NTSB Factual Report

HISTORY OF FLIGHT

On June 13, 2003, about 1730 Pacific daylight time, a Bombardier CL-600-2B19, N97325, operating as Air Shuttle flight 6277,experienced a loss of horizontal stabilizer trim command response during the initial climb out from Phoenix Sky Harbor International Airport, Phoenix, Arizona. The captain declared an emergency and manually flew the airplane back to Phoenix Sky Harbor International Airport. Mesa Airlines, Inc., operated the airplane as a scheduled domestic passenger flight under the provisions of 14 CFR 121. The airline transport pilot rated captain, the copilot, the single flight attendant, and 49 passengers were not injured. The airplane was not damaged. Visual meteorological conditions prevailed, and an instrument flight plan had been filed. The flight originated at Phoenix Sky Harbor International Airport about 1613.

The first officer told the National Transportation Safety Board investigator that the captain was at the controls, and they were passing through 23,000 feet for flight level 290 (29,000 feet). The stab trim/mach trim caution light illuminated on the instrument panel indicating that both channels 1 and 2 were inoperative. The captain maintained the airplane's airspeed below 215 knots, and diverted the flight back to Phoenix Sky Harbor Airport. To maintain the airplane in the proper attitude the first officer said that about 30 pounds of backpressure was required on the yolk to counteract the trim forces. The crew exchanged control of the airplane at least once during the 25-minute return flight in order to rest. The captain landed the airplane uneventfully.

Mesa Airlines' maintenance control traced the trim failure to the captain's stab trim disconnect switch, which they suspected stuck in the cutout position. Maintenance replaced the switch and sent the component to the Safety Board investigator. The Digital Flight Data Recorder (DFDR) was removed and sent to the Safety Board's Vehicle Recorder Division, Washington, DC.

AIRCRAFT INFORMATION

The airplane was a twin-engine commuter jet with a cockpit crew of 2, a flight attendant, and a passenger capacity of 50 seats. The airplane's serial number was 7325, and examination of the maintenance records revealed a total airplane time of 11,087.2 hours at the time of the trim failure event. A routine line check had been performed 22 flight hours earlier, on June 11th, which included an operational check of the horizontal stabilizer trim system. There were no discrepancies noted on the service check documentation. Further examination of the maintenance records for the previous 6 months of operation did not reveal any stabilizer or mach trim discrepancies. On December 24, 2002, a discrepancy was reported that the mach trim and stab trim disconnected on landing; the Horizontal Stabilizer Trim Control Unit (HSTCU) was replaced on that date.

A second trim system event occurred on July 15, 2003, in the same airplane (sn: 7325), on a flight from Billings, Montana, to Phoenix. The captain reported that during a descent from flight level 290, both channels of the stab trim and the mach trim failed with a corresponding "AP Pitch Trim" message. After about 2 minutes, the captain was able to reset both stab trim channels and then the mach trim. The airplane landed in Phoenix uneventfully. The maintenance crew removed and replaced the HSTCU to correct the discrepancy. The HSTCU was sent to the Safety Board investigator for examination. The DFDR was removed and sent to the Safety Board's Vehicle Recorder Division, Washington, DC.

The maintenance records were reviewed after the July 15th event. The airplane total airframe time was 11,330.6 hours. No stab trim or mach trim discrepancies were recorded for the period between June 13th and July 15th.

Horizontal Stabilizer Trim Control System (HSTCS)

The horizontal stabilizer trim control system (HSTCS) supplies pitch trim and operates in one of four modes; manual (pilot, copilot) trim, autopilot trim, auto trim, and mach trim. A Horizontal Stabilizer Trim Control Unit (HSTCU) receives all of the above inputs and signals a motor control unit (MCU), which controls the functions of the Horizontal Stabilizer Trim Actuator (HSTA). The HSTCU is supplied by two independent airplane electrical power sources, and contains two channels, channel 1 and channel 2. It also contains BITE (built-in-test-equipment) capability. The MCU is an electronic unit that controls the speed and the direction of the motors in the HSTA. The HSTA is an electromechanical unit that controls the horizontal stabilizer surface position in response to electrical signals from the MCU. The pilot and copilot disconnect is a double-pole, normally closed, momentary action switch located on the outboard top end of the control wheel beside the manual trim switch. The switches are normally closed without a disconnect command. The HSTCU is configured such that either pole must open to disconnect the system. A disconnect command disengages the HSTCU, and removes electrical power from the HSTA.

FLIGHT RECORDERS

The Flight Data Recorder (FDR) for the event was removed from the subject airplane by the operator and received by the Safety Board's Vehicle Recorder Division on June 18th. The recorder was undamaged, and the data was extracted normally from the recorder. The FDR recording contained approximately 51.3 hours of data. Timing of the data is to the nearest second, referred to as FDR Subframe Reference Number (SRN). The incident flight was the last flight of the recoding, and began at SRN 181088. The duration of the incident flight was approximately 1 hour. The stab trim failure event occurs about 17 minutes into the flight at SRN 182104. Near this SRN the recording shows the pitch trim stabilizer at -1.4 degrees and the stabilizer trim discrete changes from state of 'on' to 'off.'

The FDR for the second occurrence (July 15th ) was removed from the airplane by the operator and received by the Safety Board's Vehicle Recorder Division on July 18th. The recorder was undamaged, and the data was extracted normally from the recorder. The FDR recording contained approximately 51.2 hours of data. The incident flight was the last flight of the recording, and began at SRN 176442. The duration of the incident flight was approximately 2 hours. The stab trim failure event occurred 24 minutes prior to landing at SRN 182700. Near this SRN the recording shows the stabilizer trim discrete state change from 'on' to 'off.' At SRN 182815 the stabilizer trim discrete state changed from 'off' to 'on.'

The entire Flight Data Recorder Specialists Factual Report is contained in the official docket of this investigation.

TEST AND RESEARCH

Captain's stab trim disconnect switch

The stab trim disconnect switch was removed from the captain's control yolk by the operator and sent to the Safety Board investigator. The stab trim disconnect switch is a dual channel switch operated by a single thumb plunger button. The single thumb plunger engages two separate switch plungers. Each switch is completely independent and operates a single trim channel. Each switch in its normal at rest position is closed, completing the circuit. When the switch plunger is depressed the switch opens, breaking the circuit. The single thumb plunger, therefore, opens both switches (circuits) when it is depressed. The switch was taken to the manufacturer, Mason Electric Company, San Fernando, California, for examination and testing under the supervision of the Safety Board investigator on July 1, 2003. The switch was tested for authenticity, connectivity, and functionality. Mason engineers verified that the switch was a Mason product, part number: 996-0013, manufactured in 1998, week 48. The switch was functionally cycled over 10 times and required a 0.75-pound actuating force. Electrical connectivity/resistance tests revealed that the switch opened and closed both circuits on each cycling of the switch. Mason engineers did not identify any discrepancies with the switch.

Horizontal Stabilizer Trim Control Unit (HSTCU)

The operator removed the HSTCU (part number: 7060-7, s/n: 412), and sent it to the Safety Board investigator. HSTCU serial number 412 was installed on the airplane during both trim failure occurrences. On August 8th, the HSTCU was taken to the manufacturer, Sfim Incorporated, Grand Prairie, Texas, for examination and testing under the supervision of the Safety Board investigator. Technicians physically examined the HSTCU noting that the access panel door cam type fastener shoulder guide was missing from one of the two door latching fasteners. They found the shoulder half loose inside the HSTCU circuitry area of the box. They performed a BITE (built-in-test equipment) test successfully with no anomalies. They then connected the HSTCU to a test bench that contained a calibrated Motor Control Unit (MCU) and Horizontal Stabilator Trim Actuator (HSTA). The test bench effectively reproduces the airplanes stab trim system. They subjected the HSTCU to an acceptance test profile (ATP2) that Sfim uses to test newly manufactured HSTCU's. They performed the test at room temperature, and identified no discrepancies. They then disassembled the HSTCU, and examined the circuitry with no discrepancies identified. Sfim received approval from the Safety Board investigator to continue to use the HSTCU as their test bench article and report any further anomalies. On August 15th, Sfim tested the HSTCU at 53 degrees C and at 0 degrees C, and reported that there were no discrepancies. Sfim continued to use the HSTCU as their test bench article with no anomalies noted through October 3, 2003.

Sfim engineers described the electrical circuit boards as having a heavy coat of protective/sealing 'varnish' that would protect the circuits from shorts caused by debris. Additionally, they stated that each trim channel is on a separate circuit board in the HSTCU, and the simultaneous failure of both trim channels due to a single piece of debris would be highly unlikely.

Service Difficulty Report (SDR) Data

The Safety Board IIC requested a search of the Federal Aviation Administration Service Difficulty Report (SDR) database for Stab Control System anomalies in the Canadair CL-600 from 1995 to July 10, 2003. The flight control ATA (Airline Transportation Association of America) codes that were searched were 2740-stabilizer control system, 2741-stabilizer position indicating, and 2742-stablizer actuator. The search identified 67 dual channel stab trim failures that were reported to the SDR database. Twenty-seven percent of the dual trim failures were identified as cockpit trim disconnect switch issues; 52 percent were identified as HSTCU/HSTA/MCU issues; and 21percent were not resolved or could not be duplicated on the ground. The airplane in question, serial number 7325, was not listed in the results as having any reported flight control discrepancies in the database.

CRJ 700/900 Trim Disconnect Issues

Dual trim failures similar to this incident have been identified in the CRJ 700/900 series aircraft. Bombardier has isolated the anomaly in the CRJ 700/900 to the Hold Relay Monitor, which is not robust to very short stab trim disconnect signals. The CRJ 700/900 contains two Spoiler/Stabilizer Control Units (SSCU) each having two channels, while the CRJ 200 (CL-600-2B19) has one HSTCU with two channels. The SSCU has significantly newer technology, and has advanced operating and monitoring software. The MCU on the CRJ 700/900 is also advanced and monitors the health of the HSTA itself, and reports the status to the SSCU. The CRJ 200 MCU is merely a relay station that receives commands and provides power out to drive the motors. Due to the differences in the hardware and software between the CRJ 700/900 SSCU and MCU, and the CRJ 200's HSTCU and MCU, Bombardier engineers determined that the trim failure issues in the incident airplane and the trim disconnect issues in the CRJ 700/900 series airplanes are not related.

NTSB Probable Cause

the simultaneous failure of both horizontal stabilizer trim channels on two separate occasions for undetermined reasons.