Malaysia Airlines plane shot down in Ukraine near Russian border[Mod warning post#193

Maximus Alexander

smurfjed wrote: »

Maximus, would you have considered dealing with TB capacities 20 years ago?

No, but the plane wasn't shot down 20 years ago.

I don't know the specifics of what happens to commercial airliners throughout their lifetimes. Do they not upgrade these systems to keep them current? I can't speak to that.

All I can say is that the idea of generating terabytes of logs in a short time in the 21st century isn't outlandish at all. I can't say whether or not this plane was equipped to do that.

eatmyshorts

Maximus Alexander wrote: »

No, but the plane wasn't shot down 20 years ago.

I don't know the specifics of what happens to commercial airliners throughout their lifetimes. Do they not upgrade these systems to keep them current? I can't speak to that.

All I can say is that the idea of generating terabytes of logs in a short time in the 21st century isn't outlandish at all. I can't say whether or not this plane was equipped to do that.

The processor in the FMS runs at about 256k!

Jack1985

The 777 has immense amount of data, to an extent where the NTSB had a ''overflow'' of information when dealing with the Asiana incident. The amount of parameters recorded by the millisecond is staggering, so if I'm honest 7 tb's doesn't surprise me. The B777 is an absolute beauty of an aircraft, it's had a terrible year, where It hasn't been necessarily responsible for any accident in it's history (granted we are waiting to find out what happened to MH370).

-

For the rebel/putin sympathisers I thought this might be of interest,

British intelligence has evidence to suggest pro-Russian separatists discussed scattering other plane parts at the crash site.

Rebels planned to sabotage the investigation into Malaysian Airlines flight MH17 by scattering parts of other planes at the crash site, according to Sky sources. UK intelligence officials intercepted communications between pro-Russian separatists, seemingly discussing deliberate attempts to tamper with the scene of the disaster. Evidence indicates rebels talked about removing victims' remains from the fields of Grabovo, where the plane was shot down.

Sky sources in Whitehall added that the rebels also considered sending the black boxes to Moscow. Sky's Defence Correspondent Alistair Bunkall said the sources claimed the separatists wanted to confuse the situation further for investigators.

He said: "I think this will play into the fears that many people have that not all of the bodies will be returned, that parts of the plane have been tampered with in order to try and hide any evidence.

"It has been incredibly difficult for international monitors to gain access to the crash site.

"This is all a part of UK intelligence working with intelligence agencies around the world to try and build up a picture not only of what happened at the point of impact, but also what happened in the hours and days that followed on from that."

The news came as Ukraine said rebels had shot down two of its fighter jets. A spokesman for the country's military operations said the aircraft were brought down near Savur Mogila in eastern Ukraine. No information was given about those on board, but it is thought both planes were carrying up to two crew members each. Fierce fighting between Ukrainian forces and separatists has continued in the Donetsk region since the MH17 disaster.

The bodies of some of the victims have started being flown back to the Netherlands from the crash site. Two military aircraft - one Dutch and the other Australian - left Kharkiv Airport for Eindhoven Airport this morning. The rest are expected to be repatriated during the course of this week. Some 200 bodies have been handed over by the rebels but 80 are still missing and are thought to be among the flight debris at the crash site. All 298 people on board were killed when the jet was brought down in a region where Ukrainian forces have been battling separatists.

Maximus Alexander

eatmyshorts wrote: »

The processor in the FMS runs at about 256k!

Surely things like sensor data aren't routed through the FMS on the way to the FDR? I don't know the answer to this, it's a genuine question. It just seems like a massive oversight to leave a single point of failure like that.

Jack1985

http://www.liveatc.net/flisten.php?mount=eham2&icao=eham

Nobody moving in EHAM within the next two minutes. Everyone going quiet. RIP.

eatmyshorts

Maximus Alexander wrote: »

Surely things like sensor data aren't routed through the FMS on the way to the FDR? I don't know the answer to this, it's a genuine question. It just seems like a massive oversight to leave a single point of failure like that.

They don't, but I'm just giving you an idea of the kind of technological systems installed.

Maximus Alexander

eatmyshorts wrote: »

They don't, but I'm just giving you an idea of the kind of technological systems installed.

Ah but logging sensor data requires really minimal processing power. I don't suppose you happen know the capacity of the FDR? Because then this discussion would really be pointless.

Jack1985

Live feed from Eindhoven Airport - The amount of cars for each coffin really hits home how many Dutch were killed.

http://www.rte.ie/news/player/live/3/#live=listings

eatmyshorts

Maximus Alexander wrote: »

I don't suppose you happen know the capacity of the FDR? Because then this discussion would really be pointless.

The FCOM doesn't go into that amount of detail. Mr Boeing doesn't consider it a need to know for crews. I'm sure its in the AMM, but I wouldn't even know what that looks like!

eatmyshorts

Copy and paste from a Honeywell document on their FDR and CVR models.
..........,...

Audio Memory Utilization. Each narrow band audio channel records data at a rate of 3k bytes per second, for a protected memory utilization of 5,400 Kbytes per channel in 30 minutes. The wide band recording occurs at a rate of 6k bytes per second and requires a total protected memory of 43,200 Kbytes for 120 minutes of audio. Note that the critical Cockpit Area Microphone is recorded at full bandwidth (6,000 Hz) for the entire 2 Hour recording duration. In addition, the 3 crew member audio inputs are summed in a “mixed” recording channel for the total 2 hour duration, requiring another 21,600 Kbytes of memory. Thus, the total 2 Hours of audio utilizes 80 megabytes of memory.

Flight Data Memory Utilization. Flight data memory is allocated based on utilization with a standard ARINC 717 interface for flight data recorded to ensure more than 25 hours of recording duration for a standard 256 words per second, 12 bits/word, recording rate. Thus, 36 megabytes of memory are allocated for this function. In order to ensure full and complete recording for incident investigation no data compression is used in the recording of flight data information. All critical flight data information is recorded on a bit-for-bit basis.

Digital Communication Message Memory Utilization. The Digital Communication Message memory is allocated based on utilization with a standard low speed ARINC 429 interface to ensure 2 hours of recording duration for 32 bit words at a peak data rate of 15 K bits per second recording rate. No data compression is used in the recording of Digital Communication Message information. Labels of the incoming messages are filtered by means of look-up table stored in the SSDVDR’s EEPROM. The table consists of up to 256 labels. Ten (10) Megabytes of memory are allocated for this function.



'Tis all techno gobbledygook to me. !!

eatmyshorts

Full document available from link below.

http://www51.honeywell.com/aero/common/documents/Data_-_Digital_Communications.pdf

knucklehead6

Jack1985 wrote: »

Live feed from Eindhoven Airport - The amount of cars for each coffin really hits home how many Dutch were killed.

http://www.rte.ie/news/player/live/3/#live=listings

Exactly. Seeing a row of 20 hearses there, just waiting...And knowing that those hearses will be doing that trip repeatedly.

As an aside (and not disagreeing with it at all) but did ayone else think it kind of strange to see Military honours being delivered, saluting the coffins, dress uniforms etc., or was that just me?

Maximus Alexander

eatmyshorts wrote: »

Full document available from link below.

http://www51.honeywell.com/aero/common/documents/Data_-_Digital_Communications.pdf

Well that puts and end to that then.

eatmyshorts

Maximus Alexander wrote: »

Well that puts and end to that then.

I'm even surprised by the low amount of memory.

80MB for audio
36MB for flight data
10MB for digital comms

Cookiemunster

knucklehead6 wrote: »

As an aside (and not disagreeing with it at all) but did ayone else think it kind of strange to see Military honours being delivered, saluting the coffins, dress uniforms etc., or was that just me?

I'd say it's because they're the first of the bodies to be returned. It's a national day of mourning and the King and Queen are there.
I think there's only around 30 coffins on each flight, so it'll take a few day to get them all back to the Netherlands.

Maximus Alexander

eatmyshorts wrote: »

I'm even surprised by the low amount of memory.

80MB for audio
36MB for flight data
10MB for digital comms

I'm not too surprised about the audio or digital comms, but I'm stunned by the flight data. Probably because I was imagining a few hundred sensors logging data 60 times a second or similar, but I guess they must not need that kind of resolution for their purposes.

smurfjed

Eatmyshorts, you beat me to it :) I really didn't expect it to be that small.

Calina

This actually makes sense to me at a high level

eatmyshorts wrote: »

Flight Data Memory Utilization. Flight data memory is allocated based on utilization with a standard ARINC 717 interface for flight data recorded to ensure more than 25 hours of recording duration for a standard 256 words per second, 12 bits/word, recording rate. Thus, 36 megabytes of memory are allocated for this function. In order to ensure full and complete recording for incident investigation no data compression is used in the recording of flight data information. All critical flight data information is recorded on a bit-for-bit basis.

Based on this, it looks to me as though they write out 256*12 bits every second. A bit will be a 1 or 0 and a word will look something like this:

1100 1110 1111

It's entirely possible that the component whose data is being written out is identified not in the datastream but by its location in the datastream and it's possible that a lot of components only have one bit to describe something like on or off. 256*12*60*60 is 276 480 000 bits; 36 megabytes = 288 000 000 bits so yeah, they have storage for more than 25 hours there.

12 bit words are new on me - I'm used to four byte (ie 32 bit) words.

Jack1985

Clear evidence today of the propaganda put out by Russia Today, UK's ofcom has announced an immediate investigation:

eatmyshorts

Calina wrote: »

This actually makes sense to me at a high level



Based on this, it looks to me as though they write out 256*12 bits every second. A bit will be a 1 or 0 and a word will look something like this:

1100 1110 1111

It's entirely possible that the component whose data is being written out is identified not in the datastream but by its location in the datastream and it's possible that a lot of components only have one bit to describe something like on or off. 256*12*60*60 is 276 480 000 bits; 36 megabytes = 288 000 000 bits so yeah, they have storage for more than 25 hours there.

12 bit words are new on me - I'm used to four byte (ie 32 bit) words.

This series of posts demonstrates just how this type of thread should be proceeding.
To me, all the technical posts about bytes and bits and data words is bewildering. Goes right over my head, but then that's why I'll defer to you guys for this stuff.
I can imagine just how the nuances of flight ops, which seem simple to me, are lost on others!!

Growler!!!

eatmyshorts wrote: »

I'm even surprised by the low amount of memory.

80MB for audio
36MB for flight data
10MB for digital comms

Is that all? Feck that's not much. I have more in my phone:eek:

Maximus Alexander

Calina wrote: »

12 bit words are new on me - I'm used to four byte (ie 32 bit) words.

It probably uses relatively ancient but robust technology, similar to how many space probes these days still use radiation hardened versions of chips that were developed in the late 80s (albeit those are 32-bit :eek:).

amen

eatmyshorts wrote: »

I'm even surprised by the low amount of memory.

80MB for audio
36MB for flight data
10MB for digital comms

compression, efficient data structures etc

eatmyshorts

amen wrote: »

compression, efficient data structures etc

But most importantly, 1990's tech.
If you read my previous post, it mentions "no compression" for the flight data information.

smurfjed

"no compression"

Honeywell state that the recordings are done without compression in order to reduce the possibility of error.

Irish Steve

I'm not surprised by those memory sizes, they are about the sort of sizes that would be needed and more significantly, they are the sort of sizes that would have been available at reasonable cost when the 777 was being designed.

The concept of data structure identifying the data is correct, and significant when saving space and more significantly, processor power is required. Most of the current generation of computer users have been misled by the manner in which Microsoft and others have bloated their operating systems and data structures in ways that only those with an eye on lucrative extra hardware sales could, while things like XML coding and structures are portable between applications, hardware and languages, they are massively inefficient in terms of storage space and CPU usage, if the application is dedicated to one specific task and device, it is more than possible to use very powerful compression and data structure techniques that make the space and power required to store the information orders of magnitude less than most people are used to. The data that is then seen on screens and displays has been processed by the application designed to read it, the raw data contained on the device would not be intelligible to a human eye without significant processing, and most modern CVR's are also digital, so they store the captured data using a similar concept to that of an MP3 file, highly compressed, and without the right program to decode it, meaningless to a human.

There are also non hardened devices on most modern aircraft, the QAR, or quick access recorder, which usually has a lot more storage space on it than the FDR, and can often record many more parameters than the FDR, configured to meet the individual customer's requirements, the downside of that being the QAR is not crash proof, so while the data on the QAR can and has been recovered at a much later date, if the aircraft crashes, then there is no protection of the data on the device, and it may well all be lost.

So, while there may well be terabytes of data moving around the aircraft in the course of a flight, the quantity of data captured by the FDR is relatively small, and that's partly down to the age of the aircraft, it's my understanding that later aircraft like the 787 and A380 have more recent devices with larger storage, and more complex recordings

And yes, it's complex, but 99.9% of the time, the crew operating the aircraft have no involvement at this sort of level, there's no need for them to be involved with it.

Jawgap

Ukraine rebel commander acknowledges fighters had BUK missile

DONETSK, Ukraine, July 23 (Reuters) - A powerful Ukrainian rebel leader has confirmed that pro-Russian separatists had anti-aircraft missiles of the type Washington says were used to shoot down Malaysia Airlines flight MH17.

In an interview with Reuters, Alexander Khodakovsky, commander of the Vostok Battalion, acknowledged for the first time since the airliner was brought down in eastern Ukraine on Thursday that the rebels did possess the BUK missile system.

He also indicated that the BUK may have originated in Russia and could have been sent back to remove proof of its presence.

the_monkey

eatmyshorts wrote: »

I'm even surprised by the low amount of memory.

80MB for audio
36MB for flight data
10MB for digital comms

36 MB in text is HUGE !!

That's why I knew the TB in guardian article gas to be a mistake

Calina

Growler!!! wrote: »

Is that all? Feck that's not much. I have more in my phone:eek:

the_monkey wrote: »

36 MB in text is HUGE !!

That's why I knew the TB in guardian article gas to be a mistake

I don't believe it's stored in text either. Binary would not surprise me.

the_monkey

Calina wrote: »

I don't believe it's stored in text either. Binary would not surprise me.

sure .. text isn't that efficient anyway.