they were confused
Ross Detwiler has an article in the October copy of B&CA titled "Dome of Confusion". Excerpt.
A discussion of handling an airplane in an emergency has to center around maintaining control while fighting to get out from under what I term the "Dome of Confusion".
One expert quoted in "High Altitude Upset Recovery" referenced using pitch and power as the right method of maintaining control during instrument malfunctions. The Air Force drilled into us where to look to ensure we were using those tools. We had steps to start working out of the confusion when the probability of what just occurred went to “one."
The instruments on every airplane panel are divided among three purposes: (1) control via the ADI and tachs; (2) performance, with altimeter, airspeed and heading indicators and VVI; and (3) navigation, using the HSI, RMI, GPS, VOR, ILS and ADF. This is an excellent concept as it prioritizes where to look.
When you are trying to maintain or regain control, you go to the two control instruments. Even though some, including me, fault certain auto-throttle systems for their lack of tactile feedback, the tachs still report what the engines are doing.
Looking at the attitude indicator and the tachs will give you a basis for either maintaining or regaining your control given that you have not yet stalled the airplane. I had an instructor once tell me in the F-4 Phantom that if you depart controlled flight (stall), you use the g meter to fly. “Keep it at zero and hope the rest of the gauges eventually come back to where they belong. If not, step smartly over the side at 10,000 ft.” Unfortunately, we don’t have the "step over the side” option in our world, so our job is to avoid departing controlled flight.
Our modern airplanes have the ability to greatly increase confusion when we start working on the back side of the information power curve. We can be presented with so much "performance” and “navigation” data, that "control" information may not be processed. As with engine thrust being ineffective when behind the power curve, the amount of data coming at the pilot in a high-stress situation is so massive that it can overwhelm the ability to process and therefore result in less and less output.
Here’s a personal example ofthe "Dome” at work. On a recent recurrency simulator session at CAE, instructor Bob Hare told us, "I’m going to give you an ADS [Air Data System] problem that you’ll have to solve. Pm just going to try and put you into a confused state in which I Want you to handle a simple problem. I’m going to give it to you on downwind, in night VFR conditions."
Of course we rolled our eyes at Hare’s unwillingness to let even a simple VFR landing sim go by without missing the opportunity to embarrass us, but hey, "Bring it on." Embarrassments at his hand have always resulted in knowledge gained. Besides, what could he do to our ADSes that we couldn’t handle in VFR conditions, and after he told us about it, no less?
What he did was introduce an "ADS 3 mis-compare" fail light in the panel. Our checklist said basically that if left and right ADS presentations were valid, to ignore the light. Swallowing the bait in one gulp, we unwittingly replaced the word "valid” with "the same" and continued on. We immediately assumed that something was wrong with the ADS 3 system since both the pilot’s and copilot’s presentations were identical. And for that reason, we also didn’t lend proper weight to what the standby system, indicator Number 4, was doing. It was accelerating. But here again, we didn’t give that information the processing it needed because the two "main" indications agreed we were OK.
That anchor of reasonableness – that the two big screens were good — led us down the path and caused us to write off the clue coming from the standby system. In our minds, Number 3 was bad, because Numbers 1 and 2 agreed, and the warning light on the panel had a "3" in it. Although Number 4 had a problem, too, we ignored it and continued on downwind. Then, after a few seconds, we noticed that there seemed to be an awful lot of noise for the airspeed we were showing.
At that point, Hare mentioned something about the flaps (he never rests) and we considered moving them in case our speed was actually slower than we thought. Finally we began, through basic pitch and power, to fight through the confusion while maintaining control. We were in level flight on the ADI, not nose high as we would have been had flaps been needed, and the power was more than sufficient to maintain flight.
I’m proud to say that at least we spit out the flap hook. Finally, amidst the what’s- going-on-here? confusion, we looked to the Number 3 ADS and saw that it was reading about 100 kt. higher than the indications on the panels in front of us. Number 3 was high, but Numbers l and 2 were low and within a knot of each other. The “decider" finally became the standby instrument panel (Number 4), which also showed us to be at a very high airspeed —- an airspeed that agreed with our attitude, power settings and Number 3.
In the Falcon 7X there are four ADS systems. Number 1 for the pilot, Number 2 for the copilot and Number 3 as a monitor/comparator. Number 3 can be switched to the Number 1 or 2 position or will automatically go to Number 4 (standby) if needed. When we got the Number 3 mis-compare our assumption was Number 3 was bad, but that was unimportant since we weren’t using it anyway. (Something tells me Hare has seen crews make this assumption before.)
What it was actually telling us was that Number 3 "disagreed" with the primaries. It was only as we continued, in confusion, wondering why things weren’t right and so noisy with normal power and attitude, that we caught it. The point of the demo was to show how the introduction of confusion can make a simple problem much more difficult to solve.
His comment on 447
The challenges the crew of Air France 447 must have been facing on the night of June 1, 2009, over the South Atlantic are known. Sensory systems appear to have iced up, negating their input. From what I’ve read, the airplane was held in a high angle of attack (AOA) position due to back stick pressure being exerted. With a frozen pitot static system, indicated altitude does not change and indicated airspeed increases with altitude. So they may have initially thought they were going fast and maintaining altitude. If true, that might explain why they continued raising the nose.
Did they remain in "controlled” flight all the to the water because the FBW wouldn’t let the airplane exceed the stall AOA? I’m not recommending stalling and spinning down through a thunderstorm when you ice up, but a pre-stall shudder and roll off would have at least given them the clue that they were going down. They may never have had that clue.
What other confusions could be added to this situation? Were they wondering if the airplane’s FBW system was functioning correctly? Was it causing them to pitch up because it, not them, thought they were going too fast? Did FBW add to the confusion by making inputs of its own? When FBW systems ice up, what do the computers think is happening and how will that affect the possibility of “putting the nose where you want it”? I’m not saying that’s what happened because I don’t know and no one else has pointed to that fact, but it could sure add to the confusion.