Vis Trap Pumping Results Product¶

Metadata¶

Data product name	DpdVisTrapPumpingResults
Data product custodian	VIS
Name of the Schema file	euc-vis-TrapPumpingResults.xsd
Processing Element(s) creating/using the data product	Trap pumping calibration pipeline.
Processing function using the data product	VIS

Proposed for inclusion in EAS/SAS¶

Yes

Data product elements¶

Header	object of type sys:genericHeader
Data	object of type vis:visTrapPumpingData
QualityFlags	object of type dqc:sqfPlaceHolder
Parameters	object of type ppr:genericKeyValueParameters

Detailed description of the data product¶

This product contains the results of the trap pumping (TP) calibration pipeline. The product has one or more DataContainer elements, each of which holds a CSV file containing results from analysis of the vertical (VTP) or multiple-readout serial (MSTP) trap pumping raw frames. These frames contain “dipoles”, which result from shuffling charge back and forth between pixel phases containing traps.

There is one CSV file per injection gate 1 (IG1) voltage and TP “shuffling mode” combination, and this configuration is indicated by the CSV filename and header lines (see below). The CSVs contain data on how many traps are present in each quadrant, their emission time constants and “pumping efficiencies”, and their locations (CCD, quadrant, pixel and phase). Each row describes one trap.

The CSVs are headed by several lines prefixed with “#”, which contain metadata describing the TP configuration, input data and processing PPO. The metadata present is:

Metadata key	Description
`tp_type`	The TP “type”: one of “`Parallel`” (VTP) or “`MultipleSerial`” (MSTP).
`ig1_voltage`	The IG1 voltage used in the input data. This affects the signal level present in the device, upon which the size of dipoles is dependent.
`shuff_mode`	The shuffling mode. This defines the starting phase and the pattern of charge transfer, which determines the phases in which traps can be found.
`input_frames`	A comma-separated list of filenames of the input frames, using the internal VIS pipeline naming of the format `EUC_VIS_EXP_XXXXXX-YY-Z`, where the Xs, Ys and Z correspond to FITS header keys `OBS_ID`, `DITHOBS` and `EXPNUM`, respectively.
`obs_dts`	A comma-separated list of date-times (UTC) of the acquisition of each frame in `input_frames`, in the format `%Y-%m-%dT%H:%M:%SZ`.
`ppo_id`	The ID of the PPO in which the file was created.

Below the final #-prefixed line is the normal CSV columns header, followed by the data.

For VTP, the CSV columns are:

Column name	Description
`ccd`	The CCD identifier, e.g. “1-1”.
`quad`	The quadrant identifier, e.g. “E”.
`row`	The pixel row within the quadrant where the trap is located. This is with the origin at readout (i.e. rotated from the FPA orientation).
`column`	The pixel column within the quadrant where the trap is located. This is with the origin at readout (i.e. rotated from the FPA orientation).
`phase`	The phase within the pixel containing the trap.
`model_name`	The name of the best-fit model class. (These are Python class names from the trap pumping analysis code.) `TphOriginalModel` is the only valid model for VTP. `ConstantModel` indicates a bad fit – a constant fitted better than the model – and is retained for debugging. In the edge case that the pixel is flagged as a cosmic in >6 of the TP frames, this column is empty (no fit is performed, since the data is considered too sparse).
`tau_e`	The emission time constant, in microseconds, if `model_name` is `TphOriginalModel`.
`P_c`	The pumping efficiency, if `model_name` is `TphOriginalModel`. Also called probability of capture.
`k`	Additive constant, should be close to zero.
`c`	Constant, only defined if `model_name` is `ConstantModel`.
(Numerical)	There are 11 columns (nominally, fewer if frames are missing) whose headers are the phase time \(t_\mathrm{oi}\), in microseconds, and whose values are the dipole intensity (half the difference between the two sides of the dipole) at that \(t_\mathrm{oi}\). These are retained for further analysis and debugging.
`cr_flags`	Bit flag indicating in which frames dipoles have been flagged as CR-affected and excluded from the fitting. The most significant bit flags the lowest \(t_\mathrm{oi}\) (nominally 8 µs) frame and the least significant bit the highest (nominally 16 ms). Retained for further analysis.

For MSTP, the columns are:

Column name	Description
`ccd`	The CCD identifier, e.g. “1-1”.
`quad`	The quadrant identifier, e.g. “E”.
`column`	The pixel column within the quadrant where the trap is located. This is with the origin at readout (i.e. rotated from the FPA orientation).
`phase`	The phase within the pixel containing the trap.
`model_name`	The name of the best-fit model class. `DwellPhase1Model` or `DwellPhase2Model` are the only valid models for MSTP. Other models indicate neither model fitted well – these rows are retained for debugging.
`tau_e`	The emission time constant, in microseconds, if `model_name` is `DwellPhase1Model` or `DwellPhase2Model`.
`tau_c`	The capture time constant, in microseconds, if `model_name` is `DwellPhase1Model`.
`P_c`	The pumping efficiency, if `model_name` is `DwellPhase2Model`.
`k`	Additive constant used in both models. Should be close to zero.
`a`, `b`, `c`	Parameters used by other models indicating a bad fit (`LinearLogModel`, `ConstantModel`). Retained for debugging.
(Numerical)	There are 18 columns (nominally) whose headers are the dwell time \(t_\mathrm{dwell}\), in microseconds, and whose values are the dipole intensity (half the difference between the two sides of the dipole) at that \(t_\mathrm{dwell}\). These are retained for further analysis and debugging.
`cr_flags`	Not used. Currently set to zero.

The product contains two QualityParameterStorage elements, which point to archived tarballs with quality control stats and plots. The stats archive contains ASCII files with “frames” stats (VIS_TrapPumping_Frames*.txt, VTP only) and “output” (VIS_TrapPumping_Output*.txt) stats.

The frames stats are, for each VTP frame and quadrant, the mean (tp_reduced_mean_col_means) and standard deviation (tp_reduced_std_col_means) of the mean column values, after CI reference subtraction and sigma clipping at 3 sigma, and for each CI reference frame, the mean (ci_mean) and standard deviation (ci_std) of the sigma-clipped image.

The output stats are, for each quadrant, the number of traps found (num_traps) and, where applicable (VTP), the number of traps where at least one dipole was flagged as being CR-affected (num_cr_affected).

The plots are histograms of tau_e (tau_e_histogram*.png) and 2D histograms of P_c vs. tau_e (tau_e_P_c_heatmap*.png), each presented in a 6x6 (FPA) grid per file, with separate files for each quadrant, and a heatmap of the number of traps per quadrant (ntraps_heatmap*.png), presented in a 12x12 grid. The tarballs are also included in an associated DpdVisAnalysisResults product.

DpdVisTrapPumpingResults is currently used only for monitoring and is not used by any downstream products.