Go Back  PPRuNe Forums > Flight Deck Forums > Tech Log
Reload this Page >

AF 447 Search to resume

Tech Log The very best in practical technical discussion on the web

AF 447 Search to resume

Old 21st Nov 2010, 22:43
  #2421 (permalink)  
Join Date: Jul 2009
Location: France - mostly
Age: 81
Posts: 1,689
RE: Loss of SATCOM. Question to our comm specialists: Does this mean that for loss of signal the airplane must be banked 67 degrees away from the satellite (assuming satellite position was 28 deg from vertical for the AC position)?

CMA-2102 Data Sheet

Jane's Avionics
Since the CMA-2102 is mounted on the top of the aircraft, its coverage pattern does not suffer from keyholes/blindspots, and installation is greatly simplified. The system conforms to ARINC Characteristic 741 and Inmarsat SDM.The CMA-2102 covers 360° in azimuth and -5 to +90° in elevation and operates over the frequency range 1,525 to 1,660.5 MHz.
HazelNuts39 is offline  
Old 21st Nov 2010, 23:36
  #2422 (permalink)  
Join Date: Jul 2009
Location: Not far from a big Lake
Age: 78
Posts: 1,459
I guess I have a question about the Beam steering unit. Does it rely on data from the aircraft's system for orientation, or does it use satellite signal for orientation? Probably a related question is whether the antenna has sufficient gain to communicate by way of sidelobes as well as the main lobe? (I.E. can it lose lock and still communicate to some limited extent.) Not really more than a baseline description in those documents.
Machinbird is offline  
Old 22nd Nov 2010, 02:20
  #2423 (permalink)  
Join Date: Jun 2009
Location: I am where I am and that's all where I am.
Posts: 660
HN39, the analysis is not quite so easy as to admit a precise angle.

If I presume there is no navigational feedback to the antenna aiming algorithms a gentle roll that leaves the antenna no more than (conservative bogey number) tilted less than about 75 degrees will be tracked and communications can transpire. A rapid roll would destroy antenna lock once the antenna fell out of the beam - say 40 to 50 degrees of roll even if the beam could be steered to point to the satellite. If there is navigational feedback then you're back to the the detail that the antenna cannot effectively steer to smaller angles to its plane than 15 degrees or so. (Somebody familiar with polar routes may correct me on precisely what latitude is the limit for SatCom capability on aircraft. My experience is with land/maritime SatCom and I know that works adequately down at angles around 10-15 degrees. That's with physically steerable antennas, though.)

I believe navigational aiding exists for the antenna. So I suspect that a rather dramatic event would be required to shadow the satellite. My memory insists the pertinent satellite was somewhere "roughly" near the African coastline. So I'd expect such an unfortunate event to be a roll with left wing down in this case, perhaps after the plane was oriented well off course to the North. ("Oh Shirt! What's that!" as the pilot pushes stick forward and to the left to avoid that suddenly appearing bright light ahead of the plane getting larger fast. (Grumble) We cannot even rule out a UFO sighting and a move to avoid striking it. Ball lightning does freaky things.)

I didn't answer your comment with a number. But I hope the data above helps visualize what it would likely take.

The downside to a conspiracy for BEA is high with regards to exposure and loss of faith in France, French airlines, and French airframes. That's so high a risk it's hard to envision them taking it.

I can envision then "cheaping out". One airframe saved is XXX megabucks. That loss has some small probability per year. (Decades before one plane went down in these circumstances?) What's the cost in airframes per year for continued search? The risk of "cheaping out" is much the same as above, though. So they are under immense pressure from two sides to not only solve it but to solve it cheaply. I can understand there being considerable deliberation before spending more money in bulk measure.

Regarding nav aid - I am SURE the antenna can self track the satellite. The marine antennas I worked on tracked fairly easily. Once I rewrote the algorithm I had a lab unit tracking within 0.1dB (about 3%) of the peak signal. I described a square in the sky. I moved it until the signal level from all four corners was equal. I then knew the antenna was pointed very near the satellite at any given time Precise pointing was worth that fraction of a decibel in signal level. The antenna itself had accelerometers that allowed it to track in remarkably rough sea states.

For the airborne antenna the antenna gain is less meaning it has a wider beamwidth. In fact it is so wide that a simple navigation feed from the plane can keep the antenna aimed properly. Attitude feedback can ensure lock even if the plane is tilted. I'd expect that. I can't guarantee that it is there. Again, it's such a simple thing to add that correction to the antenna aim that the thought of not doing it feels "dirty" to me. If it comes rapidly enough, say once every 10 seconds or less, it'd take a REALLY violent move on the part of the plane to make communications impossible.

Last edited by JD-EE; 22nd Nov 2010 at 02:33.
JD-EE is offline  
Old 22nd Nov 2010, 06:10
  #2424 (permalink)  
Join Date: Aug 2009
Location: Germany
Age: 64
Posts: 1,809


I can envision then "cheaping out". One airframe saved is XXX megabucks. That loss has some small probability per year. (Decades before one plane went down in these circumstances?) What's the cost in airframes per year for continued search? The risk of "cheaping out" is much the same as above, though. So they are under immense pressure from two sides to not only solve it but to solve it cheaply. I can understand there being considerable deliberation before spending more money in bulk measure.
Cost of a new A330-200F: $194.8 million (official price)
Cost of researches so far (as know by the press) : $28 million
jcjeant is offline  
Old 22nd Nov 2010, 08:47
  #2425 (permalink)  
Join Date: Jul 2009
Location: France - mostly
Age: 81
Posts: 1,689

Thanks for your detailed reply. As to the satellite location, BEA has established that the ACARS messages all passed through Inmarsat Atlantic Ocean West (AOR-W) which is at 53 or 54 degrees West (sources differ), and would put it West of the aircraft at an elevation of about 72 degrees.

HazelNuts39 is offline  
Old 22nd Nov 2010, 12:07
  #2426 (permalink)  
Posts: n/a
Good morning. Your explanation with visual was excellent. Given HN's stated SAT position, I figure a roll either way gains the a/c access with the Satellite via its dorsal antennae placement.

What would not be conducive to signal maintenance would be (imo) a Pitch Down.

Roll limit in autoflight before involuntary uncouple is 45 degrees. The Pitch Down limit is 15 degrees (9 in PU). With a signal at ~70 to horizon and perhaps 8 o'clock (?), some combination of Roll right and nose down would cut the signal? The cycling in HN"s graph suggests exactly what, to you? In either extreme of attitude, whether only PD or Rolling with reversals, what would the related aerodynamic picture of this signal behaviour be in your opinion?

Old 23rd Nov 2010, 08:40
  #2427 (permalink)  
Join Date: Jun 2009
Location: I am where I am and that's all where I am.
Posts: 660
bearfoil, "Deponent sayeth naught."

(Meaning I am a EE not a pilot. You guys are the experts on how planes can misbehave in the air.)

I will note that we've not established the minimum "allowed/possible" time between ACARS messages. For that I'd need a copy of the Inmarsat specifications for that mode. I suspect the channel is very low data rate and there may be a hold-off interval. That hold-off interval MAY vary with "history". Until we know these details I'm not sure we can find most of the gaps HN39 found. The two really long ones need explanation. The others may be artifacts, especially if a message had to be resent for whatever reason. (I can almost see a potential for such congestion if we consider all planes that might be in AOR-W's footprint and the toilet configuration errors from AF-447 suggesting some small level of traffic is normal.)

This is one area BEA needs to explain a little more, IMAO. Can they explain the precise timing or is there something to those two long gaps. (And as I wrote this I had a "duh" moment. Yes, there IS a minimum interval between messages. During that interval data is coalesced into more compact messages than if they were sent one at a time. So an extra few seconds could very well be a transmission that needed to be sent a second time.
JD-EE is offline  
Old 23rd Nov 2010, 17:25
  #2428 (permalink)  
Join Date: Jun 2009
Location: NNW of Antipodes
Age: 77
Posts: 1,330
ACARS Sequencing

HN39 has raised an interesting point regarding the sequencing and 'dead' time slots in the time apparently available, i.e. up to the handshake received back on the aircraft following the last message.

As a short refresher, here are the ACARS messages as time-stamped to the nearest minute, along with their receipt times:-

An interactive web page with the above list is available from here. You can highlight one or more lines by clicking on them, and clicking again will return the line to normal.

Below is a revamped version of how the BEA described what was displayed in the cockpit, and what wasn't.

  • The ACARS time stamp has been given a different background color for each minute.
  • Warning messages take precedence over Fault Reports, and the 0210 position report had overall precedence in the ACARS sequence.
  • Sequencing within the FMS of the AOC messages is not precisely known.
  • The receipt timing (not provided by the BEA) of the 0200 position report in the ACARS system will provide a good indication of when the 0210 report was probably ready in the FMS to be inserted into the AOC sequence.
  • Highlighted cockpit effect messages are secondary faults, i.e. not causative.
  • The note associated with TCAS cockpit effect message has been inserted by me (from BEA Interim Report No.2).
The gaps in the message sequencing are for either of the following reasons, i.e.
  1. there was no AOC traffic queued in the FMS,
  2. data corruption required a retransmit,
  3. loss of data link.
Total time for each message was approximately 6 seconds, which comprised synchronization, message header, data and receipt confirmation.

Finally, for those that remember the old Telex/Teleprinter systems, the ACARS protocol was a product of that time, and the data transmission speeds are similar.

mm43 is offline  
Old 23rd Nov 2010, 20:31
  #2429 (permalink)  
Join Date: Jul 2009
Location: France - mostly
Age: 81
Posts: 1,689

For the record: When comparing time-stamp to time-received, I have assumed time stamp is minute without seconds, i.e. 2:10 covers everything between 2:10:00 and 2:10:59, rather than 'rounded to' nearest minute.

About 'gentle' and 'rapid' roll and pitch rates:

Roll axis: Regulation requires a minimum roll rate capability of 5 deg/sec at speeds above Mmo for recovery from lateral upsets. I believe the roll rates typically used when initiating routine heading changes are of similar order of magnitude. In the accident near Perpignan the A320 exhibited uncontrolled roll rates in stalls of 30 to 40 deg/sec.

Pitch axis: Typical pitch rate during takeoff rotation is 3 deg/s, in cruise probably somewhat less, say 2 - 2.5 deg/s. In the QF72 accident the airplane involuntarily pitched down at 8.5 deg/s, then pitched up in response to the pilot's control input at 5.6 deg/s.


Last edited by HazelNuts39; 23rd Nov 2010 at 20:49.
HazelNuts39 is offline  
Old 23rd Nov 2010, 21:28
  #2430 (permalink)  
Join Date: Jun 2009
Location: NNW of Antipodes
Age: 77
Posts: 1,330

I've taken everything at face value, and each time-stamped minute represents 30.1~60 secs and 00.1~30 secs, which is as you mentioned the "rounded" time.

As I mentioned in the earlier post, it would be interesting to know the actual timing of the 0200z AOC position report. It could conceivably help determine more closely when the 0210 time-stamped events possibly commenced.

Regarding the pitch and roll periods, the electronically phased array antenna is quite capable of tracking the satellite (provided it can see it) at far greater angular speeds than that recorded in the QF72 incident. Think of ships at sea; they often pitch and roll at remarkably high rates and even physically steered antennas manage to keep "locked on".

mm43 is offline  
Old 23rd Nov 2010, 22:18
  #2431 (permalink)  
Join Date: Jul 2009
Location: France - mostly
Age: 81
Posts: 1,689

I agree that whether or not the time-stamps are rounded is just one of many things we don't know. Note that the BEA doesn't speak of 'nearest minute':
la datation des messages par le CMC est précise à une minute près

HazelNuts39 is offline  
Old 23rd Nov 2010, 23:20
  #2432 (permalink)  
Join Date: Sep 2009
Location: Around the World
Age: 71
Posts: 88
mm43 wrote
Total time for each message was approximately 6 seconds
I hope you are speaking about 6 seconds between each message
(anothers planes have had to transmit own messages).

If not, It seems difficult for me to understand a so long period to transmit a few bytes, even if each "transmission packet" needs an ACKnowledge signal before transmiting the folowing.

Even if ACARS transmiting rules are comming from RTTY, I don't think a message transmission burst needing more than a quarter of second.
NeoFit is offline  
Old 23rd Nov 2010, 23:28
  #2433 (permalink)  
Join Date: Jun 2009
Location: NNW of Antipodes
Age: 77
Posts: 1,330

"... to the nearest minute", doesn't provide an exact meaning, and rightly or wrongly I construed that to mean "rounded". The alternative would be to say the events occurred within the minute in which they are time-stamped.

Lets face it, the plus or minus 30 secs difference represents less than 4NM, and we haven't located (so far) any bottom debris within a 40NM radius of the LKP!

rgds mm43

Last edited by mm43; 24th Nov 2010 at 01:30.
mm43 is offline  
Old 24th Nov 2010, 01:59
  #2434 (permalink)  
Join Date: Apr 2010
Location: Spice Islands
Age: 55
Posts: 97
First, mm43: Thanks so much for your great work (as usual)! Those depictions make it so much easier to get one's head around the messages and timings.

Speaking of locating the remains of the aircraft...

As it may relate to AF447, I have been trying to get a handle (as a lay person) on sonar systems, technology and capabilities. From my reading it seems the capability (and possibly the in situ equipment) exists to have recorded the noises of AF447's initial impact, descent of pieces, and even the impact with the ocean floor of some portions.

Apparently there may well have been stationary (permanently installed) sensors in various parts of the Atlantic, as well as onboard sensors of ships / subs at sea of course, that would have recorded the above mentioned noise.

We know the time (within a few minutes), so if there were even two or three widely separated monitoring / recording stations, then it follows that it should be relatively easy to arrive at a fairly precise location. This of course is predicated on the respective governments / agencies sharing data with BEA.


Sam Asama is offline  
Old 24th Nov 2010, 02:07
  #2435 (permalink)  
Join Date: Jun 2009
Location: NNW of Antipodes
Age: 77
Posts: 1,330

The timing sequence is carefully described by the BEA in its Interim Report #1.

Possibly a gap or two was caused by the system reverting to its prime task of trying to do an auto-connect with Dakar Oceanic using ADS-C. I'm sure that the Dakar log would have recorded another "failed connect" if that had occurred.

The BEA in its reports has made no mention of satellite traffic queuing, so that can probably be discounted.

You are right about the speed! RTTY, Morse code and then semaphore comes to mind!

mm43 is offline  
Old 24th Nov 2010, 22:12
  #2436 (permalink)  
Join Date: Mar 2010
Location: Sweden
Age: 83
Posts: 67
Mayday or not

There has been some arguments posted here about why the pilots of AF447 would not send any mayday when in beginning trouble.
On a.net you can find a description of an event on AF445 between Paris and Rio, where the pilot called mayday upon hitting severe turbulence.

AF445: Mayday Call Due To Severe Turbulence — Civil Aviation Forum | Airliners.net

The handling of CVR and other info by AF after the flight sound strange or sloppy to me. Furhermore, the disk problem with the QAR is also strange. Such things can often be solved (at least partly) by disk-specialists if one is willing to pay.
Nor do we know if any ACARS similar to AF447 were sent.
BEA is very silent about this incident and I had to dig around a lot to find a short comment saying that the incident was being investigated because it might help to shed some light on AF447.
Diversification is offline  
Old 25th Nov 2010, 07:40
  #2437 (permalink)  
Join Date: Jan 2008
Location: France
Age: 57
Posts: 9

AF 445 GIG-CDG hit some severe turbulence, the crew was unable to obtain descent clearance and applied oceanic procedures by transmissting MAYDAY to perform a level change.
Clearance to a new level was later granted by ATC.
There were no ACARS messages.

As to AF447 not sending a MAYDAY, just think how the events might have unfolded.
Pilots are in their seats in cruise, they may not be wearing their headset.
All the alarms go off and one is handflying in turbulence during the night with limited informations, as they both try to solve their problem.
To make a transmission, one has either to put on the headset and use it's mike or to use the handmike wich is hanging on its hook under the side window , that's if it's still there in the turbulence or upset.
pgroell is offline  
Old 25th Nov 2010, 15:20
  #2438 (permalink)  
Posts: n/a
Perhaps journo/drivel, but very early on, an "AF" person was quoted as "saying" that as well as lightning reports, they had received a comm from 447 that the Pilot had reported "Turbulences Forte".

This came in staccato fashion, as the Press was white hot, and reporting a blizzard of "Information".

Then again, it may have been "real", as the Public Face of AF was grasping for ways to immediately blame anything but themselves or the Bus. Gourgeon is in there somewhere.

Perhaps the time is right to introduce one of my new words. In honor of the French........ "Jourdrivelle"

Old 25th Nov 2010, 15:23
  #2439 (permalink)  
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,581
I seem to recall the report had been via ACARS?
BOAC is offline  
Old 25th Nov 2010, 18:09
  #2440 (permalink)  
Join Date: Jun 2009
Location: Brazil
Age: 72
Posts: 35
AF447 search to resume Feb 2011

O ministério francês dos Transportes anunciou nesta quinta-feira (25) que realizará uma nova etapa de busca, no fundo do mar, dos restos do voo da Air France AF447, que caiu no dia 1º de junho de 2009 no oceano Atlântico, com 228 pessoas a bordo, no percurso entre o Rio de Janeiro e Paris.
"A quarta fase da investigação no mar deverá começar em fevereiro de 2011, segundo as hipóteses apresentadas no dia 5 de outubro durante a última reunião do comitê de informação das famílias", informou o ministério do Transporte em comunicado.

Deficiências na manutenção do Airbus-A330 podem ter contribuído para o acidente. A informação consta de um relatório preliminar elaborado por peritos judiciais franceses, cujo teor parcial foi divulgado pelo jornal parisiense "Libération". Segundo o jornal, os sensores pitot, que ajudavam o piloto a saber a velocidade da aeronave, podem ter falhado porque não estavam sendo limpos com a devida frequência.

França lançará nova busca dos restos do voo AF447 Rio-Paris - 25/11/2010 - UOL Notícias

I am leaving it in Portuguese because it is the first time in following this incident that I have seen poor maintenance of the Pitot tubes as a factor rather than the design..
Gringobr is offline  

Thread Tools
Search this Thread

Contact Us - Archive - Advertising - Cookie Policy - Privacy Statement - Terms of Service - Do Not Sell My Personal Information -

Copyright © 2018 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.