SOLVE AMPLITUDE

        CLIC\SOLVE AMPLITUDE  [/PLOT] [/BREAK kind time [kind time ...]
        [/POLYNOMIAL [degree]] [/WEIGHT]

    *** THIS HELP HAS NOT BEEN UPDATED FOR THE NGRX ***

    This command fits a mathematical function into the measured amplitude of
    the  sources  in  the current index (presumably calibrators).  This data
    must have been selected and plotted in axes : SET X TIME and SET  Y  AM-
    PLITUDE for the baselines and bands of interest, specified by the corre-
    sponding SET commands. The calibration function is kept in memory.  Com-
    mand  STORE  AMPLITUDE should be used next to store this function in the
    header of source observations, after a change in the index to select the
    appropriate scans.

    SOLVE  AMPLITUDE  internally  and  temporarily  resets  SET AMPLITUDE to
    SCALED.

    In antenna mode (selected by SET AMPLITUDE ANTENNA), the averaged  phase
    and  amplitude  closures  are computed, as well as their standard devia-
    tions. The phase closures should be close to zero, while  the  amplitude
    closures  should  be  close to 100%. Strong deviations of amplitude clo-
    sures from 100% are an indication of amplitude loss on  long  baselines,
    due to phase decorrelation during the time averaging. The fit then shows
    strong systematic errors; if this occurs, baseline based calibration  of
    the amplitudes might be preferred.

    Option /PLOT will plot the fitted curve over the data.

    Fitted curves may be of two kinds:

      -  Cubic splines (the default).  By default knots are regularly spaced
        with an interval between knots set by the SET STEP  command.   Addi-
        tional  knots  may  be  introduced with the option "/BREAK kind time
        [kind time ...]"  which  introduces  a  break  at  abscissa  'time';
        'kind'  is an integer in the range 0-3; 0 means that a discontinuity
        will be present, 1 that the first derivative will be  discontinuous,
        and  so  on.  Several breaks may be introduced. The program will de-
        tect an error if too many breaks are  introduced,  compared  to  the
        density of data points.

      -  Polynomial  curves  may  be  used instead.  For this the option is:
        /POLYNOMIAL [degree] indicating the degree of  the  polynomial  (de-
        fault 0).

    Normally  the  data points are all assigned the same weight for the fit.
    With option /WEIGHT, the data points are weighted according to their er-
    rors.