BOEING 767

Full Narrative

On June 16, 2010, at 1330 UTC, Delta Air Lines flight 139, a B767-332ER, registered as N196DN, experienced a cockpit fire during preflight preparations at Eleftherios Venizelos International Airport (LGAV). The fire occurred prior to passenger boarding, and there were no injuries to the 3 flight crew and an unknown number of cleaning crew personnel on board the aircraft. During the preflight process, ground maintenance personnel performed a "push to test" on the right sidewall accessory panel (P61). After an undetermined period of time, while the first officer continued with his preflight procedures, he smelled smoke and then observed a flame coming out from the right sidewall accessory panel. He extinguished the fire by using a halon fire extinguisher. The investigation was delgated to the NTSB by the Hellenic Air Accident Investigation and Aviation Safety Board.

The panels involved with the fire (the squib test panel and the miscellaneous test panel) were removed and sent to Boeing's Engineering Quality Analysis (EQA) laboratory in Seattle, WA. On September 16, 2010, the panels were examined under the direction of the NTSB.

The existing condition of the panels submitted for examination was documented, tests were conducted where possible, and chemical analyses were conducted on some samples. The damage on the squib test panel was most severe around the light in position L3.

During tests of the L3 light, the resistance between terminals 2 and 3 varied from 0.6 ohms to 4.6 ohms. Based on the damage to L3, it could not be determined if a short initiated as a result of physical damage, a conductive path of contaminants, a workmanship error that may have been created when L3 was removed and replaced, or a manufacturing defect with the lamp itself.

While only 14 ohms were measured between terminals 1 and 2 of L3, it is believed that this short occurred after the press to test was performed per the pre-flight process and a result of the heat generated from the short between terminals 2 and 3. If this value actually existed during the pre-flight, the current would have been significant enough to discharge the fire bottles. The fire bottles were not discharged during the incident.

Switch S5 on the miscellaneous test panel showed evidence of heat damage on the bottom of the switch. Switch S5 on the miscellaneous test panel is almost directly above light L3 on the squib test panel.