|
|
Thread Tools | Display Modes |
#1
|
|||
|
|||
ASTRO: OT/ Dark column formation in KAF series sensors
The Dark columns we see affecting KAF6303, KAF09000 and KAF16803 sensors are
referred to as blocked or block channels in the literature. They occur "where charge is lost or held back at a troubled pixel". They can occur within the array in random places such as I have seen form in my KAF6303E over the past two years or they can occur at the interface between the horizontal and vertical shift registers due to design induced traps. One such example was a narrowing of the channel as it fed into the horizontal shift register and due to the fringing fields associated with the channel stop, there was a small potential barrier that formed that caused a small amount of charge to be held back. These dark columns DO NOT FORM AS A RESULT OF THERMALLY INDUCED MECHANICAL STRAIN ON THE SENSOR ARISING FROM COOLING TOO RAPIDLY. I understand that a number of people are recommending a half hour cool down from room temperature to -25C. All that will do for you is waste time. If you happen to have RBI too, and need to thermal cycle your camera, this ill-advised thermal ramp will cost you an hour of imaging time: a half hour to warm to room temp to liberate the RBI, and a half hour to cool down again. Go ahead if you like to waste precious imaging time. The dark columns form due to traps not mechanical strain. There are generally four types of traps: Design Induced Traps, Process Induced Traps, Bulk State Traps and Radiation Induced Traps If there was mechanical strain causing an effect it would be more global rather than localized: you would not see individual pixels that got darker. There simply is no mechanism for causing dark columns due to mechanical strain, so get over it if you think that true: it isn't. Many different things can cause the trapping including ionic contamination, radiation effects, charge trapping in the gate oxide causing flat band shifts and so on. CCDs are very sensitive to damage from radiation. If you are lucky enough to get one with 8 nines CTE (charge transfer efficiency) it will not stay that way for long. CCDs will age in the presence of radiation and about all you can do is bury one underground in a lead vault if you want to slow down the aging. My KAF6303E was a flawless Grade 1 when it was new but over the past two years it has picked up a number of defects including some very hot pixels and a handful of darker columns that extend partway through the array. when you examine the data very carefully you will discover that most of the darker columns have an origin pixel in the array just like bright point defects. These are dark point defects and are caused by trapping sites. I have prepared a page that shows example dark and bright columns that formed in my '6303 after deployment as well as showing the behavior at -15C, -30C and -45C. Additionally it contains some scanned pages that speak to the origin of the dark columns and shows examples http://www.narrowbandimaging.com/block_column_page.htm I have placed JPEG images on the webpage to better show the problems in my dark frames and then have linked FIT data to permit quantitative examination of the darks. All darks were 10 minutes. Please also pay attention to how nicely the dark frame cleans up by cooling to -45C instead of -15C or -30C There was a recent discussion nbout cooling o another group and this is an excellent example of the sorts of things that are improved by deep cooling to minimize dark current which in turn minimizes dark fixed pattern noise. You can regard these defects as a form of dark signal fixed pattern noise. |
#2
|
|||
|
|||
ASTRO: OT/ Dark column formation in KAF series sensors
Richard Crisp wrote: The Dark columns we see affecting KAF6303, KAF09000 and KAF16803 sensors are referred to as blocked or block channels in the literature. They occur "where charge is lost or held back at a troubled pixel". They can occur within the array in random places such as I have seen form in my KAF6303E over the past two years or they can occur at the interface between the horizontal and vertical shift registers due to design induced traps. One such example was a narrowing of the channel as it fed into the horizontal shift register and due to the fringing fields associated with the channel stop, there was a small potential barrier that formed that caused a small amount of charge to be held back. These dark columns DO NOT FORM AS A RESULT OF THERMALLY INDUCED MECHANICAL STRAIN ON THE SENSOR ARISING FROM COOLING TOO RAPIDLY. I understand that a number of people are recommending a half hour cool down from room temperature to -25C. All that will do for you is waste time. If you happen to have RBI too, and need to thermal cycle your camera, this ill-advised thermal ramp will cost you an hour of imaging time: a half hour to warm to room temp to liberate the RBI, and a half hour to cool down again. Go ahead if you like to waste precious imaging time. The dark columns form due to traps not mechanical strain. There are generally four types of traps: Design Induced Traps, Process Induced Traps, Bulk State Traps and Radiation Induced Traps If there was mechanical strain causing an effect it would be more global rather than localized: you would not see individual pixels that got darker. There simply is no mechanism for causing dark columns due to mechanical strain, so get over it if you think that true: it isn't. Many different things can cause the trapping including ionic contamination, radiation effects, charge trapping in the gate oxide causing flat band shifts and so on. CCDs are very sensitive to damage from radiation. If you are lucky enough to get one with 8 nines CTE (charge transfer efficiency) it will not stay that way for long. CCDs will age in the presence of radiation and about all you can do is bury one underground in a lead vault if you want to slow down the aging. My KAF6303E was a flawless Grade 1 when it was new but over the past two years it has picked up a number of defects including some very hot pixels and a handful of darker columns that extend partway through the array. when you examine the data very carefully you will discover that most of the darker columns have an origin pixel in the array just like bright point defects. These are dark point defects and are caused by trapping sites. I have prepared a page that shows example dark and bright columns that formed in my '6303 after deployment as well as showing the behavior at -15C, -30C and -45C. Additionally it contains some scanned pages that speak to the origin of the dark columns and shows examples http://www.narrowbandimaging.com/block_column_page.htm I have placed JPEG images on the webpage to better show the problems in my dark frames and then have linked FIT data to permit quantitative examination of the darks. All darks were 10 minutes. Please also pay attention to how nicely the dark frame cleans up by cooling to -45C instead of -15C or -30C There was a recent discussion nbout cooling o another group and this is an excellent example of the sorts of things that are improved by deep cooling to minimize dark current which in turn minimizes dark fixed pattern noise. You can regard these defects as a form of dark signal fixed pattern noise. How about partial white columns in the KAF-0402 and KAI-11000. The 402 was constantly hit by cosmic rays that are rather rare in the 11000. I'm not sure why one was so much more sensitive. It developed partial columns that were permanent. I've only had 2 develop but once there they don't go away and are fully saturated so you can't dark them out. The worst starts from a several pixel fully saturated head. Then a single fully saturated column going down. Over the last 18 months it's shrunk to a 2 pixel head that's fully saturated. The other two have faded greatly and easily dark out now. Think a cosmic ray might be involved as it showed up the night they were hitting constantly with several being fully saturated. The 11000 on the other hand has them coming and going all the time. A few have been permanent but even they vary greatly in intensity over time. All dark out if you are using very current darks. Since they age rather rapidly, after a week or so you need new ones to fully erase them. Other columns just seem to come and go. They are faint and often last only a week or two then you never see them again but others will have taken their place. Even weekly darks won't help much with these but they are so faint they are only a problem if going well down into the noise. They may be 5 or 6 counts above background on either side. Rick -- Correct domain name is arvig and it is net not com. Prefix is correct. Third character is a zero rather than a capital "Oh". |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
ASTRO: Jupiter (series of pics) | J.Warren | Astro Pictures | 5 | September 17th 07 08:55 PM |
ASTRO: a case for high dynamic range sensors and deep wells | Richard Crisp[_1_] | Astro Pictures | 1 | August 8th 07 07:01 AM |
BOINC bug (display): suspended projects should show 0% resource use on the "work in progress":"resource share" column, and "no new work" should show up in the 'status' column | Max Power | SETI | 0 | October 2nd 05 10:22 PM |
Knotty Jets, the Tell Tales for Early Dark Matter and Spiral Galaxy formation. | Leo | Amateur Astronomy | 0 | October 16th 03 06:53 AM |