CCD-world: Rockwell Hawaii question - stars abound !

[Peter Moore]

The following was posted to CCD-world:

Hiya Barry,

Thank you very much for your very detailed help that you gave us a month
ago. I've been waiting till after we commissioned the camera to reply so
as to collect the most information about the problem from the most recent
modifications and tests. With regards to the hot pixels we've been able to
determine the following :-

1. Hot pixels are spatially stable and are time dependent i.e they get
'hotter' as integration time increases.

2. The levels are pedistals and photon flux is still collected up to
saturation level.

3. There is a decrease in the number and severity of them with a decrease
in temperature.

This leads me to conclude that we have some form of leakage from the pixel
site itself that degrades the reset charge in time. This is also born out
by the fact that some sites with severe discharge influence adjacent
pixels (column and row). In addition the pedistal level shows a near
expodential growth with time, very much like a discharge curve. I will in
the future play with the clock, 'high', and vcc values to see if this
influences them.

I've gone through your helpful list and all points are in the green. BTW
we are using 240 ns clock rise and fall times, pixel time is 1700 ns
and reset time is 1000 ns.

We have found that taking a bias of the same integration time as the
science integration time and then subtracting this from the science cds
image data gives near perfect subtraction of the hot pixels. The only
concern then is that they limit the dynamic range of those affected
pixels. We are getting a useful 'full well' capacity of just over 85K -e
with a Vrest of 0.55v. At 68 Kelvin we see about 7% of pixels affected.
Noise is a nominal 12 -e with a single read cds (mndr = 1).

With regard to the ramp seen on early lines during readout, Derek sorted
this out by doing a dummy read of n columns before actual readout. This
exercised the internal mux electronics and flushed leakage (?) charge out
minimising the effect. We now see about 10 ADU difference on the early
columns that dissipates by column 50 when flushing 64 columns before
readout. It was 100's of ADU before and now the residial dissappears
completely by the cds subtraction.

Thanks very much for the advice Barry and Craig,
All the best in your endeavours,
	Peter.

   =+===+===+===+===+===+===+===+===+ ING +===+===+===+===+===+===+===+=
 =     Peter Moore, Isaac Newton Group, La Palma Observatory, Spain.     =
=   E-mail pcm at ing.iac.es                  Voice Office +34 922 405566    =
=   S-mail Apartado de Correo 321,           Fax Office +34 922 405646    =
=          38780 Santa Cruz de La Palma,     Voice Home +34 922 435042    =
 =         Canary Islands, Spain.                                        =
   =+===+===+===+===+===+===+===+===+ LPO +===+===+===+===+===+===+===+=

- -- CCD-world -- --
CCD-world is fully moderated. Send posts to CCD-world at astro.ku.dk
Standard replies will go to the list; address personal replies manually.
For more information, please go to:  http://www.not.iac.es/CCD-world/