Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.naic.edu/alfa/galfa/docs/aoprpsep03_galfa_test.ps.gz Дата изменения: Thu Feb 5 18:09:17 2004 Дата индексирования: Sat Dec 22 20:39:05 2007 Кодировка: Поисковые слова: sun halo

TESTING TECHNIQUES AND OBSERVING MODES FOR GALFA 21-CM
LINE MAPPING
1. INTRODUCTION
1.1. Background
The 21-cm line group of the GALFA project has ambitious plans for mapping large areas of sky
for various scientic purposes. The scientic areas include studies of weak line wings as indicators
of energy sources, the Magellanic stream, disk/halo interaction, HI envelopes of molecular clouds,
HI in cold clouds, and HI self absorption (particularly in the Galactic plane). The details are in
the GALFA white paper (see \The GALFA Consortium 2003").
We intend these data to be archival in quality, usable for many purposes. As such, we need
to develop optimum observing procedures that minimize instrumental eects. This proposal is
directed towards studying and ameliorating three such eects: baseline ripple, zero level osets,
and stray radiation pickup from far-out sidelobes. The Arecibo telescope is well known to suer
from these problems, which are functions of both zenith angle and azimuth. We intend to simplify
them by observing only on the meridian, which makes some of the problems dependent only on
zenith angle and, thus, declination.
1.2. Purpose of this Proposal
The purpose of this proposal is to explore the instrumental eects and develop techniques and
observing modes to minimize their in uence. To this end, we will use the four banks of the interim
correlator placed \end to end" in frequency space to cover a total of 48 MHz and sample all four
Stokes parameters. This provides excellent sensitivity for continuum observations, and provides
enough bandwidth to characterize the ripple.
Using all Stokes parameters is important because the baseline ripples occur partly from re-
ections on the telescope structure. Consequently, they are likely to be polarized; we expect (and
hope) that the baseline ripples are nearly 100% polarized so that we can determine their parameters
accurately in the polarized Stokes parameters (Q, U, V) and apply them to the one of interest, the
total intensity Stokes I.
2. MAPPING TECHNIQUES
We wish to develop a good standard mapping technique. When we observe on the meridian,
the straightforward, naive mapping technique is successive drift scans, which are easy and provide

{ 2 {
the required approximate integration time per point. However, each successive scan, done on a
dierent day, is subject to dierent zero level shifts from equipment variability and, also, the
diering positions of the Sun and Moon from one day to the next.
2.1. The Basketweave Technique
This leads us to experiment with covering an area by back-and-forth zenith-angle scanning as
the sky moves past the meridian. On any one day, the scan wiggles back and forth several times
within the area. We repeat this on successive days, beginning at a dierent position so that, by the
n th and nal mapping day, the back-and-forth scans cover the whole area with Nyquist sampling.
The scans on dierent days have many crossing points. At each crossing point the observed
result is the sum of two contributions, the true astronomical one plus the instrumental one. The
instrumental one changes from day to day and, in addition, on any one day it depends slowly on
time. The instrumental contribution also depends systematically on zenith angle. The number of
crossing points is large enough that the problem is overdetermined, meaning that the astronomical
and the instrumental contributions can be distinguished by a least squares t. This is the method
of \basketweaving".
The above two paragraphs lead to two aspects of our proposal
1. A rst assessment of zenith angle dependence of instrumental eects. We will set azimuth at
zero degrees and continuously sample as we drive the Gregorian over its full range of zenith
angle. We will then set azimuth to 180 ф and repeat. We will drive at 1.25 degrees per minute;
this requires 16 minutes to cover the full 20 ф range for each azimuth setting. It takes another
10 minutes to rotate the azimuth 180 ф . All this adds up to 36 minutes. We add 24 minutes
for setup for a total of 1.0 hours.
We wish to perform this experiment once during the night and, also, several times during the
daytime with the Sun at dierent elevations so that we can get a feel for the Sun's eect. We
expect to collaborate with Phil Perillat and do the daytime observations during maintenance
time. However, this won't work for the nightime observations, so our requested time for this
aspect is one nightime hour.
2. Gain experience with Basketweave. We will map a 2 ф  1 ф region using the basketweave
technique, where the region is elongated in the direction of right ascension. We divide the
region into two 1 ф  1 ф halves. One half will contain the strong continuum source 3C273,
located at declination  2 ф ; this will allow us to fully characterize the intermediate-distance
sidelobes in all Stokes parameters. Moreover, using such a strong source will provide an
acid test of the ability of the basketweave technique to provide accurate results when piecing
together observations on dierent days.
We anticipate the other half of the region to be suфciently far from 3C273 that its sidelobes

{ 3 {
will be very weak, so we can use this half as a test for mapping weak 21-cm line emission
with the basketweave (the HI column density is very weak near 3C273).
We choose 3C273 specically for two reasons: rst, it is strong, so that we can measure
weak sidelobes with high S/N; second, with its declination of  2 ф we are observing it near
zenith angle 17 ф , which is fairly high. Here the re ecting \skirt" of the Arecibo telescope
intercepts some of the radiation and scatters it into the feed, thus exacerbating the eects of
stray radiation. We will use Kalberla's version of the stray-radiation corrected LDS survey
to develop a method to evaluate Arecibo's stray radiation.
Scanning at 2.5 degrees per minute, covering this area with fast basketweave scans takes
about 8 minutes per day and requires a total of 14 days. There is also setup time, which
will mainly consist of the telescope moving from the arbitrary position we inherited to the
required position; this can take an additional 10 minutes maximum. Thus we require 18
minutes per day.
We will work with Phil Perillat to develop a command-line driven observing procedure that
minimizes the day-to-day eort required in these repetitive observations.
2.2. The typewriter technique
The classical basket weave drives the telescope at a fast rate|or its maximum rate|back and
forth across the area of interest. This is shown in the left panel of Figure 1. We are concerned
that the repetition of this operation day after day for months might overly stress and damage the
telescope. We re ect on the day of the demise of the 300-foot: it was being used in this very fashion.
Accordingly, we want to try driving the telescope slowly. In fact, in our basketweave of Figure 1
we restrict the rate to 1.25 deg/min, which is half the slew rate.
The problem with slow scans is that it takes longer to cover the area. One gains integration
time, of course, but sometimes one might wish to at least have the option of covering more area
with less integration time. This can be accomplished with the \typewriter scan".
In case the reader of this proposal might not be familiar with the concept of a typewriter, we
brie y digress. A \typewriter" is a mechanical device that was widely used in prehistoric times
for the same purpose word processing computer software is currently used, namely to prepare
neat copies of documents. (Some typewriters were assisted by electric motors, and were therefore
electromechanical, not just mechanical, in nature). A typewriter has a pinched roll to hold the
paper, called a \carriage", together with keys that activate dies with letters that strike an inked
ribbon on the carriage, transferring the image of the letter to the paper. As the typist comes to
the end of a line, he reaches up with the hand to a lever and slams the carriage very fast back to
the starting point of a new line. Thus, the carriage moves forward slowly during the typing and
moves very rapidly backward.

{ 4 {
Fig. 1.| Scanning patterns on 2 ф  1 ф area with 3C273. Left panel, basketweave technique; right
panel, typewriter technique. Basketweave scan rate is 1.25 deg/min; typewriter scan rates are 0.14
deg/min and 1.25 deg/min.
Our typewriter mapping technique works the same way. The telescope scans slowly down in
zenith angle and, when it comes to the lower boundary of the region, moves back up very rapidly
to the upper boundary 1 The next day we move up by one half beamwidth and repeat the process.
Figure 1, right panel, shows the scan pattern. The slow portions of the scans provide complete
Nyquist sampling of the region. The fast portion of each scan ties it together with the other scans,
which were made on dierent days. It takes about twice the number of scans to cover the same
area as basketweave does, and of course, in compensation we gain twice the integration time per
Nyquist sampled point.
We will compare the traditional basketweave and the typewriter methods on the same area.
The typewriter time requirements are similar to the basketweave ones, except that we need 32 days
instead of 14.
3. SUMMARY OF REQUESTED TIME
In summary, we request:
1. One hour at night for a rst-look determination of zenith angle eects.
1 We currently believe that it's better to make the rapid motions up in zenith angle, contrary to both gravity and
intuition, because of the hydraulic braking system (see Perillat 2002). We are currently investigating this question
with the relevant experts.

{ 5 {
2. 46 days at 18 minutes per day, for a total of 13.8 hours, for the comparison basketweave and
typewriter mapping techniques.
REFERENCES
The GALFA Consortium 2003: the GALFA web page alfa.naic.edu/alfa galactic.html and click on
'Draft of White Paper')
Perillat 2002, see http://www.naic.edu/ phil/hardware/vertex/vertex.html
This preprint was prepared with the AAS L A T E X macros v5.0.