MIKE GERZANICS / WICHITA
Regional jet is a misnomer for Bombardier's stretched CRJ900: it's a fully fledged airliner to rival its bigger competitors. But how does it fly?
The 80- to 100-seat segment is perhaps the most problematic in the airliner market. While the original Douglas DC-9 series targeted this niche, its derivatives grew larger and larger. With BAE Systems stopping the Avro RJ series, there is no aircraft left in production to fill the gap. While Embraer and Fairchild Dornier are moving to fill the void with new designs, Bombardier has become the first to market by stretching the 70-seat CRJ700.
First flown on 21 February 2001, Bombardier's CRJ900 is a stretch of the original Canadair Regional Jet and, at the same time, something totally different. Significant changes to successive models as Bombardier has developed its CRJ line have resulted in the CRJ900 being a full-fledged airliner. The regional jet definition attached to the CRJ900 is more a reflection of airline economics than aircraft characteristics and, with its arrival, regional airlines now have an aircraft nearly as capable as those of the carriers they support.
The CRJ700 was more than a simple CRJ stretch. While it was a derivative for certification purposes, the CRJ700 incorporated only 15% of the CRJ200's airframe components. A larger wing and 50% more powerful engines were among changes made in stretching the 50-seater into a 70-seater. Bombardier's decision to stretch the aircraft to fill the 90-seater market, when other manufacturers were developing new designs, was a matter of money and time.
Fewer changes
Deriving the CRJ900 from the CRJ700 required fewer changes than the previous stretch. Two fuselage plugs, 2.29m forward and 1.57m aft of the wing, are the only major changes, ensuring a high degree of commonality. Depending on customer preference, two General Electric CF34-8C5 engines can produce around 5% or 7% more thrust on take-off than the CRJ700's -8C1s. Two extra overwing emergency exits and an aft service door on the right-hand side complete the major external changes. Structurally, the wing and landing gear have been strengthened to handle the increased maximum take-off weight of 36,515kg (80,500lb). The cockpit is virtually identical to the CRJ700's and Bombardier is proposing a single type rating.
The CRJ900 addition allows Bombardier to offer one aircraft family, ranging from 40 to 90 seats. In the same way that Airbus has sought to achieve with its A320 series, the CRJ family allows seat capacity to be tailored to specific market demands. Common crew qualification reduces training costs, while increasing crew scheduling flexibility. Common ground support equipment and spares further reduces costs.
Operating costs
But where the CRJ900 shines is in direct operating costs. In the North American market, based on a nominal 925km (500nm) stage-length, the 86-seat CRJ900 is projected to have an 11% lower cost per available seat mile than the CRJ700. It will be 21% cheaper to operate than the CRJ200, says Bombardier. In the European environment, savings are almost as dramatic: 9% less than the CRJ700 and 18% less than the CRJ200.
In preparation for our flight, Bombardier test pilot Chuck Ellis led the walk-around of the second CRJ900, C-GRNH, representative of aircraft to be delivered to launch customer Mesa Airlines in the first quarter of next year. During the routine pre-flight, Ellis pointed out two small ventral strakes below the tail section. Added to increase directional stability of the stretched airframe, they had another unplanned benefit: the strakes act as vortex generators, energising airflow over the tail section and reducing drag.
Noise from the tail-mounted auxiliary power unit (APU) was low. Three cargo doors on the fuselage's left-hand side are easily accessible from ground level, and should lead to rapid turn-arounds.
Integral stairs on the forward entry door lead to the cabin, allowing operations at fields without ground support equipment. The cabin had 86 seats in a four-abreast single-class configuration. The addition of a third seat rail will allow operators to install premium class seating, giving the option of employing the CRJ900 on routes previously reserved for mainline jets. An aisle 1.89m height should allow passengers easy access up and down the single aisle. The redesigned overhead bins, while not as large as those of a A320 or Boeing 737, are larger than on most commuter aircraft.
Entry to the flightdeck is via a full-size lockable door, not a trivial consideration in today's environment. Once settled in the four-way adjustable left seat I found the cockpit to be well arranged. Field of view out of the four windows was good. The small overhead panel contains controls for aircraft systems and is logically laid out. The instrument panel is dominated by the six cathode ray tubes (CRTs) of the Rockwell Collins Pro Line 4 electronic flight instrument system. While not the company's latest offering, the 150mm x 180mm CRTs proved more than adequate for safe and efficient operations. Two control display units (CDUs) located on the wide centre console facilitated flight management system (FMS) operations.
Avionics packages can be tailored to the customer. To meet required navigation performance (RNP) standards, European carriers are ordering aircraft equipped with inertial navigation systems (INS). US regionals can still meet RNP standards with less expensive attitude/heading reference systems. The test aircraft had dual INS and dual global positioning system (GPS) receivers. INS alignment took about 7min. APU start was rapid, and pre-start actions and FMS programming were straightforward.
Once we were cleared by ground personnel, the right engine was started. APU bleed air quickly spun up the engine and throttle was brought out of SHUT OFF at 20% N2 (compressor speed). Light-off was nearly immediate, and the starter cut out at 50% N2. Idle RPM was reached 45s after starter engagement, and inter-turbine temperature (ITT) peaked at 562¼C (1,045¼F), well below its limit of 955¼C (1,752¼F). Left engine start was much like the right, with a peak ITT of 566¼C (1,052¼F). Throughout the start sequence, bleed air and air-conditioning pack operation were automatically controlled, not unlike on the Boeing 777.
Moving off
After the parking brake was released, idle power was enough to get the aircraft moving. Tiller-controlled nosewheel steering (NWS), with ± 80¼ travel, allowed me to follow taxiway centrelines easily through several 90¼ turns. Rudder-pedal NWS, with ± 8¼ authority, was enough on straight taxiways. Toe brakes were easy to modulate to keep the 26,700kg (58,800lb) aircraft down to a reasonable taxi speed.
Flaps were set to 8¼ and thrust reversers checked during the taxi out. The APU was left on for take-off, its bleed air used to pressurise the aircraft. Engine bleeds could then be closed automatically by the bleed-air management system to improve take-off performance.
While the CRJ900 can use a derated N1 (fan speed), we elected to do a full-thrust take-off. Lined up on Wichita's runway 19R, I advanced the throttles to the TOGA (take-off and go-around) detent. N1 stabilised at 90.7% and acceleration was quite brisk as each engine produced its maximum available thrust. Ellis called rotate at the V1/VR of 122kt (226km/h) indicated airspeed. Control forces were low as I rotated to a 15¼ nose-high pitch attitude. Assisted by a 25kt headwind, we left the runway 14s after brake release and a ground run of only 610m. Gear and flap retraction caused little change in yoke forces as the aircraft accelerated to 250kt. Approaching 250kt I set climb power by putting the throttles in the CLIMB detent. Once there the full authority digital electronic control (FADEC) constantly adjusted N1 as temperature and altitude varied. While the CRJ900 lacked an autothrottle, I found the FADEC and CLIMB detent efficiently maximised climb thrust, while reducing pilot workload.
Clear of the traffic pattern I engaged the autopilot in heading and speed hold modes. While the aircraft is certified to operate up to 41,000ft (12,500m), a more typical cruise altitude will be in the mid-30,000ft region. A climb from 5,000ft to 33,000ft took 9min 42s and burned465kg of fuel. The autopilot was logical and easy to operate. Engaged modes were announced on the primary flight display (PFD) and also highlighted on the respective engagement button on the flight control panel. The autopilot smoothly captured our cruise altitude of 33,000ft.
Once level I retarded the throttles out of the CLIMB detent. At a weight of 25,815kg, the aircraft accelerated smoothly to Mach 0.85, maximum operating Mach number for that altitude. I further retarded the throttles and stabilised at 300kt/M0.842. Total fuel flow was only 1,840kg/h (4,050lb/h) at 491kt true airspeed. Bombardier expects a maximum cruise speed for the CRJ900 that is slightly faster than the CRJ700's M0.825. Stretching the fuselage has increased its fineness ratio and reduced drag, allowing a faster cruise. At M0.81, slightly faster than the M0.80 predicted long-range cruise speed, total fuel flow was 1,650kg/h (3,630lb/h) and the aircraft held 473kt at 33,000ft.
A 60¼ bank-angle turn was accomplished at 260kt indicated airspeed. No buffet was felt, and the yaw damper kept the aircraft in co-ordinated flight. A sharp rudder input, designed to excite the aircraft's Dutch Roll response, was quickly damped out with the yaw damper engaged. With the damper off, a similar rudder input showed the aircraft's unaugmented response to be undamped. Unlike the Boeing 727, which needed well-timed control inputs to damp out damper-off Dutch Roll oscillations, those in the CRJ900 dissipated when the rudder or ailerons were deflected.
During a descent to 15,000ft I accelerated the aircraft to the maximum operating airspeed plus 5kt. A sharp control input in each axis revealed a deadbeat response. Deploying the flight spoilers caused two panels on each wing to extend, increasing the rate of descent. The flight spoilers are multi-function devices, also assisting in roll control. Once level at 15,000ft I stowed the spoilers and began to slow the aircraft for an idle-power clean configuration (gear and flaps up) stall. No buffet was felt, and the stick shaker went off at 131kt for the 25,140kg aircraft.
Control in all three axes was good in the shaker region; the aircraft showed no tendency to roll off. Had I ignored the shaker and continued to slow the aircraft, a stick pusher would have fired to break the stall. Recovery was effected by relaxing yoke back-pressure and advancing the throttles.
Take-off configuration (gear down/flaps 8¼) and landing configuration (gear down/flaps 45¼) stalls were then carried out. In both cases light airframe buffet was felt 5kt before the shaker went off at 112kt and 99kt, respectively. Recovery was similarly benign, as a relaxation of back-pressure and advance of power broke the stall.
With the high angle-of-attack manoeuvres complete, the gear was retracted and flaps set to 8¼. I advanced both throttles to the TOGA detent to simulate a take-off condition where the gear had just been retracted. Ellis retarded the right engine to idle, simulating an engine failure at 135kt. Two things happened automatically to ensure a safe recovery of the aircraft.
Rudder authority
The FADECs sensed a greater than 15% N1 difference between engines, and increased the N1 on the good engine from TOGA to automatic power reserve (APR), a 2% gain. Some aircraft, the 727 for example, schedule rudder authority based solely on flap position. In the CRJ900, however, the electronic spoiler stabiliser control unit (SSCU) continuously schedules rudder authority based on air data computer and FADEC inputs. The SSCU sensed the simulated engine failure, and immediately increased available rudder authority to counter the ensuing asymmetric thrust.
Less than 45DcaN (100lb) of left rudder pressure and about one-third rudder displacement was required to maintain co-ordinated flight. Electrically actuated rudder trim was more than sufficient to relieve pedal pressure, allowing me to fly the aircraft with my feet on the floor. Satisfied with the CRJ900's engine-out flight characteristics, the flaps were retracted and I turned the aircraft towards Newton-City Airport, a small general aviation field 40km north of Wichita. The APU was started before the approach, allowing its bleed air to take over pressurising the aircraft and improving go-around performance. First came coupled-autopilot instrument landing system (ILS) approach to Runway 17. With 2,300kg of fuel remaining, Vref was 123kt for a 45¼ flap approach. The autopilot smoothly captured and tracked the localiser despite a right-quarter headwind gusting from 25kt to 35kt.
The landing gear was lowered as we approached the glidepath from below, with 30¼ flaps selected on glideslope capture. Final landing flaps of 45¼ were lowered at 1,500ft above ground level. During all configuration changes, the autopilot smoothly tracked the glideslope. A speed trend vector on the PFD's airspeed tape helped me manually maintain Vref.
At the 200ft decision altitude, I disengaged the autopilot with a yoke-mounted thumbswitch. At around 100ft I aligned the aircraft with the runway centreline and flew a wing-low approach to compensate for the stiff winds. At 30ft radar altitude I slowly began retarding the throttles to idle and began the flare manoeuvre. My initial flare was somewhat high, but I was able to easily step it down, main gear alighting at the far end of the touchdown zone.
Final approach
Once on the runway, I smoothly lowered the nose gear to the runway, and Ellis selected 20¼ flaps for our touch and go. I advanced the power to the TOGA detent and rotated as airspeed increased past approach speed. After unstick, the gear was retracted and we entered the visual circuit for another approach. On downwind, Ellis simulated an engine failure by retarding the right engine to idle.
The simulated single-engine approach was flown with 20¼ flaps, and a Vref of 135kt. I flew the approach with rudder trim set to zero, and had no difficulty maintaining the final approach path. The rudder allowed me to maintain runway line up precisely as I pulled the good engine to idle at about 25ft above the ground and started the flare. The second landing was better than the first, as my initial round-out was closer to the runway. Both engines were used to accomplish the touch-and-go.
Recovery to Wichita's Mid-Continent airport was via radar vectors to an ILS approach to Runway 19R. As was the case with the first ILS approach, Vref was 123kt for a 45¼ flap setting. I hand-flew this approach, following the flight director's guidance. Little change in pitch force was noted when flaps and landing gear were lowered during the approach. Winds at Wichita were down the runway at 15-25kt. By now I was able to better judge my round out, alighting on the runway centreline in the middle of the touchdown zone.
After several more touch-and-gos in the visual circuit, we set up for a full-stop landing. With flaps 45¼ and a gross weight of 23,465kg, Vref again was 123kt. My final landing was the best, gently touching down in the heart of the zone. Pitch attitude at touch down was 6¼, an 8¼ change from the slightly nose-down attitude on approach.
Armed for automatic operation, four flight spoilers and four ground spoilers deployed on touchdown. While a display message and aural caution would have alerted me if they failed to deploy, I would have preferred more direct feedback, such as the spoiler handle moving aft. Using moderate braking and partial thrust reversers I brought the aircraft to a stop after a ground roll of only 365m. Taxi back to the ramp and shut down took only 6min. Post-flight actions were easily accomplished after the 2h 23min flight.
During the flight I found the CRJ900 to be a responsive aircraft. Control feel and harmony were good, making it enjoyable to fly. While I would have preferred an autothrottle, the FADEC and throttle quadrant detents made thrust control an easy task. While the CRJ900 can trace its lineage back to the pioneering CRJ100 regional jet, it felt every bit like a full-size airliner. Bombardier has 30 firm orders and options for 62 more, and accountants and pilots are likely to agree the CRJ900 is a welcome addition to the popular CRJ family.
Source: Flight International
