In the mid-1950’s, I was an in-house consultant and technical supervisor of the Hoffman Laboratories. In that capacity, I had designed the
Sonobuoy that we manufactured for the U. S. Navy.
What is a Sonobuoy? A Sonobuoy is a cylindrical device, which was deployed (dropped) from a low-flying airplane. The buoy portion, which contained
a two-way radio and antenna, would float vertically on the surface of the ocean. Depending upon the model, either mounted directly on the bottom of
the buoy or hung at the end of a cable, there was a hydrophone (the sono portion, from the word “sonic”), a piezo-electric transducer
which either passively or actively would listen acoustically for submarines. By placing several (at least three, usually five or six) Sonobuoys in
a specific pattern, one could triangulate the position of the submarine.
The two-way radio communication between the airplanes that deployed and then controlled and listened to the Sonobuoys employed “frequency
hopping”. This frequency-hopping was implemented in analogue hardware. As such, it was too slow and had too few frequencies to provide much
noise-immunity (reliability). It provided only slight immunity to interference among themselves. However, even this crude implementation did
provide the vital security (prevention of unauthorized reception of the radio signals, by the enemy military). This, probably the first,
utilization of frequency-hopping predates, by at least a decade, any use referenced in the open literature.
As the person in technical supervision of the design and manufacturing of these Sonobuoys, I wish to express my belated respects to the Inventor of
the frequency-hopping concept.
In the early 1960’s, at Aerojet-General, I was the System Manager over all of the contractors designing and building the surveillance
drone which eventually flew over Vietnam. (Subsequently, the drone was purchased and manufactured by the Ryan Aircraft Corporation, in San Diego.)
Among other things, I personally designed the reliable and secure two-way radio communications system. It was implemented digitally. It employed a
hierarchical digitally-simulated spread-spectrum modulation, which then was high-level double-sideband suppressed-carrier quadrature-modulated upon
a coherent (with the clock of the digital computer) carrier. A double-sideband suppressed-carrier quadrature-modulation is phase-modulation, which,
except for a difference in pre-emphasis, is frequency-modulation. For the first time, we had the ability to switch frequencies rapidly; thus, we
called it “spread-spectrum”. It is the same concept as “frequency-hopping”, only performed much faster.
The coherency leads to destructive interference effects in a subsequent square-law demodulation, as would be employed in an attempt to intercept
the transmission by an adversary. Instead, a locked phase-loop demodulator, which is synchronized with the known frequency-hopping pattern employed
in the transmission, has to be employed to demodulate the signal.
The radio signal required very little transmitter power and was immune to noise and interference from other drones, which would employ the exact
same carrier frequency. The radio signal was secure – it was completely undetectable by the square-law demodulator employed by the analogue
spectrum analyzers, then available to the Counter-Intelligence community, both ours and that of our potential adversaries.
This spread-spectrum two-way radio communication system was crucial to the success of the drone.
As the System Supervisor during the design phase of the drone, I wish to express my belated respects to the Inventor of the spread-spectrum
concept.
Who invented the spread-spectrum (formerly called “frequency-hopping”) concept?
None other than Hedy Lamarr, made famous by her starring role in Ecstasy, a very x-rated movie, filmed in
Czechoslovakia in 1933.
Hedy Lamarr invented in 1940 and patented in 1941, under her married name of “H. K. Markey”, US Patent number 2292387, granted in Aug.
1, 1942, “Secret Communication System”, filed June 10, 1941. She signed it “Hedy Kiesler Markey.” Kiesler is her maiden
name. Gene Markey was her second Husband’s name. Lamarr is her stage name. Her full name is Hedwig Eva Maria Kiesler. She
unwittingly disguised herself by not employing her stage name. It is only this year (1997) that the connection has been pointed out by Dave Hughes.
At present, Hedy Lamarr lives in retirement in Florida.
Each of the two aforementioned applications of the spread-spectrum communications concept was shrouded in military secrecy, at the time. We just
knew the concept; but the identity of its originator was not disclosed to us, under the restriction of “need to know”. The US
Government was the assignee of the Patent; thus, legally, it was the owner of the concept. It had no reason to disclose to us, and we had no
“need to know”, the originator of the Patent.
Currently, the spread-spectrum radio transmission is employed extensively in both civilian and military communications: Radio links with many
satellites, portable telephones which operate in the 900 MHz region, wireless network connections, high-density cellular-telephones, ….
Belatedly, Hedy Lamarr deserves credit for her stroke of pure genius in inventing the spread-spectrum concept in one evening, out of thin air! She
deserves to be admitted to Mensa, as an honorary member.
There are numerous references on the Internet: the Patent, her nomination by Dave Hughes, in 1997, to the EFF Pioneer Award (which she subsequently has received), the Ecstasy film may be obtained from several places Foreign classics of the '30s. In the eight years since this page was written, link rot has set it. Today, I am deleting the broken links. Employ your favorite search engine to find your own links.
In the print media, her autobiography Ecstasy and Me, published in 1966, is out of print. I have been unable to locate a copy as yet. Two recent articles about her invention are: "Advanced Weaponry ..." American Heritage of Invention & Technology (ISSN: 8756-7296), Spring 1997. "Go Wireless Today, Not Someday...." PC Magazine, 22-nd April 1997. A copy of this page probably has been published in Lament, the Greater Los Angels Area Mensa Commentary (ISSN 1043-6294 USPS 303-680), in August 1997.
Another important reference -- recently brought to my attention by its author -- is an article (including a photograph of Hedy Lamarr), "Spread Spectrum", by Dave Hughes, in the April 1998 issue of Scientific American. He also informed me that on Friday 16-th October 1998, Hedy Lamarr received the Victor Kaplan medal of the Austrian Academy of Sciences, at Eisenstadt (capital of the Austrian province Burgenland).
Hedy Lamarr died on Wednesday 19-th January 2000, in Florida, USA.
Now that I know who invented the spread-spectrum concept, once again, I, who was the designer of- and probably the only person who remembers those early applications of- the spread-spectrum concept, want to express my sincere admiration and belated thanks to Hedy Lamarr.
These technical details have been added, on Wednesday 18-th June 2003, at the request of Mr. Dave Mock, who provides extensive tutorials on wireless technology.
The Hoffman Radio Corporation was a major manufacturer of radio communications equipment: TV, AM/FM radio, automobile radio, short-wave radio receives for the consumer market; various esoteric radio transmitters and receivers for the military; the Tacan navigation system for military and commercial aircraft; we designed all and manufactured most of the spectrum-analyzers for the military counter-intelligence.
A military contract often includes a stipulation as to a technology that is to be employed. Such a stipulation may serve either of two purposes:
Because of no "need to know", the contractor would not be informed of which motivation is applicable.
When we received the contract to develop the Sonobuoy, we were provided with a copy of the H. Kiesler Markey patent. Since it was dated a decade previously, we assumed that it was an existing secret technology, devised by some clever electrical engineer, working under a Navy contract and thus obligated to assign the patent to the Navy.
As requested, we designed the radio communications following the concept of the patent. It worked and worked very well, from the beginning. Thus, we went on to design the difficult: the hydrophone, the batteries, and the pattern for the optimum deployment of the Sonobuoys. That was half of a century ago. Since the radio communications design was so easy, I do not remember any details. I do remember that we worked hard to develop a satisfactory hydrophone. The conclusion as to the pattern was that a straight-line is undesirable, regardless of how many Sonobuoys are deployed. While I do not remember the specific design of the spread-spectrum communications radio, I do remember the techniques that we were employing at that time.
I can reconstruct the design, now. We had been employing various cams to control the tuning. These cams were designed employing digital computers. To perform the frequency-hopping, we would have employed a cylinder (spool) with protrusions. Each frequency would have been assigned an individual follower riding in a row parallel to the axis of the cylinder. There would have been perhaps a dozen frequencies. How fast would the cylinder rotate? 3600 rpm (= revolutions per minute) would be unrealistically fast, for a complicated mechanical device. 900 rpm still would be unlikely. 90 rpm sounds about reasonable. Anything slower would compromise security. The Sonobuoy and the listening aircraft would employ the same spool. When the aircraft wanted to listen to a different Sonobuoy, it would have to change spools. The carrier-frequency would have been in a band that the military owned at that time in what now is channel three in the low VHF television.
In retrospect, now, I realize that the Navy asked us at Hoffman (and me in particular) to design a frequency-hopping radio system; because, they considered that if anybody could, we would. Thus, rather than existing technology, this was intended to investigate a new (but, neglected) concept.
To compute the hopping rate, we need only one additional piece of information -- the multiplicity m of each of the frequencies on the spool. Two sounds about reasonable. Let us define a few variables:
Then, we have:
Could a hopping rate of 36 per second have been achieved with the technology that we had available at that time? Let us compare to the pulse rate of the ignition on an automobile engine -- Otto, with a Kettering ignition. The ignition-points of an eight-cylinder engine begins to float at about 4800 rpm. It is stable at, say, 3600 rpm = 60 rps. With 8 cylinders, that is 480 ignition-pulses per second. Hence, if we had employed leaf-springs, a hopping rate of 36 per second would have been easy. However, our attitude had been that an open leaf-spring is unreliable, especially in a marine environment. We would have wanted to employ our favorite -- a micro-switch -- for each frequency. With the greater moving mass of a micro-switch, it would begin floating at several hops a second. Thus, a hopping rate of 36 per second would not work with micro-switches. We then would have realized that a Sonobuoy has a very short life, anyhow. It employed a pre-charged battery that was stored without any electrolyte. The shelf-life would have been very good. Once the Sonobuoy had been deployed into the ocean, the salt-water would flood into the battery, providing the electrolyte. Then, the battery would have to operate both the hydrophone and the radio-transmitter. It would last for at most a day or two. When the battery became exhausted, the Sonobuoy would be abandoned. For such a short life-time, the leaf-springs would not corrode. We would have settled upon leaf-springs, instead of micro-switches. Then, the aforementioned hopping-rate is reasonable.
While the Sonobuoy worked very well as a listening device and for either active or passive ranging of submarines, it was not practical as a system.
With effort, over time, one would expect improvements. However, a better system-design emerged. Permanently anchor the hydrophone to float submerged. Run a cable to the shore. Solves each of the foregoing problems. Regrettably, without the necessity of radio communication, there was no more spread-spectrum involved. To this day, the Navy has hydrophones deployed along the whole coast of the USA.
A swept-frequency spectrum analyzer sweeps the local-oscillator in a saw-tooth. A fixed signal becomes a frequency-modulated IF signal. At the time, we had derived the applicable equations. Qualitatively, the conflicting requirements are:
Thus, the performance of a swept-frequency spectrum analyzer depends critically upon its design and the skill of the operator to tune it. It is very tricky. Now -- but, not in then --, one would have a microprocessor to aid in the process of tuning.
Since I was involved in the design of all of the spectrum-analyzers for the US counter-intelligence, I was well aware of their inherent weaknesses. Thus, when it came to design the secure communications system for the drone, I designed to these weaknesses.
I was the System Manager for the drone aircraft.
The transmitter on the drone aircraft best may be visualized as a digital computer with an antenna.
We had been allocated a band of (probably) six megahertz. It had been expected that we would assign a narrow slot to each drone. Doing so, however, would lead to two problems:
We were requested to provide secure communication; but, it was left up to us to come up with a system.
My solution was to employ spread-spectrum. The out-going signal from the drone consisted of:
All this was multiplexed hierarchically by the digital computer into a bi-polar pulse at a rate of some three megahertz -- exactly half of the available band-width. If memory serves; the frame-rate was 30 Hertz. Today, with the availability of microprocessors, one would employ a much longer period for the pseudo-random multiplexing.
For security and to prevent cross-interference among the drones, we had a 63-bit shift-register. It had a toggle-switch for each bit. Just prior to launch, a pattern would be set manually into the drone and automatically copied into its base station. That gives us 2 to the 63-rd power -- approximately 8 times 10 to the 18-th power -- of choices. It is most unlikely that any drones would ever employ the same pattern. Thus, we prevent interference among the drones and make their signal secure.
Then, this signal was double-side-band, suppressed-carrier high-level quadrature modulated upon a synchronized carrier frequency. Usually, when one goes to the effort of suppressing the carrier, one also employs the single-side-band style of modulation. However, it had to be double-side-band. Single-side-band would not yield the desired result. The carrier-frequency would have been somewhere in the 300 MHz to 500 MHz vicinity to which the military went when they lost the channel three TV band. When the transition of the signal occurs precisely at the instant that the carrier crosses zero, the resulting radiated signal is especially clean. Thus, is becomes harder to detect by unauthorized radio receivers. It also produces less cross-interference among the drones.
However, to be synchronized, the carrier has to be derived from the master oscillator of the computer. Hence, the stability of the carrier frequency is that of the master oscillator. By the requirements of the FCC, the carrier has to be very stable. Until that time, digital computers employed an LC-controlled master oscillator, which has very poor frequency stability. The designers of the digital computer maintained that a digital computer cannot possibly operate with a crystal-controlled master oscillator! I had to tell them to quit complaining and to come back to me in a month with the computer running with a crystal-controlled master oscillator. They did so and said that the computer became much more stable, as a result of changing from an LC to a crystal-controlled master oscillator. Henceforth, they resolved to employ crystal-controlled master oscillators in all future digital computers. Now, you know who/when.
As you see, we had spread-spectrum. It indeed was spread-spectrum; because, all of the drones employed the same band of frequencies, at the same time. But, the spreading was the result of the high data-rate of the signal. The non-interference and security resulted from the shift-register. The performance depends critically upon:
Rhetorical question: Does anyone employ a synchronized, double-side-band, suppressed-carrier, high-level quadrature modulation in any current designs? I doubt it; because, doing so would require a combined design of the computer and the radio transmitter. A pity, because the synchronization produces a much cleaner signal, resulting in less interference and greater security.
Webmaster@rism.com copyright (c) 1997, 2000, 2003, 2005 by R.I. 'Scibor-Marchocki. last revised on Friday 29-th July 2005.