NRAO Home  >  Green Bank  |  Wiki Topic:    GB > Obsreports > KaBandGFMCalibrationProblem

Changes | Index | Notify | Search | Go

Ka-band GFM Calibration Problem - Spring 2005

It appears that there is a problem with GFM processing Ka-band pointing/focus data. The fitted pointing/focus offsets should be good (except perhaps under poor weather conditions), so this should not have significantly affected anyone's observing. The manifestation is that the calibration and ratio of relative beam heights may appear incorrect. Under poor conditions, the problem may also have made the weather conditions (continuum baselines) look rather worse than they really were

To avoid this problem, we recommend switching to "raw" mode for data processing, after you have performed your first peak/focus in an observing session. The absolute calibration will then be arbitrary, but the relative peak heights, and baseline variations, should be more realistic. This option needs to be selected separately for peaks and focusses. The process is as follows:

Peak:

• peform a peak scan
• ensure the "pointing" GFM tab has focus (it should if you have just run a peak).
• on the top menu bar, select Tools->Options.
• In the Options pop-up menu, select "Data Processing" and "Raw"
• Your next peak scan will be processed in raw mode.
• You can confirm you are in "raw" mode as this will be displayed in the text line at the bottom of the plot window.

Focus:

• as for "Peak", except make sure that the "Focus" tab has focus.

For those interested, the problem seems to be that GFM combines incorrect samplers, and possibly Tcal values, when performing the calibration for processing options other than raw mode. In many cases, the results can look reasonable, so it is not clear whether this was working and has broken, or whether it has never worked correctly, and the test data simply fortuitously looked correct.

In fact, Brian believes the Tcal values currently in the calibration database are rather suspect anyway. Hence "raw" mode (where the gains of the feeds are assumed constant) is probably a more reasonable estimate than calibrating the data using uncertain Tcal values.

It may be some time before we have the resource available to find and fix this problem, but it certainly will be added to the GFM task list.

-- RichardPrestage - 28 Apr 2005