Subject: DSC news #2 (The CCD)


Submitted: Fri, 26 Jan 96 17:47:57 -0600

Message number: 18 (previous: 17, next: 19 up: Index)


This is update our knowledge of the CCD in the DSC.

The chip is a 2048x2048 SITe UV anti-reflection coated device.
In September, 1995 this chip was contaminated with water. The water 
spots are visible on a flat field image. Originally the spots had 
an amplitude of about 25%. After a cleaning of the device by SITe, 
the amplitude diminished to roughly 5%. For the most part, the water 
spots flat field out nicely. (Currently, the best flat fields are 
twilight flats.)

The bias image has structure in it, probably due to clocking
noise in the dewar. A typical bias image is available in the
new dop documentation. I believe this will subtract out, given
a good enough bias image. (Take many bias frames and median them

The chip has 4 working amplifiers:
Amp 	Bias (Dn)	read noise (DN)		gain (e-/DN)
0	  4865		  3.5 			    *
1	  5292		 33.6			    *
2	  8499		  3.9			    *
3	  8276		  4.0			    *
The * means that we have not measured it yet. The gain
used to be 2.5, but probably has changed. The most current
values will be at $DOP_DIR/doc/www/dop.CCD.html on fnstar.

We currently drift scan using amps 2 and 3, and stare through
amp 2. If you really want to use amp 0 for staring, try the command
	fnstar-OAC> cc loadProgram stare_amp0
	fnstar-OAC> goStare ...

There remains small amounts of spurious charge in both the
parallels and the serial registers. This is noticeable as
ramps along both rows and cols. The amplitude has been much
reduced by changes made at Fermilab by Connie Rockosi, and
is now 10-15 DN in amps 2 and 3,  and 8 DN  in amp 0.
These should subtract out with the bias frame, leaving a
maximum noise term of 2.5 DN, assuming the gain is 2.5.
(There may be spurious charge induced by the transfer gate,
leaving another 1.5 DN of noise behind after subtraction).
In the end, after subtraction, the DSC has a readnoise in the 
vicinity of 5 DN, or 13 e-, again assuming a gain of 2.5.

A further fix while at Fermilab eliminated noise in the 
temperature sensor, which should produce a stable operating
temperature. Before, the operating temperature in effect
depended on the readout program and since spurious charge
is worse at lower temperatures this created a situation
where the amount of spurious charge could vary with time.

Finally, we have been having problems with RF noise on images.
This is the zig-zag or curvy pattern on images, distinct
from the purely vertical banding of the bias structure.
We think we have found the source of the RF noise (the camera
electronics outside the dewar) and have taken steps to
minimize it. In the near future Jon Brinkmann expects to
rebuild the DSC support structure for Tim Mckay's new filter
wheel box, and then we will mount the electronics in a place
to eliminate the noise completely. Of course, then the RF
noise will be dominated by other sources, examples of which 
we have already seen.

James Annis

APO APO APO APO APO  Apache Point Observatory 3.5m  APO APO APO
APO  This is message 18 in the apo35-dsc archive. You can find
APO  the archive on
APO  To join/leave the list, send mail to
APO  To post a message, mail it to