Aerospace and Electronic Systems Magazine February 2018 - 28

MH370 Burst Frequency Offset Analysis

Figure 4.

BFO errors as a function of track angle at 00:11Z.

was done to illustrate that under the assumption of level flight
during the unanswered telephone call period of 18:39Z-18:41Z,
MH370 would have been tracking in a southerly direction. The
same model can be used to examine the relationship between the
BFO error and the track angle at 00:11Z, when MH370 crossed
the sixth arc, 8 min before the last messages were received from
MH370.6 To generate the curve of BFO error vs. track angle, it is
assumed that the aircraft crossed this arc at 38.67° S, 85.11° E. A
previous sensitivity analysis has revealed the curve is relatively
insensitive to the actual crossing point. In addition, calculations
revealed that (as shown in Figure 5.6 of [2] at 18:40Z) the track
angle of the aircraft influences the BFO to a greater degree for
a faster assumed ground speed. To assess the maximum possible
effect of track angle on the BFO at 00:19Z, a ground speed of
500 knots is assumed, and the BFO error vs. track angle is then as
shown in Figure 4. Another curve assuming 450 knots of ground
speed is shown for comparative purposes. The peak-to-peak
variations of the BFO difference are similar for both considered
ground speeds.

BFO TREND DURING THE MH370 FLIGHT
The measured BFOs from 19:41Z to 00:11Z are shown in Figure 5,
together with a line of best fit. This line is extended forward to the
00:19Z log-on, resulting in an expected BFO of roughly 254 Hz.7
With reference to the track angle curves presented in Figure 4, the
00:11Z BFO error value for the southernmost track is roughly 6 Hz,
meaning the measured BFO of 252 Hz was 6 Hz lower than the
lowest value it could have been (assuming roughly level flight at approximately 450 knots). If we assume the same error value at 00:19Z
and use the extrapolated BFO, assuming a south track with level
flight and similar ground speed, the expected BFO value is 260 Hz.
6

The sixth arc crossing is considered here to isolate effects of ground
track angle variation as opposed to vertical velocity, because the
measured BFO at the sixth arc is still consistent with level flight.
BFOs observed at 00:19Z were much lower than the expected
value.

Figure 5.

Measured BFOs during the MH370 flight, with a trend line extrapolated
back to 18:40Z and forward to 00:19Z.

The linear trend line in Figure 5 is also extended back by 1 h to
the time of the first unanswered satellite telephone call to MH370.
The mean of the measured BFOs during that call attempt is also
shown. Because the mean of the measured BFOs from the 18:39Z-
18:41Z call attempt is in broad agreement with the linear trend
observed in the BFOs from 19:41Z to 00:11Z (for which the BTOs
were consistent with straight and level flight [2]), this supports the
finding in [2] that there were most likely no major turns after the
unanswered call attempt (see [2], Figure 10.5).

EFFECT OF SDU START-UP ON THE BFO
The BFOs logged for messages from the plane received at the Perth
GES at 00:19:29Z and 00:19:37Z, 8 March 2014, were much lower
than the expected value of 260 Hz. In the absence of other factors affecting the BFOs at these times, this suggests a large uncompensated
negative vertical velocity component in the BFOs. In determining
how to interpret this data, it is important to carefully consider other
known factors that could affect these BFOs. One potential factor is
an SDU oscillator start-up transient. This section analyses the range
of expected effects on the BFO caused by a period of power outage followed by a restart of the SDU, as was a possibility between
00:11Z and 00:19Z on 8 March 2014 (see [3] for further details).

DETAILED ANALYSIS OF SDU START-UP EFFECTS ON THE BFO
At 18:25:27Z, 7 March 2014, the Inmarsat GES in Perth received
a SATCOM log-on request from 9M-MRO. A series of messages
was exchanged in the following few minutes as part of a standard
log-on sequence. The BFOs over those minutes displayed somewhat unusual behavior in that (barring the first BFO) the BFO appeared to be exhibiting a transient settling behavior. This is shown
in Figure 6. It was noted in [1] that the spike in BFO observed at

IEEE A&E SYSTEMS MAGAZINE

FEBRUARY 2018

Table of Contents for the Digital Edition of Aerospace and Electronic Systems Magazine February 2018