Confusion over BTO and Ping Angles
When the Signaling Unit Logs for 9M-MRO were released, Inmarsat described the BTO as being the round trip delay "relative to a terminal at a nominal fixed location". There were two different interpretations as to where that terminal sat - one had it at a fixed point on the earth, the other had it at a fixed point from Perth (and thus acting as a single constant offset to the round-trip delay.) These two interpretations led to different predictions for ping elevation angle as a function of time, up to 2 degrees.
The first interpretation seemed to make sense because its predictions matched well with the "7 Data Point" graph that showed ping angles for several times during the flight (aside from a puzzling constant offset in angle.) At the time, this graph seemed to provide the most reliable ping angles.
However, it is now clear that this first interpretation is wrong. Equation 1 in the ATSB report can be rewritten to show that the BTO is, indeed, just the round-trip delay plus a constant bias. Tables 1 and 2 of the ATSB report reinforce this view. Table 2, in particular - BTO Calibration at Kuala Lumpur - shows that the BTO and path length were both increasing at a time when the plane was sitting on the ground. The increase in path length was consistent with the increase in round-trip path between Perth and Kuala Lumpur in Table 1, and was due mainly to the satellite-Perth leg.
As a further check, I recomputed ping angles for both the portion of the flight where we have ADS-B data and for the later portion where one can make use of the ping ring plots in Figure 18 of the report. I get consistent agreement between these ping angles and the newly-interpreted BTO, with a spread of 0.2 degrees in elevation angle. These angles differ from those on the 7 Data Point graph by a peak-peak of 2 degrees - ouch!
Thus, we conclude that the "7 Data Point graph" is flawed. Perhaps it was drawn in a rush by someone for one of the family briefings. Its purpose was to show that the plane did not go to Beijing, for which high accuracy was not needed, but was not intended for anything else.
Duncan Steel's Ping Ring radii in message #549 posted on April 5, 2014 are derived from the 7 Data Point graph and thus are wrong. Unfortunately, Richard Cole's BTO model is also the wrong one, and his Appendix B distances are likewise wrong.
For what it's worth, I was following the wrong interpretion as well, until being confounded by the Kuala Lumpur BTO measurements. Since there is interest, I will give my values for BTO and such in a follow-on message.