HISTORY OF FLIGHTOn May 09, 2023, about 1750 Pacific daylight time, a Rans S-21 homebuilt airplane, N468MM, was substantially damaged when it was involved in an accident near Creswell, Oregon. The pilot, the sole occupant, was not injured. The airplane was operated as a Title 14 Code of Federal Regulations Part 91 personal flight. The pilot and his son had recently finished building the airplane. The accident flight was one of a series of test flights the pilot was performing after making cowling modifications. After a normal departure, the pilot flew to a test area and was in flight for about 20 minutes. While the pilot was maneuvering at about 700-900 feet above ground level (agl), the engine made “rough” sounds and vibrated. About 20 seconds later, the propeller seized, and the pilot made a forced landing in the field below. During landing roll out, the airplane flipped over and came to rest inverted, damaging the vertical stabilizer (see Figure 1 below). Figure 1: Wreckage AIRCRAFT INFORMATIONThe UL Power 520T six-cylinder engine was new when installed on the airplane and had amassed a total time of about 55 hours. The pilot noted that for about the last 20 hours of engine operation, the engine monitoring unit showed that the cylinder head temperatures (CHT) in the Nos. 3 and 4 cylinders were high (occasionally the Nos. 2 and 6 cylinders would also be high). In response, the pilot changed to a larger oil cooler and was modifying the cowling to increase airflow. AIRPORT INFORMATIONThe UL Power 520T six-cylinder engine was new when installed on the airplane and had amassed a total time of about 55 hours. The pilot noted that for about the last 20 hours of engine operation, the engine monitoring unit showed that the cylinder head temperatures (CHT) in the Nos. 3 and 4 cylinders were high (occasionally the Nos. 2 and 6 cylinders would also be high). In response, the pilot changed to a larger oil cooler and was modifying the cowling to increase airflow. TESTS AND RESEARCHInvestigators conducted a postaccident examination of the wreckage, which included a complete disassembly of the engine. The cowlings had been removed and a trace amount of oil on the bottom cowling was noted in the area under the Nos. 2, 4, and 6 cylinders. Removal of the top spark plugs revealed that all showed signatures that the engine representative stated were consistent with normal operation. Investigators were unable to rotate the crankshaft by hand using the propeller; it would only rotate a few degrees before binding. Clean, uncontaminated oil was observed at all six rocker box areas. Removal of the cylinders revealed that the combustion chambers and pistons were undamaged. Removal of the oil sump drain revealed that the magnetic plug was full with ferrous metal; metal pieces were also found in the oil sump. The oil pump was disassembled and there was limited metal found inside, with the exception of the bypass spring. Upon applying force to the propeller, the engine would not rotate through. Removal of the pistons revealed that the crankshaft was fractured at a web (crank cheek), just aft of the No. 2 connecting rod journal. The thrust bearing mount was able to move forward and aft. Disassembly of the crankcase revealed that the six connecting rod journals still had their respective rods and caps attached; all were able to be easily rotated about their respective journals (see Figure 2 below). Figure 2: Crankshaft The fracture extended through the crank cheek in the area between the No. 2 rod journal and the No. 3 main journal. The fracture surface was relatively shiny and reflective, with a portion of the cheek showing severe deformation damage and heat tinting on the forward face. The preliminary assessment of the fracture surfaces was consistent with a fatigue fracture at the aft radius of the No. 2 rod journal (see Figure 3 below). Figure 3: Mating Fracture Faces The National Transportation Safety Board Materials Laboratory examined the engine crankcase, crankshaft with attached connecting rods and bearings, seven pairs of main bearing halves, two filters, and sectioned oil tube. The No. 2 rod journal's aft fracture face displayed crack arrest marks consistent with fatigue cracking, originating from multiple points at the surface of the aft radius. Circumferential scoring marks were present on the surface of the aft radius, intersecting the fatigue origin area, with the fatigue crack measuring about 1.6 inches. The crack propagated through 30% of the rod journal cross section. The remaining fracture extended axially through the crank cheek between the No. 2 rod journal and the No. 3 main journal. The major portion of the crank cheek fracture exhibited damage consistent with post-fracture mechanical damage. The mating fracture exhibited similar fracture features. Evidence of superficial cracks at the radii locations was observed in the No. 1 main journal, No. 1 rod journal, No. 2 main journal, and No. 2 rod journal. The remainder of the crankshaft revealed no evidence of cracks at the radii locations. The bearing shell halves for the No. 2 main journal at the inner surface showed evidence of severe circumferential scoring that exposed the steel core layer, and the inner layers of copper, aluminum, lead, and tin, whereas the inner surface of the remaining main journal bearings showed evidence of moderate circumferential scoring. Examination of two oil filters (one oil filter was mounted on the engine at the time of the incident; the other oil filter was replaced during the last oil change) revealed that the accordion-like fabric section contained material embedded within them. SEM examination of the exterior surfaces of both fabric portions indicated the presence of metallic particles. Analysis using a handheld alloy analyzer detected metal particles containing aluminum, lead, copper, iron, zinc, nickel, molybdenum, and chromium in varying proportions, consistent with metal from the crankshaft bearing shells. The sizes of the thrust bearing disc springs were measured and found to be within specification and no wear or damage was detected. The specified torque for the propeller flange bolt was 221 foot-pounds (ft-lbs) and it had not been removed since its initial installation. Disassembly of the propeller flange bolt with a wrench revealed that the bolt required less rotational torque to remove than expected and the threads from the disassembled bolt showed no signs of mechanical damage. It was not possible to check the propeller balance due to damage incurred after impact. The postaccident download of the ECU revealed that the turbocharger pressure and fuel pressure were below operational limits, which the engine manufacturer stated would have resulted in a leaner mixture. In addition, the wide-open throttle RPM was, in certain moments of the flight, below operational limits, which the engine manufacturer stated would result in additional stresses on the piston/crankshaft.