Lockheed Martin has dominated the tactical transport mission in Western air forces virtually since the C-130's first flight in 1954, with more than 2,500 Hercules delivered to multiple customers, including Brazil.

As the Brazilian air force began searching for a replacement about a decade ago, Embraer launched studies for what was to become the KC-390. Initial studies posited a straightforward freighter derivative of the E190 regional jet. But with its potential domestic customer's demand for a new aircraft with performance at least on a par with the Hercules, the company scrapped this idea and launched a clean-sheet, jet-powered design.

A $1.6 billion contract awarded to Embraer in 2009 by the Brazilian air force launched development of the first two KC-390 prototypes. Its first test aircraft flew in February 2015, followed by the second 14 months later.


Marcio Jumpei

Since the type's debut sortie, Embraer has conducted an aggressive flight-test campaign. The two prototypes have now accumulated more than 1,400h in the air, many times achieving two flights on a single day. Meanwhile, the company has completed static loads testing, with a ground test article reaching 150% of limit loads with no failures. Fatigue testing is set to begin shortly.

The Brazilian air force’s development strategy called on Embraer to certificate the KC-390 twice. The basic platform is expected to achieve airworthiness certification by Brazil’s national civil aviation authority ANAC by the end of this year, followed by a certification of military-specific qualities – such as cargo drop and aerial refuelling – in 2018.

In an early October interview, Paulo Gastão Silva, vice-president of the KC-390 programme, said no major issues had been encountered during development. However, our flight test was performed before a 12 October event during which a KC-390 returned to base early after encountering higher-than-planned altitude loss during a stall test.

The initial portion of my evaluation flight was spent in Embraer’s flight control development simulator located near São José dos Campos. My host, senior experimental test pilot Eduardo Alves Menini, concentrated on the KC-390's fly-by-wire (FBW) flight control schemes, envelope protections and the innovative active side-stick (SS) during our nearly 2h session.

KC-390 flight deck

Marcio Jumpei

The initial KC-390 concept proposed using significant components from E-Jet-series aircraft. While this approach was later abandoned, it did not stop Embraer from leveraging technology from its other aircraft. The company had put a lot of effort into developing an unique FBW control scheme for the Legacy 500 business jet family. I was fortunate to fly both the Legacy 450 and 500 two years ago, and found they were a delight to fly. The FBW control scheme for the KC-390 is largely derived from the Legacy’s system, with modifications to suit the tactical nature of the aircraft.

As with the Legacy’s FBW system, the KC-390 approach has two main modes: Normal and Direct. Normal, as the name implies, is the primary operational mode, with Direct as a back-up. In Direct mode, flight control laws (FCL) in pitch, roll and yaw are “stick to surface”, with speed-based gains. In Normal mode during up-and-away flight, the pitch axis FCL are Gamma Dot, pitch rate command.

In most instances, a Gamma Dot scheme approximates a g-commanded style. In roll, the Normal mode FCL are pure roll-rate command, while in yaw they are side slip (beta command). The Normal mode FCL are designed to keep the KC-390 within its “Normal” flight envelope (NFE). Embraer defines this as speed between 1.13VS and maximum operating limit speed (VMO), and less than 45° of bank. Flight outside the NFE is allowed, with a number of envelope protection schemes installed: in pitch, angle of attack limiter, high speed protection, load factor limiter and low speed functions; in roll, soft bank limitations; and in yaw, beta limiter and thrust asymmetry transient reducer.

The KC-390’s envelope protection features are the same as the Legacy’s, except for two notable changes. In the Legacy aircraft pitch attitudes are limited to between +30° and -15°, while the KC-390 has no defined pitch limits. With enough energy and altitude, the KC-390 could be looped. In roll, the Legacy will maintain bank angles up to 33° with no SS pressure. Reflective of the tactical nature of its missions, the KC-390 demonstrates neutral spiral stability at bank angle up to 45°. Like the Legacy, the KC-390 can do a complete aileron roll.

While the KC-390 FCL logic is very similar to the Legacy’s, there is one notable difference in flight control system architecture. For redundancy purposes the Legacy has three hydraulic systems; an approach Embraer initially chose for the KC-390. After studying vulnerabilities to live fire, Embraer determined that three hydraulic systems might not provide the needed levels of redundancy. The KC-390 has two hydraulic systems, with electrohydrostatic actuators on control surfaces providing the needed back-up.

Once seated in the simulator’s left seat, I adjusted the side console-mounted wrist rest so that the SS felt comfortable. Additionally, the outboard seat armrest can be adjusted to facilitate SS usage. The SS has three switches located above the grip: touch control steering (TCS), trim hat and autopilot (AP)/PB disconnect. The TCS switch is used primarily to set the trim reference speed when in a landing configuration. The trim hat is like that found on many control sticks, allowing hands on throttle and stick pitch and roll trim inputs. The red autopilot controller push button was used to disengage the autopilot as well as cut out control inputs and take command from the other SS. The stick's longitudinal axis is parallel to the aircraft’s, with the lateral axis perpendicular to it. Some SS I have flown with have their axes rotated slightly inboard, but I found the KC-390’s arrangement quite satisfactory.

The SS displace 15˚, plus or minus, in each axis and are “active”, with servo motors available to drive them. Because they are back-driven, the force gradient can be changed to suit the task, offering some modicum of artificial feel. Currently, however, Embraer has chosen a fixed force gradient throughout the SS range of motion. The other innovative feature of the back-driven SS is the ability to electronically connect them. Moving the left SS causes a corresponding motion in the right SS, and vice versa. I would say this is revolutionary, but conventional flight control aircraft have always done this.

The advantages offered by the active SS became clear during our simulated air-to-air refuelling (AAR) exercise.

After rendezvousing with another KC-390, I tried to stabilise just behind the refuelling drogue extended from the left-hand Cobham refuelling pod. I could easily remain stable fore and aft, but lateral stability proved challenging. We slid from side to side, almost wingtip to wingtip, as I tried to stabilise behind the basket. After working up a sweat, I backed out of the approach and stabilised about 100m (328ft) behind the tanker.

At that point, Menini selected the flight controls to air refuelling mode. Like many aircraft, the KC-390’s default FCL are not optimum for air refuelling. In AAR mode, pitch and roll axis FCL were tweaked for the fine tracking task of air refuelling. As we again advanced toward the basket I did not perceive any differences in the pitch axis, but it was an almost opposite feeling for lateral tracking.

While I was certainly no “Ace of the Base”, with AAR mode engaged I could approach and make controlled stabs at the basket. After a few solo attempts Menini joined me on the controls. While DUAL INPUT sounded intermittently, I could feel his stick inputs as we closed and made several solid contacts. Having been an instructor pilot, Menini’s ability to feel my control inputs greatly enhanced his ability to teach me how to air refuel the KC-390. With the air refuelling exercise complete, we rounded out the simulator session with visual approaches to normal landings.

Our preview fight took place at Embraer’s Gavião Peixoto (SBGP) facility. Poised on the ramp, the KC-390 looked large and solid, its 6˚ anhedral-swept wings giving it the look of a bird of prey mantling over a kill. I accompanied flight test captain William Souza as he performed the preflight exterior inspection on the first prototype, PT-ZNF. This is production-representative with respect to aerodynamic and propulsion characteristics. The aircraft’s cargo hold, however, was not finished and had tonnes of flight test equipment installed. The exterior was festooned with radar warning and infrared countermeasure receivers, confirming that this was a military transport.

The nose and main landing gear had large balloon tires, designed to allow the KC-390 to operate from unpaved airfields (CBR 4 or higher). I remarked to Souza that the main landing gear sponsons appear to be a bit oversized. After we entered the aircraft and ventured into the expansive cargo hold, the reason for the oversized sponsons was clear. The cargo hold has constant width of 3.45m (11.6ft) throughout its length, the retracted gear not impinging on the cargo hold. The cargo floor features longitudinal floor strips that reverse to reveal cargo rollers.

While serving in the US Air Force I had been a parachute-qualified forward air controller, and had trained with the army as a jump master. It was with this in mind that I approached the KC-390’s aft crew doors. I was impressed by their size: 0.95m wide and 2.01m high. Outside the door, a large step is cut out of the sponson profile – one that appears to offer a stable platform from which to shout “Geronimo!” and leap into the slipstream. With my parachute flashback complete, I closed the crew door in preparation for opening the main cargo ramp and door. Souza talked me through their operation; as the long ramp lowered and door retracted overhead. The ramp is quite long, having space for two standard-sized pallets. Its length also gives the KC-390 favourable ramp angles, easing the loading of large cargo items.

After my familiarisation of the cargo compartment was complete, I climbed up the stairs to the KC-390’s flight deck. Once comfortably seated in the left-hand seat, I checked out the expansive field of view outside the KC-390’s six framed cockpit windows. In addition to their large size it is notable that the windscreen and cockpit side windows are designed to protect against ballistic impacts, such as 7.62mm bullets. The avionics suite is Rockwell Collins' ProLine Fusion, similar to that found in the Legacy 500 family. The KC-390’s has five, 15.1in high-resolution displays: four on the instrument panel, with the fifth on the forward portion of the centre console.

The KC-390 comes standard with two large 42° x 30° field-of-view head-up displays, which were inoperative for my flight. As with the Legacy, each pilot had a scratchpad, keyboard and cursor control device. Logically formatted system sub-panels are located on the overhead panel. Overall the flight deck is well arranged, employing a “quiet and dark” philosophy.

From the right seat, Souza guided me through process of initialising the flight management system, which is straightforward and logical, just as I had found in the Legacy. Also joining us on the flight deck and acting as a safety pilot was test pilot Marcelo de Lima Camargo. When cleared by ground personnel, both of the full authority digital engine control-equipped International Aero Engines V2500-E5 engines started using bleed air from the upper fuselage-mounted auxiliary power unit.

KC390 V2500 engine

Marcio Jumpei

The start sequence was essentially automatic, with each engine reaching IDLE in about 75s. Souza set the hydraulically actuated flaps to position 2, extending the slats and setting the flaps to 20°. A slight bump on both thrust levers got the airlifter moving. Tiller-controlled nose-wheel steering (NWS) allowed me to negotiate the two 90° turns as I aligned the KC-390 for takeoff on runway 20.

With a gross weight of 68,500kg (151,000lb), including a 13,120kg fuel load, take-off speeds were 108KIAS (200km/h) for abort, 122KIAS for rotation and 125KIAS for rotation if one engine fails. I advanced the TL, and the auto-throttles (AT) engaged and set take-off power at 1.595EPR. Pedal-controlled NWS allowed me to easily track centerline, and when Souza called “Rotate” about 5kg of aft stick pressure was needed to set a lift-off attitude of 7-8˚. Had I grossly overshot the desired take-off pitch attitude, the tail-strike avoidance (TSA) feature of the FCL would have kicked in to limit pitch attitude to 13° whilst on the runway: 2° below the physical limit. Operation of the TSA is transparent to the pilot, but it can be defeated by an overly aggressive pull.

Once airborne, Souza retracted the landing gear and flaps as an initial climb speed of 200KIAS was established. Passing 10,000ft, a climb speed of 250KIAS was captured. I hand-flew the KC-390 during the climb to 25,000ft; well below the KC-390’s ceiling of 36,000ft. During the climb I did bank-to-bank rolls to assess lateral directional flying qualities. I found response in roll was crisp, and that desired angles of bank up to 45° were easily captured and held. With about two-thirds lateral stick defection, roll rates of 15°/s were noted. As with the Legacy, the KC-390’s flight control system decouples roll and yaw axes. Turn co-ordination is automatic, allowing roll manoeuvres to be executed with my feet flat on the floor. With the aggressive rolls complete, I levelled the wings and released the SS. Next, I gently put in about half left pedal, with the KC-390 yawing to the left while the wing remained level. Souza next selected the “Direct” mode of the flight controls. This time when I put in left pedal the wing initially dropped slightly to the right, before it reversed and rolled to the left, as would be expected for a conventional flight control system.

While still at altitude we slowed to 220KIAS, where Souza extended the flaps to “2” to set up for a 2g turn. I rolled the aircraft and let the nose drop slightly as I increased back pressure on the SS. While we were only pulling to 2g, I found stick force per g to be linear. During the pull I especially liked the g-meter shown in the primary flight display. After recovering to wings level flight the flaps were retracted and AP and AT engaged for a long-range cruise performance point. At M0.65 a total fuel flow of 3,120kg/h gave a true airspeed of 402kt. Indicated airspeed was 271kt on this slightly hotter-than-standard day. After the flight, Sozua stated that M0.72 at 30,000ft would have been a more representative cruise condition for the KC-390. On a standard day, total fuel would be about 3,200kg/h with a true airspeed of 425kt.

With our work at altitude complete, we received air traffic control clearance to descend to medium altitude to explore the KC-390’s slow speed handling qualities. The descent gave me the opportunity to explore the aircraft’s high-speed protection features. With the power set to mid-range, I lowered the nose and accelerated the aircraft past the VMO of 300KIAS. At 304KIAS, “high speed” was sounded to alert me of the overspeed. At 306KIAS the overspeed protection feature kicked in and decreased the pitch attitude to slow us below VMO. This protection feature can be overpowered, and I added forward stick pressure to keep the speed above 306KIAS. At speeds above VMO/MMO the FCLs limit maximum bank angle to 45°, and a full left lateral SS input confirmed this feature was active.

While still keeping forward pressure, I centred the SS. In the normal flight envelope the aircraft would have stayed in a 45° bank, as the SS commands roll rate. In an overspeed condition the FBW system will roll the wings level to help slow the aircraft into the normal flight envelope. Satisfied that the high-speed protections worked as advertised, I next slowed the KC-390 to 300KIAS, where I did a series of sharp pedal and SS inputs. Aircraft response to each individual input was well damped, with no residual oscillations.

While in the simulator I had practiced several descent profiles. The first was an emergency descent to simulate a depressurisation incident. At 300KIAS with speed brakes extended, the KC-390 was able to maintain 7,000-8,000ft/min in the descent from 36,000ft to 25,000ft. To simulate an arrival at an airport where hostile fire was a concern, Menini recommended we do a tactical descent.

After selecting “Max Effort” on the Master Mode switch, the flaps were set to the “S” position, which extended the slats. Starting from 25,000ft at IDLE power and speed brake extended, holding a steady 300KIAS the KC-390 hurtled towards the ground at 11,000ft/min. If the tactical profile involved spiralling down over the landing field to avoid small arms fire, the added g would have increased the descent rate even more. All the while the KC-390’s FBW control system ensured the manoeuvre could be performed in a carefree manner, allowing the crew to look outside for threats.

After the high-speed investigation was complete, a more typical descent was performed at 250KIAS to 19,500ft for a look at the KC-390 in the slow speed regime. The first manoeuvre was an approach to stall in the landing configuration, with gear down and flaps “Full”. The “Full” position is the same as “4” (40° flap extension), except that you are telling the aircraft you are going to land. The AP was engaged and TL retarded to set a deceleration rate of 1kt/s. Approaching the top of the yellow band on the airspeed tape (104KIAS) the AT woke up and re-engaged to keep us from slowing further. Next, I disengaged the AT and pulled full aft on the SS. The tactical airlifter settled into stable, wings-level descent. Even at 99KIAS aircraft response to small control inputs was smooth and predictable. Recovery to the level flight was accomplished by advancing the TL to the take-off/go-around (TO/GA) setting (1.54EPR) and retracting the flaps to “2”.

KC-390 cargo hold

Marcio Jumpei

Once established in a climb at 130KIAS, Souza simulated an engine failure by pulling the right TL to IDLE. I kept my feet off the pedals to observe how the KC-390 responded to a power loss. As with the Legacy I had flown earlier, the response could not have been more benign. The right wing dropped several degrees and the nose yawed slowly to the right. Sensing the engine failure, the slip indicator in the primary flight display became a beta target, and directed side slip for optimum single-engine climb performance. Only a light amount of left pedal pressure was needed to capture and track the target beta. Leandro Gustavo Vieira Bigarella, the flight test engineer onboard our flight, reported that actual rudder deflection was less than 4°: well under its maximum of nearly 30°.

After experiencing the high and slow speed regions of the KC-390’s flight envelope, it was time to return to Gavião Peixoto to see how it handled in the landing pattern. All approaches would be flown with flaps set to “Full”. Average gross weight for the approaches would be about 65,500kg, which gave a reference speed off 122KIAS and target speed of 127KIAS. Souza loaded the RNAV (GNSS) for runway 02 on our first approach. I let the AP fly the approach with AT engaged until we reached the final approach fix, 1,600ft above the touch-down zone, where I clicked them both off and hand-flew the approach.

In the landing configuration (Flaps Full), like the FBW Legacy, the KC-390’s pitch flight control laws change to add artificial speed stability. SS deflection still commands pitch rate, but there is no auto-trim to maintain the new pitch attitude. Pushing the TCS button on the stick sets the new trim reference speed (TRS), displayed as a green arrow on the speed tape. After we were fully configured on final, I used the TCS button to reset the TRS as the KC-390 slowed to 127KIAS. One nice feature the KC-390 has is that the SS-mounted trim hat can also be used to set and change the TRS. While noticeable and positive, I found the KC-390’s speed stability on final was not as strong as the FBW Legacy’s.

As I had practiced in the simulator, at 50ft radar altitude I raised the nose from its 4° approach attitude to the flare attitude of 7.5°. Once the pitch attitude was set, I retarded the TL to IDLE. The KC-390 gently settled on the runway, with a slight relaxation of SS back pressure lowering the nose wheels to the tarmac. In the three-point attitude Souza set the flaps to “2” and I advanced the TL to the TO/GA position. As we accelerated through 127KIAS he called “Rotate.” As the pitch attitude approached 8°, the KC-390 lifted off the runway. Before I could call “Gear Up” he pulled the right TL to IDLE. As the nose swung to the right, I smoothly applied left rudder to capture the beta target in the primary flight display. As I had seen earlier at altitude, very little pedal pressure was needed to maintain the optimum side slip. During the climb to pattern altitude, I used the console-mounted rudder trim switch to trim out the pedal forces. Once level on downwind I centred the rudder trim, as that is how I typically fly simulated engine-out approaches.

Flaps were selected to “3” turning base, with the landing gear extended and flaps set to “Full” on the turn to final. Single engine approach speeds were the same as those for the all-engines operating approach. Very light rudder pedal was needed on final to maintain co-ordinated flight at 127KIAS. I flared at the same altitude I had on the first approach, but delayed retarding the good engine to IDLE for a few seconds. As the engine spooled down, left pedal pressure was relaxed to keep the nose aligned with the runway at touchdown. After lowering the nose to the runway, Souza again set the flaps to “2” and I advanced both TL for the take-off.

Once level on downwind with the gear retracted and flaps still at “2”, Souza selected the KC-390’s back-up flight control laws for an approach in Direct mode. I would now be flying the aircraft without any advanced FBW features. Pitch changes caused by landing gear and flap extension to “Full” were easily countered with the stick-mounted trim switch. On final approach in Direct mode the KC-390 felt stable, and to be quite frank not much different than my first approach in Normal mode. I kept the same flare picture as the prior approaches, with a smooth touchdown ensuing. On the runway, the flaps were again set to “2” while the flight controls were reset to their Normal mode.

The last approach would end with a maximum effort stop. The KC-390’s auto-brake system had yet to be cleared for use on this prototype, so I would have to apply wheel brakes quickly after main gear touchdown. After rolling out on final, I used a 3° reference line on the primary flight display’s synthetic vision display to start the approach, aiming for the threshold of the runway. After a normal flare and touchdown, I lowered the nose to the runway while simultaneously applying moderate toe braking. When the nose settled on the runway I applied maximum reverse thrust, the engines still spooling up as we slowed below 20kt. In all, the ground run was less than 600m. After a short taxi, we turned off the runway to Embraer’s hard stand on the east side of the field. With the engines shut down, Souza guided me through the quite simple post-flight checklists.

Overall I was very impressed with the KC-390. Its FBW control system made it a joy to fly, while offering envelope protection features that should ease mission accomplishment. From a logistics mission standpoint the KC-390 is in a class by itself. Additionally, Embraer has cleared it for a number of the typically more demanding tactical roles. Heavy air drop and AAR clearance are two capabilities on the near-term horizon that are essential for the KC-390 to achieve before it can be truly mission ready. I think it is safe to say the KC-390’s robust architecture and modern systems will ensure that these capabilities are attained.

It would be unfair to our readership not to address the question that is undoubtedly on many minds: “Is the KC-390 better than the Hercules?” My response: it depends. From a flying qualities standpoint, I would much rather fly the KC-390. As far as current proven mission capabilities are concerned, the Hercules has the upper hand. Turbofan versus turboprop is not just a preference, it gives each aircraft inherent advantages and disadvantages. From a global perspective, I might liken future medium-lift competitions to those that played out between the established Boeing 737 and the upstart Airbus A320. There will be ebbs and flows, and the KC-390 need not achieve an even market split to be considered a huge success.

Source: Flight International