Dear Sir or Madam,
I am using AD9826 building a CCD imaging circuit. I found there are some missing code in the image data. Some pixels are significantly different with other pixels around them. The different pixel seems to have a wrong bit in its high bits. The different pixel always one point and has not fixed position, appears just ones. For example, when we test a dark image, all data of the pixels are less then 1024, that means all data should be zero in there high 6 bits. Sometimes, there is a different pixel with a “1” bit in the high 6 bits. And the different pixel has the almost same low bits with its adjacent pixels. The frequency of the different pixels appearing depends on the frequency of AD converting. At 256K sampling rate, there are about 10 different pixels per 100 million pixels. At 500K sampling rate, there are about 5 or 6 different pixels per 100 million pixels At 1M, 3 or 4 per 100 million and at 2MSPS, less than 1. Although the probability is very low, even less than 1 pixel per 100 frames(1,000,000 pixels/frame in the CCD we are using) at 2 MSPS readout rate, However, it is still a problem for our data analysis. Have you ever encountered such a problem and how to address it?
Many thanks and best regards