Whilst on the romantic subject of coaxial attenuation -

Attenuation is the number-one characteristic of all transmission
lines. From power frequencies and upwards. Yet, quantitativly, it is
the smallest parameter per mile and the most difficult to measure
accurately. An innocent observer might think it is hardly worth
bothering about.

It is inextricably mixed up with system economics. An exact knowledge
even of the temperature coefficient of attenuation is vital to
communications system design.

Around the 1950's I was involved with measurement of attenuation (and
other characteristics) of the first oceanic submarine telephone
cables. A transatlantic coaxial cable, 2000 miles long, has an overall
attenuation at 5 MHz of around 4000 decibels. The temperature
coefficient of attenuation is half of the resistance temperature
coefficient of copper which is 0.4 percent per degree C. Which was
well known to Oliver Heaviside around 1875.

Thus, a change in temperature on the ocean bottom of 0.1 degree
results in a change of 80 dB in the signal level at the far end.
Unless corrected in the repeaters (of which there were about 100) this
is enough to shift signals between the thermal noise level of an
amplifier and its overloaded cross-modulation level.

One of the cable factories was located in Southampton Docks. As cable
came off the production machinery at about 1 mile per hour, it was
coiled in giant circular concrete tanks below ground level, the same
size as an 8000-ton cable-laying ship's hold. Attenuation and other
measurements were made in the tanks by automatic testing equipment.
The cable was then loaded onto a cable ship waiting for it in the
nearby dock.

I designed a special attenuation and phase-shift tester for research
purposes. It did not incorporate an SWR meter. It did incorporate a
phase-locked-loop but it was not until several years later that I came
by chance upon a learned paper by Gruen and discovered how a PLL
really works. The equipment was all tubes. A whole mobile rack of it!
The final output meter was a moving-coil instrument with a scale
calibrated in 0.001 decibels. There were also home-brewed 0.001 dB
stepped attenuators which I had to calibrate myself.

To determine attenuation temperature coefficients in was necessary to
bring tons of ice by lorry from Billingsgate fish market in central
London. One of the concrete tanks was flooded with sea water and the
ice dumped in. It took 24 hours for the temperature to stabilise. I
spent much of the waiting time in a pub in Southampton Town. I never
knew who organised and paid for delivery of the ice which must have
been the most intricate and illegal part of the whole operation.

The data accumulated was rushed to the boffins who immediately began
designing even higher frequency oceanic systems. I was rewarded with a
pat on the back and told to keep my mouth shut. Politics were
involved somewhere.

My tester should have eventually been installed in the Science Museum,
Kensington, London. But long after the job was finished it was stolen
by some unfeeling person and cannibalised for the spare parts. There
was a BC221, straight-line frequency variable tuning capacitor built
into it. I would have liked that for myself.

Hope you enjoyed the story. From what I remember it's mostly true.
Southampton makes a change from Manchester and Leeds. The Queen Mary
was berthed not far from the Cable Ship Monarch.
----
Reg, G4FGQ