It’s beginning to sound more and more like the plot of a blockbuster by Dale Brown – or more possibly Michael Crichton’s thriller Airframe.
Airbus drives the agenda
The story goes back to the first decade of the millennium when Boeing was working on an all-new narrow-bodied airliner to replace its best-selling short-haul 737 twin jet. Its replacement was widely expected to have a light-weight fuselage with an elliptical cross section, advanced systems, greater capacity than the then current 737s, and space under the wings to install the latest fuel-efficient turbofan engines.
Meanwhile, Airbus was cracking on with its own project to update the A320 family, a type that was two decades younger than the original 737. The European manufacturer launched the A320neo in December 2010 with a timeline that was planned to lead to service entry four years later. The updates were relatively modest, primarily offering customers a choice of two new engines – the Pratt & Whitney PW1000G or the CFM International Leap-1A – each of which were somewhat more efficient than those they replaced. Crucially as it turns out, the original A320 stood on fairly tall undercarriage which left Airbus plenty of room to hang the new, larger engines below wing level.
Enter the MAX
Boeing was backed into a corner. It needed to do something to compete with the updated Airbus, which burned less fuel than the equivalent 737s. It was in the midst of its pre-production problems with the 787, which were delaying the program and costing the firm a great deal of money. The advanced new widebody was taking much longer to bring to market than expected, and Boeing already knew from the 787 roll-out fiasco that ground-up programs sometimes run into problems that delay delivery of early airframes.
An all-new aircraft was preferred by Boeing, but finding a cheaper solution was a potential way out for the firm when its more loyal customers were pressing for a counter to the updated Airbus. Boeing’s third generation 737, the NG (next generation), was selling well – around 370 delivered in each of 2009, 2010 and 2011 – and the firm was anxious not to cede the market to its rival. Short-haul narrow-body jets are the lifeblood of aircraft manufacturers and Boeing also had to figure out how to work through an inevitable dip in orders and output of the outgoing aircraft, if it transferred production to an all-new airframe. There was another issue Boeing had to consider too. All new aircraft have to undergo extensive testing – often taking as long as two years – before they can be awarded a type certificate of airworthiness. But under certain circumstances, updates to older designs can be ‘grandfathered’ approval – the testing is still rigorous, but not as quite as extensive or time consuming as an all-new aircraft. Grandfathering also means that certification can be a little less rigorous than it would be for an all-new aircraft, as some aspects of aderivative design are measured against the original rather than current and potentially more stringent standards.
The 737-MAX was a cheaper and more expedient solution than an all-new design that enabled Boeing to compete against the Airbus A320neo (Clemens Vasters)
While Boeing wanted to design an all new aircraft, an update to the venerable 737 would be cheaper and could be brought to market more quickly. Boeing’s customers also liked the idea of a 737 update because pilots could easily be switched between the 2nd generation classics and the newer generations. Differences training is required for new models, but it isn’t as time consuming or expensive as transferring crew to a completely new aircraft type. More to the point, pilots could be scheduled to fly either the 2nd generation classics or the NG.
1st generation (classic) – 737-100 / -200
2nd generation (classic) – 737-300 / -400 / -500
3rd generation (NG) – 737-600 / -700 / -800
4th generation (MAX) – 737-7 / -8 / -9 / -10
So some of Boeing’s major customers pressed for yet another update – the third – to the venerable 737 as a cheaper, quicker and more easily adopted solution to their need for better economy and greater capacity. However, there was a problem – the 737 was designed when jet engines were long and thin. That meant the powerplants of the original version could be slung under its wings. And that allowed short undercarriage legs which enabled ground crew to access the baggage holds without the need for complex and expensive ground equipment.
When the original 737 classic was upgraded with CFM-56 engines in the 1980s, the nacelles had to be flattened at the bottom to provide sufficient ground clearance. By the time the fourth generation 737 came along, the latest engines were larger still - as well as being much more efficient again. Boeing selected a derivative of the CFM International Leap engine for its updated 737, dubbed the 737-MAX. The Leap-1B it used had to have a smaller diameter front fan than the -1A hung on the competing Airbus to enable it to fit on the 737. Reducing the diameter of the fan wasn’t enough though - it had to make other changes too.
Why didn’t Boeing just increase the height of the 737’s undercarriage? Well, the 737 main wheel legs fold inwards into bays in the lower fuselage.
The larger diameter CFM Leap-1B engines on the 737-MAX were installed further forward and canted upwards slightly to ensure they would not hit the ground. (Anna Zvereva)
There wasn’t much lateral space between the main undercarriage mounting points on the original 737. And if longer legs had been installed, essential structure in the lower centre fuselage would have obstructed the wheels before the gear was completely stowed in the under-fuselage bays. Boeing could have moved the legs further outward on the wings, but that would have interfered with the placement of the engine mounts. That in turn would have necessitated moving the engines further out too, and would have led to a major wing redesign. And that would have jeopardised Boeing’s ability to certify the 737-MAX as a derivative of the original 737, potentially throwing away all the benefits of grandfathering certification and pilot commonality.
As a result the Leap-1B engines on the 737-MAX were instead moved forward, raised a little, and canted upwards slightly at the front to provide sufficient ground clearance. That approach worked. Boeing was able to offer its customers a cheap and timely competitor to the Airbus A320neo, and the airlines got an aircraft that could be flown by current 737 pilots following minimal retraining. The 737-MAX is actually a family of four aircraft (737-MAX7 to -MAX10), each with successively longer fuselages and greater passenger capacity. Airlines signed up in droves – there were commitments for almost 5,000 by the end of 2018, making the MAX Boeing’s fastest-selling aircraft in the firm’s history.
In to service
The A320neo flew for the first time on September 25, 2014 and received joint certification from the European Aviation Safety Agency and the US Federal Aviation Administration on November 24 the following year. The new Airbus derivative entered service with Lufthansa on January 25, 2016. Four days later – 49 years after the original 737-100 took to the air - the first 737-MAX8 flew from Renton near Seattle.
Launch customer Southwest Airlines received its first MAX at the end of August 2017, and put it into service around a month later. Its pilots, all of whom were familiar with the 737 already, received differences training. As more airlines took delivery of MAX aircraft they mirrored this approach. In most cases the conversion comprised a one hour session with an iPad or computer based training package which explained the differences between the 737NG and the 737-MAX. It was a very simple process – the content mainly concerned changes to the aircraft systems, as Boeing had strived to ensure the 737-MAX handled as much like an earlier NG as possible. In fact, the differences between the 737NG and the MAX were so insignificant that some pilots went through the course then didn’t fly a MAX for quite some time, only to suddenly find they were scheduled to fly one at a later date. It didn't seem to be a problem because the controls and the flying characteristics of each were so similar.
Everyone was happy – Boeing, its customers and the regulators – although the Federal Aviation Authority had baulked at the idea of pilots flying 737 Classics, NGs and MAXs concurrently, and instead limited aircrew interoperability to only two of the 737 sub-types at a time. The airlines that operated all three versions keenly started to retire their older 737 Classics to maintain operational flexibility.
Lion Air 610
The story then moves forward 14 months. Headlines were made on October 29, 2018 when Lion Air flight 610, operated by almost new Boeing 737-MAX8 PK-LQP, crashed into the Java Sea off the north coast of Karawang, Indonesia. All 189 on board – 181 passengers and 8 crew – lost their lives. It took time to recover the flight data and cockpit voice recorders, but eventually investigators were able to produce an interim report into what was behind the tragedy.
A fully detailed explanation of what caused the aircraft to drop into the sea has yet to be revealed, but a causal factor was found to be a new system Boeing had implemented on the 737-MAX, known as MCAS (Maneuvering Characteristics Augmentation System). Until MCAS came to the attention of the world’s media, little had been heard about it - indeed, so little that most pilots flying the 737-MAX weren’t even aware of it.
Fixing a defect
To understand why Boeing introduced MCAS, its necessary to roll back the clock to when Boeing was testing the 737-MAX. During test flying, the firm unexpectedly found that the new, larger engines that were slung forward of the wing and pointed upwards slightly at the front changed the aircraft’s flying characteristics. It wasn’t a huge thing, but it affected one unusual situation which would very rarely be encountered in passenger service. When the aircraft’s nose was pitched up significantly and it was flying at higher speeds, there was a change in handling that led to some ‘unusual’ behavior. Instead of load on the pilot’s control columns increasing as the nose up attitude became steeper, it became a little lighter.
Boeing realised that the engines, further forward than those on older 737s, were contributing to the nose-up pitch and lightening the load on the pilot’s controls. But certification standards for passenger airliners demand that as the aircraft’s nose up angle of attack becomes steeper, pilots must pull harder on the column if they want to keep the nose up. It’s a classic ‘feel’ required from all airliners - the closer they come to stalling (the point at which the wings cease to generate enough lift to remain airborne and the airframe starts to drop), the heavier the controls should become. This basic control law helps pilots identify a potential problem in the making and take steps to resolve it before it becomes more serious.
Boeing was stuck. The 737-MAX was well into its pre-certification test flying program, so a major redesign to the airframe and relocate the engines was untenable. But with the flying characteristics the aircraft was displaying, albeit in only an unusual situation that most airline pilots were unlikely to ever encounter, airworthiness authorities would not sign off on the design and allow it to enter service with the airlines.
Enter stage left: MCAS. The software system basically detects when a 737-MAX is being flown at high-speed with a significant nose-up angle of attack. For MCAS to activate the flaps on the trailing edge of the wing also have to be up and the autopilot has to be off – the aircraft will be being hand-flown. If MCAS detects this rather unusual configuration, it adjusts the horizontal stabilisers on the rear fuselage to push the nose down. It does it repeatedly until the nose goes down and the risk of a stall is averted, even if the pilots countermand it by altering the stabiliser trim manually.
The MCAS was designed to intervene inobtrusively during a critical but highly unusual manoeuvre and once its job was done it would become passive again - and its affects would be almost unnoticable and rarely if ever encountered anyway. Current philosophy suggests that pilot training should focus more on the actions necessary to resolve problems, rather than on troubleshooting and diagnosis. As a result Boeing and its airline customers didn't include details of MCAS in the information existing 737 pilots received in preparation to fly the 737-MAX. The manufacturer subsequently suggested that if MCAS develops a fault it can be disabled with the same procedure used to resolve runaway stabiliser trim, a problem that pilots practice dealing with during recurrent training in simulators.
All well and good then. MCAS resolved a minor problem with the 737-MAX’s handling characteristics that satisfied the Feds, enabling the design to be certified airworthy as a derivative of earlier 737s, and it entered service with the world’s airlines. Boeing felt that MCAS was so insignificant – and indeed the chances of it activating were so remote – that it did not include mention of the system in the differences training pilots received before they transferred onto the 737-MAX. The regulatory authorities either accepted this view or didn’t fully understand MCAS either.
Lion Air 610: MCAS activation
Actually all was indeed well and good until Lion Air 610 dived into the Java Sea. It turns out that the ill-fated 737-MAX8 aircraft, PK-LQP, had a faulty angle of attack probe which had been generating erroneous readings for at least two days. These probes, exposed to the airflow on either side of the nose, send signals to the aircraft’s systems telling them how much the nose is pitched up. On the fateful flight, the faulty probe sent incorrect signals to the aircraft's computers, which interpreted them as indicating the aircraft was flying with a significant nose-up pitch at high speed. MCAS then did what it was designed to do – it detected the likely onset of a stall and pushed the nose of the aircraft down using the horizontal stabilisers on the tail (albeit under false premises brought about by the erroneous data from the faulty probe).
It’s already been mentioned that the 737-MAX has two angle of attack probes. It turns out that MCAS takes input from only one of them at a time. And that means that if MCAS is using a probe that is faulty, it will react to erroneous data.
Angle of attack probes on the nose send critical data to the aircraft's computers. Most airlines have at least two (one on either side) while some have more. (Dtom)
There is an option on the 737-MAX to have the readings from both probes cross-checked and if there is a difference, flag an alert to the pilots. But its only that, an option that many airlines have not taken up. Presumably the authorities accepted angle of attack data discrepancies would be unlikely, or weren’t important enough to demand they were always flagged to pilots.
That’s as as far as the investigation has gone so far - to be certain of the rest, we have to wait for the official report into the Lion Air 610 disaster. However, we do know that MCAS pushed the nose of the aircraft down. The pilots hauled back on the control columns numerous times, and each time they got things almost back to normal MCAS intervened and pushed the nose down again – because MCAS, with its faulty data feed, still thought the aircraft was pitching up too steeply. But it wasn’t. It seems likely that the crew, confused by the sudden repeated pitching down of the nose, did not recognise the true cause of repeating problem soon enough to disable the automatic trim and manually take over control. The end result was that an almost new aircraft entered the sea nose down and at high speed, with tragic consequences we are now so aware of.
Almost four and a half months later in March 2019, another 737-MAX flew into the ground shortly after take off from the Etheopian capital Addis Ababa with equally tragic consequences. The preliminary report on the loss of Ethiopian Airlines flight 302 is still some way off, but observers and airworthiness authorities have noted similarities in the flight tracks, including oscillating speed and height, between JT610 and ET302. After a short pause, all Boeing 737-MAX aircraft were grounded worldwide.
Firstly, 737-MAX aircraft will remain grounded until any deficiencies in the design are resolved to the satisfaction of the world’s aviation authorities. Boeing has been working on an update to the MCAS system since the loss of Lion Air flight 610, which it says will be ready to go in a week or two. Whether that means ready for testing, or ready to be released to airlines is unclear at present. It is unlikely that regulators will take just Boeing’s word for it so there will be some flight testing takes place – if it isn’t happening already. As yet it is not a foregone conclusion that MCAS is implicated in the Ethiopian crash, but it is almost certain that a return to service will also depend on progress being made on the investigation into the loss of the second 737-MAX.
What seems certain is that the 737-MAX will with ‘fixed’ by software. MACS will be redesigned to take input from both angle of attack probes, and disabled when they disagree. It could be that further enhancements and crew alerts are necessary. But unless the challenges of resolving the issues are so insurmountable that computers can’t address them all, a redesign of the airframe to relocate the engines is out of the question. Since the dawn of the jet age, quirks in the aerodynamic design of aircraft have been resolved through fixes – in some cases these have taken the form of stick pushers that force the pilot’s control columns forward, while in others it has been done with the computers that sit between man and machine. A major redesign of the 737-MAX airframe would set the program back two year, cause chaos for airlines that need new aircraft and would be unlikely to achieve anything more than correctly functioning software. Boeing isn’t going to give up on the problem either so the 737-MAX will fly again – the firm still has more than 4,500 new aircraft to build and deliver, representing an order book value of trillions of dollars.
So the 737-MAX will be fixed. It’s unlikely to take days, it might even not be weeks, but the aircraft will be back in the air within months. Lessons will be learned and the next aircraft to be upgraded will be scrutinised more closely. Meanwhile aircraft remain groaned. Some airlines are planning for a summer without the 737-MAX while others are already threatening to seek compensation for their loses from Boeing.
Text © The Aviation Oracle