Cooling any ZWO camera - really really easy. update and opinions please on if a video would be useful?

[powerlord]

Hi chaps, I did find a few articles/videos on adding a cooler to zwo camera, but they were all a bit.. janky. either focusing too much on theory, or very rough DIY attempts with no real analysis of results or performance benefit.

So.. I thought I'd set myself a problem statement:

1. make a tutorial than even my granny could follow on how to add cooling to any ZWO camera, and must be cheap: no soldering, no fancy kit needed, not drilling and fashioning bits.

2. ideally must not cause irreverable changes to camera.

3. measure the amount of cooling achieved, and power needed

4. take darks and pictures at various pictures to actually clearly measure how temperature affects them. Bottom line - unless really good improvements in noise, whats the point ?

I've did  1, 2 and 3 yesterday. And have 4 left to do today... I'm hoping I haven't wasted my time...but doubt it.

Assuming I've not - do you think there would be much interest in an easy to follow tutorial with parts, suppliers and assembly for people ?

To give you an idea of the results I've achieved I can cool the sensor by approximately 30 degress below ambient, for 12 quid. And the whole thing can be built by my granny in about 10 mins.

stu

 

Edited July 30, 2021 by powerlord

[AstroMuni]

Definitely interested

[Priesters]

Me too 👍

[Dean Hale]

Yes please. Looking forward to no1 as my diy skills are zilch.

[vlaiv]

I'm following with interest, and must say that I'm a bit skeptical as far as expense goes.

Apart from being cooled, cameras greatly benefit from being set point temperature. I'd rather have camera at 20°C that is exactly at 20°C than camera at somewhere between -20°C and -10°C.

[powerlord]

while that's a fair point vlaiv, you ain't got a set temperature with an uncooled camera to start with - it's whatever ambient is and/or changes to through the night.

so, it's no worse from a variable POV, just colder.

If you're the sort of person that covers the camera in layers of insullation to keep it at within a narrow range, then that same insulation solution will work.

I did consider this,  but adding a temp controller into the mix adds to the complexity too much imho. the measure of success here is 'is it considerable better than it was', not 'is it as good as my actual zwo cooled camera'.

If right now, my camera varies between 15-20 degrees during the night, and with the cooler varies between -15 and -10 during the night, bottom line is (I hope) noise will still be a lot less. Your flats/bias/darks will have the same issues you already have (which temp do you choose). But they will at least have a base level of noise that's a damn site lower.

stu

[vlaiv]

4 hours ago, powerlord said:

I did consider this,  but adding a temp controller into the mix adds to the complexity too much imho. the measure of success here is 'is it considerable better than it was', not 'is it as good as my actual zwo cooled camera'.

If right now, my camera varies between 15-20 degrees during the night, and with the cooler varies between -15 and -10 during the night, bottom line is (I hope) noise will still be a lot less. Your flats/bias/darks will have the same issues you already have (which temp do you choose). But they will at least have a base level of noise that's a damn site lower.

That's what I meant. For me "considerably better than it was" is set point temperature and not coooler camera.

Ability to properly calibrate data trumps dark current.

Here, look at ASI183 dark current vs temperature.

Dark current for 25°C is less than 0.0625 e/px/s.

Say you take 300s exposures - which is very long for uncooled camera - you'll accumulate only 18.75e of dark current and associated noise will be 4.33e.

In comparison, read noise of this camera on low gain is 3e and since it has very small pixels and one will probably bin at least x2, even if one uses unity gain (2.2e read noise ) - total read noise for bin x2 climbs to 4.4e.

This means that in 5 minute exposure at 25°C - dark current noise is same / comparable with read noise (and both are much lower than read noise of cooled CCD camera that usually has about 8e of read noise).

See the point? Dark current noise is not the main issue - main issue is ability to properly calibrate data. Odd hot pixel is dealt with by dithering and sigma stacking - in the same way one deals with cosmic hits and satellite trails.

[david_taurus83]

Show and tell!

[Dr_Ju_ju]

What are you actually cooling ??  the body or the sensor, which ideally is what is required to be cooled ....

[powerlord]

1 hour ago, vlaiv said:

That's what I meant. For me "considerably better than it was" is set point temperature and not coooler camera.

Ability to properly calibrate data trumps dark current.

Here, look at ASI183 dark current vs temperature.

Dark current for 25°C is less than 0.0625 e/px/s.

Say you take 300s exposures - which is very long for uncooled camera - you'll accumulate only 18.75e of dark current and associated noise will be 4.33e.

In comparison, read noise of this camera on low gain is 3e and since it has very small pixels and one will probably bin at least x2, even if one uses unity gain (2.2e read noise ) - total read noise for bin x2 climbs to 4.4e.

This means that in 5 minute exposure at 25°C - dark current noise is same / comparable with read noise (and both are much lower than read noise of cooled CCD camera that usually has about 8e of read noise).

See the point? Dark current noise is not the main issue - main issue is ability to properly calibrate data. Odd hot pixel is dealt with by dithering and sigma stacking - in the same way one deals with cosmic hits and satellite trails.

Huh. So why is this different for cooled cameras then ? I mean the difference when I take a shot with my asi1600 at 20C vs -20C is night and day on that - vastly less noise ?

I assumed it would be the same with most sensors ? I mean folk do this for DSLRs too for example ?

Mine as an asi224. ZWO themselves used to sell  a cooled version - so there I assume is a good reason ? If the point was just to keep it at the same temperature, casing it in an insulating material rather than a block of aluminium would surely have been a much cheaper solution to that problem ? They themselves seem to think there is a good reason to cooling it to 30C below ambient ?

what am I missing ? As I've done it now, I might as well take the practical approach and shoot some shots at different temps and see I think.

 

 

 

Edited July 30, 2021 by powerlord

[bottletopburly]

Sounds good to me , if darks are within a few degrees of lights temperature  are you really going to notice, As long as  temperature doesn’t fluctuate vastly then cooling your way should give you a better result than an uncooled dslr     I’m sure the results will be satisfactory for 99% of people , look forward to seeing  your results .

[powerlord]

well here's some darks at various gains and temps. and one light. let me know if you think its really worthwhile or not. If not, not a lot of point in it.

light is a bit rubbish, as I had to sort of make a dark place under my desk, and use the allsky lens - but as thats exposed it is of course, going to get condensation, etc - but anyway - that's the data

captures.zip

[powerlord]

sorry, I kinda screwed a few of the names up there, and my 300+ ones are pretty bad as just had lens cap on.

So here's the best 2 - 100s, 135 gain. 2C vs 29C

Now, looking at that histogram, to me that looks like quite an improvement - 29C lots of noise coming up to around 1500 or so. 2C - hardly any just amp glow.

 

 

Now to

[vlaiv]

3 hours ago, powerlord said:

Huh. So why is this different for cooled cameras then ? I mean the difference when I take a shot with my asi1600 at 20C vs -20C is night and day on that - vastly less noise ?

It is not different. Did you actually measure noise levels in 20°C vs -20°C?

Looking at dark frame is not proper way to measure how much noise there is due to dark current.

In order to measure how much noise there is due to dark current, you first need to determine read noise and e/ADU for given gain - so you can use proper numbers and subtract read noise. Then you take 2 darks at -20°C and two darks at +20°C.

You subtract two darks at -20°C. This removes dark current and bias signal (things that are not noise but rather fixed signal). You then convert ADU to electrons and measure standard deviation of resulting image. You then remove read noise and what is left is dark current noise.

Do the same for 20°C and compare the two.

By inspecting visually or statistically 20°C vs -20°C subs - you might get the impression that 20°C are more noisy - but that is not necessarily the case. A lot of hot pixels is very very tiny percentage of whole sensor and will get calibrated out. If you dither you won't have issues with those hot pixels - no more than say satellite trails or cosmic rays.

Issue with DSLR is the same as with any other camera that does not have set point cooling - problem of calibration. If you can't properly calibrate your image - you can't remove respective signal and that then starts acting as noise (not really random in mathematical sense - but poor looking in image).

There is a reason why we have diagrams like I linked above - one that shows dark current vs temperature. It shows you how serious dark current is so you can compare it to other things and decide if is something you need to concern yourself with. It also shows that dark current is not as much of an issue as one might think.

[Astro Noodles]

Oh come on Vlaiv, stop being a spoil sport.😄

I don't think it's possible for those using un-cooled cameras to accurately tell what their sensor temperatures are anyway. Surely, cooler is less noisy than warmer, otherwise why bother having cooled cameras at all. I can see that you are saying that it would be better to have accurate temperature control than just cooling, but maybe that is another project ?

 

 

[powerlord]

OK, I didn't and don't want this to get into a theoretical argument about whats and whys Vlaiv. You clearly know what you are talking about - my take on this is much more simplistic:

I thought: "camera can be cooled and uncooled. cooled better." can I cool it cheap. is it worthwhile ? that's it. I'm not building a new camera from scratch, or doing some fancy thing with 2nd order temperature feedback. It's a 12-20 quid project for a camera. that's it. Not an attempt at creating a world leading astrophotography camera...

you say 'For me "considerably better than it was" is set point temperature and not coooler camera. " - that's fine - go build one of them.

That wasn't the project though  - it was off shelf zwo vs same zwo cheaply cooled.

The result above seem pretty obvious to me - a faint star in that 29C one above could be lost in the crap* you see in the histogram. It would not be lost in the 2C historgram. Why make things more complicated than they need to be ?

Looking at pictures is why I take photographs, I don't take them to analyse that they have the right amount of this and that - it just isn't where I get my jollys.

You've still not answered why cooled cameras are a good thing? They're not 'set temperature cameras' - they are cooled cameras.

Unless you really don't think they are good ?

When I got my asi1600 I focused on a nearby tree and leaves. I saw the crap. I turned the cooling on. and I watched the crap disappear. end of. Could I have got the same result by taking bias, darks, subtacting and stacking them up the wazzo - I don't know. But more importantly I don't care. I just used my eyes. -20 was a hell of a lot less crap.

But look, this topic has got side tracked. It really is just about trying to cool a camera down. Not analyse theory. It is not going to be for you, I get that.

I must admit, I didn't expect this to be controvertial 😪

I'm not sure this is worth doing now whereas I thought it was before...

*im using crap as I think you get fussy about what definitions are, where as to me, I don't care - its stuff thats not there.

[Ags]

@powerlord I am very interested in cooling my ASI178MM and I would be intrigued to to see how you did it. 

@vlaiv I don't think there is much point getting too theoretical about this - a camera is fantastically complex and there is plenty of room for confounding variables not accounted for in the unadorned mathematics. If it costs 20 euros to do this, I would rather just do it and see the results for myself. Worst case scenario: I have done a fun project and learned a bit about my camera. Best case scenario? APODs for all of us 😀

Edited July 30, 2021 by Ags

[pete_l]

11 hours ago, powerlord said:

To give you an idea of the results I've achieved I can cool the sensor by approximately 30 degress below ambient, for 12 quid. And the whole thing can be built by my granny in about 10 mins.

Excellent news. Please send your granny round.

[Ags]

11 hours ago, powerlord said:

1. make a tutorial than even my granny could follow on how to add cooling to any ZWO camera, and must be cheap: no soldering, no fancy kit needed, not drilling and fashioning bits.

 

Although a vid would be fun to watch a written tutorial is much better in my opinion (ok, I am a technical writer...). 

[JamesF]

1 hour ago, pete_l said:

Please send your granny round.

Unless she has big ears and big teeth.

James

[souls33k3r]

So looking forward to a tutorial. I'm a video kinda of a guy and some here like to see it in a written form. I guess both would be awesome 😁

[happy-kat]

I'm keen to see the idea was well please as this might be fun to try on the neximage 5 which is great on the Moon but very noisy when pushing the exposure length.

[JamesF]

4 minutes ago, happy-kat said:

I'm keen to see the idea was well please as this might be fun to try on the neximage 5 which is great on the Moon but very noisy when pushing the exposure length.

That's very true.  The ZWO cameras do have a metal case however, unlike the Neximage models, which is probably quite helpful from a DIY cooling point of view.  It may well be that some of the ideas would transfer over though, even if exactly the same method wasn't possible.

James

[vlaiv]

1 hour ago, powerlord said:

You've still not answered why cooled cameras are a good thing? They're not 'set temperature cameras' - they are cooled cameras.

But they are. All cooled cameras are in fact set point temperature cameras. You set your temperature and cooling lowers the temperature and keeps it there.

That is why you can build dark library at any time and reuse it later - you tell your cooled camera - go to -10°C or maybe even tell it to go to +10°C and it will maintain that temperature.

1 hour ago, powerlord said:

The result above seem pretty obvious to me - a faint star in that 29C one above could be lost in the crap* you see in the histogram. It would not be lost in the 2C historgram. Why make things more complicated than they need to be ?

Just try simple thing - take two dark subs with cooling off - and subtract the results. Histogram will be very nice looking and not at all the way you see it now - that is the point of calibration in the first place.

@ At all participating the thread

To be honest I don't see simple calibration to be too technical. We do it every time we shoot image. It is basic understanding of how camera works and not some crazy convoluted science.

I just wanted to point out what sort of benefits this mod will provide. If it does not provide set point cooling - well, it's just like imaging in winter. Do you see great difference when imaging in winter with non cooled cameras? They can easily end up being in single temperature digits if outside temperature is around zero or maybe slightly below - I'm sure most of us imaged in those conditions.

[Shimrod]

18 minutes ago, vlaiv said:

But they are. All cooled cameras are in fact set point temperature cameras. You set your temperature and cooling lowers the temperature and keeps it there.

That is why you can build dark library at any time and reuse it later - you tell your cooled camera - go to -10°C or maybe even tell it to go to +10°C and it will maintain that temperature.

 

I think there is a confusion between fan cooled cameras (like those offered in the lower range of Altair cameras) and the TEC cooled cameras (such as ZWO and higher-end Altair). The former offer an advance over uncooled cameras but you cannot set and maintain a temperature like the TEC cameras. I assume @powerlord is building the former, although with a 20 degree difference from ambient, possibly with slightly more effective cooling.

Can we agree TEC is better than cooled, and cooled is better than uncooled?