MIA98IA100

History of the Flight

On March 10, 1998, about 0950 eastern standard time, an Aerospatiale ATR-42-300, N275BC, registered to AMR Leasing Corporation, and operated by Flagship Airlines, doing business as American Eagle, as flight 5634, a 14 CFR Part 121 scheduled domestic passenger flight from Tampa, Florida, to Miami, Florida, experienced an electrical malfunction in cruise flight near Fort Myers, Florida. The crew declared an emergency with air traffic control (ATC) and diverted to Fort Myers. Visual meteorological conditions prevailed and an IFR flight plan was filed. The airplane was not damaged. The airline transport-rated captain, first officer, 1 flight attendant, and 44 passengers reported no injuries. The flight originated from Tampa about 20 minutes before the incident.

The flightcrew stated after the incident, the airplane had been in cruise flight at 15,000 feet, for approximately 20 minutes when the No. 1 generator fault light illuminated. They attempted to reset the No. 1 generator unsuccessfully, at which time the No. 2 generator went off line. They then declared an emergency and initiated a descent below 10,000 feet, while they attempted to reset the Nos. 1 and 2 generators. They reported that the main battery dropped off line within 1 minute, at which time they placed the standby-under-voltage switch in the override position and re-established communication with air traffic control. The flightcrew reported that they made a no flap landing and utilized the emergency brakes to stop. The captain reported that the main battery was removed by fire department personnel, subsequent to the fire chief's report that the battery was hot.

Maintenance personnel for Flagship Airlines reported that after the incident they installed a serviceable main and emergency battery, replaced the No. 1 starter/generator, replaced the No. 1 generator control unit, and replaced a wire terminal lug, broken loose from the field excitation return wire, connected to the No. 1 generator terminal block "E" post. The technicians reported that subsequent to these repairs they were able to start the No. 1 and 2 engines using battery power and bring both generators on line normally.

Flight Data Recorder

The airplane Digital Flight Data Recorder (DFDR) was read out by the National Transportation Safety Board in their Vehicle Recorders Laboratory located in Washington, DC. The Safety Board's DFDR Factual Report indicates that the aircraft was cruising at approximately 14,750 feet, when the incident occurred and that the recorded flight parameters were sporadic for approximately 45 seconds thereafter, at which time all power to the flight recording system was lost. The report further indicates that the DFDR recorded about 90 parameters; however, only 16 were presented in a chronological plot and in tabular format for about 100 seconds of the power malfunction. A complete copy of the DFDR data was supplied to the aircraft manufacturer (ATR) for their independent evaluation. (See the NTSB Flight Data Recorder Specialist's Factual Report of Investigation, an attachment to this report)

Aerospatiale's evaluation of the DFDR revealed that several subframes of data were completely missing. By comparing the Greenwich Meridian Time (GMT) recorded by the DFDR in sequential subframes, the manufacturer determined that on at least two separate occasions several seconds of recorded information were also missing, thereby indicating that the FDR had stopped due to a power supply interruption. The manufacturer noted that the power supply to the DFDR is the 115 volt AC Standby Bus, which is normally supplied by DC Bus No. 1 via the No. 1 inverter. In the case of under voltage on DC Bus No. 1, the Standby Bus is powered by DC Bus No. 2 via the No. 2 inverter. In the case of under voltage on DC Bus No. 1 and 2, the Standby Bus is supplied by the main battery via the No. 1 inverter. The manufacturer further explained that when the input voltage to a static inverter is above 16 volts, its output voltage is normal; however, between 13 and 16 volts its output voltage is outside nominal values, and below 13 volts no power is available from the inverter.

Aerospatiale also determined that on at least three separate occurrences (subframes), the flight parameters (i.e. altitude, speed and elevator position) were inconsistent with the previous flight phases of the incident flight. The manufacturer concluded that during these subframes the FDR ran but the Flight Data Acquisition Unit (FDAU) was de-energized, and therefore the data corresponded to a flight recorded 25 hours previous. Aerospatiale noted that the FDAU requires a minimum of 18 volts for proper operation and is supplied by the DC Essential Bus. Note that the DC Essential Bus is normally powered by Main DC Bus No. 2 via the main BCC. However, upon the loss of both generators the DC Essential Bus is supplied directly from the Hot Main Battery. Aerospatiale confirmed that Main DC Bus No. 1 was de-energized on two separate occasions, approximately 12 seconds apart, just prior to the complete loss of power to the recording system. (See the NTSB Systems Group Chairman's Factual Report, an attachment to this report)

Tests and Research

Review of aircraft maintenance records revealed that N275BC, serial number 275, had 13914.8 total hours and 14,861 cycles at the time of the incident. The airplane completed a maintenance check (PS-2) on March 8, 1998. This included a change of the No. 1 generator and the nose landing gear. The airplane has flown 1 hour 45 minutes and 2 cycles since the maintenance check.

After the incident, the Nos. 1 and 2 starter-generators (Mfr. P/N 8260-121, S/N's 197 and 983 respectively) were tested on a test stand which allowed them to be run up to 12,000 rpm (maximum rated speed) and electrically loaded to 400 amps (100% rated load). No anomalies were noted.

Testing of the No. 1 generator control unit (GCU), Auxilec part number (P/N) 102003-11, serial number (S/N) 1151, per Auxilec Component Maintenance Manual (CMM) 24-32-11, revealed that resistor R3 and zener diodes CR3 and CR17, installed on the Z500 "Power" Printed Circuit Board (PCB) assembly1, had failed. Subsequent to the replacement of these components, the GCU was re-tested and functioned properly. Note: R3 and CR3 are used to reduce the input voltage to a 12 volt power supply internal to the GCU. CR17 is used to limit DC voltage within the regulator excitation circuitry. Testing performed at the manufacturer's facility showed that separation of the field excitation return wire from the "E" block of the starter generator would cause the failure of the above components in the GCU. The No. 2 GCU tested normally in post incident testing.

Inspection of the No. 1 engine wiring harness (Mfr. P/N S95404003VB32) revealed that the installation of the repaired field excitation return wire (P/N 2432-0090), which was repaired at Fort Myers, after the incident, prior to removal of the harness, was not in compliance with the latest modification (No. 1348) incorporated during the production of N275BC [i.e. the replacement terminal lug was unapproved and missing its protective sleeve (P/N NSA937493-001)]. The terminal lug which had initially failed was unavailable for examination. The aforementioned modification was developed by ATR due to operator reports of in-service breakage of wire /terminal lugs at their S/G connections. This modification replaced the original excitation return wire terminal lug with a stronger one, installed a protective sleeve, and required that the wire itself be secured to an adjacent main feeder cable. The general condition of the examined terminal lug was poor. The ring tongue was extremely deformed around its perimeter and was no longer flat.

Testing of the main battery (Saft P/N 4078-2, S/N 21444) was performed by Aviall, an FAA approved repair station located in Hollywood, Flo