Why are black boxes made the way they are?

June 4, 2009 6:51 AM

The recent tragic loss of the Air France flight and the talk of how hard it may be to locate the "black boxes" raised some questions on this technology in an office discussion and we wondered why the solutions aren't somewhat simpler

If an aircraft is lost in the ocean, why aren't the flight data and voice recorders designed to just float to the surface and give off an easily locatable signal so the search and rescue crews can just pick them up? One of our group thought that they have the same technology for small ocean-racing sailboats where their emergency beacons will float and give off the radio signal, so why not the same for these?

Also, if the aircraft (a modern Airbus) was capable of sending automated text messages back to the engineering department by satellite to inform them of problems, why can't the data and voice recorders continuously send a stream of information back to the engineering people by the same method and just have it deleted after the flight arrives safely at its destination? Then there wouldn't be the constant searching for the "black box" every time there is a disastrous crash of this type.

It seems that this technology is decades behind the design of these 'fly-by-wire' aircraft.

posted by 543DoublePlay to technology (12 comments total)

This Wall Street Journal article addresses these issues.
posted by veggieboy at 6:53 AM on June 4

Well, in the first situation, consider the difficulties in making a robust system that continues to function after 400 tons of aircraft lands on it.
posted by electroboy at 7:19 AM on June 4

Regarding the second part of your question, that's theoretically possible in near real time using Iridium data services. It's about 9600 baud but data compression can speed things up a bit. The problem is that the data costs would add up, it's charged per kilobyte. You could send basic telemetry on what the aircraft is doing but the speed of LEO satellite networks is nowhere near fast enough to transmit a constant stream of cockpit voice recorder and full onboard computer / fly-by-wire diagnostic information.
posted by thewalrus at 7:32 AM on June 4

Also, if the aircraft (a modern Airbus) was capable of sending automated text messages back to the engineering department by satellite to inform them of problems, why can't the data and voice recorders continuously send a stream of information back to the engineering people by the same method and just have it deleted after the flight arrives safely at its destination?

This turns out not to be all that simple to do. Modern cockpit voice recorders and flight data recorders handle many channels of audio and data, respectively, so the bandwidth requirements to transmit all this information continuously are not trivial, especially multiplied by the thousands of airliners in the air at once. Compare this to very occasional "maintenance required" kind of messages that are sent now.

I also suspect that communication technologies that work fine in normal flight (like satellite uplinks with little steerable dishes) are going to fail at the exact time they are most needed -- when the aircraft is maneuvering violently or in an unusual attitude. Consider for a moment how any satellite antenna located on top of the fuselage will be able to communicate with the satellite when the plane is upside-down.

I think the safety authorities would be happy if we could just get modern CVR/FDR equipment on board every airliner. This seems to only be required of newly-manufactured planes, so old planes still have ancient technology of limited usefulness.
posted by FishBike at 7:37 AM on June 4

If an aircraft is lost in the ocean, why aren't the flight data and voice recorders designed to just float to the surface and give off an easily locatable signal so the search and rescue crews can just pick them up?

Mainly because they'd need to be on the outside of the plane and easily detachable to do this - sea vessels can have these because they stick the on the top of the mast and out of the way.

If the black box was on the outside of the plane it would be relatively unprotected for the majority of crashes where the plane impacts something. So it's potentially a game of percentages - better to have most of the information from land based crashes than from the unusual sea based ones.

The impetus to have all the information from every single plane crash isn't necessarily high enough to justify the extra cost and complexity (plus weight) to ensure black box retrieval in every single incident, as well (no matter what they say in the papers at present). While it is all useful information in terms of crash investigation, there will always be 'freak accidents' and this may be just within that percentage. The main difficulty with recovering the black box seems to be mostly how far it is from everything - ie from dry land and from teh surface of the sea. Even if it was possible to physically get to the black box, then it may take longer than the 30 days the batteries will last for. In just the same way, finding an airliner and it's black box in that time frame in a remote mountainous region may also be considered one of those edge case accidents that means the true cause remains unknown - it's balanced between the value of the information versus the cost and complexity of raising the percentage of times it will be available to them.
posted by Brockles at 7:50 AM on June 4

There was a bill introduced in Congress in 2005 which would require a second combination voice/data recorder on all planes, to be ejected from the rear of the plane among impact (so as to avoid the wreckage/forces of impact itself as much as possible) and to "float indefinitely." Looks like the bill never made it out of the Subcommittee on Aviation.

(Even if it had passed, since the Air France flight neither originated in, was destined for, or passed over the US, nor is it a US based airline, such a law wouldn't apply to it directly anyway, although it might have had an indirect influence as far as affecting worldwide aircraft standards.)

Also, here's a good LinkedIn thread discussing some of the same issues. Eric M notes there, "Personally, I don't think you could justify the expense and risk associated with retrofitting the existing fleet. Oceanic crashes are relatively rare, and crashes where the CVR/FDRs have been unrecoverable are even rarer still. "
posted by DevilsAdvocate at 7:57 AM on June 4

Imagining features is a fun hobby, and it's where invention sometimes starts. Most people never progress beyond that.

For those that do, an early lesson is often that features are seldom free. There are technical and more often, economic hurdles to overcome to realize a design.

Generally, designs are optimized. This means that there is a tradeoff between a given feature set and the circumstances of the environment in which it will be used.

What you've come up with is a solution to ONE of the circumstances in which a plane may find itself. The number of times such a feature is likely to be used is fairly small, it is safe to say. The additional complexity, the relatively scarce need for the feature, and the cost of developing/testing/verifying/deploying/maintaining the feature set you are imagining would be very unlikely to favor its creation.

Technology and its availability has very little to do with implementing what you and your office mates have imagined. That is usually the case. Economics is the elephant in the room in any product design, as it dictates how much something is worth. (That does not mean that super duper expensive things are impossible or never pursued... they often are. The cost/benefit mix has to say it's worth it to do so.)
posted by FauxScot at 8:32 AM on June 4 [1 favorite]

The current FAA rule on CVR/DFDRs (the first D stands for "digital" to give you an idea how old the electronics standards are) was passed in April of last year and was substantially watered down from the original proposal, principally on the grounds of costs (scroll down to section "T. Comments on Cost/Benefit Analysis" in the first link for an idea of the numbers). This is the state of the art and it is costing significant amounts of money to retrofit most large airplanes (generally those with 20+ passengers, but the operations rules are complicated) to this standard by 2012. Part of the problem is that the older the aircraft design is, the more antiquated the existing DFDR system, and the more expensive the fix is.
ICAO is said to be working on further improved standards for release in a few years (that will then need to be adopted by major government certification agencies before they are mandatory).

All that is to say that these things take time, because they take a lot of effort, because the intent is that they get designed and built with a minimum of error, because the term "crash" has very different meanings to the aerospace and commercial electronics industries. Streaming data from aircraft to satellites in the event of a crash would take a lot of power and money for a very rare information payoff. Money certainly pays a role here, because budgets are always finite, and the $1 million you spend on a better DFDR might save more lives if you spend it on something else in the airplane. The black boxes really just record what went wrong to let people build better airplanes in the future; they don't make the aircraft they are on any safer.

As for why the boxes don't float, they are bolted into the aircraft in such a way to withstand substantial impact loads (not those of a crash, but a very hard landing where you don't want things breaking free and hitting passengers). In the event of a catastrophic crash, they are generally inside the aircraft and potentially still attached to something that won't float. "Why not make them deploy from the aircraft before/as it crashes?" you ask? The worry is that a false alarm might accidentally deploy a black box onto your house, and that false deployments would happen far more often than actual crashes. Well, that and it would cost a lot.
posted by cardboard at 8:41 AM on June 4 [1 favorite]

One interesting objection I saw somewhere is that the pilots are not particularly interested in having all their conversation transmitted and recorded.
posted by smackfu at 9:10 AM on June 4

I think a floating black box is probably less desirable than a sinking black box in many cases. If it floats, currents can carry it far away from the crash site very rapidly. That's a problem. If it sinks, you'll find it more or less wherever the plane went down. That's only a problem when a) you don't know exactly where the plane went down and/or b) the plane went down in water 15,000 feet deep.
posted by chalkbored at 9:45 AM on June 4 [1 favorite]

I work for a defence contractor that installs crash data recorder systems on defence force aircraft.

There are definately deployable CDRs. We installed a deployable version on a helicopter. Upon certain conditions being met, for example the helicopter became submerged, or glass capsules located at strategic points near the airframe skin were broken, a small charge would ignite, and the deployable part sprang off the side of the aircraft. Within the deployable part was a hard drive recording the last 20 hours of flight data (positions of the controls, GPS position, engine temperatures etc.) and the last 10 hours of cockpit voice data. On deployment, the unit floats, and transmits the last obtained GPS position (i.e. where the aircraft probably crashed) over a high-power satellite-based communications system. The unit can be located post-deployment using the strength of the signal so the data can be recovered.

Deployable systems tend to be larger, and their installation is more complicated, and so are not typically used on high-speed aircraft. Deployable units need to provide their own crash-protection, and therefore are larger and heavier. This is acheiveable for low speed helicopters, but not as easy for high speed aircraft.

Yes the standard of this equipment is way behind "cutting edge" - it just takes way too long to get regulations and standards updated.
posted by trialex at 10:38 PM on June 4

First of all, the current black box is a reasonably proven technology. Engineers, especially in aviation, like proven technology. They like it a lot.

Second, it's a technology that works almost all the time that it is needed. I'm not sure where one would look to find the non-retrieval statistics (or if you're being even more conservative, the non-retrieval-of-information statistics), but they may not number more than you have fingers.

Now, that all said, this may be the incident that gets some kind of floating black box requirement in place for ocean-crossing planes, but even if that passes in the US, it will be many years before it is an internationally recognized standard, and it may be decades before all planes are retrofitted, whereas some old planes may never get them.
posted by dhartung at 10:38 PM on June 4