Manual override and Occam’s razor

Occam’s razor: the simplest explanation is usually the correct one.

I heard a National Public Radio story the other day about a paper that had come out regarding the Boeing 737 Max accidents. The crux of the interview was the quality of training, or lack thereof, pilots receive and the thought that maybe cockpits and cockpit procedures could be made easier, so recognition of problems would therefore be easier.

My initial reaction was one of defensiveness. I believe pilots as a whole are a fairly robust group. We’re conscientious. We’re obsessive in matters of safety, convention, and procedure. We’re intrinsically averse to the notion of tombstone technology. We’re… competent. It’s not an ego thing. Not even necessarily a confidence thing. It just is. If you endeavor to take to the skies, to leave terra firma behind, you commit to a level of competence and focus necessary to complete the task safely. All the time. Every time.

787 cockpit
Those avionics are great – really great – when they work. Then what?

But then I took a mental fork in the road, defenses down, and followed another branch on the tree of possibility. Could there perhaps be something to the notion of making what we do, dare I say, easier? And let’s not confuse easier with easy. We’re not talking couch-potato easier here. We’re talking easier in a safety sense, an efficiency sense. What if there were an easier way to revert to manual control? To remove the so-called “envelope protection” algorithms built into modern flight control systems.

We’ve all heard the adage: Aviate, Navigate, Communicate. But can you really aviate when control inputs are analyzed thousands of times a second and then spit out to the control surfaces? (Yes, I know, aviate is an all-encompassing word that means more than just flight control manipulation.) But when an aircraft is doing something you didn’t expect, and doing it relatively fast, especially at low altitude, you need to act NOW. Unlike some emergencies, this type rarely gives you the time to sleuth out a solution. When everything seems to be going awry and nothing makes sense, aviate means FLY THE PLANE. If a cargo door opens on one of our smaller GA aircraft, no big deal. Fly the plane. I think I heard Rod Machado once say, “Don’t drop the airplane to fly the cargo door.” Exactly! FLY THE PLANE.

What if there were a “Manual Override” button? Not a procedure. Not a checklist. A button. This button wouldn’t be a get-out-of-jail-free card or a save-me-now button. In fact, quite the contrary. It would return control – full control – to the pilots, giving them the authority needed to fly the aircraft, sans auto throttle, electric trim, augmentation, envelope “protection,” etc. The button wouldn’t be a solution in and of itself. It would simply provide the pilots with a chance to take desperate measures in a desperate time. I can think of several incidents/accidents that could possibly have been averted had such an option been available. (Air France 296 comes to mind.)

The pilot is still the most important safety feature on any aircraft. So, there needs to be a way he/she can actually pilot the aircraft when circumstances call for it. Barring some gross mechanical malfunction (e.g., Alaska Airlines Flight 261), if the aircraft remains flyable (e.g., United 232 and US Airways 1549), pilots will revert to their own internal autopilot mode. Their intuition as it were. The computer that is his/her brain will kick in.

Their innate flying sense, training, experience, and muscle memory will all coalesce in subconscious harmony to create a desired result; a state athletes refer to as the “zone.” Great NFL wide receivers don’t think about catching the ball. They just do. It’s almost like an out-of-body experience (see the book The Inner Game of Tennis by W. Timothy Gallwey). Instead of being dumbfounded by incongruent aircraft behavior which only feeds the fear/panic cycle, the pilot will instinctively do what is necessary to right the ship.

This state of selfless awareness is not something mystical. We’ve all felt it, if only briefly in the fleeting few seconds during the flare. It’s almost Zen-like. The temporary meditative state we feel when total concentration and focus push everything else away. It’s when we’re operating at our highest level, and a place where we can find the wherewithal necessary to calculate a way out of extreme danger.

Obviously, fly-by-wire is integrated in most, if not all, modern airliner flight control systems. The pilot/computer interface can’t necessarily be separated nor eliminated entirely. Aircraft use computational “laws” (automation, augmentation, protection modes), thus true raw-input control is virtually impossible. As it turns out, though, the computer seemingly doesn’t always know best in extreme cases.

Would better pilot training and aircraft manufacturer transparency help? Absolutely. But under certain conditions, the laws meant to protect and serve simply run interference. It is in these extremely rare instances a new sheriff needs to be instated. That sheriff is the pilot. The pilot must be given the opportunity to drop the computer and fly the plane.

MOR lever
Single-engine turboprops have a manual override lever – do fly-by-wire airliners need one too?

We’re so caught up in the marvel and complexity of modern aircraft we may think it couldn’t possibly be that easy. There’s got to be more to it. Of course, the fly-by-wire/pilot control discussion is nothing new. Perhaps that is why we have such a hard time accepting accidents that seem to be so clearly avoidable. Sometimes we just want a plane to be a plane. There are indeed procedures in place to revert to some semblance of basic aircraft control. And Cirrus has their “LVL” button which is similar (though not the same) as what’s being proposed here. But still, when an aircraft’s flying qualities deteriorate and pilots find themselves in a Twilight Zone of altered perception and unfamiliar territory, there’s got to be a better out.

I used to work for a company that had an issue with a specific product from one of our third-party manufacturers. We were getting an inordinate number of returns due to component failure. When I approached the manufacturer about it, they took it personally: immediately dismissing the issue, stating their design and materials were not the problem. They insisted that customers were not using the product correctly and applying too much pressure, causing the failures. We went back and forth on these points to no avail.

In the end, my final word on the matter was that this was not personal. Whether you think your product is designed and built properly doesn’t really matter. The failures are occurring, the product is being returned, customers are unsatisfied and inconvenienced, and we can no longer sell the product unless it is redesigned and/or beefed up.

We can deliberate the deficiencies of pilots and pilot training ad infinitum. But that would deflect from the fact that these accidents did happen, and accidents like them can still happen. Up until the first Space Shuttle accident, the assumption of safety was misconstrued from the fact that nothing catastrophic had occurred in the previous flights to question it. But there is a clear demarcation between the absence of trouble on one side, and safety on the other. Though safety is unquestionably paramount in the minds of all involved, they are not equivalent. This could be an opportunity for change. Aviation, like everything else in life, is a progression.

Is the Manual Override button a gross oversimplification? Total naivety?

Sometimes the simplest explanation is usually the correct one. And sometimes, the simplest solution can save lives.

Occam’s razor.

9 Comments

  • The pilots of both 737 MAX aircraft did have a manual override—for the trim system. This is the same system that’s been installed on Boeing aircraft since the first Dash 80 prototype flew back in 1954. One crew chose not to use it, the other chose to use it too late, and they lacked the training to fully understand how the system works.

    One of the saddest facts about these two accidents (to me) is that just seconds before the Ethiopian Air flight made its final plunge to earth the captain made four attempts in rapid succession to engage the autopilot. He was clearly looking for that magic, get-out-of-jail-free card.

    A manual override button will not help a pilot who doesn’t understand when to use it.

  • Blaming the 737 manufacturer from not highlighting (in extremely bold words in the POH and or advising for a simulation training) the way pilots should act in case of MCAS fail and, consequently, triggered the normal behavior of the plane, is equivalent to put a blanket in duties that airlines should be focusing in the skills of all and every pilot, regardless the degree of automation of the plane and, also, independently of the standard of the country of the airline’s company. To me, that’s the true that most commentators ignored or tend to ignore, because is easer to endorse guilts to the manufacturers than accepting the lack of skills of human brain to overrun the dictatorship of computational imputs to the commands of the plane.

  • With so much automation now a days, we have actually automated ourselves right in to the ground !!! Get rid of all the computer stuff in the cockpit and fly the airplane, what’s so difficult about that? after all that is what you are getting paid to do… Having a very good autopilot in the aircraft is what is essential !! The rest comes automatically. I keep seeing all these pilot’s saying how many thousands of hours they have as PIC when in essence they only have about 1/6 of all that written time, the rest is auto pilot time, not even a comparable sum..When you keep introducing computer stuff into the cockpit it will eventually become too much, and that is where we are today, and it does not work !! Fly the darn airplane, it’s more fun any way than sitting there texting your friends and family members at home. If those two airlines that crashed recently had not had all that pc stuff in the cockpit they all would be alive today..sorry, but some one had to say it….

  • An overly simplistic solution,which might help in some situations and complicate others. There is absolutely no reason to believe that a tired pilot facing an unexpected catastrophic problem at hour 6 of a 9 hour leg, back side of the clock, will instantly grasp the full complexity of his situation (AF407, for example). Going full manual would save some and finish off others. (B767CA Ret)

  • One of the things never mentioned is why the 737 Max exists in the first place. The massive bureaucratic system that must be navigated (at great expense) makes it tempting to keep stretching the life of (and modifying) old designs.

  • Descending out of 9,000 on auto pilot over southern Ontario my stall warning went off in a 30 day old Cirrus SR-22. I immediately thought “faulty sensor” and continued to fly the plane on AP, with the warning coming on and going off intermittently. I shot an instrument approach to Muskoka and at DA kicked off the AP — but it didn’t turn off. The ESP had activated due to the faulty stall sensor and kept turning the AP back on. I couldn’t turn it off in the 15 or so seconds I had before landing, so I manually overrode the controls and landed the plane. Being a relatively new pilot (500 hrs TIT), I consulted with Cirrus and my mechanic on the ground and was told to pull the stall sensor breaker and program the MFD to turn the ESP off. Problem solved until a new sensor could be installed. I also didn’t know that if I held the red AP shut off button on the yoke down it would prevent the AP from coming back on. I refer to this situation as my “very own MCAS experience”, although not to downplay the severity and tragic consequences in the Boeing planes. Fortunately for me, I had the ability to overpower my AP and land the plane, but I wished at the time that pressing that little red button did what it was suppose to do. Lessons learned and fortunately no harm done.

  • Some accidents are due to pilot error. Some accidents are due to technology failure. The decision to either have a manual override button or an automatic autopilot intervention to overrule pilots depends on whether you have more faith in humans or more faith in technology. I think there is a lot more discussion and outside the box thinking required before making that decision.

  • Can the pilot actually “aviate, navigate, communicate” in any commercial passenger aircraft? Can the PIC actually turn off the entire computer system and fly by mechanical means? I think not, especially in the fly-by-wire birds. So if Boeing/Airbus aren’t training pilots precisely how to turn off significant parts of the computer system then they have to take responsibility for what may happen. And if not installing two $2500 sensors instead of one was the cause, shame on them.

Comments are closed.