DSB versus Separate Sideband Calibration

CLIC can store independant calibration curves (phase and antenna) for the LSB and USB, but can also work in a mode where a single curve is stored using command

STORE PHASE or AMPLITUDE /BAND DSB or USB or LSB

Each case may have its own advantage and drawbacks:

• Separate Sideband Calibration
  This is the default mode, but it does not preserve the relative phase (and/or amplitude) of the two sidebands. On the other hand, it allows to correct for time variations of this receiver gain ratio and relative phases. Note that atmospheric phases are not chromatic (to first order), so that these relative phases are purely instrumental.
• DSB mode
  This works only when the above gain ratio and relative phases are constant in time. In case the observations were done with double sideband tuning, this mode is more sensitive. This may be an important advantage for example for CO J=1-0 observations, where the system noise is high and calibrators are weak.
• LSB or USB mode
  When nearly single-sideband tuning was used (e.g. LSB where high rejection is possible), the noise in the image sideband may be too high to allow proper calibration. DSB calibration is not optimum either, since it is more noisy than the LSB calibration. Applying the LSB calibration is then the best way to recover the USB band (should it contain anything useful...).

  USB mode is also more effective than DSB or separate sideband calibration for the 3mm receivers above 110 GHz, where the sideband rejection is 5-6 dB.

In practice, we recommend to check whether the relative phases of the two sidebands significantly differ. If no, DSB or LSB calibration depending on receiver tuning is best. If yes, this also implies that the RF calibration is time dependent: more thorough work is required... Be sure the variation is really significant: a small bias is better than a larger noise...