I had earlier posted an simple high-school-geometry analysis, which given (1) handshake timings from satellite (2) assumption those controlling plane were not idiots (i.e. flew it on near-straight course and near-max-efficiency speed after it left radar coverage, did not try to commit suicide) conclude: the plane had to have ended up near the Pakistan/Afghanistan/Kazakhstan area. Also happen to conclude that the plane's course would have taken it mainly thru probably-low-radar areas, kind of as it it had been designed that way. I never managed to get any press or government attention for this simple analysis at all, far as I could tell. In the meantime, the officials in charge concluded several times it had to be in various watery places, which kept changing but ultimately ended up in the South Indian Ocean, at a place quite similar to where my geometry-analysis had predicted -- except actually there are two geometry solutions, and they picked the southern one I'd discarded as suicide, not the sensible northern one. They spent somewhere around $50M searching for it. At one point they announced they'd heard the "pinging" sounds and if so they should have located the plane soon after. Later they, much more quietly, announced they'd decided those "pings" were not genuine and they'd somehow confused themselves into "hearing" them, or they were a hoax. No trace of the plane, or debris, or anything, has been found to this day. So anyway, at that point I had several questions whose answers I did not see in the media I read, at least not in any comprehensible way: (Q1) why did they decide those pings were not genuine? (Q2) why did they decide it took the southern route? A NOVA television-science program just happened. It did not answer Q1. As far as I can presently tell, Q1 has no answer, they never had any evidence that the pings were not genuine, but I guess they eventually decided they would have found the plane if they had been. Kind of an (inherently unsatisfying) deduction from absence of evidence. For Q2, they *did* provide a claimed answer, which was "Doppler." The plane and a satellite exchanged handshakes every hour. And those handshake timings showed, at each hour, the plane had to be on a certain circle drawn on the globe. I already knew that. The new idea is that the frequency of those signals got shifted by doppler effects depending on the relative velocities & locations of the plane & satellite. The satellite logged frequency-error measurements for troubleshooting purposes.
From these logs, they deduced that the plane had to have taken the southern route. This analysis supposedly was carried out by the satellite company Inmarsat, whereupon everybody else believed it.
However, do *we* believe it? I have been unable to find a published technical analysis on the internet. (Why was it kept secret? Why not release it for possible criticism?) Can anybody find one? The closest I found was http://aviationweek.com/commercial-aviation/routine-data-analysis-helped-pin... http://physicsbuzz.physicscentral.com/2014/03/how-did-inmarsat-really-find-f... which both are pathetically un-detailed and offer no link to any real study. I will say this. Timing handshakes seems pretty easy technically, and if you could do it with 1 nanosec accuracy (which considering my computer runs at 2 GHz, seems possible) you'd measure the satellite-plane distance to +-1 foot. So I'm totally willing to believe those timing->distance deductions. But frequency measurements seem to me to be hard technically and to distinguish whether a plane is moving toward or away from you at 600 miles/hour you need to measure frequency with accuracy better than 2 parts in a million. Actually the plane in this case would NOT be moving toward/away from satellite, but almost perpendicularly to it, so more accuracy would be needed. In the limiting case of exact perpendicular motion at 600 miles/hour (or stationary) you'd need better than 1 part in 10^12 frequency accuracy to decide 600 versus 0. None of the media sources even says what the ping radio frequency was or how long the ping signals each lasted. If the ping was at 10 MHz and lasted 1 second, then clearly you are going to get, at best, about 1 part in 10^8 frequency accuracy, and that is probably too optimistic. But even that is assuming the plane's transmitter actually has an accurate fixed frequency at all. Actually, it is probably based on a quartz crystal, whose frequency varies with temperature and pressure (and with age since manufacture). Wikipedia says "Standard-quality quartz resonators of this type are warranted to have a long-term accuracy of about 6 parts per million at 31C (87.8F) temperature." It goes on to say that by means of thermal and other kinds of compensation, more expensive kinds of quartz resonators, e.g. used in chronometers, can attain 10X better accuracy. My guess would be probably some cheesy aircraft comm pinger (not intended for anything like the purpose it is here being used for) is not going to be using a fancy compensation scheme on its quartz crystal -- but again, none of the media I found even mention this issue. (A real study should.) So... do I believe in this doppler deduction? I have doubts. I'd want to see a genuine, published, available, and independently criticized study, before I would have those doubts removed. How about some tests with other planes? Were any done? For example, instead of concluding the plane went south, a perhaps equally valid conclusion might be "the pinger got a bit colder during a couple hours." -- Warren D. Smith http://RangeVoting.org