The Bell Labs history*--"1875-1925" volume covers how the Bell System transmission media evolved over time, including detailed discussions about loading coils, repeaters, and carrier multiplexing. The "Switching 1925-1975" volume continues the discussion and also covers various long distance signalling techniques.
Long distance transmission was difficult until improved vacuum tubes were developed, perfected, and deployed in the 1920s.
To help offset the high cost of long distance telephone lines, they were also used to carry telegraph traffic. Arrangements had to be made to keep the two bands separate, especially in relation to repeaters and signalling.
Specifically: "A History of engineering and science in the Bell System / prepared by members of the technical staff, Bell Telephone Laboratories" Vol 1. The early years (1875-1925) Vol 3. Switching technology (1925-1975).
I highly recommend both volumes as a resource for telephone history. They should be available in large libraries.
In addition, the Bell Labs Technical Jounral, which is all available on-line, provides technical background from 1922 onward.
Thanks HAncock4, I'll try to track those references down.
Perhaps I should have qualified my original statements.
"Self proclaimed historians", not neccessarily technicaly minded, who state that something happened but not how; and who dont always distinguish between telphone/telegraph or a single cable with 12 lines as opposed to a single wire with 12 conversations.
"not uncommon" instead of common. eg,
- 1884 New York to Boston ~ 200 miles
- 1891 London to Paris ~200 miles, but links anyone in UK that can access London to anyone in Europe that can access Paris
- 1892 New York to Chicago ~700 miles (1000 miles cable -ny times October 19, 1892) and consequently Chicago to Boston. ....
The suggestion seems to be that you just need to use heavier wires (more surface area?), or did they need to do more than this?
I agree and hence my question. I can't find the FDM reference : ( However, I try not to underestimate the ingenuity of engineers of those days. Radio telegraphy was developed in the 1890's, and radio transmission of audio in the 1900's, all with no active components. (both from wikipaedea) I just don't think like they did.
I've had a few more days to ponder this, Actually they could have used an electro-mechanical device to convert voice signals to a varying resistance. This in serries with an AC signal could be used as a modulator (AM). It could even be used for amplification, if the electro-mechanical device consumes 0.5W to go from extreme to extreme but the AC signal it modulates could deliver 50W. Any one know what they actually did?
I'm still not sure how they would recover those conversations. A heap of passive serial tuned circuits, all in parallel seems unworkable.
You carry multiple conversations in a pre-electronic system through using phantom and simplex circuits. You can take two pairs and get three lines across it... two differential signals across each pair and then a common mode signal between the two. But that's the degree to which you can go.
Time division and frequency division multiplexing WERE used for telegraph circuits, using mechanical distributors and mechanical tuned reeds. But not for telephone.
By using widely spaced pairs to reduce shunt capacitance, using very large diameter wire to reduce series resistance, using loading coils, and shouting very loudly into the telephone.
Phantom circuits were used in actual service and additional schemes were not, however, they were investigated and documented. For example, a Ghost circuit was a simplex circuit over two Phantom circuits and a Wrath circuit was a Phantom circuit over two Ghost circuits. Ghost and Wrath circuits were impractical as the balance required could not be achieved in the "real" world outside of the laboratory. Neat stuff, though...
It should be noted that implementations of new technology back then often had three levels:
--Initial experiment. This would be making a long distance call over a long circuit, the first transmission of television, the Morris IL ESS, making a working transistor, etc. It is a long way from successful experiment to actual commercial installation.
--Commercial installation: a regular installation for paying customers, such as the Susc. NJ ESS, public TV broadcasting, use of transistors in production productions.
--Widespread usage: It takes time after the first commercial installation for a new product to become widespread. For instance, the Bell System continued using some open-wire transmission into the
1970s before more modern methods replaced it. While commercial TV came out in 1940, then more so in 1948, it took time before all cities had it.
In the case of early long distance transmission, the Bell System may have had a successful test of a long distance conversation early on, but it took time before all the elements necessary to make that happen were developed to an adequate level to be affordable, reliable, etc to use in real service. For instance, many improvements to vacuum tubes were required to make repeaters possible--a higher vacuum, physical arrangement of the components, materials used in the components. For example, considerable experimentation was required to develop the optimum metal alloy of the grid and filament, as well as techniques to manufacture those components in quanity.
When the transistor was developed, it took about ten years to understand its operating principles, figure out the optimum materials and 'doping', and development high-volume reliable production manufacturing. It took a few more years to get the cost down below tubes (there were plenty of consumer electronic products of the early
1960s made of tubes, not transistors).
The first volume of the Bell Labs history (op cit) goes into detail about all of these issues.
Some of this history is in a column by Mischa Schwartz in the IEEE Communications Magazine, May 2008, p. 20. "The Origins of Carrier Multiplexing: Major George Owen Squier and AT&T" He says, among other things,
"George O. Squier...was the inventor...in 1910, of telephone carrier multiplexing...His invention was initially rejected by AT&T engineers as not being commercially viable. After prodding by [John Stone] Stone, AT&T officials began a reappraisal of the 'wired wireless' system, as Squier chose to call it, and by 1914 development of a commercial system was underway."
"It wasn't until 1918, however, that the first commercial implementation of carrier multiplexing for telephony was introduced by AT&T in the United States. This system enabled five simultaneous conversations to be carried over a single pair of wires..."
"...the first demonstration of a carrier multiplexed system was carried out by then Major George Owen Squier of the U.S. Army Signal Corps on September 18, 1910 in Washington, D.C. ... transmitted two simultaneous analog voice signals over a single seven-mile-long private telephone circuit...One signal was that of a normal telephone circuit conversation, commonly referred to today as being sent at baseband. The other signal was a modulated high-frequency signal, with the high-frequency carrier used in the experiments varying in frequency from 20 kHz to 100 kHz. ....Four patents titled Multiplex Telephony and Telegraphy were issued to Squier on January 3, 1911, with the work being presented publicly as a paper read at the 28th Annual Convention of the American Institute of Electrical Engineers in Chicago, Illinois, June 28, 1911."
"By fall 1914 an experimental vacuum-tube-based carrier-multiplexed telephone system carrying two simultaneous signals over a single circuit had been set up in the laboratory of R. A. Heising."
Suggesting that the original didn't use active components.
Turns out that the problem was phase distortion, more than attenuation. (I need to study transmission lines :)
I'm still not sure how signalling worked. The US seems to have sent tones in the voiceband that could be recovered as signals? But I'm not sure if they did the same in Australia.
In fact, after automatic exchanges were introduced (especially in the Step-by-Step system used in Australia), when did it become possible for the subscriber to make inter-exchange calls (say from one side of a city to the other) without operator intervention? How did the signalling work in that situation?
Signalling was almost always ringdown with no supervision. Much later straightforward trunks were adopted on some busy routes, but were substantially more expensive.
In the U.S. it was normal common battery type supervision in most cases. Each step-by-step operation went just as before internally, and if the next selector was in a different office it just went forward as before.
This 1922 article explains a long distance cable--the physical construction, loading coils, and repeaters.
This 1923 article futher discusses transmission.
As to signalling, different ringdown frequencies were used at different times and locations because of the need to pass through repeaters and loading coils, to not interfere with telegraph signals superimposed on the wire, and within the limits of electronics of the era. The Ball Labs history Vol 1 goes into considerable detail on this.
Bell System Technical Journal, 1922-1983 index page: