When Instruments Lie

Familiarization with partial-panel operations can get to the truth.

A pilot was climbing through 800 feet in instrument conditions in a Cirrus SR22 when he reported that his vertical speed indicator was jumping between zero and 2,000 fpm. Shortly after that, his altimeter began acting erratically, with large deviations. The pilot requested to return to his departure airport. Then he reported his attitude indicator did not agree with his turn coordinator, followed by a report that he was losing his instruments and was going to activate the airplane parachute system. The airplane was substantially damaged when it came down in some trees, but the pilot escaped injury. The NTSB determined that water in the static system from washing the airplane caused the erratic indications.

One of the mantras drilled into our heads is “always trust your instruments.” For pilots who go on to get their instrument rating, the message becomes even more serious: “If you don’t trust your instruments, you will lose control and possibly die.” Over the years, and hundreds or even thousands of hours, trusting the instruments becomes second nature. Our brains can take in the information from a number of different sources on the instrument panel and form a very accurate picture of what the airplane is doing at any given moment and what effect any control inputs we make will have.

As the following accidents demonstrate, the current trend toward integrated glass cockpits, fly-by-wire and even computer-controlled aircraft has made it even more difficult to sort things out and maintain control when there are erroneous input.

Birgenair Flight 301, a Boeing 757-225, crashed into the ocean shortly after taking off at night from the Dominican Republic. The flight data and cockpit voice recorders showed that the captain’s airspeed indicator did not work at first, then showed increasing airspeed as the airplane climbed until it was indicating an overspeed situation. It turned out that the Pitot tube for the captain’s instrument was blocked, probably by a mud dauber wasp.

The autopilot only received information from the blocked Pitot tube, so it began to raise the nose to try to slow the airplane. Many other erroneous warnings began going off, then the stick shaker activated as the airplane approached stall. The captain continued to try to raise the nose to slow the airplane, ignoring the copilot’s accurate airspeed indication that the airplane actually was close to stalling. Finally, the captain added full power, but at that attitude the left engine failed, the plane stalled and it plummeted into the ocean.

In the case of Aeroperu Flight 603, all the static ports on the Boeing 757 been covered with tape before the airplane was washed, and the mechanics had failed to remove the tape. Besides all three altimeters indicating the airplane was not climbing, the pilots also began to get overspeed warnings. The copilot tried to warn the pilot that the airplane was not stalling and that the “horizons” were “the only thing that was right” before the airplane crashed into the ocean.

In some advanced fly-by-wire airplanes the computer decides how to fly the airplane based on inputs from the pilot and many sensors. There is no backup mode that allows the pilot to circumvent the computer and fly the airplane directly.

That was the case in February 2008, when a B-2 stealth bomber, Spirit of Kansas, crashed on takeoff. It was discovered that maintenance crews, following approved procedures, recalibrated the air-pressure sensors. The sensors had moisture in them, and when the flight crew activated the sensor heat system prior to takeoff the moisture evaporated, leaving a system that was slightly out of calibration. This led the Flight Control System (FCS) to cause a premature takeoff followed by a 30-degree nose-up pitch attitude. Nothing the pilot did was able to overrule the confused FCS. The crew ejected safely, but the $1.4 billion airplane was destroyed.

In all other aircraft, it may be possible to control the airplane and land safely if you respond to this difficult situation in a controlled fashion.

1. Cancel any warning alarms if that is possible, as it is hard to think with warning alarms blaring in the background, providing an artificial impetus to hurry.

2. Remember that it is highly unlikely for both the attitude and pressure instruments to fail at the same time, so if the pressure instruments are going crazy, ignore them and focus on the attitude of the airplane. Conversely, if the attitude instruments seem to be failing, ignore them and concentrate on the pressure instruments.

3. The whole process of integrating various inputs becomes second nature to us, making it very hard for our pilot brains to disregard information that is false, so immediately cover up any instruments providing false information or information you suspect is false.

4. Use the remaining information to fly straight and level. Just because we put computers in airplanes does not mean the basic laws of aerodynamics have changed. In any airplane, a particular attitude combined with a particular power setting will produce essentially the same outcome every time. If you have lost your Pitot/static instruments but still have accurate attitude depicted, put the nose on or slightly above the horizon with the wings level and set the power to normal cruise power. Nearly all of us fly with GPS, and the groundspeed reported by that system will be unaffected by Pitot-static failures. A quick check of the groundspeed can tell you if the airspeed system has failed and can keep you in the ballpark of a safe speed.

If the Cirrus pilot had taken this approach, he would have realized that it was obvious the pressure instruments were not working correctly, and by putting the nose on the horizon and leveling the wings, he would have calmed everything down. While I can’t fault his decision to activate the parachute if he felt he was about to lose control, he might have realized that activating the alternate static source would have been a much less expensive alternative.

An accurate assessment of the situation accomplishes nothing if you can’t fly the airplane with the remaining information. These accidents emphasize the importance of regularly practicing partial-panel operations. The odds of experiencing such an extreme situation may be very remote, but if it does happen to you, you will be glad you can fall back on your training rather than trying to figure things out when your life and the lives of your passengers depend on it.


New to Flying?


Already have an account?