Two young men, both aspiring for aviation careers, were killed on March 24, 2014, at about 1738 when the Piper Seminole PA-44-180 they were piloting broke up in flight and crashed into a salt marsh near Brunswick, Georgia. The NTSB determined that the accident probably resulted from the pilots losing control of their aircraft while flying in stratiform instrument meteorological conditions (IMC) at 8,000 ft. MSL.

The NTSB’s findings present a cautionary tale with several chapters including small aircraft minimum equipment lists (MELs), pilot judgment, partial panel competency and crew pairing during training exercises. Specifically, the Safety Board determined:

An inflight failure of the airplane’s only operating vacuum pump, which resulted in the loss of attitude information provided by vacuum-driven flight instruments. Also causal was the pilots’ failure to maintain control of the airplane while operating in instrument flight rules (IFR) conditions, likely due to spatial disorientation, following the failure of the vacuum pump. Contributing to the accident was the operator’s decision to dispatch the airplane with a known inoperative vacuum pump into IFR conditions.

The airplane was owned by Airline Training Professionals (ATP) and based at Jacksonville Executive Airport (CRG) at Craig, Florida. The flight had departed from Concord Regional Airport (JQF), Concord, North Carolina, at about 1551 and was heading back to Craig. Day VMC prevailed below 3,000 ft. MSL. FAA and Lockheed Martin Flight Service records indicate the right-seat pilot filed the IFR flight plan with an intended departure time of 1530 and had requested an en route cruise altitude of 8,000 ft.

The right-seat pilot held a private pilot certificate with ratings for airplane single-engine and multiengine land, and instrument airplane. His first-class medical certificate was issued Feb. 4, 2014, with no limitations. At the time of the examination, the pilot reported 73 total hours of flight experience and no flight experience within the six months prior to the application. The right-seat pilot’s logbook was located within the wreckage, and the last recorded entry was dated March 24, 2014, recording a flight from JQF to CRG in a different PA-44-180. His total recorded flight time was 155.3 flight hours and 8.8 hr. of actual instrument flight experience.

The pilot in the left seat held a private pilot certificate with ratings for airplane single-engine and multiengine land, and instrument airplane. He held an FAA first-class medical/student pilot certificate issued May 20, 2013, with no limitations. At the time of the examination the pilot reported no previous flight experience. His logbook was also located within the wreckage and the last recorded entry was dated March 23, 2014, recording a flight from Daytona Beach, Florida, International Airport (DAB) to JQF. Including that entry, his total recorded flight time was 163.7 flight hours and 1.8 hr. of actual instrument flight experience.

The NTSB was unable to determine which pilot was manipulating the controls when the aircraft descended from level flight.

The Accident

The crew obtained a preflight weather briefing from a DUATS provider at 1516 and filed an IFR flight plan through with Lockheed Martin Flight Service.

The airplane departed at about 1551 and climbed to its cruise altitude of 8,000 ft. While en route, the crew contacted Flight Watch at 1647:32, obtained the latest weather for the destination, obtained the updated AIRMET Sierra for IFR conditions over Florida, and issued a PIREP stating that the flight was 11.8 mi. southeast of Barnwell Regional Airport (BNL), Barnwell, South Carolina. The PIREP indicated 10 mi. visibility, no turbulence, no visible weather and light haze. The recording ended at 1652:28.

The GOES-13 infrared image indicated a layer of low to mid-level stratocumulus to nimbostratus type clouds with tops near 20,000 ft. No defined cumulonimbus clouds were identified in the vicinity of the accident site.

The flight contacted Jacksonville Center at 1727:23 and reported the Piper twin’s altitude as 8,000 ft. The controller provided the flight with the local altimeter setting. At 1727:32, the flight confirmed the altimeter setting, and no further communications from the flight were received. Controllers made several attempts to contact the accident flight between 1739:27 and 1740:36, all of which were met with no recorded response.

The Center controller solicited assistance from a nearby airplane asking its crew to fly in the vicinity of the last radar return to see if the accident airplane could be located. The controller reported to the other airplane that the accident flight was “. . . at eight thousand in your vicinity uh I uh saw him go down about three hundred feet and then uh completely lost him on radar. I don’t know if it was full electrical failure or what, but I lost him right over the Golden Isles airport.”

At 1743:14 the assisting airplane stated “I’m gonna need you to descend me to probably two thousand five hundred so I can get out of the clouds.”

At 1745:07, the Center controller reported to the other airplane “. . . he was just east of the Golden Isles airport southbound at eight and then uh I showed him at three hundred feet low and then uh I I lost [everything didn’t even get a] primary on him.” The accident flight was unable to be located and no ELT was audibly observed.

The 1755 recorded weather observation at Brunswick Golden Isles Airport (BQK) included wind from 050 deg. at 10 kt. with gusts up to 16 kt.; visibility, 10 mi.; overcast clouds at 3,800 ft. AGL; temperature, 12C; dew point, 7C; and barometric altimeter setting of 30.08 in. of mercury. The regional radar mosaic from the University Center for Atmospheric Research for 1745 indicated an area of very light intensity echoes in the vicinity of Brunswick, Georgia, in the range of 5 to 15 dBZ.

With the help of ground witnesses, the main wreckage was found inverted in a marsh area, in about 15 ft. of water at an elevation of 4 ft. MSL about 3.5 mi. east of BQK. The debris path was about 2 mi. in length along a linear path. The engines remained attached to the airplane with lines and cables, and were colocated with the main wreckage. Both pilots had died of multiple blunt force injuries; their post-mortem blood tests were clear.

The wreckage debris path was oriented on 225-deg. magnetic heading from the main wreckage and about one-half mile from the last radar return. Both outboard wing sections were located about nine-tenths of a mile from the main wreckage and the vertical stabilizer was located about 2 mi. from the main wreckage.

Investigators said examination of the recovered airframe and flight control system components revealed no evidence of pre-impact mechanical malfunctions.

Examination of the engines revealed both vacuum pump shafts (couplings) were sheared; however, no other evidence of pre-impact mechanical malfunctions was noted. The only recovered air-driven gyroscopic flight instrument was the horizontal situation indicator (HSI), which exhibited minimal damage to the housing and no evidence of crush damage. Disassembly of the gyro found no scoring or rotational damage to the pendulous vane housing or rotating assembly. That finding suggests that the gyro was not turning at impact.

Review of FAA radar data revealed that the flight had been cruising at an indicated an altitude of 8,000 ft. MSL. The subsequent four recorded radar targets, which occurred over a span of 38 sec., indicated an altitude of 7,900 ft. to 7,200 ft. MSL. The last recorded radar data, at 1738:10, indicated an altitude of 0 ft. and was in the vicinity of the accident location. The last four radar targets, which indicated an altitude above 7,000 ft. MSL, were consistent with a descending right 180-deg. turn.

The airplane had been manufactured in 1976 and had accumulated a total time of 6,664 hr. It was registered to ATP Aircraft 2 LLC on Jan. 29, 2008. The airplane had no autopilot.

Maintenance records showed the left engine vacuum pump was installed on May 13, 2011. The records indicated the right engine vacuum pump was installed on Aug. 31, 2013; however, the unit’s serial number in the maintenance logbook did not coincide with the serial number on the vacuum pump in the accident airplane.

The operator told the NTSB that the right vacuum pump that was found on the accident airplane actually had been installed on Jan. 24, 2014. “However,” said the Safety Board, “the information submitted to the NTSB did not contain a serial number nor was the information located within the airplane’s maintenance logbooks provided by the operator. The Airworthiness Approval Tag, associated with the right vacuum pump, that was at the accident scene, indicated that it had been overhauled on June 18, 2013.”

According to maintenance records for the accident airplane, the right engine vacuum pump had accumulated about 65 total hours since overhaul. “However, information provided by the operator,” said the Safety Board, “also revealed that this vacuum pump was noted as inoperative and placarded in accordance with FAR Part 91.213(d) and their [ATP’s] maintenance procedures prior to the accident flight.”

Safety Board investigators said the operator had planned to have the airplane flown to the maintenance facility at CRG to repair the inoperative vacuum pump.

“Additionally, the pilots had been verbally advised of this discrepancy and were provided the opportunity to refuse the airplane, prior to departing on the accident flight. According to a representative of the operator, no flight limitations existed, for their airplane, for an inoperative vacuum pump beyond those prescribed by applicable FARs and the airplane Pilot’s Operating Handbook.


Investigators undertook bench examination of both vacuum pumps. They had remained attached to their respective engines, but both drive shafts had sheared and salt deposits (from the marsh water) were noted throughout the units.

According to maintenance records for the accident airplane, the left engine vacuum pump had accumulated 1,207.8 total hours since overhaul. The rotor was found fractured in multiple locations, and cracks were present on four of the six slots and a transverse crack was present across the hole on one of the rotor flanges. The fracture surfaces of the rotor were consistent with brittle overstress fracture and were consistent with multiple crack initiations locations. Horizontal and vertical sliding marks were present on the faces of the vanes and were consistent with sliding of the vanes that occurred during rotational operation of the rotor. The investigation could not conclusively determine how long the left engine vacuum pump operated after the coupling was sheared.

Service Letter RASL-005, with a revision date of Oct. 6, 2010, contained guidance to clarify the recommended replacement intervals of new and overhauled pneumatic vacuum pumps. The Service Letter noted that the left and right vacuum pumps had a time in service life limit of 500 flight hours, or six years from the date of installation. Upon reaching the service life limit, the manufacturer specified that the pump should be replaced.

Master MEL

The Piper PA-44 Master Minimum Equipment List (MMEL) for Part 91 operations lists all items of installed equipment that are permitted to be inoperative. The MMEL states, in part, that “it is incumbent on the operator to endeavor to determine if O [operations] and/or M [maintenance] procedures for that equipment must be developed . . . the MMEL is intended to permit operations with inoperative items of equipment for the minimum period of time necessary until repairs can be accomplished. It is important that repairs be accomplished at the earliest opportunity in order to return the aircraft to its design level of safety and reliability. . . . Operators are responsible for exercising the necessary operational control to ensure that an acceptable level of safety is maintained.”

According to Section 37 “Vacuum/Pressure,” sequence title “1. Vacuum Pump” that two vacuum pumps are installed and one is required for dispatch. It further provided, under the “Remarks or Exceptions,” that “(M) One may be inoperative for day VFR flights.”

The PA-44 Scheduled Maintenance Manual includes a checklist to be utilized for inspections every 50 and/or 100 hr. Subsection B.36 states “Inspect and operationally test vacuum pumps and lines.” It further requires that the vacuum pump installed on each engine is inspected every 100 hr. Note 7 further states “Replace or overhaul, as required, or at engine overhaul.”

The Pilot’s Operating Handbook states in Section 7.19, “The vacuum system operates the air-driven gyro instruments. The vacuum system consists of a vacuum pump on each engine, plus plumbing and regulating equipment. . . . A shear drive protects the engine from damage. If the drive shears, the gyros will become inoperative. The vacuum gauge mounted on the right instrument panel to the right of the radios provides valuable information to the pilot about the operation of the vacuum system . . . (a low vacuum indicator light is provided in the annunciator panel). . . . In the event of any gauge variation from the norm, the pilot should have a mechanic check the system to prevent possible damage to the system components or eventual failure of the system. . . . A vacuum regulator is provided in the system to protect the gyros. The valve is set so the normal vacuum reads 4.8 to 5.2 in. of mercury. . . .”

Should suction drop below 4.5 in. Hg, pilots are cautioned to increase rpm to 2,700, descend to maintain 4.5 in. and use the electric turn indicator to monitor directional indicator and attitude indicator performance.

Safety Board Analysis

The NTSB offered this analysis of the accident:

Examination of the wreckage at the accident site revealed signatures consistent with an inflight breakup of the airframe. The horizontal situation indicator (the only vacuum-system-driven flight instrument that was recovered) exhibited signatures showing that it was likely not operational when the airplane impacted the ground.

Both of the engine-driven vacuum pumps exhibited fractured rotors. Although physical examination of the vacuum pumps could not determine whether the rotors fractured before or during impact, the inoperative horizontal situation indicator suggests that both pumps had failed before the impact.

The operator reported that the vacuum pump mounted to the airplane’s right engine was not operational before the airplane was dispatched on the accident flight and that the pilots had been advised of this deficiency. The operator used the Part 91 minimum equipment limitations for flights, which permitted dispatching the airplane with only one of the two engine-driven vacuum pumps operational. However, the FAA’s master minimum equipment list for the airplane for Part 91 operators advises operators to limit the airplane to daytime visual flight rules flights when only one of the two vacuum pumps is operational.

The operator’s decision to dispatch the airplane with a known mechanical deficiency and no operational limitations reduced the safety margin for the flight and directly contributed to the accident.

It is likely that the left vacuum pump failed en route, rendering the vacuum-driven flight instruments inoperative. Given the pilots’ minimal flight experience operating in IFR conditions combined with the difficulty of detecting and responding to the loss of attitude information provided by the vacuum-driven flight instruments, it is likely that the pilots became spatially disoriented and lost control of the airplane, resulting in the subsequent inflight breakup.

So we’ve got a number of lessons here including the importance of understanding the airplane’s maintenance-related limitations before flight, and the need to include attitude reference subsystem health indications in your scan (old spinning iron or new AHRS systems).