Douglas DC-3 accident at Michigan City, Indiana, 28 Dec 1946

Accident narrative

On December 28, 1946, American Airlines Flight 2207 crashed in the northeastern part of Michigan City, Indiana, about two miles from Lake Michigan. The accident occurred at about 0919 CST. Both pilots were killed, the eighteen passengers were injured in varying degrees, and the Douglas DC-3 was extensively damaged.

**The flight**

The flight was en route from Buffalo, New York, to Chicago, Illinois. After a scheduled stop in Detroit, Michigan, the aircraft departed at 0748 on an instrument flight plan to cruise at 2,500 feet. The DC-3 carried 550 gallons of fuel: 210 gallons in each main tank, 130 gallons in the left auxiliary tank, and none in the right auxiliary tank. The crew consisted of Captain Frank Mapes Ham, Jr., and First Officer Harman Edwin Ring.

**Sequence of events**

The flight made routine position reports at 0800, 0807, and 0822. At 0840, Chicago Air Route Traffic Control instructed the flight to climb to 4,000 feet, which it did immediately. Chicago later cleared the flight from South Bend, Indiana, to the Chicago Tower, instructing the crew to maintain 4,000 feet until ten minutes after passing South Bend. The flight reported passing South Bend at 0902.

There were no further radio contacts until 0918, when the flight transmitted, "Both engines going bad descending at 1,000 feet per minute looking for opening." This was followed almost immediately by, "Altitude is at present 900 feet." Chicago suggested the flight proceed to South Bend immediately if able, but this was not acknowledged. The plane crashed at or about that time. The direction of initial impact was just south of east, approximately opposite the course from Detroit to Chicago.

**What the investigation found**

Weather at the time included an overcast with its base at about 1,000 feet and its top at about 4,000 feet. Within the overcast, conditions were favorable for light to moderate icing. Passenger testimony confirmed that wing ice formed while climbing through the overcast at Detroit and while descending through it just before the crash. Passengers also stated that power impairment occurred only a very few minutes before the crash, while the aircraft was cruising above the overcast. Witnesses on the ground testified that the engines were running improperly or not at all from the time the plane emerged from the overcast until it crashed.

Examination of the wreckage showed no structural failure in flight. The wheels and flaps were not extended. Fuel valve settings indicated the left engine was feeding from the left main tank and the right engine from the left main and left auxiliary tanks, though the Board noted these positions did not necessarily reflect their settings before impact.

Investigators found alcohol at the fittings to both carburetor air scoops, indicating it had been applied to dispose of carburetor ice. Heat controls indicated that heat had been applied to the right carburetor but not the left.

The investigation revealed a small piece of masking tape or a similar substance wedged in the right engine tank selector valve, at the port for the empty right auxiliary tank. The material was squeezed to a thickness of about .005 inch, holding the valve open to that extent. Tests showed that while a .005-inch opening did not cause engine stoppage, a larger opening could stop the engine on the same side as the empty tank. The tests indicated that air flowing from an empty tank into the fuel system could cause irregular power output and delay the power recovery of a stopped engine.

The Board noted the possibility that the pilot accidentally exhausted the left auxiliary tank before changing to the main tanks. If this occurred, air entering the fuel system through the fouled valve may have delayed power restoration until the plane lost enough altitude to enter the icing overcast.

While carburetor ice was improbable at the 4,000-foot cruising altitude, improper use of carburetor air preheat could cause ice to form. A subsequent loss of engine power would rapidly decrease the preheat temperature, allowing rapid ice accretion. Although the pilot applied alcohol, the Board noted it takes approximately three minutes to dissipate such ice, during which a powerless DC-3 would descend about 3,000 feet. The Board believed that carburetor ice impeded power recovery, and that improper use of carburetor air preheat must have been involved.

**Probable cause**

The Civil Aeronautics Board determined the probable cause of this accident was "the accumulation of carburetor ice following the loss of power in both engines as a result of fuel starvation. The reason for fuel starvation has not been determined."