The Training Dilemma

The best companies in the flight training industry have done a good job of making jet pilot training reasonably realistic. LOFT (line operation flight training) is the acronym for simulator sessions that emphasize seemingly small problems that can grow into major emergencies if the crew mishandles them initially.

And simulator-based training is working wonders in jets. The major airlines in the United States have not had a passenger fatality in more than six years, a period three times longer than any other in history. And the regional jet operators have done nearly as well with only one fatal crash. Most impressive is that many regional jet pilots are low on experience compared to the historical standard, but they are products of modern simulator and LOFT. The results speak for themselves.

But none of this good safety news has spread to the general aviation piston fleet. Yes, the total number of accidents, and fatal accidents, was down last year. But any experienced industry observer knows that flying by the piston fleet was also down, no matter what the official estimate from the FAA says. We saw fewer accidents only because pilots flew less in piston airplanes, not because underlying safety changed.

I thought about this safety and training difference when I attended FlightSafety's Baron recurrent training course early this year. Three years ago my insurance company, USAIG, decided that keeping one of my jet type ratings current with a full simulator course each year was not good enough to insure me in my Baron. So now I have to train in the Baron, as well as in a business jet each year. I don't know if this makes me a safer pilot, but it does point out the difference-and the dilemma-between how we fly and train in jets, versus how it is done in pistons. In jets, the axiom holds and we train the way we fly, but that is not true in pistons.

As I progressed through the Baron training it dawned on me that the fundamental difference between the jets and pistons is that in the jets, the training program assumes the pilots already know how to fly. In the pistons, it is the opposite and the program is designed to make the pilot prove he can fly the airplane.

This difference shows up immediately in the training attitude toward automation. In the jets it is expected, actually demanded, that you use the flight director all of the time beginning on takeoff roll. Autopilot use is emphasized, and if you are training for single-pilot jet operation, the autopilot must fly the airplane for all phases of flight except the actual takeoff and landing. All of this equipment fails a few times in the simulator, forcing the crew to recognize the failures and deal with them correctly, but in between failures, the automated systems are used.

Contrast that to the Baron course. For the first day, the autopilot could not be used at all. And the flight director also failed. And on my second instrument approach, the one and only heading system failed, forcing me to fly a non-precision approach to minimums using only the GPS ground track for heading guidance. And unlike in the jets, I was by myself, and the simulator, while reasonably faithful to the Baron, had no motion or other advanced features that make it more realistic, and thus easier to fly.

Train the way you fly had just gone out the window. Any Baron or other piston pilot flying by themselves should be using the autopilot virtually all of the time flying IFR. The autopilot's unwavering attention to altitude, heading or course tracking gives the human pilot time to talk to controllers, look for traffic under VMC, manage the avionics and keep up with the big picture. If it is a crew of two and one of the human pilots is flying, the pilot flying operates like a human autopilot doing nothing but flying while the other pilot talks on the radio, checks the charts, manages the avionics, works the numbers and performs all other tasks. It is a myth that a lone human pilot can do a precise job of IFR flying by hand while simultaneously performing every other necessary cockpit chore.

But what is the training system supposed to do with the human pilot who can't actually fly IFR consistently and competently? And that is sadly the case in many piston airplanes. The pilot shows up for simulator training but the instructor needs to find out first if he can fly the airplane competently, while that is assumed in jet recurrent training. If the sim instructor allows immediate and complete use of the autopilot in the Baron, for example, he has no way of knowing that the pilot can hold a heading and altitude by hand. So the training can't proceed the way you should fly, at least not until well into the curriculum when hand flying competence has been established.

I don't have an answer for this dilemma because as much as I know that automated systems make flying safer, it is also a fact that the human is the backup for the automation and must be prepared to immediately take over. In most jets the autopilot system has multiple channels and sensors, so the failure of single items does not disable the autopilot. That is not the case in any piston airplane I know of, so the failure of one sensor-the compass system, for example-will disable the autopilot. The same is true for the servos that move the controls and the electronics that compute the autopilot functions. It is all single string with the human as both monitor and standby system.

However, I don't think autopilots fail as often as a lone, stressed human pilot can, so I absolutely believe that piston pilots should use the autopilot at all times in IFR, and the worse the weather, the more the autopilot should be used. But how do you teach that in the simulator, and also measure the hand flying proficiency of the trainee? The only thing I can think of is to promote more routine autopilot and flight director use, with failures thrown in. But if the pilot doesn't handle the autopilot failure well, will he become a better hand flying instrument pilot with more autopilot practice? I doubt it.

The business jet and airline safety record proves without a doubt that advanced cockpit automation and thorough and regular training in its use greatly enhances safety. Most piston airplanes used for routine IFR travel have enough automation to enjoy similar safety benefits, but the human pilots in most cases don't have the same level of hand flying skill and proficiency as those in jets. Do you train the piston pilots to rely almost totally on automation as we do in jets and trust that the electronics won't fail? Or do you spend the training time available on hand flying practice and hope the pilot uses the automation fully and correctly in normal flying? That is the dilemma and I don't have an answer. Maybe it's the technocrat in me, but when forced to choose I trust the machine more than the human. What do you think?