On May 12, 2005, about 1650 Pacific daylight time, a Beech C35, N709D, experienced a loss of engine power and collided with a fence during a forced landing in an open field near Rancho Paseana, Rancho Santa Fe, California. The private pilot/owner operated the airplane under the provisions of 14 CFR Part 91. The airplane sustained substantial damage. The pilot, the sole occupant, was not injured. Visual meteorological conditions prevailed for the local area flight that departed Montgomery Field Airport (MYF), San Diego, California, about 1630. The flight was destined for Lake Riverside Estates Airport, Anza, California. No flight plan had been filed. In the pilot's written statement he reported that as he initiated the climb to 5,500 feet, he advanced the propeller and throttle controls to obtain 25 inches of manifold pressure and 2,500 rpm. When he checked the fuel flow gauge he noted that it read 10.9 gallons per hour (gph), when he expected to see between 16-17 gph. He checked the propeller and throttle controls believing that he had read the gauges incorrectly; he said they both indicated "25 square." He advanced the throttle and the engine immediately went to an idle condition coincident with the fuel pressure gauge dropping "dramatically." The pilot reduced the throttle to its previous position; however, the engine remained at idle. He then switched fuel tanks (3 times) and pumped the throttle "in [and] out on each tank." The pilot stated that there was only a small reaction. The pilot contacted Southern California Approach Control (SoCal) and began to look for a suitable landing area. He saw an open pasture, and noted that he was too low and the field had obstacles (trees) that he would have had to maneuver around had he been able to set up for the landing. The pilot stated that he "carried extra speed (80 kts) in case he got to low and needed to pop up over the tree line." He lowered the landing gear on the "base" turn and recalled that he airplane barely cleared the treetops. The pilot reported that because of the extra airspeed, the airplane floated in ground effect and he could not get the airplane stopped before it impacted two fences. The airplane came to rest in a nose down configuration; the nose landing gear had collapsed. The National Transportation Safety Board investigator-in-charge (IIC) interviewed the pilot. The pilot reported that after clearing Class B airspace, he initiated a climb, and the engine went to idle. He changed fuel tanks, and "pulled the power in and out." The pilot stated that he made a forced landing in an open area. The airplane went through a fence before coming to rest on its nose. According to a deputy from the San Diego Sheriff's Department, the fuel tanks were not compromised during the accident sequence. According to a Federal Aviation Administration (FAA) inspector who responded to the accident site, when he drained fuel out of the left, right, and center fuel tanks, there was water and debris mixed with the fuel. The pilot also informed the FAA inspector that he had a supplemental type certificate (STC) for auto fuel. The Safety Board IIC, and Teledyne Continental Motors, a party to the investigation, inspected the airframe and engine at Aircraft Recovery Service, Littlerock, California, on June 3, 2005. The visual examination of the engine revealed no preimpact anomalies and investigators determined that a ground run was possible. Investigators were able to start the engine, but it would not maintain power. Investigators used an alternate pressure source installed on the pressure-inlet side of the carburetor. An engine run was completed for approximately 30 seconds, with 1,400 rpm and 15 inches of manifold pressure. The engine driven fuel pump was disconnected and inspected. The engine manufacturer noted that the drive coupling was "worn excessively," and would not engage when manually rotated. He also noted that the retaining clip, disintegrated when removed. The engine driven fuel was retained and shipped to the Safety Board materials laboratory in Washington, D.C., for further examination. AIRCRAFT INFORMATION The airframe logbook revealed that an annual inspection had been completed on May 5, 2004. The engine logbook indicated that San Diego Airmotive, Ramona, California, had installed a field overhauled engine on May 5, 2004. Cruiseair Aviation in Ramona had completed the overhaul. The pilot reported that the airplane had flown 149.38 hours since the annual inspection. A Thompson TF-1900 engine driven fuel pump had been installed on the engine. During the engine field overhaul done by Cruiseair Aviation, the fuel pump had been shipped to Thunderbird Aircraft Parts, Inc., Bethany, Oklahoma, for an overhaul. Thunderbird disassembled and inspected the core fuel pump and exchanged it for an overhauled Thunderbird Aircraft Parts, Inc., fuel pump. The fuel pump was assembled using Thunderbird's TAP repair procedure TAP200. The invoice indicated that the repair procedure cancelled Beech and Continental Motors 300-hour mandatory inspection, overhaul, and test. According to the TAP200 repair procedure titled SPLINE DRIVER, in the NOTES section it indicated that the splined drive couple was heat treated to 50-52 HRC, which is the specification for the part. TEST AND RESEARCH Engine Driven Fuel Pump The senior metallurgist in the Safety Board's Materials Laboratory noted that the drive gear had splines on one end, a square key portion on the end inserted into the pump, and a reduced diameter round segment between the splined and square ends. The visual examination revealed that the square key portion showed evidence of wear and metal flow through most of the axial length. He noted that the spline teeth of the other end of the drive gear were intact and showed no evidence of cracking or wear. He assembled the square key portion of the drive gear and inserted it into a square slot located on the fuel pump. The surfaces of the square slot showed evidence of minor wear when compared to the square key portion of the drive gear. He manually rotated the assembly and noted no engagement between the square key and corresponding slot. He was able to achieve only intermittent engagement. Metallurgical Examination The fuel pump was submitted to an independent laboratory for a Rockwell Hardness Test (HR) of the gear. The laboratory used a specified hardness of 50-52 HRC (C-scale) as the baseline for the testing per the overhaul facilities FAA approved type data for repair specifications. The HRC test results were below the acceptable limits of the C-scale. A Rockwell B test was performed with an average gear hardness of 94.5 HRB (B-scale). The laboratory determined that the gear had a cadmium coat. Another hardness test was conducted after the cadmium coat was removed. Rockwell C tests were performed with an average gear hardness of 40 HRC. ADDITIONAL INFORMATION On February 13, 2006, the FAA released a suspected unapproved parts notification No. 2004-00053 regarding the overhaul/repair procedure utilized by Thunderbird Aircraft Parts, Inc., for the Thompson Fuel Pump TF-1900. At the request of the Safety Board IIC, TCM issued service instruction letter SIL06-2, on May 28, 2006, notifying owners and operators of the FAA's suspected unapproved parts notification regarding the fuel pump.