ACS CCD parameters
The new ACS CCD electronics installed in SM4 permit four A/D gain settings for the Wide Field Channel: nominally 0.5, 1.0, 1.4, and 2.0 e-/DN. All are available to the observer, but only gain 2.0 is presently supported by STScI with calibration reference files. The default gain is 2.0 e-/DN, which critically samples the read noise and spans the full pixel well depth of ~84000 e-. Consult the ACS Instrument Handbook for the precisely measured value of each gain setting.
The High Resolution Channel of ACS was not recovered during SM4, so it remains unavailable for any science or calibration uses.
The user can specify the number of distinct frames (exposures) composing an observation to mitigate the deleterious effects of cosmic-ray hits and bad pixels. Cosmic-ray hits in long exposures can be remedied with the "CR-SPLIT" parameter, which allows (1) fewer numbers of detected cosmic rays per exposure, and (2) the identification and removal of cosmic-rays during data reduction. Dithering is recommended over the use of CR-SPLIT because dithering mitigates the effects of both cosmic-ray hits and bad pixels and enables greater sampling of the point-spread function. The user should beware that the ETC uses a default setting of CR-SPLIT=2, whereas the default setting for Phase 2 proposal preparation with APT is CR-SPLIT=NO.
To avoid an excessive number of cosmic ray detections with the CCD, longer exposures should be split (CR-split) in order to:
- keep the number of detected cosmic rays low
- be able to remove the cosmic rays during data reduction
The default CR-split is 2.
Quantum Yield correction for CCD detection of UV photons
For CCD detectors in the optical, each detected photon usually generates a single electron (i.e., photons absorbed × the gain correspond to the total number of electrons). However, in the near UV, shortward of ~3200 Å, there is a finite probability of creating more than one electron per detected UV photon (see Christensen, O., 1976, J. App. Phys., 47, 689). The throughput curves adopted in synphot correctly predict the number of electrons generated per incident photon and implicitly include this UV quantum yield correction. However, since multiple electrons are generated from a single photon, the actual number of photons detected, and therefore the S/N obtained, is less than the number of electrons detected would suggest.
To take this into account, the ETC corrects the number of electrons calculated by synphot by dividing the results of the synphot calculation by Q, the mean number of electrons generated per photon. For imaging mode calculations, this correction is calculated by applying the correction appropriate for photons at a wavelength equal to the "effective wavelength" determined for the synphot calculation. For spectroscopic CCD observations, Q is calculated correctly for each wavelength bin. The "source count" rate reported by the ETC for CCD observations is actually this corrected count rate rather than the true number of electrons predicted by synphot. However, the true uncorrected number of electrons is used for comparison with the CCD saturation limits and for the "Brightest Pixel (single exposure)" quantities.
STIS CCD Parameters:
For the STIS CCD, the default gain value of 1 offers the lowest read noise, but the analog-to-digital converter then limits the maximum signal that can be detected without saturation to about 33,000 e- per pixel. Using a CCDGAIN setting of 4 allows the full well of the CCD to be used (about 144,0000 e- per pixel near the center of the CCD and about 120,000 e- per pixel near the edges). The CCDGAIN=4 setting has the disadvantage that an additional large scale pattern noise is imposed on the image. It has the advantage that, the CCD response when using CCDGAIN=4 remains linear even beyond the 144,000 e? full well limit if one integrates over the pixels bled into, and for specialized observations needing extremely high S/N, this property may be useful.
The STIS CCD also allows pixels to be binned by factors 1, 2, or 4 during the readout. This will reduce the read noise at the expense of spatial information. For spectroscopic observations, seperate binning factors can be specified for the spatial and dispersion directions.
For the STIS CCD, CR-SPLIT=2 is the default.
STIS ACQ modes always use the CCD with CCDGAIN=4 CR-SPLIT=NONE and no binning, so the STIS ACQ mode ETC does not include any user adjustable additional CCD parameters.
For calculations using the STIS MAMA modes, the settings specified for the additional CCD parameters are ignored.