Yesterday I finished another 8 hour observing session at the Giant Meterwave Radio Telescope. I wrote about observing here about a week ago, but I wanted to add a bit more information.

Naturally, the control room is the same as before. But I neglected to show what I actually look at when I am “monitoring” the observations. An important part of observing is making sure you are actually getting useful information, and not wasting the telescope time. When optical observing this usually means doing simple calibrations of the images (or spectra) you are taking and comparing it to your expectations based on the rough capabilities of the telescope. (It’s almost important to make sure your telescope is pointed at the right object!). When observing with a single-dish radio telescope you are combining the spectra to look at the amount of energy emitted by an object at a given wavelength. Again, you are checking the quality of the data to be sure the systems are functioning normally.

Arrays of radio dishes combined into a single array are monitored a bit differently. Instead of doing a quick calibration of the data and checking the result (as done frequently in many optical and single-dish radio observations), the raw data products are easier to monitor. This is due to the fact that making images from an array is computationally intensive and somewhat time consuming. The time consuming portion can be alleviated by running the data through a pipeline, but it can still take some time to arrive at a “raw” data product suitable for monitoring the observation. In other words, it can take so long to get the “first look” data that it is too late to fix your observation.

These arrays work by comparing the signal recorded from pairs of dishes. So, you can monitor the health of the antennas by looking at these pairs of antennas. Below is a screen capture of a plot I viewed to monitor the status of the observation. The plot shows the amplitude (strength) of the signal against time (IST is Indian Standard Time). This amplitude is measured by comparing the signals received between two telescopes.

The key to each plot tells you what data you are looking at. For example: “W03 – USB 130:C09″ denotes the amplitude measured by the antenna pair W03 and C09 (an antenna on the Western arm of the array and an antenna in the core of the array). For this plot, I am only showing amplitudes on baselines with C09. Can you tell which antennas aren’t working properly? (Hint: they are reporting very low amplitudes.)

Antennas S04 and C01 both appear to have issues, based on the low amplitude of the signals.

There is also a range of amplitudes in the signals from the other antennas paired with C09. The highest amplitude values (at the beginning) are observations of the flux calibrator (for descriptions of the calibrators, see this post). The next highest values are observations of the phase calibrator. Finally, the values that are nearly zero at this scale are observations of the science target. The amplitudes are lower because the science target is over 100x fainter than the flux calibrator.

This is of course only one way to monitor observations. You can also look at the phase, because light is a wave. The phase contains information about the structure of the source, but rapid variations of phase with time are signs of a problem.

These plots are updated as more data is taken, so monitoring these throughout the observing run can allow you to detect some potential problems with the observation. Naturally, you can’t catch everything. But, the big problems will usually be visible. Once the observation is concluded, the more subtle issues can usually be resolved.