Summary of chop-nod analysis telecon of July 22, 2009

Martin, Hiroko, John, Giles, Lero

_____________________________________________________________

Hiroko's analysis of IRAS 20126:

- we discussed the third in the recent series of analysis logbook postings 
by Hiroko

- first map shown here: contours at 5.5% (chosen to include all vectors), 
10%, and 20%.

- counting the number of points in the chi2 fits files:  2601 is (51)^2 
which is the total number of points in the file ... what we need is to 
recount these, this time including only reasonable values (i.e., values 
not inflated by Mike by a huge number like 10000).  See telecon summary of 
June 17, 2009 for explanation of how chi2 handles masking.

- the two bins agree well (2.3 sigma threshold was chosen to exclude large 
vectors)

- last two maps shown are same as in previous memo

_____________________________________________________________

Lero's analysis of M82:

- we discussed Lero's two e-mails to sharp-software, sent July 20th and 
July 21st

Two main points were discussed:

1. Why is scv4 giving somewhat different results from scv5?  Which do we 
believe?

Giles points out that we have fewer large vectors with high nominal 
significance in scv4.  This could imply that the large vectors are 
incorrect.

But John argues that scv5 is the better way to do things since the ratio 
of the averages is more reliable than the average of the ratios.

Lero points out that the large vectors are still there in scv4 but with 
lower significance.

2. Do we believe the 3-sigma P~18% vectors?  Does their presence shed 
doubt on the similarly-angled 3-sigma vectors in the southeast quadrant 
which otherwise look very believable?

(a) is it synchrotron?  (Lero will check)

(b) is it just "really well aligned" dust?

(c) are the vectors wrong because the I is wrong?  Along these lines, note 
that the 18% vectors correspond to I/sigma-I < 2

Martin points out that if expanation (c) is correct then the answer to the 
second question listed under "2" above is "no".

Clearly we need to understand the long-neglected issue of how 
sharp-combine computes sigma-I and whether this is believable. As a first 
step, Martin will provide a better explanation of why he thinks that there 
may be a problem with the calculation of sigma-I in sharpcombine.

_____________________________________________________________