CA01 bandpass ripple

Is the ATCA misbehaving? Let us know!

Moderator: Mark.Wieringa

Post Reply
ste616
Site Admin
Posts: 220
Joined: Thu Feb 04, 2010 3:27 pm
Location: Paul Wild Observatory Narrabri NSW

CA01 bandpass ripple

Post by ste616 »

For a long while I have noticed that bandpass calibration did not work very well for CA01 at 5.5 GHz, polarisation B. There would always be a large ripple in roughly the middle of the band, but it would never be seen in the bandpass calibrator itself (which is usually 1934-638). The shape of this ripple was not always the same, and would vary depend on the source that the bandpass solution was applied to.

An illustration of this effect is as follows. If one plots a spectrum of the bandpass calibrator on a baseline to antenna 1, you might see something like this:

Image

This spectrum looks normal and quite reasonable. When the bandpass calibration is transferred to a strong calibrator, and this same baseline is plotted as a spectrum, you might see something like this:

Image

The ripple is very obvious in Y polarisation. We know this is a problem with CA01 because it appears on all baselines to it.

I always thought it was a problem with my reduction, and considered it annoying, but not terribly important as the spectral shape can still be seen on all the rest of the baselines, and this is sufficient to measure the spectral index. But I suspect that imaging these data would cause issues, since the phase is also affected (and gpcal will be unable to fix it without using a sufficient number of nfbins):

Image

Also, the effect shows up when considering the amplitude as a function of uv-distance, which would suggest some structure that isn't actually there:

Image

I eventually decided that this was a real problem with the hardware, and on 18-Sep-2013 we decided to investigate this more directly and try to fix the problem. With the assistance of Jock McFee and Christoph Brem, we tracked the problem down to a connector on a cable between the C34 module and the CABB rack. This cable was fixed by Jock and the ripple has now gone away:

Image

Image

Bandpass ripples of this type should now no longer be visible from CA01, but I haven't yet figured out the best way of dealing with data that is affected by this problem. I have seen this ripple in data going back as far as July 2012, and may be there even earlier. Ripples of this type have also occasionally been seen on other antenna. This thread will be updated as techniques are found to properly reduce data affected by this ripple. We encourage users to post here if they see this problem in their data.
cheers
Jamie Stevens
ATCA Senior System Scientist
raynorris
Posts: 1
Joined: Thu Feb 11, 2010 5:35 am

Re: CA01 bandpass ripple

Post by raynorris »

Well done for pinning it down and fixing it!
ryd010
Posts: 4
Joined: Tue May 31, 2011 6:01 pm

Re: CA01 bandpass ripple

Post by ryd010 »

Hi Jamie,

In reducing data from C184 from 8 June 2013, when CA01 was unavailable, I have encountered a similar ripple in all baselines to CA05:
ripple.png
ripple.png (9.63 KiB) Viewed 18005 times
Any help on how to work around this would be appreciated. Without CA01 I'm missing 1/3 of my baselines already, and even though it's a point source I don't particularly want to lose almost another 1/3.


Cheers,

Stuart Ryder
Post Reply