The ATCA Forum

Hi,

I've been using uvaver (with the line command) to split up my 16 cm data into 128 MHz chunks to do the polarization calibration. I recombine the chunks after calibration with uvaver but the header file it creates is from one 128 MHz chunk (ie. 128 channels instead of 2049). Is there a way in Miriad to recombine the data with the correct header information applied?

Cheers,
Shane

Hi Shane,

Firstly, you could just use the file you get after using uvaver or uvcat to combine the sections of the spectrum, all the data is in there, and correctly labelled. The data is just divided up into spectral windows. What you see with prthd in the header is just the header for the first spectral window. uvindex will show the full thing.

Secondly, I just found a task called uvglue does what you want. But it is black belt usage only: it doesn't check you have put the inputs in the right order, so you can thoroughly scramble your spectrum if you're not careful.

It requires the input files to be named like uvfile_1, uvfile_2, uvfile_3, etc., and all to have exactly the same spectral shape and time range.

uvglue vis=uvfile nfiles=3 out=uvfiles.glued

will then concatenate the spectra into one again.
Note if you use uvsplit with maxwidth the last output file will have an extra channel (because the continuum has 2049 channels) so you may need to use uvaver with line=chan specification instead to make this work.

Cheers,

Mark

Hi Mark,

Thanks for your help.
I used uvaver to split up the data into exactly 128 MHz chunks, did the calibration, then used uvaver again to apply the calibration solutions (because uvglue doesn't) and finally uvglue to join the files with the correct header information.
Everything is looking much better so far.

Cheers,
Shane