ISSCC just published its 2006 program. Session 27 has quite a good collection of image sensor papers.
The most interesting one is Samsung half-inch 7.2MP sensor. The pixel size 2.25um with 57% fill factor (geometrical? effective?) due to 4-pixel shared architecture. The full well is 14Ke while noise level is 8e - a very nice result for such a small pixel. They use 0.13um Cu process.
Sony presents 6.4MP sensor with their masterpiece 2.5um zig-zag pixels. The fill factor is 38% - quite good assuming they use 0.18um 1P3M process. Readout noise is 7e.
Another Sony presentation is 2MP sensor with ~4um pixel. It looks like the bigger pixel buys them much lower readout noise - just 5.2e. Another achievement is 0.5e FPN - by far the best in the industry.
Another interesting paper came from Cypress - probably by ex-SmalCamera guys. They use a feedback noise reduction idea to bring the 3T pixel readout noise down to 13e. This looks similar to what Altasens is doing. However, Altasens noise is lower - just 8e. But both numbers are quite good to an old-fashion 3T pixel.
As a side note, the session chair Boyd Fowler is CTO and VP of Technology of Fairchild Imaging. He used to be with Agilent for some time. It's interesting to see him re-united with Pixel Devices assets that Fairchild Imaging aquired almost two years ago.
No comments:
Post a Comment
All comments are moderated to avoid spam and personal attacks.