Quick Access Recorders

[tubby linton]

Would anybody be able to help me with some questions on QARs as my aircraft tech manual doesn't even mention them!
1 Where do they obtain their electrical power from,and do they keep functioning when there is limited electrical generation?
2 Where do they obtain the data to be recorded from and what data is recorded?
3 Is there any way spurious data could be recorded,e.g data from a faulty adc?
4 If the data is used for FDM and possibly disciplinary action would there be any way to prove the recorded data was incorrect?
My old Airbus sometimes gives conflicting analogue and digital readouts of information so perhaps the data being sampled may not be accurate.

[vapilot2004]

QARs get power from the AC bus. During a partial power loss, the QARs are not always load shed. Typically they will be un-powered when operating on standby power.

On aircraft that have non-integrated airframe computer systems (such as the A300 series, 737 etc) there is a data acquisition box (DAU) that converts selected analog inputs into digital signals and combines this data with any ARINC data bus signals for QAR and DFDR consumption.

Spurious data is very common in the outgoing data stream of most DAUs due to limited signal conditioning and lack of 'smart' A/D conversion. These acquisition boxes also use a relatively slow sample rate. Sampled analog inputs can be inaccurate to the point of requiring integration math (terminology?) to 'mold' the data to fit a known profile upon readout - not something I would wish to be judged by.

The majority of QAR (and DFDR) recorded data is of course reliable, but where there is an imperfect system used to collect and convert analog signals, there can be a 'perfectly recorded' stream of inaccurate data.

[LurkerBelow]

How is the QAR data actually used by airlines?

[tubby linton]

Vapilot thank you for the info.I thought that it was just a recording device and therefore could potentially record rubbish.Flight data monitoring is being used in some companies to discipline crews for poor operation.My reservation is that a pilot could be falsely accused based on qar data.It would be hard to prove that the data is false.

[vapilot2004]

More information on your particular aircraft's QAR data collection setup is available from your friendly ground engineers shop. A look at the supplemental type certificate for the DAU installation may provide some limited info.

There are more than a few pilot unions here in the states that have vociferously fought against performance reviews and sanctioning based solely on these digital witnesses' testimony. The Allied Pilots Association (AA) are one of the more motivated groups in this area.



My reservation is that a pilot could be falsely accused based on qar data.It would be hard to prove that the data is false.

I agree with you completely on this Master Linton.


I know of a fellow 737 pilot who faced sanctions after the aircraft he flew out of Denver later developed a 'premature' hot section failure. Airline had QAR records clearly showing an over-boost with EGT over-limit on his watch. Our man recalled things a bit differently and despite the F/O's recollection of no N1 nor EGT telltale on the day in question, company pressed on. The DFDR data had been long over-written and may not have helped anyway depending on where the discrepancy existed in the recorded data stream.

Thanks to a mechs saved FADEC readout there was finally more evidence that yes - no over-boost happened, at least on that particular day. I would have bought that gent a case of the best were it me. Engineering was rumoured to have later discovered that the DAU retrofit was feeding bad data to the QAR due to a minor mis-calibration. Imagine that! :ugh:

[woodyspooney]

It always amazes him why techies swear that e-data form QARs and AIMS/ACMS are the gospel truth. Years ago, a colleague had an inadvertant overspeed during descent from a warm airmass into a colder one coupled with a massive increase in headwind. It was a flight from downunder to SEL during winter where there was the ever present headwind. He reported the incident in the techlog and ASR, but the techies came out with a QAR/AIMS digital report showing tailwind and almost constant OAT!!! The safety dept declared him lax in monitoring his descent profile and accused him of deliberately increasing descent speed to get back on descent profile as he was " boosted " up above his descent profile by atmospheric factors!!! He challenged the dept. with met reports, forecast winds etc, but the techies stood their ground. He was reprimanded with no further investigation.

Later on another colleague experienced a fuel leak and informed maintenance and dispatch that he was diverting. The company tried dissuading him from diverting AS THEY CLAIM REAL TIME DATA LINK REPORTS SHOW NO ANOMALY; thankfully he stood his ground and countermanded company instructions, diverting to ANC. Flight Ops and dispatch tried to flame him for that even though a fuel leak was actually found; finally I believe some belated advice from Boeing persuaded the company to back down...the fuel leak checklist in the QRH was changed!

[Dani]

hmmm, me it seems that you guys don't really have a solid information on how flight data monitoring happens in a modern airline. The data in question doesn't go on the desk of chief pilot, maintenance or even dispatch :eek:

There is a safety pilot office, where there is a team of qualified pilots who run those programs over the data. They only report to the management with names crossed out. Their target is not to blame someone, but to find tendancies in the fleets and thus increase safety. If one guy still diverts from all the others, then of course the name will be forwarded. But this happens in a transparent, written procedure, and only after several talks with the guys concerned.

I've been in several airlines where they introduced flight data monitoring, and lots of guys argued like you do, but in the end it all proved to be unfounded. If you treat the data this way. Of course I don't know how it is in your airline. Maybe not only the line pilots don't know how, but also yours managements...

Dani

[oztrekker]

Dani, I believe Woody was referring to an Asian carrier in north east Asia where until recently, despite company management's denials, QAR and AIMS data was available to chief pilots, safety depts, maintenance people and even ordinary national pilots who attended their so called " safety meetings ". When I was there my national copilots used to tell me in great details about reports of deviations from normal flights, data of which were obtained from the FOQA dept AIMS / QAR extracts. I was flabbergasted and I noted my misgivings to the VPs and managing VPs ( even expats ). I was treated like a nosy, noisy fool by both national and expat management goons who were only too happy to use such data to lambast line pilots' deviations which prove that they themselves ( management pilots ) were superior as they fly only to SOPs perfectly.

It's only recently that I heard that they established the necessary protocols w.r.t. to FOQA/AIMS/QAR data to prevent wanton disemmination of such information to all and sundry. I certainly hope they are really serious and stick strictly to the established protocols so that the whole FOQA programme is held to a high standard reflecting the spirit in which it is based on.

[Dani]

That's what I mean - it's the line pilots and management who do not know what flight data monitoring means.

This information in the right hand is an enormous increase in safety.

Happy flying,
Dani

[tubby linton]

Does anybody know what the sampling rate is?

[vapilot2004]

Depends on the input. Some continuously variable inputs can be as low as one sample per second.


This kind of low resolution is not an equipment limitation, it is per approved and certificated design. Regulatory bodies (FAA, JAA) pressed to widen data collection parameters aboard commercial airliners back in the 1990s. Certain manufacturers resisted citing cost,complexity and data reliability issues with these retrofit systems. An imperfect compromise is what we have today.


hmmm, me it seems that you guys don't really have a solid information on how flight data monitoring happens in a modern airline. The data in question doesn't go on the desk of chief pilot, maintenance or even dispatch


One hopes that the majority will always follow this excellent example of the ideal. The reality, however, can be different. Once data is gathered and recorded, the home office owns it and can mine at will the entire collection for whatever suits their purpose du jour.


This information in the right hand is an enormous increase in safety.

Happy flying,
Dani

Of this there can be no doubt. :ok:

[Brizeguy]

Tubby,

Have a look at www.avionica.com for more info.
If this is the type on your aircraft the tech rep. is very helpful.

BG

[73AIRBUS73]

How long is data recorded for does it only store what is on the black box ie up to 28 hrs eg or is it seperate?

[alright jack]

Latest QAR's have a data "flash" card .The QAR signals the central maintenance computer when the card is full . Usually about 6-7 days worth on the one's we change .

[Daysleeper]

Latest QARs..

The even later ones will either do a Wifi downlink every time they hit home base or for the more nomadic aircraft will use a mobile phone to phone the data in following arrival.

The number of flights recorded therefore is unlimited as the data is stored on a company server.

Does anybody know what the sampling rate is?

How long is a bit of string, usually base four so some things ... (very rough examples) will be four times a second (AoA), some 16 times a second (momentary action switch) some once every 128 seconds. (CofG)

would there be any way to prove the recorded data was incorrect?

In general to prove the validity of the data is a time consuming task.... depending on what precise parameter you wish to verify. At its simplest you can go an function test it, then see if the recorded data matches with what you just did.

does it only store what is on the black box

Often a slightly different (usually expanded) set of parameters are recorded with near a thousand not unusual on modern aircraft.