First of all, the g limits for a passenger jet with the flaps up are -1 to +2.5 g. As soon as you roll inverted, just flying level, you are right on the negative g limit. Any push to climb will exceed it. Will the wings fall off? No, there’s a buffer, but you’re operating on tiny margins and a ham fisted push might be problematic.
Second, I think you’d struggle to physically move the column forward as much as needed. Once inverted, you’re hanging off your straps and given you never expected to be inverted you would not have reefed your seatbelt super tight so your body is hanging much further from the column than usual. That said, maybe the already nose down stab trim would be all that’s needed. I don’t really know. I know from experience flying a Pitts Special that being trimmed for upright level flight does not equal being trimmed for inverted level flight, you needed a healthy dose of forward trim to just hold it level inverted, and that’s in a plane designed for it with close to a symmetrical airfoil.
In upright level flight the stabiliser on a normal plane generates downforce. In inverted flight I’m not sure if it would have enough stabiliser trim movement to be able to generate enough or any downforce in the opposite direction. The wings themselves are far less efficient and will stall at a much lower angle of attack. In short, an aircraft that is trimmed significantly nose down while upright, is not necessarily capable of climbing while inverted.
Third, the fuel supply has some capacity to cope with negative g briefly but I don’t think it would work with sustained inverted flight. Unfortunately the more complicated a plane becomes, the less detail they give us pilots. I used to know exactly how the Dash 8 fuel system worked but there’s nothing like the same detail for an Airbus or Boeing.
My suspicion is, and I’m happy to be corrected, that once rolled inverted there would be insufficient elevator and stab trim authority to do anything other than maintain level flight and that the engines would flame out. You’d be left with a very heavy glider flying upside down without enough height to recover.
At least with the roller coaster method described in the PPRuNe thread you are working with the aerodynamic design of the aircraft instead of against it.
The discussion about flight control laws was a tangent. Non-FBW jets like the B737 don’t have control laws. The cockpit controls operate hydraulic jacks via a mechanical system of wires, pulleys, springs etc, and the hydraulic jacks move the control surfaces. If you turn the control wheel fully in one direction then the ailerons will go to full deflection. There is no computer in the loop to moderate the aileron response to the control wheel movement. As such, there was no “direct law” or equivalent that would’ve been of any use to the B737 crews.
On a FBW aircraft you can have computers helping the pilot fly the aircraft. In an Airbus there are three main control laws and most airbus pilots would only ever see one outside the simulator.
Direct Law is the most basic. Control surface deflection is proportional to side stick demand. Move the stick hard left and you will get full aileron deflection in the appropriate direction. It is essentially the same as flying a fully hydro-mechanical plane. In this sense a B737 is always in direct law.
Normal Law is the full help mode, and as it’s name suggest is what is normally active. In Normal Law the side stick demands a roll rate and G load rather than a surface deflection. Pull the stick full back as hard as you like and the computers will ensure that you only get 2 Gs. You can set an attitude up to 33° angle of bank and the computers will hold it there (in a conventional aircraft you’d have to hold it there yourself against the aircraft’s natural tendency to return to level flight). If you set more than 33° and let go of the stick, the bank will come back to 33°. If you try to set more than 67° it just won’t do it. It has high speed and low speed protections that will prevent the pilot from over speeding or stalling. It’s like a computer game with aim assist, automatic targeting, etc.
Alternate Law is similar to Normal Law but some protections are not available or simplified.
I think it’s just not physically possible to manually move a huge stabiliser against large air loads. It’s not designed to be that way, it just IS.
That’s the big question isn’t it? Why didn’t they use both vanes? As you say, it should be a simple task to use both inputs rather than just one.
Thanks for correcting Magiver on the stick force issue. I think it can take 50-100 lbs force on the yoke to try and counteract full nose down stab trim.
Apparently there is a known condition in the 737 (not just the MAX) where if the stabilizer is trimmed severely nose down and the yoke is pulled back (elevators causing nose up) this can bind the stabilizer jack screw so it can’t be moved with the manual trim wheel: Satcom Guru: Trim Cutout with Severe Out-of-Trim Stabilizer can be difficult to recover
Supposedly Boeing engineers designing the 737 MAX were told they can’t use any design features which would require additional pilot training. Using two AOA sensors for MCAS would have required an “AOA disagree” scenario, which in turn would have required pilot training when transitioning to the 737 MAX. Yet a foundational concept of the 737 MAX was it flies exactly like the 737 NG and does not require additional pilot training.
Allegedly Boeing’s design solution was use a single sensor and treat a sensor-induced MCAS failure as a simple trim runaway from a procedures standpoint. I don’t have a reference for that but it’s been widely reported. I’m sure the FBI investigation will scrutinize this.
Apparently Boeing is still strongly resisting additional training, since it will cost them big time. According to this article, they’ve agreed to pay a per-plane cost if pilots have to get into the simulator.
From the article: [Former Boeing engineer Mr. [Rick] Ludtke [who worked on 737 MAX cockpit features] recalled midlevel managers telling subordinates that Boeing had committed to pay the airline $1 million per plane if its design ended up requiring pilots to spend additional simulator time. “We had never, ever seen commitments like that before,” he said.
Nobody is suggesting artificial feel would mimic a joystick. It’s simply feedback. It’s pointless to make it too diffficult to handle. If input would exceed structural integrity the computer would simply ignore it. The only example I can think of where it was mismatched was the A-300-6 rudder pedal and the real flaw there was the computer allowed deflection rates that were detrimental at excessive speeds. Stiffer pedal feel may have avoided the accident but but ultimately the computer was programmed to allow the input.
If this is true that is an astronomically stupid arrangement. If the MCAS system has created a nose down scenario then it places the plane in a situation that cannot be resolved quickly without electric trim. An overspeed situation quickly develops making manual trim adjustment impossible. The proper solution is to counter computer initiated down trim is with pilot input for up-trim. The only time you would ever disable trim is in a runaway situation and you’re pretty much screwed at that point anyway. The aircraft would accelerate too quickly.
Again, this would be an incredibly stupid solution to the wrong problem. Unless it’s a runaway trim condition it’s crazy to turn off the trim. What has been described is a situation where the computer repeatedly continues to trim down AFTER the PIC trims up.
Yes, Centaurus is describing what I’ve been talking about exactly except there is a cutoff point where stabilizer trim creates more force than elevator trim can compensate for. Then you’re screwed. It becomes the classic fight leading up to mach 1 where pressure on the control surfaces exceed pilot input. Electric trim is needed because the clock is ticking at low altitudes. You don’t necessarily have the luxury of what Centaurus described if they can’t operate manual trim in time to avoid breaking the low-altitude record. Rolling the plane is the last desperate act of a doomed aircraft. It’s going down. What Centaurus said about training is probably true. I suspect there are quite a few pilots who don’t understand the concept of mechanical load on control systems. It’s something that should be taught in detail. It’s pretty simple to comprehend. It’s like changing gears on a 10 speed. You take the mechanical load off the pedals first so the derailer can move smoothly.
AoA vanes are cheap and easy to damage. Their reliability becomes part of the risk assessment. 99% reliability is great until you hit the 1%. If the 1% happens at low altitudes then what is the probability of a serious accident? 346 people know the answer to this particular issue. You’ve already pointed out that this can be reduced with programming changes and that may be all it takes. I personally would expect a switch that deactivates the MCAS and would be surprised if it doesn’t already exist.
It sounds to me like the crew tried to correct a runway trim situation instead of an MCAS error input situation. As long as electric trim responds to pilot command the solution is to continue to use the trim switch until they can stabilize the plane and THEN turn off the electric trim if the MCAS makes final approach too erratic. They could trim the plane (at a higher altitude) as if they were on final and then proceed land it normally.
I think this likely pilot error in response to a serious MCAS system flaw made worse by a lack of training.
Is it fair to call it “pilot error” if the pilots have not been trained in the issue, or, in the case of the initial report of the Ethiopian crash, they did what the training manuals said they were supposed to do, and it didn’t work?
When I watched videos of 737 pilots the trim wheel motion was almost impossible to miss visually. However, the camera view was from behind the pilots. Maybe it’s harder to see but IMO it’s high enough up that it should be visible. The stripes on the wheel are added to make it more noticeable.
If this is the case, I think the crew should have seen this movement as the cause of the nose-down pitch. Their first reaction should have been to re-trim using the trim switch on the yoke. Since they deactivated the electric trim then it’s likely they thought it was a runaway trim issue. I would think it obvious to hit the trim switch to counter it. I expect some confusion ensued while they tried to diagnose the problem.
Yes, I’m calling it pilot error. And yes I’m blaming Boeing for a design flaw. However, pilots are not getting a healthy paycheck to fly the plane. They’re getting paid to fly the plane when crap happens.
If it was a runaway trim situation it’s still up to the pilots to arrest the situation by disabling the electric trim before the horizontal stabilizer angle exceeds the mechanical input of the elevator. However, that’s probably measured in seconds at low altitude.
Most airline fatalities are a series of events that when combined lead up to a crash. A design defect, poor training, misdiagnosis of the problem, and low altitude could all be factors. So in that respect blaming the pilots may seem unfair. At some point the number of events will exceed human ability to deal with it.
You keep going on about computers. The 737 is not FBW. It doesn’t have computers deciding whether or not a control input is allowed. If you can physically move the controls and the hydraulic actuators can move the surfaces, then the surfaces will move. It is supposed to be very difficult to move the controls at high speeds or high AoA to prevent the pilots from breaking the wings off or inadvertently stalling. This is how artificial feedback on mechanical aircraft work.
BTW the A300 was also a non-FBW aircraft. There were no computers limiting or not limiting rudder input. Even the A320, a genuinely FBW aircraft does not have a FBW rudder, it is mechanical with hydraulic actuators.
The Boeing approved method for countering MCAS activation is to follow the trim runaway procedure, ie turn off the electric trim. They should have countered the trim first before turning it off, but that is not strictly part of the trim runaway procedure they’d been told to follow.
And yet, Boeing advice for countering MCAS is to follow the runaway trim procedure.
Reliability is probably more like .0001 %. And if one fails, nothing happens because there are no critical systems that rely on a single AoA vane. That is until some genius designed MCAS. It’s MCAS that is the problem, not AoA vanes that have been serving a useful purpose on airliners for many decades.
There is no switch to disable MCAS other than the switches that disable the entire electric trim system.
They followed the Boeing procedure for a rogue MCAS. There are lots of things they could have done better if they’d had the brain space to think of it, but they followed the procedure given to them and it didn’t work.
The expectation of airline pilots is that they are normal pilots, not Chuck Yeager. They are not expected, and in fact not wanted, to come up with their own solutions to a problem if an official solution already exists. You’re only supposed to get creative when the official solution hasn’t worked, and that’s exactly what they started doing when the FO apparently turned the electric trim back on to try to get the trim moving near the end of the flight.
Don’t forget also that while this is all happening they had constant stick shaker plus airspeeds that didn’t agree and then later the overspeed warning.
Doesn’t the fact that, within months, two different pilot crews crashed in a manifestly similar way pretty much disprove the possibility of “pilot error”?
What is the likelihood that two pilots, from two airlines, would make a series of irrecoverable, fatal errors on the same type of plane in so short a time? A new and idiosyncratic plane at that.
Wouldn’t using two AoA sensors in a voting system be very dangerous? That would mean that one malfunctioning sensor would disable MCAS and subtly change the handling characteristics of the aircraft. Wouldn’t you want at least 3 so you can use the inputs from 2 working ones if the third one is wrong?
ETA Obviously it would be better than the current arrangement but would it really be acceptable?
I haven’t said anything about FBW and I don’t know why you’re going on about it. I merely pointed out that stick pressure is arbitrary feedback for the crew and as such would not create pressures they can’t handle.
does the manual say to ignore a working trim button and turn off electric trim in a nose-down attitude? I’m going with NO. The report shows the crew used the trim to bring the nose back up.
They weren’t in a runaway trim situation as long as the system responded to their trim inputs. Which it was according to the report. This is the entire problem. They responded to the wrong issue and that’s going to fall on Boeing AND the airlines for their training procedures. But it doesn’t absolve the crew of common sense. They didn’t have the altitude/time to manually crank up the trim. They should have balanced the tail plane first and then continued to climb to give themselves time to work it out.
No, the expectations of aiirline pilots is that they train for emergency procedures and can use the equipment that is functional to maintain control.
I can believe they were confused by the MCAS inputs but I expect them to also realize they have equal control with their pilot controlled trim. The correct procedure would have been to trim it neutral or for approach speed trim and THEN shut it off. Thus locking the horizontal stabilizer into a trim setting that made the plane fliable.
It depends on how critical the system is. If MCAS stopped working mid flight it would have no effect on anything and would be fixed on the next turn-around. Remember that MCAS was only ever expected to activate in extremely rare situations. Most pilots would never be in a situation that resulted in MCAS operating as designed.
I liken it to a stick pusher in terms of its criticality. On aircraft I’ve flown fitted with a stick pusher, the system relied on two AoA vanes to activate. If the pusher became unserviceable for some reason, eg a damaged AoA vane, then you continued the flight knowing that the system was broken. The checklist had something like “don’t stall” as a note, well you’d never do that intentionally anyway.
It’s different to a FBW aircraft that has computers with full authority over the flight controls that rely on AoA data. In that case you need a more robust voting system because the system is more critical.
You talk about “computers” and what the computers will and won’t let the pilots do. My point is that the “computers” you talk about don’t exist, at least not in the way you imagine. I am saying that the feel system can in fact create pressures the pilots can’t handle.*
Are you guessing “no” or do you know what the manual says?
Yep they should’ve. We know that now.
You are mostly correct that if they had trimmed to a reasonable setting before disconnecting the electric trim things would have gone a lot better for them. Now consider the time, relaxed environment, and knowledge you have, post accident, that has enabled you to come to that conclusion. Compare that with the time, environment, and knowledge of MCAS crashes this crew had.
We are talking about the failure of a system that up until a few months ago no pilot knew existed. There is no emergency procedure for it, just one for a similar but critically different failure that Boeing decided was good enough in the mean time. These pilots never had a chance to train for this because no one knew about it. They will have trained for trim runaway and had been told to use the trim runaway procedure in lieu of a dedicated MCAS procedure but what they presumably hadn’t been told is that the trim runaway procedure alone is not enough.
Whether or not the pressure is more than the pilots can handle depends on their strength. It can certainly take both pilots pulling to overcome the forces though.
I have re-read the changes to the manual put out by Boeing after the Lion Air crash and they do essentially say in long form to do as Magiver has suggested.
But that just raises the question, if they used the electric trim a few times prior to disconnecting, why didn’t they trim all of the forces out? Is it difficult to use the trim while the stick shaker is going? What human factors came into play that prevented the pilot from continuing to trim up?
This is my concern about the “pilot error” argument.
If the cockpit warning systems are operating correctly, and there are a wide variety of warnings bring given simultaneously, is it fair to the pilots to say they didn’t understand all the multiple warnings?
I say this as someone with absolutely no pilot experience or knowledge, trying to understand the various technical arguments being made here (thanks, all!)
But at some point, if all the warnings and inputs are overwhelming, is it really realistic to expect the pilots to figure out it all out in a few seconds or minutes? And to blame them for the crash, if the manufacturer has created an “overload” situation that no-one can reasonably be expected to handle?
Indeed. How else to explain two crashes by two crews only months apart with highly similar, if not identical, terminal flight control problems.
Unless pilot error at this level is a common occurrence, the very fact that it happened twice would seem to be evidence that it can’t just be pilot error. It’s far more likely to be something independent of the pilots.
I think it’s the case that pilot error is not all or nothing. It is inevitable that some errors are made. The question to ask is “are the design and procedures sufficiently tolerant of errors?”