Contributions to T$_{sys}$ From Strong Lines

Galactic HI emission will fill the beam of each antenna and therefore increase the system temperature. But, over the 50 MHz wide front-end passband, even this strong line makes a contribution of only a few percent to T$_{sys}$. However, over the spectral line channels (i.e., the much narrower back-end), T$_{line}$ can be a significant fraction of the total T$_{sys}$. In this case, the back-end T$_{sys}$ is the more appropriate one to use. There are two possible solutions. The simplest is to specify the use of the back-end T$_{sys}$ in the observe file. This involves the use of a modified SYSxROT file (where x is replaced by the band used, such as SYSLROT for L-band) which contains sub-reflector and front-end parameters. Any changes need to be entered by the VLA Operator. This should be discussed well in advance of your run with your VLA Staff Contact. Measurements of T$_{sys}$ over the front-end will still be made automatically, but not for every time step, and will require a bit of effort to access since they will now reside in the MODCOMP (on-line computer) monitor database and won't be written to the VLA archive tape which contains the data. There are two drawbacks: by using the back-end T$_{sys}$, one is committed to using a quantity with a significantly smaller S/N. Secondly, measurements of the back-end T$_{sys}$ are affected by compression and temporal instabilities of the waveguide system.

The alternative is to scale the visibilities calibrated with the front-end T$_{sys}$ (i.e., the default mode) after the fact to reflect the desired back-end T$_{sys}$. A VLA Test Memo explaining how to determine the appropriate corrections is in preparation.