[daip] Re: A short and basic question

[Eric Greisen]

papadop at phys.ethz.ch writes:
 > I am now imaging the K-band CO 1-0 data on 4C41, and I  estimate
 > a theoretical noise of ~(20-30) microJy/beam for Nvis=1621683, Nif=2, 
 > Df=45 MHz.
 > There was a range of Tsys over the 4 days of observations, mostly around 
 > 70 K
 > but with several excusrions to 100 K. I used a Tsys=80 K as typical and 
 > an aperture
 > efficiency of 0.4 to get the number quoted above.
 > 
 > Using IMEAN on a CLEANed image I get rms=50 microJy/beam, it is  a bit
 > high and I was wondering that I may have used the wrong Tint per visibility
 > record. I used 10 secs but I am not sure and I was wondering how could I 
 > find out.
 > 
 > If I did the theoretical estimate would you consider the above difference
 >  serious of a discrepancy for this band (where elevation and humidity 
 > effects are
 > much more pronounced),  mind you I have done very  good data editing and
 > there are no apparent  artifacts in the images that could  act to raise 
 > the rms.
 > 
 > PS: If I use a shorter DT_vis=3 secs I get a theoretical rms=(35-52) 
 > microJy/beam,
 > in better agreement with the images estimates.

IMAGR itself raises the noise by weighting the data - it puts out an
estimated factor for that increase.  Depending on the imaging
parameters - taper, robust, etc. this factor can range from 1.0 to 1.5
or more.  That plus the atmospherics you mention may well be the sole
source of the discrepancy.

Cheers,

Eric Greisen