Priorities for SPARO observations/tests and analysis as of July 7, 2000.

Updated July 12th to reflect: new suggestions for checks relating
to bad reproduceability, revised plan for 20 km/s calibration, and
third set of Galactic center fields.

Updated July 19th and 20th to reflect:

--> added a description of second Moon tests (1C) and new Sgr B2 tests (1F) that we've

--> added a description of how much more we need on fields gc2 and gc3 (see item 2)

--> added items for after gc3 is done (see 3A and 3C)

Priority 1: Checks, tests, calibrations:

(1A)  grid tests - DONE

(1B) Repeat polarimetry of 6334-I  -  Done, but...

...4 more hours at tau = 1 (or equivalent) is desireable just to
be sure that the April 18th anomalous NGC 6334 result is really
wrong.  Pixel 5 pointed at the peak.  Pointing checks
periodically, as usual.  For NGC 6334, each increase of 0.25 in
tau doubles the required integration time.

(1C) Polarimetry of the Moon -  DONE

(1D) Polarimetry of 20 km/s cloud - perhaps the best test for
instrumental polarization of central pixel.

OK, we can do this without changing the chop.  Here's how it
works: We'll chop in 'software' by differencing neighboring
pixels.  So you can leave the chop at 0.5 degrees for these
observations.  When we go to CSO we'll chop at a 22 degree angle
from the horizontal (approx. perpendicular to the Galactic plane),
and we'll chop by 3.5 arcminutes.  Well process the SPARO data by
differencing pixel 5 from the average of pixels 4 and 6. There
will be additional errors involved, so I've increased the
integration time from what we'd need if we were chopping only 0.1

Also, this measurement will be somewhat sensitive to the
instrumental polarization of pixels 4 and 6, but still it will be
mostly sensitive to pixel 5, as the flux from the 20 km/s cloud is
quite a bit bigger than from the neighboring positions, I'm pretty
sure.  A factor of two I hope - Sharon can tell us as she knows
how to do photometry on polarimetry files.

All we really care about for this 20 km/s calibration is that the
center pixel be pointed at the 20 km/s cloud.  The best way to do
this is to use Sgr B2 as a reference pointing position.  So we
just need a single polarimetry pointing for the 20 km/s cloud. For
purposes of calibration, we will only use the results for pixels
4, 5, and 6.  At CSO in mid-July, we will map about 200 pixels and
average them together to compare with what SPARO gets in that
central pixel.  That will give us an independent check on
instrumental polarization of that central pixel (independent of
Moon).  We will use a 3.5 arcmin chop at the CSO, and do the SPARO
chop by differencing pixels.

I think while we are at it we should do polarimery of Sgr B2 just
as a check that nothing is changing.  So the routine is similar to
the Sgr A science observations - only just two fields, not four,
this time.  I.e.  we do pointing on Sgr B2 about 4 times/day (3 by
3 scans are best, I'd guess), and the polarimetry of the 20 km/s
position and of the peak of Sgr B2.  Just two polarimetry fields. 
It would be good to spend about 4 times more time on the 20 km/s
polarimetry as on the Sgr B2 polarimetry, as Sgr B2 is way
brighter, and way less important.  So, keep the same chop of 0.5
degrees throughout these 20 km/s & Sgr B2 observations.

See "Plan for Winter 2000: science", under "Plans" for coordinates of 20 
km/s cloud (and link to Sgr B2 coords also are found there too).

This is the most time-consuming of the calibration observations. 
I think we need the equivalent of 18-36 hours of tau=1.0, which is
like 1.5-3.0 days of tau=1.2, or 3.0-6.0 days of tau=1.4.  The
goal is a 0.1-0.2% calibration, but even a 0.3-0.5% measurement
would have significant value, as a Moon-independent sanity check
on instrumental polarization.  So even if we can only get a
fraction of the integration time we want, the test still has value.

(1E) A check on whether we might have been pointed at the "negative
reference spots" during our recent GC polarimetry and/or photometry.

This is unlikely, and its OK if we did this error.  All it means
is we wasted half of our observing time.  But if this is what we
were doing we need to know about it so that we can do the
instrumental polarization calibration properly.  Here's how to
test if we were mis-pointed in this way.  Next time you set out to
take a scripted set of observations on GC fields, increase the
size of the photometry maps of Sgr B2 from 3x3 to a map that is
wide enough to catch both of the refence beams.  maybe 9 by 3, or
11 by 3.  This will waste some observing time but will provide
this good sanity check if you do this for one remote observing

OK, so that should work for photometry, but what about
polarimetry?  How can we be sure that the polarimetry is done at
the same pointing as the photometry?  Maybe a bug in your script
or some weird difference between Bob's polarimetry and photometry
code could mess this up? Well, its easy to check...start your
usual scripted observations, with the extra-wide photometry map,
but hang around long enough to watch the photometry execute and,
when the script tells you that you are observing the central
position of the map, write down the left and right beam offsets as
they appear on the comsoft screen. Also write down the Mac
file-number during that file - the central file of the map. Then,
when the polarimetry of Sgr B2 starts, do the same: write down the
left-beam and right-beam postions as they   They should match.
Record everything in the logbook.

(1F) Cross calibration of i.p. using Sgr B2  

- measures i.p. of outer pixels relative to center pixel  

The idea is to successively put each pixel on the peak of Sgr B2
and do polarimetry.  some points to keep in mind when writing the

...cycle through all the pixles rather than staying on one pixel
for a long time.  That way the "i.p.-difference" will be less
vulnerable to any potential changes in i.p. due to

...spend about 3 times as much time on pixel 5 as on any other
pixel. (that's best way to reduce overall errors in the
"i.p.-difference" between an outer pixel and pixel 5.)

...we don't need to do pixel 3.  it is not used for polarimetry.
We don't need to do pixels 2 and 7, as the i.p. for these pixels
has been pretty well determined by Moon tests.  That leaves 1, 4,
6, 8, 9 and of course 5.

...We need the equivalent of about 1 hour at tau=1 for each pixel.
(3 hours for pixel 5). Thats a total of 8 hours at tau = 1.  For
GC, each increase of 0.2 in tau increases the required integration
time by a factor of two.

...The occasional 3 x 3 pointing check (as usual) is always a good
thing, as needed, to keep us reaonably well pointed at peak.

...Pointing is more impt. for this test than for science
observations. But, even if we are off by an arcminute, the tests
will still be quite valuable and will give us most of what we

Priority 2: Getting more vectors on gc2 and gc3:

We have finished gc1.  As of July 19th we still want the
equilvalent of 24 additional hours at tau=1 on each of gc2 and gc3.

Field designations refer to Greg's tiling that is posted at "Plan
for Winter 2000: science", under "Plans".

gc2: (-1/2, 0) (-1/2, 1) (0,2) (0,4)
gc3:  (1/2, 0) (1/2, 1) (0,3) (0,4) (-1/2, 2) (0,5)

Note that each increase of 0.2 in the tau makes the needed
integration time double, for declination of GC. 

Priority 3: Beyond gc3...

(3A) Here is gc4

gc4:  (0,-3), (0,-2), (0,6), (0,4)

48 hours at tau=1 should be enough for gc4

(3B) If the tau is above 1.6 or so, it is best to extend the
polarization map of 6334, rather than integrate on the GC.  The
penalty for higher-tau is less for NGC 6334 than it is for GC.  I
aoplogize that we have not worked out pointing postions for 6334,
but if you look at the IRAS map you can see there is a whole ridge
of emission trailing off from the peak we have been mapping down
towards the SouthWest. Just follow the ridge.  Ten hours at tau =
1.6 should secure some detections at any of these new positions.
Another valuable thing to do is just to make a photometric map of
the whole 6334 ridge.  Look at the IRAS map on the web page and
make a judgement as far as the size of map needed.  It may be
better to make the map first as a way to pick pointing positions.
So, bottom line:  if 1.6 < tau < 2.0, do 6334.  Of course use 0.5
degree chop for photometry/polarimetry of 6334.  It will also be
good to repeat the polarimetry of that main field.  It does not
take long to get a good result as that main field is very bright.
We don't need much more than the equivalent of 12 hours of data at
tau of 1.3 for this "main" NGC 6334 position, called NGC 6334-I.

(3C) After gc4 (if we get that far) we want to try to increase the
chop and see how bad the microphonics are.  This test will tell
us what we need to do next summer, and if we manage to increase the
chop successfully this winter, we will take a peek at some off-plane