Friday, February 27, 2009

Casa and high frequency data

A question from reader Adam:

Do you folks do any work with high frequency (K and Q band) data? Do you use CASA at all? I've been trying to do both.

My (Amanda) response is below. Feel free to add yours in the comments.

I've just started playing with CASA, but not with high frequency data. My impressions are that the core utilities are there and that they're making great progress debugging everything and adding functionality. That being said, be extremely careful using CASA for production data reduction and make sure you double-check everything.

I've heard from other people that there are performance issues associated with large data sets on 32-bit computers.

Read more!

Wednesday, February 25, 2009

This error paves the road to madness

Not to be melodramatic or anything.
But this error:


has started popping up all over the place ever since i updated to 31Dec08 version of aips. The relevant task always crashes after giving me this error.

It often pops up when I'm just doing an innocent UVPLT-- and sometimes if I load my data up in TVFLG and wiggle my mouse around and push a few buttons, when i quit out and try UVPLT, the error will be gone. But sometimes it won't.

And i was just trying to run BLCAL and got the error for that script too! I've checked to make sure my uv data is not being read or written. I don't understand what it means. Any thoughts?

Read more!

Tuesday, February 10, 2009

CVEL and DBCON with ATCA Data

I'm reducing some archival, multi-configuration HI data from the ATCA and am having trouble getting the velocity axis quite right. I've put the correct velocity definition in the header using CVEL and then am combining the individual data sets using DBCON with DOPOS=1,1 so it checks the position and frequency information in the header. However, when I make the image and compare to the originally published image the velocities for various features are way off (~40 to 100 km/s). Any words of wisdom from the peanut gallery on using CVEL and DBCON together?

I've mostly been following The Analysis of ATCA data with AIPS, but it's a little light on detail on this issue.

Read more!

Monday, February 9, 2009

Watch out for those FreqIDs!

If you're a continuum observer (especially if you use the single-channel, 50 MHz correlator setting at the VLA), it's important to remember when reducing line data that each different set of frequencies in line data is given its own FreqID number. Some tasks work fine with datasets with multiple FreqIDs (e.g., LISTR), but other tasks (e.g., TVFLG or SPFLG) need to be told what FREQID to use or they won't find the data. You can control how the FREQIDs are assigned in FILLM.

Read more!