# Cable propagation

• posted
I've been reading a lot about early trans-Atlantic telegraph cables,
circa 1858. These cables only transmitted a few words per minute, and
because they had no in-line amplifiers, *all* the energy for the signal
had to be supplied from the transmitting end. Because the cables were
so long and small, there was tremendous attenuation - so that a
sensitive mirror galvanometer had to be used to detect the signal.
Looking at the specs of the first trans-Atlantic cable, it would be the
a little smaller than a No.12 AWG wire (only 80 lbs of copper per
statute mile!). WIth today's copper standard this would be about 20,000
ohms of resistance one-way - and copper wire of that era was made for
mechanical, not electrical, purposes and could vary in resistance by a
factor of 2 in any given run. I haven't seen or worked out what the
total capacitance was, but assuming it was comparable to Belden 9913
what *is* the dielectric constant of gutta-percha, anyway?), I get over
200 microfarads of total capacitance. This goes a long way to explaining
the slow keying rates.
I then got to thinking how well radio waves run through a cable.
Supposing you bought a couple of miles of Belden 9913, a low-loss,
premium grade coaxial cable. Suppose you hooked up a pair of those
little FRS radios to the cable. They put out about 1/2 watt - call
this -3 dB(W). I don't know what the sensitivity of these little guys
is, but it can't be much less than 1 uV in 50 ohms - which works out
to -136 dBW. So, an FRS radio has a power budget of 133 dB to burn.
Here's the part that surprised me...if you hooked up a pair of "2-mile"
radios to this very good grade co-ax, you wouldn't even get 1 mile
range! The cable has an attenuation at 400 MHZ of 2.7 dB in 100 feet -
so you would use up the whole 133 dB of power margin in only 4900 feet.
(And in free space the "2-mile" FRS radios are good for a lot more than
two miles.)
Well, I thought it was interesting...until doing this calculation I
think I would have guessed that the range in a cable would be better
than free space.
Bill
( Check out John Griesemer's "Signal & Noise: A Novel" for the epic
brawling saga of the men who laid the trans-Atlantic cable and the women
who loved them...cliched, but accurate description. Would make a great
movie, though there may not be enough telegraphers and electrical
engineers out there to make it a profitable movie)
Loading thread data ...
• posted
in article dRJte.5020\$ snipped-for-privacy@news1.mts.net, Bill Shymanski at snipped-for-privacy@mb.sympatico.ca wrote on 6/20/05 6:14 PM:
This is certainly an interesting subject!
I think that there is an error in your estimate. 20K and 200µF give a time constant of 4s, not taking into account the distribution in the circuit. You will probably find the writings of Kelvin and Heavyside of interest. In particular, there was great prejudice against adding series inductance. My guess is that steel protection had some benefit in increasing such inductance.
On phenomenon noted by Heavyside was that failing insulation on a cable enabled greater channel capacity at the same time the signal was waning. That is easily understood these days.
Bill
• posted
The 1858 TransAtlantic Cable failed because the Chief Engineer on the European side kept cranking up the DC voltage in an effort to improve performance. Eventually the higher voltages punctured the Gutta-Percha dielectric to the seawater ground and the cable was effectively shorted out after about one month of operation.
A Scientific Board of Inquiry documented these reasons and subsequent cables were outfitted with the high-sensitivity Galvanic detection methods that allowed lower voltages to be used, but with very weak currents.
A reliable TransAtlantic Cable was not established until after the American Civil War.
Beachcomber
• posted
BTW, an excellent read, although not technically detailed book on this subject is "The Victorian Internet", by Tom Standage,
formatting link
covers the laying of the transatlantic cables (including things like them snapping whilst being laid), problems with driving them incorrectly at high voltage, etc.
It's full of amusing anechdotes. The first undersea gutta-percha cable was ordered from the Gutter-Percha Company (which nowadays calls itself Cable and Wireless) by John and Jacob Brett to lay across the English Channel. They didn't get very far laying it before they hit the first snag -- it floated! When they finally got it laid by weighting it down, it lasted a day before a french fisherman 'caught' it, thought he'd discovered some new type of eel which was so long he couldn't find the end, and so just cut a piece out of the middle to take back to shore for analysis.
Some years back, BBC Radio 4 serialised it into 5 broadcasts, and I've heard them repeated at least once. An excellent book.
• posted
Here's an interesting story about the first cable using vacuum tubes:
formatting link
I worked on the tubes for TAT-1 at Bell Laboratories in 1949. They were modeled on tubes which had several tens of years on life test. The glass vacuum envelopes were modified so the plate and grid connections were at opposite ends. The heaters were wired in series and fed with high DC voltage from each cable end. The plate voltage was the voltage drop across the series heaters in each amplifier. They lasted over twenty years!
PS: I don't believe the part about the pizzas. (The bake-out ovens weren't large enough!) :-)
• posted
"Andrew Gabriel" skrev
formatting link
He covers the laying of the transatlantic cables (including
There is also a book by Bern Dibner about the subject online:
formatting link

• posted
explaining
circuit. You
inductance. My
Yep, *huge* time constants...run the numbers: 9913 has about 24 pf per foot, the 1858 cable was 5280 * 2000 = ten and a half million feet, 240 million picofarads is 240 uF. No. 10 AWG wire (larger than the first cable!) has a resistance of 1 ohm per thousand feet, so 10,000 ohms - and a No.10 AWG is 166 lbs/statute mile, twice the weight of Mr. Field's first attempt.
Turns out gutta-percha has a dielectric constant of 2.6 - so it will have more capacitance per foot than air-spaced 9913 (which would squish flat at 14,000 feet depth). Someone writing about cables of a much later era (1922) said that a typical submarine cable of that time had between .2 and .5 microfarads per mile of capacitance.
Early cables ran as slowly as 7 words/minute (at about \$10 in 1866 purchasing power per word) - in fact in Mr. Grisemer's novel one character remarks that the cable would still be profitable at 4 words per minute. Cable Morse was bi-polar - a negative pulse for a dash, a positive pulse for a dot.
There's tons of stuff on the Web including a transatlantic cable Web site with all kinds of historical data. Apparently the 1866 cable did a steady business at \$2500/day at \$10 a word, a year later they dropped the rate to half-price and made \$2800/day, still not running the cable at full capacity. Only governments and newespapers could afford these rates - one war report cost over \$5000 in contemporary money (kind of like today's journalists with the \$6/minute satellite phones).
The 1858 cable was capitalized at 350,000 pounds sterling, about \$1.7 million US (in 1858 money). At \$2500/day, simple payback is 680 days...after that the cable is a pure money spinner. Though many of the first generation of gutta-percha cables didn't last long, often failing within 5 or 6 years or never getting into service.
Truly amazing and audacious - to go from the first gutta-percha imported into England in about 1842 to using ton quantities of the material to insulate a trans-atlantic cable only 16 years later is an astonishing pace for technological innovation. It took home computers, what, almost 20 years to really catch on?
Bill
• posted
in article Tc2ue.5101\$ snipped-for-privacy@news1.mts.net, Bill Shymanski at snipped-for-privacy@mb.sympatico.ca wrote on 6/21/05 5:24 PM:
I am not doubting high time constants. With a four second time constant it would be difficult to transmit at 7 wpm. Do you know what the inductance was? I am sure it was low, but even a little bit might help speed up transmission.
Bill
• posted
It's still a transmission line. The losses are great and it will have a complex characteristic impedance. That it would take a certain time to "charge" the line really doesn't matter.
• posted
Yep!
Lots of foolish folks think that the 20th century was the technolocy century.
But in reality, the 19th century saw the greatest advance in technoloy and the change in the lives of "normal" people.
In 1800 it took WEEKS to MONTHS to send a message 3000 miles. In 1900 it took minutes or seconds (depending upon how close your were to the telegraph office).
The same thing can be said about transportation. Transcontinental travel time went from 3 months to about 3 or 4 days in the 19th century. In the 20th, it went down to about 3 or 4 hours.
• posted
With all the airline delays now, the time to cross a continent has gone up!
• posted
Bill Shymanski schrieb:
Hello,
read here:
formatting link
Bye

PolyTech Forum 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.