Background.

It is possible to check almost all of the equipment at a VLBA site using available status information, pointing, readback tests, gps, lock tests, and the pulse cal system. Basically, if all of these tests give good results, about the only things that can still be wrong are the receiver polarization and and just possibly something in the maser. In practical experience, the polarization is the main item that has been wrong. I'm not sure of the statistics, but something like 10% of receivers end up being mounted with the polarization crossed. In the past, it was thought that the only way to check this was with real time fringe tests or with actual recorded observations. But the real time fringe tests have been very unsatisfactory for reasons that are not fully understood and anything involving recording takes days to complete.

A few months ago, I realized that we do indeed have a simple test for polarization available in data that we are already collecting. With the offset feed geometry used on both the VLA and VLBA, there is an offset between the pointing positions of the RCP and LCP beams. This offset is called the squint. It amounts to about 5% of the beam full width, half maximum (FWHM) and is easily seen in the pointing observations. In fact, one of the first steps that the pointing analysis program, PTANAL, does is to measure the squint and correct all data points to correspond to the desired pointing position half way between the beams. The squint is in the direction of the position of the feed on the feed ring. The test for polarization is to check the squint direction. If the system is cross polarized, the squint direction will be reversed.

A recent incident, where crossed polarization at MK at 7mm was not detected for over 4 months, has prompted me to first investigate the use of squint to detect polarization problems, and to implement a check in PTANAL. We are also considering changes to the sniffer to help catch crossed polarizations faster using any programs correlated with full polarization products.

It turns out that the squint can even be used at 50/90 cm to detect polarization. While the prime focus feed system does not have the full offset geometry of the other bands, the asymmetric subreflector empirically gives a systematic squint. The direction of the squint agrees well with that seen at 20 cm which is not surprising because the rotation position for 50/90 cm is the same as 20 cm to allow for multi-band observations. There is one current exception at BR where the squint is low in magnitude and deviates from the expected angle by more than 90 degrees at 50cm. Sniffer tests indicate that this system is not crossed, so I have built this exception into PTANAL. Of course, this means that a code change will be required if the situation at BR is altered.