| John Farley first flew the P.1127 in 1964
while a test pilot at the Royal Aircraft Establishment. He spent 19 years
contributing to the development of the Harrier, retiring as Chief Test Pilot
BAe Dunsfold. He then spent five years as Manager of Dunsfold and a further
two as Special Operations Manager at BAe Kingston. In 1990 he became the
first Western test pilot to fly the MiG-29 fighter. He is currently part
of the Farnborough
Aircraft team developing the the F1 air taxi.
This lecture was delivered to the Munich
Branch of the Royal Aeronautical Society in May 2000. It is reproduced
here by kind permission of John Farley.
The Harrier Development Story - by John Farley OBE AFC CEng
From the Hawker P1127 in 1960, the Harrier family
of single engined vectored thrust military aircraft has developed
to the point where, today, the aircraft is in service with six nations
and seven air arms. Over thirty distinct versions have seen service
during this 40-year period. The Harrier has allowed the Royal Air
Force to land swept wing jet fighters vertically for 31 of
its 82 years. Over a third of its history. Such
freedom from the problems inherent in high-speed landings of tactically
deployed aircraft represents a major achievement. I believe this
long-term success has been due to three things:
- The basic V/STOL concept employed was innovative
and had no flaws.
- The V/STOL elements have been continually improved
by sound development.
- The operational equipment fit has kept up to
date and has often led the way.
I will take a few highlights from this last point
first, because Harriers are built to fight, and not just to do V/STOL.
Operational military competitiveness has been vital for the RAF,
Royal Navy and USMC units who have fought wars with their Harriers.
When the RAF formed its first Harrier GR1 unit in
April 1969 (fig 1), it was their first aircraft with:
- A self contained inertial navigation system (INS).
- A moving map display that showed the present
position of the aircraft.
- A head up display (HUD) of flight, navigation
and weapon aiming information.
- A totally self contained starter running off
the main engine fuel system (GTS).
Fig 1 Harrier GR.1
Fig. 2 Harrier GR.3
the mid seventies the RAF Harrier fleet was updated to GR3 standard
(fig 2) which included a new nose having a laser range-finder and
marked target seeker that greatly helped single pass attack capability
Jumping ahead to today, the Royal Navy single seat
Sea Harrier FA2 (fig 3) was the first European aircraft capable
of using the beyond visual range US AMRAAM missile. This capability
at once made it the best interceptor with any European air arm.
Fig 3 Sea Harrier F\A.2
Fig. 4 Harrier GR.7s and T.10
|Today’s RAF single
seat GR7 / T10 fleet (fig. 4) is operational at low level at night
in the close air support role using the latest forward looking infra
red HUD and night vision goggle equipment. This is a role that many
said was impossible only a few years ago. Such military headlines
are a topic in themselves and form a very important part of the Harrier
story. But for this paper I will concentrate on the V/STOL technology
incorporated in these aircraft.
The vectored thrust story
The vectored thrust story started in the mid 1950s
when a Frenchman, Michel Wibault, proposed a single seat fighter
that he called the Gyroptere. Wibault
proposed to vector the thrust of four separate centrifugal blowers
driven by a single 8000 HP Bristol Orion engine (fig 5).
He chose the Bristol engine because it was then
the most powerful turbo-shaft engine in prospect. He was unable
to interest the French authorities with this idea, but in 1956 he
left a brochure with the US officer running the Paris office of
the Mutual Weapons Development Programme Colonel Chapman USAF.
Fig 5 Origins of thrust vectoring
Chapman was working on the Orpheus engine with Bristol
Aero-Engines' Technical Director Stanley Hooker, so he naturally
turned to Hooker for advice on Wibault's brochure. At Chapman's
request, Hooker directed a study on Wibault's idea that led to an
all new Bristol engine, the Pegasus, having four rotating nozzles.
By 1959 the Pegasus 1 was running on a Bristol test-bed and Hawkers
were making the P1127 airframe.
The starting point
Fig. 6 Powerplant installation
October 1960 Hawkers were ready to fly the first P1127. It was decided
to look at the unknowns of hovering before flying the aircraft conventionally.
The engine, by now the Pegasus 2 of 11,000lb
thrust, was installed as shown in fig.6. The main characteristics
were a large bifurcated intake and four swivelling exhaust nozzles
interconnected together which enabled the thrust vector to be moved,
as required, from aft to about 18 degrees forward of the vertical.
The LP fan (fig.7) had two stages and was fitted
with inlet guide vanes. It supplied the front nozzle efflux and
the eight stage HP compressor. A most important feature was that
the HP and LP components of the engine contra rotated to remove
A motor under the engine, (fig.8) driven by compressed
air from the final HP stage, was used to rotate the four exhaust
nozzles through a simple and reliable mechanical system of shaft
and chain drives.
This air motor was signalled mechanically by a slim
lever mounted alongside the outboard throttle (fig.9). The throttle
was a conventional fighter type, push forward for more thrust. The
sense of the nozzle lever was similar - push forward for more speed.
So lever forward selected the nozzles aft and lever back rotated
Fig 7 Engine gas flow
Fig. 8 Nozzle actuation system
Fig. 9 Nozzle selector lever
Fig. 10 Initial reaction control system
Attitude control in the hover was
by a jet reaction control system (fig.10). With the nozzles deflected,
high-pressure air was continually bled from the engine at the rate
of 12lb/sec to supply roll and pitch nozzles at the nose, tail and
Pitch control was obtained by increasing the aperture
of one of the pitch valves and reducing the size of the other -
the total lift force on the aircraft remaining constant. Roll control
used differential control of the roll jets. Yaw control came from
swinging the pitch jets laterally. The control valves were operated
by the stick and rudder pedals in the conventional sense. A 20%
authority autostabiliser was fitted, giving artificial damping and
stiffness in pitch and roll. The ailerons and tailplane were powered
the rudder was not.
The airframe was clearly representative of a typical
fighter aircraft (fig.11). It had been designed as such from the
start and not as a pure research vehicle. The undercarriage, however,
was unusual. The layout was determined by the presence of the four
engine exhausts under the high wing, especially the two hot rear
nozzles. It was considered that although a wide track tricycle could
have been employed it would have incurred a greater structural weight
penalty than the bicycle with the outriggers. You are warned that
it took from 1960 to 1967 to completely eliminate the adverse handling
and performance of this undercarriage during take-off and landing.
Fig 11 P1127
|For the initial hovering trials the aircraft was
fitted with tethers and was operated from a 40 x 50 ft grid platform
fitted flush with the ground surface above a pit used to duct away
exhaust gases. The tethers had a certain nominal length. When this
was used up an ever-increasing number of metal disc weights was lifted
up, so smoothly increasing the restraining force on the aircraft.
The first lift was carried out with 1 ft. tethers and the aircraft
was restrained by the tethers most of the time. For the next lift
the tethers were increased to 4 ft.
Fig. 12 P1127 on grid platform
Immediately the first handling problem due to the
undercarriage came to light. As the aircraft
lifted, the main undercarriage leg extended more than the outriggers
until lateral freedom was available, (fig.12), then the aircraft
banked and the resulting horizontal component of lift force moved
the aircraft sideways and swung it round in yaw, about either the
nose or main wheel.
Clearly this loss of control in roll and yaw could
have been fixed with greater control powers, but bleed air was at
a premium as its provision reduced engine performance. So it was
essential to use the absolute minimum possible. This is a basic
fact of life with jet V/STOL and must be continually borne in mind.
Additionally, the horizontal component of a slightly inclined vertical
force gets significant very quickly if that force is equal to the
weight of the aircraft. Only six degrees of tilt is needed to achieve
0.1g horizontal acceleration for example.
After a second attempt with 4 ft tethers, during
which the tethers broke a few yards from where the design executives
were standing, the tethers were rapidly doubled in strength, halved
in length and the outriggers were propped up on platforms. The thing
could then only tilt a little, but it was still helpless after the
first landing since it then sat wing low again and could not be
levelled by the pilot.
Temporary extensions were therefore fixed to the
outriggers so that they supported the aircraft level until just
after lift-off. This action greatly improved the handling and two
successful hovers, one with 2 ft and one with 4 ft tethers, followed.
A second pilot then joined the programme. Investigation
of autostabiliser behaviour at this point showed that it was saturated
most of the time so some lifts were made with the autostabilisers
off, the results being rather inconclusive within the tethers.
Initial taxying trials were then started. Directional
control was by nosewheel steering and only the main wheels had brakes.
The undercarriage, with by now well established
tradition, fought back hard when taxying, this time through the
medium of the nosewheel steering mechanism. This proved to have
a large dead-band around centre and was harsh in operation. Once
the dead-band was overcome, the sensitivity was then far too great.
Thus a generally unsatisfactory weaving progress was made between
A and B.
Following some small changes to the autostabiliser,
intended to correct a drift in the stiffness term, the aircraft
was returned to hovering. The final hovers in this series included
a tethered lift over a solid surface, some more tethered lifts over
the grid and then finally two free lifts over the grid. This programme
took five weeks and included 21 hovering sorties. The engine had
many limitations, in particular RPM bands where steady running was
prohibited and a total hover life of 35 minutes. It was marginal
on thrust, resulting in the ability to carry only about three minutes
fuel and a need to strip from the airframe about 700 lb of electrical
equipment, cabin conditioning, instruments and undercarriage components
in order to get it into the hover. However it was a start.
First conventional flight
aircraft was then prepared for more taxying tests and its initial
conventional flying. The nosewheel steering was still unsatisfactory,
a brake induced torsional oscillation developed and broke a main leg.
The castoring outriggers showed a tendency to shimmy and finally had
to be locked for flight. The aircraft banked badly when taxying especially
in a cross wind. The first conventional flight took place in March
1961 and was followed by ten more flights.
These took the aircraft (fig 13) to 400 kt,
0.8M, 30,000ft and 4.7 g. It was concluded that longitudinally the
aircraft was unstable and suffered from pitch up. Laterally it was
good and directionally was satisfactory, except with the undercarriage
down when there was slight instability. The flap and airbrake trim
changes were too great. There was a large nose down trim change
in ground effect so full back stick was needed to land.
Fig 13 P1127 3-view
|The engine operated satisfactorily but only with
many handling restrictions necessary to avoid fan blade resonance
behind the very short intakes. Strip examination after the trials
showed that an HP turbine blade had been shed at some time. The fuel
and hydraulic systems worked well, the cabin conditioning discharged
water at times. The nosewheel steering system was barely useable,
the brake judder was severe, the anti-skid often failed and the pilot
could not feel that the wheels were skidding at any time. Several
tyres were burst. Thus encouraged the team prepared the aircraft for
further taxying and hovering trials.
For the next hovering trials the aircraft was fitted
with the 12,000 lb thrust Pegasus 3 engine and the reaction controls
were modified to give a variable bleed of 9 - 15 lb/sec depending
on control deflection. This was in an attempt to achieve greater
max control powers and at the same time to reduce engine thrust
losses due to control demands. 51 hovering sorties were carried
out for a total of 61 minutes using two engines between mid May
and mid June 1961.
The aircraft was hovered at heights up to 100 ft
and translated forward by changing its attitude at speeds up to
50 kt. Numerous landings were made on concrete and tarmac, although
roll control was inadequate for dealing with jet induced disturbances
close to the ground. Previous operations over the grid did not allow
ground effects to be sampled as the exhaust gases passed through
the ground plane and did not rebound off it.
A major characteristic of the P1127/Kestrel/Harrier
family was established at this stage. This was that at low speeds
the aircraft was directionally unstable. This instability resulted
from the high intake momentum drag of the large engine acting ahead
of the CG, the fin of course had only a weak stabilising effect
at low speeds.
Due to the change to partially variable bleed the
yaw reaction control power was reduced since it was obtained by
swivelling the pitch jets, but they were now worth only 7 - 10 lb/sec.
This yaw control power was insufficient to overpower the directional
instability in cross winds of 10 kt. (Or if you prefer, the aircraft
turned backwards uncontrollably if it was translated sideways in
the hover at 10 kt)
Following an increase in the roll reaction control
gearing it was found possible to dispense with the outrigger extensions
although untidy lateral behaviour persisted. Satisfactory un-autostabilised
hovering was also demonstrated. Free from tethers, height control
could be looked at for the first time and it was found that a linkage
giving low sensitivity at high RPM was best removed and the pilots
were happier with a more sensitive throttle. A hand rest was incorporated
outboard of the throttle to enable more precise hand movements.
The typical thrust weight ratios flown were 1.07 to 1.
The aircraft was easiest to hover well out of ground
effects at some 50 - 100 ft. Below 20 ft ground effects increased
rapidly until below about 5 ft control was not really adequate and
full control was often used.
By now a great deal had been learned about the hovering
business. The pilots were asking for a better longitudinal feel
with less friction and force feedback from the front reaction control
that was connected directly to the stick. They also wanted the datum
shift trimming system deleted and full authority over tailplane
travel to be available to the stick at all times regardless of the
The vital nose wheel steering system was proving
unreliable as well as deficient in handling. The aircraft still
needed 500 lb of its full compliment of bits removed for these trials.
Importantly for the future, experience had been obtained of the
strong lateral out of trim that could be obtained with sideslip
at even modest forward speeds. When combined with the directional
instability this was to become a very significant handling characteristic
of the P1127/Kestrel/Harrier family.
In July 1961 a second aircraft joined the programme.
It expanded the conventional envelope to 40,000 ft, 538 kt, 1.02
M and 6 g. It crashed on its 35th flight when one of the Pegasus
cold nozzles fell off in conventional flight. Prior to this the
aircraft had started the speed reduction programme from conventional
speeds and reached a minimum speed of 95 kt by deflecting its nozzles
and flying partially jetborne. The speed reduction programme was
stopped at this speed, as by then the first aircraft had reached
the same IAS, starting from the hover and was thus cleared to continue
with full transitions.
Flight to high mach was temporarily held up at 0.89
M to 0.90 M due to a bad left wing drop, which happened very suddenly.
The fix for this was a row of 13 vortex generators on the wing.
However despite these VG's the improvement was limited and the following
list of major conventional flight problems was noted:
- Transonic wing drop, allied with poor useable
lift coefficient characteristics.
- Longitudinal manoeuvring instability under most
flight conditions, coupled with pitch up at high AOA.
- Unacceptable directional behaviour with the undercarriage
- Engine restricted running conditions as a result
of intake conditions.
If these problems are examined it will be seen that
they all resulted from the V/STOL capability of the design.
The wing had been designed as a delta to give the
maximum structural strength for the vital minimum weight. It undoubtedly
needed a swept trailing edge. The longitudinal instability was largely
due to having to keep the tailplane above the jets, thus exposing
it to the wing wake. The engine needed very large intakes, which
had to be horribly short by normal standards if excessive volume
was not to be occupied in a small aircraft. The reasons for the
undercarriage layout have already been mentioned. A moment’s
reflection on fig. 13 shows the constraints inevitable with mounting
a Pegasus close to the CG. But the eventual success of the programme
shows how right the designers were to face these problems and solve
The first aircraft had flown in the hover with a
metal intake having a bulbous shape for low speed efficiency, but
for the conventional flights this was replaced by one with a slimmer
lip. The second aircraft had this slimmer version. Since the sharp
leading edged shape produced an unacceptable thrust loss at low
speed and the other too much drag at high speed, it was intended
that the aircraft would be fitted with variable geometry intakes.
These took the form of rubber bags on the intakes, inflated for
low speed and sucked down for high speed (fig.14).
One flight was made to 335 kt with them sucked down,
but at this speed they started to lift and flap. This behaviour
persisted even after speed was reduced to 250 kt and so the search
for a higher level of suck was on.
Fig 14 Inflatable air intake
The nose wheel steering system on the second aircraft
incorporated a gear change system so that full rudder bar gave either
+/- 3 deg of nosewheel steer, or +/- 30 deg. Like this it was better
but still poor. The wheel brakes again suffered from V/STOL. They
had a lower share than usual of the aircraft weight acting on their
wheels, they were light in the interests of weight saving and yet
had to cope with high levels of idle thrust from the big Pegasus
Still on the undercarriage, directional handling
on the runway in crosswinds was very poor due to the continual leaning
tendency. The wing AOA in the normal ground attitude was about 8
deg - deliberately so in order to generate significant lift for
short take-off's, as rotation cannot easily be achieved with a bicycle
layout. Thus, when leaning during a landing run, the not inconsiderable
lift force present was tilted, producing a side force which was
only too willing to pull the aircraft straight off the runway should
the pilot be imprudent enough to add a crosswind of 10 kt or so
to his task.
On top of this when leaning heavily on one outrigger,
only one main wheel tyre was properly on the ground so the other
usually burst - they were not locked to the axle until later. Never
mind, a measure of reverse thrust was available instead of relying
on the brakes, so that was used instead.
Given that the nozzles only moved 18 degrees forward
of the vertical, powered nozzle braking as it was termed, only worked
at high RPM. This high RPM increased the total lift on the aircraft,
so it rose up the oleos, obtained greater lateral freedom, developed
more side force and left the runway even quicker. About this time
the number of "See me about the undercarriage" notes sent by the
pilots to the designers increased!
|Finally, as mentioned earlier, this aircraft was
lost when the left cold nozzle dropped off in wingborne flight and
the aircraft crashed in the final stages of an emergency approach
to land, after lateral control was abruptly lost at about 170 kt and
300 ft. The pilot ejected and was unhurt. The front nozzle material
was then changed from fibreglass to steel regardless of the weight
Fig. 15 Modified reaction control system
Before the initial transitions were carried out,
in September 1961, a further important change was made to the reaction
control system, with the fitting of a fully variable bleed system
With controls central, all valves were now shut,
reducing bleed demands on the engine, lowering jet pipe temperature,
saving engine life and increasing installed thrust. One effect of
this was that it was no longer possible to tilt the pitch jets in
the nose and tail to yaw the aircraft - since they were no longer
always open - thus separate yaw jets were fitted either side of
the tail as shown. This improved yaw control raised the limit of
control over the directional instability to lateral speeds in excess
of 20 kt.
Transitions were a major milestone although in the
event quite straightforward.
Following a vertical lift-off, the nozzle lever
was slowly pushed forward until the nozzles were aft and the aircraft
fully wingborne. If the nozzles were moved too fast the aircraft
sank, if the nozzles were moved too slowly the aircraft climbed
unnecessarily high. What could have been easier? There was the question
of directional control and minimising sideslip but with the improved
yaw control this was easier than, say, keeping a tail wheel propeller
aircraft straight during its initial take-off ground roll.
Decelerating transitions were equally no problem
although in this case it was found easier to lower the nozzles to
40 deg on the downwind leg and fly the aircraft in this configuration,
using conventional techniques, until about 1000 yards from the intended
point of hover. Here at 130 - 150 kt the nozzles were selected to
the hovering position. With all horizontal thrust removed the speed
quickly reduced due to drag and all that was required of the pilot
was to open the throttle to hold height as the wing lift reduced.
With both types of transition it was found that
AOA could be used as a height control at the higher speeds and there
was great latitude in how the pilot handled the combination of wing
lift and jet lift. The use of an AOA gauge was necessary to avoid
any risk of stalling the wing. There was also tunnel evidence that
above about 15 - 18 deg AOA, nozzles down, the aircraft would pitch
up. Two double transitions were made without autostabilisation and
it started to become clear that given adequate controls the pilot
could be expected to cope unaided.
The final initial V/STOL step occurred before the
end of 1961. This was the development of STOL techniques at weights
in excess of hover weight. For a short take-off the aircraft accelerated,
nozzles aft, to a speed calculated from its weight such that the
wing lift would provide the deficit of thrust weight ratio, plus
a margin to compensate for the fact the nozzles could not be lowered
all the way to the vertical in order to keep enough horizontal component
to maintain the speed already reached. At this speed, termed the
nozzle down speed, the nozzles were selected down to the intermediate
angle and the aircraft left the ground under the combined wing and
jet lift. The pilot then had to complete the latter part of an accelerating
transition. This STO technique proved entirely straightforward.
Slow landings were carried out by gradually deflecting
the nozzles down and increasing the power until the point where
there was only a small power reserve left and the AOA in use had
reached the maximum safe AOA below the stall. The speed at this
point was the minimum possible for the particular weight and thrust
available. The nozzle lever was then left and the approach controlled
on the throttle. Slow approaches tended to be a bit vague and imprecise.
There were two possible controls of lift (wing and engine) and two
of speed (RPM and nozzle angle) and only much later did the optimum
handling technique become clear.
When partially jet borne the airflow round the wing
and tail was much influenced by the four high velocity engine exhausts
and the longitudinal handling suffered from instability. However,
with the reaction controls bleed air supply main ON/OFF butterfly
valve being linked to the initial nozzle down travel, the pilot
had sufficient total control power, aerodynamic and reaction for
Grass operation was started, with both VTOL and
STOL operation being satisfactorily accomplished from a good quality
grass surface. So at the end of 1961 the P1127 worked. It had its
problems and there was a great deal to be improved but it only required
engineering and not black magic to progress to the point where the
vectored thrust aircraft could be seen to have some style.
P1127 to Kestrel
|Six P1127 aircraft were built and took part in a
programme that led to the Kestrel aircraft. The engine went from the
Pegasus 3 at 13,500 lb thrust to the Pegasus 5 at 15,000 lb. From
the Pegasus 3 onwards a third fan stage was used. Snubber blades were
developed for the fan to further reduce fan blade vibrations with
their attendant high levels of alternating stress. With this modification
and a change from light alloy to titanium as the first stage fan blade
material, the Pegasus 5 was the first version without restricted RPM
/ IAS bands.
Changes to the reaction controls continued in the
quest for optimum sensitivity, and just before the Kestrel went
into service the roll control power was given a further boost by
the provision of up-blowing valves. With this reaction control system
(fig.16) the need for the pitch and roll autostabiliser faded and
the Kestrel was not fitted with one.
Slow approach handling was aided by the development
of the fixed throttle technique instead of the fixed nozzle technique
(or as that had usually turned out the fixed nothing technique)
This principle was very simple. At the start of the approach, power
was increased to about 5% below full throttle and the nozzles lowered
to reduce speed. As speed and wing lift reduced, the flight path
was maintained by using an ever-increasing AOA until the optimum
AOA was reached. At this point the speed was noted and then maintained
for the rest of the approach using the nozzle lever as the speed
Fig 16 Final reaction control system
This technique had the advantage that there was
a known reserve of thrust (5% RPM) for use in the event of a late
go-around being needed. This use of the nozzle lever as a speed
control was excellent, as there was no lag since the engine did
not need to accelerate.
But, for a single seat combat pilot, the outstanding
virtue of this slow landing technique was that the pilot did not
need to know his aircraft weight or the ambient pressure and temperature
to decide his slowest possible approach speed. He just set 95% of
max RPM, slowed down until optimum AOA and then held that speed.
The only other V/STOL matter that needed to be resolved
before the Tripartite Kestrel Squadron standard of aircraft was
reached was the intake. After a very great struggle the rubber bags
were kept sucked down up to 0.75 M but only after a conventional
take-off. It was not possible to inflate them, do a VTO, accelerate
and suck them down adequately. At a very late stage a compromise
metal lip intake was produced that did not suffer from a significant
thrust loss at the hover and at the same time was satisfactory for
conventional flight needs. With this the V/STOL Kestrel had arrived.
In parallel with this V/STOL programme conventional
flight was progressing. Satisfactory transonic handling in the end
needed a completely new wing. This had a different section, leading
and trailing edge shapes and also different tips and vortex generators.
Fig. 17 Kestrel 3-view
The aircraft now looked like fig. 17. It now had
a swept trailing edge. Longitudinal stability required the tailplane
tips to be extended. The fuselage had to be made longer and the
CG changed. It will be appreciated that this was a major redesign.
The fin had 2 sq ft added to its tip to correct the undercarriage
down instability in conventional flight.
The wings were given two hard points so that it
was possible to carry 2 x 100 gallon fuel tanks or rocket pods although
in the event the rockets were never fired. A small gun sight was
fitted to allow the Tripartite Squadron to evaluate tracking. The
nose wheel steering was re-engineered to remove backlash and a selector
button was put on the stick which gave +/- 5 degrees of nosewheel
at full rudder pedal with a second button on the throttle which
gave +/- 35 degrees. The brakes were improved and changed from hand
to toe operation.
The engine development continued
with two mechanical engine failures in conventional flight. Both
aircraft were force landed but one was written off with a fire.
The other was rebuilt. The first engine failure was due to high
'g' distorting the casing and allowing the 8 th stage rotors to
touch the stators. This problem was engineered out but the second
case was more difficult. Here the engine surged at high mach (about
1.2 M) due to flow distortion from intake shocks. The surge was
sufficiently violent to again tangle up stators and rotors. Improved
clearances in the engine were arranged to prevent mechanical damage
recurring but the only way of preventing engine surge at supersonic
speeds was to limit RPM to 80% in transonic dives. The Kestrel went
into service with this limitation. It did not, however, have any
airframe IAS or Mach limits but could be dived vertically to its
terminal velocity. To avoid high N / ?¥ surges the engine had to
be throttled back at altitude. A surge limit gauge was fitted indicating
the maximum allowable fan RPM.
The Kestrel Tripartite Squadron (US, West German
and UK ground and air crews) was equipped with nine aircraft and
commenced operations in October 1964.
Kestrel to Harrier
Then in the spring of 1965 a new contract was obtained,
for six development batch aircraft, of a type that was to be known
as Harrier. This step, to the Harrier was many times greater than
the change from P1127 to Kestrel. It was literally
a new aircraft from an engineering point of view retaining only
the same basic configuration, but that was fixed by the Pegasus
engine (fig. 18).
The Harrier aircraft went into service as an operational
type four years later in April 1969. Those four years saw much development
before the airframe and engine design teams achieved their objectives.
It will be appreciated from just the subject headings that follow
that a great deal was involved.
Fig 18 Harrier GR1 3-view
Fig. 19 Auxilary air intake doors
The intake design was changed twice in the pursuit
of better specific air range in the cruise allied with good high
altitude handling, without prejudice to the hovering performance.
First 12 and then 16 auxiliary intake doors (fig.19) were fitted
round the intake lip and also various boundary bleed door configurations
were flown. The reaction controls were refined and to add a bit
of polish to the V/STOL handling a new low authority pitch and roll
autostabiliser was developed.
Then came the breakthrough - the undercarriage was
finally sorted out. The nosewheel steering was freshly engineered,
and this time by means of a cunning slotted link and slipping clutch
mechanism a vernier range of travel about the centre rudder bar
position gave excellent fine steering, without spoiling the minimum
turning radius capability of full rudder. The main wheels were locked
to the axle and the anti-skid system improved. The nosewheel tyre
was changed improving its floatation on boggy ground.
Then finally the most important fix of all, the
main leg was given a two-stage oleo. On touchdown this "self-shortening
leg” gives no rebound from the first seven inches of travel.
This means that at all weights the aircraft rests firmly on its
outriggers and does not gain outrigger freedom until lift exceeds
weight. Cross wind handling on the ground with this leg became the
same as into wind handling with the aircraft running dead straight
and wings level.
Conventional flight problems, other than the engine
ones already mentioned, centred on the wing development. It was
originally deficient in useable lift coefficient at high
mach number. The specification required 6 g at 16,800 lb, 400 kt
IAS and 10,000 ft. This represented a lifting capability from the
available wing area that was greater than any previous comparable
military aircraft. Initially the aircraft was short of 1 g. The
use of a combat flap position gave an extra 1/2 g. The other
1/2 g could be achieved by deflecting the thrust but the firm's
pilots were against this expedient since, although it demonstrated
the requirement, they were only too well aware that as well as getting
the increment of 'g', one also got the effect of the world's finest
airbrake as the rearwards thrust was lost. In the end a very considerable
wind tunnel and flight programme of different types of leading edge
and vortex generator modifications resulted in the 1/2 'g' required,
allied to good handling and without the need to vector the thrust.
Kestrel longitudinal stability was satisfactory
with a clean aircraft, but deteriorated with under wing tanks. Since
the Harrier would need to always carry a wide variety of external
weapons, further improvements were necessary to counter de-stabilising
stores, so the Harrier CG was moved yet further forward. Handling
was then satisfactory with the full range of stores.
The foregoing has been an outline account of the
nine years that turned the P1127 flying machine into the first service
version of the Harrier. It ignores the vast effort that was going
on at the same time regarding the weapons system which was itself
a very advanced programme for those days. Since then the later and
much improved members of the Harrier family have shown what more
can be done if you start with the right concept.
The later years
|During the early 1970's a package was developed to
help the pilot deal with the issue of sideslip during mid transition
where intake momentum drag makes the aircraft directionally unstable.
An autostabiliser working through the yaw reaction controls was produced
to reduce this instability. Additionally, an indication was provided
in the head up display (HUD) of a safe limit of sideforce, allowing
the pilot to limit any lateral out of trim that might otherwise result
in a loss of lateral control. Finally, in case the pilot was distracted
from noticing the HUD warning, rudder pedal shakers were fitted which
operated as the HUD sideforce limit was reached. The system shook
the pedal that the pilot needed to use to reduce the sideforce, not
only reminding him to use his feet, but also indicating which foot.
In 1977 ski jump trials were carried out, initially
from a land-based ramp that was adjustable from a 6 to 20 degree
exit angle (fig.20). This showed great improvements were possible
with regard to performance, handing, safety and ship pitch motion
limits when compared to flat deck ship STO's. It is interesting
that the quite remarkable advantages of a ski jump deck have recently
allowed the Russian Navy 's first conventional aircraft carrier
to be built without catapults.
Fig 20 RAE Bedford adjustable ramp
Fig. 21 Sea Harrier FRS Mk1
In 1978 the first Sea Harrier (fig.21) was flown.
It had a radar-based weapons system and a new raised cockpit for
better air combat visibility. Many other changes were made to this
aircraft to optimise it for shipboard operation.
The hover and VL handling was improved by yet another
lateral control sensitivity increase to aid recovery to ships in
rough water. The throttle box also included an ability to adjust
the nozzle angle, +/- 10 degrees about the position set by the nozzle
lever, using the airbrake selector. This enabled fore and aft position
adjustments alongside a ship to be made without taking the left
hand away from the throttle height control or changing the aircraft’s
In 1979 the first of the Harrier II prototypes were
tested. This branch of the Harrier tree uses a graphite epoxy McDonnell
Douglas designed wing of 230 sq ft area compared to the metal Harrier
I wing of 200 sq ft. This new wing effectively doubled the payload
radius of action from the same length of STO run.
Harrier II VTO performance was also improved by
developing larger under fuselage strakes and adding a cross dam
behind the nose leg to contain the positive ground cushion that
results during VTO from ground reflected jet impingement on the
bottom of the centre fuselage. Harrier II handling was improved
by a new three axes autostabilisation system that operated throughout
the flight envelope. Harrier II plus is the designation of the latest
radar based version now in service with the Spanish Navy.
Various uprated engines are used by Harrier II operators,
all using zero scarf front nozzles some having a thrust as high
as 24,000 lb.
Important past choices
In 1964, when I started flying the first P1127 prototype
I was asked by the RAE scientists to compare it with the more complex
lift and cruise concept of the Short SC1 research aircraft that
I also flew. The SC1 was a small delta powered by five Rolls-Royce
RB108 engines. Four were used for lift and one for propulsion.
This aircraft was easier to fly in the hover
than the P1127 because the intake momentum drag of the lift engines
acted downwards through the centre of gravity and so produced no
tendency for the aircraft to yaw out of wind. So the fin had no
destabilising force to overcome and was able to keep the aircraft
However although easier to fly in the hover the
SC1 was a nightmare to operate.
It was literally a five engine aircraft being operated
by a crew of one. In those days there was no automation so the pilot
had five of most engine controls and instruments. On top of this
was the need to shut the lift engines down after take-off and even
more importantly restart each one before landing. Because of the
high fuel consumption with the lift engines at flight idle, this
complicated light up had to be left until the aircraft was on the
final approach a mile or two from touchdown.
So, despite the cruise implications of having an
engine and intake compromised by the need for high thrust at low
speed, the P1127 simplicity and inherent reliability was preferred
to the multiple lift engine concept.
But what of the future? Given the march of technology,
the availability of reliable automation and fly by wire, together
with the benefits of not having your exhaust coming out around the
centre of the aircraft, the argument for simplicity at all costs
may become less overwhelming. For some roles, especially supersonic
ones, the advantages of more complex multi - mode powerplants or
lift and lift cruise concepts may well provide a better option.
But with the benefit of hindsight the Pegasus/Harrier
combination was clearly the correct way to launch jet V/STOL. Of
that I have no doubt.
© J F Farley 2 May 2000