This particular incident occurred during a ferry mission that involved picking up a C-123B aircraft that had been grounded for an engine change at Hickam AFB, and delivering it to the Royal Thai Air Force at Don Muang AFB outside Bangkock, Thailand. A crew of five; two pilots, a navigator, a flight engineer, and a loadmaster; met at Hickam AFB, conducted a 5-hour engine-change check flight, signed for the aircraft, and prepared to deliver the aircraft to Thailand. This ferry mission would require six flights and approximately 40 hours of flying time to reach Bangkok.

Figure 1 – C-123 Provider
All of the crew, except the navigator, were from the 317th Troop Carrier Squadron at Pope AFB and had flown together many times. The navigator was from a RC-121 Airborne Surveillance Squadron out of March AFB as I remember. While he had comparable experience to the rest of the crew, his usual navigation and communication equipment was some of the most sophisticated in the Air Force at the time. He was quite surprised, even amazed, to find that the only navigation equipment on board the C-123B consisted of a gyro compass, a VOR/TACAN system, an ADF radio, a Mark I Loran system (LORAN-A), and a hand-held sextant. He commented that he hadn’t seen, let alone used, equipment that old since navigation training school. The only Mark I Loran he had ever seen was a display unit at the Nav school. In the early 1960s, Loran was the state-of-the-art in long-range navigation position-fixing technology. It was the GPS of the day and worked on the same principle of determining position from the delta time-of-arrival from several different ground-based high-frequency radio stations. Loran accuracy was on the order of one nautical mile with the Mark I system. The navigator spent the entire engine-change check flight familiarizing himself with the navigation equipment and testing their operation.
The C-123B aircraft has a manual flight control system and does not have an autopilot system. While the aircraft can be manually trimmed to maintain heading and altitude, a pilot must be at the controls at all times. When the pilot at the controls wants to take a break, the procedure was to poke the other pilot and tell him that he has the controls. Once the “autopilot” was engaged, the former pilot could relax until he was “reengaged” as the active pilot. The C-123B is also not pressurized and does not have an air conditioning system. Cockpit ventilation is provided by a suction manifold above the windshield.
In order to understand the lack of modern, even for the time, systems and equipment on the aircraft one must understand when, and for what mission, the aircraft was originally designed. The C-123B aircraft was designed in the late 1940s early 1950s to perform the air assault mission that the C-47 towed gliders performed during WW II. The C-123B was basically designed as a powered glider that would land gear up on unprepared surfaces to deliver troops, equipment, and supplies behind enemy lines. It even had a tow ring mounted in the nose of the aircraft.. The main nacelle fuel tanks were designed to be jettisonable in flight to minimize the possibility of fire during an assault landing. With the fuel tanks jettisoned only about 90 gallons of flammable oil remained on board. With a normal fuel load of 8,000 pounds (1,000 gallons) of 115/145 high-octane gasoline in the nacelle tanks, and two pylon-mounted 120-gallon(?) external drop tanks, the normal no-wind maximum range of the C-123B was around 800+ miles. Since most of the island-hopping legs across the Pacific are at, or exceeded, the 800 mile maximum range of the aircraft, ferry mission aircraft were equipped with a 600-gallon auxiliary fuel tank pallet-mounted in the cargo bay. Since the P&W R2800 engines normally burned quite a bit of oil, a 55-gallon drum of engine oil was also mounted in front of the landing gear well to replenish the 48-gallon nacelle oil tanks in flight.

Figure 2 – Wake Island Runway
The flight from Hickam to the Midway Islands was routine and uneventful, as was the flight to Wake Island, except that the Loran system quit shortly after takeoff from Midway. The Loran system would prove to be unreliable throughout the mission, primarily due to a shortage of parts and experienced maintenance personnel. That’s the problem with actually using antiques. The navigator swore that the Loran system was wired through the landing gear down-lock switches and would only work with the gear down..
The takeoff, climb out, and cruise from Wake Island for the approximately 7-hour flight to Anderson AFB at Guam Island started routinely. The Loran system was again proving to be unreliable and navigation had to be based primarily on dead reckoning (DR); heading, airspeed, drift meter readings, and ADF bearings; augmented by periodic sun shots with the hand-held sextant. Sun shots were proving difficult due to scattered cloudiness, and the associated turbulence, which interrupted several attempted shots. The navigator was staying very busy trying to verify our DR position. Our actual location would become very important very soon.
A little over three hours after takeoff from Wake Island the crew, except for the navigator, had completed their normal in flight routines. The pilot in the right (copilot’s) seat was being the autopilot and I was relaxing with a cigarette in the left seat. The flight engineer and loadmaster were lounging in the rear of the aircraft and the navigator was calculating his next sun shot. Abruptly the right (#2) engine just quit.. no sputter or surge. It just quit, like the ignition had been turned off. The other pilot and I just looked at each other.. the moment of Stark Terror, and in the middle of the Pacific Ocean yet! We were almost exactly halfway between Wake Island and Guam and it was highly questionable if we had enough fuel remaining to get to either of them on one engine.

Figure 3 – Stark Terror
After the Moment of Stark Terror, the copilot and I started the engine failure emergency checklist. Before I could do anything other than turn on the fuel boost pumps, push the mixture levers to full rich, and the propeller control levers to maximum RPM, the #2 engine restarted and ran normally by all indications. The flight engineer visually checked the #2 engine for indications of a malfunction. The engine appeared normal in all respects which was verified by all engine indicators. After we had observed the engine for less than a minute, it quit again.. Another moment of Stark Terror. Then it started running again after about 15 seconds, and then it quit again, and then it restarted again after about 15 seconds. By now we knew that we had an engine that was having a serious problem, even though we had no idea what the problem was nor how long it would keep functioning.
We completed the emergency engine failure checklist for the left engine except we left the power at cruise setting, no need to stress it unnecessarily. Since the right (#2) engine was having some kind of problem, we decided to minimize the stress on it by setting the engine speed to a “gyroscopic” value, an rpm that the engine will turn at with minimum vibration if one or more cylinders is inoperative. As I remember this was a setting of 2100 rpm at 21 inches of manifold pressure. (Give me a break on these values, after all its been 45 years.) This put minimum stress on the malfunctioning #2 engine, yet provided enough power for it to carry its own weight without having to stress the good engine. At this point, we were babying both engines as much as possible short of sending the flight engineer outside to caress them.

Figure 4 – Runway cliffs, Anderson AFB Guam
After we reset the engine rpm, the #2 engine smoothed out and quit cutting out every few seconds. With the eminent danger of having to shut down an engine apparently over, the next question was where was the nearest place to land if the #2 engine decided to completely quit. That question immediately put the navigator as the center of attention. He was already making a DR estimate of our position from his last fix point. His best determination of our position was that we were approximately 20 minutes short of the “equal-time point”, the point at which it would take the same amount of time to continue to Guam or return to Wake Island, with both engines operating. Considering that 1) our speed would be slower on one engine, 2) the runway at Anderson AFB on Guam is on top of 200 foot sheer cliffs, and 3) the runway at Wake is maybe six feet above high tide; the decision was made to return to Wake Island. An option to divert to Saipan Island was considered if it turned out we didn’t have enough fuel to make it all the way back to Wake Island. This was a consideration because our weight might require the use of METO (Maximum Except Take Off) power to maintain level flight which would greatly increase our fuel consumption. We normally compute single-engine fuel required from the mid-point of each leg of an over water flight, but for no-wind conditions. In our particular case, we had a moderate westerly wind that was nearly a direct crosswind. This would require more fuel than the pre computed value, and be a headwind for a diversion to Saipan. So it was back to Wake Island…
As if to emphasize this decision, the #2 engine decided to cut out again for a few seconds. This would prove to be a preview of the next 4 hours of the flight. Every 15 to 20 minutes the #2 engine would stop running for 5 seconds or so and then start running smoothly again. A continuing series of Moments of Stark Terror because we never knew if it would restart again.
When the #2 engine first cut out, the copilot had called a PAN alert on the HF radio emergency frequency to advise all ground control station and aircraft that we were experiencing engine problems. Every ship, aircraft, and ground station in the Pacific should have heard that our aircraft was having trouble and approximately where we were. At least that’s what we hoped. Rule one when flying is to fly the plane first and talk last. On the first radio call we only gave our identification, nature of the problem, flight path, and time out of Wake Island because that’s all we knew at the time. After we had assessed the situation and made our decision to return to Wake, we called Wake Island Radio and advised them of our continuing malfunctioning engine, estimated position and ground speed, and intent to return to Wake and ETA (estimated time of arrival). Wake Radio acknowledged our situation and requested that we descend 1000 feet to conform to altitude/heading flight rules. Since we had by then determined that it would be marginal to make it back to Wake on one engine, and since altitude is life when flying, we requested clearance at our present altitude, and all altitudes below us, for the return flight. Wake Radio approved our request and asked if we would like to declare an emergency and needed SAR (Sea-Air Rescue) escort.
The #2 engine was still cutting out on a regular basis and for varying lengths of time. We had also recalculated our weight, position, and airspeed, and determined that if the #2 engine failed completely we would most likely not be able to make it back to Wake Island. We would have to either ditch or bail out over water short of the island. Given this situation, we declared an emergency and requested a SAR escort for the return flight. Wake Radio acknowledged our emergency and scrambled the local SAR aircraft, a Grumman SA-16 Albatross seaplane. SAR gave us an ETI (estimated time to intercept) of 2 hours.

Figure 5 – Grumman SA-16 Albatross
During the 4 hour return flight back to Wake Island, we made periodic position reports to Wake Radio and prepared for a complete failure of #2 engine. Our calculations indicated that if #2 failed more than 2 hours out of Wake that we would have to dump so much fuel to maintain level flight that we would not be able to reach the island. The sea state showed there were significant swells on the surface that would make successful ditching unlikely. The C-123B did have a ditching procedure, but the tail section tends to break off at the cargo door making survival unlikely. We therefore decided that we would not ditch but would bail out at 1000 feet above the water. Besides parachutes, our survival equipment consisted of rubber survival suits, individual dinghies, and a 6-man inflatable raft. None of the crew had been to water survival training, so we reviewed all the instructions on the equipment and our checklists. We didn’t put on the survival suits because it was fairly warm in the aircraft and it would be easy to overheat in the survival suits. We also figured that at our current altitude we had at least 20 minutes after an engine failure until we would have to bail out, plenty of time to put on a survival suit and parachute.
Having prepared for a complete engine failure we spent the next 3+ hours waiting for the #2 engine to fail at any time, and making sure that the #1 engine was not being stressed in any way. The SAR SA-16 made their intercept within 5 minutes of their original estimate and fell into a loose formation off our right wing to keep an eye on our #2 engine. Our cruise airspeed with the #2 engine at a gyroscopic power setting was slightly less than 105 knots (120 mph). The SA-16 crew commented to us that we were the only aircraft they had intercepted that they could keep up with. They generally intercepted four-engine C-124 aircraft with an engine out. The C-124 three-engine speed was faster than the maximum speed of the twin-engine SA-16. We landed back at Wake Island without further complications, other than the periodic cutting out of the #2 engine.
After landing and engine shutdown we immediately opened up the #2 nacelle to check the engine. There were no immediate visible problems, however a close inspection of the engine revealed that one spark plug on one cylinder had blown out the core which exposed the electrode to shorting-out on the engine block. When the electrode shorted to the engine, it had the effect of cutting off the ignition to the engine. The vibration caused by the engine cutting out would move the loose electrode around enough to stop the shorting. This is what was causing the repeated cutting out and restarting of the engine. If the electrode had become jammed in a shorted condition we would have lost the engine for good, and possibly the aircraft. The damaged plug was removed and the cylinder was thoroughly borescoped to make sure that there was no internal damage. The cylinder was undamaged and a couple of new plugs solved what could have been a serious problem. A prolonged engine run-up and post-run visual check verified that the #2 engine was operating normally. Clearance was received from our maintenance section at Pope AFB to continue the ferry mission the next day.
Immediately after securing the aircraft we proceeded to deliver the best bottle of liquor we could find on the island to the SAR ready room. We thanked the crew that intercepted us and told them to have several drinks on us when they got off duty. We discussed the day’s events and told them what had been the problem with our engine. After personal thanks all around, we headed to the local bar to get smashed so we could sleep that night. Needless to say we were still full of adrenalin from 4 hours worth of Moments of Stark Terror.
Apparently everybody in the Pacific was aware of our situation due to our use of the emergency HF frequency. At one point we had to relay a position report through Okinawa due to poor reception with Wake. Everyone at the bar was aware of our emergency and several people brought us rounds of drinks.
The next day, after a through pre-flight and engine run-up, we completed our flight to Anderson AFB at Guam, and the rest of the mission to Don Muang RTAFB, without further problems. While many people have flown the Pacific in twin-engine aircraft, I think most pilots will agree that the Pacific Ocean is definitely a four-engine ocean.
Louis Kirchdorfer
April 20, 2009







