TSR2
Page 21
‘TSR2 In Production’, the cover of a BAC brochure of July 1963, produced to illustrate the state of development and production of the aircraft in time for a visit from RAAF personnel. BAE Systems via Brooklands Museum
An engine tunnel under construction in the stage 1 jig at Preston. BAE Systems via Brooklands Museum
Edwards did not deny that BAC had substantially failed to keep its promises to improve slippages and keep control of costs, but he blamed it all on the ridiculous enforced amalgamation of two companies with separate cultures, procedures, staff and production locations. Furthermore, the main reasons for BAC’s failures to keep to production schedules were delays in construction of the wings and the rear fuselages; the work carried out by English Electric at its factories. In fact, one particular fuselage frame had been giving English Electric severe trouble, and they had to take inhouse the work on the tailplane spigots when a subcontractor could not handle the job satisfactorily. The company also had a culture of individual factories working to their own methods, rather than pulling together as part of a single organization. Friction between designers and production departments was all too common, frequent changes in drawings driving the production side up the wall.
Engine-tunnel skinning at Preston. With hundreds of close-tolerance holes to drill, fasten and seal, this was ‘a real watchmaker’s job’. Much of the engine tunnels’ areas were surrounded by fuel tanks, so the consequences of a hot gas leak (or, in the reverse direction, a fuel leak) would have been extremely serious. Much use of titanium was made in this area, and problems obtaining sufficient quantities of this exotic metal caused significant delays early in the build programme. BAE Systems via Warton Heritage Group
A rear fuselage nearly ready to be skinned at Preston. The rearmost frame has yet to have the tailplane and fin spigots installed; these forged items were again subject to initial supply problems. BAE Systems via Brooklands Museum
Pushed to introduce stronger management, Edwards agreed to several changes in the way BAC was running the project, and also pointed out that they were going to introduce Programme Evaluation and Review Technique (PERT) to the project as a whole in the very near future (as had been requested by the MoA months previously). By May 1963 the scene at Weybridge looked a bit more promising. Construction of the first aircraft was now well advanced, and in terms of structure it lacked only the wingtips, and the engines were about to be test-fitted. The next three airframes were nearly ready to have their front and rear fuselage halves joined together. The ‘Stage 3’ rig, simulating the aircraft’s electrical and electronic systems, was nearly complete as well. By August 1963 further delays had pushed back the first-flight date to January 1964 and an initial release to RAF service was now scheduled for October 1966. An engine for trial installation in the first aircraft had been delivered by BSEL, and a second was due shortly. The electrical distribution system proposed by BAC had come in for stiff criticism from the RAF’s Central Servicing Development Establishment (CSDE) and had been redesigned, and this delayed the installation of many systems in the first aircraft. On the second and third aircraft further delays were experienced because pressure transducers and accelerometers that were to be built into the airframe for flutter and load testing were failing at a stupendous rate during bench tests.
PERT project management
More or less coinciding with the beginning of airframe production in 1961 was the limited introduction of Programme Evaluation and Review Technique (PERT) project management, initially to cover design work only. Unlike traditional ad hoc management techniques, PERT enabled BAC to break up the design work into discrete tasks and show how they depended on each other, if at all, and thus schedule design work in the most efficient manner. Thus, hopefully, no time would be wasted carrying out certain tasks sequentially when they could be carried out in parallel, and the most efficient ‘critical path’ through the project would be revealed. It follows that PERT’s bias was towards time spent rather than cost saving, because lots of parallel tasks would inevitably require greater resources than a linear progression. Allied with PERT was the Project Execution Plan (PEP), which defined plans, procedures and control processes for project implementation and the monitoring and reporting of progress. This covered all sorts of project aspects such as personnel roles and responsibilities, risk analysis, cost management procedures, and quality assurance.
January 1964 came and went, with no TSR2 in the air. After a year under supposedly improved management the project had continued to slip. The company’s plan to continue final assembly of all TSR2 airframes at Weybridge, including the main production run, and transport them by road to Wisley for final equipment fit and first flight, had also become a problem. The airfield was only leased to BAC, being owned by the MoA. This lease ran out in 1968, and Surrey County Council was vehemently opposed to any actual use of the airfield for aviation (some things never change). It would be politically difficult to try to overrule Surrey’s wishes when there were more suitable alternatives, such as final assembly and first flight at Samlesbury. Consequently BAC was overruled and, from airframe 10 onwards, final assembly and first flight would be from Warton or Samlesbury. It seemed that English Electric, after five years’ work, would finally regain almost complete control of ‘its’ project. By now, however, English Electric and Vickers were no longer separate companies, and both names were increasingly disappearing from BAC paperwork. English Electric was now BAC (Preston), while Vickers was BAC (Weybridge). The loss of TSR2 final assembly and first flight to another division within BAC was no longer the unthinkable insult it would have been just a couple of years before.
With BAC once again being hauled over the coals by a customer who was now incandescent with rage about the delays and cost overruns, and threatening cancellation, George Edwards took a firm hand and introduced a slew of management changes. They included taking Freddie Page off the Lightning project and putting him in overall day-to-day control of the TSR2, while Ray Creasey took over Page’s role on the Lightning. Thenceforth the atmosphere changed radically, and the Ministry began to see results from the ‘new head of steam’ that had been generated by this and various other new appointments in senior BAC positions. In February 1964 the adoption of Value Engineering within the TSR2 programme indicated that the winds of change really were blowing within the company. This covered the analysis of all engineering solutions on the project to see if they could be reduced in cost while still retaining all of their functionality. For instance, while integral machining of a part produced the lightest and strongest components, traditional die forging and casting could produce a part of the same strength at reduced cost. The trade-off was that the forged or cast part would be heavier. Value Engineering would kick off a series of cost reductions on the TSR2, carefully selected for the largest savings and minimum adverse impacts.
Rear-fuselage production at Preston in full swing. The section nearest the camera has been fully skinned and is undergoing equipment installation. BAE Systems via Warton Heritage Group
X2020 aluminium-lithium alloy
In the late 1950s/early 1960s the future of aluminium for aerospace use was uncertain, and one of the various developments that took place was the creation of aluminium– lithium alloys. These were less dense, and therefore lighter, than aluminium, and also exhibited greater stiffness. They also cost three times as much as standard aluminium. Early tests showed that the material was reliable under a range of stress conditions including sustained use in a salt-water environment and at high temperatures. Unfortunately the tests did not immediately show up the material’s weakness under high impact loads; BAC finally found out late in 1962 that X2020, as well as proving to be much more difficult to form than more traditional alloys, fractured badly when large point stresses were applied. Its much-vaunted stiffness collapsed entirely if a complete penetration of a sheet of the material occurred. The reality, therefore, was that the TSR2 was more likely to suffer critical damage from a missile hit or cannon fire than
an aircraft of more conventional construction. In practice, the problem began to show up in production, and even in day-to-day maintenance of the first aircraft. It transpired that the ‘large’ point stresses BAC had been warned about did not need to be that large. Closing rivets on small panels was sometimes producing cracks radiating from the rivet location, and dropping a panel could result in cracks at corners. The company’s Guided Weapons Division was experiencing so many fabrication problems with X2020 by July 1963 that it wrote to BAC Weybridge, warning of these experiences and saying it had heard that the use of X2020 had been substantially restricted or discontinued on certain aircraft projects in the USA.
Contract KD/2L/16/CB.42(a) April 1964 Production Aircraft (Aircraft 21-50)
Aircraft
Serial No.
Aircraft
Serial No.
Aircraft
Serial No.
21
XS944
31
XS954
41
XS986
22
XS945
32
XS977
42
XS987
23
XS946
33
XS978
43
XS988
24
XS947
34
XS979
44
XS989
25
XS948
35
XS980
45
XS990
26
XS949
36
XS981
46
XS991
27
XS950
37
XS982
47
XS992
28
XS951
38
XS983
48
XS993
29
XS952
39
XS984
49
XS994
30
XS953
40
XS985
50
XS995
The fuselage of XR219 being transported south on Britain’s first stretch of motorway, the Preston Bypass, otherwise known as the M6. This photo was taken from the Cuerdale Lane bridge over the motorway, south of the A59 junction (modern-day junction 31). The view, and traffic level, are somewhat different today! BAE Systems via Brooklands Museum
This was a massive blow to a project already reeling from cost, weight and schedule overruns. While nothing much could be done about the development batch aircraft, which were already well advanced in production, a crash programme was instigated in early 1964 to investigate where X2020 could be replaced by L71 aluminium alloy on further airframes. This was a large task, as more than 16,000 parts had been designed to use X2020 in the components built by BAC Weybridge alone. Where the stiffness requirements were not so critical this could be done on a like-for-like basis, using L71 of the same gauge as previously specified for X2020. However, in more critical areas, particularly external stressed-skin panels, in-depth studies would be needed to determine what increase in gauge would be required when using L71, or indeed if L71 would do at all. By this time the cost of X2020 as a raw material was more than five times as much as L71, and fabrication costs were 10 per cent higher, so initial estimates suggested that more than £5,000 could be saved on each aircraft for a weight increase of just 40lb (18kg).
This problem of airframe strength manifested itself in a dramatic and wholly unwelcome manner when the wing on the static strength-test specimen suffered an early failure on 28 January 1965, while undergoing heavy static load testing. The wing was constructed of many integrally stiffened panels milled from solid billets of alloy, held together with small titanium bolts or rivets and liberally smeared with PRC sealant, as the wing itself was to be a large fuel tank. The underside of the wing cracked near the line where it met the fuselage, just behind the apex, and the wingtip was also damaged, though this was thought to be induced by mechanical disturbance from the first crack. The first crack began near one of the bolt holes, and a titanium rivet also fractured. The wing had been designed to cope with a maximum loading of up to 10g, with a limit of 6g to be applied in service. The failure occurred at 8.5g, or 85 per cent of the ultimate design strength. This was duly reported to the MoA, and up the line to the Cabinet, and was gleefully pounced upon by the project’s many critics. Protests by BAC that the break had occurred at well beyond the loading to be expected in service, and could probably be dealt with by local reinforcement and slightly thicker rivets rather than major changes, were pretty much ignored, and the news was used by many to criticize the project once more.
The BAC investigation into the cause of the early wing failure was never fully completed owing to the cancellation of the project, but BAC’s Mechanical Test Department produced a report in June 1965 in which it was admitted that, although results were incomplete, there was ‘no definite indication why the wing should have failed’. The investigation had found, however, that X2020’s superior strength characteristics could be reduced considerably by ‘minor defects such as typical scratches and screwdriver marks which commonly occur in aircraft manufacture’. Stiffeners added to the wingtip, which was skinned with X2020, did not help the strength of the joint of the main wing box to the wingtip, and skinning the tip with L71 did not give satisfactory strength either.
Had the project proceeded, it is certain that the wide-scale replacement of X2020 by L71 would have continued and been expanded further, introducing additional delays allied to the redesign plus weight and cost implications (though at least raw material costs would have decreased). In the end X2020 was an expensive mistake, and it is notable that the Concorde airliner project, another Mach 2 design, did not use it.
The fin of XR219 nears completion at Accrington in March 1963. The large hole for the spigot is clearly visible. BAE Systems via Brooklands Museum
Skinned, XR219’s tailplanes undergo final assembly at Accrington in March 1963. Other tailplane skins are visible in the background. BAE Systems via Brooklands Museum
All change
While most structural parts of the airframe were by now in production, with early examples undergoing strength and fatigue testing, there was one odd one out; the rear fairing. This covered the rear section of the engine tunnels, surrounding the jetpipes and incorporating the braking parachute housing and door, being manufactured by BSEL. Unfortunately BSEL had proved to be entirely out of its depth when it came to building part of an airframe rather than an engine, and stressing, strength and construction problems had dragged on for months. Eventually, exasperated by the cost and delays, BAC undertook to redesign the rear fairing itself so that it could manufacture it in-house for production aircraft. Having put BSEL’s work under the microscope, however, BAC could hardly afford to ignore its own designs for other areas of the aircraft, and the Value Engineering department found that BAC was as guilty of overengineering as any subcontractor.
Stage 3 fuselage assembly work at Weybridge in March 1963. Nearest the camera is XR219, XR220 is on the far left, and XR221 is between them. BAE Systems via Brooklands Museum
The wing of XR219 on delivery to Weybridge at the end of March 1963. The entire assembly was an integral fuel tank, with port/starboard halves and a centreline bulkhead with balancing passages. However, XR219 did not have an operational wing fuel system, and was therefore limited to fuselage tanks only. BAE Systems via Brooklands Museum
For example, most of the wing spars, which were of ‘egg box’ construction machined from the solid, could be made just as easily from corrugated plate webs, using chemical etching to save weight where needed. Much less material would be wasted, and the result would be just as strong. Fuselage stringers with varying angles over their length were also machined from solid billets at considerable expense, and could be replaced by stretched angles mounted backto-back. The flap-blowing sy
stem varied output with flap angle, yet performance would not be affected if the maximum output was simply used at all times the flaps were down, regardless of angle. Thus an on/ off valve could replace a complicated regulator. Airbrakes, intake cones and various other systems were also needlessly complex and could be replaced by simpler, cheaper alternatives. An overall review of possible cost savings resulted in the ‘Ewans Report’ of June 1964, which detailed all of the major savings that could possibly be made and concluded that the unit cost of each TSR2 could be reduced by 10 per cent, but the weight gain would be in the order of 1,500lb (680kg). For the 1,000nm sortie this would cause a reduction in combat radius of 16.5nm (18.9 miles; 30.5km), and an increase in the dispersed take-off roll of 28.5yd (26m).
Ministry response to this first Value Engineering report was typically shortsighted, concern being expressed that it actually raised development costs over the short term. Saving money on production was apparently of little interest to the bureaucracy if it meant they had to fork out a higher figure for R&D, despite the total cost being lower! The RAF, however, read the report with shock. All along, the MoA had been saying it did not have the resources to keep a close enough eye on the project to keep costs down. Now one man and his tiny team at BAC had come up with umpteen ways of reducing costs and shown that, hitherto, almost no efforts had actually been made to keep them down. The RAF had finally lost almost all faith in BAC’s management of the project, and also in the ability of the MoA to do the right thing; it was clear that it was going to obstruct any cost-saving effort that added to the more immediate R&D bill.
First flight
By the beginning of March 1964 the first aircraft, XR219, had finally been completed. There was no formal roll-out; there was no time to waste on that sort of publicity exercise, and no appetite for it, particularly with the security paranoia then endemic in both the government and the RAF. The aircraft was immediately dismantled again for transport to Boscombe Down, and its arrival there on 4 March heralded an even more stressful period for everybody on the project, as reassembly took longer than expected, and the already badly delayed taxying tests were deferred repeatedly owing to problems on the engine development side. The RAF was keen to see the aircraft finally in the air, but also made it clear that it did not want to sacrifice flight safety purely to meet a ‘political’ date for the first flight. It was concerned that BAC’s haste to get the TSR2 airborne could result in a loss of the first aircraft, and mentioned the ‘fact’ that BAC had now lost a pair of BAC One-Eleven airliners in accidents ‘due at least in part to too much haste’.