Wiggles in Stokes I spectrum at 16 cm

Got a calibration problem? Discuss it here.

Moderator: Mark.Wieringa

Post Reply
shane
Posts: 10
Joined: Mon Oct 25, 2010 8:54 am

Wiggles in Stokes I spectrum at 16 cm

Post by shane »

Hi,

After calibrating my data at 16 cm I get wiggles in the total intensity spectrum of the secondary calibrator that suggest I've not done a good job (see attached image, top left panel, where each point is the average value from a 128 MHz chunk of data).
I suspect it may be due to the non-standard (compared to the updated Miriad user guide) calibration strategy of having to split the 16 cm band into 128 MHz chunks in order to properly account for frequency-dependent leakages. The wiggles are not due to 128 MHz chunks with small amounts of data (eg. due to rfi).

My calibration strategy is:
1. mfcal on 1934-638 (full-band) for bandpass solution
2. gpcopy bandpass solution to my secondary
3. split up primary & secondary data into 128 MHz chunks using uvaver (hence applying the bandpass)
4. For each 128 MHz chunk:
run gpcal on 1934 (primary)
gpcopy to secondary
run gpcal on secondary
gpboot secondary gains using 1934
mfboot to adjust secondary for any spectral slope (but no bandpass table present! is it doing anything??)

Does anyone have any advice on an improved approach to the calibration?

Cheers,
Shane
Attachments
IQUV_secondary_cal.png
IQUV_secondary_cal.png (24.13 KiB) Viewed 80808 times
ste616
Site Admin
Posts: 220
Joined: Thu Feb 04, 2010 3:27 pm
Location: Paul Wild Observatory Narrabri NSW

Re: Wiggles in Stokes I spectrum at 16 cm

Post by ste616 »

Hi Shane,

Have you tried doing the whole sequence on the chunks? That is, have you tried doing the uvaver split before doing mfcal, and then proceeding from there? I'd be interested in seeing if this helps at all.

I have other suspicions about what might be going wrong here, but I'll save them until you've tried that approach.
cheers
Jamie Stevens
ATCA Senior System Scientist
shane
Posts: 10
Joined: Mon Oct 25, 2010 8:54 am

Re: Wiggles in Stokes I spectrum at 16 cm

Post by shane »

Hi Jamie,

Thanks for the fast response.
Yes, initially I was splitting up the data first and then running mfcal & gpcal on every chunk but doing mfcal before splitting the data actually reduced the wiggles.
Attached is the Stokes I spectrum for the secondary when I did the whole sequence on the individual chunks.

Looking forward to hearing about your other suspicions!

Cheers,
Shane
Attachments
StokesI_LPcal_wiggles.png
StokesI_LPcal_wiggles.png (8.32 KiB) Viewed 80802 times
ste616
Site Admin
Posts: 220
Joined: Thu Feb 04, 2010 3:27 pm
Location: Paul Wild Observatory Narrabri NSW

Re: Wiggles in Stokes I spectrum at 16 cm

Post by ste616 »

Hi Shane,

The other suspicion I have is with how gpcal works. We've shown recently that gpcal will get the flux scaling wrong, even when on 1934-638, if there are flagged data in the band, and the scaling will be more wrong with more flagging.

Since these are 16cm data you're dealing with, there's going to be a lot of flagging, and there will be different amounts of flagging in each chunk. I can easily see how this might lead to a wiggle in the data like the one you're seeing. Technically, mfboot should be correcting this, but to what accuracy we haven't really established.

So that's my suspicion. I don't however have an idea on how to correct this behaviour yet, but I'll think about it. To test whether this is likely to be the problem, try flagging some good data in various chunks, and see if the wiggle changes.
cheers
Jamie Stevens
ATCA Senior System Scientist
shane
Posts: 10
Joined: Mon Oct 25, 2010 8:54 am

Re: Wiggles in Stokes I spectrum at 16 cm

Post by shane »

Hi Jamie,

I flagged 50 MHz of good data in several different chunks and the wiggle did change a little but it seems to be a second order effect (although still worrying!). I'll keep trying different approaches and let you know if I have any success.

Cheers,
Shane
Post Reply