Lost at Sea: Latent Design Flaws in a USAF HC-130P Aircraft, Call Sign King 56
By George Anderson, ESEP
Systems thinking and the application of Systems Engineering (SE) processes are both necessary elements in the discovery and description of technical failures associated with some fatal aircraft accidents. The King 56 aircraft investigation was a fascinating search for forensic truth and I was part of the investigating team
My narration begins with the actual crash of the aircraft and covers highlights of a three- year period that ended with the completion of the USAF final report and its presentation to the Secretary of the Air Force.
I will not attempt to cover all details of the investigation, but focus on the investigative path taken that led to the final resolution. This path covers four distinct phases:
Gathering data and creating scenarios or hypotheses,
Testing hypotheses and discovering those best supported,
Gathering physical evidence to establish or disprove the hypothesis, and
Explaining the entire chain of causal events that contributed to the accident or mishap in the USAF terminology.
On November 22, 1996, USAF HC-130P, tail number 64-14856, crashed into the Pacific Ocean approximately 50 nautical miles west of Point Mendocino on the California coast. Ten of the eleven people on board died.
Based on the digital flight data recorder (DFDR) the accident sequence began at 1844 hours Pacific Standard Time (PST) and all four engines totally failed two minutes later at 1846 hours. After the loss, the aircraft descended without power but under control of the pilot until it impacted the ocean surface where it broke apart. The outboard wing sections and all four engines separated from the center wing section that in turn separated from the fuselage.
Subsequently, the engines and fuselage went straight to the ocean floor at a depth of approximately 5500 feet. The outer wing and the center wing sections floated on the surface for several days and sank more than 50 nm from the impact location.
What evidence did the investigators have initially?
The only survivor, the radio operator, had no substantive information about the operation of the aircraft systems. No relevant physical evidence was available but the last few minutes of flight before electrical power was lost was captured on the DFDR that had been recovered and its audio contents released to the public.
Gathering Data and Scenario Building
The DFDR was the only clue to how or why the engines quit. There were four avenues of approach taken initially for scenario building. These were:
Fuel starvation due to mechanical malfunction or crew fuel mismanagement
Electrical fault of unknown type
Fuel contamination or improper fueling
External electromagnetic field (e.g. a Pave Paws Radar operating near Marysville, CA)
The press had its own scenarios, several of which I have footnoted for perspective. 
The USAF initially convened an accident investigation board to try and determine the cause. All the service records of the aircraft were searched for clues and plenty of items were found to be irregular but nothing that could establish enough evidence to support any of the created scenarios. Finally, the board formally concluded that the most likely cause was fuel mismanagement by the crew.
Shortly after this result was announced, the C-130 fleet was afflicted with multiple reports of C-130s of all types experiencing simultaneous power surges and rollbacks on all four engines. The confidence in the fleet was rapidly becoming a news story and Congress finally intervened and demanded that the USAF investigation be reopened.
In this climate of doubt, lawsuits were filed against Lockheed, the C-130 manufacturer, and Allison (now Rolls Royce) the engine manufacturer alleging defective design, manufacture and/or oversight of their products.
The result was that the USAF was directed by Congress to reopen the investigation and appointed me, an NTSB investigator, to serve on the team.
To obtain final closure on the cause of King 56’s loss, more evidence was needed and would have to be recovered from the sea. Scenarios were created to identify possible failure modes that might have left evidence in the wreckage.
After several months of scenario building, I postulated that aeration of the fuel by an unknown condition might have caused the event. I suggested a test.
Two weeks later the Allison Engine representative reported to the Board that he had run a ground test on his own and found that the engine rolled back with just 5 psi of compressed air being injected into the fuel line.
From there, events quickly unfolded and a C-130 flight test at Nellis AFB, duplicated the symptoms. That is, they injected air into the common fuel manifold and observed that all four engines began to fail. The wisdom of undertaking this test notwithstanding, the scenario building now began to focus on how low pressure air could enter the fuel manifold.
After many dead ends, a source of pressurized air was found in the fuel system.
The aircraft was a rescue model and had two large auxiliary (Benson) tanks mounted in the fuselage to extend the range. Instead of venting these tanks to the outside air as was normally done, these tanks were vented to the interior of the cargo compartment. A pressurized aircraft in flight has a nominal pressure differential between the cabin and the outside air of 8 psi. This pressure would not pose a threat to the engines if both the fuel tanks contained fuel but if fuel was absent, the air would enter the fuel lines to all four engines under certain valve configurations of the fuel system.
How could this air reach all four engines? The short answer was for the flight engineer to not follow standard procedures for operating the many boost pumps used to transfer fuel out of the eight wing and two fuselage tanks to the 4 engines.
Our Tests showed that: (1) If all four main tank boost pumps were turned on as was directed by the flight manual, air from the cabin could not cause an engine flameout; and (2) If all four main tank boost pumps were off, then the engines would quit when either fuselage tank was run dry. The original accident investigation team was beginning to feel the road to vindication was getting shorter.
The next question was could recovery of fuel line components establish the state or configuration of the fuel system at the time failure? Very big stroke of luck here! The fuel system valves are all motor-driven using 28 volts DC. Because this power is lost when all four engines lose power, all fuel valves would be expected to remain in the last position selected by the crew.
For example, if the valves from a fuselage tank were found to be open then it strongly suggested that the flight engineer was feeding all four engines off that one tank.
Calculations of fuel burn rates indicated that if indeed this were the configuration selected after takeoff then the time of the power loss was consistent with a fuselage tank running dry.
Off we went to find the wreck. We boarded a research vessel equipped with a remotely operated undersea vehicle and a crew of salvage experts. We found the wreckage quickly and during seemingly endless 6- hour shifts we identified and retrieved parts of the wreckage. All the fuel valves we wished to examine were in the center wing section. It was raised once and broke free in heavy seas injuring several crewmembers. A second attempt using large chains was successful and we confirmed our hypothesis that the fuselage tank valve and others leading to all four engines were open.
The engineer’s fuel panel was also recovered and many of the switches for the valves were inconsistently in the closed position. One can see in Figure 3 the ease of determining the status of each valve with a quick glance. This suggested that the crew had sorted out the problem but had repositioned the switches too late to save themselves.
We had our answer but to be thorough we carefully tore down several engines and examined components that were part of other scenarios. There were no suspicious indications found.
The mission concluded after a confrontational scene in Portland, OR where the families of the dead crewmembers suggested that the investigation was all a lie designed to protect a defect in the C-130 fleet.
Actually, they were partially right but not for the reasons that they posed. The AF had never included a requirement in the C-130 design for air-starting an engine after all four engines quit. Other aircraft have this feature and it is sad that this crew was not given a second chance to recover from what was clearly a collective lapse of good airmanship.
 DFDR Transcript