Summary of chop-nod analysis telecon of Oct. 21, 2009

John, Giles, Lero

_____________________________________________________________

Giles' analysis of Serp-FIR1, L483, and L1448-IRS2:

We got hours and hours of tau = 0.06 data.

Used Sept 2009 RGM supplied Darren in "real time" on the mountain.

Used a crude pointing correction, based on linear interpolation of FAZO 
and FZAO values recorded on excel spreadsheet, rather than painstaking 
gaussfitting of each file.

These two things improved the maps a lot !  (See my recent e-mail to 
sharpsoftware list serve for current maps).

I did some internal consistency checks, though not a full-blown chi2 
analysis (see my recent e-mail to sharpsoftware list serve).

Serp is very bright and we have good, very reproduceable stuff on Serp 
(bin-maps not provided but trust me the reproduceability is good).

For L483 a two-bin analysis yields terrible results.  There was terrible 
sky noise a few hours after the bin2 data were collected.  Could it have 
started early but gone unnoticed by Nicholas and me?  There were some 
notes of electronic noise during bin1.  But could this cause the massive 
disagreement we are seeing?  In any case, we have nothing so far, though 
more investigation of these data is warranted.

For L1448 the two bins seem to agree pretty well, overall.  So I am 
cautiously optimistic that we have believable stuff.  Time will tell.

Tons of data on Jupiter...  tau was 0.1, but data is great.  John hopes 
to get to this in future to investigate possible changes in i.p. due to 
the new M3 or other factors, but warns his analysis methods are 
approximate.  John wants excel sheet posted ... this is now done, and I 
also posted Darren's 14,000 ft. rgm-file

Serp map suggests NO peak in polarized flux.  So ... we need a way to 
plot polarized flux.  John says this is easy in polsharp5 and can be 
done with about 4 lines of code and the output can be saved to a fits 
file.  Giles suggests adding a few more lines of code to discard data 
for which the noise on the polarized flux is worse than the median noise 
times some user-settable factor.  I guess this might require writing 
NaNs to the fits file; hopefully we can figure this out easily.  In any 
case this is a feasible addition to the pipeline that needs to get done.

_____________________________________________________________