SDTconvert - coordinates in the header of CLASS spectra

[pcastang]

Coordinates (RA and Dec) in the header of CLASS spectra should be the commanded coordinates and not the observed ones. All the scans of the same source/target should have exactly the same RA and Dec.

In the header of converted CLASS spectra, instead, RA and Dec of the same source change from scan to scan. Please see the attached plots of two consecutive scans of the same source.
The differences are of the order of 0.04", however, it is enough for CLASS to interpret the spectra as coming from different sources. As a result, CLASS command to average the spectra (that performs a consistency check before doing the operation) does not work.
The only way to make this command work is to disable the consistency check. This, however, might be very dangerous, especially analysing maps.

Note that the header variables corresponding to RA and Dec are "lambda" and "beta" (in radians).
This is the error returned by the CLASS command "average", for the two spectra in the plots:

LAS90> ave
Consistency checks:
Checking Data type and regular x-axis sampling
Checking Source Name
Checking Position information
Checking Offset position
Checking Line Name
Checking Spectroscopic information
Checking Calibration information
Reference spectrum:
Source Name : MRK622
Coordinate System : EQUATORIAL 2000.0
Proj. Center (rad): lambda 2.127882, beta 0.680756, tolerance 4.8E-08
Line Name : F0-420.000-M
Frequency (MHz) : rest 22.235E+03, resol 2.563E-02, offset 0.000E+00
Velocity (km/s) : resol -3.456E-01, offset 6.964E+03
Alignment (chan) : tolerance 10.0%
Calibration : beeff 0.000,gain 0.000
W-CONSISTENCY, Obs #2 differs. Inconsistent:
W-CONSISTENCY, Lambda Proj Centers: 2.127882, 2.127881
W-CONSISTENCY, Beta Proj Centers: 0.680756, 0.680755
E-AVERAGE, Index is inconsistent