Paul Rosen, an electrical engineer who in the mid-1950s helped develop the high-speed modem, spurring revolutionary progress in the nascent industry of telecommunications, died of congestive heart failure July 20 at his cottage in West Bath, Maine. He was 88.
The technology behind the modem -- a device that converts data into [analog (1)] signals that can be passed through channels such as [legacy (2)] phone lines -- has existed in primitive forms since the late 1940s. But in those days, phone lines carried data signals inconsistently, and information was transmitted slowly.
While working at the Massachusetts Institute of Technology's Lincoln Laboratory in 1958, Mr. Rosen and a colleague, Jack Harrington, patented a device that rapidly transmitted large amounts of data over phone lines.
Their invention, "Method of Land Line Pulse Transmission," helped expand computer networks nationwide by significantly accelerating the flow of data over phone lines.
------
(1) and (2) added by poster...
the rest, as they say, is history:
formatting link
formatting link
Paul Rosen, center, stands in front of a sequential decoder, an early piece of telecommunications equipment. His work was a crucial addition to a landmark Army defense project during the Cold War.
Could someone explain in layman's terms what exactly his invention did? I checked the article but it didn't say. It mentioned that the Bell System was able to make minor changes and utilize his invention without royalty.
Thanks.
***** Moderator's Note *****
I _think_ that the inventors were the first to realize that baud rate and bit rate didn't have to be synonymous: i.e., that it was possible to have more than two signal states ("Mark" and "Space") in a phone line.
In other words, they were able to increase the _bit_ transfer rate by increasing the symbol store available to _carry_ the bits; a change that allowed the higher throughputs without banging our electronic heads on the Nyquist (or is it Shannon?) limit.
The basic problem was this. How to send digital data from then-new computers over an existing standard telephone circuit. The network was designed to carry analog voice signals of varying frequency and amplitude. Digital data generally is a fixed frequency and amplitude. So something was needed to convert one to the other. Hence modems, or MOdulator/ DEModulators.
The SAGE project that Rosen was working on was the first large scale requirement for sending digital data over analog circuits for long distances. He found a modulation scheme to do that basic function. "Encoding" of "multiple bits per baud" came later as more robust modulation schemes were developed (PSK, QAM, etc) that allowed for more than 2 symbol states.
The above reference is extremely interesting* and explains what he did. Quoting a part of it below:
"I had made lots of tests on telephone lines. These were pretty lousy lines. Some of them ran on posts set in the ground used to keep out cattle. Therefore the telephone lines had propagation characteristics that were very different from one another. There were two things one really could not predict - the bandwidth and the phase response. I discovered, simply by messing around, that if I arbitrarily picked something like 2 kilohertz as the upper end of what a reasonable telephone line would do, and put a simple low-pass filter on the modem output, the phase of the filter would roll off gently before the signal encountered the severe phase distortion caused by the telephone line itself. The result was that this line could be pinged and a fairly representative pulse would be obtained at the end. The detector I used was a very simple amplitude detector, a diode and a capacitor. Of course there was amplification in between and there was digital stuff that somehow synchronized with what was coming off the radar.
At the time Bell Labs had great men working on this, and they said if one drives a telephone line over 600 bits per second, one has arrived at the end of the flat earth, and you've had it. With my simple device I did 1800 bits per second consistently."
As I understand it, back then long distance telephone lines were of poor quality (open wire) and as a result modem speed was limited. Mr. Rosen invented a circuit (as described above) that would be able to push higher speeds over the phone line despite the distortion; that is, his modem ignored the distortion. (I don't understand about "phase roll off" but that's what apparently did the trick.)
I wonder how business people felt making toll telephone calls--which were very expensive back then--over a lousy line. I suspect letters and telegrams were major media until the telephone network improved. (When I first made toll calls in the early 1960s I remember the lines had a slight 'white noise' in the background--you heard it come on when the connection was made--but the conversation was certainly comfortable and adequate. I don't know how it was in the 1950s.)
Well into the 1980s long distance directory assistance (NPA+555-1212) had lousy connections--very bad echo. But I guess they figured for a free call they'd route it over their 'cheap' trunks. I don't know if they even still offer that service--they certainly charge for it if so. Anyway, they now advertise that 411 has national listings (for a fee, of course).
*It gives his life story and his education and training.
***** Moderator's Note *****
The lines weren't "lousy" when used for voice: the problem with data transmission is that it is subject to interfernece and distortion which humans don't perceive. Normal phone calls sounded, well, "normal".
As for the "white noise" you heard on long-distance circuits, it was added deliberatly. The "L" carrier was so quiet compared to older methods that many subscribers would abandon calls because they thought the other person had hung up, so noise generators were added to supply the reassuring susurrus we all came to associate with "normal" long-distance lines. It wasn't until Sprint, which didn't want to go to the trouble of retrofitting the fiber network to sound like Ma Bell's, launched it's "Pin Drop" campaign that public expectations of long-distance circuit "sound" changed to what we expect now, namely "nothing".
It's probably worth mentioning that the "sequential decoder" shown in the Washington Post, of which the original photo is hanging on the wall in the room in which I'm writing this, was a major contribution that my dad, along with Claude Shannon (left) and my dad's long time colleague and friend Jack Wozencraft, (right) made to the modern technology landscape. The project did the fundamental work of what has become known as digital signal processing. That's my dad in the center of the photo, looking justifiably very proud of what they had done.
The patent (2,850,573) covers the actual circuit design and not just the modulation theory of a modulator and demodulator able to reach bit rates close to the carrier frequency. Some parameters are variable, but with the example configuration in the patent, the circuits reaches 1600 bps using a 2 kHz amplitude modulated carrier.
While we're on the subject of modem history, the following article in the Western Union Technical Review describes FSK technology of the early 1960s, including technical details on transmission characteristics. How much of that stuff do we take for granted wiith today's technology running at much higher speeds? [click next/ previous to navigate]
formatting link
The issue also contains other articles on data communications technology.
Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.