Re: How Satellite Pings Work
 

And @Ian W
It is the INMARSAT business to ensure that these thousands of aircraft transmissions are not mixed up. So the hypothesis that they 'tracked the wrong plane' is just not supportable.

Those who have laboured through thousands of posts this week will have heard the INMARSAT chap explaining in very clear terms that the nearest analogy is the SIM card in your mobile phone having a unique identity that carriers pick up. Similarly INMARSAT satellites pick up the unique IDs from aircraft, no confusion at all.

What if they tracked the right plane, but they messed up their calculations?

I took a look at post 7124. I have to say that I don't buy this. It basically says that the antenna on the aircraft is tied to an extremely fast computer chip that is guaranteed to send a response to the "ping" from the satellite within a few nanoseconds of receipt. I don't have any documentation to back this, but I find this extremely improbable, especially for a 1980's system (classic aero). And, more generally, no one writes networking code like this, not even in perfectly controlled conditions, let alone for a noisy 36000 km long satellite link. The computer in charge of sending the response may have other things to do, it will reply eventually, but realtime response is not guaranteed.

What I _could_ easily buy is the presumption that the satellite has a very precise clock, and the aircraft has a different clock, and the response to the ping has a timestamp that the satellite can compare against its own clock, thus estimating time of flight. We are still talking about extremely precise timing. The entire process could be rendered useless if there's an unpredictable source of lag on the order of as little as 1 millisecond between timestamping and sending/receiving, or if the clock that's attached to the Classic Aero antenna on the aircraft drifts off by 1 millisecond over the course of flight. Since this system was never designed for the purposes of tracking aircraft, there can be any number of potential unknown sources of error.

Chinese object is a 18 Mar image
Original objects are 16 Mar, and approx. 70NM NE of the latest Chinese position.

Surface drift I believe is a little North of East? i.e. wrong way.

Does not figure to me they are logically connected... if so, 1 set or other are false, and if we can get false images, or rather objects not related to MH370, seems to me P any image is MH370 is reduced?

Pings getting longer
 

Assuming longer meaning distance...
(and applying to southern red line)

..if a/c climbed back to FL350 or so and then, for whatever reason, descended gradually along the same path...pings would get longer as the a/c would be increasing it's distance from the satellite (direct line of comm at lower altitude=further from sat)(but further u are from sat, the smaller the difference becomes)

One possibility.

The other would be.... flying away, directly or tangentially, from satellite...and, if descending, would make pings even longer.

Just a thought.

Note: timestamping errors as mentioned by hamster3null #7286 would be an added problem for calculations.

The caption showing "BUTTERWORTH AB R295 200 nm" probably means 200 nautical miles from Butterworth VOR (VBT), radial 295 (VOR on the airfield).
VBT R295 200 NM is here (clearly not the point depicted on the radar plot):

VBT R295 200NM - SkyVector Aeronautical Charts

http://s28.postimg.org/7a9ibphkd/VBT_R295_200_NM.jpg

A bit late but thanks, ZAZ, for posting the link:
The sinister, scary impact of Malaysia Airlines Flight MH370 - Intelligent Travel
This is an intelligent piece of writing and I thought, bearing in mind the criticism of the amount of dross that has been posted on this thread, that is was worthwhile highlighting this paragraph: (the italics are mine)
If one is intelligent enough to spot and ignore the dross, one can learn a lot from a thread like this, as indeed I, and I am sure many others, have. Thanks to all of you.

One thing which bothers me from the very beginning of this strange "ping" discussion: who says the latency of the ping signal is the same as in clear sky and fully operational device?

These southern and northern arcs are derived from a time delay measured from a satellite and calculated straight forward. BUT, how says there is nothing obscuring the delay, i.e. covers are known to delay these signals and maybe the device aboard is not fully functional, even a torn wire will put additional latency? So, conclusion: these arcs are the max. distance from the satellite, not necessarily the most probable. With this, the plane can sit anywhere on half the world.

This speculation is not less probable as all the others, or?

BTW: 45 south? I find it hard to believe that MH370 could come that far. How many abandoned ships float there in the roaring forties?

Bayesian Theory that helped AF447
 

BBC News now carrying this easy guide to how it worked for AF447. Must admit I found this rather neat !
BBC News - MH370 Malaysia plane: How maths helped find an earlier crash

Is it possible that they are taking images at lower resolution to cover more area in each pass?

I think you will find that messages are timestamped probably synched to GPS clock.

The calculations will have been checked by a whole raft of international experts in SATCOM that have the raw data from INMARSAT. I don't think someone worked this out on the back of an envelope using iffy data and phoned it through.

VIC SES
 

Minske/SpaceJet,

I understand that the VIC SES crew heading over are qualified air observers who have been trained by AMSA and participate in aerial searches, usually with the civvy SAR aircraft.

The trained specialist Air Observers from the Vic SES have been flying out to the search area for 3 days now on Corp Jets looking.

Nothing new now, this info was posted from AMSA days ago.

Inmarsat
 

For starters, the satellite doesn't discriminate between anything. It's just a bent pipe retransmitting what it hears in an analogue sense. All demodulation and interpretation occurs on earth.

An end point in such a network would be identified by a DNIC or number, something that looks like an international phone number or an x.25 address. It certainly isn't identified by an aircraft manufacturer's serial number. If the data service isn't paid up and active (there was an early post in this thread indicating the airline didn't use the inmarsat service) who cares about keeping this number-to-plane mapping in sync. Or who cares if the non-useful bit of kit is removed for testing etc.

If the service was not paid up, there would be no polling from the network towards the terminal. There may be polling from the terminal towards the network "->can I logon"...."<-no"... Unless there is a two way flow from the network to the terminal and back to the network timing is going to be very hard to establish. That the range ring lay exactly on the 40 degree contour is telling of the precision the author felt they had. This wasn't 41.5deg, or 42deg, it was more like vaguely in the vicinity of 40deg.

That the inmarsat range calculations were calibrated by the "known" location of the aircraft pre-disappearance is also troublesome. A vague range ring drawn through the place the sat terminal was assumed to be.

Look, I'm accepting that this hypothesis can go right out the window if a few bits of information (like it being and inactive sat service) turn out to be incorrect. There might also be non-public information clearly discrediting this theory. But to me, fire causes loss of comms, then loss of control sounds more plausible than deliberately evades radars to end up in the southern ocean.

They actually say 'debris' in inverted commas, suggesting it's not actual fact. Besides, whatever it is could well be debris of some sort.

I think AT1 in his post https://www.pprune.org/showthread.php?p=7124 explains that the satellite sends out a precise timestamp as part of its ping and the receiver returns it along with the receiver's unique iD, this allowing the lag to be precisely calculated irrespective of how good the clock on the receiver is.

How satellite pings work.
 

The low data rate inmarsat services that work with omnidirectional antennas are not as sophisticated as the poster may imply. Inmarsat-C is a teletype era service.

The 'high quality' Google images are from aircraft not satellites. If you look at GE images in the outback you will see the resolution of the civil satellite images.

On the question of Ping latency, this is not an ip type network ping (which can have lots of reasons for latency), but appears to be a clock synchronisation ping to get the right time division for the TDM to work. As such it will be implemented in hardware not software, so it is unlikely there is anything to influence that latency. It is also likely to be only a rough time synch (say 10 to 100 Microseconds - which would be a 6 to 60 mile precision - so altitude changes would not meaningfully change the time delay.

Resolution
 

This is supposed to be a comparison of satellite resolutions

IMINT - Coverage - Resolution

Globalstar images are said to be 1m or 50cm I think.

Previous posters have incorrectly looked upon the aircraft's Inmarsat receiver clock as if it was a free-running quartz wall clock. It's not free running, its locked synchronous to the incoming bit stream from the satellite (which in turn locks it's own highly accurate caesium or rhubidium clock to other satellites in the constellation and timing sources on the ground).

Once the receiver is synchronised to the bit stream it can start looking for a unique (nearly) repeating pattern of bits to tell it where a frame starts, once it knows where a frame starts it can start to assemble groups of bits into bytes, which will start to assemble into a packet.....etc...

photos
 

KH-11 satellite imagery from 30 years ago had a resolution of only a few inches. It is probably even better now. The problem is that thing(s) is usually looking at items of special interest, mostly in the northern hemisphere. They have very little coverage in the south and it is a real bear to retask them. Even then, the targeting can only be changed a slight amount. Satellites don't make right angle turns in orbit.
- Another problem is that the area of coverage is VERY small for each frame of imagery. Some satellite imagery covers many miles per frame, but the trade off is resolution measured in many 10s of feet. The images presented so far in this search appear to be in the latter category.