Aerospace and Electronic Systems Magazine February 2018 - 29

Holland

Figure 6.

Measured BFOs following the 18:25:27Z log-on message from MH370.

this time8 was not fully understood, and while originally attributed
to a possible aircraft maneuver, it could be related to frequency
changes occurring during the log-on sequence. A subsequent study
[8] by the SATCOM subgroup of the MH370 Flight Path Reconstruction Group revealed this was most likely due to power-on
frequency drift and subsequent stabilization of the oven-controlled
crystal oscillator (OCXO) in the SDU. In [8], several SDUs were
tested to investigate the effects of power outages on the BFO during
SDU power-on, triggering a SATCOM log-on, such as occurred for
MH370 at 18:25:27Z. It was found that while the frequency settling
behavior was different for each SDU, any given SDU displayed
repeatable behavior for a fixed outage duration and similar settling
characteristics for outages of different duration.
Based on sequences of Inmarsat BFO logs for 9M-MRO concerning four log-ons after periods of SATCOM outage in the week
leading up to the accident flight, it was found (in [8]) that the SDU
in 9M-MRO resulted in a BFO that was initially too high at the time
of log-on, followed by a simple decay over the next few minutes to
a steady-state value. Regarding the log-on event at 18:25:27Z for
MH370, there was a nonzero bit error rate (BER) associated with
the log-on request at that time. The associated received signal level
and carrier-to-noise density ratio (C/N0) were also unusually low. As
such, the first BFO was deemed untrustworthy.9 This is not the only
time such a behavior was observed for the SDU in 9M-MRO. In the
analysis of BFOs presented in [2], there was an outlier in the BFO
(with an error of −170 Hz) found to have occurred for a SATCOM
message associated with 9M-MRO in 1 of the 20 previous flights of
9M-MRO. This was discarded as an outlier and not presented in [2].
It too had a nonzero BER and a low C/N0 (37.6 dBHz compared to
surrounding values between 41.5 and 42 dBHz). With the removal of
the untrustworthy BFO logged for the first message in the sequence
8

When viewed on a timescale of hours, as shown in Figure 9 of
[1], this behavior looks like a spike in BFO between 18:25Z and
18:28Z.
An alternative explanation for this outlier BFO could be the unknown conditions of the aircraft between 17:22Z and 18:25Z;
for instance, the temperature of the SDU might have been much
lower than the expected ambient temperature.

FEBRUARY 2018

Figure 7.

Measured BFOs for seven log-ons of 9M-MRO.

shown in Figure 6, the 18:25Z log-on from MH370 was determined
to follow the simple decay trend observed in other instances.
It was noted by DST Group that if the simple BFO decay
trend established after SDU power-up [8] also occurred during the
MH370 SATCOM log-on event beginning at 00:19:29Z, then it
would allow bounds to be established on the BFO that would have
been expected for MH370 if the plane was flying level at that time.
This in turn would allow the determination of bounds on the possible descent rates of MH370 during that final log-on event. To
build confidence that the simple BFO decay trend would hold for
the 00:19Z log-on event, DST Group reviewed additional Inmarsat logs for 9M-MRO, corresponding to the period 22 February to
28 February 2014. Two additional cases were identified in which
9M-MRO logged back onto an Inmarsat satellite after a sustained
period of SATCOM outage. The sequence of BFOs observed in
these two cases, along with the five considered by the SATCOM
working group in [8], are shown in Figure 7, and details about each
of the log-ons are given in Table 2.
The periods of SATCOM outages followed by log-on events
were identified from Inmarsat-provided ground-station logs by
identifying sequences of three or more unsuccessful log-on interrogations to 9M-MRO (suggesting the SDU was likely powered
off), followed by a log-on to the satellite system initiated from
9M-MRO at some later time. The exact duration of the power-off
period for the SDU was unable to be determined from the logs;
however, the time stamps associated with the unsuccessful log-on
interrogations and subsequent 9M-MRO initiated log-ons were
used to determine bounds on the outage time, as shown in the third
column of Table 2.10 Additional comments about each log-on sequence are given in the final column. While some BERs for logons 1 and 6 were nonzero, the BFOs did not appear to be outliers
and were therefore considered valid. The uncertainty window in
the length of the outage durations shown is roughly 60 min; this
relates to timers in the Inmarsat satellite ground stations that result
10

The bounds shown are derived under the assumption that the
SATCOM outage was exclusively due to a power outage. It is
also possible that the first part of some outages was due to a
blocked line of sight to the Inmarsat satellites.

IEEE A&E SYSTEMS MAGAZINE

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