Chapter 6:
Charge Measurement
Authors(s): James R. Janesick
Published: 2001
DOI: 10.1117/3.374903.ch6
Charge measurement is the last major operation performed by the CCD. This process is accomplished by dumping signal charge onto a small sense capacitor located at the end of the horizontal register. The capacitor is connected to an output amplifier which delivers a buffered output voltage for each pixel. Other than photon shot noise, previous CCD operations of generating, collecting, and transferring charge are noiseless processes. In theory, the charge packet can arrive at the sense node without uncertainty to the single electron! However, there are a host of noise problems that the user may confront. For example, shot noise from thermal dark current will add uncertainty to the charge packet being measured. However, this unwanted source of noise can be completely eliminated by cooling the detector. In Chapter 7, we will discuss dark current and many other on-chip and off-chip noise sources that disturb the measurement process. We will see that each noise source can be controlled and removed. This chapter discusses the charge measurement process performed by the CCD output amplifier and off-chip signal processor. Unfortunately, the output amplifier adds noise to the charge packet measured. Although the uncertainty is very small, the noise induced cannot be fundamentally reduced to zero. It is interesting to note that the output amplifier read noise has settled to one electron rms (slow-scan). As far as anyone knows, this unique level has happened by coincidence and is not set by any fundamental law of nature. However, it has taken 30 years and millions of dollars of CCD amplifier development to reduce the noise down to this level. In the last 10 years, amplifier noise has been reduced by no more than a factor of two, showing that the optimization process has slowed down and that major advancements to further enhance charge detection are not likely. Also, the need for lower noise for most CCD applications is not warranted. Recall that 1 e− signal will generate a shot noise uncertainty of 1 e−, yielding a S∕N of unity independent of amplifier read noise (an exception to this argument is when point or line sources are measured where lower read noise is always better).
Online access to SPIE eBooks is limited to subscribing institutions.

Back to Top