ERA13FA131

Maintenance personnel's failure to follow procedures and published directives in calibrating the continuous flow fuel system and failure to accurately diagnose debris in the throttle assembly, resulting in a loss of power in one engine. Contributing to the accident was the pilot's failure to comply with published engine out procedures, which resulted in an off-airport landing and subsequent impact with a tree and the ground.

Shortly after taking off on the test flight, the pilot transmitted "mayday mayday" over the control tower's radio frequency. According to eyewitnesses, the airplane was in a left-wing-down attitude when it impacted a tree, power lines, and then another tree before coming to rest in a pasture. A review of data downloaded from the engine data monitor revealed that the rear engine exhibited erratic fuel flow beginning 2 days before the accident and continuing through the accident flight. Further, the engine data monitor indicated that the rear engine's propeller was under low-to-no power with a low pitch angle at the time of impact. The front engine exhibited no abnormalities or malfunctions, and the investigation found no other anomalies that would have precluded normal operation of the airplane. During postaccident examination, the engine-driven fuel pump was removed and bench tested where it exhibited fuel flow higher than manufacturer guidelines with low fuel pressure. In order to meet bench test standards, an adjustment equal to three turns of the adjustment screw was made. The fuel pump then operated normally and was placed back on the engine; however, the engine still did not attain full power. Further examination revealed potential debris between the throttle assembly's brass and stainless steel plates. After removal of the debris and reassembly of the throttle assembly, it operated within the normal range. Although the source of the debris could not be definitively determined, it likely originated in either in the fuel or a fuel tank. Maintenance records indicate that two days before the accident, the fuel pump was removed, repaired, and reinstalled after work was completed on the rear fuel selector valve. After the pump was reinstalled, the mechanic adjusted the continuous flow fuel injection system using the airplane's JPI engine monitor system and an external low pressure gauge to set the takeoff fuel flow; he then refueled the plane from containers in which he had stored the fuel in order to service the fuel pump. The following day, the pilot and the mechanic again adjusted the fuel pump's fuel flow after conducting an unsatisfactory engine run-up . After the adjustment, the run-up appeared to be normal. Directives from the engine's manufacturer recommended using a Model 20 ATM-C Porta Test Unit or equivalent to ensure the fuel injection system meets all pressure and flow specifications. Using a JPI engine monitor and an external gauge would have given inaccurate results without a properly calibrated fuel pump, and the mechanic's recalibration of the engine-driven fuel pump's adjustment screw would have only masked the debris issue within the throttle assembly. Further, review of the manufacturer's approved engine-out emergency procedures indicated that with a rear engine failure, the propeller should be immediately feathered and the landing gear retracted after obstacle clearance. Therefore, even with a loss of rear engine power, the airplane's operating manual indicated that the airplane would have been able to climb at least 275 feet per minute with one engine, assuming the required pilot inputs were made. Had those single-engine climb performance conditions been met, the airplane likely would have been able to, at a minimum, maintain altitude until a safe landing could have been accomplished. However, as the rear propeller was found with a low pitch angle and the landing gear was found in the down and locked position after the accident, the airplane had not been configured for maximum single-engine performance as outlined in the engine-out procedures.

The airport is a publically owned airport and at the time of the accident had an operating control tower. The airport was equipped with three runways designated as runway 7/25, 11/29, and 02/20. The runways were reported as "in fair condition" or "in good condition" at the time of the accident. Runway 7/25 was a 5,000-foot-long by 75-foot-wide runway, runway 11/29 was a 4,319-foot-long by 100-foot-wide, and runway 02/20 was a 4,000-foot-long by 100-foot-wide runway. The airport was 10 feet above mean sea level. Propeller Examinations Both propellers were sent to the McCauley Propeller Systems in Wichita, Kansas, and examined on April 24, 2013 with oversight provided by an FAA inspector. According to their report the rear engine propeller was examined and appeared to be either at a low or possibly no power at time of impact and no indication of rotation was present. The propeller blade angle was at low pitch and the propeller blades and internal mechanism exhibited very little damage. The bearings and raceways were intact and appeared normal. In addition, oil integrity was confirmed at the propeller bearing. The front engine propeller was examined and one of the blades was unable to be removed due to impact damage. The other blade was examined and indicated rotational scoring between the stops in the "normal operating range." The front propeller hub had a mark from a blade counterweight impact during the accident sequence. The position of this mark indicated a propeller blade angle of approximately low pitch/ latch position at impact. The spring and bearings were intact and appeared to have no anomalies that would have precluded normal operation. A detailed report about the examinations can be found in the "Front and Rear Propeller Examination Report" located in the public docket for this accident. Engine Data Monitor An engine data monitor was recovered from the cockpit and forwarded to the NTSB Vehicle Recorders Laboratory, Washington, DC, for download. Review of the downloaded data revealed there were 11 recorded events, which began in October 2012. According to the data, the rear engine fuel flow exhibited an erratic fuel flow beginning on February 11, 2013, which continued through the accident flight. The data readouts are located in the public docket associated with this accident. Fuel Selector Valves The left and right wing mounted fuel selector valves were removed from the airplane and examined with the engines at the manufacturing facility in Mobile, Alabama, in May 2013, under the supervision of an NTSB investigator. Compressed air was blown through all of the ports and airflow was noted on the selected detents. The detents were checked and operated normally with no abnormalities noted. The right wing fuel selector valve was mounted on a test stand utilizing a slaved fuel pump, and fuel flowed through the valve unabated. Engine Examinations The front engine was examined at the manufacturing facility in Mobile, Alabama, in May 2013, under the supervision of an NTSB investigator. During the examination, several components were removed and replaced to facilitate an engine run. The engine was placed on a test stand, was started and operated at various power settings with no abnormalities or malfunctions that would have precluded normal operation noted. The rear engine was examined at the manufacturing facility in Mobile, Alabama, in May 2013, under the supervision of an NTSB investigator. The engine was examined and the valve rocker covers were removed to facilitate examination of the gaskets and the rocker arms. All cylinders appeared normal and the engine was prepared for an engine run in a test cell. The engine was started and smoke was observed coming out of the turbo exhaust while at 1000 rpm. The engine was idled for several minutes then accelerated to 1600 rpm and the fuel flow was high with low fuel pressure according to manufacturing guidelines. The engine was then accelerated to a full power setting but would not produce power above 1800 rpm and the fuel flow remained high with low fuel pressure. A noticeable surge was audibly detected and black smoke was observed continuously exiting the turbo exhaust. The engine was reduced to idle and the surging continued and was captured on a video recording. The rear engine's throttle assembly was removed and the throttle assembly from the front engine was utilized in its place. The engine was operated with the front engine's throttle assembly and was found to operate smoothly; however, was not able to achieve full takeoff power. The engine driven fuel pump was then removed and bench tested. The fuel pump bench test revealed that the fuel would have high flow and low pressure. The adjustment screw was measured at 0.4645 inches. Adjustments were made that equaled three turns of the adjustment bolt, which equaled 0.0675 total inches of adjustment. Then the fuel pump was adjusted to bench test standards and was operated normally. The rear engine would not develop full power at full throttle setting. After removal and testing of the throttle assembly and engine driven fuel pump, scoring was noted between the brass and stainless steel plates within the fuel metering valve inside the throttle assembly indicated the possibility of debris. The spring on the brass plate indicated the possibility of pinching against the sidewall of the fuel metering valve; however, after examination and reassembly of the unit it operated within a normal range. More details about the examinations can be found in the "Engine Examination Report" in the public docket for this accident. Cessna T337 Owner's Manual According to Section IV "Operational Data", after takeoff and during the initial climb, at an airplane weight of 3,700 pounds, the indicated airspeed at 50 feet agl should be 79 mph. In addition, after takeoff and during the initial climb, at an airplane weight of 4,500 pounds, the indicated airspeed at 50 feet agl should be 87 mph. According to the chart "Single Engine Maximum Rate-of-Climb Data" with an aircraft weight of 3,700 pounds, outside air temperature of 82 degrees F [28 degrees C], and sea level, the single engine climb performance with the rear engine inoperative and the propeller feathered is about 540 feet per minute rate of climb. An airplane weight of 4,500 pounds, outside air temperature 82 degrees F [28 degrees C], and sea level, the single engine climb performance with the rear engine inoperative and the propeller feathered is about 275 feet per minute rate of climb. The section further indicated that at a gross weight of 4,500 pounds, the airplane will stall at 74 mph calibrated airspeed with a 0 degrees angle of bank and flaps at one-third. The airplane will stall at 79 mph calibrated airspeed, in a 30-degree angle of bank, with one-third flaps. Then, the airplane will stall at 105 mph calibrated airspeed, in a 60-degree angle of bank, with one-third flaps. According to Section II "Description and Operating Details," it indicated that the landing gear retraction "is normally not started until one or two hundred feet of altitude have been obtained after take-off. Retraction at very low altitude should be avoided since the landing gear swings downward approximately two feet as it starts the retraction cycle. In addition, the landing gear would extend slowly in the event of an engine-out after take-off, and might not be completely down while a wheels-down landing could still be made on the runway." Section III "Emergency Procedures" provides an "Engine Out After Takeoff" checklist for airspeed above 85 mph (without sufficient runway ahead) which includes: 1. Throttles – Full forward. 2. Propellers – High RPM (full forward). 3. Determine inoperative engine (from engine RPM). 4. Propeller – Front Engine Inoperative – Feather propeller if gear is up and locked, or down and locked. Rear Engine Inoperative – Feather immediately. 5. Wing