[evlatests] test observations with IF pairs > 4 GHz apart

[Gustaaf Van Moorsel]

During software time I tested tuning the IF pairs more than 4 GHz apart.
During 30 minutes I observed the following scans in spectral line, BW
code 1, 25 MHz bandwidth, 1.56 MHz channel width.  The schedule was
created using the OPT.

1 - dummy scan at X-band, continuum
2 - reference pointing determination, X-band, continuum
3 - K-band, AC at 20.5 GHz, BD at 25.5 GHz
4 - Ka-band, AC at 36.0 GHz, BD at 30.0 GHz
5 - Ka-band, AC at 38.0 GHz, BD at 28.0 GHz
6 - Q-band, AC at 49.0 GHz, BD at 42.0 GHz
7 - Q-band, AC at 42.0 GHz, BD at 49.0 GHz
8 - Ka-band, AC at 30.0 GHz, BD at 36.0 GHz

The observations were successful: scans 1 - 6 were observed, whereas scan
7 and 8 were not; not surprising since they (on purpose) violated the req-
uirement that the AC frequency has to be greater than the BD frequency.
The OPT should probably check for those constraints.

Notes:

1 - In scan 3 (K-band) the system accepts freq(AC) < freq(BD) but AIPS
    labels the higher of the two frequencies as IF 1, an the lower as IF
    2.  How do we make sure the right data is attributed to the right IF?

2 - AIPS labels the IF pairs with a slightly different frequency than req-
    uested in the OPT.  For instance, in scan 4, AC requested is 36.0 GHZ,
    AIPS labels it as 36.00078125 GHz, a difference of half a channel.  BD
    requested is 30.0 GHz, AIPS labels it as 30.02421875 MHz, a difference
    of 15.5 channels.  This needs to be sorted out.

Gustaaf