HISTORY OF FLIGHT Use your browsers 'back' function to return to synopsisReturn to Query Page
On January 18, 2005, approximately 1128 central standard time, a Cessna T-37B, a twin-turbojet military trainer, tail number 66-8003, operating under the call sign Cider 21, and an Air Tractor AT-502B single-engine agricultural airplane, N8526M, were destroyed following a midair collision during cruise flight near Hollister, Oklahoma. The T-37B was registered to and operated by the United States Air Force (USAF). The AT-502B was registered to a private individual and operated by a commercial pilot. The USAF flight instructor pilot was not injured and the USAF student pilot sustained minor injuries. The commercial pilot in the AT-502B was fatally injured. Visual meteorological conditions prevailed and the airplanes were operating in Class E airspace. The T-37B was in radar contact with approach control and was operating under Air Force Instructions (AFI) 11-202, Volume III. The AT-502B was operating under 14 Code of Federal Regulations Part 91 for the delivery flight, and a flight plan was not filed. The local flight for the T-37B originated from the Sheppard Air Force Base (SPS), near Wichita Falls, Texas, approximately 1022. The cross-country flight for the AT-502B originated from the Olney Municipal Airport, near Olney, Texas, approximately 1100, and was destined for Huron, South Dakota, with an intermediate fuel stop in Hutchinson, Kansas.
According to company personnel from an Air Tractor dealership in Arkansas, the pilot was hired to deliver the recently purchased AT-502B to the new owner in Huron, South Dakota. Company personnel at the Air Tractor factory located in Olney, Texas, reported that the AT-502B was equipped with basic visual flight rules (VFR) instruments and was not equipped with any radios or a transponder.
During an interview with the NTSB investigator-in-charge (IIC), the USAF flight instructor and student pilot reported that they were on a routine mission training flight (C2803). After a non-eventful departure from SPS, they performed two normal overhead approaches to SPS before being cleared into the Military Operations Area (MOA). After completing the series of high altitude maneuvers, the training flight received radar vectors to the RANCH intersection and then to the Frederick Municipal Airport (FDR), near Frederick, Oklahoma, which is commonly referred to by the USAF as "Hacker." As the flight descended to an altitude of 6,000 feet, the instructor noted the bottom of the overcast cloud ceiling to be between 6,000 and 6,500 feet mean sea level (msl).
After arriving at Hacker, the training flight performed a straight in no flap landing, and requested left closed traffic for practice landings. After completing a normal overhead approach and a single-engine landing, the flight proceeded to depart Hacker's airspace to the east and climbed to 5,500 feet msl.
The student pilot reported that he performed the en route portion of his checklist and contacted USAF Radar Approach Control (RAPCON) to notify them they were en route back to Sheppard Air Force Base and requested the "home plate" arrival. RAPCON advised the flight that they had radar contact, and to descend to 5,000 feet msl on a heading of 100 degrees.
The student pilot reported that after he leveled off at 5,000 feet msl at an indicated airspeed of 200 knots, the flight instructor took control of the T-37B. The student pilot stated that as the flight instructor took control of the aircraft, he scanned outside the airplane to the left, and started to look back to the right when he saw the yellow Air Tractor heading toward the right side of the T-37B. He reported, "… I saw the crop duster and it was just an immediate … it wasn't like I saw him moving, it's just like it grew and my jaw just dropped and I didn't say anything. I was just … I just said, "Sir," because I was just shocked and as soon as I said that, we just impacted and went out of control."
The instructor pilot stated that he took control of the airplane while still in the descent and he leveled the airplane at 5,000 feet msl. He reported that he briefly scanned at the student pilot's altimeter on the left side of the instrument panel (a standard practice for T-37 flight instructors). As he was turning his head back to the right, he noticed a "high visibility yellow airplane" out of the right corner of his eye, but did not have time to take evasive action.
Both pilots reported that the T-37B was level at 5,000 feet msl for about one minute or less before impact with the AT-502B. The instructor and student pilot recalled feeling a spinning sensation and rolling inverted. Both the instructor and student pilot initiated emergency egress procedures and ejected from the aircraft.
The T-37B and AT-502B impacted farm fields about 3.5 nautical miles (nm) east of Hollister, Oklahoma. Both aircraft were partially consumed by post impact fires.
A witness located north of the accident site reported in a written statement that he observed an aircraft descending rapidly in a nose down attitude and on fire prior to losing sight of it behind a tree line. Subsequently, the witness observed a second aircraft spinning in a nose down attitude, and it was missing a wing. The witness added that a plume of smoke was originating from the airplane but he didn't see any flames. As the airplane continued to descend, he noticed two parachutes on each side of the airplane, and he decided to proceed to the area to see if he could assist the pilots.
The pilot of the AT-502B held a commercial pilot certificate with airplane single-engine land, multi-engine land, and instrument airplane ratings. Review of the pilot's logbook revealed he had accumulated a total of 7,545 hours of flight time, of which 562 hours were in multi-engine aircraft. He logged 72 hours in the last 90 days and had flown 58 hours in the last 30 days. The pilot had logged 3,255 hours in AT-502B's between February 2000 and January 2004. The pilot owned a Beech B-55 Baron and logged about 115 hours in the last 12 months. The pilot was last issued a second-class medical on March 29, 2004, with no limitations stated.
The USAF flight instructor was a qualified flight instructor for the T-37B. Review of the flight instructor's flight records revealed that he had accumulated a total of 1,142 hours of USAF flight time, of which 748.5 hours were in B-52H's. He had flown 347 hours in T-37B's. Within the last 30 days, the instructor had accumulated 9.3 flight hours. The pilot reported that he had obtained a Federal Aviation Administration (FAA) Private pilot's certificate with a single-engine land rating and had logged about 55 hours of civilian flight time prior to entering USAF flight training. He held a second class medical certificate.
The USAF student pilot was in the first stage of his initial flight training. Review of the student pilot's flight records revealed that he had accumulated a total of 34.4 hours of flight time, all of which was in the T-37B. Within the last 30 days, the student pilot had accumulated 12.3 flight hours. The student pilot reported that he had obtained a FAA private pilot's certificate with a single-engine land and an instrument rating prior to entering the USAF. He had a total of 156 civilian flight hours. He held a second class medical certificate.
The Cessna T-37B military turbojet airplane, serial 66-8003 was a low-wing, dual control trainer, featuring side by side seating with a maximum of two occupants. The T-37B was powered by two Continental J69-T-25A turbojet engines, rated at approximately 1,025 pounds of thrust each. The T-37B was also equipped with a two-position speed break, spoilers, thrust attenuators, a jettisonable canopy, and two ejection seats. The wingspan of the T-37B was 33.8 feet.
The most recent inspection that the T-37B underwent was a periodic phase inspection completed on June 17, 2004, in accordance with the Air Force Maintenance Program. At the time of the accident, the aircraft had accumulated 403.2 hours since the inspection and 17,306 hours total time. The left engine, serial number CA00321219, had accumulated a total of 10,097.6 hours, and the right engine, serial number CA00321863, had accumulated 11,413 hours at the time of the accident.
The Air Tractor AT-502B, serial number 502B-2570, was a low wing, tailwheel equipped, steel tubular frame design airplane with a fixed landing gear. The airplane was configured for a maximum seating configuration of one occupant, and was configured for agricultural spraying operations. The airplane was powered by a Pratt and Whitney PT6A-34AG turbine engine, rated at 750 shaft horsepower, with a three bladed Hartzell constant speed propeller, model number HC-B3TN-3D/T10282NS+4. The airplane's maximum certified gross weight was 9,700 pounds. The AT-502B was 33 feet 2 inches in length and its wingspan was 52 feet.
The AT-502B was issued an airworthiness certificate on January 13, 2005. According to the airframe and engine logbooks, the airplane's most recent inspection was the Air Tractor factory's production flight test, conducted on January 10, 2005. At the time of the accident, the airframe and engine had accumulated approximately 1.5 hours since the factory new aircraft flight test.
The AT-502B was not equipped with a radio or a transponder. The Air Tractor factory representative reported that radios and transponders are typically not installed on new aircraft at the factory. He reported that new owners prefer to have the equipment installed on the airplanes with the entire avionics package that are typically purchased from an avionics shop selected by the new owner. Airplanes are not required to be equipped with a radio or transponder as long as FAR 91.215 "ATC transponder and altitude reporting equipment and use" is not violated.
The closest weather reporting station to the accident site was located at the Frederick Municipal Airport (FDR), near Frederick, Oklahoma, about 9 nautical miles west of the accident site.
At 1053, the automated surface observing system at FDR reported wind from 150 degrees at 10 knots, visibility 8 statute miles, cloud condition overcast 5,000 feet, temperature 2 degrees Celsius (C), dew point -2 degrees C, and an altimeter setting of 30.53 inches of mercury.
At 1153, the automated surface observing system at FDR reported wind from 160 degrees at 8 knots, visibility 9 statute miles, cloud condition overcast at 4,800 feet, temperature 3 degrees C, dew point -2 degrees C, and an altimeter setting of 30.53 inches of mercury.
The student pilot reported that the airplane descended through a cloud layer after departing the MOA en route to Hacker. He reported the base of the clouds was about 6,500 feet msl. He stated that when the airplane was at 6,000 feet msl they were clear of clouds, but the sky conditions were hazy without a well-defined horizon. He reported that the horizontal visibility was about three miles due to the gray overcast and misty haze. He reported that the visibility was better at lower altitudes and that he could pick up ground references, but the horizontal visibility was limited.
The instructor pilot reported that the visibility was about three to five miles when they were climbing out of Hacker toward Hollister. He reported that when they were at 5,000 feet msl, the visibility was about three miles due to hazy conditions.
The Sheppard AFB RAPCON transcript of the communications between Cider 21 and Sheppard RAPCON from 11:26:32 - 11:30:23, January 18, 2005, is provided below. The following facility abbreviations were used throughout the transcript:
a. Sheppard T-37 Arrival control (SPS-37AR)
b. Sheppard T-37 Arrival control trainee (SPS-37ART)
c. Sheppard T-37 Arrival Assist (SPS-37AA)
d. Sheppard T-38 Approach Assist (SPS-38AA)
e. Sheppard T-38 Approach Assist Trainee (SPS-38AAT)
f. Sheppard T-38 Arrival Assist (SPS-38ARAA)
g. Sheppard 2 Assist (SPS-S2AA)
h. Fort Worth Center, Falls-Low (ZFW 4)
The transcript follows:
11:26:32 CD21: Sheppard Approach, CD21 request home-plate to Bridge with Whiskey.
11:26:36 SPS-37AR: CD26, were you just established in the MOA still, right? Not wanting to come home.
11:26:41 CD26: Confirm. We're established in Area 1 and 'd like to remain here for about another 5 minutes.
11:26:44 SPS-37AR: Roger, frequency change approved and uh, contact approach when you're ready to go home.
11: 26:47 CD26: CD26, wilco.
11:26:48 Cargo: Cargo flight approaching Bridge, cancel.
11:26:50 SPS-37ART: Cargo, contact Cooter.
11:26:53 Cargo: Uh, Cargo, go channel 5.
11:26:55 SPS-37ART: CD21, Sheppard Arrival, IDENT.
11:27:03 SPS-37ART: CD21, Radar Contact over Hollister, descend and maintain 5,000, fly heading to 1-0-0, vectors Bridge. Verify you have information Whiskey.
11:27:09 SPS-37AA: Arrival Assist.
11:27:10 SPS-38AAT: Four south of the VORTAC, JAVIT 55, flight, correction, a, AT38 climbing to 13,000, up the airway.
11:27:12 CD21: CD21 has information Whiskey, descending to 5,000, heading 1-2-0.
11:27:15 SPS-38AA: And that's a Point Out.
11:27:17 SPS-37ART: CD21, heading 1-0-0.
11:27:18 SPS-37AA: Oh, JAVIT 55, Point Out approved.
11:27:20 CD21: CD21, heading 1-0-0.
11:27:38 SPS-37AA: Arrival Assist.
11:27:40 SPS-S2AA: Point Out 2-7, correction, 2 miles south of Fletch, JAVIT27, ILS correction Surveillance approach, 1-5 Center.
11:27:50 SPS-37AA: JAVIT27 Point Out approved. Traffic, CD21, 5,000, eastbound.
11:27:52 ZFW 4: Sheppard on the 4, I've got one at ANNAA.
11:27:57 SPS-S2AA: CD21, traffic observed.
11:27:59 SPS-37AA: DD
11:27:59 SPS-S2AA: AW
11:28:00 ZFW 4: Sheppard, Falls-Low, 4-Line, handoff at ANNAA.
11:29:02 SPS-38ARAA: Arrival Assist
11:29:02 SPS-37AA: APRE, I'm sorry, Point Out.
11:29:03 SPS-38ARAA: Go ahead.
11:29:04 SPS-37AA: Eight miles north of Grandfield antennas, CD21, descending eastbound, 3,200 then 2,600. To get out of the way of JAVIT 27.
11:29:13 SPS-38ARAA: OK, I do not see CD21.
11:29:15 SPS-37AA: You don't see the primary out there?
11:29:18 SPS-37ART: CD21, be advised, I'm not receiving your transponder, when able, reset.
11:29:19 SPS-37AA: Uh, alright we'll keep him at 5,000 then. That's your traffic for JAVIT27.
11:29:25 SPS-38ARAA: Copy. ***
11:29:26 SPS-37AA: DD.
11:29:42 SPS-37ART: CD21, be advised, your transponder appears malfunctioning, when able, reset transponder, squawk 4-2-2-1.
11:30:01 SPS-37ART: CD21, Sheppard Arrival, radio check. How do you hear me?
11:30:10 SPS-37ART: CD21, Sheppard Arrival, radio check. How do you hear me?
11:30:23 SPS-37ART: CD21, Sheppard Arrival on guard, radio check.
SPS is operated under class D airspace and is attended continuously. All radio and telephone voice traffic to SPS Tower and RAPCON are recorded through the Digital Voice Recording System (DVRS). The type of radar utilized at SPS is a GPN-20 radar system, but it has no recording capability.
FDR, located 33nm northwest of SPS, is operated under Class G airspace. FDR is used by Sheppard Air Force Base T-37 jet trainers for high density student pilot training. This training is conducted on weekdays during daylight hours. T-37 aircraft are controlled by the red and white runway supervisory units (Call sign: "Hacker") at the end of runways 17R and 35L. Normal T-37 pattern altitude is 2,200 feet msl. Straight-ins are flown at 1,700 feet msl. All civilian traffic is recommended to contact Hacker on 122.80 megahertz or UHF 285.7 when approximately 10 miles from the field. Hacker controls T-37 aircraft, but is an advisory-only service to civilian aircraft.
The alert area surrounding FDR is depicted as the A-561 High Density Student Training area on the Dallas-Fort Worth Sectional Aeronautical Chart. The vertical boundary is from the surface to 4,000 feet msl, and is operational from sunrise to sunset Monday through Friday. There is no air to ground communications available. SPS is the controlling agency for the airspace. The aircraft wreckage sites were located outside the eastern boundary of A-561. The airspace where the impact occurred was Class E airspace.
The RAPCON procedures for handling T-37's departing Hacker en route to SPS are found in the "80th flying Training Wing supplement to the Air Force Instruction 11-2T-37, Vol. 3." The procedures are stated below:
"6.12.7. When departing the VFR pattern, advise Hacker and make a climbing turn towards Hollister. 184.108.40.206. When "Homeplate" is desired, monitor ATIS (if time and conditions permit), cross Hollister in the climb at or above 3,200' MSL but at or below 5,500' MSL, squawk 42XX, call Sheppard Arrival (Channel 7) and request "Homeplate to the desired entry point with ATIS (if received)." Maintain VFR until receiving vectors/IFR clearance. RAPCON will provide radar vectors. If unable to contact RAPCON, proceed VFR along the Area Lost Comm Procedures routing."
The normal controller actions following a T-37 call up off of Hacker en route to SPS are stated below:
"Once the Sheppard Arrival controller radar identifies the T-37 requesting to return to Sheppard AFB, the controller issues an altitude to maintain and easterly heading. The actual altitude and heading depend on other traffic, the runway in use, and the entry point or approach requested by the pilot. The typical altitude is 5000 feet. The controller sequences the T-37 with other traffic inbound to Sheppard AFB and separates the aircraft from other IFR aircraft. While en route to Sheppard AFB, aircraft will receive traffic advisories on any known or observed conflicting traffic based on controller workload and equipment limitations."
The procedures for handling non-squawking VFR aircraft transiting Sheppard airspace are stated below:
"Procedures for handling non-squawking VFR aircraft transiting Sheppard airspace are governed by FAA Order 7110.65. Traffic advisories are issued to aircraft Sheppard ATC is controlling based on controller workload and equipment limitations. If a controller determines that an aircraft on his/her frequency will be in conflict with any other known or observed traffic, the aircraft on the controller's frequency will receive a traffic advisory (workload permitting)."
Neither the T-37B nor the AT-502B was equipped with any type of flight recorder.
WRECKAGE AND IMPACT INFORMATION
The main wreckage of the Air Tractor AT-502B was located in an open field, approximately 3.5 miles east of Hollister, Oklahoma. The coordinates recorded at the accident site using a hand held Global Positioning System (GPS) unit were 34 degrees 20.731 minutes North latitude and 098 degrees 48.809 minutes West longitude, at a field elevation of approximately 1,067 feet msl. The wreckage debris field was approximately 6,900 feet long and 3,200 feet wide. The airplane impacted soft terrain on a magnetic heading of 355 degrees, and came to rest upright on a heading of 355 degrees, approximately 2,750 feet and 310 degrees from the T-37B. The impact of the airplane indicated little or no forward or lateral motion (pancaked). The post impact fire consumed the engine compartment, cockpit, and inboard portions of the left and right wings.
Examination of the airplane revealed that the right wing remained partially attached to the fuselage attach point. The aileron remained attached to its attach points. The aileron boost tab was found separated from its hinge points lying just below the aileron. The push pull rod from the aileron boost tab was found connected to the tab arm and to the wing hard point. The flap remained partially attached and was buckled throughout. The inboard mount for the flap was consumed by fire. The wing section inboard of the fuel filler cap was consumed by fire. The leading edge skins were deflected upwards from the wing spar as were the center skins. The fiberglass wingtip and wingtip skin remained partially attached with numerous rivets sheered.
Control continuity was established from the right inboard to the area of the fuel filler cap. The push tube fitting remained attached to the bellcrank. All the pushrods, rod end bearings, and bolts, were found properly assembled to the control stick torque tube bellcrank. The flap push rod was found attached to the flap hinge. The forward end of the flap push rod had the bolt that attaches the rod to the flap arm assembly installed. The flap arm assembly was consumed by fire. The bolts attaching the flap arm to the flap control tube was found in the end of the flap control tube.
The right horizontal stabilizer was found separated from its attach point. The holes on the inboard stabilizer attach brackets were found elongated. However, the stabilizer was free of structural damage. The V-strut supporting the stabilizer was found attached to the fuselage and stabilizer. The right elevator remained attached to the stabilizer. The elevator trim boost tab was found attached. The boost tab control rod was found assembled. The top of the stabilizer and elevator exhibited sooting, however no sooting was observed on the lower part of the surfaces.
The front spar of the vertical stabilizer was found attached to the airframe. The upper rear spar attach point was found assembled. The lower rear spar attach bolts were separated, but found in the vertical fin. The rudder remained attached to the vertical stabilizer at all three hinge points. The rudder trim boost tab remained attached to the rudder and the tab rod remained attached to the tab horn, and the fuselage attach point. Two holes within the lower section of the rudder were observed on each side of the rudder. These holes were consistent with impact from the inboard end of the elevator boost trim tab.
The left horizontal stabilizer was found separated from its attach point. The holes on the inboard stabilizer attach brackets were found elongated. The left stabilizer was wrinkled throughout its respective span. The V-strut supporting the stabilizer was found partially attached to the fuselage and stabilizer. The left elevator remained attached to the stabilizer. The elevator trim boost tab was found attached. The elevator was buckled in the area of the outboard attach hinge. The boost tab control rod were found assembled. Streaks of sooting were observed trailing from the leading edge aft on the top and bottom of the surfaces.
Rudder control continuity was established from the rudder torque arm to the rudder pedals. The rudder control cables were traced forward to the rudder pedal attach bracket. The attach bracket was consumed by fire. The bolts and T pins that attach the bracket to the cable and rudder pedal arm were found installed. Continuity to the elevators was established from the control stick to the idler aft of the cockpit. The forward fitting of the main push pull tube was found installed in the idler aft of the cockpit. The main elevator push pull tube was mostly consumed by fire. The aft fitting of the main elevator push pull tube was found installed in the aft idler. The aft elevator push pull tube remained intact and was installed from the aft idler to the elevator torque tube horns.
The left wing remained partially attached to the fuselage attach point. The aileron remained attached to its attach points. The aileron boost tab was found attached to its hinge points. The push pull rod from the aileron boost tab was found connected to the tab arm and to the wing hard point. The flap remained partially attached and was buckled throughout. The inboard and center mount for the flap were consumed by fire. The wing section inboard of the aileron bellcrank was consumed by fire. The leading edge skins were deflected upwards from the wing spar as were the center skins. The center skins were wrinkled from the outboard aft corners to the inboard forward corners.
Control continuity was established from the left aileron to the area inboard of the aileron bellcrank. The aileron push pull tube from approximately the center of the fuel cell to the inboard end was consumed by fire. The push tube fitting remained attached to the bellcrank. All the pushrods, rod end bearings, and bolts, were found properly assembled to the control stick torque tube bellcrank. The flap push rod was found attached to the flap hinge. The forward end of the flap push rod had the bolt that attaches the rod to the flap arm assembly installed. The flap arm assembly was consumed by fire. The bolts attaching the flap arm to the flap control tube were found in the end of the flap control tube.
The forward section of the fuselage frame was damaged. Most of the forward tubular members in the forward section of the fuselage were bent, buckled, or broken. The forward most section of the left lower longeron was separated and located approximately 30 yards east of the main wreckage. The longeron exhibited a bend inboard throughout its length. The cockpit tubular structure was damaged with the cockpit floor compressed downward. Most of the cockpit tubular members in the forward section of the fuselage were bent, buckled, or broken. The right and left cockpit frame doorposts were broken from the overhead structure with the windshield and doorposts compressed aft. All instrumentation within the cockpit area was destroyed due to fire damage. The aft section of the fuselage frame was damaged. A few of the tubular members of the aft section of the fuselage were buckled.
The left main landing gear spring was found bent aft and upwards underneath the airframe. The paint in the area of the upper elbow of the left main gear spring was scratched. The wheel and tire remained attached to the left main landing gear spring. The right main landing gear spring was separated. The wheel was separated from the right main landing gear spring and found approximately 200 yards east of the main wreckage.
The engine remained attached to the engine mount ring. All tubular members of the engine mount were bent, buckled, or broken. The compressor case of the engine was broken, exposing the first stage compressor wheel and blades. The burner can and exhaust duct were damaged. The upper left section of the exhaust duct was compressed aft. The outer turbine case was deformed, inboard of the turbine case, and power turbine blades were broken off at the root. Score marks were observed on the outer turbine case. First stage fan blades were rubbed in the opposite direction of rotation.
The propeller gearbox remained attached to the exhaust duct. The propeller remained attached to the output shaft of the gearbox. The engine control cables were found attached to the cockpit control quadrant, however they were burnt. Continuity of the propeller governor was established from the control quadrant forward to the swivel end. The forward propeller governor control rod was attached to the governor arm, but broken at the aft end.
All three propeller blades remained attached to the propeller hub. The outboard tip of one of the three propeller blades was separated. A gouge was in the trailing edge of the propeller blade just below the area where the blade tip separated. Chordwise scratching was also observed. On the second propeller blade, the blade tip was separated and curled opposite the direction of rotation. The blade also displayed chordwise scratching and gouging. The third propeller blade tip was separated approximately 18 inches outboard of the Hartzell sticker and displayed deep gouges and chordwise scratching.
The main wreckage of the Cessna T-37 was located in an open field, approximately 3.5 miles east of Hollister, Oklahoma. The GPS coordinates recorded at the accident site were 34 degrees 20.473 minutes North latitude and 0989 degrees 48.096 minutes West longitude, at a field elevation of approximately 1,067 feet msl. The airplane impacted soft terrain on a magnetic heading of 067 degrees, and came to rest inverted with little or no forward or lateral motion (pancaked). A post impact fire consumed the cockpit, the fuselage from the wing leading edge to the empennage, the right engine nacelle, and the left and center fuel cells.
Examination of the T-37B revealed that the fuselage of the airplane was crushed to approximately 23 inches in height. Yellow paint transfer was observed on the right side of the fuselage on six exterior screws just below the windscreen. The cockpit and surrounding area was destroyed by fire.
The right wing was fragmented into multiple pieces. Fragments of the right wing were found throughout the wreckage distribution area to the north-northeast area. Approximately two feet of the inboard section of the right wing was missing. The right main landing gear tire was observed torn in one area. A large indention was observed on the leading edge of the right wing, approximately two feet outboard from the wing root. Fire damage was also observed in this area. A black rub marks was also observed approximately two feet outboard of the wing root on the leading edge. An abrasion was observed on the top of the wing near the wing root.
The right horizontal stabilizer was fragmented into multiple pieces. The outboard two thirds of the horizontal stabilizer was not located during the investigation. The elevator was fragmented into three different pieces. Yellow paint transfer marks were observed in the mid point of the elevator.
The vertical stabilizer was fragmented into multiple pieces. The lower 23 inches of the rudder was found. A yellow paint transfer mark was observed on the mid point of the rudder. A tear in the rudder, approximately seven inches in length was observed on the lower right side of the rudder. Rippling was observed on the left side of the rudder in the same location.
The left horizontal stabilizer and elevator were separated into two different pieces. The most outboard portion of the elevator and stabilizer was separated approximately 26 inches outboard of the attach point. The other portion was severed at the attach point and appeared to be bent downward on the top and bottom of the component.
The left wing remained attached to the fuselage and was buckled and wrinkled throughout its entire span. The wing was bent upward at the mid span point. The left aileron was observed in the neutral position. The left flap and landing gear remained retracted.
The left engine exhibited a burned exterior. The right engine contained a burned gearbox housing and exterior. The first stage compressor blades were not visible. The rear most turbine blades on both engines were visible and no damage was evident. No turbine blade/turbine case interference was present. Fractured propeller blade tips from the AT-502B were found in the right engine nacelle. No further examination of the engines was performed.
MEDICAL AND PATHOLOGICAL INFORMATION
An autopsy of the AT-502B pilot was performed at the Office of the Chief Medical Examiner in Oklahoma City, Oklahoma, on January 19, 2005.
The FAA Civil Aeromedical Institute prepared a Forensic Toxicology Fatal Accident Report, which indicated 0.27 (ug/ml) cyanide was detected in the blood, and negative results for all other drugs and substances tested.
The T-37B featured two ejection seats. Each seat accommodates a back-type parachute, and is provided with an inertial reel-type shoulder harness, an automatic opening safety belt, and a seat-man separator. All seats have a canopy piercer on top of the seat to break the canopy for through the canopy ejection capability. Each seat features a four-point quick release restraint system. Both occupants of the T-37B were wearing standard military issue automatic opening parachutes with a zero delay deployment.
The flight instructor and student pilot successfully executed the ejection process and exited the airplane prior to it impacting terrain.
The AT-502B was equipped with a four-point harness. The buckle assembly for the combination lap belt and shoulder harness was found latched, but exhibited fire damage.
TESTS AND RESEARCH
The National Transportation Safety Board (NTSB) conducted an Air Traffic Control Radar Study of the accident flight. The FAA radar data for the flight of Cider 21 was obtained from recordings made by Standard Terminal Approach Radar System (STARS) processor located in Oklahoma City. The STARS system collects target information from a short-range (60nm) radar antenna located at Will Rogers World Airport, Oklahoma City (OKC), and two long-range radar sites located at Tinker Air Force Base and Putnam, Oklahoma. The Putnam site is a beacon interrogator only, providing no primary radar coverage. The accident site is located about 100nm southwest of Oklahoma City, which is beyond the range of the short-range OKC site. Therefore, all data supplied by the FAA was provided by the Putnam site, 102nm northwest of the accident, and the Tinker site, located about 98 miles northeast of the accident site. To supplement FAA-provided radar information, the NTSB also attempted to retrieve radar data from the Army radar site located at Ft. Sill, Oklahoma, two USAF radars located at Altus AFB and Sheppard AFB, and E-3 Airborne Warning and Control System (AWACS) aircraft based at Tinker AFB. The Altus, Sheppard, and Ft. Sill radars are not recorded, so no further data was available from those locations. However, an AWACS aircraft was airborne and within range of the accident site at the time of occurrence. USAF personnel attempted to retrieve data on the T-37B and the Air Tractor aircraft from archived mission recordings, and were able to supply a partial set of valid data that appears to be from Cider 21 just before the accident. Because of a technical fault with the radar reconstruction and archiving software, target latitude/longitude positions were corrupted and not usable. However, target altitudes and time data were retained on a limited basis, and were available for comparison to FAA radar data received during periods of overlapping FAA and AWACS coverage. The AWACS radar time source is manually set, and was reportedly restarted at least three times during the mission. As a result, the AWACS target times should be considered approximate.
According to statements from the T-37 crew, beacon codes 4000 and 4221 were both used at different points in the mission. The aircraft was assigned code 4221 for the initial part of the flight, which involved practice maneuvers in a military operating area north of FDR. At 1113:56, Cider 21 changed to code 4000 for pattern operations at FDR. According to the USAF, standard practice is for training aircraft to return to their original mission beacon code (in this case, 4221) upon return to base. Cider 21 contacted RAPCON for their return segment at 1126:32, and was advised of radar contact over Hollister by the arrival controller at 1127:03. There were no observed targets for code 4221 after 1114, nor were there any 4221 targets in the vicinity of the accident site at the approximate time of the collision, 1128.
The lowest altitude observed by the two FAA sites in the area of interest is approximately 6,000 feet true altitude. Radar altitude data shown by AWACS and the FAA is uncorrected mode C, referenced to 29.92 inches of mercury. The local altimeter setting was approximately 30.52, requiring all mode C altitudes to be corrected upward by 600 feet.
The AWACS aircraft did observe targets on beacon code 4221 at the approximate time of the collision. As noted above, the geographic position information for these targets was corrupted and irretrievably lost during the archiving and reconstruction process. However, altitude information is obtained through a separate beacon interrogation process, and was partially recovered. It was therefore possible to compare the time and altitude information reported by AWACS against FAA data from earlier in the mission.
The "valid" AWACS code 4221 altitude data for the period approaching the accident indicated that the altitude of Cider 21 at 11:27:24 and 11:27:55 was about 5,500 feet MSL. At 11:28:05, the altitude was about 4,970 feet MSL. (See Figure 3 of NTSB Air Traffic Control Radar Study) According to USAF AWACS staff, targets may be declared "invalid" for a variety of reasons, and may or may not represent correct information. "Valid" targets have passed quality checks and are considered acceptably reliable for use.
Examination of FAA and AWACS data showed no primary targets in the area of the accident site during the period of interest.
The RAPCON transcript indicated that at 11:27:50, the RAPCON controller, SPS-37AA, stated the following, "JAVIT27 Point Out approved. Traffic, CD21, 5,000, eastbound." At 11:27:57, the SPS-S2AA controller stated, "Cider 21, traffic observed." During interviews, the RAPCON controllers stated that they observed Cider 21 depicted on their radarscopes at 5,000 feet about 4nm south of Ranch, a VFR entry point for Hacker. The controllers reported that they did not observe any non-squawking primary targets displayed on their radarscopes in the airspace near the accident site.
USAF Aerospace Physiology Study
A USAF Aerospace Physiologist performed a study to determine the cockpit blind spots for each pilot based on relevant cockpit obstructions, and to determine the pilot visual acquisition/reaction time data. Based on distances and angles obtained by in-cockpit measurements, average human morphology, and measured eye to obstruction distances in normal flying positions, the following results were obtained:
The pilot's blind spot from the AT-502B left roll cage/door structure produced an eye position to obstruction angle of 49 degrees to the front edge, 56 degrees to the center, and 63 degrees to the aft edge, respectively.
The USAF instructor pilot's blind spot from the T-37B's right canopy bow produced an eye position to obstruction angle of 48 degrees to the front edge, 54 degrees to the center, and 61 degrees to the aft edge, respectively.
The USAF student pilot's blind spot from the T-37B's right canopy bow produced an eye position to obstruction angle of 73 degrees to the center (edge measurements not feasible due to cross-cockpit distances).
The total pilot reaction time was determined to be 3.7 seconds. The reaction time data is listed as follows:
Latent Reaction Time: 0.4 seconds
Stimulate retinal receptors: 0.1 second
Set up & move eye: 0.2 seconds
Foveal perception: 0.1 seconds
Recognition (0.65 - 1.5 seconds): 0.8 seconds
Decision Time (1 - 3 seconds): 2.0 seconds
Sub-total: 3.2 seconds
Movement Time: 0.5 seconds
Motor impulse: 0.1 seconds
Gross movement: 0.4 seconds
Total Reaction Time: 3.7 seconds
Based on pilot interviews and wreckage examination, the impact angle of the two airplanes was determined to be about 100 degrees. The USAF Aerospace Physiologist determined that the AT-502B's left roll cage/door structure created a blind spot that occluded the T-37B from the AT-502B's pilot's line of sight during the last 30 seconds prior to impact. He determined that the T-37B's right canopy bow did not produce a blind spot for the T-37B pilots, and that the AT-502B was not occluded during the last 30 seconds prior to impact.
A witness reported that the pilot of the AT-502B had planned to depart on the delivery flight to Huron, South Dakota, on Saturday, January 15, 2004, but because of poor weather conditions, the flight was delayed. On the morning of January 18, 2004, the pilot of the AT-502B flew a Piper Cheyenne from Marianna, Arkansas, to Olney, Texas. Prior to the flight to Olney, the pilot phoned the FAA Jonesboro Flight Service Station (FSS) two times for weather briefings. He inquired about the weather along the route of flight from Olney to Huron. During the second briefing at 0804, the weather briefer informed the pilot that the current surface observation for Wichita Falls, Texas, was 5,500 feet broken, 5 miles visibility with east winds at 4 knots. The pilot was not briefed, nor did he ask, about the status of the MOA's or the Alert Areas surrounding SPS and FDR. The pilot filed a flight plan for the flight from Marianna to Olney, but not for the flight from Olney to Huron.
The pilot arrived at the Air Tractor factory in Olney about 1035. The pilot conducted a preflight of the AT-502B. A witness reported that the pilot had a hand held aircraft radio, a hand held Garmin 295 GPS unit, and various maps prior to departure. The witness reported that the airplane's altimeter was set to Olney's field elevation of 1,275 feet. The witness reported the pilot departed at 1100.
The distance between ONY and the impact site was about 60nm. The direct course between ONY and the impact site was 001 degrees true. The magnetic course was about 354 degrees magnetic. Due to the unavailability of recorded radar data, the actual magnetic course and the altitudes flown by the pilot of the AT-502B is unknown.
The VFR cruising altitudes are prescribed in FAR 91.159. The VFR cruising altitudes for level cruising flight for 3,000 feet above the surface and below 18,000 feet MSL are as follows:
(1) On a magnetic course of zero degrees through 179 degrees, any odd thousand foot MSL altitude +500 feet (such as 3,500, 5,500, or 7,500); or
(2) On a magnetic course of 180 degrees through 359 degrees, any even thousand foot MSL altitude +500 feet (such as 4,500, 6,500, or 8,500).
The airspace at the accident location is designated as Class E airspace. Class E airspace is defined by the Aeronautical Information Manual (AIM) as: "generally, if the airspace is not Class A, Class B, Class D, and it is controlled airspace, it is Class E airspace. No specific pilot certification or specific aircraft equipment is required to fly in Class E airspace. No separation services are provided to VFR aircraft operating in Class E airspace.
The Right-of-way rules for converging aircraft are prescribed in FAR 91.113 (d). The regulation states the following:
(d) Converging. When aircraft of the same category are converging at approximately the same altitude (except head-on, or nearly so), the aircraft to the other's right has the right-of-way.
The T-37B wreckage was released to the USAF on April 12, 2005. The AT-502B wreckage was released to Frost Flying Service on April 13, 2005.
Parties to the investigation include the Federal Aviation Administration; the United States Air Force; U.S. Department of Transportation, Transportation Safety Institute; and Air Tractor, Inc.