Back to forums

Cooled camera temperature variances, how much is acceptable ?

ZWO ASIAIR Plus Processing Dark Frames Calibration Deep Sky
andrea tassellilunohodovRick VereginStefan PflegerEric Gagne
•34 replies•1.5k views
Eric Gagne avatar
Eric Gagne avatar
Hi,

I have set my new ZWO camera outside to take some darks for building a library using the Asiair plus.  I have set it to -20 degrees and after 25 minutes it still has not settled.   It keeps going up and down, ranging from -19.8 to -20.4.

How much variance is acceptable ?  If I take a series of darks with maybe 1% being a few tenths of a degree above or below my target, is it going to affect my calibrated images later ?   Should I take more than I planned and get rid of all those that are not exactly at -20 ?
Well Written Engaging
Eric Gagne avatar
Eric Gagne avatar
I just realized I didn’t post in the correct forum,  sorry.
Well Written Respectful
Quinn Groessl avatar
Quinn Groessl avatar
That temperature swing wouldn’t bother me all too much, although both my 183mc pro and my 1600mm pro are able to stay pretty rock solid on my set temp according to me ASIAIR.
Arun H avatar
Arun H avatar
Eric Gagne:
Hi,

I have set my new ZWO camera outside to take some darks for building a library using the Asiair plus.  I have set it to -20 degrees and after 25 minutes it still has not settled.   It keeps going up and down, ranging from -19.8 to -20.4.

How much variance is acceptable ?  If I take a series of darks with maybe 1% being a few tenths of a degree above or below my target, is it going to affect my calibrated images later ?   Should I take more than I planned and get rid of all those that are not exactly at -20 ?

that variation will exist in your lights too. It is not surprising that a large delta T results in greater variation than a smaller one. If you throw away the darks that aren’t exactly -20C, do you plan to do the same for your lights? I would simply accept that variation and proceed.
Rick Veregin avatar
Rick Veregin avatar
Hi Eric,
This goes back to why we do darks. We do darks to remove the bias signal and to remove the dark signal from your images to enable good calibration with your flats. They do help remove hot and cold pixels as well. These hot/cold pixels do depend somewhat on the exposure, so the dark signal removal helps with that. But if you dither then you can remove hot and cold pixels especially if you use a rejection method in your stacking. Also, the stacking software can detect hot and cold pixels, to remove them. Hot/cold pixel removal is not going to be that sensitive to temperature over a fair temperature range. So the biggest concern is the dark signal/bias removal.

Your camera will have specifications of the dark signal and its dependence on temperature. So you can actually calculate how much your dark signal (in electrons/px) for your sub changes with temperature. You also know what your bias sign is for the sub, since this is something you will have set, though that is usually in adu, so from your gain you can convert that to electrons as well. Finally, you know what your read noise is for your sub in electrons from your specifications. 

Your bias has a noise associated with it, which is the sqrt(bias signal). Your dark signal has a noise associated with it, the sqrt(dark signal). And you have your read noise. Now if the change in your dark signal with temperature is much less than these noises, you are good to go, as the error from the dark signal change is within all the other unavoidable errors in your system.

I am sure if you do these relevant calculations, you will find that your dark signal change for a small temperature shift is much much less than these noises, so you don't need to worry about this. For a modern low noise CMOS camera you will typically find that your temperature can vary by quite a bit, in fact typically by many degrees, before your dark signal change with temperature is significant.

Hope this helps
Rick
Helpful Insightful Respectful
Eric Gagne avatar
Eric Gagne avatar
If you throw away the darks that aren’t exactly -20C, do you plan to do the same for your lights?

Good point.
Christian Großmann avatar
Christian Großmann avatar
Hi Eric,

I own two ZWO cameras (183mm and 294mm) and both show the same behavior. I never had issues with that. My QHY cameras seems to be a bit more stable, although it is not sure if the shown values are really true. But a bit of swing around the -20 degree mark is normal. It somewhat depends on the environment condition and how the control is working. It also depends on the exposure time you use for the lights. The shorter the exposure time, the less impact does the amp glow have, that you mainly want to correct with many ZWO cameras or the used sensors. I often shoot 1200s subs for narrowband and here you might run into problems. But I never did and everything calibrates out fine.

There is one thing I realized in your post I want to adress. You said, you put the camera outside to do your darks. Be careful to keep the sensor really dark. When I took my darks, I realized that the brightness of the daylight effects the darks of longer exposure times. The cap I used to cover the sensor was not as dense as I thought and my mean value increased with the daylight. Maybe, there was even a small light leak in the filter wheel I had attached to the camera. Anyway, you should keep an eye on your mean value to make sure, your darks are really "dark". You may not see this problem, but it won't hurt either to double check.

BTW, since this experience, i put the camera in my (dark) cellar to take the darks. It's cold enough down there to reach the -20 degrees (at least in my cellar) and I got consistent frames.

Hope this helps…

CS

Christian
Helpful Insightful Respectful Engaging Supportive
webcubus avatar
webcubus avatar
Refrigerator flats for the win! Seriously, it works great. I took the camera off my imaging train and put it in the fridge to shoot -25C darks.

+1 on QHY coolers. I know they're different weight classes, but my QHY 168C (ZWO 071MC equivalent) is able to run so much cooler and more consistently than my dad's ZWO 533MC.
Helpful Engaging Supportive
Mickaelle (kaelig) avatar
Mickaelle (kaelig) avatar
For me, you have not an issue. It's normal, it is a cooled electronic system with a control loop. You could not have a fixed value. It varies a little.
The goal of the cooled system is to decrease the noise level, a little variation of temperature could not affect it.
Guillermo (Guy) Yanez avatar
Guillermo (Guy) Yanez avatar
That variance is totally acceptable so your subs and calibration frames should be all OK for integration. That said, if a few decimals of a deg really bother you , you may try bringing up  the temperature to -10C . I think that is the sweet spot for ZWO cameras and perhaps you will get a lower variance in the temperature.
In any case, such a small number will not affect anything  in any meaningful way.
Just for the record, my DSLR camera does not have a cooling system, so I use an external temperature gauge to roughly get the same average temperature for my darks as in my imaging run (which by the way is obviously decreasing during the night). I wish I could keep the temperature to a sub-degree precision smile
Helpful Supportive
Eric Gagne avatar
Eric Gagne avatar
I took 100 darks last night at different gains and exposure times.  Only the first 5 and 2 or 3 here and there had differences.

The camera was inside at about 20 degrees.  I took it outside at about 0 and set the temperature  to -20 so I think that’s why it took some time to settle.

Next time I will leave it outside at least 30 minutes before turning it on so it will go down to ambient temperature before.
Helpful Insightful Concise
webcubus avatar
webcubus avatar
I don't see any reason to cool the camera down to ambient before you turn the cooler on. It can either cool to the temperature you set or it can't. The camera specs should tell you how much differential from ambient it can handle. I tend to try not to push it and won't run the cooler higher than 70% or so.

Scopes on the other hand, can definitely benefit from acclimating before focus and shooting.
Well Written Concise
Christian Großmann avatar
Christian Großmann avatar
I agree with @webcubus . No need to wait here...
Barrington Russell avatar
Barrington Russell avatar
As an aside: when taking darks I actually put my camera in the refrigerator overnight and shut the door. That way it's perfectly dark, and it's not recirculating hot air in some tiny box. Never trust outdoors for getting good darks, and inside a warm room you probably can't reach -20C reliably anyway.
Well Written Helpful Insightful Concise Engaging Supportive
lunohodov avatar
lunohodov avatar
Eric, you have a temperature discrepancy of slightly over half a degree Celsius. Let me ask you something.
  • How much more dark current will accumulate as a result?
  • Wouldn't a good number of darks even it out?
  • Does it really matter?


One experiment you could make is taking a dark frame at -19C and -20C, then compare (i.e. PixInsight's Statistics process or subtract one from the other with PixelMath). The 533 sensor has very little thermal noise.

I am looking forward to your findings!

Yanko
Well Written Helpful Insightful Respectful Engaging Supportive
andrea tasselli avatar
andrea tasselli avatar
The IMX533 sensor I have has a LOT more thermal noise than the older IMX294.
lunohodov avatar
lunohodov avatar
andrea tasselli:
The IMX533 sensor I have has a LOT more thermal noise than the older IMX294.

Andrea, do you mean the 533 sensor in general? How did you come to this conclusion?

The camera specifications for ASI533 Pro and ASI294 Pro ZWO has published make it clear 533's dark current is way lower. Can you elaborate more?

On the other hand, 533 is a well behaved sensor. As a result the dark current accumulates (almost?) linearly and a 0.6C difference in sub-exposure temperature should have negligible to no effect.
andrea tasselli avatar
andrea tasselli avatar
Andrea, do you mean the 533 sensor in general? How did you come to this conclusion?

The camera specifications for ASI533 Pro and ASI294 Pro ZWO has published make it clear 533's dark current is way lower. Can you elaborate more?

On the other hand, 533 is a well behaved sensor. As a result the dark current accumulates (almost?) linearly and a 0.6C difference in sub-exposure temperature should have negligible to no effect.


Since you mentioned the thermal noise (which isn't the same of the thermal current) the statistical analysis of identical in length and temperature dark frames show that the IMX294 one has -3sigma scatter of 174 pixels while the one from the IMX533 has a -3sigma scatter of 5171 pixels. Yet both vastly superior to the CCDs of yesterday.
lunohodov avatar
lunohodov avatar
Andrea, you are correct that dark current is not the same as dark current noise. Yet, the latter is inferred from the former and given identical exposure length and temperature, a lower dark current will yield a lower dark current noise, no?

Thank you!

Yanko
Well Written Respectful
Stefan Pfleger avatar
Stefan Pfleger avatar
1C variance is completely fine. Also ur cooling at -20C, the colder you cool, the less impact it has in the read noise reduction. The read noise graphs are logarithmic btw.
andrea tasselli avatar
andrea tasselli avatar
Andrea, you are correct that dark current is not the same as dark current noise. Yet, the latter is inferred from the former and given identical exposure length and temperature, a lower dark current will yield a lower dark current noise, no?

Thank you!

Yanko


Hello Yanko,

I'm afraid isn't quite this straightforward. On paper (i.e., the ZWO website) the M533MC has an order of magnitude less dark current than the ASI294MC yet the thermal current noise of the latter, as measured by me, is more than an order of magnitude smaller than the former. And that isn't just a matter of numbers, the IMX533 sensor is quite clearly noisier than the IMC294 at just a quick glance (if you ignore the amp-glow).

Andrea
Helpful Insightful Respectful
Stefan Pfleger avatar
Stefan Pfleger avatar
andrea tasselli:
Andrea, you are correct that dark current is not the same as dark current noise. Yet, the latter is inferred from the former and given identical exposure length and temperature, a lower dark current will yield a lower dark current noise, no?

Thank you!

Yanko


Hello Yanko,

I'm afraid isn't quite this straightforward. On paper (i.e., the ZWO website) the M533MC has an order of magnitude less dark current than the ASI294MC yet the thermal current noise of the latter, as measured by me, is more than an order of magnitude smaller than the former. And that isn't just a matter of numbers, the IMX533 sensor is quite clearly noisier than the IMC294 at just a quick glance (if you ignore the amp-glow).

Andrea

Hi Andrea, would you mind sharing these test results? Thats a bold statement to claim that ZWO meassures their dark current wrong.
Greetings, Stefan
andrea tasselli avatar
andrea tasselli avatar
Stefan Pfleger:
andrea tasselli:
Andrea, you are correct that dark current is not the same as dark current noise. Yet, the latter is inferred from the former and given identical exposure length and temperature, a lower dark current will yield a lower dark current noise, no?

Thank you!

Yanko


Hello Yanko,

I'm afraid isn't quite this straightforward. On paper (i.e., the ZWO website) the M533MC has an order of magnitude less dark current than the ASI294MC yet the thermal current noise of the latter, as measured by me, is more than an order of magnitude smaller than the former. And that isn't just a matter of numbers, the IMX533 sensor is quite clearly noisier than the IMC294 at just a quick glance (if you ignore the amp-glow).

Andrea

Hi Andrea, would you mind sharing these test results? Thats a bold statement to claim that ZWO meassures their dark current wrong.
Greetings, Stefan

You are confounding thermal current with thermal noise. They aren't the same.
Well Written
Stefan Pfleger avatar
Stefan Pfleger avatar
andrea tasselli:
Stefan Pfleger:
andrea tasselli:
Andrea, you are correct that dark current is not the same as dark current noise. Yet, the latter is inferred from the former and given identical exposure length and temperature, a lower dark current will yield a lower dark current noise, no?

Thank you!

Yanko


Hello Yanko,

I'm afraid isn't quite this straightforward. On paper (i.e., the ZWO website) the M533MC has an order of magnitude less dark current than the ASI294MC yet the thermal current noise of the latter, as measured by me, is more than an order of magnitude smaller than the former. And that isn't just a matter of numbers, the IMX533 sensor is quite clearly noisier than the IMC294 at just a quick glance (if you ignore the amp-glow).

Andrea

Hi Andrea, would you mind sharing these test results? Thats a bold statement to claim that ZWO meassures their dark current wrong.
Greetings, Stefan

You are confounding thermal current with thermal noise. They aren't the same.

Ok could be that im not understanding it fully what you mean rn, I‘d still like to see those test results
andrea tasselli avatar
andrea tasselli avatar


So the dark signal from the IMX533 is 1/8 that of the IMX294, roughly in line with the ZWO website numbers. However, thermal noise is well over a decibel higher. Top IMX533, bottom IMX294.