ATCA primary beam width calculation

[and460]

Hi guys,

Just a quick question - is there an empirical equation that accurately describes the ATCA primary beam radius at full-width-half-power as a function of freq? I assume it would be something like k.lambda / d, where lambda and d are the obviously the observing wavelength and dish diameter respectively, but k is some value empirically derived? Or do you guys just use some standard value for k for a parabolic dish when doing the calculation? I basically just want to be able to calculate the freq-dependent beam radius for my CABB obs.

Cheers,
Craig.

[ste616]

Hi Craig,

When dealing with a presumed Gaussian beam, we often talk about the normalised distance x, which has units of arcmin GHz. With this unit, a given fractional level of the Gaussian will always lie at roughly the same x, independent of frequency.

For the ATCA 16cm beam, which is very well described as a Gaussian out to an x of about 35 (to the ~20% level), the HWHM x = 24.7 +- 0.6 arcmin GHz.

If you do the maths, this x(HWHM) = k * 23.4 (for a 22m dish). So here, k should be 1.06, and the uncertainty is about 2%.

Speaking personally though, I usually just use k = 1.

[and460]

Great - thanks!

[len067]

Hi all,

I seem to recall looking into the Miriad code at one stage to work out what the effect of applying a single primary beam correction on a wide band data set might be. I believe the code used a multi term polynomial fit to the beam as described in the technical memo here (you need to find the right one depending on which band you're interested in):

http://www.atnf.csiro.au/observers/memos/

You can also find a good description in the AIPS documentation of the PBCOR task:

http://www.aips.nrao.edu/cgi-bin/ZXHLP2.PL?PBCOR

I tried plotting up the full range of beam widths across the entire CABB band (I think this was used in the Wilson et al. CABB paper) and noticed that there were slight steps when shifting from one band to another when extending the polynomial fits across a wider band - so they're not perfect (they were originally defined over smaller bands in the original memos).

Cheers,

Emil.

[and460]

Thanks Emil!