The Blessed Curse of Automation

December 31, 2015

Back in the 70s I had an Air Force assignment flying the T-39 Sabreliner. The mission was to fly generals, congressmen and other VIPs all over the Pacific in the new (at the time) executive jet. We flew some challenging approaches, such as the IGS to runway 16 at Hong Kong’s Kai Tak airport.

I hand-flew every mission, from takeoff to touchdown. It wasn’t that I was proving my machismo. It was simply because the T-39 had no autopilot, no yaw damper, and no flight guidance system.  Altitude control was the responsibility of the pilot manipulating the controls. To get from point A to point B, we had to tune the VOR or TACAN, identify that it was the correct, reliable navaid by listening to the morse code identifier, and maintain course by centering the CDI, applying appropriate wind correction, and keeping it centered. And, of course, we’d switch to another navaid as we reached the limit of the service volume area.

Now, of course, that kind of workload is a thing of the past in most passenger-carrying operations. And that’s a good thing, considering safety. When I flew the same mission in the glass-cockpit autopilot-equipped C-21 Learjet ten years later, I was much less stressed and more alert at the end of a 16-hour crew duty day. So it was safer to have all the automation.

But the problem is that we pilots rely on the automation. A lot. We get lazy. At many type rating simulator courses, the expression “You don’t get style points for hand-flying” is the catch phrase. In most Boeing airplanes, the autopilot may be engaged at 200 feet on takeoff. Over the years that has morphed into “the autopilot shall be engaged at 200 feet”. And, typically, it stays engaged until landing is assured.

When the autopilot is engaged, the airplane flies the pre-programmed course. There is no centering of the CDI, no wind-drift correction, no skill required. The airplane simply follows the magenta line.

In terms of safety, this can be a good thing. The pilots, unencumbered with the task of manipulating the controls, can plan ahead, maintain a traffic watch, orchestrate all the events that go into a safe flight. Or they can get lulled into a false sense of security. And they can get lazy.

So, is the answer to regress to the old way of doing things, to hand-fly the whole flight? Not really, since we are required by regulation to have the autopilot engaged in RVSM airspace. Should we hand-fly up to FL290 (the beginning of domestic RVSM airspace), then engage the autopilot? Probably not, since we can maintain a better lookout for conflicting traffic when the autopilot is engaged.

Here’s my suggestion: hand-fly the airplane through flap retraction. This will give you a good feel for the airplane through a range of airspeeds. Once the flaps are retracted, engage the autopilot and get your head on a swivel. You’re at an altitude where there are a lot of other airplanes, some of which may not even have transponders. When you reach 18,000 feet, click off the autopilot and hand-fly again up to your cruise altitude or RVSM altitude. Above Transition Altitude everyone has a transponder and is talking to ATC, so you have less threat from other unanticipated aircraft. But keep looking around anyway. I once had a near-miss with a weather balloon at FL340 over the Pacific Ocean!

Following this protocol will give you a good chance at maintaining your stick-and-rudder skills while maximizing safety for yourself and your passengers.