An Independent Analysis of the INMARSAT-3 Data Logs

The search for MH370 An independent Analysis of the INMARSAT - 3 Data Logs (Denis Thomas)

  Introduction

This report describes the results of an independent analysis of the publicly available information related to the flight and disappearance of Malaysian Airlines Flight MH370:

* Malaysia Ministry of Transport MH370 Preliminary Report 03/2014

* INMARSAT MH370 Data Log 23 May 2014

* ATSB (Australian Transport Safety Bureau) Report AE-2014-054  26 June 2014

* ATSB (Australian Transport Safety Bureau) Report Update 8 October 2014

 

This aircraft must be found to answer numerous questions for the sake of the families of the passengers and crew, and to discover the cause of the incident so steps can be taken to ensure this does not happen again.

 

If the aircraft is not found in the present search area in the Southern Indian Ocean (no sign of any wreckage as of December 2014) then perhaps this report can help the authorities re-evaluate the INMARSAT recorded data in a new light to define further search areas. 

 

A computer flight path model was constructed which evaluated literally millions of possible flight paths for MH370, starting at KUL Kuala Lumpur International Airport, and continuing, flight leg by flight leg, "ping ring" to "ping ring", to the final "7th arc" at 00:19UTC, looking for flight paths that matched the INMARSAT recorded BTO (Burst Timing Offset) and BFO (Burst Frequency Offset) data. 

 

Any analysis or modeling exercise is based on the assumptions you make, or in other words, the scenario which is being modeled. Change the assumptions and you change the outcome.  The modeling assumptions are described as a timeline in a possible scenario for the MH370 flight. This scenario is only meant to help explain the assumptions that were used in the model and analysis, and reflect one interpretation based on limited publicly available information and speculation and is by no means meant to represent what actually happened on this flight - that is up to the Accident Investigation authorities.

 

Many different scenarios were examined, test flown in a Boeing B777 flight simulator, and discussed with a B777 pilot, looking for realistic possibilities. All I can say is that the scenario used for this modeling analysis is plausible, but certainly just guesswork.  If you assume MH370 turned south shortly after the last radar contact, with an incapacitated flight crew, and continued on a simple Autopilot setting for many hours, you end up in the southern Indian Ocean.    

 

However, if you assume that a) flight MH370 was subject to some emergency situation having just passed the IGARI waypoint which disabled all communications and initiated a turn back towards Malaysia, b) the flight crew remained in control of the aircraft for many hours doing their best to safely return the aircraft, crew, and passengers to safety and c) the INMARSAT log data is correct, just subject to alternative interpretations, then you can develop a very different flight path which takes you to a very different final resting place for MH370 on the "seventh arc".

 

 

Who am I ?  A retired Electronics and Computer Systems Engineer, President of TCS, a very small computer systems consulting company in Canada, an unlicensed amateur pilot and flight simulator enthusiast, who has followed this tragic story and would hope that this effort helps in some small way to eventually provide some closure to this saga.


 

Malaysian Airlines Flight MH370

 

Scheduled flight from Kuala Lumpur Malaysia (KUL WMKK Sepang) to Beijing China

(PEK ZBAA Chaoyang).  Departure 00:40am 8 March 2014, flight time 5½ hours.

 

227 passengers and 12 crew.

Pilot: Captain Zaharie Ahmad Shaw 53, 18,365 hours flight time

First Officer: Fariq Abdul Hamid 27, 2,763 hours flight time

Boeing B777-2H6ER registry 9M-MRO with 2 Rolls Royce Trent 892B17 engines.

Weight empty:                               304,500 lbs

Passengers/Crew/Baggage:         47,800 lbs (est)

Cargo:                                           8,000 lbs (est)

Fuel: 49,100kg                             108,250 lbs

Weight @ takeoff:                         468,550 lbs  (MLW maximum landing weight: 470,000 lbs)

Note: Endurance @ M0.83 FL350: 7h26m 3347nm (ie fuel starvation approx 08:08MYT)

Note: The cargo included 5,300 lbs of lithium batteries and associated equipment


 

SCENARIO

 

This describes a possible scenario that takes place aboard MH370 as a mechanism used to describe all of the assumptions and data used in the computer model, presented as a timeline from 00:00 MYT 16:00 UTC 8 March 2014 to 08:19 MYT. I reiterate that this is my interpretation and guesswork based on limited publicly available information and speculation and is by no means meant to represent what actually happened on this flight.

 

00:00 MYT

 

Aircraft systems are powered up by the flight crew and/or ground support personnel. This initiates the first power-on sequence with the INMARSAT-3 F1 satellite and Ground Station in Perth Australia. From the INMARSAT data logs:

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

R

16:00:00.000  00000

Power On

0x10 Req Logon

 

 

 

R

16:00:13.406  00013

Power On

0x15 Logon Ack

14820

103

 

00 F

16:00:32.406  00032

Power On

0x62 Data Ack

14820

85

2233.3

 

Legend: INX  the index reference used in the computer model

R  reference information – not used in the computer model

F  fixed waypoint – fixed time & fixed location

V  variable waypoint – fixed time but unknown location

T  heading change – time, position, heading determined by the computer

 

RADIUS is the “ring radius” in great circle nautical miles from the satellite footprint (latitude longitude) at that time, based on the conical satellite angle derived from the BTO (Burst Timing Offset) value per the formula described in the ATSB (Australian Transport Safety Bureau) Report AE-2014-054.

 

BFOs (Burst Frequency Offsets) calculations are complex and involve determination of satellite and aircraft movements and the resulting line-of-sight vector velocities which cause a Doppler shift in the uplink and downlink carrier frequencies, as well as some interesting frequency shift compensation carried out by the aircraft and ground stations.

The first few waypoints (index reference) were used to confirm the model algorithms.

KUL Gate C1 2.748N 101.712E   Runway 32R “hold short” 2.748N 101.722E

INMARSAT-3 F1 position/footprint at 16:00 0.960N 64.555E

 

Actual distance satellite footprint to KUL Gate C1:  2231 nm

Calculated distance from BTO value by the computer model: 2233.3 nm

 

Recorded BFO: 85

Calculated BFO: 84.61

​ 

 

00:28 MYT

 

MAS370 requests push back and engine start at gate C1 KUL Kuala Lumpur.

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

01 F

16:27:59.407  01679

Eng Start

0x22 Acc Req

14920

88

2234.2

 

 

 

00:41 MYT

 

MAS370 cleared for takeoff runway 32R.  Liftoff at 00:41:43

00:42:10 Cleared to FL180 direct to waypoint IGARI

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

02 F

16:41:52.907  02513

Take Off

0x22 Acc Req

14940

141

2231.4

03 F

16:42:47.907  02568

Departure

0x62 Data Ack

14920

123

2228.9

 

 

 

00:56 MYT

 

ACARS data sent.  Flightradar24 reports FL282 heading 25° speed 460 knots.

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

04 F

16:55:37.907  03338

AC climb

0x62 Data Ack

15200

156

2261.8


 

01:07 MYT

 

MH370  confirms  at  FL350.  ACARS  data  sent  indicating  location  (5.27N  102.79E)

altitude: FL350, heading: 25° speed: 468 knots, fuel: 43,800kg (96,562 lbs)

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

05 F

17:07:19.407  04039

AC cruise

0x22 Acc Req

15600

132

2305.6

 

Calculated distance from BTO value: 2305.6nm

Actual distance satellite footprint to known aircraft position: 2303nm

 

Calculated BFO: 132.66

Measured BFO: 132

 

 

01:19 MYT

 

MH370 handed off to Ho Chi Minh ATC. MH370 confirms Good Night Malaysian 370”

 

 

01:21 MYT

 

Flightradar24 reports FL350 heading 40° speed 471 knots at waypoint IGARI located at

6.948N 103.577E.

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

06 F

17:21:04.000  04864

WP IGARI

 

 

 

 

 

At 01:21:13 the radar label for MH370 disappears from the radar screen at Lumpur Radar in the Kuala Lumpur Air Traffic Control Centre. This is indicative of loss of an aircraft transponder signal.

 

 

EVENT

 

At this point something happened to MH370. We may not ever know exactly what. And here the speculation begins: and your assumptions depend on what you believe happened.

 

The following assumptions are mine (although many others share the same assumptions)

and are presented here only to explain the reasoning built into the computer flight model.

 

The key assumption is there was a serious, but not catastrophic, problem with the aircraft systems and/or cargo and that the flight crew were doing their best, based on their training and extensive experience, to handle the emergency and provide for the safety of the passengers, crew, and aircraft.  But unfortunately something did go wrong.  It seems unlikely that the aircraft suffered a catastrophic event at this point such as a fire, engine explosion, or explosive/rapid cabin decompression --- in any of these situations the flight crew would have landed at the nearest suitable airport and would not prolong the flight.

 

The assumption for this scenario is that a serious but not life threatening electrical problem occurred (perhaps due to a problem in the forward cargo hold, or a electrical component/generator failed with cascading effects which tripped multiple breakers) and/or the flight crew were presented with multiple warnings and disconnected generators and bus ties in fear of having a potential disastrous electrical fire. The B777 has a very sophisticated electrical system with multiple levels of backup, so it is highly unlikely that they lost power”. As a result of this event all radio equipment, including the transponder, appear to have been permanently disabled. Power to the satellite communications equipment was temporarily lost at this point or shortly thereafter.

 

Presumably with the Autopilot disconnected and flying on standby instruments, they concentrated  on  getting  the  aircraft  manually  stabilized  and  then  turned  southwest heading back towards Malaysia and appeared to head towards the nearest airport with a long runway, namely the Penang International Airport with a 3480m (11,410ft) runway.

 

 

01:28 MYT to 02:22 MYT

 

Between 01:28 and 02:22 an aircraft believed to be MH370 was tracked by military primary radar which does not depend on an aircraft’s transponder signals.

 

By 01:52 the aircraft was approaching Penang International and had descended to a lower altitude just south of the airport.

 

Assumptions With no radio communication and possibly no navigation radios, the flight crew may have tried to contact the Penang Airport by using a cell phone. If they could not notify airport personnel of their approach (primary operating hours at Penang are from 0700 to 2200), and had concerns about the responsiveness of emergency vehicles, and arriving near their maximum landing weight with their fuel load,  but with a perfectly good aircraft they were not having any problems handling, they may have decided it was safer to go somewhere and circle around, wait for daylight, burn off some fuel, and then return to Kuala Lumpur for a daylight visual approach and landing.

 

More assumptions  And now the biggest concern is one of avoiding a mid-air collision as the airways over the Strait of Malacca are extremely busy, and no one knows they are there, so they follow westbound aircraft to get away from the busy flight corridors in the Strait of Malacca, and transition between non-cardinal flight levels which they feel is the safest way to avoid other traffic.

 

At 02:22 the last primary radar track places MH370 at approximately 6.5N 96.5E having flown 765 nm in 1 hour 40 minutes, a reasonable average speed of 460 knots.

 

Given the track and time from Pulau Perak, the estimated course at 02:22 is 296° true and the estimated groundspeed is 476 knots (M0.80) with a light wind from the northeast. Although BFO calculations are quite insensitive to aircraft altitude, an altitude of 28,500 feet was assumed and used in the model; a non-cardinal altitude safely under FL290.


 

02:22 MYT to 02:40 MYT

 

More assumptions  Having decided to fly on and get out of congested traffic areas and circle until daylight, it would be reasonable to assume the flight crew would attempt to restore power to some of the aircraft and cabin systems, possibly by entering the E&E Bay (the Electronics and Equipment compartment) and selectively restoring power to systems. It would appear that power was restored to the SDU (Satellite Data Unit) around 02:23 causing an aircraft initiated logon at 02:25:27

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

13 V

18:25:27.421  08727

AES Logon

0x10 Req Logon

12520*

142

1900.4

R

18:25:34.461  08734

AES Logon

0x15 Logon Ack

 

273

 

R

18:27:08.404  08828

02:27 ring

0x62 Data Ack

12520

172

 

14 V

18:28:14.904  08895

02:28 ring

0x62 Data Ack

12480

143

1897.8

15 T

18:38:22.000  09420

Turn SW

 

 

 

 

* corrected for 4600 usec delay

 

There is no reason assume anything but a continuation of the flight path heading 296° and groundspeed 476 knots through this period.  The BFO calculations at 02:25 (18:25UTC) and 02:28 confirm this heading and speed.

 

Calculated BFO @ 18:25:27 (2:25MYT) 6.742N 96.119E: 142.51 (measured BFO: 142)

Calculated BFO @ 18:28:15 (2:28MYT) 6.904N 95.787E: 143.65 (measured BFO: 143)

 

The BFO anomaly at 18:27:08 (BFO=172) can be accounted for by assuming the aircraft was climbing (BFO calculated at VSI 1250fpm = 172.00) in preparation for a turn south.

 

A turn to the southwest had to occur some time between 2:29 and 2:39 as at 2:40 there was another communication with the satellite with recorded BFO values of 88 and 90 which indicates the aircraft was going in a southerly direction by that time.

 

The turn southwest at 02:38 to a heading of 205° to get away from the busy traffic areas is consistent with the flight crew keeping their course changes very simple i.e. a 90° turn. The timing and direction of the turn is not arbitrary – this is a retrofit calculated by the computer model to subsequently place the aircraft at the right time and place later on. A ground track of 206° is assumed to account for a slight wind drift.

 

 

 

02:40 MYT and BEYOND

 

At 18:39:55 and 18:40:56 an attempted Air Telephony call recorded BFO values which indicate a southwesterly flight path. Unfortunately these do not record a BTO ring value.

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

16 U

18:40:56.354  09656

C chan test

0x30 C chan test

 

90

 


 Numerous other scenarios and tracks and starting points were examined, test flown in a B777-200ER Flight Simulator, and run through the computer flight model. None yielded a better starting point than the C Channel Test waypoint described above at 02:40:56 at 7.186N 94.436E with the aircraft on a 206° track and a groundspeed of 475 knots. It is just coincidental that this location is near waypoint ANOKO - not pre-planned in the analysis.

 

The calculated BFO value at 18:40 of 91.7 differs slightly from the recorded value of 90 but is well within the bounds described in the BTO BFO Analysis section of the report.

 

 

03:41MYT

 

The next two recorded events are the logon acknowledgement “pings” at 3:41 and 4:41

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

17 V

19:41:02.906  13263

03:41 ring

0x15 Logon Ack

11500

111

1753.2

 

More assumptions  The assumption is that having gotten “off the beaten track to  avoid other aircraft the flight crew would have gone into a simple holding pattern for a couple of hours waiting for daylight before returning to Kuala Lumpur. Many alternatives were examined the only one which places the aircraft at the right time and place, and is consistent with what an experienced flight crew might have done, is described below. A straightforward holding pattern followed by a run down the west coast of Sumatra, keeping the shore line lights in sight, followed by an turn to the east, returning to KUL at first light (which was at approximately 06:50MYT on March 8 2014).

 ​​

 

The racetrack holding pattern depicted above consists of flying for 20 minutes on the southwesterly 205° heading, followed by a 180° 4 minute turn (which takes 2 minutes), another 20 minutes on the reciprocal course of 025°, etc.

 

This racetrack holding pattern is “anchored” on waypoint ANOKO this was not by design in the computer model – it just happened, but perhaps by intent by an experienced flight crew to retain a fixed navigation reference point.

 

This holding pattern places the aircraft at the right place at the right time to respond to the

19:41 (03:41MYT) and 20:41 (04:41MYT) satellite Logon Interrogation “pings”.

 

Recorded BFO at 19:41 (03:41MYT): 111

Calculated BFO track 206° at 475 knots at the crossing latitude and longitude: 111.09

 

 

 

04:41 MYT

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

18 V

20:41:04.904  16865

04:41 ring

0x15 Logon Ack

11740

141

1791.9

 

Arriving at the correct BFO value for the 20:41 crossing takes a bit more rationalization, assumptions, and explanations, as the aircraft is at the right place and the right time but going the wrong way !

 

More assumptions  At 04:40 the pilot throttles back and starts a rapid descent from 28500 in the holding pattern to much lower altitude (estimated 17500 or 11500 for the computer model) in preparation for the run down the west coast of Sumatra.

For this flight track the BFO value suggests a rapid descent from 28550 as shown below:

Recorded BFO at 20:41 (04:41MYT): 141

Calculated BFO track 25° at 470 knots VS= 0 (level flight):  217.8

Calculated BFO track 25° at 470 knots VS= -2000fpm:         169.5

Calculated BFO track 25° at 470 knots VS= -3180fpm:         141.0

 

The  flight  crew,  still  concerned  with  mid-air  collisions,  may  have  executed  a  rapid descent through standard mid-level altitudes to a lower non-cardinal altitude for the run down the west coast of Sumatra.  An altitude of 17500 ft, just below the FL180 controlled airspace, was used in the flight path model.

 

Another possible explanation is a secondary event at this point which may have triggered a new set of warnings, possibly of rapid cabin depressurization, cargo hold smoke or fire, etc. which precipitated a rapid descent, a slower speed at a lower altitude, and a decision to immediately proceed towards the nearest large airport, which at this point is Kuala Lumpur International.


 

04:57 MYT

 

There is a possibility of a sighting of MH370 at approximately 05:00 MYT from a sailboat off the north coast of Sumatra. Ms Katherine Tee on the early morning watch on the sailboat SV Aaza Dana reported seeing an aircraft at a lower than normal altitude which crossed their stern going from north to south. Some quotes from her Internet Blog:

 

“I was sitting facing the port side when I saw an orange speck at the 8 o’clock position relative to the boat. When it moved to the 7 oclock position identified as a plane with an orange glow. It disappeared towards our 5 o’clock position 5 to 10 minutes later, having crossed our stern heading north to south. I did not hear any sounds as our engine was running at the time”.

 

Excerpts from the boat’s GPS navigation logs. The log time is derived from GPS satellite data and is most probably relative to UTC+6 as that is the geographical time zone that the sailboat was in at that time; thus 02:00 log time is 04:00 MYT (which is UTC+8 based).

 

 

Log Time

MYT

Speed

Course

Position

Notes

02:00

04:00

5 kts

347°

N 37.075 E 94° 26.577

gybe

02:55

04:55

3 kts

069°

N 38.103 E 94° 27.887

motoring

03:05

05:05

3 kts

070°

N 38.302 E 94° 28.398

 

04:00

06:00

3 kts

077°

N 39.286 E 94° 31.343

 

 

 

 

If the modeled flight path is correct the aircraft would have been at the 8 o’clock position at 04:57 with a nearest approach of approximately 15 nautical miles (possibly closer) and would have disappeared six minutes later at the 5 o’clock position at 05:03 MYT.

 

Author’s speculation: the “orange speck/glow” might result from a reflection from the fuselage if the B777 landing lights located in the wings were on at the time, given the flight crews concern about possible mid-air collisions as no one knew where they were.

 

 

 

 

 

 

05:01 MYT to 05:41 MYT

 

 

 

There are actually very few flight tracks that take the aircraft from somewhere on the holding pattern to an 05:41 Ring crossing with the correct BTO BFO values while maintaining flight envelope speed ranges for this aircraft at a lower altitude. At each point and ring crossing the computer model reports a number of flight tracks that meet the  time,  distance, speed,  heading, BTO  and  BFO  constraints.  The  computer  model followed each of these tracks through subsequent possible turns and ring crossings but the majority do not meet the time, distance, speed, heading, BTO and BFO constraints.

 

The best fit modeled results indicate a turn to the southeast at 05:01:40 at 6.26N 94.00E with a groundspeed of 430 knots, turning to a heading of approximately 141° and slowing to a speed of 416 knots by 05:41 MYT.

 

By 06:41 MYT the aircraft is heading east, slowing appreciably, possibly waiting for daylight and/or in preparation for a landing approach.

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

20 V

21:41:26.905  20487

05:41 ring

0x15 Logon Ack

12780

168

1945.5

21 V

22:41:21.906  24082

06:41 ring

0x15 Logon Ack

14540

204

2187.5

 

Recorded BFO at 21:41 (05:41MYT): 168

Calculated BFO track 141° at 416 knots:  168.08

 

Recorded BFO at 22:41 (06:41MYT): 204

Calculated BFO track 100° at 247 knots:  204.25


 

It appears that the flight crew may have been navigating to standard waypoints as one of the possible tracks in each set in this timeframe intersects a waypoint (MERIM DOSIK).

 

It is interesting to note that by 05:41 MYT, at higher latitudes (which has an effect on BFO calculations) and with the INMARSAT-3 satellite now moving in a southerly direction, recorded BFO values are indicative of a flight path heading east. This was as much of a surprise to the author as it might be to everyone else, and assumes that the analysis and interpretation of the INMARSAT data in this report is somewhat accurate (which requires confirmation given the complexity of such an analysis and the limited amount of information publicly available).

 

06:41 MYT to 07:15 MYT

 

Flight path prediction between 06:41 and 07:15 and beyond to 08:11 is very complex and subjective in this scenario because there are so many possibilities and very little data.

 

A  number  of  possibilities  are  discussed  in  this  section,  all  very  speculative.  Better answers await the recovery of the aircraft and Cockpit Voice Recorder and Flight Data Recorder, which unfortunately may or may not provide much additional information when found.

 

At 23:14:01 and 23:15:02 an attempted Air Telephony call recorded BFO values which indicate a northeasterly flight path. Unfortunately these do not record a BTO ring value.

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

22 R

23:14:00.904  26041

C chan test

0x30 C chan test

 

216

 

22 U

23:15:02.032  26102

C chan test

0x30 C chan rel

15860#

219

2355.2#

 

# An approximate BTO value and “ring radius” can be estimated and interpolated by time differential between the 6:41 and the 8:11 rings. This places a “virtual 7:15 ring” very close to the eastern shore of the Malaysian peninsula, which helps to establish probable flight directions from the 6:41 ring to 7:15, and beyond to the 8:11 ring.

 

The following map shows a likely flight path if the intention was to land at first daylight at Kuala Lumpur International. The track indicates a small turn shortly after 06:41 MYT towards the waypoint BATAR, with the intention of turning left to a heading of 326° at this waypoint, which would put them into the landing pattern for runway 32L.

 

Unfortunately something seems to have gone wrong in this timeframe in this scenario which caused the aircraft to continue on past this waypoint. Potentially, loss of flight controls or with the flight crew overcome by toxic fumes or oxygen deprivation, the Autopilot carried on in a northeasterly direction until the aircraft ran out of fuel.

 

Recorded BFO values at 23:15 (07:15 MYT): 216 and 219

Calculated BFO track 78° at 23:14 at 340 knots:  216.57

Calculated BFO track 68° at 23:15 at 340 knots:  219.55


 

 

 

A speed / time / distance analysis indicates some possible flight paths (out of hundreds of possibilities analyzed in the computer model) towards the 8:11 ring.

 

The groundspeed at 06:41 was estimated to be 247 knots. To achieve an average speed of

294 knots to the 7:15 “virtual ring” the “exit” speed at 7:15 must have climbed to 340 knots, thus the calculated BFO values shown above for the 7:15 “virtual” ring crossing. The outbound 340 knot track assumes a continued left turn as indicated by the BTO values to a heading of 50° to reach the 8:11 ring at the correct time from this point.

 

Another possible flight path which takes the aircraft to a close proximity of the Oil Rig Songa-Mercur is labeled 040° --- if the Autopilot ran out of intermediate waypoints after waypoint BATAR it might have reverted to an earlier course setting (040° out of IGARI).

 

08:11 MYT to 08:19 MYT and Beyond

 

The next two recorded events are the logon acknowledgement “ping” at 8:11 and the aircraft initiated logon at 08:19, which is presumed to be an SDU (Satellite Data Unit) power-up around 08:17, following a RAT (Ram Air Turbine) automatic deployment (which supplies emergency hydraulic and electrical power) after a final engine flameout due to fuel exhaustion at approximately 08:15.

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

23 V

00:10:59.928  29460

08:11 ring

0x15 Logon Ack

18040

252

2615.9

24 V

00:19:29.416  29969

AES Logon

0x10 Req Logon

18400*

182

2656.9

R

00:19:37.443  29977

7th ping arc

0x15 Logon Ack

 

 

 

* corrected for 4600 usec delay

Recorded BFO values at 00:11 (08:11 MYT): 252

 Calculated BFO track 40° at 390 knots level flight: 224.62

Calculated BFO track 40° at 390 knots VS +1400 fpm:  252.29

 

Recorded BFO values at 00:19 (08:19 MYT): 182

Calculated BFO track 40° at 330 knots level flight:  222.11

Calculated BFO track 40° at 330 knots VS -1950 fpm:  182.82

 

The calculated BFO at 8:19 indicates the aircraft is descending at 1950 feet per minute which is within expected limits as a glideslope with no engine power and a lighter aircraft with empty fuel tanks.   The speed of 330 knots at 8:19 is based on entry:average:exit speed calculations for this leg of the flight path.

 

However, the calculated BFO at 8:11 is quite baffling – this would indicate the aircraft is climbing at 1400 feet per minute. The only explanation that comes to mind is an earlier flame-out of one engine, causing an electrical system disruption, which in turn causes an Autopilot disconnect, and with a lighter aircraft in a nose-up trim position, the aircraft climbs for a short period of time .

 

One of the most interesting flight tracks beyond the 7:15 “virtual ring” at a groundspeed of 390 knots on a heading of 40° is shown below labeled 040° terminating at 08:19 at N7.78 E108.13 having flown 3,108 nm from takeoff at Kuala Lumpur International. This takes the aircraft to a close proximity of the Oil Rig Songa-Mercur --- discussed below.

 

Flight simulations indicate that it is quite possible that MH370, with empty fuel tanks, flew for another 50 nautical miles, most likely in a slow spiral in one direction or another.

 

The map below indicates the end-of-track locations on the 7th arc” and the possible final locations of flight MH370 (shown in yellow) and a probable final location of MH370 (green circle) centered at N8.30 E108.00 if this scenario is correct.


 

 

The green circle centered at N8.30 E108.00 shows a possible final location of MH370

 

 08:15 MYT to 08:20 MYT (07:20 Vietnam time zone)

 

The 390 knot 040° track is particularly interesting because it would support an “aircraft in distress” sighting and report by Mr. Mike McKay on the oil rig Songa-Mercur off the coast of Vietnam on the morning of 8 March 2014 if the time of this sighting corresponds with this timeline.  His e-mail message is copied below:

 

 

This  small  portion  of  the  map shows  the  location  of  the  oil  rig and the estimated bearings to the “Malaysian Airlines plane” and the location of the “7th arc” (which had not been made public at the time of this sighting). The  glide path in a long shallow turn away from the oil rig would support Mr. McKay’s observation of “there was no lateral movement, so it was either coming toward our location, or going away from our location”

 It is hard to believe this is just a mere coincidence or fabrication.


 

CONCLUSIONS

 

If the aircraft is not found in the present search area in the Southern Indian Ocean (no sign of any wreckage as of December 2014) then perhaps this report can help the authorities re-evaluate the INMARSAT recorded data in a new light to define further search areas.

 

If indeed this does represent the flight path and final location of Malaysian Airlines flight MH370 then it does explain why no wreckage or debris has ever been found in the search areas thousands of miles away. It does however paint a tragic “so near yet so far” ending to a desperate situation in which everyone involved was doing their professional best.

 

Once again, the scenario presented here is only meant to help explain the assumptions that were used in the model and analysis, and reflects one interpretation based on limited publicly available information and speculation and is by no means meant to represent what actually happened on this flight - that is up to the Accident Investigation authorities.

 

Why was this aircraft not tracked by radar? It would not have registered on ATC Secondary Radar if the transponders were still malfunctioning, but may very well have been picked up by Military Primary Radar. However it would not have drawn undue attention as it would have appeared as just another commercial airliner on a flight path to or from Kuala Lumpur International.

 

If the radar tapes from 8 March 2014 are still available, re-examination of the radar data for other possible flight paths for MH370 would be a good starting point to help confirm or repudiate this and other possible scenarios.

 

This report was sent to the Malaysian Ministry of Transport, Malaysian Airlines, and to the Australian Transport Safety Bureau on November 2nd 2014.

 

One way or another, my hope is that one day the final resting place of 9M-MRO MH370 is found to answer numerous questions for the sake of the families of the passengers and crew, and to discover the true cause of this tragic incident.

 

Denis Thomas  TCS  November 2014


 

 

 


 

 

BTO BFO ANALYSIS

There are two important variables captured by INMARSAT which are in the Data Log. The BTO (Burst Timing Offset) derived from the signal transit time from/to the ground station, the satellite, and the aircraft. The BTO can be used to determine the distance from the satellite to the aircraft and thus the ring radius or “ping ring” locations of the aircraft at that time -- but of course there is no indication of where the aircraft was along this ring which spans thousands of kilometers.

 

The BFO (Burst Frequency Offset) involves a very complex calculation related to the Doppler shift of the uplink and downlink message carrier frequencies, requiring detailed knowledge of satellite location and motion, aircraft headings and speeds, and all of the compensating corrections applied to the signal. Many of the algorithms and factors involved were provided in the Australian ATSB reports and updates.

 

The BFO values were used extensively in the computer model to accept or reject possible flight paths as they crossed the “ping rings” – only tracks with aircraft headings and speeds which resulted in calculated BFOs in agreement with measured BFOs were considered.

 

This section describes the observed characteristics of measured/calculated BFO values in the INMARSAT Data Logs.

 

00:00 MYT to 00:30 MYT

 

The following chart plots all of the measured BFOs from 00:00MYT to 00:30MYT while the aircraft was stationary and on the ground at Gate C1 at KUL Kuala Lumpur Airport.

 


 

The BFO values change over time because of the satellite motion – the INMARSAT-3 F1 satellite is not in a perfect geosynchronous orbit.

 

BFO values immediately following the power-on sequence are not shown as these are not valid until the equipment warms up.

 

BFO values thus appear to have a variance of at least +/- 4 due to various factors.

 

 

00:56 MYT

 

The following chart illustrates the effect of speed on BFO values. The reference data at

00:56MYT  was  captured  by  ACARS  and  Flightradar24  and  thus  we  know  quite accurately the position, speed (460 knots), altitude (28,200 ft), heading (25°) and that the aircraft was still climbing at that time. Vertical speed, climbing or descending, has a very large effect on the BFO value – even air turbulence could thus effect the BFO value.

 

Horizontal  speed  does  have an sizeable effect although compensated for by the AES.

 

Vertical speed has a very significant effect as it is not compensated for by the AES.

 

Actual aircraft altitude has a negligible effect because of the large distances involved.

 

 

 

 

01:07 MYT

 

At 01:07 the aircraft was in level flight at a known location confirmed by MH370 and ATC communication, and reference data was captured by ACARS and Flightradar24 and thus we know quite accurately the position (5.27N 102.79E), speed (472 knots), altitude (35,000 ft), heading (25°) and fuel remaining (43,800kg  96,562 lbs).  Model calibration:

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

05 F

17:07:19.407  04039

AC cruise

0x22 Acc Req

15600

132

2305.6

 

Actual distance satellite footprint to the known aircraft position: 2303nm

Calculated distance from the BTO value: 2305.6nm (ring radius)

 

Calculated BFO: 132.66

Recorded BFO: 132


 

02:40 MYT

 

At 18:39:55 and 18:40:56 an attempted Air Telephony call recorded BFO values which indicate a southerly flight path. Unfortunately these did not record a BTO ring value.

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

R

18:39:55.354  09656

C chan test

0x30 C chan test

 

88

 

16 U

18:40:56.354  09656

C chan test

0x30 C chan rel

 

90

 

 

The polar-vector diagram below shows values calculated by the flight path model at

02:40 in the region just west-southwest of the last known location of the aircraft at 02:22.

 

The BFO values recorded at 02:40 indicate the aircraft had to be travelling in a southwesterly direction at that time, on a heading of 180° to 200° and at a speed between

430 and 480 knots.

 

This clearly illustrates why the aircraft had to be on a southern arc and not heading north at this time.  Southeasterly headings are also possible but this would not take the aircraft to the 03:41 ring crossing and are discounted.

 

 

 ​​

 

 

The  diagram  shows  the  calculated  BFO values at various compass points for the aircraft at 02:40 MYT at latitude 7.00N and longitude 95.00E at 475 knots.Only headings between 180° and 200° generate BFO values of 88-90 as recorded in the   INMARSAT   data   log   and   take   the aircraft towards the 03:41 ring. Headings of up to 208° can be considered (and were included in the computer model) when the BFO range is extended to 86 to 92 (+/- 4) as discussed in the BFO analysis of variance section of this report.

 

 

 

 

 

07:15 MYT

 

Similarly the attempted Air Telephony call recorded at 07:15MYT with BFO values of

216 and 219 is included with an estimated/interpolated ring radius of 2355 nautical miles.

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

22 U

23:15:02.032  26102

C chan test

0x30 C chan rel

15860#

219

2355.2#


 

METHODOLOGY

 

A computer flight path analysis model was constructed using a set of data variables for

26 waypoints corresponding to selected log entries from the INMARSAT data logs:

 

INX

TIME UTC     el sec

DESC

TYPE

BTO

BFO

RADIUS

R

16:00:00.000  00000

Power On

0x10 Req Logon

 

 

 

R

16:00:13.406  00013

Power On

0x15 Logon Ack

14820

103

 

00 F

16:00:32.406  00032

Power On

0x62 Data Ack

14820

85

2233.3

01 F

16:27:59.407  01679

Eng Start

0x22 Acc Req

14920

88

2234.2

02 F

16:41:52.907  02513

Take Off

0x22 Acc Req

14940

141

2231.4

03 F

16:42:47.907  02568

Departure

0x62 Data Ack

14920

123

2228.9

04 F

16:55:37.907  03338

AC climb

0x62 Data Ack

15200

156

2261.8

05 F

17:07:19.407  04039

AC cruise

0x22 Acc Req

15600

132

2305.6

06 F

17:21:04.000  04864

WP IGARI

 

 

 

 

07 V

17:28:00.000  05280

Thai Radar

 

 

 

 

08 V

17:51:15.000  06675

Penang E

 

 

 

 

09 V

17:53:03.000  06783

Penang W

 

 

 

 

10 V

18:01:49.000  07309

Pulau Perak

 

 

 

 

11 V

18:13:36.000  08016

WP VAMPI

 

 

 

 

12 V

18:22:00.000  08520

Last Radar

 

 

 

 

13 V

18:25:27.421  08727

AES Logon

0x10 Req Logon

12520*

142

1900.4

R

18:25:34.461  08734

AES Logon

0x15 Logon Ack

 

273

 

R

18:27:08.404  08828

02:27 ring

0x62 Data Ack

12520

172

 

14 V

18:28:14.904  08895

02:28 ring

0x62 Data Ack

12480

143

1897.8

15 T

18:38:22.000  09420

Turn SW

 

 

 

 

16 U

18:40:56.354  09656

C chan test

0x30 C chan rel

 

90

 

17 V

19:41:02.906  13263

03:41 ring

0x15 Logon Ack

11500

111

1751.7

18 V

20:41:04.904  16865

04:41 ring

0x15 Logon Ack

11740

141

1790.0

19 T

21:01:40.000   18100

Turn SE

 

 

 

 

20 V

21:41:26.905  20487

05:41 ring

0x15 Logon Ack

12780

168

1945.5

21 V

22:41:21.906  24082

06:41 ring

0x15 Logon Ack

14540

204

2187.5

22 U

23:15:02.032  26102

C chan test

0x30 C chan rel

15860#

219

2355.2#

23 V

00:10:59.928  29460

08:11 ring

0x15 Logon Ack

18040

252

2615.9

24 V

00:19:29.416  29969

AES Logon

0x10 Req Logon

18400*

182

2656.9

R

00:19:37.443  29977

7th ping arc

0x15 Logon Ack

 

 

 

25 R

01:00:00.000  32400

Model ends

 

 

 

 

* corrected for 4600 usec delay         # interpolated estimate


 

Legend:

 

INX    the index reference used in the computer model

R         reference information – not used in the computer model

F         fixed waypoint – fixed time & fixed location

V         variable waypoint – fixed time but unknown location

T         heading change – time, position, heading determined by the computer model

 

RADIUS is the “ring radius” in great circle nautical miles from the satellite footprint (latitude longitude) at that time, based on the conical satellite angle derived from the BTO (Burst Timing Offset) value per the formula described in the ATSB (Australian Transport Safety Bureau) Report AE-2014-054.

 

The model uses a latitude/longitude/altitude model with a spheroid earth algorithm and derives satellite position and velocity information using a curve mapping technique and the ECEF (Earth Centered Earth Fixed) satellite data provided in the ATSB Report.

 

Ring to Ring tracks were analyzed using a simpler Cartesian coordinate model which introduces a slight error (less pronounced at near equator latitudes) but significantly helps reduce the processing time for iterative calculations.   The analysis looked at ground tracks (true heading ignoring magnetic declinations which are not large in the locations analyzed) and groundspeed, with some adjustments for wind (based on very rudimentary wind condition data available for March 8th).

 

The analysis was essentially a “computer assisted” modeling process for each segment of the flight path with the model proving speed, heading, and BFO calculations for queries such as ring to ring, ring to ring to ring, point to ring, and track to ring with turns and heading changes along the track.

 

Even with all of these simplifying assumptions, there are literally millions of possible flight tracks to look at while looking for a best fit.   For example, the computer model considered over 340,000 possible flight tracks just between the 5:41 ring to the 6:41 ring.