Calibration

  Several options are available. The recommended option for most accurate calibration and best baselines is BE DATA continuous (see below). This is also known as channel-by-channel, as it uses a separate for every channel in the spectrometer backend.

1. The old methods were FE and IF, where the frontend receiver or IF power detectors, respectively, are used to give a single , etc for each receiver (recently modified to give two of each number, one each for the signal and image sidebands). This can still be used.
2. One new option is BE POWER ( BE = backend). This produces (and uses) a separate , etc for each subband.
3. The other new option is BE DATA. This gives separate values for each channel of the spectrum. Note that in the case of non continuous calibration, this should not be done on source (unless the line is very weak, ie less than 1K), as it would result in a system temperature which jumps at the line position. For normal continous calibration, the CAL can be done on-source, as the `SKY' part of the calibration is not used. It prints on the operators console an average value of the , for each subband. Note that each subband has a number for (the signal sideband) and for (the image sideband), using the correct ATM model. The value of may be much larger than , particularly in the 450--500GHz band, where the atmospheric transmission is highly frequency dependent. However, is the only number that effects the signal:noise ratio of a line in the signal sideband (ie don't worry too much if is very large).

It is possible to check and display the receiver, sky or system temperatures across the passband. Load a configuration with NON-continuous CAL, then type:

ICL> send das set das_output ***,

where *** corresponds to TREC, TSKY or TSYS. Then do a short sample (say 10 seconds in NSW mode). The output is displayed and stored in a GSD file like a normal spectrum, as in the example shown in Figure 3. Type

ICL> send das set das_output SPECTRUM

or reload the configuration again to set the output back to regular observing.

The effectiveness of the BE DATA calibration is illustrated in Figure 4. In the 760-MHz mode the bandwidth covered by the DAS is wider than the instantaneous bandwidth of the B3i receiver. This results in a strong increase in the system temperature over part of the band. However, as shown by a typical spectrum plotted on the same horizontal scale the only effect is the expected noise increase over the range of enhanced system temperature.

• Continuous Calibration