Frequency switching questions

[mwienen@…]

Marion Wienen via email:

Hi!

Last week, I stayed with Tatiana at the ATNF for a few days and now, we
are observing at the Parkes telescope. We are measuring the ammonia (1,1)
to (3,3) inversion transitions and want to use Class from the Gildas
software for data reduction, as there are fits available, which are well
suited for the ammonia lines, as they include e.g. the hyperfine structure
of the transitions. Do you know the software?
We are trying to use frequency switching for some of our spectral line
observations and to convert the rpf files into ascii files using the
script, which an operator gave us and I sent to you. We could not find
anything about frequency switching in the ASAP manual. One rpf file is
converted into two ascii files, do you know the difference between them?
For comparison, I sent you one image of an observation of Orion, shown by
ASAP (2009-06-21_0322_ON2_fqcsw.ps) and an image produced by Class using
the ascii file, where I can only see one peak, but no line
(Orion-fsw-class.eps). Do you know if there is an error in the conversion
or if we could try to convert into another format, different from fits, as
it cannot be read by Class?

[Malte Marquarding]

Me:

Hi,

I had a look at your data. Firstly, I recommend from previous experience to use position switching as most ripples are frequency-dependent. That is probably the reason it isn't documented ;-)
The data however shows some signal around channel 7200, which is heavily obscured by the ripple.

The two IFs i the data represent the offset frequencies. The formula used  to to frequency switching is simply.

IF0 = -1* (IF0/IF1) * TSYS0
IF1 = -1* (IF1/IF0) * TSYS1

No folding is done. With large bandwidth also note that you should have your line in the centre of the band as the axis is only approximately linear.

A good way to get a reasonable bandpass is to smooth the spectrum using a 'running median' kernel with a kernel width several times the width of the feature. Unfortunately, the ripple frequency seems to be on feature scales in your case.

I'll open a ticket to track this conversation and will attach the various files to it.

Not sure if I can help any more.

If you want to do hyperfine structure fitting in asap directly, open a ticket (http://svn.atnf.csiro.au/trac/asap/newticket) and I will see what I can do. This has always on the list of possible to-dos, but nobody has requested it so far though.

Cheers,
Malte.

[Malte Marquarding]

To export data into class use:

myscans.save("myfile.fits", "CLASS")

This should be readable by class. It is fits with a few additions for class.