Traffic was light at two in the afternoon on the Van Wyck Expressway approaching John F. Kennedy International Airport. I was headed to Pan Am’s flight dispatch center in Hangar 14, grateful for the quick ride in from Jersey because we had a long day – and night – ahead. The year was 1989, and I was picking up dispatch papers for a 4pm 727 ferry flight to Frankfurt with a fuel stop in Keflavik. In Frankfurt, the plane would be configured for Pan Am’s Berlin-based Inter-German Service or IGS as it was commonly known. The IGS was spawned in the Cold War after World War II in 1945 when Germany was split down the middle by occupying powers.
East Germany was a ward of the Soviet Union except for Berlin in the heart of East Germany, which was itself split between Russian control in the East sector and free West Berlin under the tripartite control of the U.S., Britain and France. Russians limited air access to West Berlin via three 20-mile wide by 10,000-ft high air corridors slicing through East Germany. Pan Am acquired the U.S. franchise in 1950, serving the city initially with DC-4s then DC-6s and later, 727s and other jet airliners.
At JFK, flights normally departed from Pan Am’s Worldport terminal about a mile from Hangar 14 but our plane, used mainly in U.S. domestic service, had to be specially configured in the hangar for the Atlantic crossing with an inertial system and a life raft and was conveniently parked outside on the hangar ramp.
At the Hangar 14 operations center, an oceanic route folder had been prepared to supplement our domestic charts. In it was an Atlantic plotting chart, Atlantic and European enroute charts, and alternate airports for the Canadian Maritimes, Greenland, Iceland and Western Europe. Next to it, another thick folder contained the dispatch release, notams and a flight log printout of the route: zone times, fuel burn, forecast spot winds, step climb points, and a blizzard of other data with blank spaces for ETAs, ATAs, spot winds, nav and fuel scores in a format that could be easily read off when transmitting position reports.
Also neatly folded in with destination and alternate forecasts were upper air met charts for the 300, 250 and 200 millibar levels equating roughly to flight levels 300, 340 and 390 and the high level FL 250 to 400 significant weather chart. Wind charts were rarely referred to in flight because the company’s flight planning computer integrated wind forecasts to determine least time track and cruising level. All that optimizing computational power was largely wasted on our flight, which was constrained to the so called Blue Spruce Route, an oceanic track for navigation or communication limited aircraft. Our 727 with its single INS, two VHF nav/coms and two ADFs was limited on both counts.
In contrast, even back then, before the age of satellite communications and GPS, airliners that regularly plied the oceans had multiple inertial long range navigation systems and HF radios and flew variable Atlantic tracks revised twice daily for optimum tailwinds and minimum headwinds. Blue Spruce routes separate navigation and communication-limited planes like our 727 from those better able to stay on course and communicate should one or more nav or communication systems fail.
As I began sorting through the folders, Bill, the first officer, showed up. He was a no-nonsense former Marine pilot and a solid aviator who had a methodical by-the-book manner about everything he did and would have been a captain long ago were it not for the foibles of airline seniority. In short, a good first officer to have on an all-night, two leg Atlantic crossing. At least that’s how the mission was planned although it wouldn’t turn out that way. Together we reviewed the paperwork, determined V1, VR, and V2 takeoff speeds and EPR thrust settings and noted them on the company’s takeoff computations form. Then, with a plotter, Bill set about lining in our oceanic track on the plotting chart. Our route approximated the great circle from Goose Bay to Keflavik, crossing compulsory reporting points at the Canadian FIR boundary, 50W, Prins Christian Sund NDB on Greenland’s southern tip, 40W, 30W, then direct Keflavik.
To identify waypoint entry mistakes before they became gross navigation blunders, Pan Am required crews to record latitude and longitude on the flight log shortly before crossing oceanic waypoints, then plot position about 10 minutes (approximately 80 nautical miles) after crossing the waypoint. If the plot diverged from the chart’s track line, the crew would know in sufficient time to correct back before they were grossly off course. Later, as events unfolded, those meticulous plotting procedures would be of little use and the less regarded met charts would command much greater interest.
At the end of the flight, the completed flight log and plotting chart, was inserted in a special “navigation envelope” with the flight number and date and co-mailed to JFK Flight Operations headquarters and retained for several months in case there was a question about an off track deviation or some other official inquiry. The remaining flight documents like met charts, weight and balance forms and fuel sheets along with the flotsam of ocean crossings like coffee cups, plastic forks and knives, napkins and food wrappers, ended up in a Pan Am logoed wax paper lined trash bag that crews always conveniently placed behind the pedestal.
Bill and I lugged our flight kits and bags up the aft stairs and through the cabin into the cockpit. On the way, we passed the inertial navigation system (INS) bolted on a pallet and secured in floor tracks where the left front row of first class seats had been, and forward of that, our over-water survival gear: an inflatable life raft strapped at the forward cabin door. The galley was provisioned and the flight engineer was at his station finishing up pre-departure duties.
A thick wire bundle taped to the cabin floor ran from the INS to the cockpit pedestal where the inertial mode selector unit (MSU) and control display unit (CDU) were installed. Since the inertial system was a temporary installation for the Atlantic crossing, it wasn’t wired into the autopilot, so heading changes had to be manually applied via the RMI heading bug or the autopilot’s pedestal heading selector knob. It was an awkward arrangement and more labor intensive compared to Pan Am 747s, L1011s, DC10s and A310s where INSs were coupled to the autopilot in navigation mode. On the other hand, the jury-rigged arrangement kept us more actively involved in the flight’s progress. In order to stay on track, our eyes didn’t stray far from the inertial’s cross track (XTK) display.
We lifted off around 4pm with the late November sun well on its way to the southwest horizon. Cruising at a mid-30s flight level, the sky was dark in the east and fading twilight marked the southwest horizon. New England towns had been dark for some time, their presence marked only by irregular light patches in a sea of black. My job now whittled down to fine tuning thrust, rudder trim, and periodic RMI heading bug adjustments to keep the inertial cross track readout close to zero on the path to Goose Bay.
The flight engineer headed to the galley for eats and, in accordance with common practice, summarized his panel’s settings: pressurization-automatic, center tank override boost pumps on, crossfeed valves open and feeding all engines. I nodded and he disappeared through the open cockpit door. Goose Bay was about an hour ahead.
Glancing down at the INS CDU, I noticed the cross track read-out flick to one mile. We were in stable air with no turbulence to signal we had perhaps penetrated a jet stream which could account for such a sudden drift off course. Seconds later, the display flicked to two miles then 5, 6, 7 and kept climbing. Believing such meteorologically impossible indications would mean we were flying faster sideways than forward. But what that esoteric display obviously indicated was a failed INS. And to emphasize the point, as the cross track numbers climbed ever higher, the CDU’s red warning light illuminated, indicating self-test circuits had detected a malfunction.
With no INS to keep us on track we couldn’t fly the Blue Spruce route at jet cruising levels and didn’t have fuel for the more northerly mid-Greenland route defined by NDBs. Diversion to Goose Bay Labrador, a joint military/civil airfield, was the best alternative. It was constructed early in World War II as a refueling stop for transports and bombers, and later expanded and updated to base military aircraft in sync with Cold War imperatives. Its two runways, the longest slightly over 11 thousand feet, and all-weather navaids made it a handy contingency field on the North Atlantic’s western shore.
At the Goose Bay operations office a phone chat with Pan Am operations yielded two options: routing north over Baffin Island thence to central Greenland following a line of NDBs or, if we could get the INS aligned and operating without that pesky red malfunction light illuminating, taking the original Blue Spruce Route to Prince Christen Sund NDB on Greenland’s southern tip, then straight to Keflavik. I liked the latter plan and with full tanks equating to five plus hours’ endurance, we had fuel for any contingency. I sent Bill out to align our recalcitrant INS and report back. He returned smiling: the INS was humming happily with no warning light.
We informed Pan Am operations who telexed the flight plan and paperwork, putting us back on the original Blue Spruce route. With waypoints loaded and clearance received, we fired up and in short order were off and climbing into the North Atlantic night. It was Bill’s leg to fly and I worked the radios and provided support.
About 10 minutes before 50W, to no one’s great surprise, the INS CDU’s red warning light illuminated. The balky INS had worked long enough to get us underway on a straight track to KEF. Now it was up to us to estimate drift and groundspeed using tools at hand. Bill unsheathed his six-inch Jeppesen CR-3 computer, commonly known as the “wheel,” and I fished out the met charts we brought on board at JFK.
Two decades earlier, crews routinely used CR-3s on 707 Atlantic and Pacific crossings to compute wind, integrating true airspeed with Doppler drift and groundspeed displays, correcting for variation (about 30 degrees west where we were at the moment) and transmitting the info with position reports at each 10 degree meridian. By the late 1980s, when cockpits displayed virtually everything pilots needed to know including wind speed and direction, the trusty six- inch Jeppesen wheel was regarded more like a flight kit vestigial. But it was still required in Pan Am flight kits for contingencies such as this night. Now, based on the INS drift and groundspeed before the failure and the met charts, Bill determined we had about a 40-knot left crosswind component and five degrees of right drift.
At our estimated time of arrival at 58N 50W, I transmitted a “position doubtful” report and informed Gander that our single INS had failed. Gander’s acknowledgment included a descent clearance below the minimum navigation performance specification (MNPS) airspace (FL 285 to FL 420). North Atlantic protocols prohibited aircraft without a long range navigation system in MNPS airspace.
Had the first INS failure occurred over the North Atlantic, a mandatory descent nearly 10 thousand feet below our optimum mid-30s cruising level would have presented an endurance and fuel reserve problem. But having topped off to 50,000 plus pounds of Jet A at Goose Bay, endurance wasn’t an issue.
About 30 minutes from Greenland’s Prins Christian Sund NDB, its OZN identifier was audible and the ADF needle oscillated to life. To be sure it was pointing accurately, I switched the receiver to beat frequency oscillator (BFO) and manually slewed the RMI pointer back and forth through the audio null. The bearings matched. And our fudgy drift estimate, derived partly from Bill’s “wheel” and 300 MB chart, put us nearly on the published Blue Spruce course to OZN.
To me, overflying Greenland on a clear day was always visually captivating. Its jagged peaks and glaciers calving into the North Atlantic were a scene like nowhere else in the world. Sometimes at night at those latitudes, the aurora’s multi-colored curtains of light were spellbinding. But this night the sun’s charged particles were slackers, imparting only a dim, white glow. And, as we approached, southern Greenland was just a dim jagged triangle on the radar, sliding closer with each sweep. Tracking outbound from OZN NDB and intercepting the inbound radial to KEF VOR was routine, just like flying domestic airways.
The KEF ATIS added a final small wrinkle to keep the night interesting. A swirling North Atlantic low pressure had swept in, bringing rain, low ceilings, fog and a stiff quartering headwind. Ceiling reports oscillated around 200 feet with 500 to 600 meters RVR. If ever there was an airport where such weather could be easily accommodated, KEF was it. Runway 11 with its CAT II ILS, 10,000-foot long and nearly 200-foot wide runway, ALSF II, PAPI, TDZ and centerline lights, was everything necessary to get down safely. And it was comforting to know, in the event of a missed approach, we had about two hours of extra fuel on board from the Goose Bay gas-up.
Bill briefed the arrival, and in accordance with low visibility standard procedures, set the autopilot for a coupled approach. Coupled to an ILS, the 727’s autopilot is not known for its smoothness but it gets the job done. Unlike some state-of-the-art, fly-by-wire airplanes that provide no control feedback, the 727 is just the opposite, jerking the control wheel fore and aft, left and right, as it tenaciously tracks the beam to touchdown. At 200 feet, sequenced flashers and approach lights pierced the rain and fog and we broke into the clear at 100 feet. Bill pickled off the autopilot crossing the threshold, de-crabbed and touched down. Two legs down, one to go. It was getting to be a long night.
Having done midnight ferry missions through KEF for the past several years, it was like meeting old friends in Icelandic Air’s Operations office. But being well behind schedule because of the Goose Bay diversion, we didn’t tarry, just exchanged pleasantries, gathered up the teletyped flight papers, and headed out to our refueled aircraft for the final leg to FRA. Thankfully, there was no need for the INS on our route to the continent.
My leg to fly again. Bill could relax, run the checklists, do the communications, tune the nav/com radios and provide support. In short order, we were off and heading southeast toward Scotland. It wasn’t long before the first sliver of light arced across the horizon. Soon we’d be stuffing Jepp airway charts across the windshield and snapping rectangular sunshades in place to shield the blinding sun. The diversion delay put us squarely in the stream of Frankfurt’s scheduled morning arrivals. Weather was misty VMC and Frankfurt Approach was running simultaneous arrivals on parallel runways. Had we arrived just after dawn as originally planned, the descent and approach would have been a more leisurely affair. Now, working us into the flow, vectors, level changes, speed adjustments and frequency changes were sprinkled at us amid the rapid fire instructions to other flights.
Dulled after hours of flying on the back side of the clock, Bill’s concentration was slipping and he sometimes missed a call. After an interval, when it was apparent he wasn’t responding, I’d key the mic, transmit the readback, and execute the instruction. Bill, in an unfatigued state, never missed a thing. When I acknowledged to prevent Frankfurt from repeating instructions, Bill, agitated at himself, would wince and exclaim, “Damn!” which only seemed to inhibit his concentration and lead to more missed calls. Fatigue can do that to anyone, no matter how competent.
A Pan Am ops van met us on the apron where we gave the maintenance supervisor an earful about the faulty INS. He assured us we’d have a different unit for the flight back tomorrow.
The next morning was clear and sunny as we headed to operations for the planned two-leg trip back to JFK. It was Bill’s leg. As simulator instructors always say when switching from one scenario to the next… “new day, new plane.” How true, and it wasn’t long before we found out what this 727 had in store. But we were on our way. Clearance received, checklists accomplished, INS waypoints loaded, we fired up, joined the departure queue and in short order were climbing northwest toward the UK with the morning sun at our back. A good night’s rest and a hearty breakfast had us alert and in good spirits.
Over Scotland, around FL350, ATC chatter subsided, the cockpit was bathed in sunlight and all was well with the world. The flight engineer, a perennially hungry fellow, unstrapped and headed to the galley for snacks. It was a tranquil setting except for the air roaring by outside at .79 Mach. As I mused about nothing in particular, enjoying the sun, three small rectangular amber lights illuminated in quick succession on the first officer’s instrument panel: Aileron System A, Rudder System A, Elevator System A. Something had ruptured in the “A” hydraulic system and, in a matter of seconds, its two 3000psi engine-driven hydraulic pumps had ported all the fluid overboard.
The 727’s hydraulic-powered flight controls have robust redundancy so basic airplane control wasn’t an issue. Nevertheless, with no “A” hydraulics, we were faced with significant flight limitations and would have to divert. The Abnormal Procedures section laid it all out: The functioning “B” system continued to operate ailerons, elevators, the upper rudder, its yaw damper and main gear brakes. But the failed “A” system took out the lower rudder and its yaw damper and some spoilers as well as hydraulic power to lower and raise the landing gear and extend and retract the flaps. The nose gear and left and right landing gear would have to be separately cranked down, flaps would be extended electrically via overhead panel toggle switches. Leading edge devices would extend along with the flaps but once extended, couldn’t be retracted.
But with the lower yaw damper inoperative, there was an immediate need to descend to denser air, at least FL300, and limit indicated airspeed to 270 kts because of the 727’s significant high altitude dutch roll tendencies. Dutch roll is a yaw-induced roll coupling condition common in highly swept wing jet airplanes. In the extreme, dutch roll can rip engines from the wing of a jet transport.
I declared an emergency and requested FL290 and routing to London’s Heathrow Airport. With London still about 30 minutes away, we had work to do. While Bill tended to the flying and got his approach charts organized, I worked the Abnormal Checklist with the flight engineer and tended ATC communications. The flight engineer also contacted London Operations, advising that we’d need a tow to the ramp after clearing the runway because nose gear steering was inoperative.
ATC was advised that we couldn’t accept a standard arrival because we needed extra space and time to crank the wheels down and lower the flaps. Also, in the event of a missed approach, we’d be speed and performance limited because leading edge devices couldn’t be retracted and the landing gear couldn’t be raised. They took it all in, and with characteristic British aplomb, vectored us to a clear area.
Now lower in the approach area, the flight engineer took the crank from its bracket on the cockpit bulkhead, opened the three small hatches on the cockpit floor and set to work, cranking each landing gear down. Next, as approach control headed us toward the inbound flow, Bill set the flap handle for the approach and ordered the flight engineer to toggle the overhead flap switches, electrically lowering the flaps to the partial flap setting in accordance with the abnormal checklist. Approach and landing speed would be about 20 knots faster than normal for our gross weight, about 160 KIAS.
Meanwhile, British aviation buffs monitoring ATC frequencies heard that a Pan Am Clipper jet was inbound with a hydraulic failure. Someone called the BBC and our arrival was covered as breaking news. It probably wouldn’t have made such a media splash had a United DC-10 not crashed while attempting to land following a complete hydraulic failure in Sioux City several months earlier.
Bill’s landing was smooth and straight and, with gentle differential braking, we turned easily onto the parallel taxiway. Crash vehicles followed along as we slowed to a stop. They did a quick walk around to assure there was no fire, fuel leaks or other dangerous conditions, gave us a thumbs up and, after checking with the tower to confirm we had no more problems, they drove off. With the crash trucks out of the way, a Pan Am tug pulled up, the tow bar was attached and we were towed to the ramp. A walk-around revealed a shiny red stain trailing back from a slat actuator on the wing’s leading edge. The seal had failed, liberating all the “A” hydraulic system’s usable fluid.
The next day, chastened by events of the last few days, we showed up bright and early in London Operations emotionally loaded for bear. Who knew what would happen next on our ferry flight odyssey? But two diversions in two days was statistically anomalous. Airline pilots often fly years before experiencing a major malfunction and some pilots never have one in a whole career. The flights back to JFK were uneventful. It was just our turn to be statistical outliers.