Aerospace and Electronic Systems Magazine February 2018 - 31

Holland
the difference between the stable and the prestabilized oscillator frequency (scaled to account for the difference between P-channel frequency and the R-channel transmit frequency). Moreover, the transmit frequency (before AES Doppler compensation) will already be
too high by the same amount because of the incomplete temperature
stabilization. The net combined effect to the BFO decay observed
due the OCXO settling will therefore be doubled compared to the
case in which open-loop AES Doppler precompensation is used.
Hence, the BFO decay results for log-ons 1 to 6 shown in Figure 8
have been halved for the analysis when compared to those inferred
from the raw measured BFOs presented in Figure 7.
Recall that the two messages at 00:19:29Z and 00:19:37Z are a
log-on and a log-on acknowledgment, respectively. From the data
used to plot Figure 8, it can be determined that the log-on acknowledgment BFOs are in the range of [0, 6] Hz lower than the log-on
BFOs. In addition, the maximum difference between the log-on
BFO and the settled BFO value is in the range [17, 136] Hz. The
only log-on that doesn't appear to be approaching a settled value
in this range is log-on 1, for which it is anticipated that if more
data points were available several minutes later, as was the case
for log-ons 3, 5, 6, and 7, the settling behavior would be similar to
that for log-ons 6 and 7-bearing in mind the outage duration for
log-on 1 is probably at least 381 min and the BFO already dropped
substantially in the first 30 s after log-on.

SUMMARY OF SDU START-UP EFFECTS ON THE BFO
It is likely that the SDU in 9M-MRO lost power sometime after
00:11Z and then regained power-presumably because of auxiliary power unit (APU) start-up-before 00:19Z, leading to the SDU
making the log-on attempt at 00:19:29Z. Previous such events for
9M-MRO have shown that this results in an initially too-high BFO
for the log-on message, followed by a simple decay characteristic
to reach a steady-state BFO after several minutes. To establish the
relevance of the BFOs at 00:19:29Z and 00:19:37Z to the motion
(including descent rate) of MH370 at those times, this section has
established bounds for the likely range of steady-state equivalent
BFOs at those times. This was done by analyzing BFOs from seven
previous SDU start-up events for a period of several minutes. The
specific results obtained suggest the following:
1. The recorded BFO for the 00:19:29Z log-on was between 17
and 136 Hz higher than it would have been if the OCXO in the
SDU was in a steady state at that time.
2. The recorded BFO for the 00:19:37Z log-on acknowledgment
message was between 17 and 130 Hz higher than it would have
been if the OCXO in the SDU was in a steady state at that time.

BOUNDING THE DESCENT RATES OF MH370
The results from the previous two sections can be combined to
provide bounds on the descent rate of MH370 implied by the BFOs
from the two last SATCOM messages for the flight. In this section, it is shown how this is done for two possibilities that could
explain the attempted SATCOM log-on from 9M-MRO at 00:19Z.
In [3], the most likely cause of this log-on was stated to be a power
FEBRUARY 2018

Table 3.

The Last Two BFOs for MH370 Under Hypothesis 1
Recorded
BFO
(Hz)

BFO Range If
Start-Up Drift
Is Removed
(Hz)

Extended
Range
Considering
BFO Noise
(Hz)

00:19:29Z

182

[46, 165]

[28, 193]

00:19:37Z

−2

[−132, −19]

[−150, 9]

Time
Stamp

interruption resulting from insufficient fuel and subsequent engine
flameout. It could also have been because of a temporary software
failure, a loss of systems providing critical input to the SDU, or a
loss of the SATCOM link because aircraft attitude was such that the
line of sight to the satellite is blocked. If it was a power interruption to the SDU caused by loss of fuel and subsequent reboot using
the APU, the SDU would be without power for about 1 min. In this
case, the summary of SDU start-up effects on the BFO discussed in
a previous section need to be considered when interpreting the last
two BFOs.13 However, if the power loss was momentary (resulting
in a reset of the SDU) or if the temporary SATCOM outage leading
to the log-on request was due to one of the other listed reasons,
there would be no warm-up drift to consider, so the effect of SDU
start-up on the BFO would not need to be applied. Both cases are
considered separately in the following two subsections, and then
combined overall bounds are presented.

HYPOTHESIS 1: SATCOM OUTAGE DUE TO INSUFFICIENT FUEL
If the SDU log-on at 00:19:29Z was due to engine flameout, followed by a restart of the SDU using power from the APU, the SDU
outage preceding the log-on would have lasted about 1 min. This
would result in some cooling of the OCXO in the SDU. The likely effect of the SDU start-up and consequent warm-up drift of the
OCXO is summarized in the previous section. Table 3 presents the
recorded BFOs and bounds on the adjusted BFOs to remove these
effects of warm-up drift. In the last column of the table, the bounds
are extended, taking into account the BFO noise bounds of [−28,
+18] Hz established in the section that reviews the BFO statistics.
As established in the previous section on the effects of aircraft
position and velocity on the BFO, the recorded BFO would be
roughly 1.7 Hz lower for every 100 fpm of descent rate. As such,
depending on whether the plane was tracking south or north (minimum or maximum expected BFOs, respectively), bounds on the
descent rate of MH370 at the times of transmission corresponding
to the last two BFOs can be determined by subtracting the values
given in the rightmost column of Table 3 from the expected BFO
values for level flight tracking south or north, dividing the result
13

Given the relatively short duration of the power-loss under Hypothesis 1, it is possible that the extent of BFO decay due to
OCXO warm-up would be less than shown for log-ons 1 to 7.
This is essentially covered within Hypothesis 2, which considers
a momentary SATCOM outage.

IEEE A&E SYSTEMS MAGAZINE

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