Sensor temperature and calibration

[Rodd]

Its always something, clouds, rain, wind, LP, work (heaven forbid), myriad equipment troubles,  and one that doesn't raise its head all that often (at least for me before this year)....heat.  A couple years ago I lowered my cooling set point from -25 to -20 thinking that it would allow me to always achieve the desired sensor temperature all year so I could use the same master dark. At that time it was essential because I was taking 30 min subs with the CCD and spending 10 hours taking dark frames was annoying.   It worked for two summers.  Not this summer.  the last few sessions I have not been able to achieve better than -17 or -18.   A couple of nights ago it happened..and the seeing was good, and it was clear.  I certainly did not want to shoot new darks and lose out on clear sky time.  So I collected the data and calibrated the subs using my -20 master dark (and my -20 degree master flat).  Then I was afforded the opportunity to run an experiment.  Clouds rolled in around 1:00, allowing me to collect new -18 darks.  collecting 300 sec darks is much less time consuming than 1800 sec darks.  Here are the results.  The first image is the blue stack collected with a sensor temperature of -17 and -18 (mostly 18) calibrated with -20 degree masters.   The second image is the same blue stack but the subs were calibrated using -18 degree masters (the appropriate calibration).  I don't know about you, but I can't see an appreciable difference.  PI's Sub Frame Selector yields readings for FWHM, Noise, Median and SNR the same to two decimal places and the third decimal place is different by a couple numbers.  A difference too small to be meaningful--probably not statistically relevant.

I am posting this in case folks are fretting over cooling accuracy to the point of forgoing collecting data.  I was surprised at these results.  Maybe everyone knows this and I am expounding unnecessarily...if that is the case....here is my next project--the blue stack....I am fairly pleased.

TOA 130 with .7x reducer and ASI 1600

Blue 40 300 sec

The data has been calibrated, cropped, gradients removed, and screen stretched (the reason the median value is so high...that and my sky is not very dark).  I looked at the images using manual stretches as well, but any differences due to temperature correlations disappeared completely under a less aggressive stretch (not that there are any, but if there were, they would be harder to spot with a darker background).

-18 degree data calibrated with -20 degree masters

 

-18 degree data calibrated with -18 degree masters

[Kinch]

6 minutes ago, Rodd said:

I am posting this in case folks are fretting over cooling accuracy to the point of forgoing collecting data.

Thanks Rodd;

I have a camera with extremely good cooling - but I too struggle here in SE Spain with my chosen working CCD temperature. When starting out as soon as possible after dark it is stilll very warm from the day. Looking at my 'delta T', I know I can get to maybe 1° or 2° above what I want (without labouring the TEC too much....though in truth, still a bit more than I would prefer).  If the outside air cools down at all, I will try to grab back that 1° or 2­°.....cooling down to the temp the Darks were done at. Most nights I can reach what I want at some stage during the night or early morning but last night was the 1st night I had to leave the TEC set at 2° above what I wanted. I did not fret too much about this as in the end there will be a lot of subs and hopefully most will get to or at least within 1° of what I want.

On top of my own reckoning (not tested) I am happy to see what you have written above.....I feel confident I can continue working the camera and continue to use my -30°C Darks in up to 25°C to 27°C night temperatures we are getting here now.

[Rodd]

5 minutes ago, Kinch said:

I have a camera with extremely good cooling

Despite the results--I now question the cooling effectiveness of the ASI 1600.  The outside temperature was 23-25 C and the camera could not achieve -20.  I think that is pretty weak, and I would say another downside of this camera.  There are 2 now--the micro lensing pattern at longer focal lengths and now this cooling issue.  I wish info like this would be more available when researching a product for purchase.

[vlaiv]

5 minutes ago, Rodd said:

I wish info like this would be more available when researching a product for purchase.

Well it does say that max delta T is 45C so you can't really expect it to stably hold -20C if outside temperature is 25C. If you want stable cooling - go with 40C delta T - or subtract 40C from outside temperature and go for that one.

[Rodd]

4 minutes ago, vlaiv said:

Well it does say that max delta T is 45C so you can't really expect it to stably hold -20C if outside temperature is 25C. If you want stable cooling - go with 40C delta T - or subtract 40C from outside temperature and go for that one.

Confused.   My camera was set at -20....it could not reach that....only -18.    the camera is supposed to be able to achieve -45.  If I subtract 40 from 73 I get 33.  So not sure what you are talking about.  

[vlaiv]

5 minutes ago, Rodd said:

Confused.   My camera was set at -20....it could not reach that....only -18.    the camera is supposed to be able to achieve -45.  If I subtract 40 from 73 I get 33.  So not sure what you are talking about.  

No - it is not able to reach -45C, it is able to reach at max 45C below ambient temperature.

Cooling specs are never given in absolute temperature but rather temperature difference - how much cooler than ambient it can achieve. On a cold winter night when ambient temperature is -5C it will cool down to -50C but on a warm summer night with ambient temperature of 26C it will only go as low as -19C. This is of course maximum but in reality it will be 1-2C less than that due to various factors like relative humidity, how dusty heat sink is and if ventilator is well lubricated (viscosity of lubricant changes with temperature as well so fan won't work at 100% in all conditions).

[Kinch]

On the test results for my FLI ML16200 it gave a measured coooling to  dT of -56.3°. I don't know how long that test lasted for, though.

I have aimed to use -55°....getting to -30°C in ambient 25°C . This I was hoping would be my worst case, but as I said, temps have been up to 27°C at night lately. This was my reason to start shooting at -28° rather than -30°, TEC running at 90% to 92%......a bit higher than I would really like if truth be known.  I have got dT of 57° at 93%-94%....but I really don't like pushing it that hard (for any length of time).....hence shooting at 1° or 2° above the calibration frames temp. 

Your post has convinced me that I can continue doing that (although I suspected that anyway) and not be bothered with a new set of Darks.

Edited July 27, 2020 by Kinch

[Rodd]

8 minutes ago, vlaiv said:

No - it is not able to reach -45C, it is able to reach at max 45C below ambient temperature.

Cooling specs are never given in absolute temperature but rather temperature difference - how much cooler than ambient it can achieve. On a cold winter night when ambient temperature is -5C it will cool down to -50C but on a warm summer night with ambient temperature of 26C it will only go as low as -19C. This is of course maximum but in reality it will be 1-2C less than that due to various factors like relative humidity, how dusty heat sink is and if ventilator is well lubricated (viscosity of lubricant changes with temperature as well so fan won't work at 100% in all conditions).

I know.  The camera can achieve delta T of -45.  In my case I had it set for a Delta T of -20 and it couldn't reach that.  That is a delta T of -25 off.   I think that is exceptionally weak.

[Rodd]

11 minutes ago, Kinch said:

(although I suspected that anyway) and not be bothered with a new set of Darks

The literature says that darks should be the same temp.  PI makes a big deal about it when scaling is employed.  I have heard it say that anything over 1degree will result in poor calibration.   Maybe I was naive for believing this.  On the other hand--I didn't lose clear sky time running the experiment, so no harm done.

[vlaiv]

17 minutes ago, Rodd said:

I know.  The camera can achieve delta T of -45.  In my case I had it set for a Delta T of -20 and it couldn't reach that.  That is a delta T of -25 off.   I think that is exceptionally weak.

No, you are not reading it correctly.

It was 25C ambient temperature. You set your camera to -20C. There is total of 45C between those two points of temperature scale. There is 25C from 25C down to 0C and another 20C when going from 0C down to -20C. That is total of 45C difference between 25C and -20C. You set your temperature as absolute value (software has no idea what is your ambient temperature and does not care) - in your case -20C and camera cooling system will happily reach it if ambient is less than 25C - for example 23C. But if ambient is 25C or 26C - it will struggle to go all the way down to -20C as it can only lower temperature by 45C max - if ambient temperature is 26C it will reach -19C at best (because difference between the two is 45C).

14 minutes ago, Rodd said:

The literature says that darks should be the same temp.  PI makes a big deal about it when scaling is employed.  I have heard it say that anything over 1degree will result in poor calibration.   Maybe I was naive for believing this.  On the other hand--I didn't lose clear sky time running the experiment, so no harm done.

Using mismatched darks can cause problems with flat calibration and show slight amp glow if camera suffers from it.

Sensor dark current doubling temperature is about 6C, so using 2C different darks will cause about x1.3 x1.26 stronger dark current.

This may be very little of dark current change or it can be a lot - depends on sensor.

ASI1600 has about 0.0062e/px/s dark current at -20C

This means that it will have total of 1.86e of dark current for 300s exposure. At -18C that number will be 2.3436e or difference of 0.4836e - less than one electron.

You won't see this if your sky background is higher than that, and I suspect that your sky background is significantly higher than that.

If you are imaging from SQM 21.5 skies (Bortle 2/3 skies - so rather dark), in single 300s exposure with your setup, sky levels will be around 50e - so about x100 stronger than difference in respective dark currents (one at -20C and one at -18C) - you won't be able to see that on the image as difference in brightness.

Where it will show however - it will show in numbers, it will show if you have amp glow and you remove background and stretch your data, or it will show if you have significant vignetting / dust - it can lead to slight problems with flat correction.

Less dark current camera has in the first place and with shorter exposures - temperature difference between lights and darks will cause less issues for image.

Edited July 27, 2020 by vlaiv

[Kinch]

6 minutes ago, Rodd said:

I have heard it say that anything over 1degree will result in poor calibration.

I have not 'heard' that but more or less just figured it.....the farther you go from the calibration files temp.....obviously the more useless they become. I reckoned I would get away with 1° for some and perhaps 2° for a few in a big stack. I would not try more than 2°......and not too many out that far either.

Just looking at your earlier post:    "A couple years ago I lowered my cooling set point from -25 to -20". To me (and vlaiv, I think)....this means that you set your CCD to cool to -20°C. If it is 25°C outside.....this means you are looking for delta T of 45°

 

OOPS....just see vlaiv's reply come in....I won't continue....

[Rodd]

5 minutes ago, vlaiv said:

It was 25C ambient temperature. You set your camera to -20C. There is total of 45C between those two points of temperature scale. There is 25C from 25C down to 0C and another 20C when going from 0C down to -20C. That is total of 45C difference between 25C and -20C. You set your temperature as absolute value (software has no idea what is your ambient temperature and does not care) - in your case -20C and camera cooling system will happily reach it if ambient is less than 25C - for example 23C. But if ambient is 25C or 26C - it will struggle to go all the way down to -20C as it can only lower temperature by 45C max - if ambient temperature is 26C it will reach -19C at best (because difference between the two is 45C).

  17 minutes ago, Rodd said:

Well my focus controller has a sensor to measure outside temps--why not a camera?    But more important to my point....Brendan (Kinch) can reach 53.6 degrees below ambient.  And I think some cooling systems are better than others even if they purport to cool to the same Delta T.   For example, the cooling on my STT-8300 did not have this summer difficulty.  And on that camera I had it set year round to (either -25 or -30--can't recall).

6 minutes ago, vlaiv said:

Using mismatched darks can cause problems with flat calibration and show slight amp glow if camera suffers from it.

This is why I said calibrated at -20--the flats were 2 degrees off as well (I had shot them a couple of days before when my camera could reach -20.

[Rodd]

11 minutes ago, Kinch said:

I have not 'heard' that but more or less just figured it

I guess I didn't hear it either...I read it.   I don't hear much as I never talk to anyone about this stuff--always type and read.   But PI makes a big deal out of either shooting darks so within 1 degree, or scaling the darks (scaling doesn't work well for me--PI always throws up errors for no correlation)

[vlaiv]

6 minutes ago, Rodd said:

For example, the cooling on my STT-8300 did not have this summer difficulty.  And on that camera I had it set year round to (either -25 or -30--can't recall).

Indeed - STT8300 has deltaT of 55C so it will reach 55C below ambient temperature. It also costs x3 compared to ASI1600 - some of the cost I guess went for better cooling - maybe two stage Peltier instead of one stage.

On the other hand Kaf8300 in STT8300 has 0.15e/px/s at 0C.  If we take that doubling temperature is 6C - that means something like 0.015e/px/s or about double that of ASI1600 at 0.0062e/px/s (both at -20C for comparison).

Double dark current means that you need to go with 6C lower temperature in order to have same dark current as ASI1600 - it needs better cooling in the first place!

Edited July 27, 2020 by vlaiv
added clarification ...

[Rodd]

2 minutes ago, vlaiv said:

maybe two stage Peltier instead of one stage.

Does the ASI1600 only have 1 stage peltier cooling?  THAT was not disclosed.  I do remember the STT-8330 reporting 2 stage now that you mention it. 

[vlaiv]

4 minutes ago, Rodd said:

Does the ASI1600 only have 1 stage peltier cooling?  THAT was not disclosed.  I do remember the STT-8330 reporting 2 stage now that you mention it. 

It turns out that ASI1600 also has 2 stage Peltier cooler according to their website:

I guess they just used smaller / less powerful Peltier elements or maybe sensor is running hotter than CCD - since CMOS sensors have A/D stage at every pixel unlike CCD that has A/D in separate electronics. More transistors - more heat?

In any case, I think that cooling system on ASI1600 is far from useless. It is very good indeed for what it's supposed to do. As you've seen even in 300s exposure if you cool it at -15C rather than -20C - noise associated with dark current will still be lower than read noise of that camera and much much less than light pollution noise for most of us (except lucky few that image from Bortle 1/2 skies).

Camera cooling nowadays is meant to provide stable temperature for subs so calibration can work without too much hassle. Thermal noise is by far smallest noise component and usually not an issue.

[Rodd]

4 minutes ago, vlaiv said:

since CMOS sensors have A/D stage at every pixel unlike CCD that has A/D in separate electronics. More transistors - more heat?

No idea...I guess learning what other CMOS cameras can do would be good.  If this is a common issue, than you might have a point.

 

5 minutes ago, vlaiv said:

In any case, I think that cooling system on ASI1600 is far from useless.

The new engineering standard! 

I agree....but it does seem weak.  The real question is the other night when it was 25 degrees C, how many camera's were able to achieve -20C. 

9 minutes ago, vlaiv said:

Camera cooling nowadays is meant to provide stable temperature for subs so calibration can work without too much hassle. Thermal noise is by far smallest noise component and usually not an issue

So much for my post.  Lets go with the "I am pleased so far" approach (no likes though, so I guess I am in for a trouncing when I post the image for real).  However, there is an ancillary test that was conducted.......I collected the above data running the cooling at 100%.  So, for though who fear this will impact the data....not so.  (it was only a few hours and I do not plan on doing it again, so I don't think the camera was damaged).

[vlaiv]

I've adopted strategy (not that it's doing me any good - I've not managed a single imaging session this year) - of using summer and winter darks. Former are at -15C and later are at -20C, because most of the time, those are figures I'm relatively certain I can reach with ASI1600 in my ambient conditions.

DeltaT of 45C is quite decent value really. I've seen cameras with less than that regularly. Most Atik models are between 35C and 25C deltaT for example.

QHY model with this sensor - QHY 163 has deltaT at 40C for example.

In fact, now that I did a quick survey of cooling performance across different manufacturers and models - only CCDs from a few (rather expensive) manufacturers offer more than 45C deltaT cooling solutions.

 

[Rodd]

3 minutes ago, vlaiv said:

I've not managed a single imaging session this year)

That is a show stealer....very distressing, and not, I think due solely to temperature?.  I suffer with-drawls after a week.  I couldn't imagine 7 months.  I hope you manage to get back at it.

[vlaiv]

5 minutes ago, Rodd said:

That is a show stealer....very distressing, and not, I think due solely to temperature?.  I suffer with-drawls after a week.  I couldn't imagine 7 months.  I hope you manage to get back at it.

Well, I was supposed to have moved to a new house by now together with obsy under dark(er) skies and all - but you know, this year has been tough on all I guess.

Build of my new house not started yet due to all this covid stuff - just some ground preparation works done. I took another project to be able to fund all of that so there is also lack of time for astronomy ... On top of all of that, I'm now in my 10th day of covid infection - luckily both wife and I have rather mild cases, still plenty of rest is needed (and self imposed quarantine as well).

 

[Rodd]

4 minutes ago, vlaiv said:

Well, I was supposed to have moved to a new house by now together with obsy under dark(er) skies and all - but you know, this year has been tough on all I guess.

Build of my new house not started yet due to all this covid stuff - just some ground preparation works done. I took another project to be able to fund all of that so there is also lack of time for astronomy ... On top of all of that, I'm now in my 10th day of covid infection - luckily both wife and I have rather mild cases, still plenty of rest is needed (and self imposed quarantine as well).

 

Sorry to hear that.....indeed, rest and recuperate.  Crazy times.

[Kinch]

1 hour ago, vlaiv said:

On top of all of that, I'm now in my 10th day of covid infection - luckily both wife and I have rather mild cases, still plenty of rest is needed (and self imposed quarantine as well).

What a bummer - our thoughts are with you and your wife. Hopefully the mild symptoms will remain as mild symptoms and that all this will just be a bad memory before too long. 🤞

[ollypenrice]

Hey Vlad, very best wishes from the south east of France. You are a giant in this community. Take care, you and your wife.

Olly

[CraigT82]

Sorry to hear this, hope you @vlaiv and your wife get well soon... all the best from Bristol. 

[knobby]

2 hours ago, vlaiv said:

Well, I was supposed to have moved to a new house by now together with obsy under dark(er) skies and all - but you know, this year has been tough on all I guess.

Build of my new house not started yet due to all this covid stuff - just some ground preparation works done. I took another project to be able to fund all of that so there is also lack of time for astronomy ... On top of all of that, I'm now in my 10th day of covid infection - luckily both wife and I have rather mild cases, still plenty of rest is needed (and self imposed quarantine as well).

 

So sorry ... @vlaivSending best wishes and virtual hugs from the UK to you and your wife