In a previous report we had a look at the crash with the Lion Air Boeing (BA) 737 MAX. It’s a report that drew quite a bit of attention with an active discussion (over 200 comments). We also know that the report reached all the way to Boeing’s desk, which we think is a good sign.
Source: The Boeing Company
At the time I drafted the report, no preliminary report had been released yet. The Indonesian safety board published a preliminary report on the 28th of November so in this report we will be looking at what that the report contains in terms of confirmations or important snippets and we will try to explain what those snippets mean or why they are useful.
Before we start…
One thing I do want to emphasize before I look at the report is that while more than 90% understood the takeaway from the report namely that this accident like many is not the result of a sole factor or a sole decision, but many factors contributed to what eventually became a fatal crash. Those factors can include suboptimal design, documentation, training, maintenance and handling. As I said most people understood that, just some are insisting that this either is a mistake that Boeing or the crew can be fully blamed for. Some people also wanted the report to focus on the stabilizer cutout switches present on the Boeing 737, which – if used – could have prevented a fatal crash. While I do believe that the use of the cutout switches would have prevented the crash it doesn’t mean that Boeing or maintenance did everything perfectly. Only if you insist on getting black-and-white conclusions, you would get to that conclusion and that really is a conclusion against better judgment. The report I wrote was there to show some of the simple physics and systems that go behind the Boeing 737 MAX and airplanes in general and show that even if either Boeing or the crew didn’t do things perfectly – in fact they did things far from perfectly – it still doesn’t mean you can blame either party fully for the crash. It also wasn’t my intention to show how a crash could have been prevented, so solely focusing on the cutout switches would be close to useless.
It's also good to take a moment and step back from the crash a bit and pay attention that some people, who have been reading and commenting on my work, didn’t know anything about the presence of the failing system before the crash. That doesn’t set them apart from others because nobody really knew about the system. What sets some of them apart though is that by some miracle in less than a month some people have apparently developed themselves from not knowing a single thing about aircraft design or aircraft stability to fully developed air crash investigators who understand system complexity and interactions between the system including augmentation systems and the crew. To burst that bubble, it's highly unlikely that in a matter of weeks any investor with little to no prior knowledge on aircraft engineering can say a lot of useful conclusive things about the crash and certainly not without input from others. Even if you do get the input it certainly does not mean you will be able to connect accurate conclusions and "solve" the case.
Preventing the crash
We are not going to present a full rehash of the systems (if you are interested in that click here), but we do want to present two and possibly three ways in which the MCAS (Maneuvering Characteristics Augmentation System) could have been turned off and in which case the crew should have had significantly higher chances of returning to Jakarta successfully.
As is known now, the angle of attack sensor sent faulty data which caused the aircraft to trim nose down to avoid a stall. That nose down trim basically is the MCAS functioning although based on faulty AoA data. There would have been a few ways for the flight crew to turn off the MCAS.
The first way would be to use the CUTOUT switches on throttle quadrant. This would require the crew to trim the aircraft manually via trim wheels, but would no longer result in any automatic nose down trimming. This should have significantly reduced the flight crew’s battle with the aircraft and reduce the forces on the control column making the aircraft less fatiguing to control.
The MCAS is active in manual flight mode when the flaps of the aircraft are retracted. That means that if the flight crew would have selected a minimum flap setting, the MCAS system would have been inactive. Also if the flight crew would have been able engage the autopilot, the MCAS system would have been deactivated. It is, however, less likely that given the reading disagreements that the crew was dealing with that it would be possible to engage the autopilot. Either way, there would have been at least two other possibilities the crew could have used to diffuse the situation if trained well or more emphasis would have existed on the use of the runaway stabilizer non-normal checklist.
Focusing on the report, we found some interesting things. Important to understand is that this is a preliminary report by the Indonesian safety board and as such contains no analysis or conclusion.
The report also states the following:
Readers should note that the information in KNKT reports and recommendations is provided to promote aviation safety. In no case is it intended to imply blame or liability.
That’s an important note, since the report is by read by many to get a clue on who's to blame and admittedly from reading the report I also am trying to deduce whether there are any findings or observations that fault one party or the other.
As expected the flight before the fatal one also had been observed and described in detail. Before commencing the flight, the pilot discussed the maintenance actions including the AoA sensor replacement with an engineer. The pilot in command [PIC] of JT043 experienced a stick shaker (stall warning) during rotation and that stick shaker remained active for the entire flight. At 400ft, he noticed an IAS disagree (speed disagree) and handed over the controls to the first officer and verified the primary flight display [PFD] with a standby system confirming that the data presented on the left PFD was inaccurate. So the decision to hand over controls to the first officer seems to have been a good one. The flight crew also noticed that the aircraft was trimming nose down, which was the MCAS system functioning based on incorrect input from the angle of attack sensor, and decided to use the CUTOUT switches to turn off the automatic stabilizing. The crew continued the flight with manual trim and without autopilot. After landing in Jakarta the pilot documented problems with incorrect readings of speed and altitude and FEEL DIFF PRESS light illuminating which normally signals that there is a differential between the pressure from the A and B units.
The engineer flushed the left Pitot Air Data Module and static ADM to rectify the IAS and ALT disagree followed by operation test on ground and found satisfied. The Feel Differential Pressure was rectified by cleaning the electrical connector plug of elevator feel computer. The test on ground found the problem had been solved.
The pilot-in-command of JT043 noted the following in the airline’s electronic system:
Airspeed unreliable and ALT disagree shown after take off, STS also running to the wrong direction, suspected because of speed difference, identified that CAPT instrument was unreliable and handover control to FO. Continue NNC of Airspeed Unreliable and ALT disagree. Decide to continue flying to CGK at FL280, landed safely Runway 25L.
The preliminary report states that on the fatal flight there was a 20 degrees disagree between the measured angles on left and right side. This information came from readings of the digital flight recorder. THe crew asked the controller to inform them of their speed and altitude. During the rotation the crew once again experienced the stick shaker and nose down trim.
The following graph with relevant parameters was compiled by the safety board:
Figure 1: Relevant graphs (Source: OTCPK:KNKT)
On the graph you can see that during the flight the autopilot was never engaged (A/P lines). Shortly after departure it's common to "hand fly" the aircraft, so that's nothing special except for the fact that if the pilot would have engaged the autopilot successfully, the MCAS would not have pushed the nose down since the autopilot trim should fully trim the aircraft. It is, however, not known whether the crew could actually have activated the autopilot with altitude, speed and angle of attack disagrees. We also see that the stick shaker was only active on the side of the captain.
What we found particularly noteworthy is the flap handle position. What's visible is that when the flaps were fully retracted the MCAS commanded of nose down trim as expected when the AoA is signaling high angles of attack. We then see that the pilot flying gave manual trim up command (via the switches on his yoke) and the crew extended the flaps again as they continued their climb. With the flaps extended there were only five automatic trim inputs and soon as the flaps settings was reduced to zero, the MCAS started functioning again and we see nose up manual trim by the pilot flying while the automatic trim commands a nose down. Toward the end we see that the pilot flying commanded shorter and less nose up commands while the nose down commands continue. It's highly likely that with the increased forces required to pull the nose up, the pilot suffered severe fatigue.
Currently our main questions remain whether the crew noticed that setting the flaps made the aircraft more stable to control. Additionally, we do not know whether a check was performed with the standby system. We do expect that either way the person sitting in the right seat was in control of the aircraft. What's somewhat surprising is that while the second in command (SIC) should have been flying which likely was quite fatiguing with the required elevator and trim input he also was communicating with air traffic controllers. Possibly the captain who was not in control was going through checklists to see how to solve the issue or verifying the flight display info with the standby system. We think that if the pilot flying did communications with the ATC while the captain was going through the manual and checklists, they both might have been too busy to note the trim wheel activity and stability improvements with non-zero flap setting partially distracted by the stick shaker.
What's also rather interesting is that even on the ground the alpha vanes which measures the angle of attack showed big differences between the left and right side. Previously, we discussed the angle of attack indicators but failed to recognize that the angle of attack indicator (the visualizer on the primary flight display) is optional. When looking for more information on the Angle of Attack indicators and lights because the preliminary report was light on substance about angle of attack indicators or warnings lights we found a report by The Air Current outlining that both these items are optional. Lion Air did not have the angle of attack indicators nor did it have the AoA disagree light that could make crew aware of a difference in observed angles of attack between the left and right side. Having the indicator and/or the disagree light would have raised awareness among flight and maintenance crew, we think it could have resulted in the flight crew deciding against taking off from Jakarta.
What remains unknown is whether pilots have insight in the electronic reporting system. If they do have access and it is mandatory for pilots to consult the system before flight, the crew of JT610 would have seen that the PIC of JT043 reported a runaway stabilizer, which should have triggered the crew of the fatal flight to connect the flight control problems experienced during their own flight to a runaway stabilizer and execute the associated non-normal checklist requiring a switch to manual trim and cutting out the automatic trim if the stabilizer continued to "runaway."
After the crash, Lion Air has sent out instructions to maintenance, flight crews and instructors requesting extra attention for runaway stabilizer procedures, decision making regarding abnormal handling and declaring emergency, to fill in as much information as possible in the maintenance log and maintenance to report repetitive problems even when coded differently.
What we found highly interesting are the safety recommendations and in particular the safety board’s view on the previous flight. While the board did say the report would not include any analysis or conclusions, they did note that because the stick shaker activity throughout the entire flight the crew should not have continued to their destination and return to the airport. For Lion Air that's a painful observation in a report that was said not to include any conclusions. We observed that a lot of blame has been put on the crew of the fatal flight, but we also now see that the previous crew didn’t do things perfectly either despite landing the aircraft safely. Though I strongly agree with remarks that you’d rather land safely and bypassing some rules than following the rules and crash, but this is more about the pilot deciding to continue flying the aircraft to the destination with a stick shaker, which signals that there might not have been enough emphasis in documentation and training at Lion Air to recognize when the aircraft should make an emergency landing.
An unsatisfying thought here is that if the pilot in command of JT043 would have returned to its airport of departure, the aircraft would likely not have operated as JT610 and the accident would not have occurred.
Boeing and stock performance
For Boeing there were no recommendations. There only was mention of none of the executed non-normal checklists requiring the aircraft to divert to the nearest airport during the previous flight. This is not surprising, since Boeing can’t outline procedures for every scenario. We think a flight crew suffering several disagreements in readings and a stick shaker on one side should use common sense and land the aircraft at the nearest airport.
However, we do reiterate our view from our previous report: Boeing introduced the MCAS system and the rationale behind it is very simple: The Boeing 737 MAX is slightly more unstable so they introduced a system that should help crews avoid stalling. We think they should have informed crews about it. Boeing said it did not disclose the system to pilots in order to avoid flooding them with information. We think it's only partly justified to follow that rationale. The MCAS is a stall prevention system and it could indeed partially defeat the purpose of the system if you tell pilots about it and tell them how to circumvent that.
Boeing CEO Dennis Muilenburg said that they disclosed everything to fly the aircraft safely and in some way he is right, since the airworthiness directive and Boeing’s Operations Manual Bulletin points at an existing non-normal checklist for deactivating the stabilizer trim and continue flying the aircraft with manual trim. This would, without the flight deck crew knowing it, result in the deactivation of the MCAS which is the system that pilots didn’t know about. What, however, was missing was the link to the angle of attack sensor in the checklist.
You could say that was critical, were it not that the Lion Air aircraft had no AoA indicators or disagree lights. That likely has resulted in decreased awareness. You could ask why the indicators and disagree lights are not mandatory, since the angle of attack has a fundamental importance in the basics of flight but we don’t view that as something Boeing should ultimately be blamed for. Instead, if the angle of attack indicators are deemed critical for flight and awareness then the certifying agencies should require them to be installed. You can hardly blame Boeing for aviation administration not implementing requirements.
According to reports Boeing is currently working on a software update that will deactivate MCAS when the crew opposes its nose down command. The jet maker also is said to be looking whether it needs to pull data from both alpha vanes rather than only using data input from the left vane.
Lion Air is currently threatening to walk away from its Boeing orders as it feels that Boeing attempts to deflect attention from recent design changes and blame the Indonesian firm for the crash. according to the report. Per Boeing’s backlog overview, Lion Air currently has 190 aircraft on order with Boeing. 190 aircraft is a lot, since the market value would be almost $10B and it would represent 4% of the order book for the Boeing 737 MAX. Lion Air also is one of the biggest customers of the Boeing 737 MAX 10 which intends to compete with the Airbus A321neo.
In the previous report and in the comment section of that report, AeroAnalysis International noted that share prices do decline after crashes, and while some people feel that a crash might become a drag on the jet maker’s share price appreciation, its effect is more temporary than many think.
That view has been confirmed if we look at Boeing’s share price since the day prior to the crash:
We see that shares are flat and from the moment the preliminary report has been released until the 3th of December share prices gained 7%, partially helped by easing trade tensions with China. We think the overwhelmingly negative coverage in mainstream media led to Boeing share prices tanking quite a bit in the days prior to the release of the preliminary report. Interesting to note is that share prices went up 13.6% from the moment Seeking Alpha released the “Premature Judgement: Public Opinion Condemns Boeing, Even Before Crash Report Is Out“ report.
Although the issued report by the Indonesian safety board is a preliminary one with no conclusions, we’re sharing our view in this report via some bullet points:
- We think that if Boeing would have described MCAS to pilots, the pilot of JT043 could potentially have accurately described the problem in the maintenance log and maintenance personal could have checked the angle of attack sensor once again and see that the sensors were still not in agreement.
- If JT043 would have declared emergency because of the continued stick shaker activity, flight JT610 would unlikely be operated with the aircraft that ended up crashing in the Java Sea.
- If the crew of JT610 would have been in the position to consult the electronic reporting system, they could have seen the solution to their flight control problem presented to them but it is not known whether crews do have access to reports others pilots file and whether checking is mandatory.
- If aviation administrations would have required indications of angle of attack to be displayed and a disagreement light to be installed it would have positively contributed to situational awareness for crews but also for maintenance personnel. The PIC of JT043 noted down the illuminance of the FEEL DIFF PRESS light and if the AoA disagree light would have been installed, it is likely that the PIC would have written that down in the maintenance log as well somebody from maintenance would have checked the AoA sensor again. The debate about mandatory installation of these indicators and amber lights have been going on for years and are nothing new.
- Feeding the MCAS with angle of attack data from two sensors would likely have prevented a crash.
- The installation test for the AoA sensor either was not performed correctly by maintenance or the test is not sufficient to rule out disagreements between the left and right sensor.
- The crew could have noticed the trim wheels turning and that (even with not knowledge of the previous flight or MCAS) should have resulted in the crew deactivating the autotrim.
- The crew, if aware of the function underlying the stick shaker, even without AoA indicators or disagree lights could have known AoA readings from one sensor were not correct since the shaker was only active on one side. In combination with more information from Boeing on MCAS, it might have triggered the crew to executive the non-normal checklist for a runaway stabilizer for the correct reason.
In our view the repetitive problems with the aircraft should at least have triggered maintenance to contact Boeing about the issue. For the last four flights there were problems with incorrect readings on the left side, each time a maintenance task was performed and the aircraft was returned to service. With serious issues such as repetitive inaccurate readings, the maintenance department should have taken the aircraft out of service. With no trigger from maintenance, the aircraft was cleared for flight each and every time. Even then, we think that with incomplete checklists the crew of JT610 should have linked heavy controls and a spinning wheel to counteracting stabilizer motion as they would with note with the speed trim system. This also is how the previous crew handled the situation.
What we still don’t know and probably won’t know until the cockpit voice recorder is recovered is the cockpit work flow, exact problem identification, proposed solution and communication. We think that Boeing could have informed the crew more about MCAS, but even in absence of that added documentation and training the crew could and should have thought about the runaway stabilizer procedures. The lack of situational awareness likely is the cause of this crash where the crew didn’t even think about things that actually are in the manual. Lion Air currently is threatening to cancel the order with Boeing as they are trying to "distract" from the incomplete documentation and information. We think that's short-sighted from Lion Air and in fact they might even be ones distracting from a maintenance unit that continued clearing the aircraft for flight each and every time, which might be an indication of a lack of awareness, not on the flight deck but in the maintenance hangars.
Currently I'm inclined to say that the while Boeing’s documentation could have been better and its stall identification, warning and prevention could have been more robust and aviation administration requirements are lacking on AoA indicator and disagree requirements, the maintenance facility of Lion Air, the previous crew and the crew handling the last flight also made mistakes that could have been prevented even without knowledge about the MCAS.
We also currently are seeing that fear is flowing out of the market and also out of Boeing with the preliminary report released. We think that the pressure on Boeing’s share prices and the dense coverage of the Boeing 737 MAX crash by mainstream media are no coincidence.
List of abbreviation used in this report:
AoA – Angle of Attack
MCAS – Maneuvering Characteristics Augmentation System
PFD – Primary Flight Display
PIC – Pilot In Command
STS – Speed Trim System
If you enjoyed reading this article, don't forget to hit the "Follow" button at the top of this page (below the article title) to receive updates for my upcoming articles. Join the AeroAnalysis Marketplace service for early looks and detailed reports and discussions.
Disclosure: I am/we are long BA, EADSF. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.