sitemap Cdr. Donaldson's Analysis of TWA FL800 Flight Data Recorder

Analysis of TWA FL800
Flight Data Recorder (FDR)

Presented at Press Conference held at the Army/Navy Club in Washington, D.C.

Cdr. William S. Donaldson - Aviation Consultant to A.I.M.
January 8, 1998

On 8 December 1997 the NTSB released thousands of pages of documents to the media at the beginning of the public hearing in Baltimore MD. On page 42 of the NTSB's factual report of TWA FL800's FDR Tabular Data (Click for Transcription of Last Two Seconds) is a last full data block recorded @ 20:31:12 which was lined out by the NTSB. It is ironic in that it explains exactly what happened to TWA FL800.

Data from three independent instrument systems:

(1) The Pitot Static System, providing input to altimeters, and airspeed indicators.

(2) The Angle of Attack System providing input to angle of attack gauges and stall warning.

(3) Inertial Navigation System that provides heading, roll and pitch information to compasses and attitude indicators

all independently show a powerful explosion occurred outside the aircraft! Data from all four engine pressure ratio gauges, the vertical accelerometer, and the rudder and elevator control position indicators provide further proof of an external high explosive event. The fact this vital data was lined out and not discussed suggests a cover up! Capture of the Last Data Line is the Very Purpose of Placing a FDR in an Aircraft!

-- The altimeters take their outside pressure readings directly from static air ports mounted on the side of the fuselage. The 20:31:12 altimeter reading suddenly dropped 3,667 feet. This can only happen if an instantaneous 1.32 pound per square inch pressure increase occurs at the static port.

-- The airspeed indicator gets its reading by comparing ram air @ the pitot tube with static pressure @ the static ports. The 20:31:12 airspeed dropped instantly by 198 knots. Again this is indicative of a sudden overpressure at the static port.

-- The heading indicator gets its information from the inertial navigation platform (3 very high speed gyro's oriented in three planes.) @ 20:31:12 the heading jumped from 82 degrees magnetic to 168 degrees magnetic indicating that the gyro tumbled from shock to the platform and continued to precess because the analog signal from the magnetic compass on the left wing tip ceased.  The signal stopped because either the hardware was destroyed or separated from the wing by warhead detonation.  
@ 20:31:12 the Roll Angle went from 0 degrees to 144 degrees and back to 0 degrees. Also @ 20:31:12 the pitch angle suddenly increased by 4.7 degrees.
This instantaneous pitch up was the product of an outside explosion below the nose and was simultaneously validated by the -0.89 G force registered on the vertical accelerometer located AFT of the center of gravity. It recorded negative G when the tail snapped down.

-- The angle of attack indicator is driven by an 'angle of attack vane' located on the left forward fuselage near the cockpit. It measures the angle of wind striking the nose of the aircraft. @ 20:31:12 the angle of attack went from a normal 3 degrees to 106 degrees indicating a high pressure wave coming from the low left side of the nose had struck the aircraft.  Two partial data blocks recorded in approximately 1/4 sec intervals after the full data block @ 20:31:12 show the angle of attack returning to 30 degrees @ 20:31:12 1/4 and to a normal 3 degrees @ 20:31:12 1/2!

Note: The altimeter is a direct air pressure reading instrument. The difference in air pressure between 13,794 feet, what it should have read @ 20:31:12, and 10,127 feet what it did read, was 1.32 pounds per square inch higher than ambient pressure. Because of instrument lag and multiple static ports this is only the minimum possible overpressure. The aircraft was ascending at 22.5 ft per second.

-- A military warhead 12" in diameter and 14" long filled with RDX, bursting 63 feet from below and left of the pitot static port located on the left side of the nose would produce a 1.32 psi overpressure.

Note: Distance to the center of burst is critical, for example if the bottom fuselage was 10 feet closer then the static port would receive 2.35 PSI overpressure which is 1 1/2 tons per square yard, 15 feet closer would be 3.21 PSI etc.

Note: The forward fuselage upper skin failed in compression (trying to wrinkle) and the lower fuselage failed in tension (trying to pull apart). These are opposite of normal loading and consistent with the warhead detonation described above.

Note: A center wing Fuel/Air explosion can not produce a 1.3 PSI overpressure at the static ports of a flying 747.   Using the 60 PSI maximum overpressure NTSB suggested, a fuel/air, explosion would only produce a  0.006 PSI overpressure @ the static ports some 75 feet from the center wing tank. When compared to the 60 psi center wing tank, military warheads can produce over 1.5 million psi in the weapon at detonation.

Note: Explosion shock waves are limited by the speed of sound in the atmosphere which is roughly 1,100 feet/second. Because TWA FL800 was already traveling 633 feet/second (380 knots true air airspeed) through the air, a center wing tank explosion shock wave could only move forward along the fuselage at 467 feet/second requiring over 2.3 times as long to traverse the 75 feet to the static port. This makes the effective radius of a CWT explosion 177 feet instead of 75 feet or about 12 times the volume of space to fill with overpressure.

Remaining Data Blocks: The remaining data @ 20:31:12 is consistent with a shock wave hitting the aircraft, the yoke comes back 11 degrees then forward, 0.89 negative g's then 1.02 positive, the rudder pedals cycled hard right then left and centered. This last full data block was recorded at least one second before the CWT explosion cut electrical power!

Calculation of maximum overpressure potential of center
wing tank explosion measured at 747 pilot pilot/static ports

(1) Highest overpressure in tank @ 60 psi (NTSB ESTIMATE)
(2) No objects in path of blast wave
(3) 2400 ft3 tank is a 8.3 ft radius sphere (highest blast potential)
(4) Distance from blast center to static port 75.5 ft.

Divide the distance to static port from center of blast, 75.5 ft., by the radius of the tank, 8.3, = 9.09. Cube the dividend = 752.6. This determines the ratio between the original tank volume and the volume of the blast sphere measured at the static port. 1:752.6. Divide the volume multiple into the 60 psi original pressure to determine the overpressure at static port.      
60 psi / 752.6 =0.079 psi

(5) Aircraft in Flight - 380 knots true Airspeed - 13,774 feet Altitude.
Because FL800 was in flight the force of any blast originating from aft of the static ports would be radically reduced. Over pressure waves in the atmosphere are limited by the speed of sound, about 1,100 ft/sec. FL800 was already moving 633 ft./sec away from the point of explosion, so because of speed limits, the shock wave would only travel forward of the fuselage at 467 ft./sec. This slowing of the relative shock wave causes the radius of the overpressure sphere measured to the static port to increase proportionally.

-- @ 633 ft./sec the aircraft travels 0.63 ft. per ms (1/1000 sec)
-- The shock wave would take 161.6 ms to travel forward to the static ports @ 0.467 ft./ms.
     75.5 ft. / 0.467 ft. per ms = 161.6 ms
-- The blast radius would then be expressed as: 161.6 ms x 0.63 ft/ms + 75.5 ft. = 177.3ft.

177.3 / 8.3 ft. = 21.36
(21.36)3 = 9746.5
60 psi / 9746.5 = 0.006 psi overpressure

Calculation of overpressure potential of anti-aircraft
warhead blast 62.5 ft. from 747 static ports

(1) 12" Diameter, 14" Long Warhead, filled with 93 lbs. RDX
(2) Overpressure within warhead @ detonation 1,500,000 psi
(3) Weapon radius = 0.6 ft.

62.5 / 0.6 ft. = 104.16
(104.16)3= 1,130,135
1,500,000  / 1,130,135 = 1.32 psi

Note: The overpressure actually measured by TWA Fl800's Flight Data Recorder on the altitude track, 1.32 psi is identical to the overpressure created by the generic warhead described above. Also note it is 216 times greater then the maximum overpressure potential of a center wing tank explosion on TWA FL800.

Comments on "Analysis of FDR Record"

Editorial Note from M. Hull - The comments below appeared after the January 8, 1998 press conference. Comments prior to the press conference appear in The Smoking Gun section. Many of the news articles on the press conference appear in Events from January 1998.

Comment #1
Subject: Re: Cdr. Donaldson Press Conference
From: Ron Katona <>
Date: Thu, Jan 8, 1998 23:16 EST
Message-id: <>

Bardonia wrote: > Material discussed at Cdr. Donaldson's press conference held on 1/8/98 is now available in "The Donaldson File" at >

I've read the page. Unfortunately, he does not discuss the implications of the missing postamble in the previous frame of data. I brought this to his attention and to the attention of the individual who presented CDR Donaldson with the FDR anomalies. I've been quite fair and honest of my appraisal of CDR Donaldson and his investigation. Why are they ignoring this major fact? I still believe they have no axe to grind, but I'm afraid their lack of objectivity is greater than I first thought. They simply are ignoring the facts.

I feel it is fair to point out that AFAIK military planes didn't (and most still don't) carry FDRs during the time CDR Donaldson was in the Navy. The man who pointed out the apparent anomalies in the FDR data to CDR Donaldson also is not an expert in the internal workings of an FDR. He told me this himself over the phone. In fact, when I asked him about the missing postamble as pointed out in the NTSB report he did not seem aware of it. So, I wonder how well they actually read the NTSB reports.

I have withheld making judgments about the Donaldson camp and the people I've spoken to up 'till now. At this point, I'm not going to be as favorable. They are not answering the point about the missing postamble. CDR Donaldson has seen my questions about it and replied through a third party that my analysis was at least credible. He did not refute it. Why was it not mentioned in this press conference?


Comment #2 - Cdr. Donaldson Replies
Discussion of TWA FL800
FDR Data Validity for time 20:31:12
By Cdr. Donaldson USN/RET.
12 January 1998

The data line @ 20:31:12 that was crossed out by the NTSB personnel prior to distribution to the press at public hearings in Baltimore is either valid or not valid.

Let's explore the second option first.  If not valid FL800 data, there are two options:

(1) It is valid data from a previous recorded flight as the NTSB is now saying or,

(2) It is invalid data (random numbers) from a previous flight.

The first option can be immediately discarded because the data line depicts a 747 @ 10,127 feet, a minimum of 70 knots below stall speed, almost inverted, Mag Compass spinning, flying @ an angle of attack 3 times greater than stall values, engines 2, 3, and 4 at abnormal engine pressure ratio's all while experiencing negative .89 G force! Flight controls are: 11° up elevator, rudder full right!

The second option can also be discarded, because if the numbers are nonsense (random data) subsequent data would not recover to "normal" values as they did!

(1) Data not derived directly from pressure instruments all return to normal

-- Pitch drops from 8.7° to 2.2°

-- Elevator goes from 11.2° to -0.2°

-- Roll angle goes from 144° to 0, the angle prior to 20:31:12!

-- Rudder goes from 77.76° right to 0.72°, the precise trimmed position held prior to 20:31:12!

-- Angle of Attack goes from 106° back to 3° The reading prior to 20:31:12!

-- Vertical Accelerometer goes from negative 0.89 G's back to normal 1.02 G's

-- Magnetic Heading, the magnetic heading, is the product of a Gyro stabilized magnetic compass (signal coming from the left wing tip) spins left (precesses) @ about 180° sec, the analog magnetic signal was lost at precisely 20:31:12, at the same instant the left wing was severely damaged.

(2) Data derived from pressure instruments all show an instantaneous over pressure of at least 1.32 PSI.

-- Altimeter drops over 3,600 feet. The analog altimeter signal then stops.

-- Airspeed drops over 190 knots. The analog airspeed does recover to previous values!

-- The elevator and rudder excursions are in part another product of the pressure wave. Once the aircraft senses it's at low speed (100 knots), flight controls go back to full throw. At higher speeds both elevator feel and rudder ratio are restricted

Now lets assume the Data is valid for FL800.

In fact the only reason to suspect the data isn't valid is to blindly accept the NTSB's assertion that the CVR and FDR power stopped at 20:31:12 when the CVR registered the 105 millisecond loud sound. If the NTSB were correct the FDR buffer and check stroke delays would preclude the FDR capturing the "Loud Sound" event recorded instantly by the CVR.

The problem with the NTSB's position, like many others they have taken, is that physical evidence proves otherwise as follows:

(1) Both the pilots and co-pilots altimeters are electrically driven geared instruments that freeze when the power is cut. They were found in the debris field reading 13,800 ft and 13820 ft.

At 20:31:11 the FDR altimeter reading was 13,772 feet and the aircraft was climbing at 22.5 feet per second. At that climb rate it would take an additional 2,133 milliseconds for the aircraft to reach the recorded altitude of 13,820 feet. This would place the time of power failure at 20:31:13.133. Stated another way, it would mean the FDR had a full 1,333 milliseconds to record the 20:31:12 data prior to power failure! In fact it would have recorded the data exactly as it appeared on the transcript, capturing the full 20:31:12 line as well as subsets of additional data prior to 20:31:13!

(2) The NTSB's assertion that the power was interrupted during an eight bit postamble is dubious because the probability of that occurring at random is about 0.0067!

(3) The pilots and co-pilots independent clocks (run on cockpit battery power) were found in the debris field. They stopped @ 20:31:20 and @ 20:31:30. Assuming one of these clocks were accurate to within 5 seconds and reflect the time the cockpit actually was torn apart, the time of aircraft destruction should be between 20:31:15 and 20:31:35

(4) The last radar transponder transmission from FL800 was recorded @ 20:31:16 on Islip's ASR8 radar. If that time was correct at least 4,000 milliseconds of power was available after 20:31:12!

Stay Tuned,


Go to Donaldson Report on Baltimore Hearings

Go to the Donaldson File Index