Hacker Read top | best | new | newcomments | leaders | about | bookmarklet login

"Aviation investigators and national security officials believe the plane flew for a total of five hours, based on data automatically downloaded and sent to the ground from the Boeing Co. 777's engines as part of a routine maintenance and monitoring program."

Well, if the engines were transmitting, can't someone figure out the direction of the plane?



view as:

I guess it depends on the data transferred. For an engine you probably have data like altitude, revolutions/second, fuel intake, pressure, etc. To get a status on the engines you probably don't NEED long/lat data. So it might not be in the transmitted bursts.

Maybe there is a way of interpreting the data that tells you when/if the plane turned. e.g. "the chamber pressure in engine 2 is 5% higher than engine 1 ==> the plane was turning left at a rate of 15° per minute". Something crazy like that


I wonder if engine data like revolutions/second etc could determine if the plane landed/crashed.

I wonder if the necessary transmission origin data wasn't recorded? I bet the fbi is right now going through boeing's network trying to figure that out though. Is it bounced through satellites? You'd think a XXX million dollar plane would have a gps beacon or something though, just like atm machines.

It would be a miracle if it where hijacked and there's even a chance those people are alive.


> It would be a miracle if it where hijacked and there's even a chance those people are alive.

That's very unlikely 240 people would disappear safely without leaving any trace. SO yeah, "miracle" is appropriate.


They have a GPS beacon. The ADS-B system is exactly that. It just stopped transmitting. Apparently it can be manually turned off from the cockpit. Or it malfunctioned.

Do they know for sure it stopped transmitting? The problem with ADS-B is that it is lost over large ranges, like oceans.

A space-based system to track aircraft globally is currently in the works. http://www.aireon.com/Solutions/TheGlobalAdvantage

edit: Here is an informative ADS-B coverage map I found: http://uk.flightaware.com/adsb/coverage


> Well, if the engines were transmitting, can't someone figure out the direction of the plane?

Unfortunatley not; the flight-data transmitting systems ( ACARS and ADS-B ) that do include position data appear to have been turned-off or disabled deliberately.

The engine telemetry feed is part of Rolls-Royce's 'Power by the Hour' system, which aims to identify faults or under-performance in the engines whilst still in-flight so that an engineer can tune and correct when the aircraft arrives. It doesn't care where the aircraft is.

Though... it might include environmental data such as inlet pressure and temperature, which could help to determine altitude.


Does the data include power information? At least you could get some sort of bound on the distance it is likely to have travelled.

Good thinking; yes, fuel flow and power settings are included:

http://www.rolls-royce.com/about/technology/systems_tech/mon...


That seems to indicate that the data are transmitted via ACARS. Is there another back-channel, if ACARS were unavailable for some reason?

If they received the data that means the engines managed to communicate somehow. So that tells you the length of time the transmission unit was powered up. Also, using differential signal strength between successive communications one can generate an approximate trajectory (x km from satellite at y time) etc etc. This would surely narrow down the possible locations.

What is it transmitting to every 30 minutes (fixed: seconds)? If more than one satellite was involved, that might give direction. traceroute?

Minutes, not seconds.

Sorry about that. Fixed. Wouldn't my question still hold? It's talking to a ground network.

Why ground? I'd do that as an L-band ORBCOMM type system, since it's going on planes operating over oceans. $0.01/character.

So why don't planes do that that? You need 24 bits for 3 meter GPS accuracy (times 2). Add some extra for altitude and airplane ID (or does a unique ID come built in?) So maybe 80 bits in total.

This would be nothing for an airplane - $0.40 per hour? Why don't they do this already?


>This would be nothing for an airplane - $0.40 per hour? Why don't they do this already?

They would likely receive pushback from private jets that don't want all of their flight path transmitted back to RR.


I don't know what RR is, but you don't have to transmit to any central authority. Just to the plane owner.

Make a law that the owner has to preserve the data for one week, but doesn't have to provide it to anyone.

And doesn't the US military track every overland flight anyway?


Rolls Royce. Make a law affecting every nation on the planet? Best of luck wth that.

Actually rules made by the FAA do pretty much affect every nation. Many nations do not have their own version of the FAA and just rely on the US.

But it's even easier than that: Just make all US owned airlines require it, and any flight landing in the US. Phase it in, etc, to give time to upgrade, but it's certainly possible.


I believe that Rolls-Royce doesn't sell it's engines; it leases them instead. So engine telemetry data from many airlines is transmitted to the engine's owner, which happens to be a central organisation.

(Of course, it is a business, rather than branch of a government, and not the only manufacturer of jet engines).


And since this plane is far larger than 3 meters, it doesn't even need to be that accurate; 50 meter GPS accuracy would be good enough.

There's a cost analysis of that by Stross in the comments. Basically "way too damned expensive - ~$M - for once-a-decade events".

Care to link to it? I didn't find anything by "Stross".

My numbers were $0.40 per hour, plus a few hundred for the equipment (it's available off the shelf, don't need anything custom). Don't know how much installation would be, but certainly not millions.


Wait a minute. WTF?

If the system sends signals EXACTLY every 30 minutes, then a very careful reconstruction of signal running times (combined of course with the positions of the recieving antennae and the EXACT times the signals were recieved) can reveal information about the heading and position.


That's a good point, if the timestamps have the necessary precision and their clock drift info hasn't been lost by a resync since then. I'd guess that since exact time isn't crucial to the purpose of these messages, though, that they probably don't have very high precision timestamps. :-(

Legal | privacy