Guy Norris/MOJAVE CFM International's latest engine, the CFM56-7, is being put through its paces on General Electric's Boeing 747 test-bed.

According to Phil Schultz, General Electric flight-test organisation (FTO) chief pilot, everything you see on GE's Boeing 747-100 test-bed - all the subtle changes - "represent ten years of flight-test experience on the [Boeing] 707 and [Airbus] A300".

Schultz is describing the aircraft, which was acquired by GE in 1991. Representing an investment of around $30 million in modifications, upgrades and test equipment, the 747 is now being used to test the CFM International (CFMI) CFM56-7 which will power Boeing's briskly selling 737-600/-700/-800 family.

Flight International joined the fifth flight of the CFM56-7B test programme, which began on 16 January and is scheduled to end in March. The pace of the -7B testing has surprised even some members of the test team. Up to 25 January, when the fifth flight took place, the engine had undergone some 20 flight hours and 31 ground hours.


To the satisfaction of the test team, data were also pouring in at unprecedented rates. Some 203 steady-state and 128 transient readings had already been recorded, compared with the entire second phase of the General Electric GE90 turbofan testing when 267 steady-state and 888 transient conditions had been recorded over 111 flight hours.

The fifth flight was a typical test sortie to evaluate in-flight performance, starting with a lapse-rate take-off and followed by sustained periods of performance monitoring at various altitudes up to 30,000ft (9,000m). Before take-off, the test engine was set at 70% N1 (low-pressure turbine) and allowed to stabilise before brake release. As speed built towards 60kt (110km/h), the CFM56-7B was set at 98% N1 to match the test-bed's three Pratt & Whitney JT9Ds.

Lifting off quickly, into a cold and blustery northwesterly wind, the 747 was flown to 1,500ft (above ground level - 5,000ft above sea level), while the lapse-rate data were collected. Fighting sometimes very severe turbulence, the crew jockeyed the three JT9D levers to keep a constant 250kt and to maintain straight-and-level flight as near as possible to 1,500ft for the 8min duration of the first test condition.

Power performance was calculated at varying schedules of the high-pressure (HP) turbine clearance-control system, with engine bleed on and off, while the crew flew the 747 in a controlled racetrack pattern to the north of Edwards AFB. The 747 was taken to slightly smoother air at 10,000ft, where the exercise was repeated at Mach 0.4 and varying N1s. Test points were set within +100ft and Mach +0.005 of target. During data recording, the crew fought to keep the aircraft within +50ft and the airspeed within +20kt. Mojave-based FTO test director Al Krejmas' test team also watched carefully to ensure that true air temperature remained within +1°C during the data-collection phase.

After a further hour, the 747 was flown to 30,000ft and recordings taken at M0.78. For this higher altitude portion of the flight, tests were conducted well away from Edwards airspace, over the Sierra Nevada mountains and adjacent to Owens Valley. The area is nonetheless busy with military aircraft and the crew keeps a constant vigil for conflicting traffic. Several US Navy McDonnell Douglas F-18s flew past en route for the nearby China Lake weapons test range and, returning to Mojave, the 747 passed above a Northrop Grumman B-2 and its Lockheed Martin F-16 chase aircraft. After undergoing more than 40 test conditions, the aircraft landed with a further 4h 30min added to the test tally.


Flight-testing, is a vital milestone for any engine, and particularly for the CFM56-7B. Although based heavily on its earlier siblings, the -7B is the first variant to include a wide-chord fan. It is also encased in a new Boeing-designed nacelle and duct. Added to the challenge of testing the new design features, the FTO faces an ambitious time-scale, which calls for certification in October 1996, almost four months ahead of the first flight of the 737-700.

Coupled with the basic certification work is the twin objective of ensuring that the whole propulsion system (as opposed to the engine itself) is ready for service as a mature product at its entry-into-service date of October 1997. On top of this is the novel, but not unwelcome, responsibility of ironing out the bugs in a derivative of what has become the best-selling commercial engine in history, for an airliner that is yet to fly.

"We want to be right as early as we can because we don't want to be in a position where we have to retrofit a huge amount of aircraft just because we didn't test something right in January 1996," says CFM56 programme manager Bruce Hughes. The folly of early over-sights, would be further compounded, by the sheer volume of -7 engines on order and the rapid production build-up. "When we go into production on this it's going to be like taking a drink from a fire hose," he says.

The flight test effort, gained new importance, even before it began. In late 1995 several tests (including a blade-out and medium-bird-strike evaluation) revealed the need for a late redesign of the fan case. Concerns over the containment margin and inlet structural stiffness led to a strengthening of the intake area. An additional containment collar, extending forward by 330mm, was bonded into the existing nacelle fairing to provide extra protection and increased stiffening. Selected areas of the existing fan-containment ring were thickened and the inlet gearbox casing was strengthened.

"We had already run two blade-out spin-pit tests, and we decided to put our inlet on the third spin test. That's when we discovered we needed to extend the containment collar," says Boeing CFM56-7 propulsion integrated product team leader, James Anderson. "[Nacelle manufacturer] Rohr put in a good effort. They had to do a lot of work to speed up the process," he adds.


The full-authority digital engine-control software governing stall detection and recovery logic is also being altered in response to the medium-bird-strike test, which will be re-run in April. During the first test, engine thrust had to be reduced to 70% before the post-strike stall could be cleared. US Federal Airworthiness Requirements call for a 75% target. "We need to make some control changes," confirms Peter Thompson, CFM56-7 project leader - but "...the fan came through it OK", he adds.

Following flight test, the next milestone month is April. "That's when we will do both the medium-bird and full blade-out," says Hughes. The engine underwent hail and ice slab tests in February and is due for over-temperature evaluation in March. Water-ingestion tests are set for May, with CFMI expected to decide in March on possible in-flight icing tests.

A small fence on top of the exhaust nozzle is also being tested. This disrupts turbulent airflow across the upper surface of the nozzle, which has been blamed for cracking in the area. Various locations and sizes of the Inconel fence are being tested.

For the moment, the emphasis is to verify that flight test performance, matches better-than-expected ground test results. Cruise specific fuel consumption was found to be around -0.6% better, while the exhaust-gas temperature margin was 20°C larger than predicted. At its 115kN (26,000lb) rating, the engine was demonstrating a 4% thrust margin in its first few flights. Much of this is attributed to the wide-chord fan which " proving to be a better performer than we expected at sea level", says Thompson.

Overall, it is not just the engine but the test-bed itself which is also performing better than expected - a trend which GE will wish to see continue as it prepares to test the 410kN version of the GE90 later this year.

Source: Flight International