The Crashes

Certainly, the trip from Jakarta to Pangkal Pinang was one they had made many times before. From almost the moment flight 610 lifted off, however, Suneja’s control stick began to shake: the aircraft was warning him that something was wrong.

What neither man knew was that this particular aircraft had done this before.

The previous day, alarm bells were ringing on the flight deck of the very same aircraft
On 28 October, the crew of Lion Air flight 43 prepared to depart Ngurah Rai airport in Bali for a nighttime flight to Jakarta. As part of the aircraft handover, the captain was informed that instrument errors had been reported on this particular 737 MAX 8 over a number of recent flights. Each time, ground crew had attempted to fix the problem but had little success. In the end, the arrival of the aircraft in Bali had been treated as an opportunity to replace the angle-of-attack vane on the captain’s side of the nose, on the left-hand side of the aircraft.

The captain of the flight noted the change and continued preparations to depart. At about 11pm local time, Lion Air flight 43 pulled up from the runway and started the six hundred mile hop to Jakarta.

As the aircraft began to climb, however, it soon became clear that something wasn’t right. Almost immediately, the captain’s airspeed indication failed and warnings of airspeed and altitude disagreement appeared. At the same time the captain’s control stick began to shake, a physical feedback warning triggered automatically when readings suggested the aircraft was about to stall.

None of these problems was, in itself, a critical issue — at least not yet. While uncommon, such issues are the reason why aircraft such as the 737 MAX are equipped with multiple instruments on both sides of the aircraft.

In essence, the duplication is because it is more likely that such a warning will be triggered by an instrument failure as by a real flight control issue. By definition, instruments such as the angle-of-attack vanes have to be exposed to the elements and effects of flight. That means they are always at risk of physical failure.

To mitigate this risk, the 737 MAX was fitted with more than one angle-of-attack vane, and the pilot and co-pilot received their readings from separate sets of sensors, with a third backup set also visible to both of them. This combination of backups allowed the flight crew to do something that the aircraft’s automated systems could not do: take in the sum of the information available and work out which of the three readings to discard as false.

On Lion Air flight 43, the captain correctly identified that the aircraft was suffering from an instrument failure, most likely a problem with the newly fitted angle-of-attack vane on his side of the aircraft. It was his readings that were out of sync, and it was only his stick that was vibrating. As a result, he ignored the warnings and continued his climb. When he deemed it safe to do so, he then transferred control to his first officer.

This should have been the end of the issue: the flight continued to Jakarta as scheduled and the faulty vane — and the accident investigation would prove that it was indeed faulty — should have been logged and dealt with there.
ET-AVJ was the aircraft involved in the Ethiopian Airlines 302 crash. LLBG SPOTTER

As the aircraft began to enter its regular flight, a new problem appeared
Aboard flight 43, the co-pilot retracted the flaps and began to trim the aircraft for regular flight, and the situation deteriorated rapidly. Without warning, the aircraft suddenly trimmed itself, pitching downwards.

To the pilots on the flight deck this must have appeared to be a sudden case of runaway trim.

Again, this is a problem that is serious but not necessarily critical. The trim of an aircraft must be adjusted to ensure level flight, and the exact level of adjustment required changes based on loaded weight and flight characteristics.

In jet passenger aircraft such as the 737 MAX this is generally handled automatically and electronically to take the burden off the pilot. Occasionally, however, the rapid micro-adjustments this process involves can go wrong and, when that happens, the aircraft will pitch nose-down or nose-up.

When this happens, the solution is normally relatively simple: minor issues can be solved by one or more of the pilots applying counter trim. What seems to have caught the crew of flight 43 off guard, however, was how violent the sudden pitch change had been. That and the fact that it started to come in sudden, brutal, bursts. The aircraft would fight them for ten seconds then stop for five, then the whole cycle would start again.

The repeated, stacking effect of these bursts soon began to take their toll. It wasn’t long before the first officer began reporting that his control stick had become so heavy he was struggling to hold the aircraft’s nose up at all. Meanwhile, the captain scoured the various emergency checklists available to them to try and find out what was wrong. It was an escalating situation that could have ended badly, but at this point a piece of luck intervened: there was a third man in the cockpit.

The identity of that man remains unknown. His presence was not reported in the initial accident report. Indeed it was a later investigation into the Lion Air accident by the New York Times that forced the Indonesian authorities to acknowledge his presence at all. The Times’ suggestion that he was another, more experienced, Lion Air pilot has never been confirmed.

Whoever he was, it seems that it was on his suggestion that the captain carried out the action that saved the flight: he disabled the automatic trim completely. This was easy enough to do, requiring only the flip of a pair of cut out switches in the cockpit. It was something that both captain and first officer had overlooked it as the situation escalated. As soon as this was done, the problem disappeared.

Shaken, but back in control, flight 43 continued on to Jakarta.

As the aircraft began to climb, however, it soon became clear that something wasn’t right

On departure from Jakarta, the cycle began again

We do not know how familiar Suneja and Harvino were with the problems that had affected their aircraft the day before. Certainly, the maintenance logs make no mention of the severity of the issue. They note the issue with the indicators, but not the near-catastrophic consequences. Nor does any serious investigation appear to have taken place into the potential issues with the left-side angle-of-attack vane.

All of this meant that as Suneja and Harvino began to experience the same problem during their own takeoff, they had no idea that this had also occurred the day before. Instead, the whole cycle of failure began again.

The crew of flight 43 may well have mastered the problem that Suneja and Harvino now faced, but neither they, nor the maintenance crew at Jakarta, had actually identified what had caused the trim issue to begin with: incorrect data from the faulty vane, which was activating the aircraft’s MCAS system.

Indeed it is possible that both crews weren’t aware of MCAS at all, as the 737 MAX was one of the few types of aircraft fitted with this feature. The existence of MCAS, and the extent to which it was communicated to pilots, is the subject of multiple ongoing investigations.

When Suneja’s control stick started shaking, he began experiencing the same problems that had plagued the aircraft the day before. As the accident report highlights, however, Suneja and Harvino handled the problem very differently to their colleagues the previous night.

It remains unclear whether the flight crew correctly identified the mismatch in angle-of-attack data, but Suneja’s actions suggest that they did not, at least not initially. Rather than continuing to climb before handing over to his first officer, Suneja retained control. Approximately two minutes later, Harvino contacted the controller and reported a flight control problem, requesting a holding altitude of 5,000ft.

It seems their intention was to reach this holding position and then try and work through the problem, although the altitude they requested was surprisingly low. This may have reflected their lack of certainty about the nature of the issue, a possibility backed up by Harvino’s comment about it being a flight control problem. As before, at this stage it wasn’t.

As the previous crew had inadvertently discovered, however, the moment they retracted the flaps MCAS kicked in. What had previously been an instrument problem was now very much a flight control one, and the aircraft pitched itself downwards.

That both crews had found themselves fighting MCAS was down to a critical flaw in its design. For reasons still not adequately explained, which are the subject of strong speculation within the industry, MCAS had been designed to take its angle-of-attack data from only one sensor: the one on the captain’s side of the aircraft.

This was as unknown to Suneja and Harvino as it had been to their fellow Lion Air pilots the day before. Suneja and Harvino found themselves locked in a similar battle against the aircraft to that which the previous crew had experienced.

Barely twelve minutes after it had taken off from Jakarta, flight 610 disappeared off radar

The same sudden violent nose-down bursts. The same increasingly desperate attempts to retain altitude as the stick got heavier and the effects escalated. The same desperate search through checklists to try and work out what was going wrong.

The solution was also right in front of them, of course. It was exactly the same as it had been the night before: switch to manual trim control.

Unfortunately for Suneja and Harvino, and for the other 187 people aboard, this time there was no extra voice of experience in the cockpit. Perhaps the two pilots would have eventually realised the solution without one. Yet by levelling out early they had arguably robbed themselves of the resource they needed most, time.

Barely twelve minutes after it had taken off from Jakarta, flight 610 disappeared off radar and crashed into the ocean.

Six minutes in Ethiopia

Four months later, at 8:38am on 10 March 2019, Ethiopian Airlines flight 302 took off from Addis Ababa’s Bole International Airport, headed to Jomo Kenyatta International in Nairobi. It would never arrive.

The flight had an experienced captain in command, Yared Getachew. Getachew was only 29, but he had already amassed over 8,000 hours flight experience. 1,417 of those hours were on the 737, although only 103 of these were on the 737 MAX, the aircraft he found himself operating. At just 25 his flight officer, Ahmed Nur Mohammod Nur, was considerably newer to the role. Although fully certified, he had only mustered 361 hours flight experience. 207 of those were on the 737, with only 56 on the 737 MAX.

As their aircraft left the runway at Addis Ababa, Getachew’s stick immediately began to shake. The aircraft, it seemed, believed it was about to stall. Unbeknownst to Getachew and Mohammod, they were experiencing the same problem seen on the Lion Air flights: the angle-of-attack vane on the captain’s side of the aircraft was faulty, and it was misreporting data.

Given that the aircraft was still at takeoff speed, Getachew must have doubted that a stall was truly likely and the aircraft continued to climb. Certainly, the flight crew’s actions, detailed in the provisional accident report released in April 2019, suggest that they didn’t initially panic.

Instead, the flight crew turned on the autopilot and Mohammod adjusted the settings, telling it to take the flight up to 32,000ft. On his captain’s orders, Mohammod then retracted the flaps and warned the controller that they had a flight control problem. It was the same initial misdiagnosis that Suneja and Harvino had made.

“Stab trim cut out”: MCAS, once again, intervened
One thing that remains unclear is why, with the captain’s (that is, the left-side) control stick shaking, it was the left-side autopilot that the crew engaged. Whatever the reasoning, the provisional accident report suggests it may have been the wrong thing to do. Roughly ten seconds later, perhaps in response to the fluctuating, faulty data it was receiving, the left-side autopilot automatically disengaged.

As soon as it did, the conditions for MCAS to activate were met. Almost straight away, looking at that faulty left-side data, MCAS attempted to pitch the aircraft violently downward.

Getachew instinctively reacted the same way Suneja had done before him. As he fought the MCAS correction, Mohammod contacted the controller and requested runway heading. As before, the captain’s initial efforts to hold altitude largely managed to counteract the MCAS effect. Five seconds after the first burst, however the MCAS kicked in again.

Getachew shouted for his first officer to help and together the two men fought MCAS. It was now, two minutes into the flight, that Mohammod realised what they needed to do.

It is possible that this was because he now believed they had a runaway trim problem. More likely, it was because the flight crew had finally recognised what was happening. Following the Lion Air incident, Boeing had issued a Flight Crew Operations Manual Bulletin flagging up that a faulty angle-of-attack vane could cause this exact effect. Ethiopian Airlines would later confirm that both men had been fully briefed on this addition.

“Stab trim cut out, stab trim cut out,” Mohammod called. Getachew agreed and the first officer flipped the switches. It was just in time. A few seconds later MCAS tried to activate again. With the cutout switches flipped, however, it was unable to do so.

MCAS may have been disabled, but the situation was already spiralling out of control. Flipping the cutouts had fixed the trim in the last position occupied, leaving Getachew struggling to hold the aircraft level.

“Pull up, Pull up, Pull up,” he called out to Mohammod. Both men tried worked together in a desperate effort to gain height, but their actions had little effect.

The only solution to this problem seemed to be to manually correct the trim, and Mohammod now attempted to do so via the manual trim wheel in the cockpit. Likely to his horror, he soon discovered that this didn’t work.

The reason for this was cold, hard aerodynamics. Whether deliberately or accidentally, the flight crew had never throttled the aircraft back from takeoff power. Combined with the bursts of altitude loss triggered by MCAS, this had pushed their airspeed beyond 300kts to the edges of the 737 MAX’s operational limit. The manual, mechanical system simply wasn’t strong enough to adjust the aircraft’s trim at this speed.

With the sound of the overspeed clacker filling the cockpit, the two men seem to have decided that they only had one option left. To trim their aircraft properly, they needed the help of the aircraft itself. They needed to trigger the stronger automatic (and electronic) trim. At 05:43:11 they flipped the cutout switches back. Then, seemingly in an effort to beat the inevitable activation of MCAS, Getachew attempted to activate the autopilot once again.

It refused to engage.

Five seconds later, MCAS kicked back in one final, fatal time. As the two men fought vainly to try and prevent it, flight 302’s nose dropped even further. Now moving at more than 500kts, it plunged into the ground.

In the aftermath, the investigation is ongoing
189 passengers and crew died on flight 610. Flight 302 resulted in the deaths of 157 passengers and crew.

Both accidents have raised questions and criticisms about maintenance and flight crew experience and training, not just for the airlines involved, but for the wider industry.

The rapid growth of budget airlines, as well as the emergence of new markets, has increased the need and opportunity for pilots and ground crew, while at the same time new regulation and other factors have created a constraint on supply.

The question is whether meeting this need has come at the cost of quality and experience, something that rarely seems to matter… until it really matters.

Important questions also remain over the role of MCAS in both accidents, as well as the decisions made by Boeing in its creation and rollout. The Flight Crew Bulletin added to the 737 MAX operations manual in November 2018, seen by Getachew and Mohammod, arguably represented the first acknowledgement by the company that MCAS had been a contributing factor to the Lion Air accident at all. The loss of flight 302 demonstrated that there were more issues with MCAS’s implementation than could be explained away by a simple lack of pilot knowledge.

Two transportation safety investigations were launched after the crashes.

The investigation into Lion Air flight 610 was conducted by the Indonesian National Transportation Safety Committee (NTSC, also known as Komite Nasional Keselamatan Transportasi, KNKT). Its final report (KNKT.18.10.35.04) was published on 25 October 2019 and is online: https://knkt.dephub.go.id/knkt/ntsc_home/ntsc.htm

The investigation into Ethiopian Airlines flight 302 is being conducted by the Ethiopian Aircraft Accident Investigation Bureau (AIB). Its provisional report (AI-01/19) was published on 4 April 2019 and is online: https://www.ecaa.gov.et/home/resources/resources/

Both Indonesia and Ethiopia are signatories to the Chicago Convention of the International Civil Aviation Organization (ICAO). In accordance with the convention, investigators from the US National Transportation Safety Board (NTSB) have participated in both investigations.

Interpretation and learning from the investigations are ongoing.

In June 2019 captain Chesley “Sully” Sullenberger, the pilot who successfully landed US Airways flight 1549 on the Hudson River, testified in front of a US Congressional Committee that he had been involved in a simulator recreation of the flight 302 accident.

“Even knowing what was going to happen,” he commented, “I could see how crews could have run out of time and altitude before they could have solved the problems.”

By that point it seemed that much of the aviation industry already agreed. By April 2019, the majority of the world’s civil aviation authorities had grounded the Boeing 737 MAX.

The aircraft remains grounded today.
I could see how crews could have run out of time and altitude before they could have solved the problems
“SULLY” SULLENBERGER