Newsgroups: comp.robotics
Path: brunix!news.Brown.EDU!agate!library.ucla.edu!news.ucdavis.edu!bunsen!nelsonk
From: nelsonk@bunsen.cs.ucdavis.edu (Karl Nelson)
Subject: CCD shuttering
Message-ID: <CLK7nM.1J3@ucdavis.edu>
Keywords: CCD
Sender: usenet@ucdavis.edu (News Guru)
Organization: University of California, Davis
X-Newsreader: TIN [version 1.2 PL1]
Date: Mon, 21 Feb 1994 05:38:08 GMT
Lines: 46

Thank you for the responces to my question of CCD shuttering, but
there seems to be a bit of confusion.  In a reply, Gordon Courage said
that all you have to do to shutter the "DRAM" is to start refreshing
it.  

On this he is mistaken.  CCD has little to do with the so called DRAM
camera.  A CCD is a charge coupled device, which collects electrons
generated by incedent photons in potental energy wells.  These wells
are then moved across the surface of a chip by dumping the wells from
cell to cell.  The rows are dumped into a serial readout area which
moves the charges to an amp which reads a voltage based on the charge
collected.  The CCD can be cleared by simply dumping all of the rows
through the serial gate into a drain. 

A DRAM camera is just as it implies, a DRAM chip which
has been cut open (or processed with a window) that accumulates a
charge. Then through readout the charge is removed.  If the charge is
enought to equal the DRAM minumum charge for a logic level high, then
the cell will read as a logical 1 else it will be seen as a zero.

Because of the differences in operation a DRAM camera does not require
shuttering, but to empty it it must readout.  A CCD must be shutter
during readout (or some other method.)  Because the image is "shifted"
across the chip an unshutted image is being eposed to light which
original created the charge, but in shifted cells!  Hence it can
result in a false reading.  The formula (I have heard) of the maximum
posible error resulting from a dark pixel being shifted through a
light exposed area across the entire chip is:

  Err = Tr /Ti   

  Err is maximum % error. Hence a Err=0.2 would mean the pixel would
have be 20% of the exposed value even when it was supposed to be
unexposed.  
  Tr is readout time. Ti is integration time during which antibloom is
active and the chip is not readout. The formula can be adjusted to
acount for partial shuttering or less travel (ie. not the farther
pixel, but one in the middle of the chip 1/2 factor) 

Thanks to John R Strohm for the flash idea, but we have to have an
audence.  We plan to instead use the rotating disk over the CCD face
instead of the light source.  I will post the results if anyone is
interested.

  --Karl Nelson 
    UC Davis Engineering student
