Moment of Silence: The crash of Midwest Express flight 105

Admiral Cloudberg
15 min readNov 16, 2019
The charred wreckage of flight 105 lies in a forest after the crash. Image source: the Milwaukee Journal Sentinel

On the 6th of September 1985, a routine commuter flight suffered an uncontained engine failure on takeoff from Milwaukee, Wisconsin. While the pilots tried to diagnose the problem and stabilize their climb, they appeared to lose control of the plane. The DC-9 stalled, rolled over, and dived into the ground, killing all 31 people on board. That such an accident would result from a simple engine failure came as a surprise — after all, many engines fail around the world every year without causing crashes. Digging deeper into the sequence of events, investigators found that the engine failure was actually only half the story. In fact, the pilots completely botched a procedure that should have been drilled into them from the first days of training, resulting in a low-altitude stall from which recovery was impossible. Underlying this total breakdown in cockpit communication were a dangerous unwritten rule at Midwest Express and an FAA inspector who had no business overseeing a passenger airline.

A Midwest Express DC-9 similar to the one involved in the accident. Image source: Wikipedia

Midwest Express Airlines began life as an internal corporate jet service for executives at Kimberly-Clark, the company behind several well-known personal hygiene brands such as Kleenex and Cottonelle. In the early 1980s, Kimberly-Clark decided to turn its private fleet into a public passenger airline, and in 1984, Midwest Express was born. The airline, which was still wholly owned by Kimberly-Clark, acquired three 1960s-era Douglas DC-9s and began offering scheduled services to and from its base of operations in Milwaukee, Wisconsin.

On the 6th of September 1985, one of these DC-9s arrived in Madison, Wisconsin for a flight to Milwaukee and then on to Atlanta, Georgia. In command of the flight were two captains: Danny Martin, 31, and William “Bill” Weiss, 37. The two captains had been paired together for a 2-day trip; on the first day, Weiss carried out the duties of captain, and then they switched places for the second day. On flight 105 to Atlanta, the less senior Captain Martin would be in command, while the more senior Captain Weiss would act as first officer.

The scheduled route of flight 105. Map source: Google

Neither pilot was exceptionally experienced. Both had been fast-tracked to promotion to Captain after accumulating only 500 or 600 hours on the DC-9, an order of magnitude less than at larger US carriers, but they had since racked up a few thousand more. Weiss was also a check airman, a line pilot specially certified to evaluate the performance of other pilots, and Martin had flown with him before in this capacity.

With Martin in the Captain’s seat, flight 105 arrived at Milwaukee’s General Mitchell Field and exchanged passengers for the next leg to Atlanta. In all, 27 passengers and four crew boarded the aircraft, putting it at less than half capacity. Six of the passengers were employees of Kimberly-Clark, one of whom rode along in the cockpit jump seat behind the pilots.

At 3:20 p.m., flight 105 taxied to the runway and began its takeoff roll. All was normal as the DC-9 accelerated past its decision speed and lifted off into the cloudless sky, climbing steadily through 200, 300, then 400 feet. Then, without warning, the right engine failed catastrophically.

The origin of the failure lay in a component called the 9–10 high pressure compressor sleeve spacer. Inside a turbojet engine, a series of rapidly rotating disks compress and accelerate incoming air to generate thrust; these “compressor disks” are all mounted sequentially along a central drive shaft. Spacers are placed between the disks to keep them properly aligned. Because the spacers are subject to extreme temperatures and rotational forces, they will break down over time, causing cracks to form in the metal. Sometime in the 1970s, such a crack took root in the outer edge of the spacer between compressor disks 9 and 10, progressing inward a little bit farther every time the engine started and stopped — a classic case of metal fatigue.

Diagrams show the location and purpose of high pressure compressor sleeve spacers. Image sources: MachineDesign.com and the NTSB

In 1981, the engine went in for an overhaul at the Miami-based engine workshop AeroThrust. During the overhaul, the spacers were inspected for cracks using a visual method involving fluorescent dye. At that time, the crack should have been easily detectable, but for some reason AeroThrust failed to spot it. One plausible reason for this may have been impaired eyesight on the part of the inspector — in fact, there were no records to indicate that he had received any of his required annual eye exams. One way or another, the cracked spacer was put back into the engine, out of sight and out of mind where it couldn’t be inspected again until the next engine overhaul.

Over the next four years, the crack continued to grow with every cycle until the spacer proved unable to hold itself together. As Midwest Express flight 105 climbed through 450 feet after takeoff from Milwaukee, it suddenly split apart, separated from the rapidly spinning turbine shaft, and flung itself outward at enormous speed. Pieces of the spacer tore through multiple layers of metal and shot out the top of the engine, their sheer momentum propelling them through everything in their path.

A map of where the ejected engine parts were eventually found. Image source: the NTSB

In the cockpit, the pilots heard a loud “clunk,” followed by a slight right yaw due to the loss of power on the right engine. Captain Martin instinctively countered with a left rudder input, keeping the plane straight and on course.

“What the hell was that?” Martin asked. There was no response from First Officer Weiss. Two seconds later, he asked again, “What do we got here, Bill?” Once again, he was met with silence.

Inexplicably, Martin then let go of the left rudder pedal and started applying a right rudder command instead. Due to the asymmetric thrust caused by the failed engine, the plane already had a slight desire to pull to the right, which his input severely exacerbated. The DC-9 started yawing heavily, flying off-kilter like a drift car going into a turn.

Martin started to say “Here — ” but was interrupted as First Officer Weiss transmitted a message to air traffic control.

“Midex 105, roger, uh, we’ve got an emergency here,” he said, without elaborating on what type of emergency. Astonishingly, it appeared that Weiss was fully aware of the abnormal situation, and had been outright ignoring his captain’s questions!

Within seconds, Martin’s incorrect right rudder input had put the plane into such an extreme yaw that the right wing, which was effectively trailing behind the aircraft, started to lose lift. The plane banked to the right as the wing dipped, entering a descent, but instead of countering with a left turn, Martin instead pulled back on his control column to climb.

Now in a very steep right turn with airspeed decreasing, the plane entered what is known as an accelerated stall. When a plane is in a steep turn, the lift component acting on the wings is significantly offset from the vertical, diluting its ability to counter the downward pull of the aircraft’s weight. This causes the aircraft to stall at a much higher forward airspeed than it would if it were in level flight. Midwest Express flight 105 reached this high stall speed so quickly that the stall warning activated for only a second or two before the right wing stalled and fell like a rock. Because the right wing stalled before the left wing, the plane rolled violently to the right, flipping upside down in a maneuver witnesses likened to a barrel roll. There was no time for the pilots to react.

My original drawing of the accident sequence.

Martin let loose an expletive as the DC-9 rolled inverted. In the cabin, a flight attendant could be heard shouting, “Heads down! Heads down!” in an attempt to get the passengers to assume the brace position. In a mere five seconds, the plane slipped into a corkscrew dive and plunged straight toward the ground just beyond the edge of the airport. The DC-9 exploded on impact, flattening a stand of trees in a nature conservancy and sending a fireball curling up over the southern end of Mitchell Field.

Witnesses and firefighters rushed to the scene, but found the wreckage engulfed in a raging inferno that precluded any possibility of survival. Emergency workers could only recover the badly burned bodies of the 31 passengers and crew, along with the remains of a deer that was in the wrong place at the wrong time.

An NTSB diagram of the flight path (likely much more accurate than my drawing!).

Investigators with the National Transportation Safety Board arrived on the scene within hours, by which point the media had already begun reporting rumors that an engine had fallen off the plane in flight. The presence of both engines at the scene of the crash proved that this did not happen, but they soon discovered that some of the right engine was in fact located elsewhere. Pieces of the 9–10 high pressure compressor sleeve spacer, along with a few chunks of the surrounding engine structure, littered the runway some distance past the point at which flight 105 lifted off. Evidence of fatigue cracking on the spacer showed that it had caused the uncontained failure of the engine. There was just one problem: an engine failure on takeoff should never cause a crash by itself. A two-engine plane can continue to climb normally with only one engine, and pilots are drilled to handle such a failure dozens of times throughout the course of their training. So what went wrong?

An aerial view of the crash site. Image source: Paul Piaskoski

A comprehensive analysis of the engine failure characteristics and the wreckage of the tail section showed that there was no possibility of flying debris having damaged the flight controls. Nor was the asymmetric thrust, brought on by the failure of the engine, sufficient to cause the extreme right yaw. The only explanation was that 4 seconds after the failure, for some reason, the captain deliberately steered into the failed engine — despite initially countering it correctly. This went completely against the most basic training received by every pilot of multi-engine airplanes. To handle an engine failure on takeoff, a pilot must always use the rudder to counter the resulting yaw, stabilize the flight path, and climb to a safe altitude before returning for an emergency landing. It’s a simple procedure that both Martin and Weiss had performed countless times during initial and recurrent training.

Investigators present one of the black boxes to the media, with the crash site in the background. Image source: Mark Reinstein

Conspicuously absent from the cockpit voice recording was any mention of an engine failure. In fact, Martin repeatedly asked Weiss what was happening, indicating that he might not have realized an engine had failed. Investigators examined the simulator training he had received and found that nearly all of the simulated engine-failure-on-takeoff scenarios took place after decision speed but before liftoff, where the runway centerline and other visual cues could be used to demonstrate the yaw effect brought on by the loss of an engine. But the engine failure on flight 105 occurred after liftoff, while the plane was pointed 12 degrees nose up, straight toward an empty blue sky. In the absence of visual cues, he might have fallen victim to the physiological effect known as spatial disorientation.

Our internal sense of balance, position, and motion is intrinsically tied to visual cues. While our bodies may feel a change in speed or direction, if there are no visual cues to support it, the brain struggles to interpret the physical signals it’s receiving. With nothing but sky out the front windscreen, Captain Martin might have felt the yaw to the right and instinctively countered with a left rudder input, only to become disoriented as his field of view didn’t change to reflect the motion that he was feeling. This would have made him uncertain that his left rudder input was actually the correct action to take. At around this time, there was a slight reduction in thrust on the left engine as well, potentially drawing his attention to the motion of the needle on the left engine power gauge. In contrast, the right engine had already stopped generating any power, causing the needle on the gauge to rest at zero, where it didn’t move and was therefore less likely to grab his attention. It’s possible that he initiated the right rudder input in an attempt to counter what he perceived to be a loss of thrust on the left engine, when really the right engine was the problem. Once he did so, the same lack of visual reference points would have hindered his ability to notice the dangerous attitude of his aircraft.

Investigators work at the crash site. Image source: Paul Piaskoski

But that was only part of the story. The cockpit voice recording also showed that First Officer Weiss never said a word to Captain Martin after the onset of the emergency, even though Martin directly asked him for input. He clearly had not been incapacitated because he spoke to air traffic control, he was aware of the emergency because he said as much over the radio, and he must have heard his captain’s questions because they came through loud and clear on the recording. And yet he did nothing to help Martin understand what he was dealing with, even though the Midwest Express flight operations manual explicitly stated that the first officer should scan the instruments and point out any problems or malfunctions to the captain. So why did he ignore his fellow pilot right in the moment of greatest need?

A salvage vehicle removes an engine from the site of the crash. Image source: Paul Piaskoski

The NTSB was shocked to discover through interviews with Midwest Express pilots that the airline had an unofficial, unwritten rule called the “silent cockpit philosophy.” The idea behind the “silent cockpit” was that a pilot’s first priority when dealing with an emergency after decision speed and below an altitude of 800 feet was to simply fly the plane. Any discussion of the situation would distract from that most critical task; the conversation could wait until the plane was stabilized at 800 feet. The policy specifically discouraged even basic callouts identifying the nature of the emergency. Despite the fact that the silent cockpit philosophy directly contradicted the flight operations manual, it was taught to Midwest Express pilots during training, and the airline’s chief pilot defended it at an NTSB hearing. The Safety Board, on the other hand, found the policy to be downright dangerous and probably in violation of federal regulations. In fact, the flight manual itself already explained why this was problematic: “Crewmembers shall never assume,” it stated, “that other crewmembers are also aware of [a problem] without verification.”

An investigator stops for a moment to gather his thoughts amid the field of wreckage. Image source: Paul Piaskoski

It wasn’t entirely clear that First Officer Weiss ignored Captain Martin because of the silent cockpit philosophy. After all, his transmission to air traffic control would have violated it. But it certainly might have had an influence on his understanding of what ought to be done when facing an emergency on takeoff, and could explain his baffling lack of answers to direct questions from his captain. The Safety Board added that confusion over what was happening could also have prevented him from coming up with an answer. If so, he had failed in his role as first officer, in which his primary duty was to monitor the instruments and advise the captain. Why he didn’t see the indications of an engine failure on his instrument panel is unclear, although perhaps he was distracted looking out the window for nearby traffic.

Regardless of the reason, Weiss’ silence likely added to Captain Martin’s disorientation as he continued to steer in the wrong direction. Had Weiss come out and said “Failure engine number two,” Martin’s training would have kicked in and he would have known exactly what to do. Instead, the pilots wallowed in their own helpless confusion until they lost control of the plane. After that, disaster came swiftly: in all, only 15 seconds passed between the engine failure and the moment of impact.

A fire chaplain gives last rites to victims of the crash. Image source: Paul Piaskoski

This tragic breakdown of communication underscored the importance of proper crew resource management. The philosophy of crew resource management, or CRM, holds that the cockpit should be an open space wherein pilots say their intentions out loud, actively discuss problems, correct each other’s mistakes, and create a shared perspective on the state of the flight. Training pilots in CRM strategies had been shown to significantly reduce pilot errors, but in 1985, the majority of US airlines had not yet incorporated it into their formal training programs, and indeed they were not required to do so. Midwest Express was no exception, and in fact its informal “silent cockpit” philosophy actively undermined effective CRM. While the airline was technically correct that the highest priority when facing a low-altitude emergency is to fly the plane, the crash of flight 105 proved that this isn’t possible if the pilots don’t work together to understand what they’re facing.

The plane’s landing gear lies at the edge of the scorched impact area. Image source: On Milwaukee

To understand how the pilots acquired their airmanship skills, the NTSB looked into the training they received and the FAA’s oversight of that training. What they found was surprising: the FAA inspector assigned to Midwest Express apparently lacked basic qualifications for the job. The inspector had never flown for a commercial airline, had no experience with jet aircraft operations, and knew little about the Douglas DC-9. She came from a general aviation background and was only familiar with small, private aircraft. As a result, she spent most of her time overseeing general aviation in the Milwaukee area, rather than Midwest Express, the only commercial airline under her purview. When it came to matters associated with the airline, she usually delegated certification tasks to inspectors in other areas who were more familiar with the DC-9 and the regulations applicable to it. She took these inspectors on their word and signed off on anything that they approved of, leaving her completely unfamiliar with the policies and procedures used by the airline she had been appointed to oversee. She even testified that she had no idea what Midwest Express’ engine failure on takeoff procedure was, or that pilots were being trained in the “silent cockpit” philosophy. In a moment of frank clarity, she stated during the hearing that she was uncomfortable with her appointment to this position, for the obvious reason that she was not qualified. Investigators believed that an FAA inspector who was capable of being more involved in Midwest Express’ daily operations would have recognized the danger of the silent cockpit philosophy and stopped it in its tracks. And this wasn’t the only thing that escaped the inspector’s attention: although she had approved Captain Weiss’ certification as a check airman, the NTSB could find no evidence that he had completed all of the required training that came with that certification.

The headstone of Captain Roger William Weiss. It is important to remember that pilots who make mistakes are victims too. Image source: Find a Grave

Regarding pilot training, the evidence showed that pilots do sometimes make inexplicable errors — and that proper communication is the best way to prevent this. DC-9 flight instructors testified that about 1 in 50 line-certified pilots pushed the rudder pedals the wrong way in response to a simulated engine failure during recurrent training. In such situations it’s critical that the other pilot make it clear which engine is failing and what ought to be done about it. This is why the principle of CRM is so critical, and why the silent cockpit was so dangerous. In its final report, the NTSB recommended that airlines provide formal crew coordination training, including explicit requirements that any pilot who detects a potentially dangerous situation verbalize this to their fellow crewmembers. The Safety Board addressed two further recommendations to the FAA, asking that inspectors ensure airline training programs include engine failure scenarios without visual cues, and that FAA inspectors overseeing commercial airlines possess training and certifications that match up with the types of carriers under their jurisdiction.

A chart shows how airline deaths have trended downwards massively since the peak in 1985. Note that this represents total deaths, not death rates, which have dropped even more drastically as the number of passengers increased several times over during this period. Data source: the Bureau of Aircraft Accidents Archives

The crash of Midwest Express flight 105 came at a moment when the aviation industry was already facing a moment of reckoning. The 31 fatalities in Milwaukee were enough to push the year’s total airline deaths over the record set in 1972, making 1985 officially the deadliest year in the history of commercial aviation, even with nearly four months still to go. Just between June 23rd and September 6th, six major commercial airline crashes had claimed 1,268 lives. When news broke of the Midwest Express crash, it was framed as “yet another airliner accident” in a long string of airliner accidents that seemingly refused to end.

But 1985 was also a turning point in aviation safety. The scale of the tragedies and the glaring safety gaps that caused them galvanized the authorities to act. In fact, despite the constantly increasing popularity of air travel, 1985 is still the deadliest year in commercial aviation history, with fatalities decreasing with every subsequent decade. By the 1990s, US passenger airlines had adopted CRM training en masse, in what proved to be the most important step in making this improvement possible.

Midwest Express, which played a small if noteworthy part in that overarching story, was able to recover from the crash and continued flying passengers until 2010. It had learned its lesson the hard way, so that other airlines might not have to.

__________________________________________________________

Click here to join the discussion of this article on reddit

Subscribe at reddit.com/r/admiralcloudberg to read over 100 more articles, check for updates on my book, and discuss aviation safety with me and others.

--

--

Admiral Cloudberg

Kyra Dempsey, analyzer of plane crashes. @Admiral_Cloudberg on Reddit, @KyraCloudy on Twitter and Bluesky. Email inquires -> kyracloudy97@gmail.com.