In quasi-real time:

• Task RDI solves the atmospheric calibration observations (with or without cold load measurement), to obtain the correction factors $F$ in $\Delta \phi(t) = F \left( P(t) - \mbox{$P_{\rm REF}(t)$}\right)$. There is one such coefficient for each side band of each receiver in each antenna. The standard atmospheric model is used, but the data is interpolated out of a pre-computed table to save computing time (the table name is the logical name GAG_ATMOSPHERE which should translate to COMP$DISK:[CONTROL.OBS]ATM-BURE-VAX.BIN). The new parameters are stored in the data files in the new header section named ``Atmospheric monitor''. This is currently implemented.

  At the same time a default is set for the quantity $P_{\rm REF}(t)$ discussed above: it is the measured atmosphere emission at the time of the last calibration. This will have to be refined by the command MONITOR.

• The correlators will soon do the correction on the spectral line data, using the scan-averaged atmospheric emission as a reference. For safety reasons, spectral data will have to be averaged both with and without phase correction. More powerful CPU's will be needed for the correlators (which are being installed and tested).
• The CLIC data format had to be extended to allow storing both phase-corrected and uncorrected data. So far the data section of each observation in a CLIC file would contain one record of continuum data for each second (or longer time records after data compression by COMPRESS), and a single average record containing scan-averaged continuum and spectral line data. In the new format there will potentially be two such average records, one with uncorrected, the other with phase-corrected data.

  In fact three observing modes are possible, selected by a new OBS command:

  • SET PHASE ON, to record only phase-corrected data.
  • SET PHASE OFF, to record only uncorrected data.
  • SET PHASE BOTH, to record both uncorrected and phase-corrected data.