Optimal sampling of charge-coupled devices

Abstract

It is shown that when the output noise spectrum of a charge-coupled device is flat, the maximum signal-to-noise ratio is obtained using a differential-averaging sampling function (positive, then negative square wave sampling). However, in astronomical applications, where image quality rather than readout speed is more desirable, 1/f flicker noise on the on-chip amplifier dominates the noise spectrum. It is shown that even when 1/f noise dominates, the same differential-averaging sampling function yields a signal-to-noise ratio which is never more than 36% from optimum. The results were obtained using a matched-filter analysis.