TSR2
Page 32
In April BSEL dropped a bombshell. Exasperated by the snail’s pace of contract negotiations with the MoS, and concerned that stretching the 15R to meet the requirement could be a long and expensive process, it warned that the predicted development costs had risen dramatically, from around £4 million to a stunning £15 million. The reaction was not one it expected. The Ministry informed them that it was instigating an immediate investigation into the circumstances of this increase, and that, if the figures were correct, the competition between engine companies would be reopened. In the meantime the Ministry would not be liable for any costs incurred by BSEL on any TSR2 engine work.
The investigation certainly succeeded in scaring BSEL into producing some lower figures, though the Ministry did not, it appears, have any serious intention of reopening the competition. The Treasury stepped in, objecting to the MoS’s edict against further consideration of Rolls-Royce proposals, because if BSEL was being given the opportunity to offer improved figures, then it was unfair not to let Rolls-Royce have another crack at it. The Ministry refused to budge. The RAF also refused to accept a drop in range, and it became clear that the Olympus 15R was not going to be suitable and that the even more expensive 22R was going to be necessary. BSEL offered two versions of the 22R, one containing an extra compressor stage and able to cope with intake air up to temperatures of 146°, and one with two extra compressor stages and able to cope with temperatures up to 180°. Development costs for these would be £8 million and £10.3 million respectively; still much more than BSEL’s original estimates, but not as frightening as that £15 million figure.
Further delay ensued while the MoS and the Treasury argued the toss over the selection of an engine supplier, nearly a year after the decision had already been taken. By May BSEL had realized that further work was going to be tricky in the face of uncertainty as to which of its engines was really going to be required, but that work on the reheat unit could at least continue, as the same size of reheat unit would be required regardless of the engine chosen. Not until August 1959 was BSEL in a position to be entirely confident that it had the job of producing the engine, although a signed contract would not be in its hands until nearly the end of the year. The company had agreed to a fixed-cost contract to develop the engine up to initial CA release (predicted at that point to be mid-1965), with provisions for profit reductions and cost penalties if certain targets were not met. The 146° 22R was the choice; and BSEL had been beaten down to a development cost of £8 million, just £1.5 million more than the amount the MoS had privately believed would be necessary.
Engine specification
The final specification for the overall aircraft, RB.192D, laid out a tight requirement for engine performance. Single-engine flight was to enable steady flight at altitudes up to 40,000ft (12,000m) in reheat or 30,000ft (9,000m) without reheat. No surging would be permissible within the aircraft’s normal flight envelope, including during weapons release, even when making large throttle movements. Slam accelerations, opening the throttle to 100 per cent as quickly as possible, were to take no more than 1sec below 10,000ft (3,000m) or 3sec above that altitude, with the engine itself responding in a smooth and rapid manner, accelerating to the required speed in a similar timeframe. Repeated slam accelerations or decelerations were to be permitted in a range of flight conditions, and when the aircraft was stationary on the ground, to enable carefree engine handling particularly in the take-off and landing phases. The reheat unit was to be fully variable, and would be required to operate in a stable manner throughout the flight envelope with similarly tight constraints on response time to slam accelerations. Running on either AVTUR or AVTAG fuel, all of the handling requirements needed to be met throughout the aircraft’s flight envelope up to an altitude of 56,000ft (17,000m), with performance to remain satisfactory up to 70,000ft (21,000m). Some limitations would be accepted when running on AVCAT fuel, these to be decided based on flight testing.
Following on from these requirements, the MoA’s Directorate of Engine Research and Development issued its own more detailed specification to cover the engines used for bench testing and development-batch aircraft, this specification being issued as D.Eng.RD.2437. The highlights of this specification were that each engine was to weigh 4,290lb (1,950kg) (for the basic unit; including the reheat unit and accessories, the weight would rise to 6,150lb (2,790kg)) and produce thrust of 30,610lb (13,890kg) in full reheat or 16,780lb (7,615kg) in full dry power, with an SFC of 1.87 (reheat) or 0.70 (dry) respectively. The fuel consumption figures in particular would prove difficult targets to meet.
The aircraft’s speed and range requirements also complicated matters. Initially, RB.192D had been rather vague about the aircraft’s top speed of Mach 2, with no solid requirement for just how long such a speed should be maintained, though it did mention the ability to cover 1,000nm at Mach 1.7. The 22R could handle the turbine entry temperatures necessary for sustained flight at Mach 2 only at the expense of a short overhaul life. Alternatively, if the turbine entry temperature could be reduced and extra reheat used to restore the lost thrust, sustained Mach 2 flight and a normal overhaul life would be possible, but at the expense of higher fuel consumption and thus a reduction in range.
Engine testing at the NGTE
The major problem facing BSEL was the fact that the existing Olympus variants were all designed for the subsonic role, and sustained supersonic performance introduced a host of challenges regarding intake design and the speed, temperature and pressure of intake air. The addition of a reheat unit was a relatively simple task by comparison, and on this score at least BSEL had already been experimenting with reheat on the Olympus 14R. The Olympus 22R, unlike the 14R, was based upon the Olympus 21 (the Mk 301 for the Vulcan B.2), and its exterior appearance showed some similarities to the 21, the major difference being the large reheat unit. Owing to the higher temperatures and pressures involved, the materials used in constructing the engine were different in some areas, notably the compressors, conventional alloys being used less and being replaced with titanium, steel and nimonic.
An Olympus 22R running in reheat in a test cell at Patchway. Without facilities like this and those offered by the NGTE, Olympus development, along with that of most British jet engines, would have been next to impossible for any single manufacturer. Rolls-Royce Heritage Trust via Brooklands Museum
While Bristol’s own plant at Filton had basic engine-test facilities, the NGTE at Pyestock near Farnborough was to be used for much of the simulated flight-testing engine runs, the Establishment’s newly uprated test cells making it possible to test-run engines in a variety of conditions, including high-altitude in the newest test cell. A relatively small establishment, NGTE Pyestock specialized in engine component testing before 1957, but underwent massive expansion from that point onward to enable the testing of complete engines in heavily armoured and instrumented test cells. Complete engines could be lowered by crane into a test cell, connected to an array of instrumentation and run continuously on endurance and power tests. The cells themselves were huge steel-walled cylinders. The first, Cells 1 and 2, designed initially for ramjet testing, were of 12ft (3.6m) diameter and 120ft (36.5m) long. Cell 3, opened in 1962, was designed for altitude testing and had a larger diameter of 20ft (6.0m), though it was shorter, at 80ft (24.4m). Each cell had an adjoining control room (Cells 1 and 2 shared a single room situated between them). Small armoured glass windows looked into the cells, which were supplied with air piped from the huge Air House elsewhere on the site. Exhaust was directed away from the cells into large diffusers, but even so, the thunderous roar from the NGTE could be heard for miles around whenever engines were being run at high power.
Initial tests were made with an Olympus 22DR (a derivative engine, basically an Olympus 201 with an added LP compressor stage), and mostly showed that Cell 2 was not really up to the job of testing such a large engine. Much work was done to improve the cell’s capability to handle the amounts of intake and exhaust airflow, and Cell 2
was later used for much of the lowaltitude/high-subsonic-speed simulation work, while Cell 3 handled the high-altitude work. The intention was that test results in Cells 2 and 3 would be verified with results obtained from mounting an Olympus on an Avro Vulcan flying test bed (FTB). However, intake temperatures for sustained supersonic flight were significantly higher than those experienced in subsonic operations, and one of the major tests was to run the engines in a suitable test cell, one with heated intake air. The MoA wanted the engine to be able to deal with intake air at a temperature of 146°, with short periods at 180°. No testing cell existed that was capable of providing intake air of this temperature, so considerable effort went into the creation, and subsequent testing, of a unit to provide heated intake air to the new test Cell 4. On at least one occasion the pre-heater that provided the heated intake air malfunctioned, allowing unburnt heater fuel to be sprayed into the Olympus and causing the engine to overheat as a result.
Engine description
Meeting the sustained 146° intake temperature requirement needed some substantial reworking of the existing Olympus design, some of which would be far-reaching in its implications, notably the relocation of the HP thrust bearing between the LP and HP compressor driveshafts to the intermediate casing, which required lengthening and expansion of the LP shaft. This simplified the supply of both cooling air and oil to the bearings. The first Olympus 22, the 22DR, initially used a single additional LP compressor stage with aluminium blades, but these proved susceptible to flutter during testing and were replaced by titanium blades that helped alleviate the problem, though did not entirely cure it.
The 22R’s LP compressor spool had eight stages (seven for the 22DR and six for previous Olympus variants), and the HP compressor spool, seven. The combustion system had eight fuel injection points as with earlier Olympus marks, but two fewer flame tubes. The reheat system was of conventional design, with a parallel-sided duct linking the engine to the reheat diffuser section, combustion section and the fully variable nozzle. Eight pairs of pneumatic jacks operated rods extending rearwards to the annular shroud, which operated the fully-variable final nozzle area. The nozzle was of convergent shape. Reheat lighting was by the hot streak method, which entailed the injection of fuel upstream of the turbine, producing a streak of flame that was conveyed to the reheat combustion zone via fuel injection points in the jetpipe. Combustion stabilization was again conventional, with three annular spray rings and V-shaped gutters. Thermal blankets wrapped the engine aft of the HP compressor casing, cooling airflow being directed in the space between the engines and the engine mounting tunnel walls.
An Olympus 22R LP compressor rotor. The redesigned LP rotor shaft of the 22R suffered several dramatic failures, resulting in expensive delays for the entire TSR2 programme. Rolls-Royce Heritage Trust
An Olympus 22R HP compressor rotor. Rolls-Royce Heritage Trust
The Cumulus auxiliary power plant was mounted in a small bay forward of the weapons bay and Doppler compartment, though the first two aircraft were not fitted with Cumulus, using the vacant bay for flight test instrumentation. BAE Systems via Brooklands Museum
Control of the engine was to be by a combined electro-hydro-mechanical system, Bristol having successfully used such a system on its Proteus turboprop (used by the Saunders-Roe Princess flying boat and the Bristol Britannia airliner). Rather than have any mechanical link between the pilot’s throttle levers and the engines, actuators on the engine would be driven by electrical signals from control units on the levers. In the event of a fault in the 115V 400Hz AC electrical system, the engines would remain in their current running state, and a manually switched backup 28V DC supply would enable emergency throttle control via a DC motor driving the throttle. The engine was fitted with various sensors to protect it against a variety of dangerous conditions, such as a turbine temperature limiter that would progressively close the throttle automatically if preset temperature limits were exceeded. In testing, the control amplifier and associated units that made up the electrical portion of the engine control circuit proved very reliable, and this contributed greatly to the ease of engine handling.
Three initial versions of the 22R were to be developed. The Mk 320X was a limited performance version, given certain concessions to deviate from the required specification, and to be fitted only to the first few airframes (nine engines to be built). The Mk 320 would then attempt to meet all the required targets and be fitted to the remaining development-batch aircraft (eighteen engines to be built). The Mk 321 was the production version of the engine, to be fitted to pre-production aircraft (an initial batch of thity-five engines to be built).
Starting the engine
The original OR had required a rapid engine-starting time of 10sec, similar to that of fighter types required to scramble in a hurry. This was revised downward to 35, to 45sec, and later to ‘under one minute’ which gave much more scope to consider different starting schemes. Ten different options for an engine starting system were investigated, including external LP air sources and AVPIN starting as used on the Lightning and Javelin, but the final choice was a Palouste carried as a permanent AAPP.
Vickers was responsible for this area of the aircraft, and its initial choice of location for the AAPP was on the top of the fuselage behind the wing, within the curved spine present on the early drawings for the aircraft. However, the spine was removed when the wing design was finalized, and the remaining space in the area was reserved for the flap-blowing pipes and flap actuating mechanism. An alternative upper location ahead of the wing was unsuitable, as it encroached upon structural members and fuel tanks, so a significant weight and fuel penalty would be incurred even if the structural issues could be resolved. The third choice therefore became the final location for the AAPP; a small bay just ahead of the weapons bay, where, in retracted position, the AAPP would be on the aircraft centre-line.
The lower location also offered distinct advantages in terms of weight and complexity. The unit could be lowered with gravity assistance rather than the hydraulic power needed to raise a unit positioned higher, and could also operate using a gravity fuel feed, needing no additional fuel pump. The AAPP would be attached to the bay door, which hinged at the port edge. When the door was opened for operation the AAPP would hang down, offset to the port side, with the intake forward and exhausting to the rear, inclined 5 degrees downwards. This was quickly found to be an unsuitable orientation, as with the AAPP in operation the hot jet exhaust would preclude any access to the weapons bay or undercarriage bays. Instead, the AAPP’s orientation was changed by 90 degrees so that the intake was from the port side with exhaust to the starboard side. Although this would create a substantial danger zone to the starboard side of the aircraft within a narrow cone starting at the port intake, it was considered to be acceptable. The AAPP would be removable, with the aircraft able to be flown without it. External LP air connectors would be available for ground LP supply if needed.
Blackburn, licensed manufacturer of the Palouste (a Turbomeca product), proposed an improved APU based on the Palouste but of higher power output, smaller dimensions and less weight (though development would see its weight rise). In August 1960 this improved unit was named the Cumulus and accepted as the AAPP for the TSR2. At this time, any thought of being able to extend the AAPP in flight to cater for restart of a flamed-out engine was also discarded, as the characteristics of the Olympus 22R were becoming more apparent and it was realized a windmilling re-start would be possible at speeds down to 200kt (230mph; 370km/h) (a year earlier, 300kt (345mph; 555km/h) had been the expected figure). A warning light would now indicate if the AAPP was still lowered when the pilot made the switch selection taking the aircraft from Ground to Flight before starting a take-off, and a further warning and automatic retraction would occur if the undercarriage began retracting with the AAPP still extended.
The Cumulus’s greater power enabled the addition of an alternator to provide limited AC power to the aircraft to run some systems wi
thout engine or external power being provided. Chief among these systems was the Stable Platform, which could then be kept running during a rearming/ refuelling turnaround, thus avoiding the 15min warm-up period, though from a flight safety point of view it was preferable that the AAPP was shut down during refuelling in peacetime. The AAPP was initially to be started via a hand-pumped hydraulic accumulator(!) but a more practical cartridge starter was eventually specified.
The auxiliary power plant bay of XR222; the first two aircraft were not fitted with the Cumulus, and the bay housed flight test instrumentation instead. Damien Burke
The engine-starting air supply system. BAE Systems via Brooklands Museum
The general arrangement of the CSDS. Rolls-Royce Heritage Trust
Engine starter motor and electricity generation
Jet engines of the era generated AC power by means of a generator driven by a con-stant-speed drive (CSD), itself driven from the engine gearbox. Effectively, the CSD was a variable-displacement hydraulic pump driving an hydraulic motor at a constant speed regardless of engine speed, the motor running the generator to provide a steady frequency.
Three CSDs were seriously considered for use on the TSR2, these being from English Electric/Sunstrand, Hobson and an entirely new unit from Plessey. The last had been developed in less than two months to a standard that already met the requirements. It comprised a pneumatic-mechanical drive with an air motor as a constant-speed trimming device, with the added bonus that the air motor could be used for engine starting through the gearbox, when operated by LP air from the AAPP. With the gearbox declutched from the engine, ground checking of gearbox accessories would also be available. Neither of the other two CSD units offered this dual functionality, and both would incur substantial weight penalties and development costs if they were combined with a starter unit. The choice of the Plessey unit was therefore easy to make, and the relatively few problems found during NGTE’s testing were straightforward to fix and resulted in a reliable unit.