The problem here was that the page was designed for someone using
a GUI browser with images on over a T1 link; I was using it with
a text only one over a V.34bis modem - it would have taken around
of 2 minutes to load the page into a GUI one, though.
This includes the part of the article defining the problem, but not the
solution. Basically it says that a signal in the waterhole region
would be spread to around 0.1Hz in bandwidth and would suffer up to 10:1
fading (not sure if this is power or field strength) as a result of the
interstellar medium. It would also have unknown doppler shifts varying
by of the order of 1kHz over a few hours.
>From the project Argus point of view, fading would be a problem, because
it might cause a valid signal to fail the verification tests. The
bandwidth would not be a problem for the current typical station, because
the bin width is much more than this. Doppler would be borderline, the
worst case is comparable to one bin width per integration time, but, in
my view, there is no reason, other than Doppler, for not integrating for
the full transit time (using rolling average logic) and the only thing
preventing 1Hz or 0.1Hz bins would be local oscillator stability; this
would mean some Doppler compensation would be needed.
At 0.1Hz bin width, the change in earth rotation Doppler shift would be
enough to take the frequency out of the bin during the period needed to
collect the samples needed to do one FFT. However, personally, I would
expect a beacon to be Doppler compensated, relative to the sender's best
guess of an inertial frame and so it would only be necessary to compensate
the receiver for a similar frame.
Doppler - change in apparent frequency of a wave due to motion of source
FFT - Fast Fourier Transform - a specific algorithm, but used to indicate
any algorithm attempting to determine the power versus frequency graph
for a signal.
bin - FFT and similar algorithms split the frequency range up into chunks,
typically of fixed length; these are bins.
integration - adding up multiple consecutive samples as the noise
(measurement error) will grow as the square root of the number but the
signal will grow linearly.
inertial - not being accelerated by any force, including gravity (note
that orbiting planets are accelerating towards their sun).
frame - frame of reference - in this case, defines the zero and axes for
defining a velocity.
transit time - the time it takes the earth's rotation to move a source from
one edge of the antenna beam to the other.
power versus field - like power and voltage in an electrical system.
waterhole - the frequency range in which the earth's atmosphere doesn't
absorb a lot of the signal.
local oscillator - microwave signals are shifted down in frequency by
combining them with a locally generated signal in such a way that the
difference in the frequencies is generated. The local signal is generated
by the local oscillator. In project Argus type systems, the frequency is
shifted from around 1420,000,000 Hz, to somewhere around 1,800 to 20,000
Hz (centre). There are many reasons for doing this, but if it weren't
one would still need an accurate local references of the same precision,
which needs to be significantly better than the bin width in 1420,000,000 Hz.
power, frequency, bandwidth, voltage - assumed common knowledge.
GUI - graphical user interface (Windows, Macintosh, X-Windows, etc.).
T1 - US standard for carrying 24 telephone circuits digitally at a little
over 1,5 million bits per second.
V.34bis - specification for a modem operating at up to 33.6 thousand bits