On June 20, 2005, about 1205 eastern daylight time, a Piper PA-31-350, N615SN, was substantially damaged during a forced landing shortly after takeoff from Beverly Municipal Airport (BVY), Beverly, Massachusetts. The certificated private pilot was not injured. Visual meteorological conditions prevailed and no flight plan was filed for the local maintenance test flight conducted under 14 CFR Part 91. According to the pilot, he was test flying the airplane after recent maintenance was performed on the engines. During initial climbout from runway 16, the left engine began to surge, and most of its power was lost. The air traffic control tower radioed to the pilot that heavy black smoke was trailing from the left engine, and that he was cleared to return to the airport and land on any runway. The pilot initiated a left turn, and continued to climb to approximately 600 feet. He then initiated a "short approach" to runway 16 because he believed that the left engine might have been on fire. The airplane touched down "hard" in a grass area to the right of the runway, rolled across a taxiway, and sheared off the landing gear, before coming to rest. The pilot additionally stated on the NTSB Form 6120.1/2, Recommendation (How Could This Accident Have Been Prevented): "Possibly more single engine training, and on a regular basis to stay more proficient." Examination of the left engine by an FAA inspector revealed that the spark plugs had "more than the normal carbon deposits on them"; however, no carbon bridging was noted. The spark plugs were tested and operated with no abnormalities noted. No chaffing of the magneto p-leads was observed. The fuel pressure to the engine was tested and found to be within nominal limits. The upper deck air pressure manifolds, air fittings, and flex lines, were absent of debris. The engine was test run in the presence of an FAA inspector. The inspector observed that on initial startup, the engine smoked for a few minutes, then the smoke cleared, and the engine subsequently ran with no other abnormalities noted. The engine was also run on both magnetos, then just the left magneto, and finally the right magneto, with no abnormalities noted. After the test run, the spark plugs were removed and examined. Carbon was observed on the electrodes; however, it was less than what was observed prior to the test. The engines had accumulated 2.2 total hours of operation since overhaul at the time of the accident.