Centrifugal or Axial?

The compressor debate

The feasibility of gas-turbines to drive aircraft occurred to several people in the late ‘20s. Amongst those in England were Dr A.A.Griffiths, a scientist with a growing reputation in the aviation and academic world after his paper “Aerodynamic Theory of Turbine Design” was published in 1926. Another was Frank Whittle, a Royal Air Force pilot. When in 1929 Whittle, age 22, first brought his ideas to the attention of his superiors, they were thought of sufficient interest to obtain an expert opinion. They turned to Dr Griffiths, whose own proposal was based on an axial-flow engine driving a propeller.

His opinion of Whittle’s much simpler turbo-jet scheme was damning. That he found a mistake in Whittle’s calculations could not have helped, although in subsequently revising them Whittle found another error which largely cancelled out the first. Griffiths seems to have also underestimated the favorable effect of low intake temperatures at high altitude and of increased compressor efficiency due to ram-effect at high speed – his turbo-jet proposals would have operated lower and slower than Whittle was envisaging.

The subsequent letter to Whittle from the Ministry also stated that gas-turbines were impractical since there were no suitable materials which could withstand the high temperatures and stresses involved, a view which was still substantially true ( as a matter of chickens and eggs) though based on an earlier (1920) report by the Air Ministry scientist, Dr W J Stern. Whittle nevertheless filed the first patent for a turbo-jet on 16th January 1930, only a few months after his interview with Griffiths, showing a 2 stage axial compressor feeding a single-sided centrifugal compressor. By the time he embarked on building an engine he had decided to simplify this to a single, double-sided centrifugal compressor.

Which was best? An axial compressor could achieve a higher efficiency under ideal conditions. It would have a smaller diameter, resulting in lower frontal area, lower blade tip speeds and the opportunity for an annular combustion chamber. Against this, it would be mechanically less robust and aerodynamically much more sensitive to both intake conditions and downstream pressure changes, such as those induced by rapid throttle movements. Both these could lead to blade stalling and resulting surges, which could in turn produce catastrophic blade failures. Pressure increase per row of blades was low, requiring at least 8 rows with intermediate stator blades. A centrifugal compressor would be simpler and cheaper to produce, would be more rugged and much more tolerant of the varying intake conditions and throttle changes encountered by a flight engine. It would need to run at higher speeds but there was the extensive experience of small centrifugal compressors developed as aero-engine superchargers, whilst there was little prior art on axials.

The year that began with Whittle filing his patent also saw the Aeronautical Research Engine sub-committee recommending that all government-funded gas-turbine projects be discontinued. When work did restart, at the Royal Aeronautical Establishment at Farnborough in 1937, it would be as a direct result of Whittle’s independently-funded success. In the meantime the Air Ministry had been sufficiently impressed by Whittle to send him to Cambridge in 1934, where he obtained a 1st class honours degree and was able to develop his engine design. Although then released from his RAF duties to develop his ideas with private finance, there would be no official financial support and he would have to rely on what the newly-formed Power Jets Ltd could raise privately.

Meanwhile Dr von Ohain ( , after studying physics and applied mechanics at Gottingen University, had in 1935 patented his proposal, which like Whittle's original patnet of 1930, featured both axial and centrifugal compressors. In later years von Ohain denied that he knew anything of Whittle's patent but his colleague Dipl-Ing Wilmhelm Gundermann, who worked with him at Heinkel from the begining of April 1936 recorede 'We kept fully up-to-date with such patents as Whittles. Both now set about building engines, von Ohain with the help of Max Hahn and financial support from Ernst Heinkel ( ), whose company he joined in 1936, and Whittle by contracting parts to various companies, particularly the steam-turbine maker British Thompson-Houston. Both successfully ran their engines in 1937, Whittle in April and von Ohain some 3 months later. Heinkel now designed and built an aircraft to test von Ohain’s engine, the HE 178 which first flew on 27th August 1939, but which was abandoned after only 3 flights.

Ohain's HeS 3B, with a single stage axial compressor feeding a centrifugal, was the first to fly

Government work on gas-turbines now began again under Griffiths at the Royal Aircraft Establishment at Farnborough, with sufficient funds to both build and contract-out axial compressors, including to Metropolitan-Vickers. However Griffiths then moved on to Rolls-Royce where he continued to work on a complicated reverse-flow turbo-jet. Metrovick, after two unsuccessful attempts to build an axial turbo-jet themselves, began development of the Hayne Constant designed F2, running it in December 1941, the first non-German axial gas-turbine. After 3 re-designs the F2/4 was built in a small series in 1945 by which time it had achieved 4,000lbs thrust. However problems persisted and the company decided to use the experience to develop new designs.

In Germany, the government had placed contracts in 1937 with 3 engine manufacturers, later including Heinkel who had acquired a large part of a team from Junkers Aircraft, who had not received a contract. The four design teams had each decided to pursue axial compressor solutions – Heinkel ( now having both an axial and von Ohain’s centrifugal project. These decisions may have been partly influenced by the German Air Ministry (RLM) having carried-out research into axials at the time the British Air Ministry had stopped gas-turbine research. This left only Whittle and von Ohain developing centrifugal engines, although quite unknown to each other.

When, in mid 1939, Whittle successfully demonstrated to the government the re-build of his original engine, he and his small team at last gained government support. He had no production facility himself, so in April 1940 the Ministry allocated production of his newly designed W2 engine to Rover, an independently-minded car manufacturer, at a stage when there was still much to be learnt about gas-turbine design and manufacture. Unsurprisingly, this significantly extended the time taken to get it into production (see “Whittle, Lutterworth and the birth of the Jet Engine” for details). Meanwhile the progress of the Rolls-Royce Griffiths-designed axial-flow turbo-jet was slowing to a halt (it had a combined compressor/turbine whereby the inner section of each stage formed the compressor and the outer the turbine) and Rolls-Royce eventually took-over Rover’s contract, the engine passing a 100 hour test at 1,600lbs thrust in May 1943 and powering a Gloster Meteor ( )shortly after. The Metrovick F2 ( began flight-tests soon after but by now it had been decided to put the major effort into producing a development of Whittle’s engine (the Derwent) and another centrifugal engine based on Whittle’s layout, the de Havilland H1 Goblin designed by Frank Halford ( , which had also reached the flight stage at this time (It powered the first Meteor to fly) and was intended to power an aircraft from the same company, the Vampire.

In Germany, von Ohain continued to develop his engine until it was dropped in 1942 in favour of an engine (S 011) with a diagonal centrifugal compressor between axial stages designed by Helmut Schelp. This was about to go into production when the war ended. The most promising of the axials, the Junkers Jumo 004 ( ).and BMW 003, ran into serious compressor and vibration problems and, although the Junkers Jumo 004 had reached production by the end of 1943 and was being produced at the impressive rate of 1000/month by mid-‘44, it continued to prove troublesome in service (in the Me 262 ) and had a service life of 10-25 hours. This was partly due to the inferior materials which had to be used but compressor stall remained a major problem.

Ohain's HeS109-011, with a diagonal centrifugal compressor followed by a 3 stage axial compressor

General Electric, supplied with the very first Whittle W1 engine when it became available in October 1941, had bench-run a GE-produced version (J31) in 8 months and had developed a series of designs to reach 4,000lbs thrust with the production J33 in 1945. An axial design, begun at the same time, the J35, reached production at the same thrust by September the following year. Rolls-Royce decided that the successor to the Derwent should be the RB41 Nene, another centrifugal. First run in October 1944, this was designed and built in 7 months (some say 5), produced 5,000lb thrust and powered a number of 2nd generation jets, with licence production in Britain, the US, Russia (RD-45), Spain, Canada and Australia. Sir Stanley Hooker, Rolls-Royce Chief Designer, commented that their axial Avon ( ) in contrast, took as many years to reach success as the Nene had taken months. DeHavilland also followed the Goblin with another centrifugal, the 5,000lb thrust Ghost.

Therefore at the middle of 1945 it could be said that there were 2 centrifugal (British) designs and 2 axial (German) designs of 2,000 lbs plus with a production and flight history. The centrifugal ones were simpler, more rugged and had a longer-life – the Welland had entered service in May 1944 with a TBO of 180 hours. Arguably they had also been developed in a shorter time with smaller design resources. Against this, the Allied blockade and bombing had seriously hindered German progress. Pilots who evaluated both British and German jets at this time had little doubt that the British centrifugal engines were more reliable and easier to handle than the German axial ones.

There were also 3 second-generation engines, one American and 2 British, of 4-5,000lbs thrust (J33, Nene and Ghost) about to enter service at this time. All of these had centrifugal compressors and could trace their lineage to Whittle’s W1.

However GE's axial J35 was only months behind and, with the demand for ever more thrust, the long gestation period of the axials now began to pay off. The Avon and the Metrovik F2- derived Sapphire which had by now been taken over by Armstrong Siddeley) both entered service in the late 40s with thrusts in the region of 7,000lbs. Both were highly successful and were produced under licence in the USA. At high thrusts the axial became universal. But what would be the choice now, with all that is known of both axial and centrifugal compressors, if one was designing an engine to produce around 2,000lbs thrust, as those early designers were? An example is the Rolls-Royce Williams FJ44-2, an engine of 2,300lbs thrust currently produced for business jets. A fan engine (an idea also first patented by Whittle, see below) it has a 3 stage axial (including the fan) driving a centrifugal compressor. Without the fan, this corresponds to Whittle’s original turbo-jet patent of 1930. The axial/centrifugal arrangement (or a centrifugal on its own) is still the most common configuration in small engines, the blades of an axial becoming too small for the later stages.


In late 1944/45 Frank Whittle and his team at now-nationalised Power Jets began the design of a fan engine. This had an 8 stage axial compressor feeding a single centrifugal compressor. There is no doubt that the axial compressor incorporated the knowledge gained over many years by the RAE, Farnborough, one of whose leading scientists was on loan to the company. The project was regrettably cancelled by the government when 50% complete and it would be a further 9 years before the first production turbo-fan, the Rolls-Royce Conway, would emerge.

Bibliography and further reading.

Genesis of the Jet, John Golley, Airlife Publishing.

World Encyclopedia of Aero Engines, Bill Gunston, pub. Patrick Stevens.