Our IFR training does a good job of teaching the instrument scan for normal flying as well as the more demanding partial-panel scan. However, most IFR training is seriously deficient when it comes to the important transition step between the two, the instrument crosscheck technique that allows a pilot to recognize an instrument failure and thus prompts the change from normal to partial-panel flight.

The cross-check concept is taught, although its importance and application are not always made clear, during ground school where we learn it as one of the three official “fundamentals” of instrument flying, the other two being instrument interpretation and aircraft control. We are also tested on knowledge of the system of Primary vs. Supporting instruments for various flight configurations. The system makes sense even though it is non-intuitive, but the important connection between this double-reverse/inverted system and its application as an instrument cross check is rarely emphasized -- and how many of us have actually been taught to regularly cross check our instruments where it counts, in flight?
In fact, we tend to fly merrily along under the hood or in the clag using the primary instruments without crosschecking, i.e., backing them up. In part it’s because of how we’re taught: normal instrument flight runs smoothly enough without an instrument crosscheck so instructors can’t tell whether or not students are backing up the primary instruments, and so they tend to gloss over it. Also, crosschecking is added work in what is already a task-saturated environment so it tends to get dropped, usually without consequences. Yet the instrument crosscheck is an important backup measure that prevents a spatial-disorientation/unusual-attitude disaster by increasing the chance of early recognition of a failed instrument. Its importance only becomes apparent when an instrument actually fails.

The failures that an instrument crosscheck is designed to detect cannot be demonstrated in flight. Simulators and computer training devices offer about the only opportunity to realistically train for gradual and/or unexpected instrument failures. Puckering liability issues dictate against installing a valve that can block the vacuum lines to simulate vacuum failure, and usually there are no switches to surreptitiously flick to disable an electric instrument. Yet the importance of mastering the transition is apparent in several studies that have shown that 1) it takes a significant amount of time, measured in minutes, for pilots just to recognize an instrument failure, and that 2) this is plenty time to get into real trouble. Coping with a failed instrument by using a partial-panel scan is an entirely different problem from recognizing the failure: the same pilots flew well enough in partial-panel mode when the instrument failure was known, suggesting that it is detection of the failure that is confusing, and that training for it is difficult, deficient, or both.

The flight instruments can be divided by whether they show roll, yaw, and/or pitch information, and theoretically you should crosscheck flight indications in all three axes. However, roll and yaw in flight almost always occur together and so they can be lumped for simplicity. In order to cross check roll/yaw indications, compare the attitude indicator/directional gyro (vacuum driven) with the turn coordinator (electrical). The imprecise magnetic compass can also be of some use in that if it is relatively stable, it indicates that the airplane is not turning even if one of the other instruments shows it is.

Cross check for pitch between 1) the attitude indicator (vacuum), 2) the altimeter, airspeed and vertical speed indicators (pitot-static instruments), and 3) power settings. Beware of the vertical speed indicator: it can wrap around far enough to give an erroneous and therefore confusing climb indication at high descent rates, and its indications lag significantly behind the actual conditions when pulling out of a steep dive.

One of the few places where the mechanics of an in-flight instrument crosscheck have been described was written a few months back by Michael Church in Private Pilot. Church suggests that turns, set up as a bank with the attitude indicator, should be backed up by checking to see that the turn coordinator agrees that the aircraft is actually turning, and in the desired direction. Simple enough so far: we do this anyway to establish the turn rate after establishing the bank. The less common flip side of the coin is that when the attitude indicator shows an unwanted turn, the turn coordinator should be checked to see that it corroborates the turn before trying to level the wings and coming back to a heading. Factor the heading indicator into these scenarios carefully since both it and the attitude indicator are vacuum-driven and could both be lying if the vacuum system is shot: crosscheck to the vacuum gauge. If there is disagreement between instruments, take the time to figure out which one(s) are lying before making drastic moves.
Likewise, if the attitude indicator shows an unwanted descent, double check with the airspeed indicator, vertical speed indicator, and/or altimeter before hauling back on the yoke. If you want to climb or descend, set it up with power and the attitude indicator and then make sure the airplane is doing what you’ve told it to do by crosschecking to the same instruments.

Usually an instructor waves a heavy paw across the panel and slaps a cover on an instrument to simulate failure: there might as well be a red flashing sign: “Go To Partial Panel, NOW!” It is much more difficult to detect the subtle and confusing indications of a real instrument failure, but it is imperative to do so because the chance to demonstrate dazzling partial-panel skill never occurs if the pilot does not first recognize the opportunity.