You’re looking at this on a very simplistic level. ALL airplanes have flight characteristics unique to their design. All of them. They must be flown according to these characteristics. A 747-8 is a much different airplane than a 747-100.
I’ll give you a personal example. I changed to a more powerful engine in my plane. It was a custom design similar to another model of the same make. It added almost 40% more power. The differences between the two models is the cowling, a redistribution of parts under the cowling, and lack of flaps. The balance of the plane was close to identical between them. The end result were flight characteristics unique to the plane. If you flew it like the original or the similar hp model you were in for a surprise.
Every single plane model is different. The goal is to make them fly as similar as possible to designs in the same general category. Boeing did this with the 757 and 767. One is a narrow body and the other a wide body yet the type certification is the same. The 737 series is no different in this respect. The goal is to make it fly as close to as many other aircraft as possible.
Commercial airlines are fly-by-wire. Virtually all control inputs are made using artificial feedback. The plane is doing what the computer interprets the pilot wants to do. The pilot is engaging the controls based on how they feel. The initial public speculation on the 737-max problem is that a negative feedback loop exists with the MCAS system. Boeing’s solution seems to indicate this is the problem. They are limiting the number of computer generated corrections in this system to one cycle.
It might have been solved with training but 2 major crashes demand something of a more tangible nature.
I don’t think the issue is elevator authority. It’s not a function of not being able to alter angle of attack. The nose of any airplane will rise and fall under various inputs. It’s a function of reacting to the change in pitch.
In this case the pilot is allowing the nose to come up higher than it should and the plane independently making adjustments.
If I understand this correctly it’s not happening with the auto climb engaged but rather when it’s manually flown.
If there are people who track such stats, does it seem that Boeing jets crash far more often, on a per-jet basis, than Airbuses?
(Some of those aren’t Boeing’s fault, of course, but were caused by malicious human action instead - all four airliners hijacked on 9/11 were Boeings, the Malaysian 370 possible-pilot-suicide was a Boeing, the Egyptair pilot suicide was a Boeing, the SilkAir pilot suicide was a Boeing, the Malaysian jet shot down by Russian missiles was a Boeing, the Pan Am Lockerbie bombing was a Boeing, the Korean Air Soviet shootdown was a Boeing, the Ethiopian hijacked crashing in the sea in the 1990s was a Boeing, etc.)
Great cherrypicking of data. Now give us a list of the non-Boeing crashes, bombings, and accidents, like Air France 447, so we can compare. Got any more?
Except it appears to be true. I think you have a naive or idealized view of the way big corporations make decisions.
Look on any forum for professional pilots, like pprune.org (thread on the Ethiopian airliner crash currently 87 pages long) or aviation24.be and see what they they are saying about MCAS.
I design and build model airplanes and drones for a hobby, I know how much of an exercise in compromising attributes it is to design airplanes.
Also know full well what a complete pig to fly a design with poor stability is, be it from static or dynamic stability reasons. For example, would you be comfortable flying in your airplane if you knew the only thing that keeps it from going into a divergent fugoid is an electronic flight controller, IOW one more thing that can fail?
The issue, ISTM, is precisely that the 737 MAX does not fly as the previous models, if it did it wouldn’t need this new artificial stability system, would it?
The design compromise Boeing made when they decided they’d go ahead an inherently unstable airplane for the sake of saving on design and manufacturing costs and trying to fix it via software, has already cost the lives of almost 350 people in just six months.
So what would the solution be? As Richard Pearce pointed out the plane could not be certified unless it had the artificial stability system, that means, I presume, that as-is the plane has unsafe flight characteristics. Yes they could change the MCAS for something that doesn’t kill everyone on board when it glitches but that is still another layer of complexity to paper over the fact that the design has, presumably, unsafe flight characteristics.
That is why I think the solution would be to fix the airframe, and if that is not economically feasible too bad for Boeing but enough people have already died because they stuffed up royally.
Well I’ve built, modified and certified a real airplane and have flown it. so, your example of models is noted for discussion. Not to be rude but models don’t mean much beyond wind tunnel data.
Every single airplane built today is fly by wire and the “stability” of the plane is software linked to engineered tactile input to the pilot.
none of the previous models fly the same as the previous models. each model is a stand-alone engineering design.
there is no “compromise” between models per se. The next generation of each aircraft is effectively a new aircraft based on previous construction techniques.
The solution is a system that reacts to anticipated pilot inputs. If there is a flaw in that system it needs to be fixed. If there is a flaw in flying parameters then that has to be fixed.
It looks like the problem with the software is not easy to fix, and may have run into unanticipated difficulties. After the previous crash, Boeing promised an update by the end of 2018. Now we are still waiting for the update, even though it’s costing Boeing a fortune. On Friday they were still saying, “no later than April”.
An article in The Air Current has some interesting insights into the financial pressures on Boeing, after the American Airlines decision in 2011 to buy 260 Airbus aircraft.
There is also an interesting light on the Lion Air crash:
Boeing has refused to comment on making the AOA disagreement alert optional, at extra cost. Since there were only two AOA sensors, the MCAS system couldn’t determine the correct AOA once one sensor failed.
And the pilots hadn’t been trained on the MCAS because Boeing hadn’t deemed it necessary…
Your first comment isn’t true. It’s nearly true and I wouldn’t have bothered to correct it except that the aircraft under discussion is not fly-by-wire itself. All the other main players are, the A320 and all subsequent Airbus models are FBW, the B777 and B787 are FBW, but the B737, being a really fricken old design, is not, not even the MAX.
The second isn’t really true either. Typically in the past, making a new version of an existing design has involved extending the fuselage by literally adding fuselage plugs forward and aft of the wing, and upgrading the engines. The end result feels much the same to fly, it might be heavier but have the same wing which means the speeds are a bit faster and perhaps it will feel a touch heavier on the controls. Any real differences tend to be at the edge of the envelope where no one intentionally flies anyway.
I’m not quite sure where I stand on the wisdom of using a software solution to an aerodynamic problem. On first glance I would agree that it is not reasonable to allow an uncertifiable airframe to be fixed with a “patch”, for want of a better word. But on the other hand that is exactly what stick shakers and stick pushers are, without these no airliner would be certifiable but we’ve been accepting of those for several generations of jet transport aircraft.
I think where MCAS falls down is that it only needs one AOA sensor to activate it. Where stick pushers are fitted they typically need both AOA vanes to sense an impending stall. Pushers can also be physically over powered, they can be shut off by dedicated shut off buttons that light up when the pusher activates, and they are a system unlike any other so they can’t be confused with anything else. My understanding is that MCAS solves a slightly different problem to a pusher, but a misbehaving MCAS is just as dangerous as a spurious stick push and so should probably incorporate similar design philosophies.
By the way, something I learned researching for this thread, Boeing’s in general don’t have stick pushers. My previous experience has been with T-tail turboprops and jets and they have mostly had stick pushers due to their ability to “deep stall”, the turbulent air off the stalled wing blankets the tail plane robbing it of the necessary control authority to pitch down and recover from the stall. This experience lead me to believe that pushers were more common than they actually are.
Fly-by-wire was a poor choice of words. All modern aircraft use augmented drive systems where the input of the pilot is translated to a system of electro/hydraulic mechanisms. The feedback to the pilot is artificially driven by design. If you look at the A/A flt 587 crash it was the result of artificially light pedal feel and their there was a change in pedal feedback plus inadequate rudder limits. If I’m remembering it correctly these were changes made to the design of the plane from one model to the next.
Well yes and no. If you look at a DC8 the segmentation is as you describe and probably had similar feel until they got to the CFM engines. The DC-9 conversion to the MD-80 looks like it was segmented forward of the wings. But conversion may be the wrong word to use. Maybe design extension is a better word. I suspect the MD-80 has a different feel to it than a DC-9 but I’ve never asked a pilot who has flown both.
Computers have blurred the interface between pilot and plane. When planes got so big that pilots could not physically operate a cabled control surface then control feedback became an artificial creation. Computers can be reconfigured easier for feedback. And as automated trim systems are added a computer is easier to reconfigure.
The word “patch” is as you suggest, a tricky word to use. All airplane designs have unique characteristics. Certification, IMO, is a process of identifying structural limitations and ensuring the plane does not operate beyond those limits. I’ve never liked the idea of a computer overriding a pilot’s input in a way that is invisible to the pilot. To me that’s a patch. But that’s what roll control does.
If it were me, I’d like a visual indication that the MCAS is overriding pilot or autopilot commands and I’d want the disconnect to the MCAS located next to this visual display.
I am not a pilot and understand only the most basic basics of aviation and piloting.
I was thinking about 2001: A Space Odyssey (bear with me a moment… ) where you often see a video monitor in the background with a random combination of 3 capital letters. I always wondered what those were supposed to indicate. No one seems to ever look at them but maybe they would be critical information in some situation.
That led to me imagine a small monitor in the cockpit that lists what changes the auto-pilot or other systems are making to the pilots inputs at any given time with significant changes high-lighted. It’d be something the pilots never even look at unless they were trying to troubleshoot really bizarre aircraft behavior where they could see at a glance what auto-system is being activated and when.
Seems like that could clear up a lot of confusion during a crisis.
Perhaps I’m missing your point here, but relating altitude AGL to the phase of a flight is only valid for normal flight profiles, which Ethiopian most certainly was not. Your statement about low airspeed during the initial climb is certainly true and borne out by my cited reference flight of a MAX 8 in a normal climb, which was only doing about 200 kts (ground speed) as it approached 1500 AGL, pretty much as you describe. But it was doing this less than 60 seconds after takeoff – by the time it had been in the air three minutes (where contact with Ethiopian was lost) it had already climbed to an altitude of over 7500 ft AGL.
I would guess (strictly a WAG) that Ethiopian’s relatively high speed was some combination of unintended powered descents and likely the pilot throttling up as he fought a possible nose-down trim problem.
The latest announcement from the Ethiopian transport minister was that the flight data recorder is already yielding evidence of “clearly similar” behavior to Lion Air – but no further details were provided.
That is a standard feature of modern aircraft. It is called the Flight Mode Annunciator (FMA) and lives at the top of the primary flight display (PFD). Rather than being something that is hardly ever looked at, the pilots live by it. In the Airbus world “understand your FMA at all times” is rule number 3 of their golden rules. The FMA is a window into the soul of the aircraft, it tells you what it is doing. Any changes to the FMA, ie a new mode becomes active, are boxed in white for a few seconds and must be called out by the pilot flying so that both pilots are aware of the current mode status of the aircraft. Any downgrade in FMA, ie a change from a high level of automation to a lower level, will be accompanied by a triple click sound.
It can go a long way toward preventing a crisis in the first place. Unfortunately, once the crisis hits the poor human brain tends to get a bit of information overload. An argument for not having humans at all? Not really, the crisis normally happens when the automatic systems have already failed to do their job.
Yes, the feel is artificial, not necessarily electronic computers by the way, mechanical systems of springs and weights do a perfectly good job
I’ve flown short and long versions of the Dash 8 and short / long versions of the BAe146. If I wasn’t told which I was flying, I wouldn’t know the difference. They don’t have any computers either, the ailerons and elevators are operated by cables and pulleys.
There’s a difference between adding feedback to a system because it is controlled remotely and the controller wouldn’t otherwise feel any feedback, and adding feedback to mask unsavoury handling characteristics.
It all depends on why the computers, if there are any, are doing what they are doing. In the Airbus A320 and A321 the normal control law does all sorts of shit in order to make it really easy and safe to fly. It trims automatically and it waggles the ailerons, spoilers, and elevators around so that when you set an attitude it stays there. When you get below 50’ during a landing it trims nose down so that you have to hold back pressure on the stick during the flare just like you do in any other plane.
The thing is though, you can turn two of three redundant systems off, the inertial reference systems, and it reverts to direct law. Direct law is exactly like a Cessna 172, there’s no auto trim at all and the flight controls are moved in direct proportion to the control input. It flies perfectly well like this, there are no vices, and it doesn’t do anything odd, this is because the fundamental aerodynamics are sound. It is not an unstable FBW fighter jet that would not be flyable without the FBW, it’s a conventionally stable aircraft that has FBW to make the pilot’s job a little easier.
It seems to me like the re-engined 737 may have lost the fundamentals at some point in the design process. That said, the MCAS is a system that operates near the edge of the envelope, it’s not like it is intruding all the time. I’d suspect that in normal ops a pilot would never see it operate.
Yeah. Or make it like the stick pusher systems I’ve seen where the visual display and the deactivate button are the same thing (buttons light up, push the buttons to deactivate). More importantly, make the MCAS require two high AOA inputs, not just one.
Yeah, perhaps we are talking past each other. You were comparing the Ethiopian speed to some other flight profiles at similar altitude AMSL which showed the Ethiopian aircraft was going a bit faster, but not massively so. All I meant was that, when compared to the profile it should have been flying reference the ground, it was going nearly twice as fast, which shows it was way out of control in terms of speed.
