Mitchells, VA, USA
N4070F
TITAN TORNADO
The pilot reported that, after flying the airplane for about 40 minutes on a personal flight, the engine suddenly lost power and ran rough. The pilot could not maintain level flight and performed a forced landing to a road. After touchdown on a roadway, it turned to the left, and the airplane departed the pavement, struck a drainage ditch, and then struck a fence. The airplane’s wingtips, landing gear, and stabilator were substantially damaged. Examination of the engine revealed the No. 3 cylinder had incurred significant internal damage. Numerous metal fragments were present, the piston had broken up, and the exhaust valve head had separated from the stem. Further examination revealed multiple initiation fatigue cracking of the exhaust valve stem, which progressed until failure. Research indicated that fatigue cracking of exhaust valves in similar engines had occurred in the past. Investigations by multiple countries revealed that valve stem separation due to fatigue, in addition to stem cracking of other valves in engines that did ultimately separate, had occurred. This evidence also indicated that overheating of the valves was at least a factor in those engine failures.
HISTORY OF FLIGHT On October 15, 2015, about 1300 eastern daylight time, an amateur-built Titan Tornado airplane, N4070F, was substantially damaged when it was involved in an accident near Mitchells, Virginia. The pilot was not injured. The airplane was operated as a Title 14 Code of Federal Regulations Part 91 personal flight. According to the pilot, after performing a preflight inspection on the airplane and receiving the weather via the Warrenton-Fauquier Airport (HWY), Warrenton, Virginia, automated weather observation system, he departed from runway 33 about 1235. After departure, he turned the airplane to 240° and climbed to about 2,000 ft mean sea level (msl). He flew the airplane outbound from the airport for about 35 minutes and then turned around to return to HWY. He climbed the airplane to 2,500 feet msl, and about 5 minutes after turning back to the airport, the engine suddenly "lost power, ran rough," and he could not maintain level flight. The Coffeewood Correctional Center was at his "4 o'clock" position and he believed that it presented the best available landing site, so he turned towards it. He then maneuvered the airplane to line it up with the entrance road to the prison. An automobile was entering the prison on the road in front of the airplane, so the pilot glided the airplane over the car and then made a "very quick descent" to the road surface to avoid striking a set of powerlines that crossed the road. After touchdown, he attempted to stay on the roadway but the airplane was unable to follow "a dogleg" of the roadway to the left. The airplane departed the pavement on the right side of the road, struck a shallow drainage ditch, struck the prison fence with its right wingtip, spun 180°, and came to rest with the left wing against the fence. AIRCRAFT INFORMATION The airplane was equipped with an air cooled, 4-cylinder, naturally aspirated, 85 horsepower, Jabiru 2200A engine mounted in a pusher configuration, driving a Prince Aircraft 2-bladed carbon fiber propeller. The airplane's most recent condition inspection was completed on June 28, 2015. At the time of the accident, the airplane had accrued 388 total hours of operation, and the engine had accrued 136.3 total hours of operation. WRECKAGE AND IMPACT INFORMATION Examination of the accident site revealed that the airplane traveled about 223 ft after leaving the roadway on a heading of about 108° before striking the perimeter fence of the correctional center and coming to rest. Postaccident examination of the airframe revealed that it was substantially damaged. The wingtips displayed impact damage, the right main landing gear was separated from its mounting location, the stabilator was bent, and the wheel pant for the nosewheel had been separated from its mounting location. Postaccident examination of the engine revealed that the No. 3 cylinder had internal significant damage, as the cylinder head dome displayed multiple areas of impact damage, and the exhaust valve head was missing. Examination of the No. 3 cylinder also revealed numerous metal fragments and that the piston had broken up and pieces of it were lying in the bottom of the cylinder. The pieces displayed evidence of impact damage and fracturing. The No. 3 connecting rod was visibly connected to the crankshaft but displayed areas of bending. Examination of the exhaust system also revealed that part of the piston ring had been captured in the muffler. Additional examination of the No. 3 cylinder revealed that the exhaust valve had separated at the stem. The remaining exhaust valve stem appeared straight, and no scoring was observed on the outer diameter (OD) surfaces. Progressive crack arrest marks were visible on the fracture surface, consistent with fatigue cracking, and had progressed through most of the valve stem prior to separation. Numerous thumbnail-shaped patterns with accompanying ratchet marks were present around the diameter of the valve stem, which were consistent with multiple-initiation fatigue cracking that likely occurred as the valve continued to operate after the main fatigue cracking had begun to progress. No defects or pre-existing damage were observed at the initiation site of the main fatigue cracking. After removal of the rocker shaft, both valves moved freely inside their respective valve guides. The intake valve stem was bent. Cracking was visible on the OD of the stem of the intake valve. The material around the cracking appeared to have flaked off and shiny silver metal was visible under the darkened outer layer of the intake valve. Similar cracking was observed in the head end of the exhaust valve near the fracture surface. However, the cracking in the exhaust valve was not accompanied by similar material flaking observed with the intake valve cracking. Metallurgical cross-sections through the exhaust valve revealed numerous cracks in the exhaust valve stem near the fracture surface, one of which extended through almost half of the exhaust valve diameter. Most of the cracks had oxidation products visible inside, which was consistent with the cracks being present during engine operation at temperatures sufficient to cause oxide scale growth. A distinct layer was also observed on the external surface of the exhaust valve. The layer was present on both the exhaust valve stem and tip. Some shallow cracking was observed in the layer near the tip. A distinct layer like that on the exhaust valve was observed on the intake valve. Shallow cracking was also observed on the intake valve. The cracking was observed at both the head and tip which were comprised of different materials. Some shallow cracking was observed in the layer near the tip and occurred more frequently than on the exhaust valve. The cracking was mostly confined to within the distinct layer, but there was at least one crack that had extended into the base material. Some of the cracking was in conjunction with areas where the distinct layer had separated, which was consistent with the flaking observed macroscopically. According to Jabiru, the valves were manufactured from 214N stainless steel with no additional coatings. 214N steel is a nitrogen-strengthened stainless steel. Semi-quantitative energy dispersive spectroscopy (EDS) spectra showed the base metal of the valves was iron-rich, with a high level of chromium, a moderate amount of manganese, some nickel, and a trace amount of silicon, which was consistent with 214N steel. Wavelength dispersive x-ray spectroscopy (WDS) line scans of the valve cross-sections resulted in elevated levels of nitrogen in the distinct layer compared to the base metal. The microstructure of the base metal of the valve stems was austenitic stainless steel containing ferrite stringers. The average core hardness of the intake and exhaust valves measured 47.5 and 48.3 HK, respectively, with a 50g load. The recovered pieces of the No. 3 piston had extensive secondary damage. The undamaged areas of the fracture surfaces had features consistent with overstress. One of the No. 3 sparkplugs was intact but had some impact damage. The insulator on the other sparkplug had separated at the threaded end. That same sparkplug had the center electrode tilted off-center and the ground electrode bent upright. The damage to the surfaces of the No. 3 cylinder head, the piston, and the sparkplug were likely secondary to the separation of the exhaust valve. ADDITIONAL INFORMATION Accident Airplane Engine History The engine involved in the accident (S/N 22A2371) was manufactured in Queensland, Australia, on September 3, 2006, and sold into the United States. According to the pilot, he had purchased the engine second-hand to replace his previous Jabiru 2200 engine (S/N 22A959), which had incurred a valve failure at 269.7 total hours of operation and resulted in an emergency landing in a field. The pilot installed the accident engine on the accident airplane on May 4, 2013. The engine had previously only been installed on a Kolb airplane and had accrued about 9 total hours of operation on that airplane since being manufactured 7 years before. The pilot had installed cooling ducts as well as cylinder head and exhaust gas temperature (EGT) probes on the No. 2 and No. 3 cylinders. The EGT probes displayed their data on an AMPtronic SKYDAT GX2 display. Because the pilot believed that the EGT readings were too low, he also would rely on the color of the spark plugs to monitor performance. On September 22, 2015, (5.2 hours before the engine failure) the pilot installed a ROTEC ignition system to improve cold starting. Use of Fuel Additives The pilot had been operating the engine using 93 Octane “MoGas” (lead- and ethanol-free automobile gasoline) since he installed it. The pilot had heard about using Decalin Run Up fuel additive in the gasoline from some friends at a fly-in; on June 22, 2014 at 61.9 hours of operation, the pilot started adding Decalin to his gasoline. Because the fuel additive was used to eliminate lead oxide fouling, the pilot saw no point in using it with lead free gas and had discontinued using it for some time prior to the engine failure. According to Jabiru, the use of fuel additives was not allowed. Air Accident Investigation Branch Identified Failures In May 2010, the United Kingdom’s Air Accident Investigation Branch (AAIB) published Bulletin 5/2010, which discussed failures in two different Jabiru 2200 engines. In the first instance, about 2 hours into the flight, the pilot of a Jabiru UL-D noticed the sudden onset of vibration, and during the emergency approach to a field, the engine stopped. Postaccident examination of the engine revealed that the No. 3 cylinder exhaust valve had failed in fatigue and the remaining three exhaust valves had fatigue cracks in the same area. In the second instance, a Jabiru 2200 engine’s No. 1 cylinder exhaust valve failed. The metallurgical examination of all four exhaust valves indicated a failure mode and fatigue cracking very similar to those from the other airplane. According to the AAIB, examination of the two failed exhaust valves showed that both failures were a result of fatigue crack propagation initiating at multiple origins at the base of the exhaust valve stems. Postaccident examination of the valve stem surfaces in the regions of failure identified pitting and general surface corrosion with secondary cracking. The fatigue cracking probably initiated from corrosion pits on the surface of the stems, which would act as stress concentrators. Examination of the intact valves also showed evidence of corrosion and cracking. The AAIB stated that the evidence from these valve failures indicated that overheating of the valves was at least a contributory factor and noted that in addition to their investigations, a number of overheat-related failures occurred in France about the same time. NTSB Previously Identified Failure On August 16, 2012, a similar failure occurred with a Jabiru 3300 engine (NTSB Case No. ERA12TA542) when, while climbing to 6,500 ft msl, about 45 minutes into the flight, the engine began to run rough “like it was developing carburetor ice,” then began to “cough” like it was starved of fuel. However, before the pilot and the other crewmember began the high-altitude engine failure checklist, the propeller came off the airplane, struck the right side of the cowling, struck the right lift strut, and fell away. Postaccident examination of the No. 6 cylinder exhaust valve revealed a fractured valve stem and multiple ratchet marks around the perimeter of the stem, both consistent with fatigue crack initiation and propagation. Examination of the other cylinders and pistons also revealed the presence of heavy lead deposits, which indicate that the engine may have been run at excessively high temperatures. The NTSB determined that the probable cause of this accident was, in part, “A failure of an engine cylinder exhaust valve due to the buildup of lead deposits on the valve stem and fatigue cracking of the valve stem, resulting in a total loss of engine power, and the loss of the propeller.” According to Jabiru, at the time of the August 2012 accident, it was aware of about eight previous in-service failures of exhaust valves. Jabiru stated that the exhaust valves appeared to be intolerant of temperatures over about 750°C and that heat stress was the most common issue that they had identified. Jabiru further stated that around that temperature, the exhaust valve would start showing indications of stress corrosion/cracking at the base of the stem. ATSB Safety Report On March 9, 2016, the Australian Transportation Safety Board (ATSB), published ATSB Transportation Safety Report AR-2013-107, Engine Failures and Malfunctions in Light Aeroplanes. According to the report, over the 6-year study period between 2009 and 2014, 322 engine failures or malfunctions involving light aircraft were reported to the ATSB and/or Recreational Aviation Australia (RA-Aus). Fractures relating to valves were the second most common failure in Jabiru engines, with 13 reported over the 6 years and another 15 valve failures coded as mechanical discontinuities. In May 2015, Jabiru conducted a root cause analysis of valve train failures from 2013 to 2015. The report stated that “valve failures in Jabiru engines are virtually always exhaust valves,” which was consistent with what had been reported to the ATSB. Jabiru’s Analysis of Valve Failures According to Jabiru, at the time of this accident, they believed that about 31 valves had failed in the 6,500 engines (29,000 valves) that had been manufactured. Jabiru stated that these valve failures included situations where there had been all manner of unapproved modifications, use of unapproved propellers, using old fuel with additives, noncompliance with service bulletins and service letters, and failures to maintain and operate the engine in accordance with their manuals. Jabiru also stated that the practice of mixing of fuels was not endorsed by Jabiru. Although Jabiru identified it as being a common factor with some of the exhaust valve failures, it was not clear how much this practice contributed to exhaust valve failures. Jabiru Safety Improvements Jabiru took several actions to address these valve train failures, including reducing the published nominal EGT, completely redesigning of the valve train to use hydraulic lifters, modifying the valve guide tolerance, introducing valve relief pocketed pistons, including double valve springs in the valve train assembly, requiring 50-hour leak down tests, and increasing the oil change frequency. Additionally, Jabiru published several service letters and service bulletins to increase awareness of the issues and prescribe correct maintenance practices. In September of 2020, Jabiru reported to the NTSB that since these safety improvements were implemented, in the worldwide fleet of about 7,000 engines, that there have been only 8 broken valves found from 2018 through 2020. Additionally, they had not been notified of any exhaust valve stem cracks since 2018.
A total loss of engine power due to fatigue cracking and separation of the No. 3 cylinder exhaust valve stem.
Source: NTSB Aviation Accident Database
Aviation Accidents App
In-Depth Access to Aviation Accident Reports
