Summary of dicussion during chop-nod analysis skypecon of Feb 27, 2008

Martin, Hiroko, Larry, Lero, Giles

_________________________________________________________________

skype:

this time it was kind of echo-ey.  We'll all bring headsets next time.  

_____________________________________________________________

NGC 1333 analysis (Nov. and Dec. 2007 runs):

Hiroko did the rgm for Nov. and posted it on the teamsite entry for that run.  There were 221 active 
pixels, similar to August.  For the high-gain case, there were 14 pixels that were not flagged which 
*were* flagged for the low-gain case.  (They were flagged as "1".)  Hiroko used ten files from 
Darren to produce the rgm, and she used median filtering.

Martin said Mike had already used this rgm and it worked well.

So the *basic* analysis is complete.  In addition to the four remaining issues from the last 
skypecon (see below) here is one more that Martin raised:

5. There are some vectors for low-flux regions.  It was suggested that we get rid of them since they 
are for very low fluxes.  Giles and John know how to set a flag in polsharp that will cut off 
vectors for areas where the flux is below a certain threshhold.  But how do we justify this?  Three 
ideas were discussed:

A. If the s/n for I is not good, then the (q,u) are suspect.

B. reference beam contamination.  This problem can be estimated from work by Schleuning et al. 97 
(PASP) and Novak et al. 97 (ApJ).  Were some of our files taken with bright sources in the reference 
beam?

C. compare the I map from SHARP with that from SHARC-II (Darren has supplied).  See if SHARP has 
background subtraction issues.  If it does, don't trust detections made where flux levels are 
comparable with SHARP background subtraction error.

(I just thought of a fourth - what is the reproduceability of these particular vectors?  Is it worse 
than for the ones at the peak?)

For reference, here are the four issues we discussed on the last telecon:

1. reproduceability?  reduced-chi-squared?  We have only 3 bins, so it dicey but we need to address 
the issue.

2. How sensitively do the results depend on rgm ?  Mike reported that the december data analyzed 
with the August rgm gives totally different results than what we see in the November data or in the 
December data when analyzed with the rgm Martin used on the mountain.  (Where did this rgm come 
from?  We don't know.)

3. are we truly getting information about the spatially extended B-field in this source or is all 
the polarization essentially coming from a point source at the center?

4. what about 4B at the edge of our field?  Do we believe it?

_____________________________________________________________

M82 analysis - April 2007 runs:

Lero sent new results for one day in early April.  He showed that the pointing corrections sharpen 
the peak, though they create some weird stuff North of M82.  Giles mailed out a comparison between 
Mike's I-map from Feb (mailed to us last week) and Lero's new stuff.  We discussed two problems and 
possible solutions:

problem 1: the application of pointing corrections adds concentrated flux North of M82:

solution: probably a bad (FAZO, FZAO).  Plot nominal (FAZO, FZAO) vs. corrected (FAZO, FZAO) to find 
the bad one.

problem 2: Lero's map looks pixelated compared to Mike's.

discussion: They used the same flags, as you can tell from the header of Mike's files (I-map).  
(Note: these flags are not the same ones as used by Mike for his posted work on M82 pointing (Oct. 
19 2007 posting))

solution: John suggests that background subtraction that explains the difference.  Lero will 
experiment with this.

_____________________________________________________________

not discussed on skypecon: sharpcombine on mac