VLA Flux Density History Database Description


This document attempts to describe the features, accuracy and limitations of the VLA flux density history database. This is a new flux history database for the VLA calibrators going back to December 5, 1996 and containing entries for over 1000 sources. Entries are generated in an automated fashion (in ANTSOL) by the on-line system every time the VLA observes a calibrator. The data is sent down by operations sometime after IAT midnight each day and is automatically incorporated into the database at 12:00 UT. This database is accessible in a simple graphical user interface via the Web at http://www.vla.nrao.edu/astro/calib/flux. Note that a Java enabled browser is necessary.

This utility is still under construction. Users may experience delays in retrieval, outright failures, and errant data points. The phrase 'caveat emptor' certainly applies.


1) Enter the source name of interest as HHMM+DDD in either B1950 or J2000 coordinates. If a 3C name exists for the source it will also be recognized. See the VLA calibrator manual for the complete source list. Click on the "Plot It!" box to invoke the plot,


2) Select the desired wavelength. This will automatically invoke the plot. The wavelength bands correspond to the following frequencies: 90cm (P) = 0.302 -> 0.35 GHz; 20cm (L) = 1.36 -> 1.54 GHz; 6cm (C) = 4.81 -> 4.91 GHz; 3.6cm (X) = 8.41 -> 8.51 GHz; 2cm (U) = 14.89 -> 14.99 GHz; 1cm (K) = 22.41 -> 22.51 GHz; 0.7cm (Q) = 43.29 -> 43.39 GHz

The default mode of operation is to plot data from the last 60 days. The number of points plotted is described at the bottom of the plot. The Start and Stop times for the plot will be shown in the boxes marked. These fields can be editted to change the time range that the user would like to search. Use the windowing features described below to zoom the plot.

Flux densities are plotted for the AC and BD IF pairs as available. Most VLA observations record simultaneously at two independently tunable frequencies.


Very little editting of the data is possible on-line. In the future we may be able to do better, but the presence of some spurious measurements is probably unavoidable, especially at the higher frequencies. During winter storms, snow or ice in the dishes can also cause flux measurements to dip down at 5 GHz and above (e.g. look at 3C84 in Dec 1997).

Accuracy of the measurements at X and C bands is better than 10%. At U and L bands it is generally within 10% although spurious measurements are more common. At P band measurements have much larger errors due to the higher system temperatures and more narrow bandwidths employed. RFI may also be problematic at P and L bands. At K and Q bands the atmosphere and antenna gain as a function of elevation introduce considerable scatter into the measurements. To reduce this scatter somewhat, only observations taken between 30 and 80 degrees are entered into the database. The accuracy here is still probably not much better than 50% but could be improved. Pointing errors at high frequencies may also introduce some poor measurements into the database.

No uv-restrictions are currently employed in the ANTSOL solutions so resolution effects _will_ be present. These effects and the effects of confusion will become more significant as we move into the smaller configurations.


Both B1950 and J2000 calibrator names (the same as are to be found in the VLA calibrator manual and OBSERVE) are recognized, as are common aliases like 3C48, 3C286, etc.

Plots can be "zoomed" to reveal a time/flux range of interest.

Above 10 GHz observations below 30 degrees and above 80 degrees in elevation are discarded.

A table of raw measurements, frow which each plot is derived, can be examined in the Java console. And since the data is contained in an AIPS++ table, various other applications can be fairly readily built as the need arises.

Limitations in coverage:

The coverage of calibrators is necessarily rather spotty since it is entirely at the whim of the user community. Your favorite calibrator may have been observed many times or not at all. In general the good class "P" calibrators get observed at least once every 30 days.

Known Bugs:

Sometimes the plotting window does not refresh before the next plot causing data points and labels to be overlaid. This can be fixed by clicking "Plot It!" (slow) or by zooming the entire plot (fast).


Wes Young built the software for this system over the past couple months. Thanks to Ken Sowinski for modifications to the on-line system without which we were stuck. Also Gustaaf Van Moorsel, Tim Cornwell, and Miller Goss played key roles in motivating this project.

Comments? Send them to mclausse@nrao.edu

Last Modified on 2006 Apr. 13

Return to the VLA Calibrator page , or to the Calibrator Flux History Database