On June 28, 2013, about 1930 Pacific daylight time, a Quicksilver MXL-II light sport airplane, N2812, was substantially damaged during a forced landing following a partial loss of engine power near Newbury Park, California. The light sport pilot and his sole passenger sustained serious injuries. Visual meteorological conditions prevailed for the local flight, which was conducted in accordance with 14 Code of Federal Regulations Part 91, and a flight plan was not filed. The flight departed the Camarillo Airport (CMA), Camarillo, California, about 1900.In a post-accident interview with the National Transportation Safety Board investigator-in-charge, the pilot reported that the purpose of the flight was to check the operation of the engine, as it had been missing and running rough on a previous flight. The pilot stated that during the flight the engine began to run rough, and when he could not obtain an increase in power with the throttle and extra electric fuel pump, he elected to make an emergency landing. The pilot added that after landing on uphill sloping terrain, the airplane nosed over and came to rest inverted. On July 18, 2013, under the supervision of a Federal Aviation Administration aviation safety inspector, a Rotax field technician performed a detailed examination of the engine. The engine was a Rotax 503, dual carburetor, dual ignition, fan cooled engine, equipped with electric and pull start systems. The engine is a twin cylinder, 2 stroke engine with a tuned exhaust and a reduction gearbox. The examination revealed that the engine remained attached to the airframe, and that the exhaust muffler had been previously removed from the exhaust Y-pipe and placed on the floor next to the aircraft. The Warp Drive propeller was also previously removed from the engine and placed next to the aircraft. The exhaust system was the correct type but was in very poor condition with previous weld repairs and several cracks were observed. The technician noted during the investigation, while Rotax does not state that the exhaust muffler cannot be welded or repaired if cracks are found, they also do not supply instructions on how to complete a repair of this nature. The technician added that if an exhaust system had an existing crack which was obviously leaking exhaust, and in addition had multiple prior cracks repaired, this should be an indication that it is time to replace the exhaust system. The technician referred to the Rotax Operators Manual, Section 10.3.3, Pre-Flight Checks, page 10-13, which provides the warning "…to repair as necessary all discrepancies and shortcomings before flight." Further, it states to inspect exhaust for cracks, security of mounting, springs and hooks for breakage and wear, and to verify the safety wiring of springs. The technician reported that during handling of the exhaust it sounded like there may have been broken baffles inside the exhaust system. Any broken internal baffles would affect the running condition and reliability of the engine, as they could block the exhaust system resulting in an immediate engine failure or loss of power. The technician further reported that after the exhaust system was cut open the internal baffles were examined. It was observed that the exhaust system was found to be very poor condition, with several broken internal components and baffles. The technician stated that the broken baffles and components can shift their position depending on the flight attitude of the aircraft and block the exit of the exhaust, which could result in an immediate engine failure or loss of power. The technician concluded that the cause of the loss of power was due to a blockage of the exhaust system due to broken engine baffles that had shifted from the original location during flight. The technician added that the internal and external examination of the exhaust system revealed that the system was in very poor condition and was not suitable for flight. A review of the engine logbook revealed that the engine was installed on the aircraft on February 12, 2013, at a Hobbs reading of 2,268.3 hours. On June 23, 2013, five days prior to the accident and at a Hobbs time of 2,323.7 hours, the engine's spark plugs and exhaust gaskets were changed. The pilot reported in a follow-up interview with the NTSB IIC that the exhaust system had been inspected for cracks about 20 hours prior to the accident flight, with none observed. The pilot stated that the system was completely removed and then reinstalled after the detailed inspection had been performed. He also stated that with respect to the welds discussed in the Rotax investigation report, they were not done subsequent to his club purchasing the system, and that he had no idea of when or where the repair took place, or by whom they were performed. The pilot added that when an anomaly with the exhaust system is detected, it is removed from the aircraft and replaced. The engine was subsequently test run. In preparation, and as the aircraft's fuel tank was breached, a secondary container was used with the same fuel that was captured from the original fuel tank. The aircraft's fuel line was then placed inside of the container to supply fuel to the engine. The battery was connected in the original location, which allowed the ignition system to be used during the test run. The original exhaust muffler was not installed for this engine test run, nor was an exemplar exhaust muffler installed. The aircraft's electric fuel pump was turned on and fuel could be seen traveling through the fuel line to the carburetors. The magneto switch was turned on and the starter was engaged. After several rotations the engine started and ran without issues. The RPM was manipulated with use of the throttle cables. After several seconds of running the engine was then shut down by turning off the magneto switches. No anomalies were noted during the test run. The original exhaust muffler, which was still in its original state, was then installed on the engine for a second test run. The engine, which was started in the same matter as the first test run, ran for several seconds before it was manually shut down. No anomalies were noted during the test run. (Refer to the Rotax Investigation Report, which is included in the public docket for this report.)