Few figures from the papers:
Tuesday, April 12, 2016
More Single-Photon Imaging Papers
IEEE JEDS publishes early open access papers "Photon Counting Error Rates in Single-Bit and Multi-Bit Quanta Image Sensors" by Eric Fossum and "Determining Conversion Gain and Read Noise using a Photon-Counting Histogram Method for Deep Sub-Electron Read Noise Image Sensors" by Dakota Starkey and Eric Fossum.
Few figures from the papers:
Few figures from the papers:
Well, as long as you posted these figures, I might as well say Fig 6 is pretty important. Look at the lower two blue diamond symbols, at ~0.04 electrons per pixel per read and at and ~0.1 electrons per read. It shows that at 0.29e- rms read noise, the average output read from the pixel will be ~0.08e- and ~0.13e- respectively. You cannot accurately count photoelectrons is this low level range with a read noise of 0.3e- rms. That leads to a systematic error in count that is substantial (e.g. off by a factor of 2!) and many of the counts will be false. We need a read noise of 0.15e- rms or less for accurate counting in this range.
ReplyDeleteThe reason for the systematic error is that while most counts should be zero (96 out of 100 at H=0.04) the read noise tail will cause some voltage reads to be above the threshold and they will give a false count. So at H=0.04 we will see about 4 true counts plus 4 false counts = 8 out of 100 total counts.
Finally the QIS will be simply a ultra-high gain mode in a conventional 4T pixel design, am I right ?
ReplyDelete-yang ni
How about superconducting nano-wires at cryogenic temperatures? :-D
DeleteYou are right, but the QIS-show must go on !
DeleteNo but some of what we have developed will apply to conventional sensors of course.
ReplyDeleteThanks for sharing interesting scientific achievement. I m wondering if there are research projects to develop QIS for depth measurement (direct time of flight, so a pixel integrating a TDC). As far as I understand, Time of flight sensor based on SPADs array are not yet compact and large enough, most probably due to pitch issue (to limit cross talk...). With QIS we could benefit from the possibility to decrease pitch. I you have any documents that you could share that could help me to understand most promising technology for depth sensor based on time of flight measurement, it is more than welcome...thanks in advance. Cedric
ReplyDelete