NTSB Factual Report

HISTORY OF FLIGHT

On July 7, 2004, at 0811 Pacific daylight time, an experimental Anderson Harmon Rocket II, N28BE, impacted a residence in Seal Beach, California, following an uncontrolled descent. Impact forces and a post crash fire destroyed the airplane. The airline transport pilot, who was also the builder and registered owner, was operating the airplane under the provisions of 14 CFR Part 91. The pilot, the sole occupant, was fatally injured; three people on the ground sustained minor injuries. The airplane departed Zamperini Field, Torrance, California, at 0802, and was flying a tower en route control (TEC) instrument clearance to the Chino Airport, Chino, California. A combination of instrument and visual meteorological conditions prevailed along the route of flight that was being flown under instrument flight rules (IFR).

According to the Federal Aviation Administration (FAA), air traffic control cleared the pilot to climb to 5,000 feet mean sea level (msl), and proceed direct to the Seal Beach VHF Omni-directional Range/Tactical Air Navigation (VORTAC). Recorded radar data depected the aircraft's target return in cruise flight with a mode C report of 5,300 feet msl, and then radar contact was lost.

According to a witness riding his bicycle southbound on a bike path, he heard a noise that "made the hairs on the back of my neck stand up." He looked over his left shoulder and saw the airplane coming out of the 600-foot cloud layer in a 75-degree nose-down attitude. The airplane impacted about 100 yards from the witness's location. The witness was an engineer, and a former glider pilot with about 800 flight hours.

Another witness stated that the airplane was flying north and the engine noise was increasing. The airplane was approximately 200 feet above ground level (agl). The wing was almost perpendicular to ground level. The airplane then arched downward and impacted the ground nose-first. This occurred over a period of about 3 seconds.

PERSONNEL INFORMATION

The pilot held an airline transport pilot certificate for single and multiengine airplanes. He also held a repairman experimental aircraft builder certificate for the Harmon Rocket II, serial number 017 (the accident airplane). The pilot's most recent third-class medical was issued on April 29, 2003.

The personal logbooks of the pilot were not recovered. The pilot's family reported that he carried his logbook in the airplane when he flew. Based on an insurance application obtained from the pilot's family, dated June 9, 2004, he had a total flight time of 5,000 hours. Two hundred hours were in the past 12 months; 50 hours were in the past 90 days.

The pilot had experience in the following categories, as listed on the insurance application: tailwheel, 550; retractable gear, 4,000; multiengine, 2,300; jet, 3,200; 1,000 instrument (actual and simulated); and a type rating for a Cessna Citation issued in 1977. The pilot had also flown P-51, T-6, and T-28 airplanes. Throughout his career in the United States Navy he flew additional military aircraft as well. From 2002 to 2003 he attended aerobatics training at a local airport.

AIRPLANE INFORMATION

The Harmon Rocket II airplane was a low-wing, single engine, experimental airplane, constructed by the pilot over a period of 13 years. The last inspection was completed on May 27, 2004, and was in accordance with the scope and detail to 14 CFR Part 43, appendix D. At the time of the accident the airplane had accrued approximately 250 hours.

The airplane was equipped with a Lycoming IO-540-C4B5 engine, serial number L-26847-48A, modified so that it was for experimental use only. The two-bladed propeller was a Hartzell HC-M2YR-1BF/F8475J-4, serial number EN 995B.

The week prior to the accident, the pilot and his wife flew to Jackson Hole, Wyoming, in the accident airplane. The entire trip was flown under visual flight rules. According to the pilot's wife, she was not aware of any mechanical issues with the airplane during or following the flight.

On the day of the accident, a witness, who was a retired mechanical engineer, was waiting for his friend to arrive at the Chino Airport about 1200. While he was waiting he walked to Square One Aviation, a P-51 restoration company owned by the pilot. As he was walking toward the Square One Aviation hangar, he noticed N28BE parked on the ramp. There was a man lying down on the asphalt under the empennage section of the airplane. The man's hands were reaching inside of the airplane but the witness could not see any tools on the ground surrounding him.

The day prior to the accident, the pilot told the general manager of Square One Aviation that he was having trim system problems with the airplane. During the morning's flight from Torrance to Chino, he was required to hold the stick in the full aft position. The electric trim would travel in one direction but not in the other. The pilot was looking at the airplane between 1200 and 1330.

A receipt dated July 6, 2004, was acquired from Aircraft Spruce and Specialty Company. The receipt listed the following purchases: SAF-Air Drain Valves; RAC Servo Relay Deck; RAC Clevis/Pushrod; and RAC Servo Only. Following the accident, all of the purchased parts were located in the pilot's hangar except for the RAC Servo Relay Deck.

According to the Aircraft Spruce and Specialty Company catalog, the purpose of the servo relay deck is to convert the SPST (single-pole, singe -throw) momentary contact action of any pair of switches, or a 4-way switch, into a DPDT (double-pole, double-throw) action that is necessary to operate R.A. [RAC] servos. The two internal relays are wired special for this and rated at 1 amp. According to a builder of a Harmon Rocket II airplane, if a builder uses normal switches to operate the RAC trim servos, they need to be wired through the servo relay deck.

The Ray Allen Company, manufacturer of the RAC Servo Relay Deck, did not report any known malfunctions of the relay deck or of the associated components based on customer service reports.

According to the owner of Harmon Rocket, LLC., he supplies parts to convert the RV-4, home-built kit, offered by Van's Aircraft, Inc., to a Harmon Rocket. The airplane modification kit was structurally tested through dive tests. The flight testing of the Harmon Rocket II consisted of flutter testing up to 300 miles per hour (mph), stability testing on the ailerons and elevators, g-load testing to 6 positive, and aerobatics training that included loops, rolls, and spins. No drawings were created for the tests. The single change made to the empennage section was an increase in the metal skin thickness from 0.016 to 0.020 inches. At the time of this report's writing, approximately 214 kits have been sold and 108 airplanes are currently flying.

The F-1 Rocket, also an experimental aircraft, is a similar design to the Harmon Rocket II. The F-1 has the same engine and is based on the RV-4 airframe design. The owner of F-1 had his airframe vendor company perform analytical testing on the structural components of the airframe to ensure adequate structural design of the RV-4 during increased airspeeds. As a result of the analytical calculations, the following findings and/or changes were made to the aircraft design:

- The middle hinge on the elevator could be overloaded at increased airspeeds; however, the hinge was not changed due to the fact that there have been no previously reported problems.

- The horizontal stabilizer's spar channel was changed from a material thickness of 0.032 inches to 0.040 inches.

- The vertical fin's forward spar attachment (where it attached to the horizontal stabilizer) was enlarged and a doubler was placed on the spar to aid in side load tolerances.

- The vertical stabilizer and the rudder were increased in length approximately 5 inches.

METEOROLOGICAL INFORMATION

The routine automated surface observation (METAR) for Long Beach was issued at 0753. It indicated the following: calm wind; visibility 7 statute miles; overcast cloud conditions at 1,600 feet agl; temperature 66 degrees Fahrenheit; dew point 59 degrees Fahrenheit; altimeter setting 29.99 inches mercury.

The National Transportation Safety Board investigator-in-charge (IIC) reviewed pilot weather reports (PIREPS) for the local area. A PIREP for Chino, located about 26 nautical miles northeast of the accident site, given at 0812, reported that the cloud tops were 2,000 feet msl. A PIREP given for Hawthorne, located about 15 miles northwest of the accident site, at 0954 indicated that the cloud tops were at 2,200 feet msl.

A Safety Board meteorologist reviewed weather data for the period around the accident. No unusual weather disturbances were noted in the area.

COMMUNICATION

The pilot was in communication with the Southern California Terminal Radar Approach Control (TRACON) just prior to the accident. The transmissions during the majority of the en route portion of the flight were unremarkable. At 0810:59, the controller queried the pilot because his mode-C target indicated 5,300 feet. There was no response.

WRECKAGE AND IMPACT INFORMATION

The FAA inspector and a Textron Lycoming representative, who was a party to the investigation, responded to the accident scene. The airplane collided with a residence, and an additional residence sustained fire damage. The approximate global positioning system (GPS) coordinates were 33 degrees 46.848 minutes north latitude by 118 degrees 5.618 minutes west longitude, approximately 4 miles from Daugherty Field, Long Beach, California.

The airframe and engine were extensively fragmented. The largest recovered portion of the airframe was the main wing spar. The right portion was bent aft from the center spar area, similar in deformation to an S-bend. The remaining pieces of the airframe were recovered for further examination.

The engine was extracted from a 3-foot crater. Two of the six cylinders were displaced from the engine assembly. The remaining cylinders were crushed and deformed. The engine casing was shattered so that the interior of the engine was exposed. The engine manufacturer's representative stated that coloration of the two displaced cylinders was consistent with normal operation. The crankshaft was intact and ran the length of the engine. The connecting rods remained secured to the crankshaft. The engine accessories were destroyed.

MEDICAL AND PATHOLOGICAL INFORMATION

The Orange County Coroner completed an autopsy on the pilot. The FAA Toxicology and Accident Research Laboratory, Oklahoma City, Oklahoma, performed toxicological testing of specimens of the pilot. The results of the specimen analysis were negative for ethanol and tested drugs.

According to the pilot's family, the family medical history included documented brain aneurisms and heart problems leading to unexpected and sudden death. Following the accident, there was no available medical evidence to ascertain whether or not the pilot was incapacitated during the flight.

TESTS AND RESEARCH

The airplane was examined on July 8, 2004, by the Safety Board IIC . A follow-up examination ensued on July 28 by the Safety Board IIC and a representative from the kit manufacturer, who was a party to the investigation.

Sections of the fragmented and fire-damaged airplane were identified through the verification of controls surfaces, spars, cables, and attachments. The center section of the wing, the flaps, and the left aileron were all identified. Fragmented pieces of the right aileron and attachment hardware for the left aileron and flaps were loose within the wreckage. The left aileron's outer attach point was connected to the wing, while the inboard attach point was not. The aileron bellcrank assembly was located in the deformed left wing structure. The elevator control tube that extends to the forward cockpit was crushed within its remaining burned and fragmented structure.

The rudder system is controlled through cable linkages. The cables run from the rudder pedals to the empennage area where they are attached to two springs. These springs are attached to the tail wheel. The tail wheel was recovered and one spring was secured to it; this spring measured approximately 12 inches and was attached to a 70-inch cable. A 3-inch spring, similar to the first, was connected to a 44-inch cable; however, it was not connected to the tail wheel. A cumulative total of approximately 293 inches of cable was located within the wreckage. A light located by design on the lower portion of the rudder, was recovered loose within the wreckage.

On July 9th, the Safety Board IIC examined an exemplar airplane that was in its finishing stages of building. The control surfaces and attachments were measured and identified. The elevator control system consists of an elevator control tube that runs aft to the empennage. The elevator tube is connected to a bellcrank and two tubes run outboard to the each elevator.

The airplane's elevator control surfaces, including their attachment hardware, were not identified within the wreckage. The rudder was identified by the light positioned by design on its lower portion. According to the kit manufacturer, in order for the rudder light to arrive at the accident scene, the vertical stabilizer and rudder would have been intact, prior to the airplane's impact with the ground.

The kit manufacturer did not have on record any reports of trim system control problems that resulted in a loss of airplane control.

Radar data from the southern California Terminal Radar Control (SoCal TRACON) was sent to a Safety Board Air Traffic Control specialist. Radar data indicated that the airplane's mode-C radar target was initially established at 5,200 feet, with a groundspeed of 200 knots, at 0811:02. At 0811:11, the target indicated an altitude of 3,900 feet. Primary radar returns not associated with the airplane's transponder returns were observed just prior to and following the 3,900-foot target.

A Safety Board engineer completed a trajectory study for the unidentified airplane components utilizing the primary radar returns and winds aloft information. The study revealed that the initiation point for the returns would have been near the Los Angeles River. Several aerial and ground searches commenced; however, none of the missing pieces were located.

A review of air traffic in the area did not reveal the presence of any larger aircraft crossing over or near the accident airplane's flight path.

A cassette tape that contained a copy of the audio information from the SoCal TRACON (Radar 070 position) was sent to the Safety Board Vehicle Recorder's Laboratory, Washington, DC. The analog recording was examined to document any engine or propeller sounds that could be heard during radio transmissions from the accident aircraft. There were several radio transmissions identified as originating from the accident aircraft recorded while the pilot was talking to SoCal approach control. The radio transmissions were examined on an audio spectrum analyzer to identify any background sound signatures that could be associated with either the aircraft's engine or the attached propeller.

During all of the radio transmissions originating from the accident aircraft, the primary blade passing frequency of the rotating propellers could be identified. This primary frequency was usually accompanied by the odd numbered harmonic frequencies (3rd , 5th, 7th, etc.) generated by the rotating propeller. No sounds associated with the actual engine could be identified.

Additionally, during the known flight time of the accident aircraft, there were three unknown radio transmissions recorded. No distinct propeller sound signatures could be identified during these unknown radio transmissions.

The three unknown radio transmissions were examined in an attempt to determine if they originated from the accident airplane. The unknown radio transmissions only contained transmitter noise with no voice. Even though there was no discernable voice or engine/propeller sound present in any of the unknown transmissions, it was determined that the first two unknown radio t

NTSB Probable Cause

the in-flight separation of the elevator, which resulted in a loss of airplane control and an uncontrolled descent. The underlying cause and mechanism of the elevator separation could not be determined.