This text describes how Saab has worked towards developing flight control systems for military fighter aircraft and how Saab has evolved in this technology area.
We will describe how different flight control systems have developed from the 1940s up to the JAS 39 Gripen (hereinafter referred to as “Gripen”) in the 2000s.
This flight control system is one of the most advanced systems found in military fighter aircraft. What really sets the operational level apart in a military fighter aircraft is the functionality and performance of the flight control system.
In order to see the development of flight control systems for military fighter aircraft in context, we first need to understand the basis of how flight control systems work. This chapter will also describe in particular how the flight control systems and system solutions have been developed for Saab’s different military fighter aircraft.
One of the most central issues for the Swedish Air Force was to reduce the vulnerability of the entire air defence system. In the 1950s, the air force had realised the vulnerabilities associated with fixed air bases, so they developed a concept for road bases along normal roads.
Such road bases also had adjacent parking areas for aircraft with maintenance capability. As a consequence, the aircraft needs very good take-off and landing characteristics, as runway length on a road is very limited.
These conditions formed the basis for the construction of the aircraft Saab has developed. The flight control system for all aircraft is a primary system which has a significant influence on an aircraft’s performance.
The author recommends the following texts that relate to this story: In chapter Creating value for customers under the heading Development of Technology Demonstrators in chapter Keeping unique development skills under the heading Designing the world’s best control system and under the heading Methodology and analysis in design work.
The text refers to the highlighted areas in the journey of change in the aircraft industry
Throughout the years, Saab has developed flight control systems for different types of military fighter aircraft.
In a conventional statically stable aircraft the aircraft has a ‘built-in’ inherent stability which requires a large wing area and causes a lot of drag that slows the aircraft under a turn. This type of flight control system was used on all Saab aircraft up to and including the Saab JA 37 Viggen.
The beginning of the 1980s saw the start of an experimental project, ESS01, to demonstrate that Saab could fly using a three-channel asynchronous electrical digital flight control system. This was to be adapted for use in the future Gripen aircraft.
This is where the journey of change towards an electrical digital flight control system with full authority began. This paved the way for a statically unstable basic aircraft as well as the dramatically improved flight characteristics of modern statically unstable aircraft.
The Gripen is a modern aircraft with performance which only a statically unstable aircraft that cuts through the air with little energy loss can achieve. The Gripen is fundamentally an unstable aircraft, and this instability must be continuously parried using active commands from the autopilot so as not to diverge.
The advantages of a statically unstable basic aircraft are thus less resistance due to the smaller wing area and forces that works together on the aircraft. Furthermore, a statically unstable basic aircraft that is stabilised by a flight control system that works continuously offers better stationary turn performance and better energy preservation.
The pilot can fly the Gripen aircraft “Carefree” without having to think of all of the different limitations to the aircraft, thereby being able to concentrate fully on his/her tactical mission.
This part describes the historical development of control systems and how Saab’s work has developed in this technology area. We describe how different flight control systems have developed from the J21 in the 1940s up to the Gripen in the 2000s. The flight control system is one of the most advanced systems found in military fighter aircraft. Its functionality and performance set the level of operational capacity of a military fighter aircraft.
In order to see the development of flight control systems for military fighter aircraft in context, we first need to understand the basis of how control systems work. The following text provides a very brief description of control systems and the different system solutions used in Saab’s various military fighter aircraft.
One of the most central issues for the Swedish Air Force was to reduce the vulnerability of the entire air defence system. In the 1950s, the air force had realised the vulnerabilities associated with fixed air bases, so they developed a concept for road bases along normal roads. Such road bases also had adjacent parking areas for aircraft with maintenance capability. As a consequence, the aircraft needs very good take-off and landing characteristics, as runway length on a road is very limited.
The Saab J 21 was developed in the 1940s and featured a design with twin tail booms and a rear-mounted propeller. It was the first aircraft in the Swedish Air Force to use nose landing gear. The aircraft design offered the pilot good visibility, but resulted in cooling problems for the engine and the propeller could injure the pilot should he need to leave the aircraft in flight. As a consequence, Saab developed a gunpowder-powered ejector seat that shot the pilot out and over the hazardous propeller area. Following the end of the Second World War in 1945, Sweden gained access to foreign jet engines and a version of the Saab J21 with a jet engine was developed as the Saab J21R.
The earliest control systems from the 1940s such as those on the Saab J21 were mechanically reversible, where the pilot could feel the force of air on the control surfaces via the mechanical control system as the forces were fed back to the control stick. The pilot could feel the force of the air in the control stick. High speeds resulted in greater forces in the control stick, and thereby smaller control surface movements. Low speeds resulted in lower forces in the control stick, and thereby larger control surface movements.
Flight control system for the Saab J21A/R – the mechanical reversible system
The 1950s saw the development of the Saab J 29, which was given the nickname “The Flying Barrel” and was a single-engine Swedish fighter aircraft. From 1954 onwards, all combat and attack aircraft were rebuilt and fitted with afterburners, enabling the aircraft to set speed records for the time. At that time, the J 29 was considered equal to the best American F-86 aircraft and Soviet MiG-15 aircraft.
Due to the high speeds, the pilot required help with controlling the control surfaces at high speed, so the flight control system in the Saab J 29 was equipped with servo assistance. The pilot could still feel the force of the air from the control surfaces fed back through the control stick.
Flight control system for the Saab J 29 Tunnan – the mechanical reversible system
The latter part of the 1950s saw the development of the Saab 32 Lansen, which was a two-seater fighter aircraft with all-weather capability which came in three principal variants. The Saab A 32A was an attack version with all-weather capability. The fighter version with all-weather capability was called the Saab J 32B and the reconnaissance version was named the Saab S 32C.
The Saab J 32B Lansen was the first ever Swedish aircraft to break the sound barrier. The aircraft was subject to very high forces, with the pilot requiring help from the servo to control the control surfaces. This meant that the pilot could no longer feel the force of the air through the control stick, so artificial control stick feel was implemented using springs and dampers instead.
Flight control system for the Saab J 32 Lansen – mechanical system with full servo
The Saab J 35 Draken was developed to fulfil the need for a fighter that could intercept the new bombers with nuclear warheads introduced during the 1950s.
The aircraft design was based on a narrow fuselage with a double-delta wing configuration. This type of wing offered good flight characteristics at high speeds thanks to the inner, more arrow-shaped wing. The outer wing was complementary, displaying good performance at low speeds.
The Saab J 35 Draken was designed to intercept supersonic enemy aircraft which meant it required full servo and, moreover, assistance from an autopilot system to improve the poorly dampened pitch characteristics at supersonic speeds. The flight characteristics were improved with the aid of the autopilot while the pilot’s sense for the control characteristics were determined by the mechanical control system.
Flight control system for the Saab J 35 Draken – mechanical system with full servo + pitch
The Saab 37 Viggen was a fighter aircraft that came in four different variants, as well as a trainer (SK 37). The aircraft had a double-delta wing configuration with so-called canard wings (extra wings forward of the main wings) which made it possible to land at lower speeds. A reversing system made it possible to change the direction of the air flow with the help of a thrust reverser when landing on the ground. This resulted in the aircraft being able to land in a very short distance: less than 500 m even on slippery surfaces.
The AJ 37 Viggen went one stage further improving the pilot’s sense for the control characteristics by using a force sensor on the control stick. The flight characteristics were improved by adding stabilising and damping signals from the autopilot to the control surfaces commands. The autopilot on the AJ 37 Viggen was analogue. The components in analogue autopilots have varying tolerances, meaning that each aircraft was unique and could display slightly different flight characteristics.
Flight control system for the Saab AJ 37 – mechanical system with full servo and full authority + limited 5 deg analogue FCS
The Saab JA 37 Viggen was the first aircraft in the world to feature a digital control system added to the basic mechanical control system. The digital control system of the JA 37 could eliminate the variations in tolerances found in analogue control systems.
Flight control system for the Saab JA 37 Viggen – mechanical system with full servo and full authority + limited 5 deg digital FCS (first digital FCS in the world)
All of the aircraft described above were statically stable basic aircraft which could be flown using their mechanical flight control systems alone.
The beginning of the 1980s saw the start of an experimental project, ESS01, to demonstrate that Saab could fly using a three-channel asynchronous electrical digital flight control system for the future Gripen aircraft. This is where Saab’s journey of change towards an electrical digital flight control system with full authority began, paving the way for a statically unstable basic aircraft and the dramatically improved flight characteristics of such modern statically unstable aircraft.
A Saab JA 37 Viggen was rebuilt into an experimental aircraft called the JA 37-21 ESS01, which was used to test a full authority electrical flight control system. The change meant that the aircraft’s flight control system became a three-channel asynchronous full authority (±30 degree trailing edge control surface) electrical digital flight control system with a mechanical backup system.
Flight control system for the Saab JA 37 ESS01 – three-channel digital full authority fly-by-wire FCS with mechanical backup
As a comparison, the Tornado and Concorde used similar control systems. The Tornado is a twin-engine fighter that was developed in three different variants designed by a tri-national consortium made up of companies from the UK, West Germany and Italy. The Concorde was a supersonic passenger airliner and was developed in a collaboration between the French company Aérospatiale and BAC in the UK.
The experience gained from the JA37 formed the basis for the design of the flight control system in later aerial vehicles; one aspect in particular is the monitoring of calculations, which was used as a basis in the development of the Gripen autopilot, designated the SA11.
The Gripen is a modern statically unstable aircraft with such good performance that it cuts through the air with little energy loss. This instability must be continuously parried using an active counter-command from the autopilot so as not to diverge.
Flight control system for the Saab JAS 39 Gripen – three-channel digital full authority fly-by-wire FCS with digital backup system
In order to understand the function of a flight control system, below we describe the difference between conventional statically stable and modern statically unstable aircraft.
The figure shows a comparison between stable and unstable aircraft.
In a conventional statically stable aircraft the aircraft has an inherent stability which requires a large wing area and causes a lot of drag that slows the aircraft when turning.
Examples of conventional statically stable aircraft include the Saab J 29 Tunnan, Saab J 32 Lansen, Saab J35 Draken and the Saab AJ/JA 37 Viggen.
In a conventional statically stable aircraft the lift of the wing acts behind the centre of gravity, making the aircraft automatically free float into the wind. The aircraft has a naturally restoring aerodynamic pitching moment and therefore natural stability.
In order for the aircraft to turn, the nose must be held up by a downwards counter-force from the control surfaces. This means that the lift of the wing must compensate for the downwards force from the control surfaces, in order to maintain the load factor and lift of the turn. This means that the wing area must be large, resulting in a lot of drag.
The pilot control the control surfaces via the control stick to turn the aircraft. A statically stable basic aircraft can be flown using only a mechanical control system as the aircraft’s stabilising flight characteristics are inherent to its aerodynamic design.
The disadvantage of a statically stable aircraft is that it requires a large wing area to compensate for the downwards force on the control surfaces, which means a lot of drag and large energy loss when turning and therefore a large reduction in turn performance. The landing speed is also higher as the control surfaces force counters the lift of the wing. Moreover, a statically stable aircraft becomes even more stable when flying at supersonic speeds, meaning the control surfaces must be controlled to greater amplitude in the opposite direction. Because dynamic pressure affects the control surfaces control surface forces, so powerful control servos must be installed, increasing the overall weight.
To summarise, conventional statically stable aircraft most often use mechanical basic control systems with a large number of non-linearities such as friction, hysteresis and dead zones, which impair the control characteristics.
A dead zone is an interval in a signal domain in which no action takes place. The output signal of a signal sent via a non-linearity as a “dead zone” depends on the magnitude of the signal. A small signal through a “dead zone” does not issue a signal, whereas a large signal goes through reduced by the size of the dead zone.
Hysteresis is the time-based dependence in a system’s output on recent and earlier input.
A modern statically unstable aircraft lacks inherent stability and so requires continuous artificial stabilisation. The wing area is smaller, and therefore experiences less drag.
In a modern statically unstable aircraft, the lift of the wing acts forward of the aircraft’s centre of gravity; a statically unstable aircraft has no natural restoring aerodynamic pitching moment, i.e. no inherent stability. The lift of the wing must be continuously parried by the flight control system. The forces on the control surfaces acts in the same direction as the lift of the wing to create a parrying counter-moment. These forces are combined into the total lift acting on the aircraft, which means that the wing of a modern statically unstable aircraft can be made much smaller which provides less drag. This enables the energy to be better preserved.
At supersonic speeds, the lift of the wing moves even further back to 40% of the chord (the chord is the straight line from the leading to trailing edge of the wing). This means that modern statically unstable aircraft become statically stable at supersonic speeds.
A statically unstable basic aircraft needs continuous stabilisation with the help of an active flight control system with full authority.
However, at supersonic speeds there is less control surface movement compared to a conventional statically stable aircraft. This means that modern statically unstable aircraft do not need the same powerful and heavy control servos.
The aircraft cuts through the air. Landing speed is reduced as all forces acting on the aircraft combine to lift the aircraft while also slowing it down and reducing its speed.
Additionally, a statically unstable aircraft flying subsonic is less statically stable at supersonic speeds compared to an old conventional statically stable aircraft. This entails lower control servo forces at supersonic speeds and thereby reduced control servo weights, as they do not need to be as powerful. This also means less weight to carry when flying at subsonic speeds.
Modern statically unstable aircraft most often have digital electrical flight control systems with full authority which eliminate the weight and non-linearities of the mechanical control system.
The disadvantage is that the control system must continuously issue its stabilising commands with the shortest possible time delay and at a frequency of at least 60 Hz.
If amplification is too low on the control system’s stabilisation, the stabilisation will be insufficient; if, on the other hand, amplification is too high, it can make the aircraft dynamically unstable at a higher frequency and, moreover, even result in aeroelastic oscillations (flutter).
The figure below illustrates the difference in wing area and placement between the JAS 39 Gripen, a modern statically unstable basic aircraft (shown in grey) and the JA 37 Viggen, a conventional statically stable aircraft (shown in blue).
In the JA 37, an automatic aiming system was implemented. In a Saab JA 37 Viggen, the automatic aiming function is built into the aircraft’s digital flight control system with a authority of 5 degrees and uses data from both the radar and the “central computer”. The radar provides information on the distance to and the position of the target. This data is process in the “central computer” in JA 37, which carries out a aiming calculations and feeds aiming errors to the flight control system so that the flight control system in JA 37 can eliminate the aiming error.
JA 37 had a mechanical primary flight control system with a digital low authority flight control added to the mechanical primary flight control system. JA 37 also had a mechanical long control stick with nonlinearities. The mechanical system together with the destabilizing mass effect from the arm affected the aiming accuracy in gun aiming. It was hard to put the pipper on the target. To improve the aiming accuracy and to quicken the aiming a Automatic Aiming control system implemented in the digital low authority flight control system for JA 37. The pilot did the gross aiming with the normal control system mechanical and digital. At a certain point he pressed the Automatic Aiming button and the Automatic Aiming control system automatically did the fine aiming and putting the pipper on the target.
JAS 39 Gripen has a digital full authority fly by wire with a ministick without nonlinearities and no destabilizing mass effect from the arm. The JAS 39 Gripen can do very accurate manual gun aiming without help from a Automatic Aiming fuction.
Automatic aiming provides support for fine gun aiming, so the pilot can aim more quickly and more accurately compared to manual aiming in JA 37. Automatic aiming supports all target angles and uses the autopilot’s pitching and yawing channels to control the aircraft automatically so that the aiming error is eliminated and the pipper is on the target. In a roll, a roll command is presented to the pilot to follow so the pipper is close to the target. The pilot can connect the Automatic Aiming by pressing a button on the control stick and follow the bank angle command presented.
The aircraft then carries out the aiming by adjusting in pitch and yaw and also compensates for any possible errors which are not handled by the pilot using the autopilot’s yawing channel. If the pilot does not follow the bank angle command, the aircraft flies with so-called “sideslip” which gives a certain type of point aiming.
A sideslip involves wind coming in from the left or right-hand side of the aircraft (for example a gust) and makes it fly into the wind like a weather vane.
This section explains how the Automatic Aiming was developed for the JA 37 Viggen and the SA 07 autopilot.
Classic control system design was used for the Normal Mode on the JA 37 digital flight control system.
For the parallel digital Automatic Aiming control system in the JA 37 Viggen Modern Control theory was used Linear Quadratic design.
The JA 37 Viggen was a conventional statically stable aircraft both when flying subsonic, and even more stable when flying supersonic. The Viggen was statically stable in supersonic flight which required great control surface movement and therefore large, heavy and strong control servos were necessary.
The JA 37 Viggen could be flown using just the basic mechanical primary flight control system without the need for additional stabilisation. A conventional statically stable aircraft has a large wing area which results in high drag when the pilot makes a turn.
The JA 37 Viggen aircraft had a flight control system which was divided into a primary mechanical flight control system with full authority ±30 degrees and a digital flight control system with an authority of ±5 degrees in pitch, roll and yaw. The digital flight control system signals were added to the primary mechanical flight control system control servo commands, thereby enabling the digital flight control system commands to add or subtract to the primary mechanical flight control system commans to improve handling and flight characteristics in the flight envelope. The mechanical control system consisted of push rods and lines between the control stick and control servos. The image below shows the different components in the flight control system with the SA 07 digital autopilot circled.
The figure shows the different components of the control system with Autopilot SA 07 circled.
The authority of the autopilot was limited to what was required to improve the aircraft’s characteristics, and the autopilot could be disengaged in case of error. Commands from the digital autopilot required comprehensive monitoring of the digital single-channel computer and comparison of signals from different sensors in order to satisfy the safety requirements. In case of an error, the autopilot was disengaged and the aircraft continued to be flown using just the basic primary mechanical flight control system.
The Normal manual mode of the digital flight control system was designed using conventional control theory, which made it difficult to put together loops on loops in the flight control laws with satisfactory flight characteristics.
Flexibility in updates to the autopilot
The JA 37 Viggen had relatively thin wings which affected the aircraft’s aeroelastic properties. This meant that complicated, advanced, non-linear digital filtering was needed to be added to the flight control laws. This affected the flight characteristics and the control law design.
The wing control servos for the trailing edge control surface were connected. This led to a force fight between the two servos. At high speeds the control surfaces had a tendency to twist, resulting in considerably decreased control surface efficiency and a lag effect. These effects had to be taken into account when designing the JA 37 Viggen’s flight control laws.
On the other hand, the function in the autopilot on the JA37 Viggen could be modified easily as it was principally defined in software which made it possible to update the control system continuously with completely new functions even once the basic development had been finished and the aircraft was already air force service. This meant these functions could be introduced before test flights and in series aircraft in the air force. In earlier autopilots, the function was defined by the hardware, analogue circuits and logic functions.
JA 37 Viggen hade vingar med relativt låg styvhet vilket påverkade flygplanets aeroelastiska egenskaper. Detta innebar att en komplicerad avancerad icke linjär, digital filtrering var nödvändig att införas i styrlagarna.
När man var tvungen att ta hänsyn till den icke linjära digitala filtreringen, påverkades flygegenskaperna och därmed konstruktionen av de digitala styrlagarna.
Vingrodren var kopplade vilket kunde ge upphov till ”kraft fight” mellan de två servona, som i tillsammans med höga farter gjorde att roderytorna hade en tendens att vrida sig (tordera) vilket medförde betydligt sämre rodereffektivitet och fördröjningseffekter i höga farter. De här symptomen hade man att ta hänsyn till vid konstruktionen av JA 37 Viggens styrlagar.
Å andra sidan kunde funktionen i styrautomaten i JA37 Viggen enkelt ändras eftersom den i huvudsak definierades av programvara, som gjorde det möjligt att fortlöpande uppdatera styrsystemet med helt nya funktioner, även efter avslutad grundutveckling, då flygplanen fanns på förband. Dessa funktioner kunde då införas både för provflygning och på serieflygplan i flygvapnet. I tidigare styrautomater var funktionen definierad av hårdvara, analoga kretsar och logikfunktioner.
As processors and memory (core memory) were expensive components at the start of the 1970, comprehensive studies on different safety systems were carried out. The initial suggestion was for one processor to carry out the calculations, and a second to monitor them on a “bit by bit level”, i.e. 2 processors. After extensive analysis and error simulations, the decision was made to use one single processor with a comprehensive self-testing system carried out in real-time in conjunction with hardware logic. This monitoring would be able to disconnect the digital autopilot at any moment before an error could critically affect the flight. This made it smaller, lighter and cheaper.
The processor which was used, the HDC-301 Honeywell Digital Computer, was well-suited for the task in a real-time system, both for analogue and discrete signals. Input and output comprised – among other things – a combined analogue/digital and digital/analogue converter which could easily be addressed in the software.
Interoperability with the rest of the avionics system in the JA37 took place via dedicated serial transmission to each of the different systems. Input was written directly to the memory – DMA (direct memory access). This hardware logic was replaced in later systems with a special processor for I/O handling which offers much greater flexibility.
Reasons behind continued development of the control system on JA 37 Viggen.
Experience from the JA37 Viggen formed the basis for the design of the control system in later aircraft. This applied to the monitoring of the calculations used as a basis for the design of SA11 autopilot on the Gripen. This autopilot has three independently-operating asynchronous calculating channels which means that the requirement for individual error detection can be lower in the Gripen.
The lessons learnt from the JA 37 Viggen development work which formed the basis for the design of the Gripen flight control system can be summarised as follows:
Testing during development of the automatic sighting
Different types of tests were carried out during the development phase for the Automatic Aiming function for the gun (AKAN) including aiming tests with pilots in a JA 37 simulator. The tests showed that the distribution from different pilots when aiming with the JA 37 Viggen’s normal control system, digital control + mechanical control with basic flight control system varied greatly.
This was based on the fact there were non-linearities in the basic mechanical control system such as slack, friction and hysteresis (“inaccuracies” in the system). The long mechanical control stick had a low pivot point in pitch, i.e. a long lever, and a shorter level in roll. There were also non-linearities in the long mechanical control stick. In addition, there was also a destabilising effect, bobweight a mass effect from loadfactor acting on the arm and control stick when turning due to the effect of the load factor.
The possibility modern control theory was investigated by Saab, the Lund University (LTH) and the Linköping University (LiTH). Saab decided to use the Linear Quadratic design technique to design the Automatic Aiming control laws.
When developing the Automatic Aiming control law, it was thought that the pilot should do the rough aiming of the target and then – when in a appropriate position – connect the Automatic Aiming function which would eliminate aiming errors rapidly. The general requirement was that 90% of the aiming errors should be eliminated within 1 second.
This could be improved at a later stage and this resulted that the pilot aimed with the automatic gun, the radar measure the aiming error to the target and the Systems computer calculates the aiming error for the pitch and yaw. The Automatic Aiming control law then automatically carried out the aiming on the target quickly and effectively.
The Automatic Aiming flight control laws were a parallel digital flight control system in the JA 37 digital flight control system computer. These parallel flight control laws for Automatic Aiming control laws were selected by the pilot via a special button on the control stick for Automatic Aiming.
These gains in the parallel control law for Automatic Aiming varied with speed and altitude. With an extra change in the flight control system, the authority of the digital autopilot could be extended from 5 degrees to 8 degrees. The roll channel was used to follow the target’s flight vector.
The lessons learned from working with modern control theory can be summarised as follows:
The flight control laws and flight control system continued to be tested using the JA 37-21 ESS01 experimental aircraft by introducing mid value select. Because the JA 37-21 ESS01 experimental aircraft was a conventional statically stable aircraft, there was the possibility of flying to a flight condition (altitude and speed) with the mechanical “backup flight control system” connected then disconnect it via a “clutch” system, thereby activating the electrical digital full authority flight control system for testing. A similar concept with the possibility of using a mechanical system as a backup could be found on the Tornado and Concorde.
Between the three-channel asynchronous computers (HDC-301) in the electrical digital flight control system mid value select on inputs from the sensors were used and mid value select on outputs from the digital control system’s commands to the control servos were used. This concept was used to ensure the aircraft could continue to fly in case of a failure and became the concept for failure and redundancy handling used later on the Gripen.
Canards were not used actively in the JA 37-21 ESS01 experimental aircraft. The canards were flapped prior to landing just like for an ordinary 37 Viggen aircraft.
Flight with an unstable aircraft in a small angle of attack region was tested with the JA 37-21 ESS01 to see whether the flight control laws for the electrical digital flight control system could cope. For this purpose, a JA 37-21 with heavy store configuration was used. The tests proved that the flight control laws could handle unstable basic aircraft.
Modern control theory and linear quadratic design were once again chosen to design the flight control laws for the electrical digital flight control system. This technology provided both good pilot handling and flight characteristics, as well as functions for the autopilot altitude hold and course hold.
The JA 37-21 ESS01 had a large control stick with non-linearities such as slack, friction and hysteresis which had an effect on the control characteristics. In addition, mass effects from the pilot’s arm and the control stick destabilized the results even further. On top of that, delays from the servos and trailing edge control surfaces affected both the design of the aircraft and its flight characteristics.
The flight control system used mid value select on the input signals from the sensors and mid value select on the flight control laws’ commands to the control servos.
Preliminary testing was carried out on a simple “breadboard simulator” of the three-channel asynchronous control system.
A “breadboard simulator” is a compact and intuitive program which helps users design and simulate circuits quickly and with the least possible effort.
In the case of digital signals, the mid value select is to use mid signal of three signals. For analogue output signals, the mid value select on the JA 37-21 ESS01 was defined as the selection of an analogue mid signal of three signals.
The placement of the voting planes is important for isolating errors on units in different parts of the system. The voting planes describes which signals are used for the control.
After the first error, the middle value is taken from the remaining healthy three so as to reduce the effect of a possible further failure by 50%. Each signal is monitored internally so that possible similar failures can also be disconnected. By carrying out this preliminary mid value selection in three channels, the requirement for isolating the next failure is reduced to much less than without this selection. This combination ensures sufficiently low probability
To test this, an electrical flight control system with full authority was implemented in the test aircraft (JA 37-21 ESS01). This test aircraft had been used previously for testing the flight control system in the JA37 Viggen.
The mechanical back up system between the control stick and the control servos could be disconnected giving the digital autopilot full authority. During the testing, the mechanical connection could be used by the pilot when necessary and the system availability requirements were reduced thanks to the possibility of possibility.
Take-off and landing were possible to do with the original mechanical system and the electrical system was only connected once at a safe altitude and with the desired speed when the effect of possible problems could be handled by the pilot. This was the same principle used when testing the new control functions in the JA37, boot for the basic control functions and for the automatic sighting.
The three-channel asynchronous system tested in flight in the 37-21 ESS01 was based on three digital autopilots called SA07 from the JA37 Viggen. All communications during flight (e.g. incoming signals from the sensors and outgoing signals to the servo motor) were analogue.
Each channel was isolated so that possible failures in one channel would not affect the others; this analogue communication resulted in comprehensive, heavy cabling between the channels. In the final system on the Gripen, the equivalent communication was digital, reducing the cabling considerably as well as simplifying isolation between the channels and making enhanced communications between the three channels possible.
Digital communication took place between the channels on the ground used for testing before flight and for updating the software in the flight-tested ESS01 system. Before flight, the communication was used to synchronise the tests in the three channels, as well as printing out all the results from the testing.
The same discrete signals were also used to enable concurrent uploading of the software to the three channels and the application of programme changes when testing in the rig and aircraft (so-called patches).
Communication between the channels used two general discrete input and output signals between the control and other channels and was fully implemented in the software. Testing, communications and software handling were based on the solutions used for the JA37 Viggen. Completed test flights showed that the system functioned as expected and could form the bases for the Gripen flight control system.
The handling and flight characteristics of the flight control laws could be tested by a pilot in a JA 37 simulator with cabin. The flight control laws ESS01 did not contain any automatic trimming in pitch and roll, so the pilot needed to constantly use the trim button on the control stick when the aircraft changed speed or altitude. There was no automatic limits in the angle of attack or load factor which could make it easier for the pilot to fly the aircraft. The pilot had to constantly monitor and manage aircraft’s limitations.
The lessons learned from the testing were as follows:
Before the JA 37-21 ESS01’s maiden flight, comprehensive testing was carried out on the ground with the aircraft connected to a closed loop with the simulator computer so that sensor signals could be fed into the aircraft’s sensor inputs. The sensor signals were handled by the flight control system’s digital flight control laws which then commanded the control servos. The position of the control surfaces were measured and entered into a simulator computer with an aerodynamic simulation which then created sensor signals such as load factor and gyro signals to the flight control system.
In this way it was possible to check the sign and size of the aircraft’s signals to and from the flight control system before the flight, as well as to ensure that the mechanical backup system did not interfere with the control servo commands from the electrical digital control system. The mechanical backup system functioned as a continuous back up in case a failure would occur in the electric flight control system command.
The conclusions and lessons learned from testing with Automatic Aiming in the JA 37 Viggen and from the testing of the electrical system with full authority fly by wire were as follows:
The Gripen is a modern statically unstable aircraft with performance which only a statically unstable aircraft can have. The aircraft cuts through the air with little energy loss.
The total lift on the Gripen is a combination of the lift from the wing, the lift from the canards and the lift on the trailing edge control surfaces. Together, they have a lifting force in the same lifting direction on the aircraft, unlike in the case of a conventional statically stable aircraft.
This means that the wing area on a Gripen is less in comparison with a conventional statically stable aircraft where the lift on the trailing edge control surfaces point downward.
The Gripen becomes less statically stable in comparison with a conventional statically stable aircraft at supersonic speeds as the lift of the wing and the canards shifts toward the rear of the aircraft; this means it is subjected to less counter-acting trailing edge control surface movement compared to a conventional statically stable aircraft.
This means that the Gripen does not need to have such powerful and heavy control servos and hydraulic systems when flying at subsonic speeds. The main aim is to have as low a take-off weight as possible for the aircraft itself so that it can carry the maximum payload.
The Gripen is a modern statically unstable aircraft with a duplication time in case of disturbance (e.g. wind gust) to grow to double amplitude of approx. 0.4 seconds without a stabilising autopilot. This should occur if the autopilot stops commanding stabilising commands in which case the aircraft should diverge. The instability of a basic aircraft depends on where it is flying within the flight envelope.
In geometry, an envelope of a family of curves in the plane is a curve that is tangent to each member of the family at some point.
The time taken for a disturbance to double amplitude (duplication time) for the ESS01 with a heavy external stores mounted without a stabilising autopilot was approx. 2.0 seconds.
The Gripen’s flight control system have to command stabilising commands with the least possible time delay and at a frequency of at least 60 Hz and if the amplification is too low on the control system’s stabilisation, the stabilisation will be insufficient. If, on the other hand, amplification is too high, it can make the aircraft dynamically unstable at a higher frequency and, moreover, even result in aeroelastic oscillations (flutter).
Based on the success of the ESS01 programme, the first version of the autopilot for the Gripen– called SA10– was constructed. When the SA10 was being constructed at the beginning of the 1980s, the electronics were very sensitive to ionising radiation. Digital technology such as a computer processor can suddenly stop working due to memory changes.
Especially processors and the memory were sensitive to the technology that was used in integrated circuits at the time; analogue technologies, however, only experienced a short ‘blip’ in the output signal. This additional insecurity in the software was the reason behind introducing an analog back up system which would work independently from the digital control system as a “backup”. Security-critical software was becoming a new area of interest at the time.
The analogue reserve mode used free floating canards (canards floated into the direction of flow of the air) which led to poor flight characteristics. The analogue backup mode also meant that the control servo loops had to be analogue.
All three digital channels in normal mode in the autopilot were in theory the same except for variations in how they communicated with other in the aircraft. Communication with the other systems were distributed among the channels in order to reduce redundancy as best as possible in other systems and sensors on the aircraft.
Flight safety critical sensors such as the angular velocity gyro signals were implemented independently in each channel so that the system would continue to control the aircraft even if one or two channels were to fail. The angular velocity gyro signals, angle of attack sensors and accelerometers are required to stabilise the aircraft. The pilot stick command from the control stick is necessary to fly the aircraft.
Control of the control servos to the control surfaces was implemented in all three channels for the same reason. Two installed accelerometers and one accelerometer signal from the Inertial Navigation System (INS) and angle of attack sensors and the output from both is compared so as to disconnect incorrect signals before they can lead to serious problem.
All incoming signals are shared between the channels via CCDL, (Cross Channel Data Link), so that all of the incoming signals are available in all the channels. This also means that the same software can be used in all three channels. Adaptation was needed to fit with the aircraft’s other systems (air data, system computer and navigation) which were not three-channel.
One of the three channels communicated with these systems, with the information being passed on to the others via CCDL. If there is an error in the communicating channel, one of the others can take over to ensure full performance with just two working channels.
The gyros, accelerometers and angle of attack sensors are used to stabilise the aircraft. There are two angle of attack vanes on each side of the nose. They measure the airflow on the aircraft and its variation in the form of external disturbances such as winds and gusts. This information is then sent to the autopilot which can command the control surfaces. The angle of attack sensor is the most useful sensor to measure changes in the airflow and its effect on the aircraft.
Speed information from the pitot tube total pressure and static pressure measurement and altitude information from the pitot tube’s static pressure measurements are necessary to stabilise the aircraft in any speed or altitude within the flight envelope. Redundant flight information is received from a fin pitot tube and the Inertial Navigation System. Flight information comes from the nose pitot tube and the pitot tube on the fin. The flight information incoming from the pitot tube on the fin then becomes a secondary (redundant) source of information. The same information from all sensor signals are fed into the control system’s three-channel computers.
The sensor information is used by the flight control system to give the aircraft the desired handling and flight characteristics. The flight control law commands the canards, trailing edge control surface, rudder, leading edge flap and air brake. Each computer puts out control servo commands to the respective control surface after a mid-value selection.
The canard is controllable and quickens the aircraft response speed to the desired level. The trailing edge control surface provides the main aerodynamic moment for the aircraft’s rotation in pitch and roll whereas the leading edge flap ensures that the airflow over the wing remains optimal at all times. Differential trailing edge control surface command results in a roll.
The rudder take care of the aircraft’s yawing movement and also if a roll is commanded to the trailing edge control surfaces the rudder will be commanded to minimize the sideslip. If a failure with the aircraft occurs, a combination of the above can provide the best possible performance in case of a failure. The control system has excellent adaptability to deal with any possible failure situations.
An accident took place in August 1993 during the Stockholm Water Festival.
Investigations of the crash memory and film showed that the accident was the result of pilot-induced oscillations (PIO). The category III PIO was caused by non-linear elements in the flight control system together with the undamped control stick.
Saab contributed to the investigations regarding the accident and a “construction group” was formed in order to find out what had happened and to find a solution to the problem. A number of foreign investigators monitored Saab’s progress.
Categories of pilot-induced oscillations (PIO)
PIO can occur when the pilot is required to make control tasks requiring high levels of concentration and this increases the pilot gain considerable for example at landing, flying in formation, flying low or performing aiming tasks.
In PIO category I, the control system is too sensitive for the pilot to be able to carry out the control task. The sensitivity does not depend on the magnitude of the command applied by the pilot, however it can vary depending on the frequency with which pilot commands occur.
In the case of PIO categories II and III, the linear sensitivity of the control system may be correct and without a tendency for PIO until the magnitude and/or the rate of the pilot’s control signals reaches the non-linear element in the control system. The pilot influence on the stability in the loop can alter dramatically.
As Saab was the first in the world to have a modern statically unstable aircraft as part of an air force, we had to investigate the problems which can arise with this kind of aircraft. We came up against the problem with the conventional rate limitations for control surface commands from the pilot and feedback. The system can become unstable if the loss of phase, i.e. the time delay, is too large for the corrective control command from the flight control system.
In the control system version which was in use at the time of the crash at the Water Festival in Stockholm in 1993, conventional rate limiters from the pilot command and stabilising commands were used. The pilot with the undamped control stick could easily create fast commands which resulted in saturation of the pilot command and stabilisation command in the conventional rate limiters to the control servos. This resulted in large time delays which meant that the aircraft became unstable and ultimately crashed.
The control and stabilisation commands required for a statically unstable aircraft became delayed and resulted in “non-linear” PIO.
Up until that point the only investigations had concentrated on linear PIO and there are many theoretical criteria to check against the tendency for PIO (e.g. Neal-Smith, Gibson, etc.). There were no theoretical criteria which dealt with PIO on non-linear elements in a flight control system.
Lessons learned from the problem with PIO were as follows:
Conventional rate limiters could not be used in the flight control laws in a modern statically unstable aircraft as a result of their large loss of phase and resulting time delay when the pilot commands were so large and rapid that they hit the conventional rate limiter.
The control and flight characteristics of the Gripen within the linear area were very good. This is when the command from the pilot and the stabilising feedback signal are not so large and rapid to be limited by conventional rate limitation.
In order to start test flights as quickly as possible and give the flight test operations the chance to test out other systems in the aircraft, flight control law versions were made with reduced performance and lower load factors and angles of attack limits to avoid to hit the conventional rate limiter. This was done to save time and to have the time and possibility to address the problem with the large loss of phase and resulting time delay of conventional rate limiter.
It did not take long to discover the cause of the accident: it was a combination of the conventional rate limiter on the pilot’s control command signals to the trailing edge control servo and to the canards and conventional rate limiter on sum of the pilot’s commands and stabilising feedback command.
The Gripen was not the only aircraft suffering a PIO incident, however the problem with a modern statically unstable aircraft combined with a conventional rate limiter was new. The construction group investigated the effect of the non-linear conventional rate limiters that were implemented into the control system.
The undamped control stick made it possible for the pilot to command large rapid commands. This led the pilot and stabilization feedback command signals into a rate limitation and thus caused the accident.
Was it possible to construct a rate limiter that would limit the rate of the pilot’s control servo commands and the stabilization feedback command signals without the major loss of phase and time delay?
The handling and flight characteristics within the linear range were very good, i.e. when the pilot’s control servo commands and the stabilization feedback commands were not rate limited.
The difference between the input signals and output signals of a conventional rate limiter was fed back via a constant and a filter, and was subtracted from the input signal to the rate limiter.
The input signal to this conventional rate limiter was thus the difference between the input signal and the signal which was fed back. The output signal from this adjusted conventional rate limiter then passed through a further conventional rate limiter.
The graph below shows the difference between a conventional rate limiter (red) and the difference in time delay between it and the patented phase-compensated rate limiter (blue). The input signal to each rate limiter is shown in green.
The purpose of the rate limiters is to limit the command rate to the servos so that the servos always are able to carry out their commands at different hydraulic pressures. The servos cannot be commanded in acceleration limitation because this leads to major time delays and losses of phase in the closed loop system’s stabilisation and pilot loop.
This is why the phase-compensated rate limiters for the pilot commands were made depending on hydraulic pressure. If the hydraulic pressure in the system was low due to many frequent commands from the pilot, the pilot’s command put out a lower control servo command rate so that the stabilising inner loop always had the authority to stabilise the aircraft.
With the improved phase-compensated rate limiter introduced into the control system on the pilot command and the sum of the pilot command and stabilising feedback command, it remained to be demonstrated that the control system and flight control laws would not cause a category III PIO or any PIO at all.
Comprehensive theoretical testing followed with full testing with a pilot in a very realistic MAHS simulator with a hydraulic system, control system and other realistic hardware. The pilot attempted to make the aircraft diverge with different combinations of pilot command in pitch, roll and yaw.
The pilot’s commands were then optimised and mechanised in the software so that batch simulations could be carried out. This digital mechanisation was given the name “clonk simulation” as a result of the “clonking” sound made during pilot simulations in the MAHS resulting from the pilot’s commands hitting the metallic stop on the control stick.
These calculations, together with theoretical studies, showed that Gripen’s control system was resistant to PIOs. Saab is the world leader in the causes of PIOs and how to correct the problem. Furthermore, the phase-compensated rate limiter was patented, as it is unique in the aviation world and is a very effective rate limiter with a low time delay.
In older aircraft such as the Saab J 21, Saab J 32 Lansen, Saab J 35 Draken and Saab AJ/JA 37 Viggen, the pilot had to constantly monitor the angle of attack in order to avoid entering a stalled or uncontrolled manoeuvre.
It is important not to get into a stalled manoeuvre at low speed where the airflow separates over the wings, etc., as this can result in a loss of control: “departure”.
So as not to overload the aircraft’s structure, the load factor must also be limited etc. at the same time the pilot is achieving his tactical mission.
The pilot can fly the aircraft to its limits without having to worry about angle of attack or load factor limits, sideslip, motor intake limitations or the risk for stalling or departure.
The image below shows the automatic limiting functions included in the Gripen’s care-free manoeuvring.
The JAS 39 Gripen has an automatic load factor limitation depending on the speed, aircraft weight and external stores mounted. Light aircraft have a maximum load factor limit of around 9G, whereas a more heavily loaded aircraft will have a lower limit depending on its structural limitations.
The Gripen also has automatic angle of attack limitation to prevent stalling or departure. Similar to the load factor, a light aircraft have a higer angle of attack which is reduced for aircraft carrying heavy external stores.
If the aircraft is to make a vertical climb with speed reducing slowly and the angle of attack increases over the limit, the Gripen has a “high angle of attack recovery function” which constantly returns the aircraft to the normal flight envelope.
In addition, the Gripen also has automatic turn coordination which ensures that the sideslip remains low during a roll command or combination of pitch and roll command; this means the load on the aircraft fin remains low and under the permitted limit. The lateral force on the fin must not exceed a level so that the strength of the fin is endangered.
When the pilot issues a roll command with the control stick, the trailing edge control surface moves and causes a rolling moment; at the same time, the rudder is controlled automatically by the pilot’s control stick command via gain called the aileron to rudder interconnect (ARI). This is used to ensure as little sideslip as possible during the rolling manoeuvre which also places less pressure on the fin.
The allowable roll angular velocity depends on the load factor operated by the pilot. This prevents the structure from being overloaded.
As the trailing edge control surfaces are split, it is also possible to control the distribution of lift across the wingspan. The best situation is one in which the lift distribution is as close as possible to the fuselage body as this increases the life length of the aircraft.
The need for larger control surface movements is more pronounced at lower speeds, the load on the canards is automatically limited to its maximum hinge moment and lift on the canards and thus the canards cannot be used to full extent at supersonic speeds. This automatic limitation on the canards when flying at supersonic speeds is also set in the flight control laws.
Similarly, the maximum trailing edge control surface hinge moment at supersonic speeds is also limited automatically for a load factor command.
The leading edge flap improves airflow over the wing giving the aircraft a higher angle of attack limit and therefore improved turning performance. The leading edge flap in combination with the canards and trailing edge control surface also provide the aircraft with lower drag in all flight conditions and for any pilot command.
The automatic trims in pitch, roll and yaw keep the aircraft trimmed in all flight conditions and for all possible external stores, even asymmetrical stores. In older aircraft such as the JA 37 Viggen, the pilot always had trim it manually using a trim control button on the control stick whenever the aircraft changed speed or altitude.
The Gripen has a damped central control stick with a maximum command back to soft stop of 10.8 degrees. If the pilot moves the control stick to this limit with the aircraft flying at over 600 km/h, this means a load factor command to its load factor limit.
The load factor limit depends on how heavy the aircraft is and what external stores are mounted. If the pilot commands the controls stick to -7 degrees and the aircraft is flying faster than 600 km/h, the pilot commands the aircraft to its negative load factor limit. The pilot is also able to achieve approx. 3G extra at speeds over 600 km/h by pulling the control stick even further back from the soft stop to the hard stop.
If the aircraft is flying below 600 km/h and the pilot moves the control stick to the soft stop at 10.8 degrees, the aircraft will be flown at the maximum angle of attack limit. The angle of attack limit depends on the external stores hanging on the aircraft. With the control stick command at -7 degrees, the aircraft will be at the lowest angle of attack of -9 degrees.
Flight testing of the Gripen’s care-free manoeuvring properties with automatic limitations was carried out in the later half of the 1990s.
A very large number of combinations of pilot commands against all the different automatic limiting functions were simulated in a 6DoF model (ARES) of the Gripen to ensure that Gripen really is “care-free”. The number of combinations of pilot command, all different external stores, flight conditions, etc. was large to prove that the Gripen is care-free to fly.
The angle of attack, sideslip, load factor, roll rate velocity and lateral load factor from the multiple simulations of different pilot manoeuvre combinations and external stores were given in a diagram in order to classify external stores with the correct alpha beta limits or load factor vs. roll rate limits.
The combined pilot pitch/roll/yaw commands with the worst results were selected for simulation with a pilot in a simulator.
The pilot commands consisted of different pitch, roll and yaw commands in different time sequences and for different external stores combinations. The combination of pilot manoeuvres which gave the worst results were selected and entered into a so-called “knäblock” for flight testing. The “knäblock” is mounted on his knee and is a testing schedule used by the pilot during flight.
Before the flight, the “knäblock” was simulated in a so-called STYRSIM simulator which had a very realistic outside world presentation. With the differing pitch, roll and yaw command sequences, it was very important that the pilot carried them out in the correct time sequence which could be exercised in the STYRSIM before flight. After the test flight, the results could then be compared to the simulator flight in STYRSIM.
In addition, the flight test results were also compared with simulations in a 6DoF model of the Gripen called ROMAC running in parallel to the flight testing. This real-time model for aerodata checks was fed with the actual flying pilot’s commands during the flight. ROMAC was synced in real time with the flight test and was fed with current speed and altitude as well as the pilot’s actual control commands throughout the flight.
This means that the aircraft’s actual behaviour could be compared with the behaviour stored in the ROMAC; as a result, the simulation models could be updated with the aerodata wherever necessary.
The start of the 2000s saw an increased need for flights involving very heavy external stores resulting in very aft centre of gravity positions. The very heavy external stores was required by new customers. Many of these new heavy external stores moved the centre of gravity very aft and this affected the flight characteristics negatively – especially at higher altitudes.
To solve the problem, the flight testing was carried out with a “flight test function” which added different levels of alpha feedback gains to the flight control laws in order to improve stability and, as a result, the handling and flight characteristics.
Analyses and flight testing showed that this made it possible to move the heavy aircraft’s poles towards a lighter aircraft’s poles, thereby making the handling and flight characteristics more like those of a lighter aircraft’s, with lower load factors and alpha limits set for the heavy external stores. The results from the flight testing proved that this method was viable.
The high angle of attack recovery (HAoA) function is an extension of the flight control system’s angle of attack limitation. If the aircraft has a low speed and high nose angle, the angle of attack will be low at first but will increase and exceed the angle of attack limit as the speed drops further. The HAoA function and command will return the aircraft to the normal flight envelope in a controlled manner.
In the normal flight envelope, the angle of attack from two alpha sensors on the sides of the nose is used in pitch; in addition, the pitch gyro signal is used to measure the angular rotation of the aircraft. The accelerometers measure the normal acceleration along the Z-axis; these signals are used to control and stabilise the unstable aircraft.
In the lateral channel, the gyro signals in the roll and yaw were used to measure the angular rotation of the aircraft. The lateral accelerometers measure sideslip.
If the aircraft has a high nose angle at low speed and exceeds the angle of attack limit, only signals for the angle of attack and pitch rate gyro are used in the pitch.
Better lateral stabilisation is required in the angle of attack range between 30 – 45 degrees. This means that the sideslip needs to be measured from a sensor located under the nose in addition to the roll and yaw gyro signals. At angles of attack over approx. 55 degrees, the sideslip can no longer be defined.
The figure below shows the coefficient in pitch moment coefficient as a function of the angle of attack for the max nose up control surface command and max nose down control surface command. As can be seen from the diagram, the aircraft is stable if the angle of attack is greater than 55 degrees; the same is valid on the inverted side. It also shows the use of sensors for stabilisation at high alpha ranges.
The most common high alpha situation is a combination of nose-high attitude and low speed which results in the speed slowing down and the angle of attack increasing, and the HAoA function will return the aircraft to the normal flight envelope.
If the aircraft should still get into a high alpha spin situation where it rotates around a vertical axis, the HAoA function will first detect this and then centre the canards and trailing edge control surfaces in order to use maximum roll command to stop the spinning rotation. After that, the recovery pitching command is commanded for the canards and trailing edge control surfaces until the aircraft returns to the normal flight envelope.
Some form of safety device was required to carry out high alpha testing with spin, etc., beyond the normal flight envelope. A large spin Shute which could take down the whole aircraft was too large and complex. As a result, a pitching moment chute was mounted on the back of test aircraft. The moment chute resulted in a nose down pitch moment so that the aircraft nose dropped, the speed increased and the aircraft could return to the normal flight envelope. The safety chute was to be used if there was a failure on the aircraft during flight testing. The safety chute detaches as soon as the speed has increased and the aircraft has returned to normal flight within the flight envelope.
The fact that the safety chute needed to be mounted on the aircraft also depended on the fact that the aerodynamic data at high angles of attack was very uncertain. In addition, the aerodynamics depends on the dynamic process involved in going into and coming out of high angles of attack.
Flight control laws were created to return the aircraft to the normal flight envelope from a high alpha situation based on the available aerodynamic information used in the ARES simulation model. The flight control laws required constant updating as the alpha flight testing progressed.
During the later half of the 1990s, flight testing started on a test Gripen aircraft. First of all the safety chute function testing was carried out to ensure that it could develop correctly and that it detached at the correct speed.
After that the actual high alpha testing began with stalling the aircraft from different pitch attitude angles. With the experience from this testing, it was possible to adjust the flight control laws and implement improved lateral stabilisation in the alpha range between 30-45 degrees. This was achieved by using the aircraft sideslip sensor as a feedback signal to the control surfaces with suitable gain.
The next stage was to move onto high alpha ranges with an alpha over 55 degrees where the aircraft is stable. See figure above.
For this purpose a flight test control law was created in the flight control system whereby the pilot’s command was sent directly to the control surfaces via a gain without stabilising feedback command signals. This mode was called the Direct Link Mode (DLM) and was a preliminary flight test function.
By carrying out a steep climb with a high pitch attitude, the pilot was able to bring the aircraft into a high alpha range over 55° and then connect the DLM and remain stable in this alpha region.
The Gripen was painted black with a white nose, canards and dorsal spine so that it could be filmed from the ground during testing. The flight testing took place at a satisfactory altitude of approximately 11,000 m.
The next stage was to test whether the flight control laws could stop an established spin and then command a return command to the normal flight envelope; the pilot flew the aircraft with a high nose position and bringing it into an angle of attack over 55 degrees. After that the test function DLM was connected, whereby the control surfaces were centred and the pilot controlled the aircraft while it still was in an angle of attack position greater than 55 degrees. In this position, the basic aircraft was stable and so could be controlled simply with the pilot’s incoming command to the control surfaces.
The pilot then started a spin using the control surfaces until reaching a 50 degree per second yaw rate velocity at which point the normal flight control system and the HAoA function were connected. The HAoA function automatically stopped the rotating spin and then commanded the control surfaces to return the aircraft to the normal flight envelope.
The high alpha testing continued without a spin Shute mounted on the test aircraft and then with an aircraft similar to production line without any shield.
Testing was also carried out on a two-seater Gripen aircraft. This aircraft has a longer nose than the one-seater version and so is easier to get out of a high alpha situation.