CALBLOCK-PV-011: NISP IPC Verification¶
Observation configuration file¶
Scheduling¶
Once, during PV phase. Can be repeated if there is significant evolution over time.
In case of failure, repeat at next possible opportunity.
Time-critical¶
No
Dependencies¶
None
Objective¶
Obtain measurements to verify the Interpixel Capacitance (IPC) matrix obtained on ground, using the single pixel reset (SPR) method.
Observations¶
The SPR method is a specific application of the Euclid Firmware, Exposure Configuration Keyword 7 (Calibration IPC). With this method, pixels can be reset in a grid to measure the IPC matrix. In Calibration IPC mode, the first ramp is averaged, and the second ramp averaged, and a difference image created and sent to the ground. This can only be done for one detector at a time per DPU.
On the ground it was performed in this way (Remi Barbier): 3 values of Vreset for ramp 2 (standard reset at 0.35V for ramp 1). 16 frames to generate the first group - the first group will be the zero point in flight - ktC noise and baseline will be computed with G1 in flight due to coadd in the FPGA. On ground, no significant different was found for different Vreset values, hence here we testonly one reset voltage.
A grid of SPR 8x8:
ipc_vresets = 0.625
x_y = [(0,0), (0,1), ... , (0,7),
(1,0), (1,1), ... , (1,7),
...
(7,0), (7,1), ... , (7,7)]
for detector in [1,...,16] (two at a time):
Repeat 3x:
cycle(run, nb_iterations = 2,
pattern = pattern(
# ipc_acquire: nb_ramps, nb_reads, offset_x, space_x, offset_y, space_y, ipc_vreset:
[ sequence(setup.ipc_acquire(2, 16, x, 8, y, 8, volt)) for x,y in x_y ]
))
Fulfilling¶
Targets and pointing constraints¶
Pointing: Straylight suppression as per MOCDA-SYS-003.
Instrument configuration¶
7: Calibration IPC
Instrument: NI-FWA closed, NI-GWA RSG000, single detector readout (check)
Methodology / Uplink procedure¶
Entity: NISP
Commanding Method: NISP: Calibration IPC
Dither Sequence: None
Inter-operability constraints¶
VS-CU off
In-orbit duration¶
In-orbit Duration: 39.4 h
The estimated time is 1.64h for a detector pair and a single Vreset voltage. Needed 8 times for all detectors, repeated three times.
Data volume¶
64*NP = 3776 MB, assuming photo-MACC file size
Processing input¶
Ground Processing Input Data: BPM
Data Description: 64 difference NISP images. Format and header unclear.
Processing method¶
To some degree, the IPC could also be extracted from cosmics. However, since cosmics usually extend over several pixels, the IPC matrix is not accessible in a straight-forward and accurate way. Also, cosmics create charges at different depths and IPC must then be distinguished from charge diffusion. Therefore, the SPR method is highly preferred, and currently no tools are in place or under development to extract the IPC matrix from cosmics.
Processing output¶
IPC matrix per pixel for all detectors based on the SPR method. Comparison to ground characterization data.
Required accuracy, constraints¶
Currently (2020-07-22) there is no CalCD-B requirement for the IPC correction. It is assumed that it can be corrected perfectly (the full error budget for electronic ghosting goes to ghost images).
Once a requirement has been set for the accuracy of the IPC correction, the accuracy of the IPC measurement during PV phase can be set (TBD % within the ground characterization).
Validation¶
The IPC is not corrected at pixel level. Therefore, the CalProduct cannot be verified directly by looking for residuals in the processed survey data. Possible validation methods are to extract e.g. 3x3 superpixels around highly significant hot pixels in the slew darks, measure the IPC and compare with the CalProduct. If necessary, hot pixels from the same detector must be stacked statistically to achieve sufficient S/N (assuming the IPC does not vary too much within a detector).
Alternatively, the Poisson noise (smoothed by IPC, after correction by PRNU) in the LED flats can be evaluated and compared to the expected Poisson noise (in the absence of IPC). The CalProduct should predict the correct amount of attenuation in the Poisson noise.
These measurements can be automatized and do not require a substantial amount of processing. The CalProduct should be validated e.g. every 6 months to check for increasing radiation damage.
Recovery¶
The IPC is an inherent property caused by local physics at the detector pixel level. It is therefore expected to be fairly stable and mostly subject to radiation damage, only.
Quality control parameters¶
DataModel: https://gitlab.euclid-sgs.uk/
Contact person¶
NISP-P calibration scientist