Aircraft Accidents and Lessons Unlearned LXII: United Airlines Flight 585

On March 3, 1991, United Airlines Flight 585 (UA585), a Boeing 737-291, registration number N999UA, while continuing an approach into Colorado Springs Municipal Airport (KCOS), impacted terrain four nautical miles from the KCOS airport. UA585 was completing a passenger flight from Denver, Colorado, when control was lost before the aircraft struck the ground at a near 90-degree angle.

The National Transportation Safety Board (NTSB) assigned UA585 accident number DCA91MA023. The original accident report published in 1992 was AAR-92/06; the amended accident report adopted on March 27, 2001, was AAR-01/01. Any archived materials for the UA585 accident through the NTSB Archives were unavailable for accident number DCA91MA023.

The original NTSB accident report’s probable cause stated, “… after an exhaustive investigation effort, could not identify conclusive evidence to explain the loss of United flight 585.”, With all evidence available on a land-based accident with a minimal debris pattern, the NTSB chose to end the investigation midstream, on December 8, 1992, twenty-one months after the accident occurred. Why?

Per AAR-92/06, the NTSB is, “… an independent Federal agency dedicated to promoting aviation, railroad, highway, marine and hazardous materials safety.” That being said, did NTSB management make a wise decision? For the only investigatory group conducting the UA585 investigation, to abandon their responsibility after only 646 days without even holding a public hearing (as stated on page 107 of AAR-92/06 Appendix A), UA585 was not an example of promoting aviation safety. If this accident were an unmanned aerial vehicle that crashed in the forest, perhaps the NTSB’s actions could be excused for deserting their assignment, but lives were lost on UA585, and more lives were endangered (and likely were) from ignoring this accident.

Then, should the NTSB have deferred UA585 to another qualified – or, in this case – more qualified investigatory agency to analyze the accident data to a satisfactory conclusion? There was no evidence to suggest NTSB management even considered another investigatory body to find cause in UA585. Ten years later, however, the NTSB determined probable cause, not through thorough investigatory methods, but by adopting another accident’s probable cause, a move that was a convenience.

The original report, AAR-92-06, stated, “The two most likely events that could have resulted in a sudden uncontrollable lateral upset are a malfunction of the airplane’s lateral or directional control system or an encounter with an unusually severe atmospheric disturbance.” It was what the probable cause stated next that contradicts the AAR-01/01 report, “Although anomalies were identified in the airplane’s rudder control system, none would have produced a rudder movement that could not have been easily countered by the airplane’s lateral controls,” an ignorant conclusion. NTSB investigators originally rejected rudder problems as minor before ‘finding’ the rudder was the problem (AAR-01/01). In 1992, the NTSB blamed the accident on a rotor, an atmospheric event found that day in the accident area.

Ten years later in AAR-01/01, the NTSB determined that UA585’s probable cause was, “… a loss of control of the airplane resulting from the movement of the rudder surface to its blowdown limit. The rudder surface most likely deflected in a direction opposite to that commanded by the pilots as a result of a jam of the main rudder power control unit servo valve secondary slide to the servo valve housing offset from its neutral position and overtravel of the primary slide.”

Indifference towards AAR-92/06 was unfortunate. The NTSB’s failure to originally pursue UA585 to conclusion, that … was tragic. The NTSB lost control of the UA585 investigation. There was no shame in not discovering UA585’s cause in 1991; every organization has its limitations. But what the NTSB did by shutting down the investigation, by not deferring to more qualified investigators, made repeated events possible. The NTSB could have taken the high road and admitted UA585 was beyond their abilities and asked for help. They did not.

So, for almost ten years, the NTSB allowed this accident to fester. Assuming that AAR-01/01 was correct, it was conceivable that a review by a separate responsible investigatory group could have determined the root cause of UA585 in time to possibly prevent other like events.

The NTSB cut-and-pasted facts from a later accident, US Air flight 427 (US427) on September 8, 1994 (NTSB accident number DCA94MA076, Accident Report AAR-99/01), in Pittsburgh, Pennsylvania, to re-adopt the UA585 report into amended AAR-01/01. When reviewing AAR-01/01, the telling difference between AAR-92/06 and AAR-01/01are the revision bars that populate the sides of the revised report. What was reported in AAR-01/01 – what was discovered during US427’s investigation and conveyed to AAR-01/01 – was an in-depth look at the 737’s rudder system control and linkages. Remember, the NTSB originally looked at the rudder system, but dismissed it. After US427’s investigation, the NTSB wrote AAR-01/01’s probable cause as, “… a loss of control of the airplane resulting from the movement of the rudder surface to its blowdown limit.” Copy and paste. On page 12 of AAR-92/06, the investigators took a brief look at the rudder system in 1.6.3 Flight Control Systems Description. In AAR-01/01, the view was expanded completely by revisions in 1.6.3 Boeing 737 Hydraulic System Information, where the revision bars run almost non-stop through the section.

Accident report AAR-92/06 sat untouched for nine years before revision. 20 years later, it’s still not clear whether UA585 was ever properly re-investigated because of questionable analysis found in AAR-01/01, which highlights three reasons why UA585’s accident reports failed to solve for cause.

1 – The events that led to UA585 were so unique that they should have been investigated to the very end. There was no reason to allow UA585 to go unanswered. Here was an opportunity for the NTSB to act responsibly and defer the investigation, if not the controls, to a more apt investigatory body – not the Federal Aviation Administration – but a more unbiased group. The NTSB failed with UA585 because they lacked the expertise, indeed the quality to perform a conclusive investigation. Passing on to those more qualified would have been a renewed dedication to the NTSB’s mission of improving aviation safety at all costs. Instead, the ball was dropped with a thud.

2 – In the US427 accident report, AAR-99/01’s probable cause stated, “… a loss of control of the airplane from the movement of the rudder to its blowdown point,” which was almost word-for-word the AAR-01/01’s revised probable cause. Did anyone ask what the NTSB did to prove US427’s probable cause applied to UA585? Reviewing AAR-01/01, the evidence tying UA585’s and US427’s probable causes was sketchy, at best. Did the NTSB verify the UA585’s new probable cause? Did the NTSB test UA585’s rudder components or did they strictly use US427’s data? Did the aviation industry question if the US427 accident could have been averted had UA585 been properly investigated three years earlier?

3 – To be clear, there were two inexcusable problems concerning qualified investigators on the UA585 investigation. The first was that the NTSB continued to employ engineers who were out of their depth. These engineers looked at issues beyond their experience. The second was that the NTSB, in its arrogance (or its ignorance), refused to employ aircraft mechanics (technicians) to investigate the areas of an aircraft that only maintenance technicians come into first-hand contact with.

What did these two points mean to solving UA585 in 1992? Engineers the NTSB hires come from either a manufacturer, an air operator’s engineering group or are recent graduates from college. The latter, the recent graduate, is as inexperienced as one can possibly be – no hands-on, no reference to past experience, no career to pull from, no mentor to teach them. In addition, the recent graduate will do whatever the NTSB manager says to do – improper in itself – because NTSB managers are past engineers.

An engineer who comes to the NTSB from an air operator or manufacturer is even less qualified. The fact is that any of these engineers will come to the NTSB understanding only one system on only one aircraft. If this engineer came to the NTSB after years of working pneumatic systems, what knowledge would he/she have of hydraulics, navigation, electronics, ice/rain protection, fire suppression, fueling, flight controls, landing gear, door rigging, or pitot static systems? None. The engineer would have zero knowledge of these systems, so he/she would have no idea what they are looking at.

What’s more is that the manufacturing engineer comes with a special engineering handicap. In addition to only working one system, he/she doesn’t know what an aircraft operates like after it rolls out of the factory. They only know what an aircraft is supposed to do. Their experience is with brand new aircraft, not aircraft exposed to salt air, hard landings, airworthiness directives, ground handlers, turbulence, long hauls, contract maintenance, engineering orders, or anything an aircraft experiences through its years of service. These engineers don’t know what to look for in a heavy maintenance check, door or flight control riggings, checking structural integrity with non-destructive inspection equipment, replacing components for time changes, or even defer per the minimum equipment list … or why.

The NTSB continues to exclude aircraft maintenance experience from investigations. Aircraft technicians work aircraft in the field, an advantage engineers do not possess. Aircraft technicians work with rudder packages; technicians work rudder component inspections, conduct operational testing on all flight controls, rigging them, inspecting for mount cracks and overall integrity. A technician knows when a rudder works correctly and incorrectly. Aircraft technicians are the most effective accident investigators.

NOTE: In AAR-92/06, page 101, section 2.7.2, the NTSB focused on problematic flight data recorder (FDR) problems, which was misdirection for those not understanding a system. FDRs are an excuse, a crutch on which the investigation should never rely on. Like all technology, the FDR is a tool, not the basis for which any or all accident investigations pivot upon. I say this from experience; the aircraft maintenance accident investigations I conducted never relied on the FDR or any recording device. They relied instead on the information provided by my investigatory teams – maintenance technicians – who always spoke to the accidents’ fundamental issues.

UA585 deserved to have been handled with more resolve. Ten years without closure was inexcusable, because every time a B737 took off, it ran the risk of meeting UA585’s tragic end.