The ballast that initiated the investigation, Day-Ray Products part number 70-94, is used in the DC-9 and MD-80 for upper and lower cabin sidewall fluorescent lighting and in the DC-10 for primary cabin lighting. The ballast are installed on top of the passenger luggage stowage bins throughout the cabin.

The ballast provides full- range dimming of two 30-watt fluorescent lamps from a 230-volt alternating current, 400-hertz source. It is approximately 8 inches long, 2 inches wide, and 1 inch tall, with a phenolic case over all of the circuitry. Other ballasts in this family control single 30-watt lamps, dual 20-watt lamps, and single 20-watt lamps. The interior of the ballast is almost completely encapsulated with a silicone elastomer potting compound surrounding the various electrical components.

MOISTURE THEORY

An investigation of two cabin fires in 1995 pointed to the ballast as a probable cause. To prove that the ballast caused the fires, a Douglas investigation team was required to devise a testing procedure that forced an unknown failure mode of the ballast before its own internal circuit Protectors could activate and disconnect it from the power source.
There are five different circuit protectors inside the ballast: three thermal fuses located on different components of the ballast, and two current fuses for two levels of over-current protection.

The investigation team theorized that a failure of the ballast might be caused by ingested moisture accumulated inside the ballast. Naturally occurring condensation in the cabin, a phenomenon that is often referred to as "rain in the plane," was suggested as the source of the moisture.
TESTING PROCEDURE

Ballasts were soaked in different mixtures of fresh and salt water. The soaked ballasts were subjected to pressure cycles similar to those produced when aircraft fly at altitude and then land at normal atmospheric pressure. The pressure cycles force water into any voids of the ballast potting compound, creating the potential for wetting the electrical components. The ballasts were connected to a cabin lighting test fixture and power was applied.

None of the soaked ballasts failed upon initial power application in the test program. They did, however, function anomalously in various ways: the lamps connected to the ballast flickered; the ballasts did not respond to the dimming control; and the ballast case became unusually hot, occasionally hissing steam and emitting bubbles that originated from within the case.

TEST RESULTS

After persistent testing, in early January of 1996, one ballast suddenly made a loud pop and produced a small flame that appeared through a rupture on top of the phenolic case. The ballast failure damage was consistent with the failure damage of ballasts suspected of causing the cabin fires. Further testing over a period of several weeks resulted in additional failures.

Test results were shared with the Federal Aviation Administration (FAA) and the National Transportation Safety Board (NTSR). Douglas invited foreign and U.S. regulatory officials and airline representatives to a demonstration of the failure testing. Two ballasts produced flames during the demonstration, providing evidence that the ballast caused the cabin fires.

Expanded testing showed that a total of eight ballast part numbers, all of the same family made by the same manufacturer, exhibited the same failure mode. The ballasts differ mainly in the size and number of lamps they power.

CORRECTIVE ACTION

Douglas issued Service Bulletins DC9-33-103, MD80-33A-107, and DCI0-33-073 in the second quarter of 1996 to provide operators with corrective action. Day-Ray Products also issued an alert service bulletin.

This Service Bulletin has been perform in Spanair MD Fleet.

The FAA issued Airworthiness Directive 96-11-13 to reinforce corrective action for the ballasts on the MD-80. The FAA has released a notice of proposed rule making (NPRM) for an airworthiness directive to preclude an uncontained failure of the ballasts throughout the industry. The action ensures the protection of all aircraft using any of the subject ballasts.

To ensure safer ballasts on all Douglas aircraft, the failure testing discussed in this article has become part of Douglas ballast procurement specifications.