Incident Overview
A Boeing 767 aircraft was in cruise flight at flight level 370 en route from La Guardia Airport, New York, to Zurich, Switzerland, when the flight crew detected an electrical smoke smell in the cockpit. Upon donning oxygen masks and smoke goggles, thick acrid smoke suddenly emanated from behind the glareshield on the first officer's side, accompanied by a very hot heat source resembling an "arc welder" and a forward equipment overheat warning. The crew declared an emergency and requested the nearest airport from Moncton Centre, which advised Sydney Airport. An emergency descent was initiated, and the in-charge flight attendant was instructed to prepare the cabin for landing. The crew later informed the attendant that an emergency evacuation was not anticipated as no further smoke entered the cockpit after the initial puff. The aircraft landed at Sydney without incident, with emergency response services positioned, 23 minutes after the smoke was first detected. The aircraft stopped on the runway; after emergency responders confirmed no fire, it taxied to the ramp where passengers deplaned orderly through the main cabin door.
Cockpit Indications and Crew Actions
Shortly after the arcing fire and smoke began, an Engine Indication and Crew Alerting System (EICAS) message "Forward Equipment Overheat - FWD EQUIP OVHT" was annunciated. The crew selected Standby on the equipment cooling switch per the emergency checklist; the arcing stopped and the EICAS message disappeared. No EICAS message regarding window heat was displayed during the flight, so no related checklist item was required. The smoke cleared, and the crew, suspecting possible fire in the forward equipment bay under the cockpit floor, continued descent for landing at Sydney. Electrical power to the window heat was not shut down until after landing during normal shutdown procedures.
Post-Flight Examination
Preliminary examination revealed no fire or smoke damage in the forward equipment bay. However, a terminal block on the lower outboard corner of the right (first officer's side) front windshield was extensively heat damaged, and the inner glass ply of the window was cracked. The circuit breaker for windshield heat had not tripped. The damage was not readily apparent until closer ground inspection. The damaged windshield, the right window heat control unit, and the cockpit voice recorder were sent to the TSB Engineering Branch for examination.
Crew Performance
Interviews and cockpit voice recording review indicated excellent coordination among flight deck crew and between the flight deck and the in-charge flight attendant. Flight attendants were in the middle of meal service; the captain commended their performance in keeping passengers calm and preparing the cabin for landing in the short time available.
Component Details
The window heat control unit (part number 624066-5, series 1, serial number 100C-1815, manufactured by Garrett Canada on 11 October 1990) controls three channels for the right forward windshield and two left-side windows. It has a self-test feature but no electronic memory to store faults after power removal. The unit was maintained on-condition and had never been returned for repair/overhaul. Testing found it passed all procedures except a minor ramp-up voltage deviation (35 VRMS at 10 seconds vs. 40 VRMS low limit), considered not problematic and requiring only a minor potentiometer adjustment. The unit was deemed serviceable.
The right windshield (serial number 92154H2287, manufactured by PPG Industries) was installed in December 1992, accumulating 14,283 hours and 2,358 cycles. No maintenance had been carried out on the windshield or wiring during that time.
Examination Findings
The windshield was shipped to PPG's facility for teardown with representatives from PPG, American Airlines, Boeing, and the TSB. The inner glass ply showed melted glass and a crater-like depression directly under the J5 terminal, with cracks radiating outward. Glass melting point is 968°C. The power connector (Wallace-Black) had a cadmium-plated steel insert; the epoxy connector was cracked but not grossly heat damaged. The steel insert had resolidified cadmium (melting point 321°C) on its surface, but no arcing or pitting. The cap, screw, and lock washer were not recovered, but a pre-removal photograph showed the cap blackened and wiring insulation charred.
The J5 terminal (PPG number 22-17-1385) consisted of a cast epoxy block with a threaded brass block, extensively heat damaged and cracked. Two flat-braided copper conductors were soldered to the brass block; no braid remained, only solder residue. No arcing pitting was found on the brass. The cavity where braids attach was filled with polysulphide adhesive (PR1425) containing carbon black; most adhesive was missing or ash. The burned window sill edge revealed about half an inch of medium copper-braided conductor (typical length 2.5 inches). Some braid ends were melted, typical of arcing (copper melts at 1,080°C).
Related Service History
Boeing reviewed service history and found three similar events between 1992 and 1993. In response, Boeing issued a message to operators in 1993 requiring increased torque on screws holding power conductors to window terminals. Maintenance personnel who removed the windshield stated the screw on the burned terminal was still tight.
Testing and Analysis
PPG tested a power terminal and J5 terminal in a 20-ampere circuit: with screw fully tight, temperature rose 5.6°C above ambient (25.6°C); when backed off to visibly loose, temperature rose 27.8°C above ambient. This rise was not considered significant, and the tight screw indicated loose screw did not initiate overheating.
The copper braid length from window edge is covered with heat shrink removed before soldering, leaving 0.5 to 0.75 inches exposed. Nicking during this operation could allow strands to flex and separate. PPG tested a heavier braid in a 32-ampere circuit: deliberately cutting the braid from 0% to 75% cut-through increased temperature 11.1°C; at 98% cut-through, temperature rose 150°C to 226.7°C. No arcing occurred during that test, indicating complete separation is needed for arcing.
The lack of pitting on the brass block and remaining solder suggests arcing did not occur at that interface. Sufficient heat melted the solder, detaching the braid from the block and melting the cadmium plate. Melted ends of copper braid in the window sill are consistent with arcing observed by the pilot, strongly indicating arcing occurred at or near that location behind the glare shield. High temperatures from arcing caused glass cratering and cracking. The loss of braid between terminal and remaining section precluded definitive findings on arcing cause, but it was speculated that nicking during construction might lead to flexing and eventual failure.