by Orr Kelly
As the engineers should have realized, stresses migrate. With the base of the tails beefed up with the cleats, the stresses went looking for somewhere else to show themselves. Soon, there were alarming reports from the fleet of new cracks in other parts of the tail. Mechanics on the flight lines speculated that the stresses from the tail were even migrating far enough to cause cracks in the engine mounts and the landing gear.
“I personally feared that we were going to have to make large portions of the vertical tail and aft fuselage from titanium,” Capellupo says, recalling his reaction when the reports of new problems with the tails began coming in. “That would have been a terrible thing. We were only halfway to where we had to be when we learned that the tail would just not live in that environment. Those were very, very tense times.”
The cleats obviously were not the solution they had seemed in the fall of 1985. But they had bought time. With them, the plane could continue to be flown safely and deliveries could continue while efforts were made to find a permanent solution. There was no need to ground the planes, with all the unfavorable publicity that comes with such a move.
Instead, this second act of the drama of the cracking tails attracted little outside attention, even in the trade press. McDonnell Douglas waited until 11 November 1987, and then quietly issued a press release announcing it was developing modifications to the plane “to sharply reduce stress on its aft sections and thereby increase its service life.” The press release was issued on the deadline for the company, as required by law, to file a Form 10-Q with the Securities and Exchange Commission, revealing that it expected fixes in the tail to result in a “material charge” to its earnings. Later, it reported that charge had amounted to $13.6 million in a single quarter of 1987.
Neither the press release nor the statement to the SEC came close to reflecting the anguish suffered by McDonnell Douglas executives when they were confronted with the fact that the problem with the vertical tails had not been fixed after all. Capellupo worried that the company might have to embark on a massive effort to remanufacture the planes already delivered. McDonnell Douglas, the biggest manufacturer of combat aircraft in the free world, was strong enough that it would not have been forced into bankruptcy. But there was no question that rebuilding several hundred planes already delivered would have, as Capellupo put it, “done major surgery to our earnings.”
By the time the press release was issued, McDonnell Douglas was well along in the development of what it hoped would be a lasting fix to the problem of tail cracks. This time, the effort focused on changing the forces at work on the vertical tails, rather than trying to beef up the tails themselves. Using wind tunnels and computers, the McDonnell Douglas engineers experimented with as many as fifty different shapes of fences and vortex generators to change the flow of air over the wings so it would not beat the tails to death. Since all these gizmos would be attached on the forward part of the plane, the work was done at St. Louis.
The Northrop engineers watched all this with concern. They wanted to be sure that nothing done up front had an adverse impact on the performance of the tail on their section of the plane. It was the vortex, after all, that helped give the plane its superb performance at high angles of attack. The trick was to make the vortex more benign, not to weaken or deflect it away from the tails.
The solution finally settled on was what is called the “LEX fence.” It is a narrow chunk of steel thirty-two inches long and a little over eight inches high, attached to the wing where it narrows and extends forward along the fuselage. The fences on both sides of the plane are interchangeable. Together with the internal structure to which they are bolted, they add fifty-two pounds to the plane’s weight.
If the fences worked as the designers hoped, they would smooth the flow of air and spread out the vortex to reduce its impact on the vertical tails while still providing enough turbulent air for the tails to do their job. But there was real fear that the fences might cause other serious problems. Wind tunnel tests hinted that the fences might make the horizontal tail—the stabilator—less effective. That would make it harder to bring the plane aboard an aircraft carrier. There was also concern that the fences might make the plane more likely to fall off into a spin.
The earlier solution involving the cleats had been approved after only half a dozen test flights. But this time, a much more elaborate test program was set up, both to assure that the fences would do what they were supposed to do and to assure that they would not cause some other problem.
The tests involved forty-five flights checking the plane’s performance at 3,000 combinations of speed, angle of attack, and altitude. F. Alan (“Al”) Frazier, McDonnell Douglas’s chief experimental test pilot, flew some seventy-five percent of the test flights in a plane loaded with stress gauges. Each day, carrying a card listing the points to be checked, he would climb to altitude, pull the nose back, and snap the stick sharply to one side in two-tenths of a second. The result was a seven-G snap-roll. Then he would change the speed or angle of attack slightly and do it again—over and over and over.
Since the fences are held on with bolts, it was possible to go through the same tests with the same plane, with and without the fences.
When the tests were complete, it was clear that the fences reduced the load on the vertical tails and extended their life, actually improved the plane’s performance at very high angles of attack, and did not make it more difficult to come aboard a carrier. The only serious adverse effect was to increase the vibration of the LEX itself. But even so, it was calculated, the LEX would last for twice the design goal of 6,000 flying hours.
The test program was conducted with a good deal of urgency. Every plane that left the factory without being fixed would have to be modified later at company expense. Even so, the flight tests continued from October 1987 into April 1988. On 20 May 1988, the navy gave its approval to the modification, and planes in service were quickly fitted out with fences.
In addition to developing and installing the fences, McDonnell Douglas also took on a major job of checking hundreds of planes already in service and repairing any damage caused by stresses on the tail.
How could such a terrible defect slip past some of the best engineers in the business?
Donald Snyder, vice president for aircraft engineering at McDonnell Douglas and director of engineering on the F/A-18 from 1982 to 1985, says: “Our management questioned if we should have found it. No one has come forward and said, ‘you could have found it.’ I was not there, in charge of engineering. I have no reason to be defensive. I’m the fellow who came in and had to fix it. It would have been easy to blame the previous administration, but I don’t think that’s fair.”
Strohsahl, the program manager during this difficult period, puts the primary blame on Northrop. “They goofed on it,” he says.’ ‘They clearly did. The navy doesn’t hold Northrop responsible. But in all fairness, we really know where it came from.”
A big part of the problem can probably be traced back to the sudden switch from a technology demonstration to a fly-off between the YF-16 and the YF-17, and the navy’s subsequent attempt to convert a lightweight fighter into a heavier strike-fighter. Northrop, in the early development of the YF-17, focused heavily on the development of new technology, and then demonstrated it in a prototype designed to fly for only a thousand hours or so.
As Tom Burger, the Northrop manager for the F/A-18 program, explains it, the company was struggling to find the best combination of the size and shape of the leading edge extension and the placement and angle of the vertical tails to give the plane superior performance at high angles of attack and low speeds—all without beating the tails to death.
Could the Northrop engineers have done a better job with the tail design? Burger shrugs: “It’s hard to say. I don’t know.”
Throughout this entire episode, the Northrop engineers remained possessive about the tail. They relied on McDonnell Douglas to tell them what stresses the tail must withstand. Bu
t they insisted on doing the engineering on the tail themselves.
When McDonnell Douglas became the lead contractor on the Hornet, however, it had assumed responsibility for the performance of the entire plane, and it did a good deal of design and test work on the vital tail structure.
In addition to thousands of hours of wind tunnel tests, the engineers also had the advantage of sophisticated new computer programs that impose a fine-mesh grid over the entire plane and then analyze the stresses in each square of the grid. Finally, actual airplanes were built and then sent to a torture chamber, where they were dropped repeatedly to test the landing gear and where the wings and tails were shaken and twisted, so they would go through all the strains of flight and expose any problems long before a pilot encountered them in the air—or a mechanic saw a crack when he opened an inspection panel.
But all of this clearly wasn’t enough.
One reason may be that McDonnell Douglas relied too heavily on the fact that the YF-17 had flown successfully, even though the F/A-18 was in many ways a new airplane. But Len Impellizzeri, a vice president of the company who was responsible for structural work during the design phase, says he doesn’t think the McDonnell Douglas engineers were misled by the work already done on the YF-17. “None of us was clever enough to see the problem,” he says.
In using the powerful vortex created by the LEX to help the vertical tails control the plane, the engineers were stepping off into new territory. In hindsight, Impellizzeri says they should have done a better job of testing. “But,” he says, “we just didn’t know enough to do the right kind of testing.”
Even now, the question remains: Have the fences really fixed the Hornet’s tail problems for good? The engineers think so, but they admit they are working at the outer edges where aeronautical engineering is as much art as science. It is still an area where the fleet sometimes learns of problems before the engineers.
CHAPTER EIGHT
“A Deep-Seated Drive to Kill”
As a pilot banks his Hornet and turns toward the carrier, ten miles ahead, for a nighttime landing, everything seems to happen in slow motion. The carrier comes into view as a postage stamp-sized apparition, floating in the darkness. A single line of strobe lights marks the center line of the deck. A vertical line of lights—the drop light—marks the stern. There is a faint glow off to the right from the sodium vapor lights around the island.
Just to the left, a round amber light—the meatball—is centered between two green lines if the pilot is on the proper guide slope.
Near the stern of the ship, two landing signals officers huddle behind a barrier that shields them from the wind and the jet blast of landing planes. Far astern, a single red light atop a destroyer provides an artificial horizon for the landing signals officers. As the incoming pilot approaches the ship, he announces his plane number and how much fuel he has left and says, “ball,” to indicate he sees the meatball. The LSO responds: “Roger, ball,” giving the pilot permission to land.
If the approach is normal, those will be the only words spoken.
The pilot concentrates on three things: his airspeed, the center line of the deck, and the meatball. On a dark night, there is no horizon, and there are no other points of reference except those disembodied lights hanging out there in the darkness.
In the Hornet, he can set his throttles, much as a motorist sets the cruise control in an automobile, and the engines automatically maintain his speed as he comes toward the carrier.
If he flies so the lights marking the center of the deck form a straight line with the vertical lights on the stern, he is properly aligned for his landing.
Most crucial of all is the meatball, which tells if he is on the proper guide slope. If he comes in too high, he will miss the arresting wires and have to go around again. If he comes in too low and does not add power in time, he risks crashing into the stern of the ship.
Inside the mechanism that holds the meatball, gyroscopes automatically compensate for the movement of the ship, sending the beam of light through a series of mirrors similar to the lenses originally designed to boost the power of a lighthouse beacon.
On either side of the amber light are green stripes. In an ideal approach, the meatball is centered between the stripes. If the plane is too high, the meatball appears to rise above the stripes. If too low, it drops below them. If the approach is dangerously low, the meatball turns an angry red.
As the pilot comes closer to the carrier, it is suddenly obvious that he is pursuing a moving target. Not only is the carrier sailing away from him into the wind at some thirty-five miles an hour, but, since he is headed for an angled deck, his landing area is moving away from him at an angle. This means he must not only fly in a curve to compensate for the movement of his landing place, but he must also compensate for a slight crosswind caused by the fact that the deck is set at an angle to the movement of the ship.
The space for which he is headed is incredibly tiny. The runway of a military jet airfield may well be more than two miles long. The landing space on a carrier such as the U.S.S. Coral Sea, one of the first to take the Hornet to sea, is only 120 feet long. For a plane crossing the stern at 200 feet a second, that is only a fraction of a second’s worth of space in which to set down. Width of the landing area is also critical. Other planes are parked on both sides of the deck, leaving a strip only eighty feet wide. This means that if a Hornet, with a wingspan of forty-three feet, including Sidewinders, is more than a few feet to one side or the other of the center line, it is probably going to break something.
Normally, the landing signals officers set the meatball so the plane will catch the third of four cables stretched across the deck. For a perfect landing, the pilot must maneuver his plane so precisely that, when he crosses the stern, his tail hook is fourteen and a half feet above the deck, and his head passes through an imaginary three-foot square.
In those last few moments, what had been an almost leisurely process abruptly changes. Everything goes into fast motion. The shape of the carrier suddenly looms out of the darkness, lights flash past, and the landing gear crunches down on the deck. If the pilot has been focusing almost all his attention on the meatball, as he should, the actual touch down will come as an abrupt surprise.
The natural instinct, at that moment, is to cut off the engines, slam on the brakes and stop. But what if the tail hook has not caught the wire? Following one’s instinct would send the plane careering down the deck and off into the sea. Instead of trying to stop, the pilot does just the opposite: he tries to fly. He slams the throttles into full military power and steels himself to go hurtling off into the darkness to make another landing attempt. If he has caught the cable, the sudden burst of power from the engines will force him backward, and then he will be thrown forward hard against his restraining harness as the plane comes to an abrupt stop.
Quickly, as the deck crew checks under the plane, he releases the cable, folds his wings, and moves forward to make room for the next plane, already heading in for a landing exactly sixty seconds later. Then follows an experience that many pilots find even more stressful than the landing itself. Out in front, a teen-aged member of the deck crew waves the pilot forward with a lighted wand. As he taxis over the catapult toward his parking place at the bow of the ship, he feels his wheels slip on the oily deck.
Onward the baton beckons him, into the darkness, part of the process of making the most use of every square foot of some of the most intensively utilized real estate on earth. At the edge of the abyss, the pilot pivots the plane and parks with the tail hanging out over the ocean. Even pilots with hundreds of landings to their credit make those last few movements with one hand on the ejection handle. They know that, if the plane slips over the side, the only realistic chance of survival is to eject and hope to be rescued.
Afterward, down in the squadron ready room, it may take half an hour or more before the adrenalin stops pumping. Studies conducted during the Vietnam War confirmed that a night carrier
landing is the most stressful part of military flying—even more than the tensest moments of combat itself.
Daylight landings, when the pilot can usually see the horizon, the carrier, and other ships and planes, are demanding, but not nearly as stressful as coming aboard in the dark of night. In the daylight, instead of making a long, straight run at the ship, pilots fly parallel to the deck, then turn toward the ship while half to three-quarters of a mile astern, and then make their approach. One trick that Hornet pilots have adopted is to put the velocity vector—that little symbol in the heads up display that shows where the plane is actually going—just ahead and to the right of the bow of the ship. That helps to compensate for the way the angled deck is moving away to the right. Landing signals officers say they can spot the pilots who move the velocity vector to the center line of the deck and use it—instead of the meatball—to land, because they usually trap on the second rather than the third wire.
With experience, daylight landings become fairly routine, even fun. But a pilot’s feelings during a nighttime landing are never far from the edge of sheer terror. Even those with hundreds of landings to their credit think about dying.
Much of what the navy does to prepare a new plane to go to sea is concentrated on that one task: landing on a carrier. The plane itself has been beefed up to withstand the impact. Engineers have worried over the responsiveness of the engines and the way the plane handles as it approaches the deck. How well a plane comes aboard the carrier is not merely a matter of safety. A squadron of planes that can land predictably on the first pass needs less fuel to go around again, and they require the carrier to keep its nose into the wind for a shorter time.
By these measures, the Hornet comes aboard well. Its relatively short wingspan and the maneuverability built into it as a fighter plane make it much easier to handle in those last moments before touchdown than the F-14, with its sixty-two-foot wingspan. And its very responsive engines make it much easier for the Hornet to maintain the proper airspeed than the sluggish response of the single-engined A-7.