Pixel-level A/D conversion: comparison of two charge packets counting techniques

Abstract — A comparison between two techniques for pixel-level ADC in CMOS image sensors is presented. As the charge represents accurately the amount of illumination, the principle of both techniques consists in counting small charge packets which come from the detector. Based on a 0.13 µm CMOS technology, simulation results and first measurements are presented showing that the LSB value can be reduced to 1900 electrons with a static power of 250nW per pixel. For the moment, due to the technology available, this technique is restricted to medium size pixels, i.e. 50x50 µm2 or 25x25 µm2 when the fill factor is close to 100%. I