TSR2
Page 30
A single firing handle was located at the front of the seat base; no face blind or side handles were provided. The face-blind handle in particular was considered too difficult to reach in high-g situations, and the blind itself was deemed unnecessary when TSR2 crews would wear helmets with visors. The helmet itself was attached to the seat by straps that pulled the occupant’s head firmly against the headrest before ejection, and leg and arm restraints pulled the limbs into safe positions. After canopy jettison, automatically triggered by pulling the ejection seat handle, or by manual selection, the main ejection gun would fire, and primary and auxiliary explosive charges would boost the seat out of the aircraft, a seat-base rocket array quickly igniting to thrust the seat forward and assist with fin clearance. A duplex drogue system would then fire a small controller drogue, followed by a larger stabilizer drogue parachute, tilting the seat into a horizontal attitude to shield the occupant from further wind-blast and ensure that deceleration forces were approximately in line with the seat axis. The main parachute was housed within the rigid headrest of the seat (which had a large recess to accommodate the head when it was pulled back against the headrest). The drogue chute was held within the left-hand guide beam on the back of the seat. The right-hand beam held the barometric release gear for sensing altitude, to open the parachute automatically at a safe height. In the event of a high-altitude ejection, no further action would take place until the seat had descended to a safer altitude (10,000ft (3,000m)) at which seat separation and main-parachute deployment could occur. The occupant would use the emergency oxygen supply on the way down. Otherwise, separation (including guillotining of head, arm and leg restraint straps) would occur after a short delay, just long enough to allow the seat to decelerate to a speed acceptable for deployment of the 28ft (8.5m)-diameter Irvin parachute.
One of the first mockups of the new Martin-Baker seat for the TSR2. Compared with the final Mk.8VA, the headbox is the most obviously different area. BAE Systems via Warton Heritage Group
Once the occupant was on the ground (or in the water) the survival pack became his next focus of interest. The seat cushion (much praised by the crews as being far more comfortable than those of other ejection seats on which they had spent time) doubled as a container for a dinghy and survival pack, and separated from the seat assembly at the same time as the occupant, remaining attached to them by a long strap. The dinghy, a simple one-man affair with a canopy and hood for protection against the elements, was inflated by a CO2 cylinder. Martin-Baker did put forward proposals to include a chaff launcher in the seat so that it could cause a bloom on radar at the point of ejection and pinpoint the site for search-and-rescue efforts, but this was not incorporated in the final seat.
Relations between Vickers and Martin-Baker appear to have been strained from the start, with Martin-Baker repeatedly being blamed for causing hold-ups in the cockpit design owing to lack of a representative seat mock-up. Seat assemblies required for test-track firing trials on various dates in 1962 were not delivered in time, and development dragged on. It has to be said that this was in common with various other systems on the aircraft. BAC Weybridge, however, got on with any escape system tests it could manage without sight of an actual ejection seat, such as canopy jettison tests using windtunnel models. These tended to show that the navigator’s canopy would follow a low trajectory and hit the fin when jettisoned at the same time as the pilot’s canopy, so a delay between the two was introduced.
By the end of 1963, when the first flight had been originally expected, Martin-Baker had still not managed to test-fire a complete seat, though an earlier mark of seat using the Mk 8’s rockets had been test fired from the company’s Meteor. Unsurprisingly, the company had also been unable to deliver a functional example, and nor had blower-tunnel tests on canopy separation been made. Even the delivery of a structural test specimen seat to the RAE, scheduled for September 1963, was pushed back to February 1964. Two functioning seats were finally delivered to BAC at Boscombe Down on 30 June 1964, by which time the revised first-flight date of mid-May had come and gone, and still no test firings had yet been carried out, so they were not flight-cleared. The Ministry’s patience with Martin-Baker was fast running out, and BAC was similarly exasperated by the lack of visible progress, hoping that the RAF would be able to make more impact and impress upon Martin-Baker the urgency of the situation.
Incorporating several significant advances in ejection seat design, the Martin-Baker Mk 8VA seat was also well-liked by the TSR2 flight-test crews for its unusual level of comfort. Brooklands Museum
On 17 July Martin-Baker finally managed to fire a seat from one of its Meteors, and BAC and the A&AEE were able to start the initial blower-tunnel tests on various escape-system components on the 23rd, beginning with canopy jettison tests and moving on to a complete automatic ejection sequence including the firing of a seat. This programme continued throughout the summer, and the resulting report was not actually finished until a month after the first flight. The system was cleared as suitable for use on the development batch aircraft, and had performed well.
Sequencing of ejection was designed to enable the navigator to escape without the risk of being hit by the pilot or his canopy. Thus a pilot-initiated ejection would result in the navigator’s canopy separating, followed by ejection of the navigator’s seat, then the pilot’s canopy separating and the finally the pilot’s seat operating. The navigator could also eject solo, in circumstances where it would perhaps be a safer option than staying with an aircraft the pilot believed he could recover. Each member of the crew could also jettison both canopies without initiating an ejection, this facility being provided to aid escape on the ground, and envisaged as being possible in the air as a precautionary measure preceding a forced landing. In practice, the results of the blower-tunnel tests had found that if both canopies were jettisoned simultaneously they would collide with each other, but would do so behind the crew compartment and thus cause no danger to either occupant. The navigator’s canopy would then hit the fin, but as both crew members were on the way out this was not a problem. As a result, the delay between canopies was removed and a fully sequenced ejection would see both canopies fire first, then the navigator’s seat, then the pilot’s seat, all in quick succession. Below 75kt (85mph; 140km/h) a fully sequenced escape was not possible, and the canopies were to be jettisoned individually before each particular crew member ejected. Above 300kt (345mph; 555km/h) only a fully sequenced escape was permissible.
A preserved Mk 8VA ejection seat at the RAF Museum Cosford. Unlike most seats, the 8VA was left in an unpainted alloy finish. Damien Burke
Canopy jettison tests using a windtunnel model and multiple exposure photography to show the trajectories. These tests were useful to verify roughly the amount of power necessary to boost the canopies high enough to clear the tailplane when they were jettisoned. BAE Systems via Brooklands Museum
A TSR2 test forebody mounted in front of the blower tunnel at RAE Farnborough for canopy jettison and ejection sequencing tests. This rig could be mounted at various angles of incidence with differing blower speeds to represent particular configurations, such as slow and nose-up in the approach configuration. This forebody is now preserved at the Brooklands Museum at Weybridge. BAE Systems via Brooklands Museum
One of many canopy jettison tests carried out with the aid of the blower tunnel and test forebody rig. BAE Systems via Brooklands Museum
Larger-scale tests on canopy jettisoning at higher speeds (420kt (485mph; 780km/h)) were carried out at the Proof & Experimental Establishment site at Pendine in early March 1965. One of the RAE’s recoverable rocket sleds was modified to accept a TSR2 pilot’s canopy, constructed in the same manner as a real canopy but with metal window panels instead of transparencies. The first test, on 2 March, was a failure, as was the second, with the booster unit failing to fire and the canopy staying on the sled as a result. The third attempt worked, and the canopy followed a trajectory that would have taken it 12ft (3.6m)
clear of the fin had the jettison been from a complete aircraft. This was a smaller clearance than predicted by the blower-tunnel results, but was still acceptable.
Ejection Sequence
Time
Action
0 seconds
Ejection initiated
0.05 seconds
Navigator’s canopy jettisoned
0.05 seconds
Pilot’s canopy jettisoned
0.50 seconds
Navigator’s seat fires
0.77 seconds
Navigator’s seat rocket expended
0.95 seconds
Pilot’s seat fires
1.22 seconds
Pilot’s seat rocket expended
Testing crew tolerance to any buffeting from ‘cabriolet’ flying was to be undertaken both in a nose section in a windtunnel and also using a real aircraft. The fifth TSR2 was to undertake some flight trials without canopies, while the first would undertake canopy jettison and (navigator) ejection trials on fast taxy runs. When the programme was cancelled BAC advanced a good case for continuing with escape-system tests, as the problem of high-speed, low-level escape was going to be a recurring one, and blower-tunnel and rocket-sled trials therefore continued as scheduled. A report dated 20 April 1965 summarized the blower-tunnel and static (zero speed, zero height) live seat firing tests carried out so far, which were remarkably successful. In July 1965 a Mk 8VA seat was fired from a rocket sled at Pendine travelling at 247kt (284mph; 457km/h). This was mostly successful, though the correct parachute deployment could not be verified as an additional line attached to the seat for tests fouled the main ’chute before deployment. A successful zero-zero firing was carried out in May 1966.
The Mk 8VA seat was cancelled along with TSR2, but the advances made in its development and production found their way into the Martin-Baker Mk 10. A less capable version of Mk 10 seat, produced for the RAF’s Shorts Tucano turboprop trainer, was designated Mk 8LC, but bears no relation to the Mk 8VA. The number 8 was used purely because it was a ‘vacant’ mark number, the TSR2’s Mk 8VA having been long forgotten.
A rocket sled modified to represent a TSR2 nose section, on the Pendine test track for high-speed canopy jettison tests. BAE Systems via Brooklands Museum
Fuel system
The TSR2’s LP fuel system was required to maintain fuel flow to the engines in all flight regimes while also keeping the aircraft’s c.g. within strict limits. The fuel tanks were divided into two groups. The forward group comprised two tanks in the forward fuselage plus the port wing, and the aft group comprised two tanks in the rear fuselage plus the starboard wing. Each group fed a 30gal (136L) fuel collector box positioned so that it could be fed by gravity flow, with non-return valves to prevent fuel loss towards a holed compartment in any particular tank. Within the collector boxes were double-ended booster pumps driven by hydraulic pumps, which could be fed from top or bottom ends (thus catering for negative-g conditions) and pumped fuel from each collector into a single two-channel mechanical flow proportioner. Based on the pumps fitted to the Lightning, the booster pumps actually used the fuel tapped from the LP system as their working fluid, and so were known as fueldraulic pumps.
The purpose of the flow proportioner was to ensure that both engines were fed with fuel at the same rate, regardless of differences in supply from the two tank groups, thus also keeping the fuel tanks in balance. Fuel flow rates could vary from as little as 180gal/hr (818L/hr) to a massive 12,100gal/ hr (54,934L/hr) in full reheat. However, as it was expected that periods of full flow would be limited and both engines would be running at similar consumption rates during these periods, the proportioner would actually be bypassed whenever the flow rate exceeded 3,500gal/hr (15,890L/ hr). The crew could also elect to bypass the proportioner manually. If one fuel group ran dry while the proportioner bypass was in use, this would result in one engine being starved, so when running on low fuel or during landing, the pilot would open a crossfeed between the two engine fuel supply lines and so keep both engines running from the single remaining fuel group. A cavitation warning would also warn the pilot of insufficient fuel pressure at the pumps, with automatic cut-out of reheat in this situation to avoid the engine falling below 100 per cent power, as was found in similar circumstances on Lightnings. A reheat failure during take-off while retaining 100 per cent dry power was far preferable to intermittent reheat with less than 100 per cent dry power – the latter combination could result in serious engine damage.
The fuselage fuel tank layout. The fueldraulic system was similar to that of the Lightning, with a complex auto-balancing system to maintain the aircraft’s c.g. BAE Systems via Warton Heritage Group
The single ground refuelling point was in the aft end of the nose undercarriage bay; it is the blue cylinder at upper centre in this photograph. Damien Burke
The wing fuel tank layout. Almost the entire wing was an integral fuel tank. BAE Systems via Warton Heritage Group
The fuel system layout as of March 1962, before deletion of the buddy refuelling pack. Several serious fuel leaks occurred during flight testing, with double-walled pipes and junctions failing, mostly owing to over-pressure caused by component failings elsewhere. Luckily the failures occurred late in the flights in question, and no harm was caused. BAE Systems via Warton Heritage Group
Fuel was fed from the flow proportioner or bypass to the LP cocks on the engine and reheat systems, each of which were fitted with flowmeters and filters. Fuel destined for the engine was used as a heat sink to cool the hydraulic oil, auxiliary gearbox oil and engine oil, and was therefore fed through three heat exchangers en route. Under certain circumstances this could raise the temperature of the fuel beyond the permissible limits for the engine burners, so a recirculation system could feed fuel back into the fuselage tanks if necessary. Under normal sortie conditions and when using AVTUR fuel, this would introduce no limitations on engine use, but, if AVTAG fuel was used, certain minor limitations in range and performance would have to be expected due to that fuel’s lower boiling point and lower specific gravity. In particular, a period of cruising flight would result in high fuel temperatures and thus introduce speed limits (for example, low speed limit above 27,000ft (8,230m) in the worst case of a Mach 2 cruise at 40,000ft (12,000m) with proportioner bypass in use).
In-flight refuelling probe details. The aircraft’s generous combat radius specification meant that IFR was originally considered to be primarily of use in the ferry role, so the IFR probe was designed as an optional fit rather than being permanently carried. Damien Burke
Ground refuelling was via a single fuselage point located at the aft end of the nose gear bay, feeding into a refuelling gallery running along the fuselage with branches running into the wing tanks. Each tank had a top-level switch which would close the tank to further fuel inflow once it was full. The port forward fuselage was plumbed for a retractable in-flight refuelling probe fitted with a standard Mk 8 nozzle, which would be housed in a removable fairing below the port cockpit area. Both ground and in-flight refuelling was specified at a maximum rate of 500gal/hr (2,270L/hr) (later reduced slightly), enabling the aircraft’s tanks to be filled in around 12min. Both crew members would have controls to extend or retract the in-flight refuelling probe, but only the navigator could open the refuelling valve, as it was his duty to manage refuelling.
Defuelling was to be carried out via the refuelling point in the rear of the nose gear bay, with low-rate emergency (suction) defuelling or higher-rate defuelling carried out by opening all crossfeeds and non-return valves and using the fuel booster pumps to drive fuel into the fuel gallery and out of the refuelling point. Fuel jettison in flight was not considered necessary, as the undercarriage was strong enough to handle a landing at maximum AUW. However, as the engine problems arose during development the question of fuel jettison was raised by Roland Beamont, who pointed out that maintaining height on a single engine if one should fail with a full fuel load would require placing the remaining e
ngine into the danger zone of possible LP shaft failure. Accordingly a basic fuel-jettisoning system was incorporated into the development-batch airframes, mounted in the down-turned wingtips and making use of existing fuel vent piping. As BAC had been forced by circumstances into adding this system, the RAF took the opportunity to ask for it to be retained in production aircraft, ideally as a faster jettisoning system than already provided, but had the cost increase been unacceptable the Service would have settled for the basic system already in place.
The weight of remaining fuel within an aircraft naturally has an effect on the aircraft’s c.g., and keeping that c.g. within limits is extremely important if the aircraft is to remain controllable. A difference in c.g. position was necessary on the TSR2 for easy take-off, transonic cruise and when carrying underwing stores. Thus the TSR2’s fuel system included an automatic balancing component (an improvement over the manual systems provided in most aircraft), and was designed to feed fuel in such a way as to maintain a slightly forward c.g. Taken to extremes, if the rear fuel group were to be entirely emptied, the forward group would still contain approximately 2,500lb (1,135kg) of fuel. Any imbalance beyond specified limits would result in warning lights illuminating in the cockpit, and the pilot would then be able to rebalance the system manually by transferring fuel from one group to another.
One of the requirements of OR.343 was to provide a ‘range remaining’ readout for the navigator. To allow for this, not only was fuel contents gauging used, but flow metering was in place too, as both methods had inherent accuracy problems at low fuel levels and the combination could help to eliminate these errors.