Jumpseat: What's It Doing Now?

The love/hate of cockpit automation.

Jumpseat Autopilot

Jumpseat Autopilot

A brief smirk passed my lips. In an effort to hide my nerves, I made a feeble attempt at humor and queried the simulator instructor, "You really want me to execute the ILS without a flight director?" His wordless, stone-faced response wasn't quite the reaction I had hoped for in alleviating my anxiety.

I was about to demonstrate my flying abilities as part of my airline's hiring process. For this part of the simulator session, an applicant was allowed only the use of good old-fashioned left/right, up/down needles. The magic guidance of the bat wings and delta bars would remain off. Raw data only. Fortunately, the story had a happy ending.

Reflecting on the experience from over 31 years ago, it was probably the first moment I realized my own growing reliance on automation despite the fact the approach was performed in a 707 simulator, antiquated by today's standards. Considering that today's single-engine, GA airplane rolling off the assembly line has GPS-magenta-line technology, flat-screen displays and push-one-button auto­pilots, it seems laughable that a flight director would be thought of as an automation epitome.

I could remember a few moments on the 727, the airplane I began my airline career flying, in which I stared in confusion at the flight director delta bars and asked the classic question "What's it doing now?" More than likely, as is often true in most cockpits, some form of data was entered in error; the machine was just attempting to perform according to the request it had been given.

The "three-holer," as the 727 was affectionately called, had an autopilot with limited capabilities. What the autopilot lacked in sophistication it made up for in personality. Two small paddles on the center pedestal were required to electromechanically engage the autopilot. An odd-looking knob that resembled Groucho Marx's nose was rotated left or right for turns and pushed forward or back for climbs and descents.

The finicky nature of the knob required finesse in order to start the airplane down from cruise altitude. Too much movement and the passengers were treated to a noticeable bump. Too little movement and the airplane was stubborn to begin a descent. Everyone had his or her own technique. And everyone knew that his or her own technique never worked consistently.

Although the autopilot of the 727 was capable of executing a very precise coupled ILS approach right down to Cat II minimums, it lacked autothrottles (unless they were added from the aftermarket). Perhaps tedious at times, having direct control of the power kept us in the automation loop.

The three-holer had limited pilot-stupidity protections. As an example, a clacker would alert an offending crew that the airplane was exceeding its max airspeed or Mach. For stall protection, a stick shaker activated and vibrated the control column in anticipation of impending aerodynamic doom. It was then up to the pilot to respond with the appropriate recovery technique.

Leaving the automation dinosaur days behind, I transitioned to the 767. The airplane might as well have been the Starship Enterprise. The concept of being able to manipulate a 300,000-pound machine from a relatively small control panel set into the eyebrow of the glareshield was foreign to my psyche. It felt awkward.

The 767 was a game-changer in airline flying philosophy. Boeing had designed the automation to be more efficient than a pilot, integrating the FMS (flight management system) computer and its performance calculations into the methodology in how the wide-body was flown. In training, we were encouraged to utilize every aspect of the autopilot/flight director system more than hands-on flying. Only when the automation seemed to conflict with the flying task in progress was it considered appropriate to reduce the level of automation.

As an example, early in simulator training my frustration peaked during performance of an engine failure procedure. My brain and hands weren't connecting with the buttons and knobs of the mode control panel, nor were they connecting with the autothrottles — which were moving like a player piano directed by an alien life-form. I declared to my sim partner/captain and the instructor, "I've got the airplane!"

One thumb pressed the autopilot disconnect button on the control yoke while the other thumb pressed the disconnect button on the autothrottles. I completed the emergency procedure, executed the ILS approach, and landed the airplane with my own hands, albeit not so gracefully.

Eventually, I got it together and life was good. But I vowed never again to let automation defeat me. That being said, as I gained experience with the 767, the question of "what's it doing now?" became a common theme for everyone flying the airplane.

The mystery of the "VNAV" button was a common source for consternation. It never seemed that the function of VNAV was consistently in harmony with the autothrottles, the FMS or the autopilot. The programmed crossing altitude would be reached too late, or the descent airspeed would exceed the entered value, or the autothrottles would revive from hibernation and re-energize with more power. And so on …

Through software updates, the VNAV function improved. Although even on the 777, a fully electronic airplane, it's still not quite perfect. Understanding the parameters of each automation function is key to safely flying the big jet.

A new philosophy of airplane operation also brought with it some protection systems. In addition to a stick shaker activating when approaching a stall, a stick nudger automatically pushed the control column forward in an effort to reduce the angle of attack. And in conjunction with an impending stall, the leading edge devices responded by automatically extending to increase lift. The autothrottles, however, did not increase power automatically. Pilots were required to either push the go-around button or override the autothrottles during a stall event.

Other protection systems warned of wind shear conditions. An electronic horn sounded when the gear wasn't down with the flaps extended to a landing configuration. On a go-around, the autopilot set the pitch attitude to a fixed rate of climb with the selected airspeed. Three auto­pilots with separate electrical ­sources operated in synchronization to perform a Cat III approach with no ceiling limitations and a visibility as low as 600 feet, touching down without assistance from human hands. All pretty cool stuff.

When I moved on to the 777, it was apparent that the automation of the 767 had been refined to the level of a well-aged California pinot noir. The 777 is a complete fly-by-wire airplane. It has its own autonomous IT department located in the E&E bay below the forward cabin floor.

Many of the 777's protection systems are similar in nature to those of the earlier-technology 767. But one common denominator still exists. Even if a pilot has induced a bad situation to the airplane, i.e., a stall, the automatic protection features are designed to assist in recovery. The airplane will not prevent nor drastically impede human intervention. I find this notion important.

An airplane manufactured with the philosophy that its automation systems will protect the airplane from the pilot rather than working in harmony with the pilot is not a product I have a desire to fly. People will make mistakes. But as a pilot, I hope you don't prevent me from correcting those mistakes by completely taking away my control. Rather than point fingers at certain airliner models, I'll leave that controversy for another discussion.

When I began my airline career, training involved mentally rebuilding airplane systems, redrawing schematics, tracing electrons and memorizing obscure numbers. This was the norm. With the design of today's complex automation systems, centered around highly advanced computer technology, such rote knowledge is not necessary and is virtually impossible to fully comprehend.

Although it's still important to have an understanding of the machine, it becomes more of a priority to understand cause and effect as it applies to automation. If you press this button, if you enter this data, if you activate that switch, what happens? These are questions that should have answers on every flight. Perhaps if these questions were answered, the dramatic event of Asiana Flight 214 in San Francisco would never have occurred.

So rather than asking "What's it doing now?" perhaps we should be asking "Why is it doing that now?"

I love the autopilot — and I hate the autopilot.

Get online content like this delivered straight to your inbox by signing up for our free enewsletter.