Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

How to minimize readout noise in a CCD camera?

  • Please log in to reply
19 replies to this topic

#1 EdM2

EdM2

    Mariner 2

  • *****
  • topic starter
  • Posts: 239
  • Joined: 08 Mar 2011

Posted 24 November 2021 - 06:55 PM

The CCD camera I am buying says the following:

 

Typical_Readout Noise 9 e- at 500 kHz; 14 e- at 2 MHz

 

I want to minimize the Readout noise, but how can I control whether the camera runs at 500 kHz or 2 MHz?   Is there a way to adjust this?  Is this something I can adjust, or something that the manufacturer must adjust?

 

Is there a tradeoff in lowering the camera from 2 MHz to 500 kHz?  The positive effect would be to minimize the readout noise, but what is the tradeoff for doing that?


Edited by EdM2, 24 November 2021 - 07:27 PM.


#2 GaryShaw

GaryShaw

    Apollo

  • *****
  • Posts: 1,484
  • Joined: 28 Nov 2017
  • Loc: Massachusetts / Wyoming

Posted 25 November 2021 - 09:23 AM

Hi Ed,

Sorry not to have an answer to your specific question - but I’m just curious what factors decided you on a CCD for EAA rather than the low noise, lower cost CMOS cameras that dominate the EAA community?

 

I’ve been an EAA observer for just over three years and never encountered mention of or discussions of ccd cameras until I began training as a variable star photometrist. I considered buying a ccd at first since all the training material seemed based on ccd cameras and that’s what most variable star photometrists seem to have. When I realized the ccd world seemed to be transitioning to the cmos cameras and that the instructional materials for photometry are applicable to either technology, I stuck with cmos and acquired a cooled version of a cmos camera I already have. 
 

Im interested though to hear your thoughts on the pluses of CCD for EAA.

Cheers 
Gary


  • Noah4x4 likes this

#3 nic35

nic35

    Apollo

  • *****
  • Posts: 1,268
  • Joined: 08 Sep 2007

Posted 25 November 2021 - 09:34 AM

You might get a better response over in the experienced deep sky imaging forum.  I suspect there are more advanced CCD users over there. 

 

From basic principles, the spec is saying that at lower sampling rates, the read noise is lower.  Which makes sense, as it is reading the "cell" less frequently.  But it means the accuracy of the measurement is less.  But at these rates, I'm not sure that's material - it would be a function of the relative accuracy of the individual reads.  Playing very loose here with statistical concepts.

 

Quite a pricey camera.  But then you have setup to die for.

 

Its only a guess, but I'm presuming you are most interested in conventional astrophotography, and not the casual sort of AP that this forum is about. 

 

john



#4 Noah4x4

Noah4x4

    Aurora

  • *****
  • Posts: 4,650
  • Joined: 07 Apr 2016
  • Loc: Colchester UK

Posted 25 November 2021 - 10:18 AM

Atik manufacture both CCD and CMOS and produce some super technical information about each at https://www.atik-cam...r-technologies/, including discussion about the merits of class 1 CCD versus Class 2 CCD sensors, which I suspect is probably more material than the OPs question.

 

However, I agree with Gary and John, that this isn't a routine debate for EAA enthusiasts within this forum as we do tend to pursue a more "casual form of AP" (I like that definition John) for which a cheaper CMOS sensor is perfectly satisfactory. Top end CMOS will produce incredible results. So, unless you want to pay top dollar then I don't see the merit in CCD except for the absolute purest form of long exposure AP when striving to win Astrophotographer of the year.  But Atik offer a more objective opinion as they manufacture both (unlike many other popular makes that have abandoned CCD).



#5 alphatripleplus

alphatripleplus

    World Controller

  • *****
  • Moderators
  • Posts: 126,701
  • Joined: 09 Mar 2012
  • Loc: Georgia

Posted 25 November 2021 - 11:08 AM

Moving this from EAA to the Experienced Deep Sky Imaging Forum for a potentially better fit.



#6 Jared

Jared

    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 7,018
  • Joined: 11 Oct 2005
  • Loc: Piedmont, California, U.S.

Posted 25 November 2021 - 11:35 AM

The lower the readout speed of the CCD camera, the lower the rear noise. Some camera drivers let you choose, others are just “set” to the lowest rate. You’ll have to look in your driver settings to see what is offered.

As far as trade offs, obviously the lower read noise is good. However, a slower readout can lower the frame rate quite a bit. My STL11000, for example, takes about 30s to read out a frame. That’s a lot. It’s enough that you will lose some efficiency in terms of data collection which might matter as much as the read noise (depending on focal ratio, light pollution, and broadband vs narrowband filter). You’d have to do the math. Auto focus takes a while with a slow readout (though using sub frames and binning can help). Flats can be challenging if you are taking sky flats since conditions can change a lot in the thirty or more minutes it takes. You might even notice a gradient in your images from dark current—part of the chip will spend more time collecting dark current than other parts due to the readout time.

Not many cameras are as slow as my STL11000. If the difference between fast and slow modes is on the order of a couple seconds or less, just choose the slower clock rate, if it’s long enough to be cumbersome in terms of focus and overall efficiency, you’ll have to do the math, but you might still be better off with the lower clock rate. For narrowband, you are almost certainly better off with the slowest readout.
  • EdM2 likes this

#7 EdM2

EdM2

    Mariner 2

  • *****
  • topic starter
  • Posts: 239
  • Joined: 08 Mar 2011

Posted 25 November 2021 - 12:02 PM

The lower the readout speed of the CCD camera, the lower the rear noise. Some camera drivers let you choose, others are just “set” to the lowest rate. You’ll have to look in your driver settings to see what is offered.

As far as trade offs, obviously the lower read noise is good. However, a slower readout can lower the frame rate quite a bit. My STL11000, for example, takes about 30s to read out a frame. That’s a lot. It’s enough that you will lose some efficiency in terms of data collection which might matter as much as the read noise (depending on focal ratio, light pollution, and broadband vs narrowband filter). You’d have to do the math. Auto focus takes a while with a slow readout (though using sub frames and binning can help). Flats can be challenging if you are taking sky flats since conditions can change a lot in the thirty or more minutes it takes. You might even notice a gradient in your images from dark current—part of the chip will spend more time collecting dark current than other parts due to the readout time.

Not many cameras are as slow as my STL11000. If the difference between fast and slow modes is on the order of a couple seconds or less, just choose the slower clock rate, if it’s long enough to be cumbersome in terms of focus and overall efficiency, you’ll have to do the math, but you might still be better off with the lower clock rate. For narrowband, you are almost certainly better off with the slowest readout.

Thank you.  It is beginning to make sense.  I hope the driver provides a function for readout speed.  I will check with the manufacturer.



#8 EdM2

EdM2

    Mariner 2

  • *****
  • topic starter
  • Posts: 239
  • Joined: 08 Mar 2011

Posted 25 November 2021 - 12:03 PM

Atik manufacture both CCD and CMOS and produce some super technical information about each at https://www.atik-cam...r-technologies/, including discussion about the merits of class 1 CCD versus Class 2 CCD sensors, which I suspect is probably more material than the OPs question.

 

However, I agree with Gary and John, that this isn't a routine debate for EAA enthusiasts within this forum as we do tend to pursue a more "casual form of AP" (I like that definition John) for which a cheaper CMOS sensor is perfectly satisfactory. Top end CMOS will produce incredible results. So, unless you want to pay top dollar then I don't see the merit in CCD except for the absolute purest form of long exposure AP when striving to win Astrophotographer of the year.  But Atik offer a more objective opinion as they manufacture both (unlike many other popular makes that have abandoned CCD).

CMOS might have been a better choice.  I chose CCD because I am familiar with it.  



#9 Jared

Jared

    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 7,018
  • Joined: 11 Oct 2005
  • Loc: Piedmont, California, U.S.

Posted 25 November 2021 - 12:03 PM

It probably does or they wouldn’t quote both levels of readout noise.

#10 Jared

Jared

    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 7,018
  • Joined: 11 Oct 2005
  • Loc: Piedmont, California, U.S.

Posted 25 November 2021 - 12:05 PM

CMOS might have been a better choice. I chose CCD because I am familiar with it.


Nothing wrong with CCD. Even 8e- isn’t a big deal for broadband. It matters a bit more for narrowband, of course, where you’ll need a lot more data and longer subs for shot noise to dominate over read noise.

#11 EdM2

EdM2

    Mariner 2

  • *****
  • topic starter
  • Posts: 239
  • Joined: 08 Mar 2011

Posted 25 November 2021 - 12:22 PM

Moving this from EAA to the Experienced Deep Sky Imaging Forum for a potentially better fit.

Thanks.



#12 EdM2

EdM2

    Mariner 2

  • *****
  • topic starter
  • Posts: 239
  • Joined: 08 Mar 2011

Posted 25 November 2021 - 01:35 PM

Nothing wrong with CCD. Even 8e- isn’t a big deal for broadband. It matters a bit more for narrowband, of course, where you’ll need a lot more data and longer subs for shot noise to dominate over read noise.

Thanks.  My primary interest is in exoplanet work where I take an exposure from 1 to 30 seconds, then take another exposure a minute later, etc.  My goal is to get the cleanest, lowest noise, highest quality readout but at a speed that fits within the exposure/timing parameters mentioned.  

 

This camera allows either 500 kHz or 2 MHz.  For the kind of work I described, which is the best setting?  And what is the tradeoff to running at 500 kHz?



#13 Jared

Jared

    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 7,018
  • Joined: 11 Oct 2005
  • Loc: Piedmont, California, U.S.

Posted 25 November 2021 - 02:54 PM

The downside to 500 KHz is simply 4x the time to download. Check the timing and see how long it takes to download. If 500 KHz is under 4s or so I’d stick with it. If it’s longer, you might start worrying about the time between exposures. Depending on your focal length you could consider using a smaller region of interest (ROI) rather than the entire frame. That will speed up downloads. Not sure how many stars you need in your frame for the photometry comparison.

Also, depending on focal length, you might consider binning 2x2. That will improve SNR which is what you need for measuring tiny drops in brightness and will also speed up downloads. If you are already under 4-5s for downloads at 500MHz just use the 500MHz setting.

#14 nic35

nic35

    Apollo

  • *****
  • Posts: 1,268
  • Joined: 08 Sep 2007

Posted 25 November 2021 - 06:05 PM

Out of curiosity what size FOV are you interested in, and what kind of resolution in arc sec per pixel is optimal?

#15 EdM2

EdM2

    Mariner 2

  • *****
  • topic starter
  • Posts: 239
  • Joined: 08 Mar 2011

Posted 25 November 2021 - 11:43 PM

The downside to 500 KHz is simply 4x the time to download. Check the timing and see how long it takes to download. If 500 KHz is under 4s or so I’d stick with it. If it’s longer, you might start worrying about the time between exposures. Depending on your focal length you could consider using a smaller region of interest (ROI) rather than the entire frame. That will speed up downloads. Not sure how many stars you need in your frame for the photometry comparison.

Also, depending on focal length, you might consider binning 2x2. That will improve SNR which is what you need for measuring tiny drops in brightness and will also speed up downloads. If you are already under 4-5s for downloads at 500MHz just use the 500MHz setting.

That makes sense.  Four times the download time is the cost of running at 500 kHz vs. 2 mHz.  That I can understand.  The benefit is then lower readout noise.  Hopefully this is something I can adjust in the driver settings and experiment with.  Thank you for explaining that.



#16 EdM2

EdM2

    Mariner 2

  • *****
  • topic starter
  • Posts: 239
  • Joined: 08 Mar 2011

Posted 25 November 2021 - 11:49 PM

Out of curiosity what size FOV are you interested in, and what kind of resolution in arc sec per pixel is optimal?

Because of the size of the pixels (13.5 microns) and the focal length of the telescope (3454 mm) the plate scale will be 0.81 arc seconds per pixel.  That is well within NASA's specification of "less than 1 arcsecond/pixel".  The CCD has an array of 2048x 2048 so the FOV will be 1659 arcseconds.



#17 EdM2

EdM2

    Mariner 2

  • *****
  • topic starter
  • Posts: 239
  • Joined: 08 Mar 2011

Posted 26 November 2021 - 01:03 AM

The downside to 500 KHz is simply 4x the time to download. Check the timing and see how long it takes to download. If 500 KHz is under 4s or so I’d stick with it. If it’s longer, you might start worrying about the time between exposures. Depending on your focal length you could consider using a smaller region of interest (ROI) rather than the entire frame. That will speed up downloads. Not sure how many stars you need in your frame for the photometry comparison.

Also, depending on focal length, you might consider binning 2x2. That will improve SNR which is what you need for measuring tiny drops in brightness and will also speed up downloads. If you are already under 4-5s for downloads at 500MHz just use the 500MHz setting.

Do you happen to know if the readout speed can be adjusted by the customer (perhaps through the driver)?  Or is it set at the factory and cannot be changed by the customer?



#18 Jared

Jared

    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 7,018
  • Joined: 11 Oct 2005
  • Loc: Piedmont, California, U.S.

Posted 26 November 2021 - 01:32 AM

I don’t happen to know, sorry.

#19 EdM2

EdM2

    Mariner 2

  • *****
  • topic starter
  • Posts: 239
  • Joined: 08 Mar 2011

Posted 26 November 2021 - 04:04 PM

I think it can be set by software.  I looked around and found this link:

 

https://flicamera.co...tartupGuide.pdf

 

There is also a plugin that allows Maxim-DL to access readout speed for an FLI camera.



#20 EdM2

EdM2

    Mariner 2

  • *****
  • topic starter
  • Posts: 239
  • Joined: 08 Mar 2011

Posted 26 November 2021 - 04:05 PM

I don’t happen to know, sorry.

I think it can be set by software.  I looked around and found this link:

 

https://flicamera.co...tartupGuide.pdf

 

There is also a plugin that allows Maxim-DL to access readout speed control for an FLI camera.  Other image capturing software is compatible with FLI as well.


Edited by EdM2, 26 November 2021 - 04:10 PM.



CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics