Tek CCD Antiblooming

[fhh at nofs.navy.mil]

Having just viewed the images on "Paddy's Problem Page" as regards Simon
Tulloch's antiblooming-induced features, I have some qualitative comments
from my own use of antiblooming on our first Tek 2K:

I too primarily say many bright spots on my own images, in repeatable
positions from frame to frame; these I interpreted as pumped pockets.  I
did not see excessive dark current anywhere on the chip as a function of
enabling antiblooming clocking.

I was running the Tek 2K CCD at -134 C, and dark current contributions of
any sort were neglible.  A completely DARK frame with antiblooming on showed
considerably fewer bright pixels than ANY frame with anitblooming on in which
a non-zero light-induced background was present.  Two conclusions I made:
1. pockets don't get pumped up from a zero-charge condition, and 2. spurious
charge effects (from impact ionization of electrons in motion) required some
source electrons to be initiated, hence the lack of pumped pockets or
spurious charge effects in the completely dark frames.  The Tek 2K was always
run in standard (non-MPP) mode.

A separate source of charge for pocket pumping and/or spurious charge
generation could be the dewar window if that window material contained some
detectable radio-isotope content.  BK7 is sufficiently radioactive to 
qualify; S1-UV fused silica is sufficiently clean.  In a recent experiment
in which a filter made of GG475 glass was placed in the vacuum of the CCD
camera approximately 1.25mm from the face of the thinned Tek 2K CCD, the
"hit rate" from energetic particles and photons went up by x6 over the
background measured in the same camera both before and after the presence
of the GG475.  This leads me to wonder whether the window used in the camera
making the images in Paddy's page might be contributing source electrons in
the CCD to be manipulated by the antiblooming clocking?

We gave up on antiblooming on the Tek 2K in the end; it was needed to reduce
the bleeding from Jupiter as we took frames of the Galilean satellites in
support of the Galileo mission.  We solved Jupiter by using a methane filter
which reduced the flux from Jupiter to less than that from the satellites;
this will be a poster paper by Alice Monet next week at the AAS meeting.

Are your "wobble-clocks" bilevel?  RC-filtered analog-switch outputs or
opamp buffered?  Trilevel antiblooming clocks as per Kohley etal's paper
from last year's SPIE conference (SPIE Vol. 2415 pg 67)?  Mine were RC-
filtered complementary-VMOS-buffered analog-switch outputs, with little
control over rise times.

Hoping I haven't added to the confusion,

Fred Harris
US Naval Observatory Flagstaff Station