TSR2

by Damien Burke

  The TFR system components. The initial requirement had been for terrain clearance, a simpler and easier concept than terrain following. BAE Systems via Brooklands Museum

  Ferranti’s Dakota, TS423, with its AI.23 radar installed in an extended nose. This aircraft carried out some of the early FLR trials using a modified AI.23B set, with little success. BAE Systems via Glenn Surtees

  The Ferranti FLR was a relatively compact unit, and is seen here with most of its exterior removed. BAE Systems via Brooklands Museum

  Early trials with a modified AI.23B mounted in the nose of Douglas Dakota TS423 were a dismal failure, primarily because the trials had been begun prematurely, before the problems of separating terrain from general ground clutter had been solved. One of the Buccaneer development aircraft, XK487, was earmarked as a more suitable testbed for the completed FLR, even though the aircraft was being pushed as an alternative to the TSR2. The Admiralty, unsurprisingly, proved less than cooperative, and negotiations for the use of the aircraft and associated spares backup delayed matters. Blackburn also dragged its heels, stating at one point that it had not expected Ferranti to progress so quickly with the FLR development, and thus had not expected XK487 to be needed so soon.

  Progressing quickly was not something of which Ferranti was often accused. Delivery of Blue Parrot for the NA.39 programme had already slipped, and progress with the TSR2 FLR was looking so poor in mid-1961 that serious thought was given to cancelling the Ferranti FLR entirely. The Air Staff had its feathers seriously ruffled by the Admiralty also asking Ferranti to conduct a study into fitting a terrain-following version of Blue Parrot to the NA.39. The Air Staff had been told on numerous occasions that Ferranti was ‘up to its eyebrows’ with work already, hence the lack of progress, and was seriously concerned that any additional loading coming from the Admiralty was going to have a negative impact on the TSR2 programme. The deeper objection was really that the Admiralty appeared to be working on the basis that it could gradually upgrade the NA.39 to the point where it was a serious threat to the TSR2, with RAF officers reduced to squabbling about whether the NA.39 should be allowed to attack inland targets as well as targets ashore!

  An MoA delegation then visited the USA to examine an off-the-shelf alternative, the Texas Instruments APN.149. This was J-band radar, already test-flown in a Martin B-26, with proven terrain-following capability, radar ranging and ground mapping and fixing. An RAF pilot had flown in the B-26 down to altitudes as low as 90ft (27m) with the TFR functioning well. The lack of tanker and airfield homing could be dealt with by the installation of an air-toair tactical air navigation and radio compass (already fitted to the first two aircraft to aid navigation). The lack of weapon-aiming capability was a more serious omission, and would require the creation of a weaponaiming computer to be linked to the APN.149. The ramifications of replacing the expected Ferranti set with the Texas Instruments alternative were studied by BAC, and it was found that while there would be a delay of two months in the flight of the development aircraft, which would cascade down to the in-service date of production aircraft, the guaranteed delivery times from Texas Instruments would enable significantly more testing of the TFR system. Additionally, Texas Instruments was offering a guaranteed mean time between failures (MTBF) of 140hr (compared with a vague promise of 50hr from Ferranti), fixed prices, and discounts if it slipped on the promised production schedule. It was an attractive offer; the saving per production aircraft was expected to be £25,000.

  The FLR installation and associated controls. BAE Systems via Brooklands Museum

  The navigator’s FLR display and control panel. BAE Systems via Brooklands Museum

  The Buccaneer selected for FLR trials, XK487, at Brough in late 1962 with the newly fitted nose section, awaiting the addition of a TSR2 radome and a new paint job. BAE Systems via Brooklands Museum

  Ferranti was furious. Its FLR was more capable in turning flight (being stabilized to a 45-degree angle of bank compared with the APN.149’s 30 degrees), integrated the functions of airfield and tanker homing, and the company believed it had ample room for further development, whereas the APN.149 had none. A new radome would need to be developed for the APN.149, not a minor job, and the TFR capability of the APN.149 was only proven at far lower speeds than specified for the TSR2. The 1/4g limit on the APN.149 would result in undershoots at high speeds, or, to put it more bluntly, flight into terrain, though Texas Instruments claimed this particular limitation would be easy to remove. Trickier to deal with was the frequency at which the APN.149 operated, as J-Band radar could be confused by rain. On one occasion during a B-26 test flight the TFR commanded the aircraft to climb over a rain cloud as if it were an obstruction. The X-Band Ferranti unit had no such weaknesses. The Prime Minister (Harold Macmillan), on finding out about the proposal to use a US radar unit, wrote to the Minister of Defence, Peter Thorneycroft, to express his own unease: ‘I must say I find it exceedingly disquieting. The Government is very firmly committed to produce the T.S.R.II as a viable weapons system in the mid-1960s. It would seem very sad if this could only be done at the price of having to purchase American equipment for it’. Elliotts had clearly done well to hide the American origin of the TSR2’s central computing system!

  By April 1962 Ferranti was offering more guarantees, such as 60hr MBTF, and was becoming much more receptive to the idea of a fixed price and delivery date. Meanwhile, airborne trials on Canberra B(I).8(mod) WT327 had begun on 27 February 1962, with extremely promising results. The modified AI.23B set on this aircraft was linked to an early model of the TSR2’s terrain-following computer, and displayed pitch commands to the pilot on a small display unit. The pilots soon trusted the system enough to enter cloud while in terrain-following flight, losing sight of the ground entirely and continuing to follow the pitch commands given by the system. Ferranti won the argument; the Minister of Defence decided on 2 May that there was no alternative but to proceed with the Ferranti FLR, which, despite being the more expensive option, would be more capable. The FLR lived on, and Texas Instruments went away empty-handed.

  By 1964 Ferranti had rewarded the faith placed in it, and had regained much of the previous schedule slippage. Several FLR test models were in use, and flight trials on both the Canberra and Buccaneer were generally going well, with some particularly impressive results in the later Buccaneer flights, using a much more representative model of the FLR. Unfortunately, the ground-mapping mode was not quite so impressive, and some serious work was going to be needed to improve this aspect.

  In physical terms the TSR2 FLR had a scanner head and transmitter/receiver unit that was basically similar to the existing AI.23B, but the various ancillary units were different, making extensive use of transistorization. The entire assembly was supported by a surrounding mounting ring. Attached to this ring at the front was the aircraft’s radome, and at the rear the FLR canister, shaped rather like a top hat. This meant that the unit was fully enclosed in a pressurized and sealed container to keep the sensitive electronics clean of dust and other contaminants, and the entire thing could be removed from the aircraft, radome and all. The Air Staff objected to this, wanting the ability to remove the radome for servicing or replacement without disturbing the FLR, so an internal dome was introduced to enclose the scanner area. This dome later came up as an item that could possibly be dropped to save costs, the FLR’s components having proved to be more reliable and less sensitive to environmental contaminants than expected.

  Ferranti’s Canberra, WT327, flies over Turnhouse Airport. This aircraft carried out much of the early FLR airborne trials before the Buccaneer testbed was available. BAE Systems via Glenn Surtees

  At the time of cancellation, the only flying TSR2, XR219, had never been fitted with the FLR, or indeed AFCS, so the real performance of the TSR2’s TFR system was never able to be demonstrated. However, the most recent flight trials at the time had suffered from several spurious pitch-up demands, and clearance height over peaks had been dangerously low on
several occasions, so it was clear that some months of further work were needed before a fully automatic system could be safely demonstrated.

  Head-up display

  The TSR2 was to be one of the first aircraft to incorporate a HUD, that is, a display superimposing various flight data and flight director cues over the pilot’s view of the outside world. This was to use a CRT display (the Precision Display Unit) buried within the cockpit coaming, focused at infinity via a 6.5in lens and reflecting off the aircraft’s windscreen. Information to be presented was to include a flight director to lead the pilot through manoeuvres as required (including terrain following and bombing) plus readouts of airspeed, altitude and heading.

  Early work on such a display had taken place at the RAE, which had also come up with suitable symbology to use, and after early tests of very basic HUDs on Meteor WL375 and Javelin XA831 the RAE commissioned Hawker to produce a one-off Hunter airframe to use in HUD development trials. This was XE531, the one and only Hunter T.12, which as well as provision for a HUD also included a nose bulge to house a new camera (the FX.126, also destined for TSR2 use). Among other changes, XE531 had been given a master reference gyro, an air data computer and an ILS (all to be linked to the HUD), along with a small joystick to be used by the second pilot to control the Flight Director. Thus the second pilot could ‘fly’ the aircraft in simulated terrain-following by using his side-stick to control the demanded flight path displayed on the HUD, hoping that the other pilot was following the director symbology correctly. This demonstrator aircraft was flying regularly throughout 1964 and proved quickly that the whole HUD concept was excellent.

  Derek Whitehead, Blackburn’s Chief Test Pilot, hands over XK487 to John Field, chief pilot of the Ferrranti Flying Unit. After months of delay, more demanding airborne trials could finally begin. BAE Systems via Glenn Surtees

  The RAE’s Hunter T.12, XE531, was used for HUD trials. BAE Systems via Brooklands Museum

  The contract for TSR2 HUD development went to Rank-Cintel, and by 1963 an early model was ready for testing at Weybridge. While it worked well given a suitable reflecting surface, the TSR2’s windscreen was anything but that. With multiple layers and an overall thickness decided by the birdstrike resistance requirement, it gave multiple ghost images, and early models also suffered from highly distracting cyclic variations in brightness owing to tiny voltage variations. A demonstration of a supposedly improved unit in August 1963 received a slew of complaints about the ghosting, and resulted in an attempt to improve matters by the addition of a slightly reflective layer on the windscreen’s inner face. A portable cockpit mock-up was built to include an elliptical reflective patch on the windscreen, and this was positioned on a hill over looking the Weybridge site so that typical viewing angles at low level with varied lighting levels could be experienced. The result was a significant improvement, but pilots who tried the mock-up complained that the obviously different elliptical area was distracting, and the reduction in light transmission through the wind screen might cause target identification problems. An alternative application, covering the entire windscreen, was also demonstrated, but this introduced additional distractions from reflections of other cockpit objects. The Air Staff decision was to apply the reflective coating on a band across the middle covering the required area, and to fade it out top and bottom to reduce the distraction factor. This, naturally, would be the most expensive option.

  By late 1964 the latest laboratory models were performing rather better than earlier models, and reliability of components was being improved with soak testing. However, there were constant complaints about wind screen quality from the pilots carrying out TSR2 taxy runs and test flights, so there was a very good chance that the final HUD would have needed to use a separate reflector plate, as used in all practical installations after the TSR2.

  Head-Up Display installation details. BAE Systems via Brooklands Museum

  The HUD installation in the mockup cockpit at Weybridge. Restricted viewing angles and the poor reflective qualities of the TSR2 windscreen was a continual problem, and the final HUD would probably have needed substantial redesign. BAE Systems

  Reconnaissance pack

  The TSR2 was designed to carry out reconnaissance as a dedicated role and also in addition to its more usual strike role. In dedicated reconnaissance fit the aircraft was to obtain information for tactical purposes, including target mapping at low altitudes and in all weather conditions, by day or by night using radar and/or photographic methods. In strike fit the aircraft would still be required to carry out the maximum photographic and/or radar reconnaissance without this adversely affecting its strike capabilities. The primary purpose of the TSR2’s reconnaissance information was to support both the counter-air battle and the land battle. For the latter task in particular it was essential that information could be got back to interpreters as soon as possible, so that it could be studied and effective use made of it before the information became stale. The defences to be expected on the European battlefield further complicated the task, as extreme low-level flying and transonic speeds would be necessary to allow anti-air units as little time to respond as possible. For coverage of wider areas, medium-level sorties would be required, and here supersonic speeds would be necessary.

  The TSR2’s reconnaissance system was made up from permanently fitted items, such as the strike cameras and SLR, and a dedicated reconnaissance pack. BAE Systems via Brooklands Museum

  General arrangement of the reconnaissance pack. Damien Burke

  In dedicated reconnaissance fit the aircraft would have the navigator’s bombing control panel replaced by a reconnaissance control panel, and a large reconnaissance pack would be mounted in semi-conformal style below the fuselage. The bomb-bay doors would be removed and fairings fitted to the forward and aft ends of the reconnaissance pack, which would hang from the bomb-bay roof racks. The pack was self-contained apart from relying on the aircraft for control, power and air conditioning supplies. Early windtunnel tests showed that the pack added no detectable drag, and therefore performance was expected to be unaffected. There was some consideration given to flight testing the pack by fitting it to a Canberra B(I).8 bomb bay – the fit was a surprisingly good match once the forward fairing was removed, though the aircraft would have needed an enlarged tail bumper to ensure ground clearance was sufficient with the tail down. However, the position of the Canberra’s engine nacelles was expected to interfere with the SLR, so this was not proceeded with. A mock-up pack was fitted to XR225 at Weybridge in February 1965, and some minor fouls were experienced in matching the pack to the airframe which necessitated both pack and airframe modifications. These were to be incorporated on airframes from XS665 onwards.

  The reconnaissance pack embodied three separate systems: optical linescan, SLR and standard photographic cameras.

  Optical linescan (ARI.23132)

  The linescan fit was from EMI, and consisted of a high-speed rotating scanner that picked up variations in brightness as it swept the ground at right angles to the aircraft’s track. Unlike traditional cameras, linescan was not limited to the field of view of a fixed lens, as the scanner would rotate and thus the field of view was theoretically a full 360 degrees. In practice, from horizon to horizon was the area of interest, and the physical make-up of the reconnaissance pack meant that the actual scan area was limited to a 144-degree sweep.

  Passive linescan was to be used in daylight, and active (illuminated) linescan would be used at night. Optimized for best performance while flying at 500ft (150m) above ground level, in active mode an extremely narrow beam of light would be projected in concert with the rotation of the scanner unit. The beam’s narrow width, 5 milli-radians, and the high-speed rotation of the scanner (12,000rpm) would combine to make the illumination practically invisible to the naked eye, and thus would not endanger the aircraft, unlike the traditional means of illuminating the ground for photography, which entailed the ejection of parachute flares and the attraction of lots of u
ndesirable attention. While daylight linescan coverage would not match the quality of traditional photographic results, in low light, before dawn and during twilight, the light-sensitivity of the photocells at the heart of the unit would be superior to the exposure characteristics of photographic film and thus produce more usable results.

  An experimental linescan unit on its servicing trolley. The TSR2’s linescan worked in purely visual wavelengths, as opposed to the IR linescan units that became more popular later. BAE Systems

  The scanner unit consisted of a motor rotating a pair of mirror drums. In active mode one of these mirror prisms would reflect a powerful Philips CS200 arc light source at the ground. The other mirror was the receiving mirror, collecting light entering the unit and directing it to the scanner for amplification and processing. The scanner unit itself was sealed (with a window to let light in and out, of course), and the drums operated in a vacuum environment to ease the load on the motor. The belly of the reconnaissance pack itself had a sliding shutter covering another window, through which the scanner operated. When the scanner was not being used, the shutter would be closed to protect the window glass from erosion or foreign-object damage.

  A linescan image of Tewkesbury, Gloucestershire, with the River Severn on the left, crossed by the A38 and the long-since-vanished railway line. The resolution of the original image is sufficient to distinguish cars from lorries, though distortion towards the horizons renders the edges of the image of far less use. BAE Systems via Brooklands Museum

  Originally it was intended that the output be recorded on photographic film with rapid onboard processing and enough film for 50nm (57 miles; 90km) of coverage (passive) or 150nm (170 miles; 270km) (active), but this was changed so that the varying electrical signals from the linescan sensor were recorded on magnetic tape instead. This enabled the linescan record to be buffered onboard the aircraft, so that delayed transmission of linescan data could be made in circumstances in which live transmission was not possible. Coverage width would vary from 0.2nm at 200ft (60m) to 1nm at 1,000ft (300m), increasing above this height, but with resolution increasingly suffering to the point that it would no longer be able to identify a three-ton truck, as required by the specification. Signal processing included automatic contrast adjustment to deal with hazy conditions and automatic gain control (AGC) to deal with extreme variations in brightness on what was a rather low level of signal produced by the scanner. The processed signal was then combined with digital data of the aircraft’s flight parameters before being recorded on the videotape recorder unit. The data track on the tape could briefly be replaced by an event marker under the control of the aircraft’s navigator, so any particularly noteworthy location could be highlighted for the interpreters looking through the tape afterwards.