Aircraft Accidents and Lessons Unlearned IV: Colgan Air 9446
I remember watching a 1967 episode of Star Trek: The Original Series called Court Martial … and, yeah, I remember watching it when it originally aired. Captain Kirk is watching a video of himself on the bridge; the video distinctly shows him ‘jumping the gun’ and pressing a jettison button during Yellow Alert, thereby killing a crewman when danger did not exist. Kirk’s reaction is (in William Shatner’s voice and style): “But that’s not the way it happened.”
Most of my Lessons Unlearned monthly series will deal with accidents I did not work, but I find them educational to revisit, in the hopes of learning what was missed. These missed lessons could very well result in future accidents and deaths because we never learned the lessons the first time.
However, Colgan Air 9446 was one that I did work; I was the Aircraft Maintenance accident investigator working at Yarmouth, MA, to find the cause. Recently, I was reading the National Transportation Safety Board (NTSB) summary of the accident; just like Captain Kirk, I find myself reading the words and saying, “But that’s not how it happened.”
The NTSB’s final report of the Colgan accident’s Probable Cause read:
The improper replacement of the forward elevator trim cable, and subsequent inadequate functional check of the maintenance performed, which resulted in a reversal of the elevator trim system and a loss of control in-flight. Factors were the flight crew’s failure to follow the checklist procedures, and the aircraft manufacturer’s erroneous depiction of the elevator trim drum in the maintenance manual.
Now, to the first question: Why didn’t I point out the true Probable Cause back in 2003? Answer: I can’t. The Inspector in Charge of the accident controls the Probable Cause; I don’t get a say.
Colgan Air was a Regional airline for US Air. Colgan Air flight 9446, tail number N240CJ, was a repositioning flight for US Air Express; a Beech 1900D that was being flown Part 91 from a maintenance base in Yarmouth, MA, to Albany, NY, to recover a flight. On August 26, 2003, it departed Yarmouth’s airport: Barnstable Municipal airport (HYA), made it 100 yards off the coast before flying into the ocean. There were no passengers; the only fatalities were the two pilots. NOTE: It is important to point out that Colgan Air 9446 followed the Air Midwest 5481 accident by almost eight months, that the circumstances were similar – not identical – and the NTSB’s first impression was the Beech 1900D aircraft was the cause of the two accidents. Although not blameless, Beech aircraft, particularly the 1900D, was not the cause of the accident.
Before we explore the accident, I wish to hammer on, yet again, an important distinction I make in my Lessons Unlearned article series: there is a clear difference between what did CAUSE THE accident and what was THE CAUSE of the accident. Yes, yes … the first use of ‘cause’ is as a verb, e.g. Engine failure did cause the accident. The second use is as a noun, e.g. The cause of the accident was engine failure. This must be distinctly understood because the NTSB focuses on the verb and not the noun. They target on what made a perfectly good airplane malfunction suddenly, fall out of the sky and hit terrain. The NTSB rarely walks the accident backwards to determine what series of events led to the malfunction to begin with. The causes are where things went wrong; they are the Lessons Unlearned; they are why planes crash and people die.
Everything else is irrelevant.
Colgan Air 9446 was pulled out of a maintenance phase check early; US Air Express needed aircraft N240CJ; they deferred the remaining part of the maintenance check until a later date – this is a common practice; it is not dangerous …. at all.
While the check was still in progress, the elevator trim system was inspected and found to have a bad actuator, so the actuator was changed. NOTE: This actuator change led to the cause of the accident.
In Risk Analysis, we normally ask ‘five whys’; the theory being that if you ask why to every answer, by the time you get to the fifth why, the answer presents itself. So, let’s utilize the five whys to determine the cause of the accident:
First Why: Why did the plane crash? The plane crashed because the elevator trim system was reversed through the airplane; in other words, the elevator trim cables were physically run backwards. The reversal meant that as the pilot trimmed for ‘nose up’, the aircraft flew ‘nose down’ and vice versa. The inputs put a load on the yoke column that fought the pilots’ attempts to fight the condition. Lacking altitude and time to recover, the pilots unwittingly continued to feed ‘nose down’ into the aircraft until they hit the water.
Second Why: Why were the cables routed in reverse? Good question, because the cables were correctly routed when they arrived for maintenance and they were not scheduled for replacement. Part of the blame lies with Beech Aircraft; its manual shows the proper way to install the trim cable drum in the pedestal. Unfortunately, the drum is not ‘Murphy proof’ and can be installed backwards, which is what happened. By being installed backwards, the cables became ‘reversed’ throughout the airframe, even though they were routed correctly. This error pointed directly back to Beech Aircraft.
The NTSB felt that the answer had been found and focused little energy to look further. I felt there was more to the problem, so asked the question that the NTSB didn’t feel needed asking:
Third Why: Why were the elevator trim cables replaced to begin with? This is where the transition between ‘cause the’ and ‘the cause’ takes place. In the second why paragraph, I said the aircraft arrived in maintenance with properly routed cables and they were not scheduled to be changed. During an elevator trim actuator change, the cables came loose and unraveled off the elevator trim cable drum in the pedestal. When they tightened the cables up, the cable didn’t sit correctly on the drum. When they tested the system, the cable got pinched inside the pedestal, damaging the cable and requiring replacement. NOTE: if done correctly, a cable should never come loose during an actuator change, so …
Fourth Why: Why did the cable come loose during maintenance? And here is where the plot thickens: the cable came loose off the drum because a newly hired mechanic did not employ proper procedures in changing out the elevator trim actuator; no one taught him practices, e.g. ‘blocking cables’ to maintain tension. Instead he used masking tape to hold the greasy cables to the side of the vertical stabilizer. Masking tape, being an inadequate ‘cable block’, allowed the cables to loosen; they unwound off the cable drum in the pedestal.
Fifth Why: Why was an untrained mechanic working by himself? This question speaks to the Training program for the Colgan mechanics and the lack of oversight conducted by Colgan’s management to make sure all mechanics receive proper training as per their approved training manuals.
The first thing people say is why didn’t the Federal Aviation Administration (FAA) get in front of this? Let’s use some perspective: Many airlines employ hundreds of mechanics and pilots; major airlines employ thousands of each, located all around the world, 24/7, 365 days a year. An FAA office in charge of an airline, employs perhaps 30 to 50 inspectors to oversee, not only several thousand mechanics, but the maintenance they perform on hundreds of aircraft in its fleet. The strain on Operations inspectors is similar: 30 to 50 inspectors overseeing thousands of pilots worldwide. One must do the math to understand.
This accident demonstrates why it is important to look beyond what caused the accident to the cause. Whatever caused the accident is over with; it ended when the aircraft arrived at the point of impact. What was the cause of the accident can continue, unchanged, until someone recognizes the problem and fixes it. This was the second time in eight months that a Beech 1900D crashed due in large part to improper training. This was the true cause of the accident.
This was the Lesson Unlearned.