NTSB Cites Maintenance Oversights, Lack Of Pilot Checks

As is often the case in air tragedies, it was a cacophony of errors that led to the downing of a Colgan Air Beech 1900D off the coast of Cape Cod last year.

In its listing of probable cause in the August 26, 2003 accident, the NTSB cited a flight crew that didn't follow its checklists, a faulty maintenance manual and mechanics who took what turned out to be fatal shortcuts.

The Colgan Air Beech had just come out of maintenance when Captain Scott Knabe and First Officer Steven Dean were assigned to deadhead from Barnstable Municipal Airport in Hyannis (MA) to Albany (NY). A team of four mechanics in Hyannis had replaced two elevator trim actuators and a 55-foot long forward elevator trim cable.

The NTSB report stated:

The two lead maintenance technicians that replaced the forward elevator trim tab cable did not use a lead wire as instructed by the AMM. They marked the topmost cable pulleys with a "T" instead. A lead maintenance technician and the quality assurance inspector stated that following the maintenance; a successful operational check of the system was completed. They added that the operational check included running the manual and electric elevator trim several times, with the quality assurance inspector at the cockpit and tailbone during different phases of the operational check.

Then there was the issue of the Raytheon Aircraft AMM manual itself:

Review of the Beech AMM Chapter 27-30-04, "Elevator Trim Tab Cables - Maintenance Practices," revealed that the trim drum was depicted backwards. Although the drum could not be installed backwards, it was possible to mis-route the cable around the drum, and reverse the trim system. The depiction in the maintenance manual showed the nose-up trim tab cable emanating from the aft end of the drum, rather than the forward end. It also showed the nose-down cable emanating from the forward end of the drum, rather than the aft. However, the "FORWARD AS INSTALLED" arrow included in the depiction would have to be ignored, and the cables would have to be crossed once along the cable run, to reverse the system and secure the cable ends into the turnbuckles.

Further review of the Beech AMM revealed that there was no procedure for an operational check contained in Chapter 27-30-04. Nor was there a referral to Chapter 27-30-09, "Elevator Trim - Maintenance Practices...Elevator Trim Operational Check;" which did contain a procedure for an operational check of the elevator trim system.

It wasn't the first time Raytheon's Beech 1900 manuals had come under fire. There were two fatal crashes that immediately followed 1900D maintenance. But Raytheon issued a statement Tuesday, denying blame and instead pointing a finger at Colgan.

"As with any case, there are many details to consider and we understand the NTSB's statement of probable cause to be factually correct," said company spokesman Tim Travis. "Had the mechanic performed a proper functional check as required by the maintenance manual, the error would have been discovered."

From The Flight Deck

In addition to the maintenance and manual problems, the NTSB report said Knabe and Dean failed to perform their "first flight of the day" checks.

Further, the cockpit voice recorder showed the copilot failed to disable the trim:

• At 1523:30, the captain called for the Before Start checklist.
• At 1523:43, the first officer stated, "preflight's complete. cockpit scan complete." The captain replied, "complete."
• At 1523:58, the first officer stated, "maintenance log, release, checked the aircraft." The captain replied, "uhhhh. maintenance and release on aircraft." The captain subsequently identified that the DFDR was inoperative, and confirmed that the minimum equipment list (MEL) was still open.
• At 1525:11, the captain began to start the right engine, before being interrupted. Approximately 1 minute later, after a conversation with maintenance personnel over the radio, the captain resumed the starting of the right engine.
• At 1529:29, as the captain was starting the left engine, the flight crew began non-pertinent conversation, which lasted about 30 seconds.
• At 1530:04, the captain called for the After Start checklist. After completing the After Start checklist items, the first officer announced the checklist "complete."
• At 1530:21, the captain continued the previous non-pertinent conversation, followed 10 seconds later with, "all right we're ready to taxi with HOTEL."
• At 1530:50, the flight crew began a conversation about the flight plan to ALB, taxiing the airplane, and which pilot would fly the airplane. The conversation lasted for about 4 minutes.
• At 1535:14, during the Taxi checklist, the first officer stated, "...three trims are set." The first officer then called the Taxi checklist "complete."
• At 1535:26, the flight crew began a non-pertinent discussion about a landing airplane. The discussion lasted about 1 minute and 27 seconds.
• At 1537:00, the airplane was holding short of runway 24.
• At 1537:17, the captain stated, "all right. forty six is ready." The flight crew then began to announce several items, which were identified as being on the Before Takeoff checklist; however, the checklist was not called for.
• At 1538:07, the controller cleared Colgan flight 9446 for takeoff on runway 24.
• At 1538:08, the flight crew initiated a takeoff on runway 24.
• At 1538:40, the first officer stated "V1...rotate."
• At 1538:46, the captain stated, "...we got a hot trim..." At that time, according to the digital flight data recorder (DFDR), the elevator trim moved from approximately -1.5 degrees (nose down) to -3 degrees at a speed consistent with the electric trim motor.

  

• At 1538:48, the captain stated, "kill the trim kill the trim kill the trim."
• At 1538:50, the captain stated, "roll back...roll back roll back roll back roll back." According to the DFDR, the elevator trim then moved from approximately -3 degrees to -7 degrees at a speed greater than the capacity of the electric trim motor.
• At 1538:56, the captain stated, "roll it back roll my trim..."
• At 1539:00, the captain stated, "do the electric trim disconnect..."
• At 1539:04, the captain instructed the first officer to, "go on the controls" with him.
• At 1539:14, the captain instructed the first officer to retract the landing gear.
• At 1539:18, the captain instructed the first officer to retract the flaps. The first officer responded that they were "up."
• At 1539:21, the captain declared an emergency regarding a runaway trim and requested to return to the airport. The controller acknowledged the emergency and offered the option of the left or right downwind for runway 24.
• At 1539:33, the captain instructed the first officer to reduce the engine power.
• From 1539:49 to 1540:03, the captain instructed the first officer to "pull the breaker." The first officer queried the captain as to its location.
• At 1540:30, the captain requested to land on runway 33. The controller acknowledged the transmission and cleared the flight to land on runway 33.
• The recording ended at 1540:47.

Still, the NTSB report showed there was very little Knabe and Dean might have done to save the aircraft. The data showed it would have taken more than 250 pounds of pressure on the yoke to pull the Beech out of its fatal dive. Further, NTSB investigators flew a similarly configured flight simulator six times, attempting to recreate the conditions surrounding the accident. Five times the pilot impacted simulated terrain shortly after take-off. The sixth time, the pilot was able to circle around for a landing -- but had to touch down at 170 kts. He ended up impacting terrain after touchdown.

"I'm glad to know it wasn't their fault, that they didn't crash a good plane," said the copilot's older brother Robert Dean in an interview with the Cape Cod Times.

FMI: www.ntsb.gov/ntsb/brief2.asp?ev_id=20030904X01459&ntsbno=NYC03MA183&akey=1