible by running it at 1750 rev/min. But there was worse to come, for at that speed, it overheated so badly that, in the right light, the upper parts of the cylinders could be seen glowing a dull red.

The R.A.E. at Farnborough designed a new cylinder head which eliminated the cooling deficiency, albeit at the expense of an increase in the overall diameter of the engine to well beyond the specified 42 inches and also increased the weight, which in production form, now totalled 662 lbs.  Now able to be run consistently, the engine’s final problem was discovered; It had been designed to run at the critical torsional-vibration frequency, a fate which other engine designers had avoided merely by chance as this form of vibration was little understood at the time. Eventually, new pistons, together with crankshaft balance weights, reduced the effect of the vibration somewhat and thus modified, the Dragonfly could usually produce 300 hp at 1800 rev/min, but failures remained frequent, and it became clear that no amount of development could ever make it reliable.

Only 23 Dragonflies had been delivered by the time of the Armistice, with a further 15 examples joining them by the end of 1918 and yet a total of 1147 were delivered before production was finally cancelled. In December 1918, sample engines were delivered to a number of companies planning to use the Dragonfly in their designs. A few examples of the Nieuport Nighthawk did enter service, but its active life was understandably short-lived, distinguished only by some success in post-war air races where the sacrifice of reliability in favour of performance was more readily accepted. Prototypes of other Dragonfly-powered designs, including the Sopwith Dragon, Westland Weasel and Bristol Badger, were delivered for testing, but only a few Dragons were ordered, and none of these saw service. Instead, the peace-time RAF was obliged to continue using the same types as it had flown during the war for some years to come. As the official history of the air-war states, “Had the war continued into 1919, the failure of the Dragonfly would have put an effective brake on the expansion of the air service.”

Fortunately, the War ended before the effects of this disaster could affect the outcome.