Early British gas turbine development

[tartle]

In 1945 it was becoming apparent to the A-S team that there was an opportunity to design a 1,000hp turboprop for one of the Brabazon Committee aircraft requirements so the engine that became the Mamba was proposed. A government contract followed and in towards the autumn of 1945 detail design began. As the understanding of combustion had increased after work on the Python and sharing the results of the wider work undertaken by Lucas, Rolls-Royce, etc. The first engine ran in April 1946 and delivered 800 hp. It was configured as a single-shaft engine with a gearbox drive on the front of the shaft. There was a 10-stage axial compressor driven by a two-stage turbine and six straight through combustion chambers. The Mamba prototype had spray burners supplied by a high-pressure fuel system and the single spool ran on two bearings. Issues with the combustion chambers and with differential thermal expansion of blades and casings that made close clearances impossible to maintain, possibly exaggerated by vibration of the main shaft meant a redesign was called for.
The layout was redesigned to incorporate a 3rd bearing and after the success of the Python vaporising combustion system this was incorporated too.
The Mamba 1 emerged from this redesign and successfully delivered the design rating of 1,010 shp for a weight of 760 lb. As the engine found possible airframes- Apollo, Athena, Balliol- it was realised that additional horsepower and development potential was needed and so the team decided to increase the AMF from 13.5 lb/sec to 17 lb/sec enabling an increase in power to 1,270 shp, which took until 1950 to achieve as a rated power. This power increase was delivered by removing two stages from the rear of the Mamba 1's compressor and adding two more at the front resulting in the Mamba 2, which as the ASM.3 went into limited small scale production.

....tbc

 

[tartle]

The modifications to the Mamba compressor meant that the A-S engineers suddenly found that the front stages of the compressor, now much larger than before suffered from a condition known as flutter.
Flutter is an aeromechanical phenomenon that usually occurs at a blade natural frequency and involves sustained blade vibration resulting from the changing pressure field around an aerofoil as the blade oscillates.
Flutter is dependent on both the aerodynamic and structural characteristics of a compressor or turbine blade. A straight stator or twisted compressor blade can vibrate in such a way as to increase the local angle of attack, generating lift which feeds into the vibration, increasing the angle of attack- feeding on itself until high stresses are reached. The amplitude will then reverse as, the blade stalls and so the cycle begins again.... high frequency fatigue soon results and cracks begin to propagate...
This form of failure was found on testing the increased mass flow version of the Mamba. The original design of the compressor had aluminium alloy blading throughout and the 'A-frame' configuration of disc was in steel. The first change was to replace the first three rows of alumiium blades with much stiffer steel ones. These generated higher centifugal forces due to higher density, so the discs were redesigned in aluminium and changed to more conventional looking configuration. This enabled the new blades to be accommodated with no increase in weight. At the same time the last two stages were also changed to steel blading as the air temperatures were getting close to the limit for aluminium to work well. The increased mass flow also meant that the turbine was called on to do more work... the blade length was increased to allow this to happen.
The compressor changes can be seen in the cutaways of 1948 and 1951 below.
Another change that happened during development of what became the ASM.2 was improvements in the vaporising combustion chamber. It had been noticed that hotspots were appearing on the outer walls opposite the 4 tertiary holes, also the small tongue across the hole was breaking off as it overheated. The turbine end cutaways show some combustion chamber mods but this was not fixed until greater changes were made in the early '50s- even then the new welded walking sticks needed modification to finally get the reliability required. Incidentally it was the flinging out of unvaporised fuel round the curved walking stick that promoted relative overheating of the inner bend and so set up differential themals that led to failures.
....tbc

 

[tartle]

Mamba and variations on a theme... a jet, a turbofan and a gas generator....all wip.
The British and Australian governments made an agreement in 1945 that meant the UK would develop guided missile and Australia would develop the test facilities. This resulted in the creation of the Woomera Test Range and the development of the target drone by General Aircraft Factory ... the first proof of concept vehicle being the manned Pika and the drone proper was known as the Jindivick. The Pika and initial versions of the Jindivick were powered by a short life engine known as the A-S Adder and later versions that followed the first batch were Viper-engined; the last production batch was in 1997 still with the Viper.
The Adder was used on the Mk1 Jindivick. The prototype Adder first ran in Nov 1948, and was flight tested in a Lancaster flying test bed before powering the Pika into the air in 1950 and the Jindivick Mk1 in Aug 1952. The Adder incorporated some changes to its layout such as air cooling of the centre bearing eliminating the need for scavenge lines used on the Mamba. The second cutaway (first 'Aeroplane'; second 'Flight') shows the bearing layout in more detail. As the simplification worked it was soon incorporated into the Mamba engines.

The cost of neutrality was, for Switzerland, technological isolation. The end of WW2 left the federation rich but woefully ill-equipped to defend its political status in the new world order that was rapidly taking shape. It embarked on two aircraft programmes, under the leadership of Jürg Branger Technical Director of F+W Emmen, Swiss Federal Aircraft Factory, designed to move their expertise forward, with the aid of several world class scientists that were leading edge.
One, the Aigullion N-20 had four engines derived from the Mamba. These were buried in the wing, Comet-style and were configured as plenum chamber burning turbofans with air bleed for flap blowing, etc. Not much has been published but we know the engines were developed by Sulzer Brothers and several versions were planned. The SM-01 engine instruction to proceed was issued in 1947 as the engine would be the critical item, in terms of development timescale.
In an article translated in the late 60s, 'Aviation History in the Swiss Transport Museum it refers to the N-20:
"Potentially the most interesting types at Lucerne are the N-20 Aiguillon STOL fighter and the Arbalète research aircraft, both evolved after the war by the Federal Aircraft Factory at Emmen, near Lucerne.
Of modified delta-wing planform, the N-20 embodied many features that were entirelly new at the time. Its powerplant consisted of four Swiss-designed turbofan engines buried in the wing. In each of these engines, the cold air from the fan was ducted through additional combustion chambers located on each side of the axial 'core', providing a reheat device that doubled the normal thrust and was intended for use during take-off and combat. For short take-off and landing, the secondary airflow could be diverted through large slots on the upper and lower wing surfaces. When only the lower slots were open the deflected air acted as an aerodynamic flap; with upper and lower surface slots open the deflected air acted as a thrust reverser.. As a substantial proportion of the airflow passed through the wing aerodynamic drag of the relatively thick section was kept low.
....
Ground tests of the Aiguillon prototype began at Emmen in 1952 but, to the great disappointment of the designers, no flight tests were authorised due to lack of funds."

So if we read the above whilst looking at the Aiguillon 3-view drawing, the wing sections (both similar to post #22 here) and the photo we can begin to visualise the propulsion set up. The aircraft photo also has the engine under its wing, with the afterburners next to the engine.. there are other photos of the aircraft and engine here.


...and the gas generator? Well originally Fairey were going to use a RR Dart with auxiliary compressor on what became the Rotodyne... Hives pleaded too much work so Fairey decided to go with a Mamba plus auxiliary compressor, known as the Cobra. Like Rolls, A-S pleaded that they were too stretched (Python, Mamba- single and Double and F9 takeover) and so Fairey had to go elsewhere.
....tbcike Rolls

 

[tartle]

Although the aircraft powered by the Mamba were not the main drivers of development of the engine the Double Mamba's steeds certainly were. The beauty of the configuration is that a large amount of development can be done on a single engine and only the twinned systems and components have to be tested in concert. The Mamba therefore grew in power up to the ASM 8 which delivered 1,950 shp.
There is a picture of a Gannet just after catapulting off Victorious here.
The development issues affecting the Double Mamba centred around its gearbox and also surge problems made apparent by the control system and the complexity of mechanical solutions to the Navy's requirement that either engine could be shutdown without loss of 'services'. Treating the engine as one powerplant enabled A-S to provide one accesory gearbox but it had to be possible to drive it from either engine; this meant a system of freewheels had to be introduced. Each engine drove one of the contra-rotating propellers, which means one engine has an idler in the train whilst the other (port) engine has different diameter wheels to absorb the radius of the idler but giving the same output speed but in the reverse direction.
Developing the second (ASMD 3) version was driven by the exclusive needs of the Gannet. The control system became one that enabled the engine to operate at constant 15,000 rpm and the stagger of the compressor blading was altered to move the operating and surge lines apart during accelaeration to operational speed. The Gannet had been grounded due to the surge problem that kicked in at 14,500 rpm and as the original control system allowed rpm to drop during demands for more power at the prop and consequent pitch increases happened... there was a huge problem. This led to a grounding of the aircraft until the solution could be found and incorporated. (The Python was having a torrid time too with Westland losing Wyvern pilots at a high rate). Gearboxes were also suffering from bearing problems brought on by metal pickup leading to worn cages and races. Changing materials and improving manufacture helped eliminate this but it all took time.
A major simplification of the gearbox design improved reliability and also dropped the vertical offset of the propshaft axis to engine axis by 5 inches to 6 inches.
Flight published two good technical summaries of Double Mamba progress 14th March 1955 p.272 onward and 22 Nov 1957 p. 815 onward; it is unfortunate that 0n p 816-7 the section of the new gearbox has been cutoff on the electronic version so the full impact of the new gearbox design is missing. If anyone can scan their edition of this it would be useful.. if not I'll chase a few institutions for help. The ASM 7 and ASMD 7 projects are interesting in that they are total redesigns of the concept to bring in a free power turbine to reduce the inertial energy in the rotating mass attached to the propellers. This concept was adopted for A-S's next project in the mid-50's.
I've scanned a couple of my Double Mamba photos that give a good idea of the size of the powerplant for the Gannet.
.......tbc

 

[tartle]

In 1956 there was agreat deal of interest in helicopters that need upto 1000 shp engines. This resulted in De Havilland taking a licence for the GE T-58 which it named Gnome, Blackburn continuing to work with Turbomeca to uprate the Turmo engine that eventually became the B-S Nimbus and Napier looking at a scaled down Gazelle- the Gazelle Junior (not built). Armstrong Siddeley began to look at concepts for their answer to the challenge and this developed over the last part of '56 and into 57. Flight did a good write-up of the progress in their 21 Feb 1958 edition.
The sketches contained in the article I've captured below as they illustrate the designers' thinking very clearly:
sketch A is essentially a scaled down Dart configuration
sketch B is similar to A but like the contemporary thinking on the Mamba reflects a free turbine layout where the turbine is flipped to be near to the propeller without a long shaft through the rest of the rotating components, keeping the first stage impeller diameter as low as possible.
sketch C shows the result of looking at Turbomeca/Blackburn and incorporating a transonic axial compressor as the first stage of compression; this allows the designers to go for a more conventional shaft arrangement; the challenge of rapidly/cheaply developing the engine meant that the actual demonstrator built was based on...
sketch D where a conventional compressor is substituted. Only one engine was built and run before the project was abandoned.
I have a copy of the SBAC brochure mentioned in Flight which outlines the current state of A-S thinking... i e sketch B. which was discussed with the airframe people at the Show.
The inner pages are attached

...tbc

 

[tartle]

Meanwhile over at Bristol and Napier the wartime efforts to get into the gas turbine business got underway. Napiers were the last to join the race but as soon as Bristol realised what was going on when they were invited to attend the second GTCC meeting in Nov 1941 they got going on a strategy to enter the business. Before Fedden left (pcl for thrown out) the business he and Frank Owner looked at what sort of engine they could develop that did not compete head on with those already in the business. The first go was to put W2/700 engines in the rear of the Buckingham nacelles to give more power than the Centaurus at mission critical events.... this soon died a death!
Owner suggested a 4,000 hp turboprop with similar fuel consumption at 20,000 ft and 300 mph. and at a weight not exceeding the equivalent piston engine. By May 1942 the team - F G Evans and P Fortescue- came back with an answer: An axial plus centrifugal compressor, heat exchanger and free propeller turbine layout. Exploring alternatives took them into a three shaft layout... LP and HP spools plus a free turbine driven propeller. This was soon dropped as they did not think three shafts for a first effort was sane!; RAE suggested driving the prop off the LP spool but they were nervous about the control problem (this would be like or a variation on the 'Clyde' theme... did RAE influence RR?).
The two approaches investigated were a high pressure ratio layout... a prerequisite being to increase component efficiencies by a considerable amount or to live with existing efficiencies but recover power through the use of a heat exchanger. The second route seemed more likely to succeed as only one 'module' was a challenge... they thought. but to keep the size of components and modules within existing testing facilities they decided to halve the target power output.
An intense study of likely heat exchanger configurations took place juggling the thermal ratio, pressure drops, bulk, weight and manufacturing challenges until a scheme was proposed that consisted of a circular outside profile tubular configuration. Daily visits to the Drawing Office would find the 'design of the day' as various detailed attempts to produce a 'lightweight' heat exchanger configuration that could deliver maximum number of tubes in the available space but with passage ways that were uniform between each tube-way. Eventually a designer at Tockington Manor , Henstridge, ingenious configuration that used flattened oval tubes with a wall thickness of .012 in. enabling a design wighing 500 lb and transferring the heat equivalent of 2,000 hp to be contemplated. Coventry Radiator and Presswork had the dubious pleasure of translating the design into hardware.
Overall size of the exchanger was 45 in overall diameter tapering to 36 in dia at the rear and 31 in. long.
The heat exchager was designed to receive 1,000 cu. ft. gas per second from the turbine exhaust, at 500 deg C., and delivered 180 cu.ft per sec. of air to the combustion chambers at 300 deg. C. An exchanger’s effectiveness is the ratio of the actual heat transferred to the heat that could be transferred by an exchanger of infinite size. Theseus effectiveness was a thermal ratio of 0.4. which was calculated to save 150 lb of fuel per hour at cruise conditions.
The turbine exhaust passed through 1,700 (yes one thousand seven hundred) straight tubes of stainless steel, 0.012 in. thick and 0.625 in. at the tube endplates. By arranging the tubes along involute curves and flattening them, air passages of uniform width were formed between adjacent involute curves. Air was taken from the eight inlet chambers, through groups of involute passages to the central space. This space formed an inlet header from which air returned to the peripheral space through other groups of involute passages alternating with the first groups previously mentioned.
The thinness of the tubes and their close pitching made any form of mechanical or expanded (flared or swaged) joint impracticable, and were, therefore, brazed in place. The joints suffered during hot starts that often happened on the test bed. An alternative welded construction withstood the starts, but cracked through excessive contraction stresses on shutdown.
As the first gas turbine to be built at Bristol the engine suffered from similar problems to other first-time developers. The compressor had been designed to deliver a PR of 5:1, any more would have been technically challenging and anyway the temperature differential across the heat exchanger would have been too low to deliver the performance required. It turned out that the predicted efficiency of the combined 8-stage axial, single-stage centrifugal compressor was about right but the mass flow was 7 % high. In May 1945 attempts were made to start testing the engine, without heat exchanger as CRP were still making this module; the engine refused to accelerate to self-sustaining speed without exceeding a safe turbine entry temperature. It was guessed that the nozzle guide vane area was too low and the turbine was choking. As a quick fix the NGVs were cutback to increase the flow area, but this still did not work so blow off valves were devised, and the engine finally ran on 18th July, 1945... this week, 67 years ago. The engine testing then went ahead in an uneventful manner. The high mass flow meant there was plenty of power but the 'modifications to the NGVs meant turbine efficiency suffered so sfc was below exoectation. The first year of testing clocked up 117 hours including a 25 hour development test. Component research on newly completed experimental rigs ensured there was a steady improvement on sfc but the biggest snag was the high Air Mass Flow.
The first complete heat exchanger (HE) was delivered in July 1945. An engine fitted with the HE ran in December and it was immediately obvious things were not right. The pressure drops at the higher flow rate were excessive, so that the fuel saving at full throttle was only 8% rather than the designed for 20%, while power loss was correspondingly greater. The part load economy was improved by upto 25% but this was not a condition that at the time people thought was a good idea for a propeller turbine.
Rig testing of the compressor made it very clear that great strides were going to be made in a relatively short time and so the slower pace of HE development meant it was better to put resources into the former rather than live with the bulk and weight of the HE. The practical difficulties of attaching the tubes was a bugbear but the expected tube fouling did not take place.
The mechanical integrity of the Theseus was excellent. The Heat Exchanger was abandoned and efforts continued to improve the power/weight ratio and achieve a reliability that would enable meaningful flight testing. In Dec 1946 the engine passed the Ministry 100 hour Type Test at 1950 ehp; the world's first successful TT for a turboprop. In February the Theseus made its first Flight in a Lincoln.
By July 1948 the engine was robust enough to become the first prop turbine to pass the MoS 500 hr endurance test. In 1949 the engine carried out its last TT at a maximum power of 2450ehp.
In order to be able to have an idea of the power being generated by the free turbine drive to the propeller a torque dynamometer was built into the reduction gear. the oil cushion that resulted from the design meant there was a damping mechanism in the drive train which probably accounted for there being no high frequency induced problems in the drive train.
Interestingly the split drum type construction adopted for the axial compressor resulted in three identical sections that constrained the numbers of blades on each row, i.e. identical as the blade roots sat in milled slots in the drum.
Incidentally, there were two Theseus Lincolns and they accumulated 1,000 hours of flying time. A Theseus was also installed in a Hermes V civil airliner, making its first flight in Aug 1948. Flight has 2 pages of pictures starting here.

 

[tartle]

So on to the Proteus.... started before the Theseus was built! and so was not necessarily as great a leap forward as it should have been.
The Parliamentary Select Committee system has operated for a great many years and the Public Accounts Committee Report for 1952-53
makes interesting reading. In the section that covers the MoS it uses the Proteus engine as an example of what goes on at the Ministry.
In 1944 the MoS started to develop a special type of gas turbine for long-range transport aircraft; by 1952 they had placed contracts to an estimated value of £13m of which they had spent nearly £10m for this work.
The Proteus ! was abandoned after £3.6m had been spent. The development of the Proteus 2, and finally, the emergence of the Proteus 3 as an engine suitable for the Britannia and likely to earn some return on the cost of public funds.
In reply the MoS, having noted that development costs had been somewhat inflated by the changing views on the ultimate use of the Brabazon and Princess aircraft, proceeded to point out that the uncertainties involved in forecasting the development costs of an entirely new form of propulsion were high; the design of aircraft are dependant on the availability of the engine and the development of the final configuration can involve the engine developer in further modification. before the needed capabilities are finally delivered.
So let us look at these three Proteus engines:
Work carried out under the leadership of Frank Owner indicated that the lower limits of speed and altitude below which Bristol felt the prop-turbine would not compete was 300 mph and 20,000ft. Balancing the engine's propulsive efficiency, weight and cost puts the upper limit at 500 mph and 40,000ft. The increased cruising speed puts a premium on weight and bulk so it was realised that the HE engine they had designed and were building had limitations in both these respects. As soon as design of the Theseus was sufficiently advanced Owner pulled Charles Marchant out of the team and gave him a new task. His assignment was to design a high pressure ratio propeller turbine of around 3,500 bhp with a diameter not exceeding 1 metre (Fran
k thought a yard was a bit too restrictive). Marchant started making layouts in September 1944 and was fortunate that the Theseus work load on the Technical Office was reducing and so attention could be switched to the new scheme. By December there was a fairly complete design scheme backed up by aerodynamic and performance calculations at a similar stage of completion. Air Commodore 'Rod' Banks arrived on Dec 8th 1944 and as head of DERD at MAP asked to see the latest Bristol project. The Bristol team were convinced the free-turbine layout was right in principle and, so, the turbines and reduction gear were based on the Theseus but the compressor had to be designed afresh... They decided to go for a 9:1 pressure ratio which was extremely adventurous considering their experience. As Bristol were still anxious about potential surge problems and so went for an axial section PR of below 5:1 and put on a two-stage centrifugal at the high pressure end to deliver the required overall PR.
By this time it was becoming clear that the Brabazon and Princess powerplants would be buried in the wing and so the location of engine inlets became a challenge. Because Bristol did not want an engine of excessive length they had opted to copy the Python layout and put the intake at the rear- something that the team would come to regret, although it did mean on the Princess they could have a long diffusing pitot entry to a plenum chamber from which the compressor took its air.
Until Bristol built their own compressor testing facility they were dependant on a shared facility. The requirement for separate compressor testing had surfaced early in the development of gas turbines. Rolls-Royce had built one, initially for Power Jets, using a Vulture engine, but this had to operate with a throttled inlet as the W2 impellers absorbed 4,500 hp -well above the Vulture's capabilities. Metrovick used a staem turbine as a temporary measure. Fortunately the RAE knew of a source of power sufficient for the engines then being planned. The Northampton Electric Supply Company had a power station of about 6,000hp and agreed to aloow use as a drive for the various compressord then being developed. The H.1, F.2 and ASX were successfully tested on the plant and Bristol followed with the Theseus and Proteus designs.
Whilst getting ready to test the Proteus's 12 axial plus 2 centrifugal stage compressor, someone had the idea of using the core as a turbojet and this was quickly schemed out and constructed. It was ready to run just as test results were coming out of Northampton in May 1946. The compressor was calibrated upto 7,000 rpm and was delivering 10% above design mass flow but at a lower than desired pressure ratio.
The Phoebus acted as an early warning system on potential troubles on the Proteus. Having designed the Proteus using the same rules as the Theseus, which had not run by the time Proteus design was finished, it suffered from NGV choking making the whole spool extremely sluggish on acceleration, even after trimming the guide vanes and putting in blow-off valves. Also it was realised the first centrifugal compressor stage was causing a pressure drop and so as a short term palliative it was removed and replaced with a diffusing duct.... there never was time to investigate the loss of pressure and so the stage was never replaced. Another shortcoming that experience in running Phoebus enabled them to tackle was was the excesively high axial velocity through the power turbine. This led to a redesign of the turbine stage with wider and longer blades and a thicker turbine disc to support the extra centrifugal rim loading.
All these changes resulted in a much modified engine which became the Proteus 2.
The original Proteus had not met its weight, fuel consumption or power targets, delivering well below 2,000 hp and the Proteus 2 was an attempt to correct this as well as providing a slightly higher rating!
Incidentally the 'early warning Phoebus' was intended to deliver 2,540 lbt sea level static thrust with an sfc of 0.79 and a weight of 1525 lb. In fact the compressor problems described above meant it did not achieve its thrust target, neither did the first and only batch of engines meet the weight target as they weighed in at 1900 lb!
When Hooker arrived at Bristol in Dec 1949 he began to familiarise himself with the Proteus... the first thing he discovered, early in 1950, was far from meeting its target weight and power of 3,050 lb and 3,200 shp plus 800 lb thrust, the Proteus 2 weighed 3,800 lb and gave 2,500 shp plus a host of reliability problems as compressors, turbines and bearings failed.
Hooker decided there was no way the engine was going to close the power and weight gap and also achieve the reliability needed for the Princess, Brabazon and soon the Britannia aircraft... all civil. Now the costs of the first 2 versions of Proteus mentioned in the Public Accounts Committee Report make sense!
Also note the twin Proteus installation that doubled the challenge!

 

[JFC Fuller]

Tartle,

Excellent thread, thanks for all the detail.

Napier Gazelle Junior: http://www.flightglobal.com/pdfarchive/view/1958/1958%20-%200160.html?search=napier%20gazelle%20junior

 

[tartle]

JFC.. thanks for Gaz. jnr link..... the text highlights what a crowded market it was for an engine of that size... and for your comments. It has amazed me that outside of an industry like aviation generations of innovators/developers seem to be doomed to make exactly the same mistakes as previous generations. The fact we can find detailed records and comments from the people involved makes it so much more interesting, and useful!

 

[tartle]

So Hooker, on arrival at Bristol, was plunged headlong into a crisis. He soon realised he had a crisis of organisation as well as a hardware problem. He tried to be a good co-operator with no job title with Owner but the team below him was problematic ... the Mansell brothers were both shy and retiring; Swinchatt did not believe in turbines and felt the future was Centaurus, his view being supported by Norman Rowbotham, the engineM/D. There was no collaboration between Design and Development with Design often acting unilaterally.
By mid-1950 the Proteus was at crisis level... Hooker and Owner were summoned to Rowbotham's office; also present was an old friend of Hooker's -Reginald Verdon Smith director and grandson of the founder. Owner was moved sideways and Hooker put in charge.
Hooker soon kicked piston engines out to production area and came up with a stronger more focussed management team.
Time was against them and it was agreed the first Princess (Brabazon had been cancelled) would fly with underpowered Proteus 2; meanwhile Bristol would go ahead with the Proteus 3 to deliver the right power at the right weight. So the third redesign began which we will cover in another post.
In order to have an engine available for the first Princess it was agreed to install ten of the Proteus 2s which would be cleared for flight at 2,500hp. The Twin-Proteus had been tested for 1600 hrs in a specially built hangar and the gearbox seemed to be one more part that gave heartache.
The first and only Princess to fly achieved a total of 100 hrs flying time.
In Saunders-Roe's opinion Development problems were of many kinds but the most challenging are included in the list attached.

Hooker's view was that both the big aircraft were doomed and the Britannia was the best bet. As they considered the redesign they thought of straightening out the flow but the Princess was still on the books so they stuck with the existing flow path.

 

[LowObservable]

The Princess - an amazing machine.

Meanwhile, some really early British gas turbine technology, from The Hospital of St Cross (medieval almshouse) in Winchester. This is looking up at a fan located in the kitchen flue, which is spun by the rising hot air and turns the roasting spit through the cross-shaft (visible), a gearbox and chains (now missing).

The hardware looks like late 18c or later to me. Supposedly it is called a "chimney jack" and Intertubez sources say that it was invented by Leonardo da Vinci (but then, what wasn't?).

 

[tartle]

LowObservable... that is a good picture!
I recollect a book on 'Mechanical Devizes' of around the mid 1820s, I think, publicised a device that the Islamic innovators had come up with in about 100 AD... ideas recycle...hopefully the technology gets better on each cycle. Maybe that is why the lunchtime food at Bristol was so good...wrong, or right, sort of turbine!

 

[robunos]

Excellent thread, thanks for all the detail.

Just to echo the above...you're writing a book here, you know....
A question. Was the Pheobus ever intended as a production engine, or purely as an aid to Proteus development?
A thought, whoever named the Proteus was prescient indeed, for like the engine, the Proteus of mythology was shape-shifter...


cheers,
Robin.

 

[tartle]

Robunos...
The bright idea was to develop it as a turbojet, but the thrust requirements were greater than the thrust delivered by the Phoebus, so it continued as a research vehicle for Proteus, which was slow to be made and tested.

 

[robunos]

Thanks for the clarification...


cheers,
Robin.

 

[tartle]

The Twin Proteus 2 looks like the first attachment. Its intake on Princess is also shown.
Stanley Hooker gave the instruction to proceed with the third attempt to get the Proteus right in mid-1950. He had already called Clarke at Lucas and handed them the job of designing and supplying the combustion chambers, freeing up resources to concentrate on getting the rest of the engine right. Lucas's team (who we have already met) of Clarke, Watson and Morley were delighted to take on the work, especially as Hives had decided to bring combustion design in-house at Derby with Arthur Lefebvre recruited to run research (which he did before going to Cranfield as Head of the Propulsion Dept, researching combustion which became World-Class; where he had me as a student!). Hooker had recruited three 'musketeers' from Derby- Basil Blackwell and Gordon Lewis and from Armstrong Siddeley- Pierrre Young. They were unleashed on the Proteus with Lewis in charge of compressor design and performance, Young i/c overall engine performance and Blackwell i/c turbine testing and analysis. All were in the same office so that they could work together on Hooker's challenge to deliver 3,200 hp (Hooker said to aim for 4,000 hp. Charles Marchant's job was to design an engine of 3,000 lb weight. This engine, the Proteus 3, first ran in May 1952 and coincided with Britannia's first flight on Proteus 2s. On its first test the engine delivered its design spec figures. At 1,000lb lighter than the 2 it was soon delivering 3,475 shp and 1,000 lbt.
By improving the turbine efficiency the Proteus 705 that went into the first 15 Britannias was replaced by the 755 at 3,780 shp and 1,180 lbt. The colour cutaway is a low res scan of my 705 picture and is representative of the Britannia engines.

 

[tartle]

The saga that really did for the Britannia as a huge seller and therefore its Proteus engine and the follow on Orion engine (what the Proteus 3 might have been without the Brab/Princess committment) was icing in the intake that could cause temporary flame out. By the time this had been discovered, on BOAC Britannia route proving the airline in its arrogant way (what happens if you think you are part of the government estabishment) had decided that Boeing 707s might be what it really wanted; the UK government would not sanction cancellation of the former and ordering the second. BOAC may have decided to make an ice-mountain out of an excresence and took Bristol to the cleaners. The new Airworthiness regs being introduced at this time required the demonstration of the effectiveness of prevention of ice build up on the wings, tail etc and extended this to the intakes as well. Consequently the thought of flying around the globe in search of the 'right' extreme weather conditions meant that Bristol utilised flying test bed techniques Napier and Rolls-Royce were using to enable the required icing conditions to be simulated by water sprays into the intake. The rates of impact and the size of ice particles could be controlled to represent any extreme weather condition. Hooker's leadership meant that a solution to the flameout was found fairly swiftly. Basically as a chamber went out due to the ice extinguishing the flame there was a period before relight that was noticeable in the cabin. By putting a glowplug in the chmber a relight happened almost immediately with a drop of only 60-300 rpm in 11,000 before the engine was running again. This was not detectable by anyone who was not listeneing out for it and was a good solution. BOAC decided that only a 'perfect' solution would do and that was to eliminate the flame out altogether... and until that was achieved no introduction to service. As Jacques Fontaine put it:

"Issues are emotional
Solutions are technical, but
Decisions are political"

Bristol assumed that their solution, a hollow platinum rhodium alloy tube some 1.5 inches long and 1.375 in. diameter poking into the combustion zone on every other flame tube, was sufficient for BOAC to carry on turned out to be a miscalculation. There are those who were around at the time who thought a reasonable palliative whilst the optimal solution was found would normally be good enough but the desire of BOAC to move on to yet another aircraft (707), incidentally making 5 different aircraft in their fleet meant that they were willing to delay and delay and so BOAC took the opportunity. delaying the introduction for another two years, did not help Bristol's case for the Brittania. As Alertkin points out in the next post the lack of urgency at board level meant that the whole programme was wallowing along and it may be that BOAC had just about lost all confidence in Bristol delivering an airframe and engine that wasn't going to surprise yet again.
The search for a technical solution to eliminate the icing problem tokk a long time as the data on weather conditions had to be improved as well as just looking at the engines. The problem was one of dry ice crystals... the sort that makes lousy snowballs unless you take your gloves off and warm up the ice slightly so it begines to stick together. The use of test rigs, ground-based and flight, determined that the Proteus installed in the Britannia could swallow 2 gm/cu metre of severely supercooled water for 20 secs before a flame out occured; this is equivalent to fling into normal cu-nim (cumulonimbus) cloud for 1.5 miles. (Thick industrial fog has about 0.1 gm/cu m. and heavy snow about 10 gm/cu m). Unfortunately the typical cu-nim around and on the way to Nairobi was typically 6 gm/ cu m. for distance way beyond 1.5 miles, and surprisingly not always visible to the naked eye. The cold Britannia airfame and skin did not cause a problem but the warm stagnation points within the engine intake did., with slush forming at them. In particular it was eventually found that slush was accumulating on the Combustion chambers and when eventually it broke away it was of suffucient size to partial block a zone of the inlet guide vanes; Eventually detection instruments sensed pieces as large as 5in by 2.5 in when flying through these tropical clouds. Two chambers in particular were the sites for these large pieces and elimination of the warmspots was achieved by fitting mufflers. This gave a satisfactory solution and the problem was eliminated enabling BOAC to continue the service introduction.

 

[alertken]

That's not how it was. Cathay Pacific, QF, Ansett, TAA, TEAL did not take Electra over Britannia just because BOAC was, ah, lukewarm; nor did QF take 707-138B just because Boeing had the wit to invent a bespoke model for a batch of 7. Though that did show awareness that customers buy utility, not product. Bristol, indeed UK Aero, indeed UK all-and-any-industry did not understand that.

(3-volume history of QF) J.Gunn,High Corridors,QUP,88,P69: Britannia: “much too late to be competitive. (negative impression in QF’s 7/55 visit) to assess Bristol(’s ability to meet schedule or to demonstrate an) organisation adequate to service the aircraft”. QF at the same time rejected Comet 4: “unable to consider (it) an economic proposition.” (Very long sectors.) You don't disagree, do you?

Not until 2000 could Airbus Industrie overcome that legacy of Brit-uncare for customer support, and pursuade QF to take their products: stringent Guarantees were needed to purge their British past: AOG response times, material burn per flight hour. Bristol...que?

Britannia's failure at market was not the fault of Proteus' sensitivity to ice. It was abysmal drift from date of opening the sluice of our money (Tender Design Conference, 14/7/47, which admitted design (we now say) software effort into MoS-recoverable overhead; BOAC order for 25 (! incredible),(Centaurus-Power) 29/7/49. Srs.102 inaugural, 1/2/57, 312, 19/12/57. Pathetic, and due to the “abysmal lethargy of the (owner-family) Board of Bristol Aeroplane” Sir Peter Masefield,(Bristol A/c’s last MD, whose contact with the firm began in 1943), Flight Path,Airlife,2002,P.209.

 

[tartle]

Alertkin.. you assume too much... I agree entirely with your valuable comments that you make in response to my half finished post.. the board at Bristol was amazingly awful. How Hooker managed to get a flyable Proteus 2 and a Proteus 3 out of the doors with the board he had was amazing if not heroic. Of course some of the credit must be given to the team he built up at Bristol. Some were Bristol-grown but pre-Hooker lacked responsibility and some migrated to Hooker from his old team at Derby. One of the Derby people was a lubrication (what we now call tribology) expert by the name of Robert Plumb, who had worked on the Merlin and then the early Derby turbine engines. He soon joined Hooker at Bristol and was appointed chief development engineer - Olympus and Proteus. When Bob Plumb realised they were redesigning the Proteus yet again he asked whether the lessons of Derby's gearboxes should be taken into account. He meant, of course, the redesign of the Trent, Clyde and Dart gearboxes to adopt helical gears in preference to straight spur ones, so eliminating the vibratory force input as each tooth engaged- a source of excitation on all three engines(helical gears engage gradually along the teeth thus reducing the shock loading of the straight spur) Hooker pointed out that the Proteus gearbox seemed to have performed well throughout the Proteus's life, unlike the rest of the engine and so they stuck with spurs. As would have happened at Derby, Plumb carried on and developed a helical gear set anyway....thank goodness!
In January 1954 the second prototype, with Proteus 705 engines, G-ALRX commenced her series of flight trials for a certificate of airworthiness from Filton. Disaster struck on the morning of 4th February 1954, when Bill Pegg, the company chief test pilot, was demonstrating the Britannia's capabilities to KLM officials. The aircraft suffered an oil fed engine fire, the result of a failure in the input pinion of the reduction gear of number three engine which caused the compressor turbine to overspeed and disintegrate. The fire raged for 19 minutes as Pegg headed back towards Filton. A few miles from his destination he decided that the wingspars might burn through and elected to belly land on the mudflats. As he touched down the mud covered the aircraft and extinguished the flames. Unfortunately the aeroplane was damaged beyond repair as the salvage crews spent 2 days trying to pull her out of the tidal estuary.
Once it was determined the gear was the culprit, Hooker ordered the switch to double helical which was done without delay due to Bob Plumb's anticipation.
The photo below is from the RRHT archive.

 

[tartle]

After the mudflat incident on 4 th Feb 1954, Bristol were under pressure to get the Proteus powered Britannia in a serviceable state so that BOAC, and hopefully other customers could get started as well. But as we have seen the icing problems were very long-winded in solving.
The first conventional icing trials were carried out in Canada on the Ambassador and showed the normal anti-icing features provided could cope with the ingress of water droplets that froze on contact with the inlet guide vanes.
The second icing challenge was the tropical cloud dry ice injestion which we have already discussed. This was a phenomenon that was little known and according to your point of view BOAC made a meal of for their own ends, or BOAC were nervous about a new type of aircraft having a perceived shortcoming that might make it more difficult to achieve success in service.
If that was not enough, it was found in 1957 that water plus ice crystals could be a difficult combination leading to a build up of ice on the outer surfaces of the intake bend. This sort of weather occured in the monsoons over India. Fortunately by this time the development team had a better understanding of the intake and were able to rapidly devise a set of fixes in the hope that one or two would prove to be effective.
The most effective fix was found to be the provision of a series of air jets mounted in the outer wall of the intake duct and fed with pressurised air tapped from the compressor of the engine. The air was injected along the wall of the intake just ahead of the position on the bend at which the ice tended to accumulate, accelerating the boundary layer to a sufficient extent to prevent ice accumulations and to remove any deposits which might have built up before the air jets were turned on. To be really sure they had fixed it the team also schemed out some 8 ducts between the combustion chamber outlets so that the areas of warmth were protected like the muffs which were no longer needed. The resulting fix was christened the 'Rabbit Warren'.
All this attention to detail in the air ducts ensured icing was no longer a service problem.
The development of the Proteus 3 took a great deal of test hours as can be seen from the graph.
It is unfortunate that 4 years were taken off the service life of Britannia which had in been developed with aggression, could have been produced in numbers greater than the 77 production versions... even double would have meant that Bristol made something from their effort.
I have clarified the last graph.... and added part of my Aeroplane Britannia cutaway for intake detail.

 

[tartle]

In 1952, as Hooker wrestled with the Proteus 2 and all its faults, the designers looked at alternative design schemes that would simplify the engine layout and make it more straightforward to develop. Unfortunately the commitment to the Princess programme meant that the next design scheme had to be suitable for that aircraft; hence Proteus 3. However the ideas that came out of the design office were not wasted as it was decided that rig work should start so that an alternative engine could be built for the Britannia if the opportunity arose. The compressor rig work was also useful when an opportunity for an engine for NATO arose. The LP compressor designed for the B.E. 25 became the compressor for the single-spool B.E. 26 - later Orpheus.
Hooker delivered a paper to the SAE in New York on the 18th April, 1955. He stated that when the concept of the supercharged turboprop was conceived, late in 1952, the underlying principles were:

(a) To produce a turboprop with a specific fuel consumption equal to that of the best compound piston engine. This implied a pressure ratio greater than 10:1, and hence necessitated the "two-spool" arrangement of compressors.
(b) To produce a turboprop having a cruising power of the order of 3,500 h.p. at 30,000ft, with the lowest possible specific weight. Such an engine would give more than 8,000 h.p. at full throttle at sea-level and it was desired to restrict this power to between 4,000 and 5,000 h.p. in order to reduce the weight of the reduction gear and airscrew and thus achieve an appreciable improvement in the specific weight under cruising conditions.
(c) To produce a turboprop the take-off power of which was independent of the altitude and air temperature of every aerodrome in the world. Naturally aspirated turbine engines (i.e., those now in use) suffered badly at the higher and hotter airfields, although the adverse effect of high temperature could be partially mitigated by water/ methanol injection.
(d) To exploit to the full the known ability of the gas turbine to produce (relative to piston engines) large power for a small bulk and weight.

Three basic layouts were considered:
1. A single spool comressor driving the propeller gearbox. This was rejected for several reasons. The use of a single spool for such a high pressure ratio would result in an inflexible aerodynamic design needing blow-off valves and variable guide vanes. High starter power would be necessary and operationally a high idling rpm was necessary in order to improve acceleration time and to allow increased propeller drag on approach.
2. A free turbine layout building on Proteus knowledge. Again the high pressure ratio design means that the compressor will have almost as many disadvantages as 1. above. Rejected.
3. 2-spool with LP compressor or 'supercharger'. The LP spool drives the propeller gearbox. This was considered the best compromise between 1. and 2., with the introduction of a moderate inertia problem from the propeller plus LP compressor, whilst preventing the perfect aerodynamic matching that happens on 2-spool turbojet engines such as the Olympus.

As the Britannia was nowhere near its Mach limit in cruise (as had been originally feared) the increased cruise power from the Orion would be easily utilised. Hooker set the target for cruise power at 3,500 eshp at a speed of 350 kts and 30,000 ft altitude; this was double the power of the Proteus 755. The target sfc was 0.4 or less and the engine was to weigh no more than the Proteus.
The Orion ()as it was known after being named on 16th May, 1956) first ran Dec 10th, 1955. It was designed as a 'power egg' and was intended to be supplied as a replacement unit for the Britannia. The target was to swap the Proteus for an Orion in the same time as swapping out a Proteus during normal maintenance, making the upgrade very straight forward.
The Orion's LP spool has a 7-stage LP compressor driven by a 3-stage turbine. The LP spool also drives the compound epicyclic gearbox with a ratio of 0.1006:1.
The HP spool consists of a 5-stage compressor driven by a single-stage turbine. The combustion chamber is of cannular design with 10 flame tubes fed by simplex fuel nozzles.
LP spool rpm is 10,000 rpm at Take-off.

By December 7th 1956 there had been six engines built; 4 were for test bed work and 2 for flight tests. A seventh was under construction. The six engines had amassed 1,00 hours running, including 50 hrs installed in a Britannia. The Britannia, G-ALBO, first flew on Aug 31st.
The MoS withdrew support for Orion development at the end of January 1958, citing financial constraints. About £4.75 m of government funding had been invested. Convair/Canadair were specifying the Orion in the CL-44 Britannia derivative but as work ceased at Bristol they had to go elsewhere (Tyne).

...tbc

 

[tartle]

The only other early engine that Bristol designed but never built was the Janus. In 1945 The MoS was asking for design studies for a 1,000hp turboprop. Frank Owner thought that such a small engine would be a challenge for an axial design so went for a centrifugal. The aerodynamic design was for a 2-stage centrifugal compressor driven by a single stage turbine, a second free turbine drove the propeller gearbox. So far it sounds like a Rolls-Royce RB 60; but here the similarity ends. The compressor impellers were arranged back-to-back, hence the name, and the outlet from the first stage was connected to the second-stage inlet by 4 curved pipes between which four highly skewed combustion chambers were located.
During the design of the engine the Ministry asked Bristol to scale the engine to 500 hp. to avoid multiplication of effort. Owner believed the design was very practical being compact and light.
This is the only picture of the projected design I have ever found... a project illustration.

 

[LowObservable]

Here's the full Flight profile of the Orion, presumably by the Great One himself...

http://www.flightglobal.com/pdfarchive/view/1956/1956%20-%201521.html

Would I be right in guessing that an Orion-Hercules would have been something quite impressive?

Mind you, reading that also caused me to wonder at the polysyllabic style of Flight in those days. I suspect that the sub-editor was closely related to Kipling's Bi-Coloured Python Rock-Snake.

Then the Bi-Coloured-Python-Rock-Snake came down from the bank, and knotted himself in a double-clove-hitch round the Elephant's Child's hind legs, and said, 'Rash and inexperienced traveller, we will now seriously devote ourselves to a little high tension, because if we do not, it is my impression that yonder self-propelling man-of-war with the armour-plated upper deck' (and by this, O Best Beloved, he meant the Crocodile), 'will permanently vitiate your future career.'

 

[tartle]

LowObservable... yes, a useful technical description... as to an Orion-Hercules.. that would have been interesting.. think of being able to override the throttling of the engine and have a guess looking at Short Belfast with Tynes or for extra payload, or a bit of both. I suppose the basic Hercules plus Orion might be analogous to Short Tyne-Belfast?

 

[LowObservable]

Was the Tyne similar in philosophy to the Orion? How would they have compared to the constant-speed, single-shaft (and rather primitive) Allison?

 

[tartle]

LowObservable... you ask an interesting question. My best answer is to look at the parameters that drove the development paths of these engines and how they ended up in the 'marketplace'.
The T56 and Proteus are more or less contemporary engines; the T56 evolving from the T38, which first ran in 1947 following on from private venture studies in 1944 that the US Navy then took on board and financed. A Double-T38 known as the T40 also was developed in the 40s.
Like the Proteus the development period was a long one as Allison wrestled with mechanisms to cope with drag on engine failure and the unreliability of the propeller gearbox caused by lubrication problems. I also reckon the remote gearbox may have allowed more misalignment than a British close-coupled layout as stiffness will be less in the US configuration.
Like the Proteus by the fifties the engine had evolved into a better design which became known as the T56 and was built, run and flown in 1954 as the engine for Lockheed's C-130.
In a NACA RM 'AN INVESTIGATION 1N THE AMES 40- BY 80-FOOT WIND TUNNEL OF A YT-56A TURBOPROP ENGINE INCORPORATING A DECOUPLER AND A CONTROLLED-FEATHERING DEVICE'
By Vernon L. Rogallo, Paul F. Yaggy, and
John L, McCloud III
The summary reads:
"An. investigation of a decoupler and a control:ied-feathering device incorporated with the YT-56A turboprop engine has been made to determine the effectiveness of these devices in reducing the high negative thrust (drag) which accompanies power failure of this type of engine. Power failures were simulated by fuel cut-off, both without either device free to operate, and with each device free to operate singly. The investigation was made through an airspeed range from 50 to 230 mph. It was found that with neither device free to operate, the drag levels realized after power failures at airspeeds above 170 mph would impose vertical tail loads higher than those allowable for the YC-130, the airplane for which the test power package was designed. These levels were reached in approximately one second, The maximum drag realized after power failure was not appreciably
altered by the use of the decoupler although the decoupler did put a limit on the duration of the peak drag.
The controlled-feathering device maintained a level of essentially zero drag after power failure. The use of the decoupler in the YT-56A engine complicates windmilling air-starting procedures and makes it necessary to place operating restrictions on the engine to assure safe flight at lowpower conditions."

So although the T56 is a single shaft engine, there are unique operational challenges that have to be solved in order to safely use the single-shaft design in practice. The A-S Python in the Wyvern was another engine that had severe control challenges.

The Orion for the Britannia was designed to allow the use of higher cruise speeds on the long routes operate by the likes of BOAC where the economic benefits of greater speed are quite clear. The Orion was therefore designed to deliver the correct cruise thrust and then throttled to prevent huge take-off powers beyond the capability of the Britannia to structurally absorb those (Take-Off) thrusts. Also it was a retrofit to an aircraft that had evolved from a 1944 Brabazon concept and had evolved into the Britannia over the years. Had the Brabazon 2 been a straightened-airflow engine the Orion may not have been necessary.

The Tyne evolved from Rolls-Royce's anticipation of the way the Viscount market would evolve and the lack of upgrade to the power needed from the Dart configuration. Rolls had looked at high-power turboprops in the 1940's first with the Clyde and then with the Tweed, which was the original choice for the Princess. As Rolls-Royce had their hands full with Dart, Avon, Nenes they decided to abandon both Clyde and Tweed in the short-term.
By 1952/3 BEA realised that the Viscount turboprop airliner was opening up the marketplace but by 1959 would be to small to economically service it. So they began to think what a replacement would look like. In 1953 Peter Masefield, chief exec at BEA wrote to Vickers with an outline of their thinking of a Viscount replacement with a targeted in service date sometime in 1959.
Vickers working on various studies with props (RB109 and similar) and jets ('baby Conways") concluding that a turboprop delivered better economics on BEA's routes.
On longer routes serviced by BOAC the economics favour a higher cruising speed and so building an engine that is oversize for take-off but optimised for cruise, then limiting the take-off power means that the pilot will have power to take-off with a normal payload even on hot and high airfields. This is the rationale for the Orion.
On the shorter higher frequency routes the higher cruising speed does not have the same payback overall, (also helps to explain why even today turboprops are still favoured for smaller capacity, shorter range routes), so it is more difficult to justify turbojets (back in 50s) but advanced turbofans can and do slowly replace them.
Therefore the Vickers aircraft for BEA, christened Vanguard had an engine developed for it that was optimised for take-off and delivered sufficient thrust at the required cruising speeds and altitudes.
I have put together some numbers for the 3 engines and put them in a table for easy comparison (see below).
The conclusion I have come to is that the T56, although a simple design, has very good thrust/weight ratios. The Orion has a built-in disadvantage in being too big for the T/O job and therefore carries a weight penalty. One wonders what would happen in a non-Britannia derived aircraft application. Or what would happen if you used a Tyne and ran it it a higher cruise thrust (not optimum sfc power).. would the overall weight reduction make up for extra fuel. An interesting situation that makes me realise why performance departments back then had were so big and had so many bright people in them...3 ft sliderule anyone?

....tbc

 

[tartle]

Whilst Bristol had their icing challenges to delay them and so miss the market window, Vickers and RR had a different challenge to lose sleep over. Fortunately the problem that emerged on the Tyne did not take as long to solve but was still a challenge to the intellect!
Opening the Flight magazine for 3rd June 1960 would have come across an article on page 744
"Tyne Investigation
ON May 27 Rolls-Royce Ltd announced that the Tyne turboprop
(described in Flight for April 22) has suffered mechanical failure
of a nature which may not be immediately rectifiable. The trouble
arose during bench testing of a Tyne at Derby and at one point
during the run a compressor disc failed. Although an isolated
occurrence, the company found "confirmatory evidence" on
another engine. A full technical investigation is in hand, which
naturally extends to the many engines already delivered, and
Rolls-Royce have recommended that flying of Vanguards and
CL-44s be suspended until the fault can be rectified. A note on
the effect which this has had on BEA is given in our Air Commerce
section."[below]
STOPPING THE VANGUARD GAP
HOPES that BEA will be able to inaugurate Vanguard services
- on the London - Paris route on July 1 [1960]are now diminishing.
The setback to the Rolls-Royce Tyne programme (see page 744 [above])
is such that BEA are planning on a possible three-month delay
before they will be able to introduce the aircraft into service.
Fortunately for the airline, the effect is not so serious as might
have been the case had the Vanguard been allocated to a larger
number of routes; in fact, at the outset two aircraft only were to
be deployed on the single London - Paris route. Nevertheless,
BEA's forward bookings are heavy on the whole of the network,
and there is no Viscount capacity to spare for redeployment to
fill in the missing Vanguard capacity. The corporation is therefore
seeking to charter, on a bare-hull basis, four Viscounts to
resolve the situation. The crews who have been converted to the
Vanguard will accordingly be reconverted to the Viscount.
As reported last week, the first Vanguard was due to be handed
over to BEA by Vickers on June 15. The machine which BEA
were to have m the meantime for crew-training, G-APED, was
ready to be handed over last Friday. But, with the seven other
Vanguards so far flown (and Canadair's CL-44s), the aircraft is
suspended from flying. The most bitter pill for Vickers to have to
swallow is that the ARB were expected to grant the Vanguard its
Certificate of Airworthiness yesterday, June 2. But assuming that
the unofficial estimate of three months is likely to prove accurate,
at least the TCA Vanguard programme—which calls for delivery
in August—should not be unduly affected."
In fact the investigation and rectification took until November but integrity testing by BEA in order to make sure all was really fixed took a month or two. Limited services began at Christmas, more at Easter 1961 but it was July1961 before BEA were back on their schedule of the previous year.
Tn May 1960 an hp7 compressor disc burst whilst a re-worked Vanguard engine was on pass-off test at Derby. As a genuine disc failure cannot be contained it is a very serious incident. Flying on route proving and prototype flights still went ahead, with severe restrictions, but passenger flights were off the agenda.
The figure below from A C Lovesey's paper 'Gas Turbine Development- Thirteen and a half years in Commercial Aircraft' Aeronautical Jnl V68 No 644 August 1964, shows the disc pieced together after the failure. It had a running time of only 68 hours!
Prolonged lab examination determined the initial failure was the vertical fracture seen in the photograph but little useful additional information was unearthed.
A further 5 cracked discs were discovered in discs all running with times close to the first one. The metallurgical team were sure the problem lay in the material itself. Many discussions with experts within and outside the aero industry but to no avail. The probability of failure of another disc was estimated to be 1 in 700 discs- too high to contemplate and so RR decided that no passenger carrying flight could go ahead until the problem was demonstrably eliminated. This decision was taken after many attempts to eliminate faulty discs, including 64 individual disc rig tests, each test being run to 20 times the number of cycles experienced by the first disc to burst. 3,500 discs were scrapped.
Incidentally this disc material was used in the Conway and the engine had 80,000 hrs of airline experience without incidence.
.tbc
The lab had been looking at the improvements that come from vacuum re-melting of this particular steel alloy, and were convinced that the variability of the steel would be significantly reduced by the process. RR had no choice but to 'forge' ahead, as to revert to a lower strength alloy would necessitate a significant redesign and development programme. The steel maker involved had not got the capacity to vacuum remelt sufficient quantities of steel so RR arranged to have this done at their titanium alloy provider who had spare capacity at this moment. Material testing demonstrated that a significant increase in uniformity and increased properties was obtained from material test pieces that are part of each disc forging. A measure of a ,aterial's ability to even out the stress distribution is the tensile elongation. The change from air melted to vacuum melted steel had a dramatic effect on the elongation as shown in the distribution of results from 300 test pieces of each type, plotted below. The vacuum re-melting process reduces scatter and the improvement in elongation results enabled the acceptance level to be increased one and a half times from 10% to 15%.
Lovesey states that the evidence of improvement outweighed any that could be obtained by other means, particularly as the effort to isolate potentially unreliable discs in the air melted material had failed. It was on this basis that the ARB approved the modification. Subsequent experience has shown their decision to be correct and by 1964 flying hours on Tyne engines stands at one and one quarter million hours with a total absence of any disc material trouble.

 

[tartle]

In the early days of development the Tyne had some serious trouble in the form of fracturing of bolts securing the LP Turbine assembly.
Lovesey described how these bolts were sometimes found with fractures originating under the heads. As often happens the circumstances surrounding the failures presented a confusing story. He goes on...
"A first approach to a mechanical failure is to make the component strong enough, by way of showing that the action proposed gives a large increase in life when subjected to loads that produce typical failures of the original design. This, more often than not, is best done independently of the engine. If the 'forced' failure is identical in character with that experienced in the engine we can be pretty sure that the treatment is the same, and one has a yardstick to measure the effectiveness of design modifications.
The evidence of failure origin, in the radius under the head at a point directed closely towards the centreline of the shaft, resulting in eventual fatigue failure. This evidence and that of fretting patterns on the mating face between flange and disc made it clear that, whatever the disturbing force, the consequent flexure of the flange would most likely be the prime source of high bolt stress, and that attention to the flange stiffness might be a powerful factor in preventing further failures.
A Sonntag fatigue testing machine was used to generate the evidence for the before and after situation. The whole LPT assembly is anchored to the table of the machine. a vibrating force is then transmitted via a rod to the shaft. Results for the original and modified design can then be collected. The flange could not just be thickened up as the gap between HPT and LPT assemblies is tight; the flange is thickened between the bolt heads. The results show a 60+ improvement in life as tests were discontinued with no sign of failure.
No failures of the modified flange have been found and much later the LP rear bearing was modified to have improved damping and the source of vibratory forces was removed, having an overall benefit on the engine as well as easing stresses on the flange....
The fourth picture below shows (top) the original LPT shaft and (bottom) the shaft with modified flange.

 

[LowObservable]

Interesting numbers...

The Orion itself would have been a big weight penalty (5600 pounds for 4 bare engines) for the Herk, but then it was a bigger engine even at T/O rating. But if you'd scaled the engine down and used the same philosophy, the weight penalty would have been much smaller, the aircraft would have cruised faster, higher and more efficiently, and the operators would have loved the hot-high-short capability that it would have given you.

 

[tartle]

LowObservable... good point... it would lower the penalty... the Tyne of course restored hot/high power by using water/methanol injection at a small weight penalty to restore power to sea level conditions at 10,000 ft and between +15 to +20 deg C.
Just as a first order excercise I have replotted Orion performance curve from the Proteus cruise power and looked to a cutoff of 5050 hp. I btlieve the Orion output was mechanically limited by the main gearbox no thermally limited so we would have to ask what the turbine entry temperature implications of having a higher than scaled sea level power would be.
We seem to have scaled Orion 8,200 ehp to 6,900 ehp i.e a .841 scale. The mass flow will then be 68 lb/sec which is a massflow scale from Tyne of 1.48 so a linear scale of 1.22... over 20% bigger. The economics of a Herc with such an engine would then have to be done, but my gut feel is that getting the engine rightsized as normal and then using water/methanol may be a more cost effective solution... but if economics do not weigh so heavily it would be interesting. Knowing operators minds a little... I would see mil/civil operators looking to use all the power they could get whenever they could get it so they would have ended up supersizing the load anyway!
But superficial calculations can be misleading!

 

[alertken]

NBMR.4 was put out to tender in 1961: NATO-Standard (V)STOL freighter, to support NBMR.3 (V)STOL strike fighter. UK tired of infighting, could not wait, so in 1963 took HS.681.

Bristol took a C-130E licence and bid that as T.222/Tyne, and again to UK-only Reqt.OR.351 that was met by 681. When 681 was chopped, 2/65, RR chose to dust off desultorily its T.222 installation scheming and submit a slim brochure for Tyne to be shipped to Marietta for C-130K. They put little effort into it because:
- no evident operational benefit; but much more significantly:
- incredible to contemplate such a certification exercise in the very short lead-time to delivery of T56/C-130K.

MoS/NGTE put much effort from 1944 into Theseus/Proteus/Orion. Brabazon Committee Type III (ultimately, Britannia) and Vickers Windsor were to have taken Theseus. Bristol's 1946 bid to (by then Medium Range Empire, to be initially Centaurus/, then Proteus/Britannia) was a Centaurus/licenced L-649 to be followed by a Theseus/L.849. Cabinet declined to make $ available: for me, a tragic What If: DC-7Orion, L-1649Orion, 1954-ish. (France negotiated a DC-6 licence, 1946: Communist influence in turgid 4th Republic politics scuppered that. Theseus could have been hung on that, too).

 

[LowObservable]

Orion-Starliner... that would have been something. Were the Mach limits similar to the Britannia, though? And then there was this project:

http://www.flightglobal.com/pdfarchive/view/1955/1955%20-%201675.html

http://www.flightglobal.com/pdfarchive/view/1956/1956%20-%200925.html?search=britannia%20600%20type%20187

We'd have got to jets eventually, but fast props with better long-range performance could well have held off the competition from early one-stop jets.

And it's a little ironic that the US banned the DC-6 for France just as we shipped Nenes to Uncle Joe. Hey Vlad, how about some royalty payments?

 

[tartle]

LoeObservable... the birth of the CL44; one wonders whether a differnent sort of political climate might have allowed the high Mach Number mods to appear in that aircraft's development programme which would have made a market niche for the Orion; it seems turbojets killed that engine as the aircraft/engine combinations were not pursued with the urgency needed to take advantage of the high MNo turboprop window... as you pointed out.
The first details of Electra in Flight magazine also had a section on engine selection which I have cut and pasted below:"Powerplants. We have previously described the provision of
power for the Electra as a "nigger in the wood-pile," owing to
the fact that the engine assumed in the original projection of the
design—the Allison 510—is not yet fully developed as an airline
engine. Before describing the installation of this engine it is
therefore worth adding a few notes on alternative units.
The Electra is designed to use any modern turboprop in the
3,000 to 4,500 h.p. class. Into these limits can be fined the
Allison 510, the Rolls-Royce Tyne, the Napier Eland and—fully
developed but of earner conception—the Bristol Proteus.
The Allison suffers from a number of fundamental deficiencies,
which become serious when considering the engine from an
airline viewpoint. At present, it is a constant-speed engine
(13,820 r.p.m.) and even the ground idling speed is as high as
10,000 r.p.m., resulting in a high noise level on the ground.
Allison have put a great deal of effort into perfecting the control
system, and have brought the engine/airscrew combination to the
point at which it can be accepted for service in the C-130 for the
U.S.A.F. General handling, however, is still not as good as it
should be for an airliner.
Another factor which should not be overlooked is that the
Allison 501-D-10 is not yet ready to deliver its full design power in
commercial service. This power is 3,460 sJi.p. (3,750 e.sii.p.)
at sea level, i.s.a., at zero forward speed, and to achieve this the
turbine inlet temperature has to rise to 1,780 deg F (1,320 deg K),
making it one of the hottest engines yet flown. Development
running is going forward with the 5O1-D-12 commercial engine
with a maximum rating of 3,015 e.s.h.p. at 1,625 deg F max.
t.i.tcmp., but the Electra needs more power than this. For future
development, Allison hope to achieve well over 4,000 e.s.h.p. in
the 501-D-8 and D-13, but it would seem that they will be hard
pressed to reach this goal.
From a number of aspects, the Tyne and Eland appear to fit
the requirements of the Electra very well. The former is an
outstandingly efficient unit and will provide more than enough
power. The latter is also well-engineered and efficient, and will
shortly deliver considerably more power than the Allison, using
air-cooled turbine blading. The view of Lockheed, expressed
directly to us earlier this year was that, although the American
Airlines' Electras were to be powered by the Allison 501, it
should not be inferred that this would become the standard, or
even the most common, Electra engine. One Lockheed engineer
said, bluntly, "we want the engine from Derby; and if we can't
have the Rolls engine in time, the Eland has a good chance."
Anyway, for the time being the Electra is wrapped around four
Allison 501s. Each is installed as a complete power package and,
except for certain handed items and accessories, all four are interchangeable.
These Allison engines have the reduction gearbox
mounted remotely from the power section (see detail drawing on
p. 715[below]), thus allowing the air intake free entry to the compressor.
In the C-130 the intake is under the spinner, but various factors
have resulted in the Electra engine being "the other way up," i.e.
the axis of the airscrew lies below the axis of the engine, and the
main air intake duct passes down from the top of the cowling
behind the spinner.....etc"
Reading all the press gives an impression of lots of activity around these engines but due to inaction the opportunity was missed... all one can say about Tyne and T56 is that ultimately they found (military) markets of decent size.

 

[LowObservable]

Flight's stylebook has changed in the last 57 years...

Interesting stuff about a momentous era. Had Lockheed had the Tyne, they wouldn't have had this problem:

http://www.airspacemag.com/military-aviation/The_Hammer.html

 

[robunos]

Tartle,
In #186 in this thread, in other threads on SPF, and elsewhere online, mention is made of thr Rolls-Royce Tweed.
However other than a purported designation of 'AJ.25', I can find no more information on this engine. As it seems to be truly a 'Secret Project', could you tell us a little more about it...


cheers,
Robin.

 

[tartle]

Robin.... the AJ25 was a jet engine project of 2500lbt; the AP25 was a turboprop version. The AJ25 had thinking in it that fed into BJ45 that eventually evolved into a Conway and AJ65 that eventually became the Avon.
Until I have found my Tweed info this picture of the mockup gives an idea of it.
Correction: I have changed AJ45 to BJ45 so the section below is also BJ not AJ!
....tbc

 

[robunos]

Many thanks for that...and the AJ.45 is completely new to me...


cheers,
Robin.

 

[tartle]

Robin,
I made a typo.. it is the BJ45 started to be schemed on the 23 Oct 1946. I knew Don Eyre, who was A A Griffith's personal designer, and he had a sense of history and so was happy to chat. The B of course stood for Bypass and this was the first scheme issued around that concept.. much work by the Main Project Office followed and the concept emerged as the Conway. The BJ80 Conway was first schemed out as an engine with the first 4 stages of the AJ65 Avon as the LP Compressor, driven by a single stage turbine and the Tweed compressor spool as the HP system, also driven by a single stage. Note that at this period of understanding RR and others were designing highly loaded turbines in order to minimise engine length and weight. The Dart was another engine that was suffering from single-stage-itis. When it first ran it had a turbine efficiency of 80%, way below target... by attention to losses due to leakage this was raised to 86%.. but that is another story. The RR turbine efficiency curves (that are very accurate and derived from model testing and are corrected for losses- i.e. the leakage factors on model are factored out) show that for a given work output dividing the work over two stages will yield a 4% improvement in overall efficiency, less the losses over the second stage... worth having if the weight issue can be tackled and length is not a critical issue. The history of the early Dart brings this out and is a fascinating contrast to the way the Proteus issues were tackled.....so I'll do that shortly.
Back to the Tweed.
This engine was an exploration of a 2,500 hp turboprop that was pure axial. It was designed by John Reed under the watchful eye of Fred Hinkley and was followed by what John called an AP12½... which, unsurprisingly was a scale-down to 1250 hp during the period when the MoS was looking for engine schemes at that power. The compressor follows the Metrovick design on the Clyde as it has constant outer diameter and a rising hub diameter. RR favoured this arrangement on the turboprop as accessories were still contained within the intake diameter; but in the case of turbojets the reverse is true, resulting in a tapered outer casing .. as it kept the installed diameter of the engine down when accessories were included in the general arrangement, as they could be included between the inlet and outlet diameters of the casing. The combustion chamber took the best of thinking at Lucas and Derby and was schemed as 8 separate chambers. 2 turbine stages were used as the spool also drove the propeller.

 

[tartle]

Lionel Haworth was the chief designer on Clyde and Dart:
In his biographical memoir (download here) it was stated:
"In March 1944 he was responsible for the design of the compressor and gearbox of the RB39 Clyde turbo-propeller engine, which first ran in 1945. This was the first two-shaft aircraft gas-turbine engine, using an axial flow compressor followed by a centrifugal derived from the Merlin supercharger on the high-pressure (HP) shaft, the propeller reduction gear being driven from the low-pressure (LP) shaft. The Clyde proved to be a powerful and reliable engine but despite its potential it was not adopted for production. However, valuable experience was gained particularly on fuel and propeller control systems.
THE DART
In 1946 a small team at Derby under Lionel Haworth designed the new and relatively simple turboprop, the RB53 Dart. It was aimed at 1000 shaft horsepower (SHP) to satisfy a Ministry requirement for a trainer aircraft. An engine of this class was also needed for a short-range turboprop airliner under study by Armstrong Whitworth and Vickers.
Haworth chose to use a tandem centrifugal compressor as in the latest piston engine superchargers, mounted on the same shaft as the two-stage turbine, which also drove the gearbox and propeller. From the HP compressor the air passed through seven combustion chambers arranged at a skew angle to reduce length and avoid a right-handed bend in the flow. At the front of the main shaft was a helical high-speed pinion connected by three layshafts to a second stage of reduction gearing with spur gears. The Dart was overweight and down on power when it first flew in the nose of a Lancaster in October 1947. It did not compare well with the
The first flight on 16 July 1948 was a great success, and by then the Dart was meeting its design criteria. Haworth had already planned the RDa3, giving almost 50% more power; this was immediately put to use in stretching the now very promising Viscount to meet the needs of BEA with an increase in seating from 32 to 53. Limited scheduled services started on 29 July 1950, the first revenue-producing service by any gas turbine in the world. An order was placed for 20 V701 Viscounts and this was later increased, followed by many others."

As discussed above, the aerodynamics and structural understanding which was derived from the remorseless development programme on the Merlin... see below for the development drivers that ensured that the Spitfire was at least on par with the Me 109 and later the FW 190. Hooker was key to driving Supercharger performance and Lovesey made sure it hung together.
Lionel Haworth was a practical hands-on engineer and designer and was also very keen on mentoring those who followed his footsteps into RR Design and Development... he therefore was always keen to lecture the new intakes of apprentices to help make sure thay realised why they were being trained so thoroughly and the opportunities that could open up for them. I was too young to have heard him speak but his team that were rising over the years spoke the same message. Lombard and Morley lectured my year.
Back to Haworth.. the notes made for one of his lectures form the backbone of the story below, and it contrasts with the lack of urgency and direction at Bristol 'til Hooker arrived.

Rod Banks in his autobiography 'I Kept no Diary' talks of Peter Masefield and his inspired backing for the Viscount that helped make it a success. Rod wrote in the context of our lack of commercial aircraft success- Trident for instance:
" It could be asked: what about the Viscount? The answer is that the '630', the Viscount prototype, with its Dart engines, was designed and built at government expense ahead of any B.E.A. requirement, experience or thinking. It was Peter Masefield, as Chief Executive of B.E.A., who eventually took it on and had the passenger capacity increased from 32 to 47, i.e. the '701'." An important player, was Masefield- his obituary is here.

Haworth covered the background to the Dart and its development
Towards the end of the War the Government set up a committee, under the chairmanship of Lord Brabazon, to prepare 5 specifications for civil aircraft that the World might need in the 20 years after the War........ of the more succesful aeroplanes within the Brabazon specifications are the de Havilland Comet and Vickers Viscount.

The Viscount was planned as a medium speed (300 mph) airliner of an all up weight of about 40,000lb to carry 36 passengers. For this job it required 4 gas turbine propeller engines of 1000 hp. We commenced work on the RB 53, the Dart, early in 1945, while Armstrong Siddeley commenced work on the Mamba. There were other aircraft competing in this field- Armstrong Siddeley built the Apollo and Handley Page also designed an aircrfat to meet this spec'n.
The RB 53 was to give 1000 hp at the propeller for take off and prop diameter was limited to 10 ft. this was fixed by consideration of wing span, wing loadings, landing speeds, and length and weight of undercarriage.

 

[tartle]

Lionel Haworth went on to say:
"It was natural that we chose for the compressor of the RB.53 a two-stage centrifugal type, the type which in fact we had used on later Merlin and Griffon engines. This was drawing on our previous piston engine experience; we thought that we knew something about reduction gearing since the Merlin and Griffon featured these components, but we had not attempted anything with such a large ratio as .106. The piston engine had a much smaller reduction gear ratio of .442. Having decided on the main construction of the engine, some calculations were made to determine the approximate size of the engine. In order to have 1000 hp at the propeller it was necessary to have a turbine that developed about 3000 hp driving a compressor that absorbed about 2000 hp.Thus the rough basic dimensions were arrived at. With the compressor we were hoping to use our latest knowledge and maybe a little beyond. With the turbine we decided to go beyond our jet experience and have a two-stage turbine. This engine was the first aero engine in the world to feature 2 stages in the turbine.
The engine was then sketched out to produce a design scheme with the main components worked up in enough detail for estimates of weight to be made. The project layout or scheme is then turned into a detail design and a General Arrangement generated."

The RB 53 design project started on or around 25th April 1945 and it took 57,048 man hours over a period of 63 weeks to complete the initial design. Detail design of component parts began 12th June 1946 and took another 29,195 man hours over the next 54 weeks. First detail drawings were issued to the shops on 1st Nov 1945 and the first engine was built and ready for test by July 10th 1946.

Haworth continues:
"As soon as the engine was completely assembled, it was put on a weighing machine. Imagine how we felt when instead of showing the predicted 700 lb, the scales showed over 1100 lb! But this was not the worst blow we suffered during that second week of July, 1946; a few days later the engine was wheeled from the shops to the test bed and the anxiety as to whether it would in fact run was completely overshadowed by an inquest in the design office on the weight analysis. The engine,however, did run even under its own power, but when the next day it was decided to open up to full power we were completely stunned by the news - our 1000hp engine could only achieve a little over 600 hp. It, however, was decided to persevere with the 'heavy' engine and within a matter of a month or so it had completed a 50 hr endurance test. We were now faced with the task of finding the power and reducing the weight and these two design problems occupied us for the rest of 1946 and well into 1947.
Every piece of the engine was carefully examined to see were metal sections could be reduced or avoided all together to save weight. Magnesium was used extensively for castings instead of aluminium, and by a process of pruning every ounce of weight on desin and making certain castings were manufactured to drawing thickness (almost 100 lb of weight was put on through overthick castings) we managed to achieve a saving of some 350 lb.
The loss in power output was not really surprising on closer examination, for instead of the turbine developing 3000 hp it only achieved about 2800 hp, and instead of the compressor requiring 2000 hp it was taking over 2200 hp leaving only 600 hp for the propeller.
Another version of the RB 53 was produced which gave about 900 hp and weighed 800 lb. This engine was developed up to 1000 hp and in Aug 1947 the first 150 hr Type Test was attempted. It was not until Dec 1948 that a clear run throught the 150 hr test schedule was achieved, at a power of 1045 hp. In April 1949 a 500 hr endurance test was completed at this rating."
An aside: At some point during the War Haworth had visited Dr. R. W. Bailey at MetroVick, Manchester and had been shown a disc in the laboratory being subjected to steep radial thermal gradients by alternately heating and cooling the rim. Dr Bailey expalined he was simulating the thermal stresses that occur transiently in an engine, saying he planned to find out if the metal would attain what he called a 'cyclic state'. He further said he believed the rim hoop stress alternated between tensile and comprssion in excess of the material yield.
Haworth decided that such punishment would fail the disc and went away to find a design solution to the problem.
The 4th attachment below shows a crack at the bottom of the blade root slot of a Derwent engine. 13 Derwent discs failed in service from cracks like this (typical across the industry at the time). While these failures were occuring Haworth was having a design solutiion fitted to the improved RB 53 being developed. This was the extended root blade.
Back to Haworth:
" The improved power derived from the turbine by blocking up the leaks under the NGVs prompted some thought to be put into the idea of blocking up the leaks over the top of the turbine blade.
This was eventually achieved by designing a turbine blade with a shroud on it and to seal off the hot gas on the front edge of the blade by a very narrow land running against a face with as small a clearance as possible. Hence this engine was the first aero engine to have shrouded blades and, incidentally, long roots. Obtaining the correct clearance was very much more difficult to calculate and we resorted to trial and error, observing the stae of the shroud upper surface through holes in the casing. We adjusted the running clearances until we only had a slight rub.
The improved turbine design gave a substantial increase in efficiency and by Oct 1948 we were able to complete a type test at 1100 hp and an sfc of .867.