Usual difficulties

**Figure:** Example of phases affected by a `baseline' problem. The separation of the phases obtained for the two phase calibrators as well as the large phase drifts are characteristics of a wrong baseline model. In this case, the position of antennas 3 and 5 were wrong in $\sim$ 5mm.

Sometimes, during the observations of a project, a good baseline model (a precise determination of the antenna positions) is not available. This introduces typical phase drifts that are often well identified by an experienced astronomer (see Figure

). Normally the AoD reports on this in the project notes and often proposes a baseline solution obtained hours or days later. Baseline solutions obtained close to each track are given in the directory baselines. Otherwise, they can be provided by the local contact. To apply a baseline model, ``set /def'' and ``find'' should be entered at first in CLIC, followed by the command ``modify antenna 1...2...3...4...5...6.../offset 99'' as proposed by the best baseline model, either found in the in the AoD notes or in the baselines directory. Note that this command refers to logical antennas, which should be well identified within the solution if the number of antennas is smaller than six. Verify with your local contact if any doubt.

**Figure:** Example of a RF calibration where we can appreciate a remaining delay in the baselines including A2. Delays can be solved with the commands ``solve delay'' and ``modify delay'' (see Sect. ).

As mentioned in Sect.

, sometimes data must be corrected for delays (see Figure

). To determine the delays we first select a polarization (with ``set polar ..'') and then plot phases (with ``set y phase'') vs. the if1 frequency (or band frequency, with ``set x if1''), preferably after selecting antenna-mode (with ``set antenna all''). Delays are determined with ``solve delay /plot /print'' for each polarization. To enter a solution, all concerned scans must be selected with ``find'' and then the command ``modify delay ...'' must be typed, as proposed by the solution. After correcting delays the RF calibration must be repeated.

Note that, for the time being, the continuum units are not corrected offline for delays. If delays are modified, we advise using spectral units (instead of continuum ones) at each step of the calibration process.

By default the atmospheric phase correction derived from the 22GHz receivers is applied according to the results obtained with the PhCor procedure. We can however remove it by typing ``let do_atm no'' in the line of commands before performing any calibration. If a previous calibration was already carried out applying the atmospheric phase correction, the option no must be also selected in Use previous settings. The whole calibration must be then repeated. On the upper-left corner of each data plot we can see Ph: Abs./Rel(A). Atm, where Atm reports on the presence of the atmospheric phase correction (according to PhCor).

If a track includes observations with different telescope configurations, for example if an antenna was removed during the observations, after clicking on Select the different configurations will appear in the main calibration widget. The calibration must be perform independently for each configuration. Use previous settings is then set off.