At the beginning of the semester there was a problem with one of the quadrants of the NOTCam detector not working. This was described in detail in my last report, where we in the end found that the problem was with a transistor on the mounting board of the detector which was replaced with a spare after which we got signal from all four quadrants again. Also, an interface board needed to be replaced as this had failed due to the problems with the mounting board, where we had to make some (undocumented!) modification to make it work. Since these changes the detector has continued to work properly.
The modified clock board for NOTCam that should correct the ``dark current'' problem with the detector has been built and tested both in the laboratory and in the controller, though with the array disconnected. We hope that the sudden jumps seen in the reset level and described with examples on
http://www.not.iac.es/instruments/notcam/staff/newpcb/newpcb.htmlwill also be solved/cured together with the dark-current problem by modifying the clock-boards. It is planned to try the board out on NOTCam in the period after the last NOTCam run in November 2009 and before the next run in February 2010.
If the modification is successful a full check and characterization of the detector (basically a new detector commissioning) will be needed as we expect all properties to changes at some level.
The bad-pixel masks provided up-to-now have not been optimal in the sense that the tool was developed for another array and electronic setup (the engineering grade array and old PCB). Also, it did not work well for clusters of bad pixels.
New bad-pixel masks are being prepared and made available. The new masks differentiate between zero-valued pixels, ``cold'' pixels, and ``hot'' pixels, where ``cold'' refers to low response compared to the nominal value of neighboring pixels (typically 20-40% lower), ``Hot'' pixels are few and mostly along the edges for short integrations, but the number is a strong function of exposure time. Zero pixels are mainly dead pixels located in one corner of the array and a few spurious single dead pixels. Also, 2048 more pixels are perceived with zero value, due to a problem the controller has in addressing the pixels in the first readout column. This should disappear with the new controller. The ``cold'' pixels are scattered over the array, many of them in the bad corner, but those over the array are usually located in small groups. Only one larger bad feature was present when installing the array in 2007, the rest being single pixels or few-pixel groups and the bad corner, in total affecting 0.2% of the array. Over the years this has slowly increased to 0.4% of the array. The increase is likely related to the vacuum problems with have, and in one case a problem with the telescope altitude amplifier which caused the telescope to shake strongly during power-on while NOTCam was mounted.
When the improved characterization of bad pixels is finished and made available, the notcam.cl reduction package will be upgraded to improve the handling of bad pixels.
Thomas Augusteijn 2010-02-09