Casualties unknown

Boeing 767-200 accident at California, 28 Jun 2008 (N799AX)

California, US

On June 28, 2008, a Boeing 767-200 (registration N799AX) was involved in an aviation accident near California. Investigators recorded the probable cause as: The National Transportation Safety Board determined that the probable cause of this accident was "the design of the supplemental oxygen system hoses and the lack of positive separation between electrical wiring and electrically conductive oxygen system… This summary draws on records from the U.S. National Transportation Safety Board (NTSB) Aircraft Accident Reports; 8 related events involving the same aircraft type or operator are linked below.

Sourcesthe U.S. National Transportation Safety Board (NTSB) Aircraft Accident ReportsPrimary reportUpdated 1786183165Data APIEditorial standards
Boeing 767-200
Photo: Tosaka / CC BY 3.0, via Wikimedia Commons

Probable cause

The National Transportation Safety Board determined that the probable cause of this accident was "the design of the supplemental oxygen system hoses and the lack of positive separation between electrical wiring and electrically conductive oxygen system components. The lack of positive separation allowed a short circuit to breach a combustible oxygen hose, release oxygen, and initiate a fire in the supernumerary compartment that rapidly spread to other areas." The Board found that contributing to the accident was the FAA's failure to require the installation of nonconductive oxygen hoses after the safety issue concerning conductive hoses was initially identified by Boeing.

— NTSB Determination

Accident narrative

On June 28, 2008, about 2215 Pacific daylight time, an ABX Air Boeing 767-200, N799AX, operating as flight 1611, experienced a ground fire before engine startup at San Francisco International Airport in California. The airplane was parked near a loading facility with all cargo loaded and the doors shut. The captain and first officer evacuated through the cockpit windows and were not injured. The airplane was substantially damaged.

### The flight

The cargo flight was scheduled to depart at 2230 for Wilmington, Ohio, operating under 14 Code of Federal Regulations Part 121. Earlier that day, maintenance personnel performed a service check that included checking the supplemental oxygen system pressure. A mechanic reported the supernumerary oxygen bottle, which provided emergency oxygen to the cockpit, had a pressure between 1,220 and 1,230 psig, and the cockpit bottle had a pressure of about 1,560 psig.

The flight crew conducted preflight checks, including the supplemental oxygen system in the cockpit. The captain stated he pushed the oxygen switch, tested his mask, and verified the pressure, reporting that the system check was normal. The cockpit voice recorder (CVR) captured a sound similar to an oxygen mask test at 2200:07 and the first officer stating "mask check" at 2203:24.

### The fire

Less than one minute before the pilots heard loud "pop" and "hissing" sounds, the first officer had been in the supernumerary compartment—located directly aft of the cockpit and forward of the main deck cargo compartment—turning off lights. He reported no smoke or fire was visible at that time.

While performing the engine start checklist, the pilots heard the pop and hissing sounds. About 15 seconds later, the first officer opened the cockpit door and looked inside the supernumerary compartment. He observed black smoke at ceiling height and saw fire near the ceiling above the right-most occupant seat and the weight and balance computer. Between 2210:28 and 2210:41, the first officer stated, "hey, there's something going on in the back... we got a fire... got a big fire." The CVR then recorded sounds similar to the lavatory smoke detector alarm and a fire warning bell.

### Evacuation and emergency response

At 2211:04, the first officer contacted the ground controller to request aircraft rescue and firefighting (ARFF) for a cargo fire. The controller confirmed the report and stated at 2211:24, "we're gonna roll the trucks right now." The pilots began the fire and evacuation checklist.

The normal exits available to the flight crew were an entry door on the left side of the supernumerary compartment and a service door on the right side. Because of the location and intensity of the fire, the flight crew evacuated through the two cockpit window exits using escape ropes, with the captain utilizing stairs positioned by ground personnel.

Twenty-two ARFF personnel responded, with the first vehicle arriving by 2215:14. Firefighters could not immediately access the fire because the forward doors were inoperable; the fire had melted the mechanisms and tracks that allowed the doors to open upward. The driver of the first vehicle attempted to use a high-reach extendable turret with a skin-penetrating nozzle through the right cockpit window, but the extinguishing agent sprayed outside the airplane. After the vehicle was repositioned near the left door, the driver successfully used the nozzle to apply agent through burn-through areas above the supernumerary compartment. The fire was contained about 25 minutes after the first vehicle arrived and extinguished about 43 minutes after arrival.

The driver stated he had received familiarization training on the nozzle 14 years prior and practiced on cars and vans. The Board concluded the type of training he received was not sufficient to allow him to successfully insert extinguishing agent on his initial attempts, and that personnel not sufficiently trained on the device may not be able to use it effectively.

### What the investigation found

The fire completely consumed the supernumerary compartment seat cushions and most of the seat structure. Most ceiling panels were consumed, and large burn-through areas were located on the crown skin. The fire spread forward into the cockpit, consuming the separating bulkhead, and aft into the main deck cargo compartment after melting aluminum strips on a smoke barrier wall.

The brief time in which the fire developed, its intensity, and the pop and hissing noises indicated a source of pressurized oxygen was involved. The supernumerary oxygen bottle was found empty, while the cockpit bottle was found almost full.

Above the supernumerary seats, three oxygen mask stowage boxes each included a flexible hose made primarily of polyvinyl chloride (PVC) and rigid stainless steel supply tubing. The PVC hoses contained a loosely attached stainless steel coil spring to prevent bending, making the hoses electrically conductive.

The National Transportation Safety Board performed tests on similar PVC flexible oxygen hoses. When pressurized with oxygen and subjected to an electrical current, the internal spring heated. At low energy levels, the PVC softened and ruptured. At higher energy levels, the spring became an ignition source, causing the flexible hose to ignite and sustain a fire. Time to failure ranged from 6 to 180 seconds. Each ignition caused by heating the internal spring was preceded by a loud pop sound, followed by a loud hissing sound during sustained burning. Tests using an external heat source produced hissing and an intense fire, but no pop sound. The Board concluded the sounds heard by the crew were consistent with the ignition of an oxygen hose by an internal rather than external heat source.

In the wreckage, investigators found a stainless steel 90-degree elbow assembly from an oxygen stowage box melted into two pieces. Aluminum fittings attached to its ends showed only incipient melting. Because stainless steel melts at 2,500°F and aluminum at 1,200°F, the Board noted this indicated a highly localized oxygen-fuel flame occurring before heat could conduct to the ends, with the flexible oxygen hose as the source.

Boeing guidance stated oxygen installations should include a minimum of 2 inches between oxygen lines and electrical wiring. However, inspections of other ABX Air 767s found electrical wiring in contact with or routed near the stainless steel oxygen supply tubing. The Board concluded the fire most likely began when a combustible and electrically conductive oxygen hose became energized by a short circuit, causing the hose to ignite and burn through, releasing oxygen that caused adjacent materials to burn at an accelerated rate.

### Previous knowledge and maintenance

Boeing had previously received reports of electrical energy causing leaks in 737 and 757 oxygen hoses. In 1999, Boeing began installing nonconductive hoses on new airplanes and issued alert service bulletins recommending operators replace existing cockpit hoses. However, the Federal Aviation Administration (FAA) did not mandate compliance with airworthiness directives. The accident airplane was converted to a cargo configuration in 2004 by Israel Aerospace Industries, which adapted original Boeing-numbered parts, including the conductive flexible hoses. The Board concluded it is likely the modifier would have recognized the use of a potentially combustible part if the FAA had issued an appliance airworthiness directive citing the part manufacturer.

Maintenance records showed the airplane's oxygen system was serviced 50 times between January 2007 and June 2008, indicating a chronic leak. ABX Air's reliability department issued four advisory notices for the oxygen system during this period. However, the company's continuing analysis and surveillance program only reviewed a rolling 3-month period, which prevented them from recognizing the total accumulation of discrepancies. The Board concluded the number of discrepancies was excessive and that the surveillance program did not include adequate actions for resolving them. While no evidence showed previous leaks contributed directly to this fire, the Board noted oxygen leaks are a safety hazard.

The investigation also found that the supplemental oxygen system was grounded to the airplane structure through the PVC hoses and a convoluted hose, creating an undesirable ground path that could allow adjacent wiring to energize components. Additionally, passenger service unit reading lights in the supernumerary compartment lacked rubber boots to isolate electrical parts, creating a potential short circuit hazard. This was ruled out as the ignition source because external sparks do not produce the popping sound heard by the crew, but the Board noted it as a hazard. The Board also concluded that installing smoke detectors in supernumerary compartments would help flight crews identify fires in accessible, possibly unoccupied spaces.

### Probable Cause

The National Transportation Safety Board determined that the probable cause of this accident was "the design of the supplemental oxygen system hoses and the lack of positive separation between electrical wiring and electrically conductive oxygen system components. The lack of positive separation allowed a short circuit to breach a combustible oxygen hose, release oxygen, and initiate a fire in the supernumerary compartment that rapidly spread to other areas."

The Board found that contributing to the accident was the FAA's failure to require the installation of nonconductive oxygen hoses after the safety issue concerning conductive hoses was initially identified by Boeing.