doflag, n_phase, n_amp Control the identification and fate of un-correctable visibilities. EXAMPLES -------- 1. To specify that a minimum of 3 telescopes for phase, and 4 telescopes for amplitude selfcal must exist in a closed loop of a given solution bin, or the solution bin will be flagged, type: 0>selfflag true,3,4 - In phase-only self-cal, good data on baselines that are not in closed arrays of at least 3 telescopes will be flagged. - In amplitude self-cal, good data on baselines that are not in closed arrays of at least 4 telescopes will be flagged. 0> 2. To specify the same as two, but instead of flagging the bad solution bins, just don't correct them, type: 0>selfflag false,3,4 - In phase-only self-cal, good data on baselines that are not in closed arrays of at least 3 telescopes will not be used. - In amplitude self-cal, good data on baselines that are not in closed arrays of at least 4 telescopes will not be used. 0> 3. To specify that all un-flagged data shoud be used in selfcal, type: 0>selfflag false,0,0 - In phase-only self-cal, all un-flagged data will be used - In amplitude self-cal, all un-flagged data will be used 0> All but the first line of each example are responses from 'selfflag'. PARAMETERS ---------- doflag - Flag un-correctable visibilities? This controls what happens to visibilities that are found to be uncorrectable in subsequent invocations of selfcal. If this argument is false, the visibilities are left un-corrected. If true they are flagged. n_phase - The minimum number of telescopes required for a phase solution. In theory the minimum number of connected telescopes required to form a constrained phase self-cal solution is 3. See below for further discussion. n_amp - The minimum number of telescopes required for an amplitude solution. In theory the minimum number of connected telescopes required to form a constrained amplitude phase self-cal solution is 4. See below for further discussion. If no arguments are provided, the current settings are displayed. Whenever the observe command is used, all the above values are reset, to the values that would be set by typing: selfflag true ,3 , 4 CONTEXT ------- This command determines the behavior of subsequent invocations of the selfcal and gscale self-calibration commands when they encounter un-constrained visibilities. Self-calibration implicitly relies on the existence of closure phases for phase solutions, and closure amplitudes for amplitude solutions. To form a closure phase requires at least 3 connected telescopes in a closed loop, and to form a closure amplitude requires at least four. Before attempting to ascertain corrections for a given integration, the current self-calibration algorithms attempt to determine if there are sufficient un-flagged visibilities to constrain solutions for each telescope. They do this in the following manner. First they identify which telescopes lie on closed baseline paths. Only visibilities from these telescopes can be used in self-calibration. If a prior call to selfflag set dogain to true, then the rest of the un-flagged visibilities are flagged at this point, since they are totally un-correctable by self-calibration and will come to dominate the goodness of fit and prevent the fit from converging. The number of correctable telescopes is then counted and compared against n_phase if performing phase-only selfcal, or against n_amp if amplitudes are being corrected. If there are insufficient telescopes to attempt a solution and 'doflag' is set to true, then the correction is flagged, and when the corrections are applied to the data, such correction flags are propagated to all affected the visibilities. These flags can be globally removed with the 'uncalib' command, or individually applied or removed with the 'corplot' command. Since it is impossible to have fewer than 3 telescopes forming a closed array, if the p_mintel (a_mintel for phase selfcal) is set to less than or equal to 2 then the above checks will be turned off, no flagging will be done and all visibilities will be used in selfcal. Note that if n_phase and/or n_amp are lowered below their defaults (3,4) then because there are too few constraints in integrations that benefit from this, the resulting visibilities will fit too well to the model. Needless to say, this is not recommended. RELATED COMMANDS ---------------- selfcal - Self-calibrate data with the current model. selftaper - Used to down-weight short baselines during selfcal. selflims - Used to set limits to amp and phase corrections in selfcal. selfant - Set antenna based constraints in selfcal. gscale - Finds overall amplitude corrections for telescopes. startmod - Phase selfcal to a starting model then discard the model. corplot - Display and allow editing of corrections. uncalib - Remove selected parts of corrections.