dofloat
Self calibrate to determine overall telescope amplitude corrections.
EXAMPLES
-------

0>gscale
Computing UV representation of latest clean model (147 components)
The established model now contains 147 components and 7.0511 Jy
Performing overall amplitude self-cal

Correcting IF 1.
 Amplitude normalization factor in sub-array 1: 1.01723
 Telescope amplitude corrections in sub-array 1:
  BONN      0.97     WSRT      0.95     JBNK1     0.99     BOLOGNA   1.01     
  TORUN     1.00*    HART      1.01     HSTK      1.01     NRAO      1.02     
  VLBA_KP   0.96     VLBA_PT   1.00*    VLBA_LA   0.96     VLA27     0.98     
  OVRO      0.96     
 Fit before self-cal, rms=0.498946Jy  sigma=3.396564
 Fit after  self-cal, rms=0.438911Jy  sigma=3.160276
 
Reinstating the default IF.
0>

All but the first line are responses from the 'gscal' command.
Some of the reported corrections are post-fixed with a '*' symbol.
This means that no solution was possible for that telescope.
These corrections are not 1.0, only because of amplitude
normalization.

Without amplitude normalization (dofloat=true) the same data-set gave:

0>gscale true
Performing overall amplitude self-cal

Correcting IF 1.
 Telescope amplitude corrections:
  BONN      0.97     WSRT      0.95     JBNK1     0.98     BOLOGNA   1.01     
  TORUN     1.00*    HART      1.02     HSTK      1.02     NRAO      1.01     
  VLBA_KP   0.96     VLBA_PT   1.00*    VLBA_LA   0.96     VLA27     0.98     
  OVRO      0.96     
 Fit before self-cal, rms=0.500856Jy  sigma=3.412605
 Fit after  self-cal, rms=0.365222Jy  sigma=2.861346
 
Reinstating the default IF.
0>


PARAMETER
---------
dofloat - If this parameter is omitted or given as false, the
          telescope corrections are normalized to maintain the current
          total flux. Alternatively, if dofloat=true then the
          telescope amplitudes will be allowed to float and the
          observed flux may be changed if not sufficiently
          constrained.

CONTEXT
-------
When one or more telescopes are miscalibrated by a constant factor,
this command may be used to solve for the correct telescope
amplitudes. In order to do this a reasonable model of the source
structure must have been provided. Such a model may be provided using
CLEAN or using a model from the VLBI modelfit program read into this
program via the 'rmodel' command.

This command should be used with caution. For instance, comparison of
the current model, in 'vplot', and the observed visibilities may
appear to indicate a number of consistently high baselines. This often
indicates data that hasn't been cleaned deeply enough. In particular,
check to see if just the shortest baselines are too high by examining
the plot of visibility amplitude verus UV radius given by the radplot
command. Such missing flux implies the existence of un-cleaned low
surface brightness extended emission that hasn't been cleaned yet. Try
switching to natural weighting (see help uvweight) and/or apply a
UV-taper before cleaning (see help uvtaper). This may reveal such
emission, which may then be cleaned.

If it turns out that artifacts due to poor amplitude calibration,
prevent you from cleaning low surface-brightness extended emission
then you should use the selftaper command to weight down the
significance of the short baseline visibilities before calling gscale
(see help selftaper). Failure to do this will make gscale attempt to
force the visibility data to fit the incomplete model at short UV
radii.

Amplitude self-calibration relies implicitly on the existence of
closure amplitudes. To form such a closure amplitude requires that
there be un-flagged visibilities on that lie baselines in closed
arrays of at least 4 telescopes. Gscale attempts to detect
visibilities that do not satisfy these conditions and exclude them
from the fit. How it does this, and whether these visibilities are
then flagged, is controlled by the selfflag command. See help selfflag
for more details.

NB. Observations that contain multiple sub-arrays and/or multiple IFs
    will produce nif * nsub lists of telescope corrections, instead of
    the single list given the examples above.

RELATED COMMANDS
----------------
selfflag - Used to control the fate of un-correctable data.
selfcal  - General self-calibration command.
selftaper - Used to down-weight short baselines during selfcal.
selflims  - Used to set limits to amp and phase corrections in selfcal.
startmod  - Phase selfcal to a starting model then discard the model.