MHO_ManualPolDelayCorrection

Purpose

MHO_ManualPolDelayCorrection applies a user-specified polarization-specific delay offset to visibility data. The operator computes a frequency-dependent phase correction phasor from the delay (in nanoseconds) and multiplies all matching visibility elements. The correction is applied independently to each channel, accounts for sideband inversion (LSB vs USB), and distinguishes between reference and remote stations.

Control File Trigger

  • Keywords: pc_delay_x, pc_delay_y, pc_delay_r, pc_delay_l

  • Category: calibration

  • Priority: 3.5

Each keyword selects one of the four polarization states (X, Y, R, L). The keyword name determines the polarization via the suffix (_x, _y, _r, _l).

Parameters for pc_delay_<pol> keywords

Parameter

Type

Description

value

real

Delay offset in nanoseconds applied to the specified polarization.

Input Data

This operator acts on the visibility_type container.

Algorithm

MHO_ManualPolDelayCorrection has no Initialize method; all work occurs in ExecuteInPlace.

The builder (MHO_ManualPolDelayCorrectionBuilder) parses the polarization from the keyword name, retrieves the reference frequency from the parameter store (/control/config/ref_freq), and resolves station identifiers.

Execution (``ExecuteInPlace``):

  1. Iterate over the reference (index 0) and remote (index 1) stations.

  2. For each station, check applicability via IsApplicable: the operator matches if any configured station identity equals the station’s Mark4 ID (1-character match) or 2-character station code (2-character match). Wildcard ? matches any Mark4 ID, and ?? matches any station code.

  3. If applicable, retrieve the polarization-product axis and channel axis.

  4. For each polarization product:

    1. Check polarization match via PolMatch: the character at the station’s index in the pol-product label must equal the configured polarization (case-insensitive).

    2. Compute the delay in seconds: tau = delay_offset * 10^-9.

    3. Store the delay offset (in nanoseconds) as metadata on the pol-product axis under the key ref_delayoff_<pol> or rem_delayoff_<pol> (depending on station index).

    4. For each channel ch:

      1. Retrieve the channel frequency from the channel axis: nu_ch (in MHz).

      2. Compute the frequency offset from the reference frequency:

        \[\Delta f = (\nu_{\rm ch} - \nu_{\rm ref}) \cdot 10^6\]

        where \(\nu_{\rm ref}\) is the reference frequency in MHz and \(\Delta f\) is in Hz.

      3. Compute the phase angle:

        \[\theta = 2\pi \cdot \Delta f \cdot \tau\]

      4. Construct the correction phasor: Phi = exp(i * theta).

      5. Apply sideband conjugation: if the channel’s net_sideband tag equals L (lower sideband), conjugate the phasor: Phi = conj(Phi).

      6. Apply station conjugation: if the station index is 0 (reference station), conjugate the phasor: Phi = conj(Phi).

      7. Multiply the visibility sub-view at (pol-product, channel) by Phi.

Effect on Data

For each matching station and polarization, the operator applies a frequency-dependent phase rotation to every channel in the visibility data. The phase rotation encodes a linear delay offset in nanoseconds relative to the user-configured reference frequency. The correction phasor is conjugated for both lower-sideband data and reference-station data (so the net sign depends on which combination applies). The delay offset value is stored as metadata on the polarization-product axis for later inspection.