CCD Detector Lab

CCD Detectors

September 30, 2015

Ryan Hall

Introduction

            In this lab we obtained darks, flats, and bias images from CCD cameras. A CCD, charge-coupled device, camera takes in incoming photons that excite electrons in the pixels. These excited electrons are what is eventually used to build an image. Along with this, these cameras have other qualities that can work against the actual image you are trying to build. Darks, flats, and bias images are taken to help rid the scientific images of this background noise. Darks are images taken with the camera’s shutter closed to account for the noise generated by the electronics of the camera. This lab took a variety of exposure times while taking out dark images to show a relationship between the counts and exposure time. Normally when dark images are taken the exposure time is consistent. Bias images are taken to account for the base noise that will be generated for the camera, so it will have an exposure time of zero. Flats are images of a relatively equally lit screen. This finds which pixels seem to be more or less sensitive to incoming photons. These exposure times are not required to be the same as darks; however, if they are, it will make the work easier when the obtained data is being reduced.

Procedure

            This lab begins by opening up Maxim DL on your computer. In the Maxim DL window go to view and click camera control window. This brings up the window that allows images to be taken and saved by the CCD camera. Once your CCD camera has been plugged into the computer with a USB cable click connect under the settings tab of the window. These cameras are designed to be used for taking astronomical images; however, this lab is designed to understand how to account for the background noise of the cameras. Given this, these cameras are not needed to be connected to a telescope to accomplish the goals of this lab.

                At this point images can start being obtained. All images must be saved individually after they are taken and 10 images of each type are required. For bias images, the exposure time is set to zero and ten images are taken and saved. The shutter on the camera should remain closed for these images, if it does not, the lens cap can be put on to prevent light from entering. Our flat fields were taken by pointing the camera at a blank, white piece of paper. A few exposures were taken at first to make sure the counts were at appropriate levels. An appropriate level meaning an unsaturated image while having counts that are higher than the darks or bias. The flats we obtained ended up having .001s exposure times. Darks were taken with an exposure .05s and had the lens cap on. We also took dark images at exposure times of 1, 2, 5, 10, and 20 seconds and saved one image for each integration time.

Results and Discussion

                This lab saved 10 images for each type of image, darks, flats, and bias.  Using IDL we can use the command READFITS to allow the image to be viewed as a 648 by 486 array. At this point the tvscl command in IDL can be used to show the image in Xming. An example of each image type is shown in the figures below.

Figure 1: Bias Image

Figure 2: Dark Image

Figure 3: Flat Field

The flat field image shows darkening in lower corners of the image; however, the counts are fairly consistent throughout.

A mean function can be used on each image to find the average number of counts for each image. By taking the average of each of the 10 images of each type, the average counts for the darks, flats, and bias is found as shown in the table below.

Type	Exposure time (s)	Avg. Counts
Bias	0	1923
Flats	.001	6241
Darks	.05	1114

                This table shows that the counts in darks are lower than the counts in the bias. This is not supposed to be the case. This error seems most likely to have been caused by the fluctuating temperature of the CCD camera. These averages also include bad pixels within their counts. A find command in IDL was used to find a rough estimate of the number of these bad pixels. (as shown in image below)

Figure 4: Bad Pixel Count

This image shows there to be 27 bad pixels, assuming that any pixel with a count over 10,000 is defined as bad.

                We obtained our dark current rate by taking dark images at varying exposure times. It is expected that with increasing exposure time yields an increase in counts in a linear relationship. The slope of this relationship would then be the dark current rate while the y-intercept would be the average bias level. The results from our data are shown in the graph below.

Since we already discussed that there was error in our bias count, it is not surprising that the y-intercept of this graph is not around what our bias level is. Even with that, the y-intercept here has a value of 1021.8 which would be an appropriate value since our average dark counts was found to be 1114. It should also be noted that since there are only six data points on this graph, and that they are scattered rather heavily, that this trend line cannot be taken to be very accurate.

Conclusions

                This lab had us take flat field, dark, and bias images to find how much background noise is created by the camera. Our data yielded an average dark count of 1114, flat field count of 6241, and bias count of 1923. We obtained a dark current rate of a little over one, 1.17, and the bias level calculated from this graph was found to be around a more appropriate 1021.8. The approximate count of bad pixels was found to be 27. This data can be used to make a more accurate reduction of data obtained from future observations.