Re: [ccd-newastro] Dark frame integration times
- We operate at -55C, and optimize for minimum read noise - slow.I haven done any metrics, but images are exceptionally clean. Never use dark frames.
Sent from my iPad
On Sep 29, 2013, at 9:08, knelson@... wrote:
I was curious about how you operate your Pixis. The spec shows 0.005e-/p/sec @-60C. Read noise is listed as 3.5e- RMS at 100kHz and 12e- RMS at 2MHz. What temperature do you typically cool the sensor to, and what read rate do you use for your imaging? Does your measured performance equal or exceed the specs? Published specs are sometimes purposefully conservative.
---In firstname.lastname@example.org, <email@example.com> wrote:The actual specification is quoted thus: 0.005 electron/pixel/second. Shortened link to data sheet:http://goo.gl/AsFkxkOn Sep 28, 2013, at 11:17 AM, k3fit@... wrote:
Thanks Guys, Good info.
About the one electron / hour. Is that one electron/hour/pixel? One per the whole chip?
---In firstname.lastname@example.org, <email@example.com> wrote:Colder temps are always beneficial, but only so long as the camera's cooler can maintain that temperature in a stable way. If your camera is using 95% of it's cooling capacity, it doesn't have enough headroom to keep the temperature stable. For most amateur cameras, 80% of capacity is a reasonable max - and I usually like to be at 70% for really long exposures, but that's pretty conservative.Professional cameras and high-end amateur cameras tend to always keep extra capacity in reserve; they don't even let you approach 100% of capacity.As Stan noted, QE is QE; it's not affected by temperature.What lower temperature buys you, mostly, is lower noise, mostly in the form of lower dark current. Professional cameras are run at extremely low temps for this reason. Our Princeton Instruments Pixis camera has less than one electron per hour of dark current, for example. If the camera is also designed either with a very large full well or very low read noise, then you also get much better dynamic range, which allows you to use shorter (sometimes much shorter) exposures, which improve efficiency.On Sep 28, 2013, at 4:24 AM, k3fit@... wrote:
Related question: What temperature should you set your camera to? As low as it will go?
Does quantum efficiency drop off as temperature decreases?
---In firstname.lastname@example.org, <s24man@...> wrote:
You're right, I meant "darks" not "flats", and no doubt there are better statistical methods to use. I'm not someone who just likes to push buttons and watch things go without understanding really what I'm doing.
CCD stack has three ways to combine dark images. A simple mean, a min-max clip and sigma reject mean. The first two I understand, and the third I don't--yet. (And I believe Pixinsight has a few more options). I'm slowly dissecting Adam's videos because he's very good at explaining what he does--but he hasn't the time to go over absolutely everything.
---In email@example.com, <firstname.lastname@example.org> wrote:The bottom line is that there are probably better statistical methods for lowering the total noise in darks. (I presume you meant darks...)On Sep 26, 2013, at 9:47 AM, Stuart Forman <s24man@...> wrote:Thank you guys, That's what I've been doing. I've just been taking a lot of flats (a library of 40 or so) with the same integration time of the lights. I just wanted to know if anybody was doing it the way Adam was talking about, and if there was an advantage to it.RegardsStuart