When a noticeably tail-heavy Air Midwest Beech 1900D lifted off from Charlotte-Douglas International Airport, N.C., on the morning of January 8, the twin turboprop was carrying the heaviest payload of its seven previous flights since emerging from a D-6 maintenance check two nights before.
According to one pilot eyewitness, it smoothly increased pitch “like [at] an airshow,” until–at about 1,000 feet–it did a wing-over to the left and began descending nearly vertically. About halfway to the ground, he said, it rolled to the right and the nose lifted to almost level flight before it appeared to roll over again to the right and dive into the corner of a maintenance building at the airport.
The crash of Flight 5481 is the subject of an ongoing NTSB investigation, which is looking at company and contract operations, training, surveillance, management and oversight. It also prompted the FAA to issue an AD on elevator rigging for the Beech 1900 series and to form an Aviation Rulemaking Committee to conduct a survey on average weights of passengers and baggage.
A two-day NTSB hearing in Washington, D.C., in late May revealed that maintenance provider Raytheon Aerospace, which accomplished the D-6 check at an Air Midwest facility in Huntington, W.Va., used the services of mechanics employed through Structural Modification and Repair Technicians (Smart) of Edgewater, Fla.
NTSB member John Goglia, himself a former airline A&P mechanic, said during a hearing break that Air Midwest’s reliance on third-party mechanics “flew under the radar” of the regional airline’s FAA principle maintenance inspector (PMI). Thomas Wellcome, director of recruiting for Smart, admitted under board questioning that his company often hired contract mechanics over the telephone with no face-to-face interaction.
Under questioning by Goglia, Wellcome said Smart’s mechanics stay an average of two or three months at a posting before moving to another job or being hired permanently by the company that had contracted for their mechanic services. He characterized Smart as a human-resources service supporting the industry. Of about 275 mechanics it had under contract, he estimated that about 100 held A&P licenses.
Wellcome said Smart’s mechanics learn by on-the-job training (OTJ) provided by the maintenance companies that hire them, and he did not dispute a statement by Goglia that five of the mechanics working the night of the D-6 check had “virtually no Beech 1900 experience.”
When Air Midwest Flight 5481 took off at about 0848 EST on January 8, the aircraft was near its mtow and close to the aft limit of its c.g. envelope using then-accepted weight averages for passengers and baggage. The Beech 1900D load manifest form for the accident flight listed a takeoff weight of 17,018 pounds, and the flight crew spent several minutes discussing and recalculating the weight-and-balance. The mtow of the 1900D is listed as 17,400 pounds, and using the recently revised weight averages would have put it at least 300 pounds over mtow.
During the course of the NTSB investigation, a ramp agent said he asked Capt. Katie Leslie if she wanted to do the calculations for fuel, passengers and baggage and tell him if he needed to remove any bags because of excess weight. About 10 minutes later, he again went to the cockpit window and asked if any bags needed to come off the airplane. The captain told him that the number of bags loaded was fine. He estimated the aft baggage compartment was 98 percent full by volume. In addition, all 19 seats were occupied.
Several pilot eyewitnesses who saw the accident airplane taxiing to the takeoff runway said the airplane appeared to be heavily loaded and sitting tail-low. One pilot stated that the nose strut was “extremely” extended, and he remembered that the airplane “bounced heavily” as it taxied.
An American Airlines MD-80 first officer who had 2,000 hours PIC time in the Beech 1900 stated that her impression was that the accident airplane was “heavily, heavily loaded,” describing it as “tail-low and nose-high.” Another pilot said the main landing gear struts were compressed, the tires were compressed and the nose strut was almost fully extended.
The accident occurred on the first flight of the flight crew’s scheduled one-day flight sequence, which consisted of three flight legs. They were scheduled to fly from Charlotte to Greenville-Spartanburg International Airport, S.C., and then continue on
to Raleigh-Durham International Airport, N.C., before returning to Charlotte.
According to Air Midwest records, on January 7 the accident flight crew flew the accident airplane on six flight legs for a total flight time of six hours. Another flight crew met them when they finished their day in Charlotte and flew the airplane to Lynchburg Regional Airport/Preston Glenn Field, Va.
The next morning, January 8, that same flight crew flew the accident airplane back to Charlotte, arriving at 0715. In interviews with Safety Board investigators, neither the captain nor the first officer on those two flight legs noticed anything unusual about the handling of the accident airplane.
The accident flight, which was operating as US Airways Express, departed the gate at Charlotte on time at 0830 and taxied to Runway 18R. It was cleared for takeoff
at 0846:18.2 and at 0846:35 Leslie called for takeoff power. FO Jonathan Gibbs announced V1 and Vr at 0846:49.6, but less than 14 seconds later Leslie said, “Help me,” followed by “you got it?”
In the next few seconds, the cockpit voice recorder picked up sounds of grunting and exhaling, along with warning horns, changes in power and exhortations from both pilots to push the nose down. At 0847:16.4 Leslie radioed the tower, “We have an emergency for Air Midwest 5481.”
NTSB investigator-in-charge Lorenda Ward testified at the May hearing that following the D-6 “phase check” at the Huntington facility in the overnight hours of January 6 to 7, the range of motion of the control column was “severely limited.” She said that the pitch-control turnbuckles had been adjusted to “abnormal” length, and “the airplane lost about two-thirds of its down elevator travel.” She added that Flight 5481 was loaded the heaviest and had the most aft c.g. of the 10 flights it made after the maintenance procedures were performed.
“We believe the crew filled out the paperwork properly, but that the assumptions used were incorrect,” said Ward, who suggested that the actual weights were greater than the average weights used by airlines in their weight-and-balance calculations.
On January 24 the FAA requested all Part 121 operators of 10- to 19-passenger airplanes to survey some actual passenger weights. A standing FAA advisory circular recommends an average adult weight of 180 pounds in the summer and 185 in the winter, including 20 pounds of carry-on baggage and one personal item.
But the survey found that average passenger weights were higher by 20.36 pounds, carry-on bags were higher by 5.72 pounds and domestic checked bags were higher by 3.81 pounds. The FAA announced in mid-May that the 10- to 19-seat carriers that conducted the survey had already adjusted their weight-and-balance assumptions to reflect their findings, where appropriate.
Further, the agency said that by August 12 all airlines using an average weight-and-balance program must add 10 pounds to account for passengers’ personal items and five pounds to the average weight of checked bags. “The recent survey results indicated these short-term changes would be prudent until the FAA can gather more extensive data about passenger and bag average weights,” the FAA said. After the crash, Air Midwest increased the total weight in its passenger/baggage calculations to 200 pounds.
The mechanic trainee who did the flight-control rigging the night of the inspection was an employee of Smart. He was assigned to the Huntington base less than two months before the accident. Although he held an A&P certificate and had previous flight-control rigging experience on other airframes for other employers, he had no previous experience working on Beech 1900s.
He told investigators that he felt properly trained for the rigging task, but admitted that he omitted several steps in the maintenance manual because he considered them unnecessary. When asked whether he believed he was properly overseen during the task, he stated, “Yes, when I needed help, there was somebody around.”