The full story
This is the episode's narration, word for word. Headings jump to that point in the video.
The question 0:00
Every airliner carries two boxes that nobody hopes to open. Everyone calls them black boxes, but they're painted bright orange. And the idea for one key part, a recording of the voices in the cockpit, came from a fuel chemist in Melbourne in the 1950s. So how do you build a memory that survives a plane crash? And why did it take most of a decade before anyone wanted one?
Three jets fall apart 0:28
In the early 1950s, the de Havilland Comet was the world's first jet airliner. Then, in less than a year, three of them broke up in the air. Near Calcutta in May 1953, forty-three people died. Off the island of Elba in January 1954, thirty-five died, and near Naples that April, twenty-one more. There were no survivors to ask what had happened.
"Did you get my" 0:58
The Elba Comet was chatting by radio to another airliner when its last message stopped mid-sentence: "did you get my". The Calcutta inquiry admitted its work involved "a good deal of groping in the dark".
Solving it the hard way 1:12
The Royal Navy dredged up about seventy per cent of the Elba Comet from the seabed. At Farnborough, engineers sealed a whole Comet fuselage in a tank of water and pressurised it, flight after flight. After the equivalent of three thousand and sixty flights in all, counting its years in service, it split open at the corner of a window. The verdict was metal fatigue: the cabin had burst.
A chemist in the room 1:40
Back in 1953, after the Calcutta crash, Australian experts met at the Aeronautical Research Laboratories in Melbourne. One of them was David Warren, a fuel chemist, there in case a fuel tank had exploded. His mind wandered to the Minifon, a German pocket recorder that stored sound on steel wire, which he'd wanted for recording swing bands. Then someone asked whether a Comet could have been hijacked, and someone else asked, "How would we ever know?". Warren said that was when it clicked.
His father's crash 2:14
In October 1934, Warren's father, an Anglican minister, was on the airliner Miss Hobart when it vanished over Bass Strait. His last gift to his son was a crystal radio set, which started David on electronics. But the sources don't say grief drove the invention; in Warren's own telling, the spark was the Comet meeting.
The memo 2:39
In 1954, Warren wrote the idea up as a short technical memo. In almost any accident, he wrote, the pilot gets some warning, and usually says something about it to the co-pilot. Unless someone happens to be on the radio, there's usually no time to get a message out before the crash, so the answer is to record what's said.
A loop that forgets 3:02
Warren proposed a closed loop of steel wire running past an erase head and then a record head, about every two minutes. The record head is a coil on an iron ring that magnetises the passing wire in step with the sound, and the erase head scrambles the old magnetism just before. So the wire always holds the last two minutes, and in a crash it simply stops. And if nothing goes wrong, he wrote, everything said is wiped "during the time taken to taxi in". Recorders still work on that endless loop today.
No, thank you 3:39
For years, almost nobody wanted it. In 1955 a research council said that "in view of the practical difficulties," no action should be taken. In July 1958, asked to assess a working unit, the civil aviation department said it had "little immediate use in civil aircraft", and the air force joked that it "would yield more expletives than explanations". And in 1960 the pilots' federation warned it would object if the recorder were used as "a spy flying alongside the pilot". A few months later, it said it backed recorders in principle.
Fair to the sceptics 4:21
It's easy to make them villains, but Jeremy Sear, in a 2001 honours thesis, argues that's too simple. Australia hadn't had a major airline crash for years, and the same civil aviation letter offered flight trials once the lab developed it to a stage ready to install in an aircraft, though it wouldn't pay for that. Warren said later, "We didn't have the accidents in Australia… I don't blame them".
Built anyway 4:49
Before most of those letters were written, Warren had built one anyway, more or less unofficially. By his own account, his new boss, Tom Keeble, got a Minifon bought as if it were office supplies. A Melbourne instrument maker, Mirfield, turned Warren's sketches into a working recorder for six hundred pounds, using parts from two Minifons, and it was ready by the end of 1957. It could hold four hours of cockpit voices plus eight instrument readings, wiping the oldest as it went. It became known as the ARL Flight Memory.
Voice and numbers on one wire 5:28
But how do you put voices and dial readings on the same wire? Give them different pitches. The speech sat in one band of frequencies, and the instrument readings went in as bursts of tone in another, so a ground station could filter them apart afterwards. In the first version, each reading was the loudness of a beep, accurate to only about five per cent.
The lunchtime visitor 5:53
The break came by luck in June 1958, the month before those two letters. The lab's chief walked in at lunchtime with Robert Hardingham, secretary of the UK's Air Registration Board. As Warren remembered it, Hardingham said: "That's a damn good idea. Put that lad on the next Courier and we'll show them in London". In London, officials loved it. Who first called it a "black box" is disputed: Sear's thesis points to an air-accident official that August, Warren's family to a BBC journalist, and engineers already used the phrase for sealed electronics. And Britain began moving to make recorders compulsory.
The team that made it work 6:39
Warren was a chemist, not an electronics engineer, so in 1960 the lab gave him three who were: Lane Sear, Wally Boswell and Ken Fraser. Their problem: wire doesn't run past the head at a steady speed, so a recorded note comes back sharp or flat. Their fix was to send each reading as a number of cycles of a tone, rather than a pitch. A wobble changes the pitch, but it can't change the count. By the thesis's figures, accuracy improved to about one per cent, at twenty-four readings a second. The team's version first flew on a Fokker Friendship out of Essendon in March 1962.
Mackay, and a rule 7:24
In June 1960, a Trans Australia Airlines Fokker Friendship crashed into the sea off Mackay, killing all twenty-nine people. The inquiry couldn't find the cause, and it recommended flight recorders. In 1961 the government announced that Australian airliners would have to carry them, with cockpit voice, from 1963. By the Defence scientists' account, Australia became the first country to require cockpit voice recording. The safety bureau says, more cautiously, "one of the first".
Who gets the credit 8:02
So did Warren invent the flight recorder? Not exactly. America had tried to require simple data recorders as early as 1941, and did so from 1958. An American engineer, James Ryan, had filed a patent for a crash-resistant data recorder in August 1953, before Warren's memo. What was new was Warren's idea of recording the cockpit voices, together with the instruments, on a loop that forgets. Jeremy Sear's thesis concludes that Warren's 1954 memo was the first proposal for a cockpit voice recorder, and none of the sources we checked describes an earlier one. Australia couldn't patent it, because the idea had already been published. A British firm, Davall, built the design into its "Red Egg", while Australian airlines bought American recorders.
What's on the voice recorder 9:00
On a cockpit voice recorder, microphones at each pilot's seat, plus an area microphone, each feed a separate channel. So it catches more than talk: alarms, switch clicks, engine and wind noise. A modern one keeps the last two hours, with the newest sound overwriting the oldest, just as Warren's loop did.
What's on the data recorder 9:21
The flight data recorder keeps numbers instead of sound. Since the 1960s the bare minimum has been five: altitude, airspeed, heading, acceleration, and when the radio mic was keyed, which lets investigators line the data up with the voices. A modern Airbus A380 records more than a thousand, and keeps the last twenty-five hours.
Why orange 9:47
So why orange? Because it has to be found in a field of wreckage. The rules say bright orange or bright yellow, with reflective tape. It's mounted near the tail, which usually survives best. And if it lands in the sea, a beacon starts pinging at thirty-seven and a half kilohertz, too high to hear, for at least thirty days.
The memory in the middle 10:11
The box itself is allowed to be wrecked; only the memory inside, in a case built to resist impact and fire, has to survive. The test is brutal: a jolt equal to hitting a wall at five hundred kilometres an hour and stopping in forty-five centimetres. And an hour in a fire at 1,100 degrees, hot enough to wipe an unprotected magnetic recording, because iron loses its magnetism above about 770 degrees. Modern recorders use memory chips, where each bit is kept as a tiny charge sealed inside a glass-like layer of silicon dioxide, which holds for years without power.
Do not open 10:55
David Warren died in 2010. At his daughter's request, he was buried in a casket labelled "Flight Recorder Inventor: Do Not Open". Next time, an Australian device that listens for a person. When the tiny hair cells in someone's inner ear die, can you send sound straight into the nerve of hearing?
Sources
Every factual claim in the episode is tied to one of these. Spotted an error? Tell us.
- Australian Transport Safety Bureau, "Black box flight recorders" (2014) — atsb.gov.au
- US 14 CFR § 25.1457, "Cockpit voice recorders" (Cornell LII) — law.cornell.edu
- J. Sear, The ARL 'Black Box' Flight Recorder – Invention and Memory, BA (Hons)… — web.archive.org
- K. Fraser and D. Warren, "The Black Box: An Australian Contribution to Air Safety",… — web.archive.org
- DSTO, "Dave Warren – Biography" (last modified 6 June 2006) — web.archive.org
- Lord Cohen (Commissioner), Civil Aircraft Accident: Report of the Court of Inquiry… — web.archive.org
- N. S. Lokur, Report of the Court Investigation on the Accident to Comet G-ALYV on 2nd… — web.archive.org
- Wikipedia, "Flight recorder" and "South African Airways Flight 201" (secondary; used… — en.wikipedia.org
- R. Seales, "This little-known inventor has probably saved your life", BBC News, 18… — bbc.com
- D. Crotty and M. Churchward, "ARL Flight Memory Recorder", Museums Victoria… — collections.museumsvictoria.com.au
- K. Cole, "Warren, Hubert Ernest de Mey (1885–1934)", Australian Dictionary of… — adb.anu.edu.au
- Wikipedia, "de Havilland Express" (DH.86; Miss Hobart lost 19 October 1934; secondary) — en.wikipedia.org
- DSTO, "Dave Warren – Inventor of the black box flight recorder" (last modified 29… — web.archive.org
- HyperPhysics (R. Nave, Georgia State Univ.), "Tape recording" / "Tape heads" (erase,… — hyperphysics.phy-astr.gsu.edu
- ATSB, "Black box flight recorders" fact sheet (PDF) — atsb.gov.au
- B. R. Allen and J. S. Leak, The Potential Role of Flight Recorders in Aircraft… — commons.wikimedia.org
- J. J. Ryan, "Flight Recorder", US Patent 2,959,459 (filed 4 Aug 1953, granted 8 Nov 1960) — patents.google.com
- N. Campbell, "The Evolution of Flight Data Analysis", Australian Society of Air Safety… — web.archive.org
- NTSB Vehicle Recorder Division, Cockpit Voice Recorder Handbook for Aviation Accident… — web.archive.org
- US 14 CFR § 25.1459, "Flight data recorders" (Cornell LII) — law.cornell.edu
- HyperPhysics, "Ferromagnetism" (Curie temperature of iron ≈1043 K) — hyperphysics.phy-astr.gsu.edu
- Halbleiter.org, "The Flash Memory Cell" — halbleiter.org
Image credits
- ARL Flight Memory recorder, c. 1958 · photo: Rodney Start / Museums Victoria · CC BY 4.0 · licence: https://creativecommons.org/licenses/by/4.0/
- Cockpit voice recorder · photo: Meggar, CC BY-SA 3.0, via Wikimedia Commons · licence: https://creativecommons.org/licenses/by-sa/3.0/
Researched and scripted with AI assistance, fact-checked claim by claim, with synthetic narration and diagrams drawn in code. How we make episodes.
