ICEINSPACE
Moon Phase
CURRENT MOON
Waning Gibbous 93.3%
|
|
09-10-2011, 08:46 AM
|
Quietly watching
|
|
Join Date: Jul 2007
Location: Yarra Junction
Posts: 3,044
|
|
Cooling of CCD cameras, what temp do you use and why?
Long title, but pretty much what I'm interested in.
I've got myself a qhy9 which has quite a range for cooling, I started at -20 as I figured it could be used all year round, and then a set of calibration files would last longer. ( extra question how often do you replace calibration files), there is still some noise when it's stretched, more than my qhy8 but it's also more sensitive so I guess that evens it out.
So..... Is there any benefit in going lower, and I mean real benefit as opposed to theoretical. And where do most of you prefer to set your cooling?
If there is a noticeable benefit I would be tempted to go there, the other option might be longer subs so signal overrides noise..although there has been some comment regarding bloated stars on this chip (I went from 20 min to 10 min so I can go back that long for dimmer targets)
Thanks for your input
Clive
|
09-10-2011, 10:32 AM
|
|
Country living & viewing
|
|
Join Date: Mar 2006
Location: Armidale
Posts: 2,790
|
|
On my ST10XME I use -25 deg as I can achieve that all year round. I do need to use water cooling though in summer to assist it.
I have tried -15 and it has more hot pixels as espected. At -30 it is cleaner but I can't maintain that temp in summer.
I use the same darks, offsets for a long time and update flats as needed so like to keep to the same temperature.
|
09-10-2011, 10:46 AM
|
|
Big Scopes are Cool
|
|
Join Date: Jun 2007
Location: SE Tasmania
Posts: 4,558
|
|
I use -10 on the qhy8pro as it can achieve this temp all year round so same calibration files can be used. I started off comparing stretched darks at -10,-15 and -20 and couldn't notice any difference so decided -10 would do. I think it is going to depend on the type of CCD the camera has as to how much benefit you are going to get by going colder. The QHY8 has the Sony ICX453 CCD whereas the QHY9 has the KAF8300.
|
09-10-2011, 10:55 AM
|
|
sword collector
|
|
Join Date: Oct 2006
Location: Mount Evelyn
Posts: 2,922
|
|
I have a QHY9m and have it set to -20 all year round.
Even in summer it goes to -20 and runs at 60% power.
I am running it at 13.8Volt so i set it to not go higher than 75% power.
|
09-10-2011, 02:05 PM
|
Registered User
|
|
Join Date: Aug 2010
Location: Silicon Valley
Posts: 124
|
|
Cooling: some considerations
key reasons to cool
1) management of dark shot noise
2) management of cosmetic defects/Dark FPN
3) management of RBI
There's a relationship between the level of cooling used and the maximum time you can expose and still have the dark shot noise less than the read noise of the camera. That time point is usually regarded as the maximum practical exposure time (see pages 6-9 of the linked pdf below)
http://www.narrowbandimaging.com/inc...ling_crisp.pdf
Cosmetic defects can and do accumulate with time; usually they are pixels that have higher dark signal than surrounding pixels. The bad effects of such leaky pixels can be minimized by deep cooling (page 10)
RBI is a problem to varying degree with KAF series sensors. The KAF09000 is particularly bad. The best known method for managing such RBI is to flood the sensor with NIR light, flush it and then expose. This fills the traps prior to the exposure, placing the sensor into a known state. During exposure some of the charge leaks from the traps into the image. The charge leakage will leave patterns observable in the image that are a form of Dark Fixed Pattern Noise and these are removed completely by proper dark subtraction. However the average signal level leaking from the traps does contribute to the dark shot noise (typically dominates versus thermal dark signal) so this needs to be considered along with the thermally generated dark signal when selecting a cooling target operating point.
Surprisingly, the KAF series require substantial cooling in order to support longish exposures and low read noise target limits (pages 11-19)
Hope that was helpful.
|
09-10-2011, 02:24 PM
|
|
Highest Observatory in Oz
|
|
Join Date: May 2006
Location: Canberra
Posts: 17,424
|
|
Another concise analysis Richardo
So two key things I saw there were
1) Dark shot noise is NOT removed by dark subtraction
Only cooling can reduce dark shot noise for a given sensor
and
2) Lower read noise in a camera is certainly desirable but you need good cooling available to exploit the improvement.
The example you included of managing cosmetic defects with cooling was pretty telling too
Great explanation mate
So in the end is it basically the colder the better really..?
To answer your question Cliveus, I use -30C year round so I can be lazy with the same set of darks but given Richards explanation I should probably exploit the low read noise of my ProLine better and go lower temps in the winter months
Rich, I understand dark libraries should be re-shot periodically, how often does one need to redo dark libraries to keep exploiting low read noise and cooling?
Mike
Last edited by strongmanmike; 09-10-2011 at 02:38 PM.
|
09-10-2011, 02:57 PM
|
Registered User
|
|
Join Date: Aug 2010
Location: Silicon Valley
Posts: 124
|
|
Quote:
Originally Posted by strongmanmike
Another concise analysis Richardo
To answer your question Cliveus, I use -30C year round so I can be lazy with the same set of darks but given Richards explanation I should probably exploit the low read noise of my ProLine better and go lower temps in the winter months
Rich, I understand dark libraries should be re-shot periodically, how often does one need to redo dark libraries to keep exploiting low read noise and cooling?
Mike
|
you really need to measure it to know how much difference the extra cooling will make. In that data I posted is some measured data from Tim Khan's ML8300 running at -25C with and without light flood (Page 17)
He has very low read noise so his cooling at -25C and no light flood is sufficient for a 400 second exposure
of course none of this matters with bright objects shot in bright skies with broadband filters. The background signal levels are way larger than the noise contribution from dark signal components.
as far as libraries go: I have seen that the ambient temperature can affect the bias level a tad. Since you should NOT use dark scaling when using the light flood, you might want to take darks at extremes of ambient temperatures encountered during a year of imaging. There's really no reason to toss them unless you are picking up more costmetic defects or you have a long term bias drift issue which I doubt.
|
09-10-2011, 03:27 PM
|
|
Highest Observatory in Oz
|
|
Join Date: May 2006
Location: Canberra
Posts: 17,424
|
|
Quote:
Originally Posted by ptc
you really need to measure it to know how much difference the extra cooling will make. In that data I posted is some measured data from Tim Khan's ML8300 running at -25C with and without light flood (Page 17)
He has very low read noise so his cooling at -25C and no light flood is sufficient for a 400 second exposure
of course none of this matters with bright objects shot in bright skies with broadband filters. The background signal levels are way larger than the noise contribution from dark signal components.
as far as libraries go: I have seen that the ambient temperature can affect the bias level a tad. Since you should NOT use dark scaling when using the light flood, you might want to take darks at extremes of ambient temperatures encountered during a year of imaging. There's really no reason to toss them unless you are picking up more costmetic defects or you have a long term bias drift issue which I doubt.
|
Ok, so:
1) Colder the better is valid but most useful under dark skies where it is advantageous and noticeable.
2) A library of darks is usable for quite a while (even years?)
So perhaps I should keep the set of -30C darks for spring/summer and do a new set of -40C darks for use in winter?
Mike
|
09-10-2011, 04:08 PM
|
Registered User
|
|
Join Date: Aug 2010
Location: Silicon Valley
Posts: 124
|
|
Quote:
Originally Posted by strongmanmike
Ok, so:
1) Colder the better is valid but most useful under dark skies where it is advantageous and noticeable.
2) A library of darks is usable for quite a while (even years?)
So perhaps I should keep the set of -30C darks for spring/summer and do a new set of -40C darks for use in winter?
Mike
|
if -30C is sufficient to meet the exposure criteria then -40C will offer no additional advantage
my comment about darks taken in extremes of ambients expected was directed more to one operating temperature (say -25C) but at extremes of environmental ambient temperatures (like summer and winter ambients)
the change in ambient temperatures can make a few DN difference in the average bias level so that is where it might be a good idea to have a couple sets of darks. One for winter and one for summer temperatures
|
09-10-2011, 05:08 PM
|
|
Highest Observatory in Oz
|
|
Join Date: May 2006
Location: Canberra
Posts: 17,424
|
|
Quote:
if -30C is sufficient to meet the exposure criteria then -40C will offer no additional advantage
|
This relates to Clives original question in a way ie what real difference is there in going colder. Going significantly colder is good but only if the camera has lower native read noise in the first place - correct? If your camera has higher native read noise going colder won't help that much?
Quote:
my comment about darks taken in extremes of ambients expected was directed more to one operating temperature (say -25C) but at extremes of environmental ambient temperatures (like summer and winter ambients)
the change in ambient temperatures can make a few DN difference in the average bias level so that is where it might be a good idea to have a couple sets of darks. One for winter and one for summer temperatures
|
Hmm, this is new to me....so because my camera is low read noise in the first place and once I have determined the lowest all year achievable chip temperature it is then the ambient temperature at which the darks are taken (rather than the chip temperature) that needs to vary ie I should take a set of -30C darks when the ambient is say +20C (simulates a typical summer night) and another set of -30C darks when the ambient is say +5C (simulates a typical winter night)..?
Or am I out of my mind...?
Mike
|
09-10-2011, 05:43 PM
|
Quietly watching
|
|
Join Date: Jul 2007
Location: Yarra Junction
Posts: 3,044
|
|
A few things I'm not used to,
Dark shot noise ? Is this the bias frame ?
Light flood.... Never done that, or even heard of anyone here doing it, however given the 8300 chip had a 400 sec for -25 , I might have to lower my temp somewhat.
Mine runs to -50 of ambient , fortunately nights here generally aren't over 20 too often. So -30 is doable on most nights.
Lots to consider there, thanks for the replies from all.
|
09-10-2011, 06:06 PM
|
|
Registered User
|
|
Join Date: Jul 2005
Location: Melbourne Australia
Posts: 957
|
|
Page 17 says the data is from KAF 16803 chip not 8300 ? did I miss something ?
Also interested in this light flood you speak of. How does one do this ?
|
09-10-2011, 07:48 PM
|
|
Registered User
|
|
Join Date: Dec 2010
Posts: 573
|
|
Quote:
Originally Posted by cventer
Also interested in this light flood you speak of. How does one do this ?
|
The camera does it. Your camera either has this feature or it doesn't. If your camera had this feature, you'd know about it.
|
09-10-2011, 10:17 PM
|
|
Waiting for next electron
|
|
Join Date: Mar 2008
Posts: 2,427
|
|
Keep my QHY9 at -20 C all year round without any problems. Exposures are usually around 600 sec and calibration removes anything left over. Have run it as low as -35 C on a cold winters night but gained little from doing so so will leave it where it is.
Mark
|
09-10-2011, 10:51 PM
|
Registered User
|
|
Join Date: Sep 2006
Location: Melbourne, Victoria
Posts: 3,916
|
|
Cool thread, no pun intended. Thanks for all the info.
It does make me think I should be running cooler with the Apogee U16M with D9 cooling package. During winter running at a -30c set point, TEC power is only at 20%, so I wouldn't have any challenge in getting the camera down to -40c or -50c. A few months ago I had the camera serviced by Apogee. They tested the camera at -62c below ambient according to the test report, no doubt this is in a controlled environment. During summer, I don't like running the TEC power higher the 80%. If I run at a set point of -30c all year round, TEC never reaches this. I refresh the calibration library every three months with a quick script that I run on a night where the weather is bad.
|
09-10-2011, 11:10 PM
|
|
Registered User
|
|
Join Date: Jan 2009
Location: Adelaide
Posts: 9,966
|
|
QSI583, during winter it holds -20C easily. Spring the same. Summer most nights -15C, rarely -20C. Darks done every 3-4 months. The images start to show black squares and this tells me it is time for new darks.
Clive if you can do -20 or -25C all year around then do that.
|
10-10-2011, 09:23 AM
|
Registered User
|
|
Join Date: Aug 2010
Location: Silicon Valley
Posts: 124
|
|
Quote:
Originally Posted by cventer
Page 17 says the data is from KAF 16803 chip not 8300 ? did I miss something ?
Also interested in this light flood you speak of. How does one do this ?
|
yes, I mistyped. it was a ML16803 not ML8300
|
10-10-2011, 09:27 AM
|
Registered User
|
|
Join Date: Aug 2010
Location: Silicon Valley
Posts: 124
|
|
Quote:
Originally Posted by jase
Cool thread, no pun intended. Thanks for all the info.
It does make me think I should be running cooler with the Apogee U16M with D9 cooling package. During winter running at a -30c set point, TEC power is only at 20%, so I wouldn't have any challenge in getting the camera down to -40c or -50c. A few months ago I had the camera serviced by Apogee. They tested the camera at -62c below ambient according to the test report, no doubt this is in a controlled environment. During summer, I don't like running the TEC power higher the 80%. If I run at a set point of -30c all year round, TEC never reaches this. I refresh the calibration library every three months with a quick script that I run on a night where the weather is bad.
|
I regularly run my cameras near 95%. so long as it has enough headroom to maintain a constant temperature there's no risk at higher power numbers.
what is your reason to avoiding running higher power numbers than 80%? You are only needlesslty limiting yourself by erecting artificial barriers.
|
10-10-2011, 09:37 AM
|
Registered User
|
|
Join Date: Aug 2010
Location: Silicon Valley
Posts: 124
|
|
Quote:
Originally Posted by Alchemy
A few things I'm not used to,
Dark shot noise ? Is this the bias frame ?
Light flood.... Never done that, or even heard of anyone here doing it, however given the 8300 chip had a 400 sec for -25 , I might have to lower my temp somewhat.
Mine runs to -50 of ambient , fortunately nights here generally aren't over 20 too often. So -30 is doable on most nights.
Lots to consider there, thanks for the replies from all.
|
well obviously if you haven't heard of it there are things that are commonly done that you need to learn how to do.. or at least understand what they do and decide if it is appropriate for your usage.
The Cassini and Galileo probes both had light flood to manage RBI. it is implemented in FLI ML and PL cameras too.
Except for FLI and to a lesser extent Apogee, no one else that I am aware of implements light flood. These camera companies are in denial as to the issues associated with RBI and it is a very real phenomenon observed in ALL KAF series sensors: with some significantly worse than others.
Dark shot noise is completely unrelated to bias frames. It is the noise you get by virtue of a non-zero signal level (average) in dark frames. It is numerically equal to the square root of the number of electrons in the frame.
So 100 electrons per pixel of average dark signal level means you have 10 electrons of dark shot noise. This is random noise like read noise. so you manage it by averaging many frames together. For dark signal you can do one better and that is by cooling... so you can make it smaller in a per frame basis by cooling or you can average many frames together to 'beat it down' (like read noise).
When you take 4x the frames then you get to reduce the noise by a factor of two....
sounds like some of you guys could benefit from my basic lecture series of the basics of electronic imaging... noise sources, management thereof and how to get high SNR images and how to tell when good enough is good enough!
Keep in mind that if you are imaging something faint where read noise is a consideration, then dark shot noise will also be a consideration and if you cannot run as cold as you would like to keep the dark shot noise less than the read noise, you may be better off to reduce your exposure time and take more exposures....
this stuff isn't terribly complicated but there are a few basics that I cover in my lecture series that I think anyone that is serious about imaging ought to know... sort of like the basics of how an internal combustion engine works and the importance of lubrication, gearing and tires by way of analogy....
|
10-10-2011, 10:19 AM
|
Registered User
|
|
Join Date: Sep 2006
Location: Melbourne, Victoria
Posts: 3,916
|
|
Quote:
Originally Posted by ptc
I regularly run my cameras near 95%. so long as it has enough headroom to maintain a constant temperature there's no risk at higher power numbers.
what is your reason to avoiding running higher power numbers than 80%? You are only needlesslty limiting yourself by erecting artificial barriers.
|
Voltage
When you are running the camera near 95%, what is the voltage draw? If I hit 95%, I get a low voltage warning as the TEC consumes considerable power. Running at constant 98% or more, the camera becomes unstable. I've ensured the camera has ample power, I've even connected it to a regulated 12vdc 10amp source, same result.
|
Posting Rules
|
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
HTML code is Off
|
|
|
All times are GMT +10. The time is now 03:11 PM.
|
|