Ethiopian Airlines Crash; Another 737 Max

If the counterweight detached internally, for whatever reason, and the vane itself wasn't damaged, the AOA sensor input would be reversed?
Yes. It would have failed to full low-AoA then shown movement under higher positive-g loads.

Somewhere earlier on this thread I asked about how often an AOA sensor fails, I think I got a "they don't" answer.
Not often. In 29 years of airline flying I've had one failed AoA.

From what we know at this point, the Lion Air and Ethiopian AoA failures were from different causes.
 
I’d fly on one today.
 
Considering the impact of this grounding and considering that it's a software fix what on earth is taking so long? You'd think they'd have sorted it all out by now, they've had time to do hundreds of hours of testing. I mean, I know there is bureaucracy involved but for cripes sake is this not just a wee bit ridiculous?

Let's take, for instance, today. What are they doing right now, today, to move this forward? Are there people just sitting idle and doing nothing? Is there a form on someone's desk that's just sitting there and not getting processed?
 
Are there people just sitting idle and doing nothing?
FYI: back in May it was determined that one part of the MAX equation (manual stab trim at speed) might also be applicable to the NG models. With over 20 different Civil Aviation authorities "involved" in the process now, everyone is checking to ensure their personal/political issues are covered. It was stated at the end of April the MAX fix was going through flight tests with a possible release in June.
 
I believe that the fix is done; has been done. It's in the approval process.
Wonder what happened to the original FAA requirement MCAS was meant to satisfy? That's always been the other shoe to drop.
 
Wonder what happened to the original FAA requirement MCAS was meant to satisfy?
In general, the MCAS remains intact except it will receive data from two AoAs (instead of one) and the rate of correction has been reduced. Plus there are updated checklists and additional training requirements.
 
In general, the MCAS remains intact except it will receive data from two AoAs (instead of one) and the rate of correction has been reduced. Plus there are updated checklists and additional training requirements.
The only change is that explanatory text has been added to the end of the checklist to give additional context on an MCAS-driven runaway situation. The steps for countering a runaway stabilizer, regardless of cause, remain unchanged.
 
Pilots, including sully say more training was and is needed. Simulator training.
https://www.rawstory.com/2019/06/pi...-us-congress-more-training-needed-on-737-max/

Proper training is always valuable.

I’d suggest that, in the first instance, a better culture of accountability would have prevented the accident ab initio and, in the process, liberated sufficient information on the root cause in time to head off any other incidents (although, with regard to Ethiopian, the jury is still out and will, for political reasons, stay out).
 
https://www-bloomberg-com.cdn.amppr...-on-doomed-jet-came-down-to-two-cockpit-wheel

Seems like a fairly accurate description of what happened based on known information. The over aggressive MCAS system with a AoA failure leads to the manual stabilizer becoming very difficult if not impossible to operate under specific flight conditions.

I know somewhere much earlier in this thread somebody mentioned that the manual trim wheel in 737's (and in most large jets), can become very difficult for a pilot to turn due to aerodynamic forces on the elevator. The article mentions this is a know problem in 737's since basically the start of the model, and the manual mentions aerodynamic alleviating procedures may be needed to relieve the pressure on manual trim so it can be turned in certain situations.

I have never flown anything but single engine bug smashers, but it seems somewhat clear what has happened here. A faulty AoA lead to the MCAS system being over aggressive. The pilots were possibly slow to reacting to a runaway trim situation. By the time they reacted (and it may have happened so quick it was not their fault) the trim reached the point where it became very difficult (possibly impossible) to turn manually. The pilots paniced (understandably), and failed to enact a second emergency procedure, that after the trim was disconnected, they may have failed in doing the steps needed to move an essentially stuck manual trim wheel, by adjusting power and adjusting pitch so they could turn the wheel.

No delusions that I could have reacted better, and it seems clear the initial failure was poorly understood, and therefore slowed reactions leading to a cascade of other failures.

Sent from my SM-G950U using Tapatalk
 
Seems like a fairly accurate description of what happened based on known information. The over aggressive MCAS system with a AoA failure leads to the manual stabilizer becoming very difficult if not impossible to operate under specific flight conditions.
The manual trim wheels each have a fold-out handle. The handles are mounted 90 degrees of rotation apart to ensure that one of them is always in a position which gives the pilot good leverage to turn the wheel. This system is not unique to the MAX or even the larger 737 fleet. This was the backup trim system for all of the B707, 727, and 737 aircraft which all have long records of operating safely.

When the airplane is close to in-trim, as you would be throughout a normal flight, the wheel is easily moved by the pilot-flying (PF) using the fold-out handle. As you get farther out of trim, the forces required to turn the wheel increase so it might be necessary for the PF to call for trim and have the pilot-monitoring (PM) turn the wheel. i.e. "Trim down" ... "Stop trim"

When an out-of-trim airplane's airspeed increases the aerodynamic force on the stabilizer increase which increase the pressure on the stabilizer jackscrew which increase the force necessary to turn the trim wheel. If the airplane is significantly out of trim and at high airspeeds, it may be necessary for both pilots to work together to turn the trim wheels. The Captain using his right hand, and the F/O using his left hand, on the extended handles.

If you allow the situation to progress to the point that the trim is at the nose-down limit, and your airspeed is excessive, you may have to unload the stabilizer to be able to move the trim wheels. The stabilizer is introducing a strong nose-down moment so you unload by relaxing back-pressure and allowing the nose to drop. While the back pressure is relaxed, the forces on the stabilizer are reduced and you both trim together to progressively move the trim away from the full nose-down position. You alternate between unloaded to trim and raising the nose to regain the altitude that you're giving up in each cycle. With each cycle, the stab trim is moved closer to neutral and the forces required for the next cycle are reduced.

The key here is to avoid situations where you have full nose-down trim and excessive airspeeds. If you do, you'll never be in a situation where the more extreme measures are necessary to get or keep the airplane in trim.

The DFDR data from all three flights (Lion Air incident flight and the Lion Air and Ethiopian accident flights) show that the primary electric trim overrides MCAS (as it is designed to do) and can be used to keep the airplane in-trim through multiple unschedule MCAS activations. The Captain of the Lion Air accident flight kept the airplane in-trim through 21 unscheduled MCAS activations. Each time MCAS started trimming down, he applied nose-up trim with the thumb switches on his yoke which stopped the MCAS activation and allowed him to return the airplane to an in-trim state. There was nothing that would have prevented him from continuing to do this indefinitely. Unfortunately, he didn't continue flying the airplane. The F/O couldn't find the applicable checklist so the Captain transferred the airplane to him and began looking for the checklist in the quick reference handbook (QRH) himself.

While the F/O was flying, he had 5 additional unscheduled MCAS activation. He stopped the first 4 with a quick bump of the trim thumb switches on his yoke but he did not continue to hold the switches long enough to remove the nose-down trim that MCAS has applied as had the Captain. With each MCAS activation the stab trim moved progressively closer to the nose-down stop. After the F/O's fifth MCAS activation the stab trim was at the full nose-down stop and they were unable to maintain control. The stabilizer runaway procedure was never accomplished and the use of manual trim was never attempted.

On the Ethiopian flight the Captain was again flying the airplane. This time there wasn't a fault in the left AoA sensor, the sensor was damaged by an apparent bird strike almost immediately after takeoff. The effect was similar. The left-side stick-shaker activated and "AOA DISAGREE" messages appeared on both primary airspeed indicators. The Captain engaged the A autopilot (A/P) while still below the minimum A/P altitude engagement altitude of 800'. The A A/P draws its flight data from the left sources which were the ones which were compromised by the loss of the AoA vane. (If he had selected the B A/P it would have flown from the right data sources and MCAS would not have activated as that would have switch from the compromised left FCC and ADIRU to the right FCC and ADIRU which were unaffected) The A A/P attempted to follow the invalid left-side air data which proceduced the unstable initial climb which, prior to the release of the DFDR data, suggested that the problem had been something other than MCAS which is inhibited with flaps extended. After approximately 30 seconds, the A A/P disconnected due to the increasingly erroneous data it was receiving.

Once the flaps were retracted, the unscheduled MCAS activations began. The Captain re-extended the flaps and that temporarily inhibited MCAS. For some reason, the flaps were again retracted and the unscheduled MCAS activations resumed. Many of the MCAS activations were stopped with the primary electric trim but the Captain was not fully re-trimming out the MCAS trim inputs. Meanwhile, the Captain had called for the Level Change (LVL CHG) vertical mode on the Mode Control Panel (MCP). The MCP selected altitude was well above their current altitude so this put the auto-throttle system (A/T) into "N1" mode with climb thrust (CLB) commanded. The flight directors would then be commanding pitch attitudes to track the airspeed selected in the MCP window. In LVL CHG, the A/T will not be reduced to correct for overspeeds. The F/D, and A/P if engaged, will command higher pitch attitudes in an attempt to correct the overspeed. The A/T were left engaged throughout the flight and they maintained the very high CLB power setting. This high power setting caused them to accelerate quickly. Actual airspeeds (indicated correctly on both the F/Os and standby indicators) reached the 390 KIAS range. Vmo for the airplane is 340 KIAS and a more appropriate speed for their situation would have been in the 210 KIAS to 250 KIAS range.

Within a short time, the uncorrected unscheduled MCAS activations had progressively moved the trim to an excessive nose-down position. At some point the F/O recognized an MCAS/stabilizer runaway and the memory items were partially accomplished (they failed to disengage the A/T and never reduced thrust) including flipping the primary and standby trim cutout switches. This removed all electric trim from the stabilizer system but, when they did this, the trim was already close to the full nose-down stop. The Captain (PF) told the F/O to use manual trim but it is unclear if he tired to use the trim switches (now deactivated) or turn the wheel. If he tried to turn the wheel it isn't clear if he unfolded and used the handle. There was no indication on the CVR that the crew ever attempted to use both trim wheels together or to unload the stabilizer to reduce trimming forces.

At this point, they decided to try the electric trim again. They deactivated the two trim cutout switches but either failed to hold the primary trim switches in the nose-up position or the combination of excessively nose-down trim and excessive airspeed produced loads that were too high for the electric system to overcome. What did happen was that the unscheduled MCAS activations resumed and it pushed the trim to the full nose-down stop. At that trim setting, and around 390 KIAS, they were unable to maintain control.

The sim trials that have been reported have been in NG sims which don't have MCAS. You can't accurately simulate an MCAS runaway in an NG sim because as you introduce the nose-down runaway the pilot will instinctively apply back pressure which will cut out the trim with the trim brake. The trim brake is not active on MCAS activations because it would prevent MCAS from doing its intended job of introducing a nose-down bias in high AoA situations. Since they couldn't replicate the MCAS runaway, they started the trial with the trim already full nose-down and at around 250 KIAS. Those who have tried this have said that the recovery is difficult due to the factors that I've already discussed.

The key is not letting the trim get all the way to the full nose-down position before you take action to recover. Use the yoke trim switches to keep the airplane in-trim until the electric trim system is disabled.
 
Larry in TN

Thank you for the lengthy explanation. I was 80% of the way to understanding what happened and am now fairly sure I understand. Having not delt with A/T among items with the shear level of automation in modern jets, this is nowhere near something I have needed deal with, and certainly have not been taught. It is also something that will be nearly impossible for people without pilot training or engineering experience to remotely understand.

It seems the MCAS has interacted with existing systems in a way that was either unintended or poorly understood because the series of faults and cascade of events following, had not been fully tested or even imagined. In theory, a well trained and rated pilot could deal with the problem, but as usual in a stress situation, where a new (at that time undisclosed) system caused unexpected problems in an unexpected sequence, they unfortunately did not, or perhaps could not, react fast enough.

Seems to go back to an old adage I have heard about in engineering and computer programming for years. That despite of your best testing, design, and intentions, users or unexpected events will stress a system and cause failures in a system nobody even imagined when it was developed and tested.

Seems like that is what happened here.

Sent from my SM-G950U using Tapatalk
 
Having not delt with A/T among items with the shear level of automation in modern jets, this is nowhere near something I have needed deal with, and certainly have not been taught. It is also something that will be nearly impossible for people without pilot training or engineering experience to remotely understand.
The second and third items (memory items) for the runaway stabilizer checklist are A/T and A/P disengaged.

The first two items (memory items) for the airspeed unreliable checklist are A/T and A/P disengaged.
 
Was wondering though, when they can correct with the thumb trim, indefinitely as mentioned, or is it "inching closer" each time?
What I mean is, if the MCAS is bringing the nose down for X+y seconds, but they only have Y seconds to raise the nose, is it losing stability such that it evenetually will not continue to correct?
 
Was wondering though, when they can correct with the thumb trim, indefinitely as mentioned, or is it "inching closer" each time?
What I mean is, if the MCAS is bringing the nose down for X+y seconds, but they only have Y seconds to raise the nose, is it losing stability such that it evenetually will not continue to correct?

Nope.

As long as the electric stab trim switch is used, it will always offset the MCAS. That was how the pilot on the flight prior to LionAir’s loss did it - and the Cpt on the LionAir flight did, as well, but he turned it over to the FO, who apparently stopped.
 
Was wondering though, when they can correct with the thumb trim, indefinitely as mentioned, or is it "inching closer" each time?
What I mean is, if the MCAS is bringing the nose down for X+y seconds, but they only have Y seconds to raise the nose, is it losing stability such that it eventually will not continue to correct?
When either of the yoke trim switches are activated it stops the current MCAS activation and has full control of the stabilizer.

When the thumb trim switch is released, MCAS is inhibited for an additional five seconds after which MCAS can activate again if the triggering conditions are met.
 
When either of the yoke trim switches are activated it stops the current MCAS activation and has full control of the stabilizer.

When the thumb trim switch is released, MCAS is inhibited for an additional five seconds after which MCAS can activate again if the triggering conditions are met.
I know the proposed changes to the Max area reducing the rate of pitch change making it less aggressive, making the MCAS read both AOA sensors, and Boeing has made the disagreement light standard. Do you think they will add a MCAS cutoff switch individually or will the trim cutoff still be the only one. Just with relabling and manual clarification that that switch disables MCAS.

Sent from my SM-G950U using Tapatalk
 
Well, since an MCAS failure acts like a runaway trim issue, there is no need for a separate MCAS cutout switch since the trim cutout switches accomplish the same objective.
 
I know the proposed changes to the Max area reducing the rate of pitch change making it less aggressive, making the MCAS read both AOA sensors, and Boeing has made the disagreement light standard. Do you think they will add a MCAS cutoff switch individually or will the trim cutoff still be the only one. Just with relabling and manual clarification that that switch disables MCAS.
The AOA DISAGREE message wouldn't have helped. Even if they saw it, which they probably would not have, it would have been the lowest priority item. 1. Runaway trim, 2. IAS DISAGREE, 3. Stick shaker, 4. AOA DISAGREE. And, it's not a light. It is a message that appears on both primary flight displays under the bottom-right corner of the ADI.

What Greg said. Unscheduled MCAS activation presents as a runaway stabilizer. That is the appropriate procedure to use to address it.

With the software change, a single bad AoA input will inhibit MCAS. If I remember correctly, a disagreement of more than 5.5 degrees between the two AoA readings will inhibit MCAS.
 
I haven't seen the newest problem being discussed, which is apparently that the 737 Max's flight control computer is too slow to properly handle the "fixed" MCAS software.

First reported by NY Times and analyzed by an obscure but well-researched blog:
https://www.moonofalabama.org/2019/...x-problem-overwhelms-the-planes-computer.html

The blog also discussed the problem that started with the 737NG regarding impossibly large manual trim forces:
https://www.moonofalabama.org/2019/...severe-problem-with-older-boeing-737-ngs.html
 
Last edited:
I’ve always thought it an odd marriage the FAA and big aviation biz. I think it’s absurd when a Continental blows to crap we ask Continental to be in on inspecting why it went wrong...
 
Back
Top