Project Cyclops, A Design... - Department of Earth and Planetary ...
Project Cyclops, A Design... - Department of Earth and Planetary ...
Project Cyclops, A Design... - Department of Earth and Planetary ...
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
sec,therebyobtaining100samples<strong>of</strong> thecomplete<br />
powerspectrumwith0.1-Hzresolution. These are then<br />
added in perhaps 400 different ways to synchronize with<br />
the assumed drift rate, but since additions with almost<br />
the same slope are correlated, there are only about 100<br />
independent signals obtained from the adder, each<br />
having 2X109 independent Nyquist intervals per polarization.<br />
Thus, a total <strong>of</strong> 4×10 _ tests is made per<br />
observation, <strong>and</strong> to keep the probability <strong>of</strong> a false alarm<br />
to 20% for the entire observation, the probability <strong>of</strong> a<br />
false alarm per test must be about 5X 10-_ 3<br />
With the threshold set so that the probability <strong>of</strong> the<br />
noise alone exceeding it in any one Nyquist interval is<br />
5× 10-l 3, the probability <strong>of</strong> missing the signal is 50% if<br />
the received signal-to-noise ratio is unity (0 dB) in the<br />
0.1 Hz b<strong>and</strong> or -93 dB for the 200-MHz total b<strong>and</strong>.<br />
l'or only a 1% chance <strong>of</strong> missing the signal an additional<br />
2 dB <strong>of</strong> signal power is needed. We know <strong>of</strong> no other<br />
method <strong>of</strong> processing the signal that can even approach<br />
this performance. The technique is believed to be an<br />
optimum one, though <strong>of</strong> course other <strong>and</strong> better means<br />
<strong>of</strong> implementing the operations may be found.<br />
IMAGING THE RADIO SKY<br />
The proposed <strong>Cyclops</strong> array is orders <strong>of</strong> magnitude<br />
more powerful than any existing fully steerable radio<br />
telescope. As such it would be a magnificent tool for<br />
finding <strong>and</strong> studying distant weak radio sources <strong>and</strong> for<br />
mapping the structure <strong>of</strong> known sources. Because all the<br />
IF signals are returned independently to the central<br />
headquarters, they can be combined to form more than<br />
one simultaneous beam. In fact, with n elements we can<br />
form n independent beams, or more than n correlated<br />
beams. By arranging the beams in a closely packed array<br />
<strong>and</strong> portraying the sky brightness measured by each<br />
beam as the brightness <strong>of</strong> a corresponding point on a<br />
screen, we can form a real-time high-resolution map <strong>of</strong> a<br />
portion <strong>of</strong> sky within the beamwidth <strong>of</strong> the antenna<br />
elements. In this way the time required to map the sky<br />
or to search for new sources is only one nth as great as<br />
with a single beam.<br />
Several ways <strong>of</strong> doing this imaging were studied, none<br />
<strong>of</strong> which is ideal in all respects. The proposed method<br />
involves reradiating the signals from the antenna elements<br />
as electromagnetic waves from antennas in a<br />
scaled down model <strong>of</strong> the receiving array. If the scale<br />
factor is o, the angular magnification <strong>of</strong> the telescope is<br />
l/o. The imaging, if broadb<strong>and</strong>, must be done at the<br />
original received frequency; otherwise, the angular magnification<br />
is frequency dependent <strong>and</strong> changes appreciably<br />
across the received b<strong>and</strong>.<br />
72<br />
The model, or signal, array focuses the radiation on a<br />
second array whose receiving antennas pick up samples<br />
<strong>of</strong> the signal in the image plane. The energy collected by<br />
each receiver is then detected <strong>and</strong> converted to a<br />
proportional brightness <strong>of</strong> an element <strong>of</strong> a display<br />
screen.<br />
The useful angular field <strong>of</strong> the image is inversely<br />
proportional to frequency. Thus, to keep the number <strong>of</strong><br />
points in the image constant, the separation between the<br />
two arrays must be proportional to the RF center<br />
frequency. In the proposed system, the signal <strong>and</strong> image<br />
arrays are 20-m in diameter <strong>and</strong> their separation varies<br />
from 40 to 120 m over the tuning range from ! to 3<br />
GHz. lm_,ging is not attempted from 0.5 to 1 GHz.<br />
The size <strong>of</strong> the proposed imager is awkwardly large.<br />
To prevent coupling to the receiving array it must be<br />
housed in a completely shielded space, which, in turn,<br />
must be lined with microwave-absorbing material to<br />
prevent undesired reflections. The dimensions <strong>of</strong> the<br />
anechoic chamber needed are about 75 ft wide, 75 ft<br />
high, <strong>and</strong> 550 ft long, a building roughly the size <strong>of</strong> the<br />
turbine house<br />
<strong>of</strong> a large power plant.<br />
The building volume needed is directly proportional<br />
to the number <strong>of</strong> antennas in the array, to the ratio <strong>of</strong><br />
the high to the low frequency limit <strong>of</strong> operation, <strong>and</strong> to<br />
the cube <strong>of</strong> the wavelength at the low frequency limit. If<br />
we were content to do the imaging over the b<strong>and</strong> from<br />
1400 to 1700 MHz, which includes the H <strong>and</strong> OH lines,<br />
a chamber 60 by 60 by 125 ft would suffice. However,<br />
even the cost <strong>of</strong> the proposed imager is small compared<br />
the total system cost, so we have proposed the large<br />
version, leaving possible compromises to the judgment <strong>of</strong><br />
future study teams.<br />
The shielding requirements, though severe, involve<br />
only a few mils <strong>of</strong> copper. Most <strong>of</strong> the problems are<br />
associated with bringing leads into <strong>and</strong> out <strong>of</strong> the<br />
chamber. All shielding problems are greatly reduced if<br />
the imaging frequency is <strong>of</strong>fset from the receiver b<strong>and</strong><br />
by the IF b<strong>and</strong>width. This introduces an appreciable, but<br />
tolerable, amount <strong>of</strong> lateral chromatic aberration.<br />
The proposed imager can image both polarizations<br />
simultaneously. By combining the intensities, we can cut<br />
the integration times in half. If the selected polarizations<br />
can be varied, then subtraction <strong>of</strong> the intensities would<br />
reveal polarized sources. This might be a powerful search<br />
tool for unusual types <strong>of</strong> emissions in the universe. In<br />
any event, the imaging capability would reduce by about<br />
three orders <strong>of</strong> magnitude the time needed to make<br />
certain kinds <strong>of</strong> radio astronomy studies, <strong>and</strong>, for the<br />
first time, would place radio astronomy on a more<br />
nearly equal footing with optical astronomy so far as<br />
data collection rates are concerned. Who can say what<br />
major discoveries might ensue.'?