Measuring DEIMOS Detector Parameters

Background

This document describes the procedure for measuring the gain and readnoise of the DEIMOS science mosaic detectors. It is to be run by a WMKO Staff Astronomer.

Prerequisites

1200G grating loaded.
Z filter loaded.
Instrument interface running as any DEIMOS numbered account.

Procedure

Acquire images. Open an xterm window on deimosserver (polo) and execute script
```
gain_check
```
to acquire 3 biases and 5 flats in both spectral and direct mode.
Analyze images. Run the following commands:
```
deimos_mosaic_gain_check -m direct
```
Exit IDL terminal
```
deimos_mosaic_gain_check -m spectral
```
This creates logfiles in the directory /s/sdata1004/ccd_gain.
Update web page. Execute the command:
```
LoadDetectorDataInWeb Auto All
```
to regenerate the web page describing gain and readnoise.
Note that it will take several minutes until the external web page is regenerated.

Description

Acquisition

We configure DEIMOS for a mode that provides near-uniform illumination across the entire detector plane by acquiring internal slitless flats in spectral mode. This turns out to be challenging to accomplish, because even high-resolution spectral flats with the internal quartz lamp saturate. I found that one mode which succeeds in providing generally uniform illumination is to insert the Z filter, thus blocking most of the light, and select a central wavelength of 5000 Å, which is well below the cutoff of the filter. For some reason (possibly internal scattered light) this produces relatively uniform illumination across the middle of the detector plane; at the edges there is an excess of light.

Analysis

We employ the "Janesick method" (employed in the IRAF findgain task) for measuring gain and readnoise simultaneously from a set of biases and flats. In this case we consider only the half of each detector which is closest to the middle of the focal plane in order to avoid the regions of nonuniform illumination near the corners. On each amplifier, we derive the gain and readnoise within numerous independent 256×256 pixel regions and compute the median value across the array to derive the adopted gain and readnoise for that amplifier. For each region, we determine which pair of the 5 flats has the closest match in terms of illumination level, and we use that pair of images to derive gain and readnoise for that region. A bootstrap analysis of the array of gain and readnoise values for individual regions provides the quoted uncertainties on each quantity.