CCD Selection Guide

Choosing the correct CCD for an application requires a knowledge of the wavelength of the incident photons, the spatial resolution required, the dynamic range of the scene, the optical connection from the experiment to the camera, and other experimental details such as experiment time, signal intensity and trigger issues.

  This article will summarize the most important components to be considered.

Retriever Technology delivers cameras with a wide variety of CCD chip sizes.  We supply cameras with CCDs from E2V Technologies, Fairchild Imaging, Atmel and Kodak.   The number of pixels in our cameras typically ranges from 1024 x 1024 to 2048 x 2048 to 4096 x 4096, but also includes a wide variety of intermediate and outlying sizes.  Depending on the manufacturer, pixel sizes range from approximately 9 microns up to 24 microns.

Most static imaging applications utilize full frame CCDs.   These CCDs are available in both front illuminated and back illuminated configurations.   Back illuminated devices are used in light starved applications where high detection efficiency (quantum efficiency) is required.   They also are used for certain wavelength ranges where the back illumination has higher quantum efficiency for that particular wavelength.

A generalized selection guide is shown below, in which the various optimum CCD types are shown for particular wavelengths.

Note that there are other coating options and configurations that can be selected as well.   Please discuss your particular need with Retriever Technology.

Retriever Technology provides cameras with a variety of optical front ends.  

• Lens coupled.   A window at the front of the camera allows lens-directed light to illuminate the CCD.    Standard and custom lens mounts are available.
• Fiber optic coupling.   Fiber optic faceplates (1:1 magnification ratio) and fiber optic tapers (up to 4:1 (de) magnification ratio) are available.   Much higher light transmission is possible with fiber optic coupling vs. lens coupling.

Dynamic range indicates how many light intensity values can be captured in a single scene.  A high dynamic range allows for faint signals to be seen in the same scene as extremely strong signals.

For a CCD sensor:           Dynamic Range = Full Well Capacity/Ö (Read Noise)2 + (Dark Noise)2

Typical camera values are:

Full Well = 120,000 electrons

Read Noise = 4 electrons

Dark Noise = negligible

Which gives typical dynamic ranges of = 30,000: 1, or in terms of optical density = 4.47 OD, which also = 89.5 dB.

Read Noise -- For scientific grade cameras, the largest noise source is typically generated by the CCD amplifiers themselves, with an additional contribution from the camera controller’s electronics.   Read noise is a function of readout rate, decreasing as read rate decreases.  

Retriever Technology has the lowest noise electronics available.   Typical read noise values are 4 electrons at 500 kHz effective Read Rate.

Full Well Capacity

Dark Noise

CCDs build up thermally generated signal over time.   This is called Dark Current, and can be subtracted out of the image because it is a repeatable phenomenon.  However, there is a noise component in the Dark Current called Dark Noise, given as:

Dark Current can be greatly reduced by cooling the CCD, typically dropping 50% for every 5 – 7°C of cooling.  
The following plot shows dark current for a typical Retriever Technology camera:

Read Rate

Read Rate is defined as the speed at which the CCD is read out (clocked).    Typical electronic circuitry allows read rates from 20 kHz to 1 MHz.    Many CCDs have multiple readout amplifiers (ports), which speed up Read Rate.  Total time to read out a CCD is given by:

Read Time (seconds) =  # Pixels / (Read Rate (Hz) * # Ports)