On June 14, 1996, at 1057 eastern daylight time (edt), an Airbus A320, N347NW, operated as Northwest Airlines Flight 395, from Boston, Massachusetts, to Detroit, Michigan, experienced an uncommanded pitch excursion during climb. Neither the six crewmembers nor the 100 passengers were injured. The airplane was not damaged. The 14 CFR Part 121 flight returned to the Boston Logan International Airport, Boston, Massachusetts, without further incident. The flight departed from Boston, Massachusetts, about 1055 edt. In a written statement, the pilot reported that "shortly after takeoff, the first officer noticed that pitch response seemed sluggish. The trimmable horizontal stabilizer (THS) had a tendency to overcompensate for pitch inputs." The crew received a fault message of "ELAC 2 PITCH FAULT." The pilot reported a two to three degree "pitch down" occurred about 600 feet above ground level at the same time the message was received. "The THS seemed to overcompensate throughout the approach and landing." The crew declared an emergency and returned to Boston without further incident. During troubleshooting, the maintenance technician reported the rigging of the position transducer monitor was checked within three degrees. He reported the "right half of the stabilizer trim control pkg appears to be by passing fluid, (hot to touch) left side cool ck'd trim movement, appears slow compared to another a/c. Suggest stab trim pkg replacement." The THS actuator was replaced and operationally checked. The airplane was returned to service without further incident. Flight data recorder readout by the NTSB laboratory revealed an ELAC #2 pitch fault occurred as the airplane climbed through an altitude of approximately 500 feet. The pitch dropped from an attitude of about 17 degrees nose up to about 13 degrees nose up, oscillated through two cycles, then stabilized at 15 degrees nose up. Airbus Industries evaluated the flight data recorder data for this flight. The Airbus Industries Flight Safety Manager reported "no erratic THS movement could be observed at any time of this flight. An ELAC2 pitch fault occurred just after takeoff, but the reason of this fault could not be identified via the recorded data... . We highly suspect that the trouble was caused by either the electronic module or the pitch trim motor, both removed after the second flight." The THS actuator was tested on July 11, 1996, on the Airbus Industries iron bird flight control test laboratory. The Airbus Industries Flight Safety Manager reported that the actuator functioned normally. The pitch trim actuator servo motor was laboratory tested. No fault or discrepancy was observed. The electronic module was tested by Lucas Aerospace. The module was inoperative at voltages between 14.1 and 15.5. The discrepancy was attributed to a defective diode. An engineering study was conducted by Lucas Aerospace. An Airbus Industries engineer concluded, "despite a discrepancy on the diode characteristic the electronic module can be operative or not at low supply voltages (below 16 VDC). The engineering analysis have also confirmed that a defect of the diode characteristic could not affect the electronic module behavior. Taking into account that a drop of the electronic module supply voltages provided the A/C computer is very unlikely we can not conclude that the origin of the oscillations reported experienced on MSN 408 was originated by the electronic module... ." Examination of the airplane logbooks revealed an entry on June 13, 1996 "momentary/transient ELAC #2 pitch fault. FL 310 1 1/2 after t/off smooth flt conditions. Pitch wheel moved rapidly 1/8 to 1/4 turn pitch up to pitch down. A/P disconnected - OK changed to AP#1." Corrective action was listed as "replaced #2 ELAC... ." Airbus Industries evaluated the flight data recorder data for the flight on June 13, 1996. The Airbus Industries Flight Safety Manager reported "in cruise at 37,000 feet in calm air conditions, THS erratic activity counteracted by the elevators could be observed, prior to the loss of pitch control by ELAC #2 which led to the warning #ELAC2 PITCH FAULT#." A logbook entry on June 14, 1996 read "at cruise 370 rapid trim change (.8) to (1.3) up several times. No faults or disconnects. Autopilot #1 or #2 or hand flown. No change." Corrective action listed in the logbook included removal and replacement of ELAC #1, the 44CE relay, the electronic module, the THS actuator position monitor, and the pitch trim motor. Airbus Industries evaluated the flight data for recorder data this flight. The Airbus Industries Flight Safety Manager reported "during climb and cruise, the load factor appeared to be perturbed. The THS activity was identical to the previous described behavior. However in this flight the discrepancies between the THS order and the actual THS position did not reach the threshold for triggering the warning #ELAC2 PITCH FAULT#. (This warning is triggered if the discrepancy between the THS order and the actual THS position becomes greater than 0.6 degrees for at least 100ms)."