Wednesday, August 10, 2022

Loss of Control in Flight: Socata TBM700N (TBM850), N965DM; fatal accident occurred October 02, 2020 in Curfu, Genesee County, New York

Attorney Elizabeth Barnes

Aviation Accident Final Report - National Transportation Safety Board

Investigator In Charge (IIC): Kemner, Heidi

Additional Participating Entities:
Federal Aviation Administration / Flight Standards District Office; Buffalo, New York
Daher Aircraft; Paris
Hartzell; Piqua, Ohio
Bureau d'Enquetes et d'Analyses; Le Bourget
Transportation Safety Board of Canada
Pratt & Whitney Canada; Bridgeport, West Virginia

Investigation Docket - National Transportation Safety Board:


Location: Corfu, New York
Accident Number: ERA21LA003
Date and Time: October 2, 2020, 11:44 Local
Registration: N965DM
Aircraft: Socata TBM700 
Aircraft Damage: Destroyed
Defining Event: Loss of control in flight
Injuries: 2 Fatal
Flight Conducted Under: Part 91: General aviation - Personal


The airplane was in cruise flight at FL280 when the instrument-rated pilot failed to contact air traffic control (ATC) following a frequency change assignment. After about 25 minutes, and when 30 miles east of the destination airport, the pilot contacted ATC on a frequency other than the one that was assigned. He requested the instrument landing system (ILS) approach at his intended destination, and the controller instructed the pilot to descend to 8,000 ft and to expect vectors for the ILS approach at the destination airport. The controller asked the pilot if everything was “okay,” to which the pilot replied, “yes sir, everything is fine.” The controller
then observed the airplane initiate a descent. About 2 minutes later, the controller asked the pilot where he was headed, and the pilot provided a garbled response. The controller instructed the pilot to stop his descent at 10,000 ft, followed by an instruction to stop the descent at any altitude. The pilot did not respond, and additional attempts to contact the pilot were unsuccessful. The airplane impacted terrain in a heavily wooded area 17 miles from the destination airport.

All major components of the airplane were located in the vicinity of the main wreckage. Examination of the airframe and engine revealed no preimpact mechanical malfunctions or failures with the airplane that would have precluded normal operation.

The investigation was unable to determine why the pilot was not in contact with ATC for 25 minutes. The pilot’s eventual contact with ATC about 30 miles from his intended destination, while still operating at his cruise altitude, suggests a clear breakdown in awareness of his position through distraction or impairment. However, upon re-establishing contact with ATC, the pilot’s communications were clear, nominal, and timely, which did not suggest impairment or use of an oxygen mask. Additionally, in response to a direct query from ATC the pilot did not indicate any difficulty. Further, there was no sign of airframe depressurization and examination of the wreckage did not reveal deployment of the passenger oxygen masks. Toxicology results were positive for ethanol at a low level, which was likely due to post-mortem production. 

Meteorological data and a performance study indicated that the pilot initiated a descent through multiple cloud layers about 15 seconds after acknowledging the descent clearance. During the initial portion of the airplane’s descent, its airspeed and rate of descent appeared to be nominal. About 2 minutes later, excessive airspeeds, descent rates, bank angles, and pitch attitudes were achieved. The performance study depicted the airplane entering a spiral dive during which the airplane exceeded airspeed, maneuvering, structural, and autopilot limitations. At 6,000 ft above ground level, and about 10 seconds before ground contact, the airplane descended through a final cloud layer, the descent profile shallowed, and the rate of descent decreased to 6,800 ft/min before radar data ended.

In addition, there were no clearances issued by ATC that would have required the pilot to change either the airplane’s rate of descent or track about this time; however, the airplane’s proximity to the destination airport may have created a heightened sense of urgency for the pilot to descend and or configure his avionics for the approach, which may have served as an operational distraction. Although it was possible that restrictions to visibility during the descent may have affected the pilot’s ability to maintain positive airplane control, there is insufficient information to determine how or why the pilot lost control.

Probable Cause and Findings

The National Transportation Safety Board determines the probable cause(s) of this accident to be:
The pilot’s failure to maintain control of the airplane for undetermined reasons during the descent to the destination airport.

Personnel issues Aircraft control - Pilot

Factual Information

History of Flight

Enroute-descent Loss of control in flight (Defining event)

On October 2, 2020, at 1144 eastern daylight time, a Socata TBM 700, N965DM, was destroyed when it was involved in an accident near Corfu, New York. The private pilot and passenger were fatally injured. The airplane was operated as a Title 14 Code of Federal Regulations Part 91 personal flight.

The airplane was serviced with 173 gallons of Jet A fuel on the morning of the accident, which filled the tanks. The first flight departed Buffalo Niagara International Airport (BUF), Buffalo, New York, at 0747 and landed at Manchester Airport (MHT), Manchester, New Hampshire, at 0914, where the passenger was boarded. No services were received before the airplane departed on the accident flight.

Radar data provided by the Federal Aviation Administration (FAA) depicted its departure at 1019, and a climb to flight level (FL) 280 where it remained in level flight. Air traffic control (ATC) voice communication data revealed the pilot checked in with the Boston Air Route Traffic Control Center following a frequency change issued by a previous controller, and about 25 minutes elapsed before the pilot established communication with Buffalo Approach Control when he was about 30 miles east of BUF.

On the initial contact, the pilot told the radar controller he had lost communication with center and was level at FL280. About 1141, the controller instructed him to descend to 8,000 ft and to expect the instrument landing system (ILS) runway 23 approach; radar data show the airplane began its descent about this time. A few seconds later, the controller asked the pilot if “everything was okay...” and the pilot responded, “yes sir, everything’s fine.” The controller then cleared the pilot to fly direct to BUF and to expect vectors for a left downwind to the runway 23 ILS. The pilot acknowledged this clearance.

About 2 minutes later, the controller asked the pilot where he was headed, and the pilot provided a “garbled” response. The controller then instructed the pilot to stop his descent and to maintain 10,000 ft, followed by an instruction to stop his descent and to maintain any altitude. The pilot did not respond, and several additional attempts to contact the pilot were unsuccessful.

Witnesses in the area surrounding the crash site reported that the engine sound was “very loud” before the sound of impact was heard.

Pilot Information

Certificate: Private 
Age: 62, Male
Airplane Rating(s): Single-engine land
Seat Occupied: Left
Other Aircraft Rating(s): None
Restraint Used: 4-point
Instrument Rating(s): Airplane 
Second Pilot Present: No
Instructor Rating(s): None
Toxicology Performed: Yes
Medical Certification: Class 3 Without waivers/limitations
Last FAA Medical Exam: May 29, 2020
Occupational Pilot: No 
Last Flight Review or Equivalent: March 30, 2019
Flight Time: (Estimated) 960 hours (Total, all aircraft), 239.2 hours (Total, this make and model)

The pilot reported having flown 60 hours during the 6 months before his last FAA medical examination. The pilot’s logbook reflected a flight on May 20, 2019, and completion of a SimCom recurrent course for the TBM850 on February 9, 2020. A summary of the logbook review revealed the pilot’s hours of flight experience during the 4 years prior to the accident.

Aircraft and Owner/Operator Information

Aircraft Make: Socata
Registration: N965DM
Model/Series: TBM700
Aircraft Category: Airplane
Year of Manufacture: 2009 
Amateur Built:
Airworthiness Certificate: Normal 
Serial Number: 527
Landing Gear Type: Retractable - Tricycle 
Seats: 6
Date/Type of Last Inspection: December 6, 2019 Annual 
Certified Max Gross Wt.:
Time Since Last Inspection:
Engines: 1
Airframe Total Time: 1181.8 Hrs as of last inspection
Engine Manufacturer:
ELT: Installed
Engine Model/Series:
Registered Owner: 
Rated Power:
Operator: On file 
Operating Certificate(s) Held: None

According to FAA records, the pilot purchased the airplane in June 2016. It was powered by a Pratt & Whitney Canada PT6A-66D, 850-horsepower engine equipped with a Hartzell constant-speed propeller.

According to the airplane flight manual, the airplane was equipped with an altimetric valve that provided an automatic passenger mask actuation function at a cabin altitude between 13,000 ft and 14,000 ft when the oxygen switch was set to ON.

Meteorological Information and Flight Plan

Conditions at Accident Site: Visual (VMC)
Condition of Light: Day
Observation Facility, Elevation: BUF, 716 ft msl
Distance from Accident Site: 16 Nautical Miles
Observation Time: 11:54 Local
Direction from Accident Site: 265°
Lowest Cloud Condition: Few / 2000 ft AGL 
Visibility 10 miles
Lowest Ceiling: Broken / 6000 ft AGL 
Visibility (RVR):
Wind Speed/Gusts: 9 knots /
Turbulence Type Forecast/Actual:  /
Wind Direction: 250° 
Turbulence Severity Forecast/Actual:  /
Altimeter Setting: 29.97 inches Hg 
Temperature/Dew Point: 13°C / 8°C
Precipitation and Obscuration: No Obscuration; No Precipitation
Departure Point: Manchester, NH (MHT) 
Type of Flight Plan Filed: IFR
Destination: Buffalo, NY (BUF)
Type of Clearance: IFR
Departure Time: 10:18 Local 
Type of Airspace:

The 1154 recorded weather observation at BUF, located about 17 miles west of the accident location, included wind from 250° at 9 knots, visibility 10 miles, few clouds at 2,000 ft above ground level (agl) and 3,500 ft agl, broken cloud layers at 6,000 ft agl and 14,000 ft agl, temperature 13°C, dew point 8°C, and an altimeter setting of 29.97 inches of mercury.

According to high-resolution rapid refresh model sounding for the accident site around 1200, the wind at FL190 was from the west about 25 knots. Relative humidity about FL190 was 33% and the relative humidity was above 90% for the entire atmosphere above about 9,000 ft. The freezing level was identified as about 5,200 ft. Calculations made by the Rawinsonde Observation Program (RAOB) did not identify any potential for icing, clouds, or significant turbulence between about 9,000 and 27,000 ft.

Wreckage and Impact Information

Crew Injuries: 1 Fatal
Aircraft Damage: Destroyed
Passenger Injuries: 1 Fatal 
Aircraft Fire: On-ground
Ground Injuries: N/A
Aircraft Explosion: None
Total Injuries: 2 Fatal 
Latitude, Longitude: 42.971879,-78.382376

The airplane impacted a heavily wooded, swampy area. The airplane was fragmented and damaged by post-crash fire. Wreckage and components of the airplane were buried 15 ft below the surface. All of the major components were accounted for at the scene. An odor of Jet A aviation fuel was noted at the accident site by first responders.

The fuselage was fragmented by impact forces. The door locking pins on the main cabin door were extended and the door handle was in the closed position. A passenger emergency oxygen mask was located separated from the airframe, but the mask remained inside its housing. Control cable continuity could not be confirmed due to the fragmentation of the fuselage. All sections of the cables and push-pull tubes located exhibited tensile overload fractures. The autopilot aileron and trip servos were located, and all cables exhibited tensile overload failures. A majority of each primary and secondary flight control surfaces were in the fragmented wreckage.

Both the forward and aft sections of the center wing spar were bent symmetrically about 43° aft from their original positions. The left elevator trim tab was separated from the left elevator and located in four sections. The inboard section of the right elevator trim remained attached to the elevator. The elevator trim actuators were located and measured 16mm, consistent with a neutral position. The rudder was impact separated from the vertical stabilizer. The top section of the rudder and rudder trim tab were not located. The tail-cone was impact separated and located in the wreckage.

Examination of the engine revealed that the accessory gearbox and inlet case were impact separated, and they were not located. Multiple parts of the power turbine section and compressor section of the engine exhibited rotational scoring. Multiple blades were fractured, smeared, and bent the opposite direction of travel. The fracture surfaces that were not smeared exhibited features consistent with overload.

The propeller hub was impact-separated from the engine. All five blades were impact separated from the hub. A portion of each of the five blades was found in the wreckage.

Additional Information

Performance Study

A performance study was conducted using airplane weight, radar, communications, and weather data. The results indicated the airplane was in cruise flight at 28,000 ft and about 190 knots calibrated airspeed (KCAS) until 1141:17, when over a 30-second span, the nose pitched down about 10° and the rate of descent increased to 4,500 ft/minute. The descent was initiated 15 seconds after the airplane was cleared to descend from its cruising altitude. The airplane maintained its assigned heading until 1143:31. At that time, the airplane descended through 19,300 ft and accelerated through its maximum operating speed of 271 KCAS.

The study showed that the airplane then entered a right banking turn that reached nearly 90°, with a pitch-down attitude of 47°. Both airspeed and rate of descent increased in the descending turn. Airspeed reached 350 KCAS and the rate of descent peaked at 28,400 ft/min at 1144:13, and then decreased to 6,800 ft/min at the end of the data. The computed normal load factor increased to about 3.8 G during the descent. The study noted the descent profile was consistent with a spiral dive.


According to the manufacturer, the airplane autopilot engagement limitation for angle of bank was +/-45° and -15° for descending pitch angle. In addition, the autopilot operational envelope (commands limit) was +/- 25° for bank angle and -10° for descending pitch angle. Furthermore, the airplane was equipped with an emergency descent mode that, when engaged, would signal the autopilot to initiate a 90° left turn and descent up to 4,000 ft/min (depending on airplane loading, power lever positions, and weather conditions) until the airplane reached the target altitude of 15,000 ft msl.

Medical and Pathological Information

According to the autopsy report from the Office of the Medical Examiner, Monroe County, New York, the cause of the pilot’s death was multiple blunt force injuries, and the manner of death was accident. The examination was limited by the extensive injuries.

Toxicology testing performed by the FAA Forensic Sciences Laboratory detected ethanol in the pilot’s muscle tissue at 0.022 grams per hectogram (gm/hg). No other tested for drugs were detected in his muscle tissue. Toxicology testing performed for the medical examiner’s office was negative for ethanol and tested for drugs in the pilot’s muscle tissue.

Ethanol is a social drug commonly consumed by drinking beer, wine, or liquor. It acts as a central nervous system depressant; it impairs judgment, psychomotor functioning, and vigilance. Ethanol is water soluble, and after absorption it quickly and uniformly distributes throughout the body’s tissues and fluids. The distribution pattern parallels water content and blood supply of the tissue. Ethanol can be produced after death by microbial activity; sometimes in conjunction with other alcohols, such as propanol. Extensive trauma increases the spread of bacteria and raises the risk of ethanol production after death.


  1. Pilot logged many operations in/out of BUF, should have been just another familiar, routine flight. Seems like malfunctioning cabin pressurization would have generated enough alarm indications to get noticed by the pilot. Maybe a medical episode is the simple answer to the mystery.

  2. A puzzle that may never be solved, but a medical episode seems top of the list. Hypoxia appears less likely due to the TBM systems which -- if properly configured and functioning normally -- would prompt the pilot towards the correct action.

    1. Looks like about 25 minutes of NORDO, followed by "yeah, everything's alright". Something doesn't make sense.

    2. The NORDO period followed a failure of the controller at 1515Z to respond to pilot's "say again" after being told he was on the wrong frequency after the 1514Z change and "to return to his previous". The fact that he came up on BUF approach freq at 30 miles out suggests no distress or impairment.

      Seems like a simple miscommunication caused the NORDO interval. He was asked if everything was okay because controllers had not been able to reach him after the 1514Z frequency change was messed up.

    3. No way he was NORDO for 25 minutes by a simple mistake. This is an experienced IFR pilot. No way does he get a handoff, not understand the frequency, ask for clarification, not get an answer, and do nothing about it for half an hour. He’d simply find the frequency on the chart. Something happened long before the descent began.

  3. Too much money, too much aircraft, no brains.

    1. Yet, somehow he had piloted that aircraft in/out of BUF many times since mid 2016, as reflected by his logbook entries. Seems that your disparaging comment lacks factual support:

    2. What a low class thing to say, you are some kind of turd.

    3. I am astounded by this one. I have flown scores of aircraft types, including jets, turboprops, and piston pressurized aircraft since 1971. I don't have a clue as to what might have happened. The jealousy of this clown aside, it is a sad case.

    4. >The jealousy of this clown aside, it is a sad case.

      Yes, people LOVE to somehow disparage accomplished doctors and lawyers flying to implicate that themselves, a schlub who can barely fly a 152, are somehow the "real pilots."

      Not to say they aren't some bad rich pilots. But they generally aren't bad because they are rich, they are bad because they think it would be easy.

      These days I'd be surprised if a pilot is not a lawyer, doctor, businessmen, engineer or something like that in the cockpit: I just paid $10/gal for 100LL, an all time high for me.

  4. I've never flown single pilot pressurized. Can someone enlighten me about cabin pressure alarms. Shouldn't they be so loud and urgent they could wake the dead? Can the transponder be set to squawk something unique if the cabin pressure is low? Then ATC could urgently call the pilot "get on oxygen NOW. Descend immediatley!" Can autopilots be set to descend to a safe altitude? I don't want to read any more of these stories. Mercifully they may not have been aware of the last few seconds.

    1. TBM series aircraft cabin pressure system features blinking warnings, continuous chime, display of cabin pressure, built-in cabin altitude limit in the outflow valve design and automatic mask drop.

      A 2014 TBM info sheet will help you understand:

  5. I'm surprised either someone here (or NTSB) didn't reveal the typical descent profile this pilot took coming into KBUF (since it appears this is where he was based since 2016). When Approach cleared him to descend, he appears about 15 seconds later to start a 10-degree pitch down 4,500'/min descent, obviously a rather high rate of descent that would likely start pushing the plane into its NTE speed? Had he done this often (or even sometimes) in the past or was this a first? Some pilots with these fast aircraft like to start a descent and just touch the barber pole (and sometimes run up into it). Not saying that's the case here, but it would be interesting to see if he'd done this before, occasionally? Often? or never? Wouldn't NTSB have done that?

    1. NTSB wouldn't have cause to comment on previous BUF approaches unless they exhibited a descent profile of interest. Have a look at some prior approaches and see what you think:

      You can step back further through the history and find more.

  6. There are 4 tracks you show (last one doesn't appear and 1st one is actually two: one outbound from BUF to Long Island and one back to BUF). But the Not To Exceed design speed of the TBM 700 is 266 knots IAS (above which you're into the barber pole red stripes). The accident plane had an 850 turbine engine on it (essentially making it a TBM 850, but it was still a TBM 700 airframe. An 850 "claims" 320 knots true at the upper 20's FL altitudes, these flights were all lower 20's, so assume 300-310??). That said (assuming all upper-atmosphere winds were "westerly" as usual), his ground track on 1st: With-Wind-Eastward is 364 knots just after starting descent (even if FL winds were 80+ knots, doubtful but possible, then he may have been above NTE speed). Coming back Against-Wind-Westward after starting descent he's up at 282 knots groundspeed. Pretty compelling that he was a fair amount above NTE speeds after starting descent in both flights. 2nd Flight he's running Against-Wind-Westward and he's even higher at 291 knots after starting descent. 3rd flight running Against-Wind-Westward back to BUF seems almost most compelling of all at 322 knots and 339 knots when if he was against the wind even at 20 knots and safely below NTE, he'd probably be at 240 at most? Am I having a brain-fart and missing something? What do you think since you pulled this data. But it seems like he'd run it up above NTE each time he'd start a descent (pitch down, 10 degrees, as NTSB states). My gut tells me he may have been overstressing airframe each time? It's hard to make these assumptions without knowing "actual winds aloft", pressure and temp. Another TBM 700 came out of the FL's in Urbana, Ohio nearly straight down in 2021. These fast single-engine turbines sort of have a black-eye for being overstressed coming out of the FL's. Wouldn't NTSB have looked into that? They could pull all that info and make a more definitive "likelihood" that these planes maybe being flown over stressed. Anyway interested in posters thoughts who posted the tracks above. Note: I fly something similar from time to time, hence my interest. I already fly well below NTE, but am thinking of building in even more margin. Wish NTSB would dig into this one more.

    1. The track post should have pointed out that those tracks captured for this aircraft are all MLAT data (multilateration) instead of ADS-B. While ADS-B transmits GPS coordinates originating from the aircraft that are used to calculate ground speeds, MLAT relies on timing of received signals reaching receiving stations to construct position locations. The raggedy MLAT plot that results causes speed errors in the ground speed calculations that are based on those constructed positions.

      Look at this zoomed in view of the September 4th MLAT data. Notice the continuous heading changes and speed variations, artifacts of the MLAT data. The continuously varying speeds of 255, 231, 298, 254, 244 from right to left don't represent what the aircraft is doing.

      Once you learn to hover-check tracks and watch out for MLAT data, you won't be misled again by erroneous speeds such as that 298 knots seen at 14:15:29 and similar.

  7. Also throw some light-to-moderate (to a quick severe bump) turbulence while above NTE and you might easily be overstressing a spar or something else. Just my gut here. Wondering if anyone else sees it possibly that way?