Glass Class: Automated flying

This is the fourth in a series of articles about learning to fly a Technically Advanced Aircraft (TAA). Ninety percent of production aircraft rolling off the assembly lines of the five largest general aviation manufacturers have glass cockpits that meet the TAA definition. This series will help you to learn how to fly them more efficiently.--Ed.

Using a Technically Advanced Aircraft's autopilot on a cross-country trip makes it easier for the pilot to monitor the flight's progress; scan for traffic; and look at weather, terrain, or other information on the multifunction flight display (above). Of course, in order to use the technology, the avionics must be properly configured; for example, on some TAA, you must enter the altimeter setting in three different places--including the autopilot (below).

This S-Tec autopilot, installed in a Cirrus SR20, is positioned above the aircraft's transponder and below its two Garmin GPS receivers (above). Buttons below its display are used to select operational modes, including Heading and Navigation, as well as Altitude and Vertical Speed (below).

The Avidyne Entegra displays autopilot mode at the top of the primary flight display; heading (above) and course tracking with GPS steering (below).

Disconnect the autopilot by pressing the autopilot disconnect button on a Cessna yoke, Diamond stick, or by pressing down on the hat switch of a Cirrus sidestick--and if that doesn't work for any reason, pull the autopilot circuit breaker.

We have found that there are two key challenges to safely learning to fly Technically Advanced Aircraft (TAA): programming knowledge and cockpit fixation. Each needs to be addressed early during training. Last month, we discussed the TAA navigation and flight planning process and keeping ahead of the aircraft in flight.

Most TAA aircraft use a moving-map GPS to depict and organize waypoint information, an integrated autopilot that is coupled to that navigation guidance to execute a predetermined flight path or procedure, a flight management system (FMS) as the primary interface between the pilot and the avionics, and an overriding computer program or data controller to coordinate all those pieces.

Earlier in this series, we discussed the Garmin G1000 and Avidyne Entegra primary flight display (PFD) and multifunction flight display (MFD) and how those subsystems' computers make assumptions about how the information in the flight plan should be processed. Then we discussed the flight management process of how the pilots gets information into the navigation and flight planning portions of the system (using the "bump, scroll, and twist" process).

One thing that really sets the TAA aircraft apart from the conventional-cockpit airplanes still used regularly for flight training is the presence of a capable, coupled autopilot. The various systems continue to bedevil pilots and worry educators concerned that a lack of thorough knowledge about autopilots will lead to preventable accidents. This installment will look at features of the most typical autopilots, manufactured by Honeywell Bendix/King, Garmin, and S-Tec.

The presence of an integrated autopilot in the cockpit is not new. General aviation aircraft have come with capable autopilot systems for many years. What makes the TAA aircraft unique is how the autopilot is integrated. TAA aircraft cockpits have autopilots that are connected with the other navigation and guidance systems through software and system processors, which offer very powerful capabilities. Previous autopilot systems had very limited integration with the rest of the avionics stack, and there was no overriding automation to coordinate functions--the pilot simply set desired courses or navigation signals on the heading indicator or horizontal situation indicator (HSI) with a heading bug, and configured a series of switches on the autopilot control panel to obey those signals. If the pilot entered improper information, the autopilot would fail to arm, or might follow the improper settings to the detriment of the pilot or the chagrin of air traffic control. The pilot needed to be the "brains" that coordinated what was on the instruments and what was set into the autopilot.

Today's advanced aircraft have started down a road toward computer automation in the cockpit that allows the pilot to develop a very trusting relationship with the navigation and autopilot equipment. This relationship could quickly be spoiled by a simple failure of some of the electronics features. The FAA and the insurance companies want to make sure that the pilot is fully in control of the automation and is supervising it at all times while it is performing its programmed function. Pilots who fail to master the system--think about people who let the digital clock on their VCR blink because that's easier than figuring out how to reset it--will face a crisis if the system decides to fail, especially in instrument conditions. The other major issue is the fact that these systems allow a low-time students to access technology that was available only in jets and space vehicles just 20 years ago.

What does all this mean to the student? First, it requires the pilot to learn these systems and how to control them, as well as how to override them, and how to manipulate them in the event of an emergency. These training processes have never been offered in private pilot courses, or for that matter in most instrument-pilot curricula, but now they must. If flight instructors don't understand them, then they will not teach them effectively, and we could see pilots who are trying to learn systems while they are taxiing the aircraft or while they are blazing across the ground at four miles a minute.

The current crop of autopilots found in TAA aircraft falls into two categories: digital and analog with digital faces. The difference between these two is an amazing amount of technical capabilities. The pilot must know the functions and capabilities of the autopilot installed in the aircraft that he or she flies.

The first generation of TAA autopilots are analog. The S-Tec Fifty-Five X is a very common autopilot platform found in most Avidyne Entegra-equipped Cirrus aircraft and many 2004 and 2005 Mooneys with G1000 panels. The Honeywell Bendix/King KAP 140 autopilot is found in 2004 through 2006 Cessnas and Diamonds equipped with the Garmin G1000. Both of these autopilots feature advanced functions and are reasonably integrated into their host glass-panel cockpits.

The challenge to keep the pilot from getting into trouble is understanding the limits of overriding automation that is built into the interface. Both of these autopilots depend upon a hidden turn coordinator, which provides these systems with analog turn-rate information. That means that if the connection to these hidden components fails, the autopilot can lose its ability to determine how to make standard-rate turns. In addition, these units depend upon a sensor that samples ambient static information to help them determine how to climb, descend, and hold altitude. In a KAP 140-equipped aircraft, the pilot must input the altimeter setting into three places: the autopilot, the G1000 system, and the standby altimeter. The S-Tec autopilot does not require an altimeter setting to be input into its control panel because it can get this information from the glass cockpit at the time of autopilot engagement, so pilots flying these aircraft need to enter the altimeter setting only twice.

The pilot must monitor these systems to avoid a situation in which the autopilot attempts to maintain a predetermined vertical climb rate, set at a lower altitude, that could stall the aircraft at a higher altitude. This stems from the lack of digital information flowing back and forth between the glass cockpit's processors and the autopilot. If there is not sufficient or valid information, the overriding computer programs that guide and coordinate the systems cannot perform basic advisory or corrective functions, leaving the pilot to fend for himself.

There are other limitations with these systems. There are a number of instrument procedures that these units simply cannot execute. An example is the standard holding pattern depicted on nearly all instrument approaches. The early versions of these autopilots do not have the integration necessary to fly these procedures and require the pilot to hand-fly the aircraft, or use the heading mode to steer the aircraft around the holding pattern like a go-cart at a racetrack. Thus, the pilot must keep constant vigilance over the systems, watching to make sure that the autopilot is doing what was intended. This need for pilot oversight leads to an increased training requirement, despite the fact that the increased level of cockpit technology might imply to the pilot that his or her function has been reduced to one of a cockpit automation manager.

The digital autopilot, such as the Garmin GFC 700 installed in Beechcraft, Columbia, and newer Cessna, Diamond, and Mooney aircraft, incorporates a high level of automation coordination between the systems. This system not only has the ability to watch over the pilot's operational parameters and provide certain advisories, but it also is capable of flying very complex instrument procedures--19 to be specific--such as holding patterns, procedure turns, and a wide variety of other procedures that experienced pilots will be thrilled to automate. It's difficult to appreciate the differences until you've had a chance to fly using both types of autopilot.

A glass-cockpit student should know several autopilot attributes. First is the concept of axis control. Typical single-engine autopilots control the aircraft on two axes: pitch and roll. Every action that the autopilot can do for the pilot depends upon its ability to control special servos along the control cables leading to those respective flight controls (or their trim tabs). The roll modes are the easiest to understand. The autopilot is instructed to follow a heading or a navigation signal by the pilot's selection of the HDG or NAV button on the autopilot control. The pilot will know that the autopilot has armed in that mode by a steady indicator on the autopilot control or the PFD autopilot mode display strip. Blinking indicators on the autopilot control unit or a display reflecting ROLL tells the pilot that the unit has not achieved the desired state and in fact may never correctly arm if the pilot doesn't intervene. Many times the pilot must use the HDG mode to steer the aircraft to a point close enough to the desired navigational track so that the NAV mode will arm. Pilots should watch these indications closely to make sure that the autopilot is in the desired mode.

Regarding the pitch mode, the pilot must be careful to understand all of the possible modes of the autopilot's pitch channel. Usually, there are two modes: vertical speed (VS) and altitude hold (ALT). If the autopilot has an altitude intercept feature, then the autopilot can be configured to climb or descend at a specific rate of speed and then level off at a predetermined altitude. Most of the errors we see students make regarding autopilot use are in this area. A pilot who expects the autopilot to level off at a preset altitude can be surprised when it keeps climbing right past that altitude. This happens from not arming the autopilot correctly or by disarming it accidentally. Work with your instructor and determine the precise set of steps required to properly arm the autopilot to accept a level-off command and then mentally count down the last several hundred feet as that altitude approaches, so that no surprises occur.

There is one way to turn each autopilot on and multiple ways to turn it off. Know exactly how to turn off the autopilot in your aircraft and always verify that it is off before takeoff and landing. Perform the prescribed autopilot and trim checks before each flight as instructed in the autopilot supplement. Some aircraft checklists do not describe this process in enough detail, and then pilots begin to skip this step. This check is your final opportunity to determine that your autopilot is ready for your flight.

You should now have a better understanding of the autopilots used in glass-cockpit aircraft such as the Garmin G1000 and the Avidyne Entegra. The pilot must learn these systems and become completely skilled at preflight testing, normal operation, and emergency operation modes. Their instructors must test the pilot's knowledge of autopilot limitations and ability to recover from autopilot stalls and unexpected or uncommanded responses. The pilot will find that the skillful use of the autopilot is not difficult--and that it can, in fact, be the key to staying ahead of the ever-faster TAA aircraft by allowing more time to monitor cockpit, navigation, weather, and air traffic conditions.

Michael G. Gaffney is president of Skyline Aeronautics at Spirit of St. Louis Airport in St. Louis, Missouri. A Master CFI and a Master Ground Instructor, he was named the 2007 national Flight Instructor of the Year. Gaffney, author of ASA's The Complete G1000 Course, also holds airline transport pilot and airframe and powerplant certificates.

Glass class online

See accompanying multimedia at ft.aopa.org/glassclass4/.