bandpass 12h limit

[cat001]

Hi Jamie,

You told me last week that the bandpass needs to be taken within 12 hours of the gain and flux calibrator, otherwise MIRIAD would not apply the bandpass correction.
I was re-reducing data with the new script, and I noticed that after gpcopy of the bandpass to the flux cal (planet), the flux cal points were all flagged. After doing some investigating, I realized that the bandpass was taken 13 hours before fluxcal.
This data had been reduced with no problems before. When was the 12 hours limit implemented? Is there a way around it?

Many thanks,
Catarina

[ste616]

Hi Catarina,

I'm afraid I may have mislead you here. I didn't mean to say that the bandpass is not valid for 12 hours, but rather it is the gain solution that can only be extrapolated for up to 12 hours before Miriad decides the answer it would come up with is worse than useless.

And in a sense, your data here is not really flagged, but rather uncorrectable by Miriad given the information in the dataset. You can check this in a number of ways. In most tasks, you can use "options=nocal,nopol" to prevent the task from applying gain and leakage corrections before using the visibilities. If you were to use these options, I think you'll find that the data appears as expected. You could also delete the solutions using the task delhd. Or you can specify "options=nocal,nopol" during gpcopy from the bandpass calibrator.

The idea here though is that you should be copying a gain solution from a source nearer to your planet than the bandpass calibrator that was observed 13 hours beforehand. The bandpass is unlikely ever to change by very much (nor should the leakages really), but gains change very rapidly especially at mm wavelengths. And although you don't actually need to phase calibrate the planet observations to use them for flux calibration, you need to have the gain scaling set to that of the source you want to bootstrap to the planet.

I hope this makes sense!

[cat001]

Hi Jamie,

That sorta makes sense. Though when I used uvplt to see the data, there was only flagged data.
thanks,

Catarina

[MaximVoronkov]

Although it is probably not very wise to apply >12h old gain solution, if you really need this, you can also use puthd task to create a variable "tolerance", which would allow to
apply gain solutions created at any time in the past. This option is often used in conjunction with the self calibration task which sets the gain validity interval to solution interval (i.e. typically 1 min or less).