Aloft

by William Langewiesche

  Kukar was flying by the big artificial horizon in front of him, the so-called attitude director indicator, or ADI. The copilot, whose tasks included monitoring the captain’s flying, had his own independent artificial horizon. At the center of the instrument panel, easily visible to both pilots, was a smaller standby horizon, which also was independent. In addition, each pilot had a gyroscopic compass and turn indicator, along with the standard clusters of nongyroscopic instruments – airspeed, altimeter, and vertical speed, among others. Over history these instruments have been steadily consolidated in an effort to reduce the difficulty of the pilot’s visual scan – the ghostly ‘herding’ that kept Lindbergh busy during his flight across the Atlantic. Nonetheless, even in the most recent ‘glass’ cockpits with their crisp and minimalistic displays, the instruments have changed surprisingly little since those first developed by Sperry and Doolittle in the open-cockpit biplanes of the 1920s. For all its apparent complexity, Kukar’s 747 panel was largely just an answer to every biplane’s needs.

  What has changed is the way the instruments are used. Flight in the early years was an immediate and tactile experience, a short step away from the ground, and it required instrumentation only during passage through the clouds. Flight in a jet is a more distant experience and a condition more completely of the sky. Isolated not only by the quiet of the cockpit but by climb performance and high cruising altitudes, pilots are taught to ‘fly the numbers’ and to rely on their instruments for all stages of flight. The amateurish distinction between visual and instrument conditions, like the antique one between ‘flight weather’ and ‘flier’s weather,’ seems hardly to matter anymore. Airplanes shrug off most weather. And we have come to a time when the least curious (and usually least competent) pilots hardly bother to look outside at all except during takeoff and landing. Even so, the physics of flights have not changed. The bank still cannot be felt. When it also cannot be seen outside the airplane, even peripherally, it must still be treated with great care – measured and controlled on the inside with gyroscopes. Ocker and Crane are long forgotten. Working pilots now accept their teachings as their own. But when it comes to a showdown, some still fall apart.

  Kukar’s trouble started when he began the right turn, at about the same time he wished the departure controller a happy New Year. The airplane was still heading west, into the blackout conditions of the night sky. A moment later, according to the flight-data recorder, Kukar made an unnecessary and unusually large control wheel movement to the left. The airplane answered appropriately by rolling out of the right bank, passing through wings-level flight, and dipping into a left turn. Kukar must have thought he was still turning to the right, because he now steered the control wheel even farther to the left. The airplane’s left bank angle passed through thirty degrees, and the nose began to drop, flattening the climb. All this happened in three seconds.

  Kukar knew something was wrong. The cockpit-voice recorder picked up the surprise in his voice. He swore, ‘Arey Yar!’ and said, ‘My instrument!’

  His ‘instrument’ was of course Sperry’s device, the all-important artificial horizon.

  Kukar brought the control wheel back to neutral – a position to which the airplane would have responded, if only temporarily, by holding a steady bank angle. But Kukar was confused. Again he steered hard to the left.

  The copilot looked down at his own artificial horizon, and saw the indication of a steep and unexpected left bank. He said, ‘My… Mine’s…’

  Kukar said, ‘Mine’s just…’

  The copilot said, ‘Mine’s is also toppled!’

  Kukar said, ‘Check your instrument!’

  This conversation took four seconds, during which Kukar continued to swing the control erratically to the left. Though he did not know it, the airplane was banking past seventy degrees and accelerating through 300 miles per hour, and the nose was dropping swiftly below the unseeable horizon. It was the moment of maximum altitude, 1,462 feet, twelve seconds past ‘Happy New Year.’

  It is possible that Kukar remembered the bird strike of the day before, and he may have wondered if his controls had quit working. He kicked the rudder pedals once, hard, as if to check the response, although in an airplane the rudder is not normally used to get into a turn and cannot be used to get out of one. Kukar was flailing. Evidently he thought the airplane was still banked to the right, because he continued to steer and roll hard to the left.

  If people on the shore noticed a falling comet, the passengers aboard the airplane probably noticed nothing at all. They sat easily in their seats and felt none of the heaviness of a well-flown turn as the Boeing, dropping its nose in perfect 1.0 G synchronization, rolled past knife-edge flight, and began to turn upside down. If the passengers had looked outside and behind, they would have seen a lit-up Bombay turning silently on its side and then floating above them like some strange city in the sky. They might have noticed the lack of feeling where feeling should be. Just then they could have poured themselves drinks without spilling a drop. For all the good it did them to remain seated, they could have stood up and danced the length of the inverted aisles.

  That lack of feeling was of course precisely the problem for Kukar and his crew. The airplane was expressing a perfect spiral dive. Simulator studies at Boeing later showed that this moment, when it first rolled inverted to the left, was the pilots’ last chance for recovery. If they had rolled fast to the right to level the wings and had aggressively raised the nose, subjecting the airplane to 2.5 Gs, they could have pulled out of the dive about 100 feet over the ocean. But that would have required them to know which way to turn.

  In his blindness, Kukar did swing the control wheel hard to the right, and the airplane wobbled back to a ninety-degree bank.

  The copilot repeated, ‘Mine has also toppled!’

  The flight engineer may have leaned forward to point at the third, standby horizon. He said, ‘No, but go by this, Captain!’

  But Kukar was still confused. He swung the control wheel back to the left, and again the airplane rolled inverted. Seventeen seconds into the upset, the nose had dropped twenty-two degrees below the horizon. Though the airplane was still 1,100 feet above the water, its descent rate was shooting past 10,000 feet per minute, and before the end would reach three times that.

  Only four seconds remained. Kukar said, ‘Just check the instrument! Yar!’

  The copilot asked desperately, ‘Check what –’ and was answered by the impact.

  In the course of its spiral dive, Air India’s Flight 855 had turned from west to south. It hit the water doing 380 miles per hour, with its nose pitched thirty-five degrees down and its wings banked inverted, eighteen degrees past the vertical. Kukar remained confused to the end and died steering left.

  The Indian government convened a court of inquiry to establish an official cause for the accident. The hearings lasted several months. They took place in the heart of Bombay, in a sweltering courtroom whose latticework walls let in the dust and noise of the street outside. Pedestrians peered in through the openings. The inquiry was presided over by Justice M. N. Chandurkar, of Bombay’s High Court, who had no previous experience in this area. A clerk recorded the testimony by pounding on an antique manual typewriter loaded with carbons. Flight 855’s instruments had been shattered and dispersed by the impact, so the case was complicated by a lack of good physical evidence. Chandurkar, however, turned out to be an effective and intelligent questioner, unintimidated by the technical complexities of the case and unwilling to brook the double-talk of the interested parties – Air India, the pilots’ union, Boeing, and their various expert witnesses, along with a slew of American attorneys who had converged anxiously on Bombay to monitor the proceedings. Chandurkar sorted through the conflicting testimonies with a fairness and certainty that ultimately won over all but the most partisan of the observers. The final report, written by Chandurkar himself, still stands as a quiet and credible piece of work, the expression of an almost naïve belief i
n the possibility of truth – this in contrast to the cynical rewriting of the story that took place later.

  Chandurkar’s first conclusion was simply that Kukar’s artificial horizon had failed. During the hearing, expert witnesses had advanced three theories about such a failure. The first was that the instrument had frozen during the gentle right bank and had never moved again. The second was that the failure was ‘ratcheted,’ meaning that the instrument had responded normally to right rolls but had jammed whenever the airplane had rolled to the left. The third and most likely was that the instrument had not stuck at all but had failed by showing a steady and fictitious roll to the right, to which Kukar had slavishly responded by rolling to the left. Chandurkar did not try to decide among the three but reduced them to the essential: Kukar had rolled left and had died while trying to fly an indication that had gone bad.

  Which raised the question of the warnings. The best artificial horizons, such as those on the 747, constantly test themselves; and when they detect a problem they drop a red warning ‘flag’ across the face of the display. If such a flag had dropped, Kukar would certainly have seen it, and would have known to switch his focus immediately to the small standby horizon, or to hand over control of the airplane to the copilot. The airplane would hardly have wobbled. But as all pilots know, the flags themselves fail, and such failures, of safety systems generally, are especially dangerous because of the trust invested in them.

  It was never Chandurkar’s intention to absolve Kukar, and in the end he did not. At the same time, however, maybe because of his lack of previous involvement with the flying business, he was determined to describe the crisis in the cockpit not as others insisted it should have been but as he himself believed it was. After uncovering cases of flag failures in other 747s, he concluded that Kukar’s instrument had deceived him twice over.

  Chandurkar then went farther. The Air India cockpit was equipped with another safety device called a ‘comparator,’ which continuously compared the flight attitudes shown on the pilot and copilot artificial horizons and in the case of a disagreement between the two flashed a light on the master warning panel. Chandurkar agreed with the manufacturers that the comparator must have functioned correctly and that in the dimness of the cockpit the warning light must have been visible. But he pointed out that the warning had obviously been of no use to the pilots anyway, and he questioned the value of such a device in this particular case, when survival depended on the immediate interpretation of the panel’s main instruments.

  Chandurkar had glimpsed an especially modern aspect of human flight, and something that crews themselves need clearly to understand – that the cockpit’s automated warnings, horns, and flashing lights provide largely just the appearance of safety and that for a variety of practical reasons no amount of automation can yet relieve pilots of the old-fashioned need to concentrate and think clearly in times of trouble.

  Having examined the failures of the instruments, Chandurkar turned his attention to those of Kukar, who may once have been a good pilot but who at the time of the accident clearly was not.

  ‘Arey Yar! My instrument!’

  There it was on tape. Kukar had suspected a failure of his artificial horizon in plenty of time to keep the airplane under control if only he had been able to convince himself of it. If this lack of conviction seems hard to understand, remember that in the complete blackness of that night, a toppling artificial horizon line would have looked level to Kukar as he, in equal reaction, toppled the airplane to follow it. As an experienced airman, he seems to have noticed a bewildering discrepancy between his hard left steering and the lack of normal response on the face of his instrument. ‘Arey Yar!’ Such a disconnection would have given him an intuitive perception of trouble, perhaps of control failure, and would explain his single exploratory kick of the rudder. A momentary confusion was inevitable. The sadness is that he sustained his confusion when directly in front of him a host of secondary instruments in a little stampeding herd clearly showed that the airplane was turning to the left, diving, and gaining speed. Kukar ignored them all. He had fixated on his artificial horizon, and in the urgency of the moment he could not summon the discipline to look away from it. His visual incapacitation is the most frustrating part of the story. He flew as if shackled to a single indication of the turn.

  As the captain of the flight, Kukar was to blame for his copilot’s errors as well. Pilots communicate with each other in unspoken ways, by the way they slouch in their seats, or throw the overhead switches, or handle the controls of the airplane in flight. The airlines fight back with standardized procedures and try to encourage enlightened regimes of teamwork and safe behavior. Nonetheless, the cockpit is like a club, and in the privacy of flight the sloppiness of a senior pilot will encourage the same attitude in his subordinates.

  Kukar’s copilot must have noticed what the airplane’s ‘black boxes’ later indicated – that Kukar made sloppy and unnecessary control motions on the runway, that he lifted off the pavement late and fast, that he rushed the flap retractions, that he climbed at too flat an angle, and that he started the right turn 500 feet below the altitude called for by the departure procedure. Less experienced pilots would not have made such mistakes. Kukar had proved capable of greater precision during his periodic flight checks, and he would no doubt have disapproved of such flying among his juniors, but he may also have felt that his improvisations were the prerogatives of a high-time pilot. And in the immediate sense he was right, too: Nothing he did during that takeoff was unsafe. But on the basis of his subsequent performance, his looseness now seems to indicate that he had grown careless. And to make matters worse, his copilot clearly did not think less of him for it.

  Relaxing in the right seat, the copilot did not even bother to monitor the flight. When he heard Kukar’s ‘My instrument!’ he looked down at his own artificial horizon, expecting to see it showing the standard right turn, and was shocked to find it showing something quite the opposite – a nearly vertical bank to the left. Feeling nothing of that turn, checking none of the other instruments, he declared like some early airmail pilot that his indication was in error.

  His stuttering ‘My… Mine’s… Mine’s is also toppled!’ was not what Kukar needed to hear, and it marked the actual point – no matter what the simulator later showed – from which no recovery was possible.

  The confusion was now absolute. Kukar said, ‘Check your instrument!’ But against what? The circumstance that causes the spiral is the very circumstance that prevents its solution – in this case the collapse not just of one gyroscopic instrument but of the two pilots who refused to cross-check all the others.

  From his position in the back of the cockpit, the flight engineer had the largest view of the instrument panel and therefore in this case the most accurate. The last seconds must have been frightening to him. The passengers and flight attendants knew nothing of the dive and had their lives extinguished so mercifully that they will in some sense forever be flying to Dubai. The pilots knew something, but not what, and they were fully occupied with their confusion. Only the flight engineer clearly saw the errors being made. It was not a nightmare, though that possibility must have crossed his mind. After so many years during which these things had happened only to others, he was the one now actually going down, out of control, crashing. There was no reason for this, and it didn’t feel wild, but the instruments told an undeniable story. And yet he could do nothing but point.

  ‘No, but go by this, Captain!’

  Kukar did not even answer him. It is impossible to know what denials he was engaged in. Pilots train in simulators to handle all sorts of failures, sometimes heaped one on top of the other. But even the best of the simulators require a suspension of belief that never quite overcomes the understanding that they are pretend airplanes built for the purpose of experiencing failures and that no matter how poorly the pilots perform they will walk away unscathed. That, of course, is not true of failures inside the sky, where the firs
t challenge is to suspend disbelief and where the urgency is real.

  In any case the flight engineer’s advice came too late. I refuse to turn away from the thought that the airplane’s lights illuminated the ocean’s surface at the last instant, that the surface appeared to surge at the airplane from somewhere above, and that the flight engineer flinched as the water exploded through the cockpit. It does not help to be polite about these details. The tangible consequence of any serious failure in flight can be just such an unstoppable insider’s view.

  Chandurkar was appropriately severe in his final judgment. He placed responsibility for the 213 deaths squarely upon Kukar and blamed the tragedy entirely on his inability, in a cockpit full of functioning instruments, to handle the failure of just one gyroscope. He wasted no emotion blaming the instrument itself, since it was obvious to him that even the best equipment can fail, which is one reason why pilots are needed. Instead, he recommended changes to Air India’s recurrent training program and a renewed emphasis not only on the basics of instrument flying but also on the principles of communication within the cockpit. It all made perfect sense.

  Nonetheless, Boeing was unhappy with Chandurkar’s work – and for good reason. Armed with the official finding of an instrument failure, the families of the victims hired a New York law firm to bring suit in the United States against Boeing and the instrument manufacturers. The plaintiffs contended not only that the artificial horizon had caused the accident but that the comparator had failed as well, and furthermore that the designers of both devices had been negligent.