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

QHY163 mono first shots

  • Please log in to reply
75 replies to this topic

#26 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 103
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 02 November 2016 - 02:23 AM

I can test camera with another power supply and other cables. But is this pattern somehow visible in the frame (I cannot see it even in much stretched bias frame)? How it should be interpretted?



#27 dvalid

dvalid

    Ranger 4

  • -----
  • Posts: 320
  • Joined: 02 Apr 2014
  • Loc: Georgia

Posted 02 November 2016 - 05:52 AM

Thank you very much Lukas

 

Here it is the FFT of a central crop (1024 x 1024) of the read noise file you supplied:

 

attachicon.gifRead noise FFT - QHY163M.jpg

 

Is this some kind of fixed pattern noise? Would this be from the camera, or some electrical interference from the power supply?

 

Any thoughts would be appreciated.

 

Regards,

 

Alfredo

 

Here is mine at a gain setting = 5. 

 

Workspace02.jpg

 

Subtracted version is on the right side, while FFT of a single bias on the left, just for a comparison.

 

Cheers.


Edited by dvalid, 02 November 2016 - 05:57 AM.

  • drjolo likes this

#28 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 103
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 02 November 2016 - 05:56 AM

I have checked also for my bias frame at gain=10 and it looks the same like above - strong vertical line and very faint horizontal line. 


  • Maudy likes this

#29 Alfredo Beltran

Alfredo Beltran

    Viking 1

  • *****
  • Posts: 854
  • Joined: 01 Aug 2010
  • Loc: Bogota, Colombia

Posted 02 November 2016 - 02:50 PM

I can test camera with another power supply and other cables. But is this pattern somehow visible in the frame (I cannot see it even in much stretched bias frame)? How it should be interpretted?

According to Craig Stark's article the FFT of a read noise frame with just random noise should have only a white dot in the center.

 

Any fixed pattern bands are not random noise introduced by the camera when reads the image, leaving a fingerprint on each image. So, in order to have a clean read noise from the camera, it would be really good to know where this banding is coming from and if QHY can eliminate it.

 

Regards,

 

Alfredo



#30 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 103
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 02 November 2016 - 03:26 PM

Thanks Alfredo - it makes perfect sense. So the bias noise is not uniform for gain=0 setting, and the question is how much it affects final images. 

Next time I will calculate photometric transformation coefficients for some standard field using V filter and check how camera behaves in this area. 



#31 dvalid

dvalid

    Ranger 4

  • -----
  • Posts: 320
  • Joined: 02 Apr 2014
  • Loc: Georgia

Posted 02 November 2016 - 06:04 PM

 

According to Craig Stark's article the FFT of a read noise frame with just random noise should have only a white dot in the center.

 

Any fixed pattern bands are not random noise introduced by the camera when reads the image, leaving a fingerprint on each image. So, in order to have a clean read noise from the camera, it would be really good to know where this banding is coming from and if QHY can eliminate it.

 

Regards,

 

Alfredo

 

 

Is not it more reliable to subtract SuperBias frame instead of Master in order to estimate what the read noise looks like with a removed fixed pattern noise?

 

Here are the results for Super Bias subtracted from a single bias frame:

QHY_B_SB.jpg

 

and from the Master Bias:

 

QHY_MB_SB.jpg



#32 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 103
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 03 November 2016 - 02:45 AM

Thats very interesting David. So first FFT is single bias frame minus superbias frame, and second FFT is single bias frame minus master bias, right? But what is the difference between masterbias and superbias you created? Did you use gain setting 5 again?



#33 Alfredo Beltran

Alfredo Beltran

    Viking 1

  • *****
  • Posts: 854
  • Joined: 01 Aug 2010
  • Loc: Bogota, Colombia

Posted 03 November 2016 - 06:40 AM

I think master bias is more accurate to know the read noise behavior of a camera since the super bias is generated by running an algorithm on the computer as stated here. Nevertheless superbias should be a very good tool for calibrating.

 

Alfredog


Edited by Alfredo Beltran, 03 November 2016 - 06:47 AM.


#34 dvalid

dvalid

    Ranger 4

  • -----
  • Posts: 320
  • Joined: 02 Apr 2014
  • Loc: Georgia

Posted 03 November 2016 - 09:13 AM

Thats very interesting David. So first FFT is single bias frame minus superbias frame, and second FFT is single bias frame minus master bias, right? 

That is correct Lucas. Gain setting was 5 again, Offset 35. 

 

The Superbias is a process in Pixinsight, as I know it removes random noise from the master bias while leaving fixed pattern noise in the frame. Actually calibrating with superbias does not inject any additional noise, just removes the pattern noise.

 

Masterbias vs Superbias:

 

Screenshot 2016-11-03 16.53.05.jpg

 

Cheers


Edited by dvalid, 03 November 2016 - 07:08 PM.


#35 DesertRat

DesertRat

    Fly Me to the Moon

  • *****
  • Posts: 6142
  • Joined: 18 Jun 2006
  • Loc: Valley of the Sun

Posted 04 November 2016 - 01:18 PM

FFT's run on bias frames earlier showed a vertical line in the DFT magnitude.  This is a result of a faint row oriented pattern.

 

The PI "superbias" using a default column pattern for this camera is not useful.  A superbias for a low noise high speed camera such as a QHY163 is not recommended.  This is really intended for cameras which for whatever reason (slow download, mechanical shutter, etc) a good master bias is not achievable.  You could take any image, even one composed of pure gaussian noise, run the PI superbias process and you would get a pristine column oriented pattern.  Unless you select "Columns and rows", in which case you get two frames which can be added in some way to give a screen-door pattern.  This would be poor calibration for this camera.

 

With the QHY163 or similar camera one should generate a high quality master bias using lots of frames.  If you have the cpu power, memory and patience try to integrate 256 frames or at the very least 64.

 

Stretching that master will show any faint patterns, whether linear or periodic structure oriented.  Its nothing to be concerned about as long as you have good calibration data.  Any FFT should be run on this master to detect any periodicity.  Sometimes cropping may be beneficial in this type of analysis, since beginning and ending rows and columns can have artifacts - not sure about this camera - perhaps not.

 

Glenn


  • drjolo and Shiraz like this

#36 Pauls72

Pauls72

    Vanguard

  • *****
  • Posts: 2226
  • Joined: 28 Oct 2007
  • Loc: LaPorte, IN USA

Posted 05 November 2016 - 12:18 PM

QHY163M first light.

Temp=-15C

Gain=11

Offset=30

50 @ 120 seconds

 

Scope 6"RC - I had given up on the RC over a year ago and went to a refractor. So this is the first decent image I have gotten from the Ritchey-Chrétien too.

M81 - Bode's Galaxy

 

Sorry no filters yet.

 

M81_120s-15C_11g_30o_Stacked_STs.jpg

 

get.jpg


Edited by Pauls72, 05 November 2016 - 12:27 PM.

  • artem2, dvalid, Tyson M and 2 others like this

#37 Herra Kuulapaa

Herra Kuulapaa

    Viking 1

  • -----
  • Posts: 509
  • Joined: 10 Dec 2014
  • Loc: Vantaa, Finland

Posted 16 December 2016 - 01:29 AM

I think there is a lot of good information already on this thread so I propose we use it as an official one.

 

Anyway, I attach the test data here as well:

 

Sensor -20C, environment +8C, stabilized for 1 hour, taken to room temp for FLIR photo:

QHY163_FLIR4.jpg

 

QHY163_FLIR6.jpg

Sensor same -20 degrees C, sensor window heating ~19 degrees C, which is warmer than both frame and outside air (8 degrees C).


  • drjolo likes this

#38 Herra Kuulapaa

Herra Kuulapaa

    Viking 1

  • -----
  • Posts: 509
  • Joined: 10 Dec 2014
  • Loc: Vantaa, Finland

Posted 16 December 2016 - 01:39 AM

This is what was posted to Stargazerslounge, but I think you find it interesting here as well.

 

Disclaimer:  Below is strongly stretched uncalibrated data just to illustrate sensor noise characteristics and primary meant for the people who like to dig into data/noise. Image calibration is the tool to deal such minor deviations and one shouldn’t give too much weight to them. Overall this camera produces very clean darks.

 

A strongly stretched BIAS, Gain 12, Offset 75, Temp -25C, 160 frame stack:

BIAS_G12_OS75_-25C_160f.jpg

 

A strongly stretched DARK, 300s, Gain 12, Offset 75, Temp -25C, 100 frame stack:

DARK_G12_OS75_-25C_100f.jpg

 


  • drjolo and dvalid like this

#39 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 103
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 16 December 2016 - 02:45 AM

That is interesting! So the camera case is 8*C warmer than ambient. I noticed it getting warm when cooling is on (it must, because the heat needs to be distributed somehow), but I didn't thought that is 8*C more. Do you remember what was the cooling power in % for this test conditions?

 

A few more pictures from my backyard. All made with Photoline 130/910 refractor on EQ6 mount under suburban sky.

 

Rosette nebula HaOiiiRGB

2016-12-04-N2238HRGB1.jpg

Large color version https://goo.gl/VQzqDh

Ha mono enlarged crop https://goo.gl/WdAoRq

 

Crab nebula Ha band

2016-11-28-M1Hcrop.jpg

1:1 crop version here https://goo.gl/Dxr8p0 (1"/px scale)

 

Barnard 174 area LRGB through high clouds, so star halos are there https://goo.gl/acChjS

M71 globular cluster in Sagitta LRGB https://goo.gl/YvQzrt


  • artem2 and dvalid like this

#40 Herra Kuulapaa

Herra Kuulapaa

    Viking 1

  • -----
  • Posts: 509
  • Joined: 10 Dec 2014
  • Loc: Vantaa, Finland

Posted 16 December 2016 - 03:24 AM

That is interesting! So the camera case is 8*C warmer than ambient. I noticed it getting warm when cooling is on (it must, because the heat needs to be distributed somehow), but I didn't thought that is 8*C more. Do you remember what was the cooling power in % for this test conditions?

 

Very nice images!

 

Yes, in that setup the case was only 8C warmer. Pumping the heat with peltier effect produces always more heat. In other words moving one unit of heat produces one extra heat unit (or more). Warmer the environment, harder the heat dissipation is.

 

Here is -10 sensor temp at 22C environment.

QHY163_FLIR1.jpg

 

QHY163_FLIR2.jpg

 

Cooling sensor to -25C at -5C, 0C, 10C 20C range needs very different power inputs. I haven't tested it, but at 0C the power requirement is quite low (maybe 20-30%[?]), 10C it's roughly 50% and at 20C it's 100% [delta max 40C] - Just to give an example. Pumping full amps to cooling at 20C also makes case much hotter and I wouldn't see surprised to see >50C temps.


Edited by Herra Kuulapaa, 16 December 2016 - 03:54 AM.

  • drjolo likes this

#41 Herra Kuulapaa

Herra Kuulapaa

    Viking 1

  • -----
  • Posts: 509
  • Joined: 10 Dec 2014
  • Loc: Vantaa, Finland

Posted 17 December 2016 - 11:58 AM

Hmmm., odd :question:

 

I get unity gain setting to be a bit over 13 instead of 12.4 what some of you say.

 

Chart_2.jpg

 

I messed up the gain setting point 9 when took darks and it has been calculated/approximated there, but still it's only a one point and doesn't affect any other so it shouldn't be that.

 

Full well capacity seems to correlate in my test what at least ZWO says (20ke-), QHY (18-20ke-)

Chart_1.jpg


Edited by Herra Kuulapaa, 18 December 2016 - 01:19 AM.


#42 Herra Kuulapaa

Herra Kuulapaa

    Viking 1

  • -----
  • Posts: 509
  • Joined: 10 Dec 2014
  • Loc: Vantaa, Finland

Posted 19 December 2016 - 12:14 AM

And the high gain part included by gain steps of 5.

 

Chart_3.jpg

 

Chart_4.jpg

 

Chart_5.jpg

 

I'm thinking to limit my high gain setting at 25, which will result 1,2e- read noise, but still have ~1000e- of well capacity. To be honest there isn't that much difference to gain 13 (1,5e- / 4300e-).



#43 Herra Kuulapaa

Herra Kuulapaa

    Viking 1

  • -----
  • Posts: 509
  • Joined: 10 Dec 2014
  • Loc: Vantaa, Finland

Posted 27 December 2016 - 03:52 AM

First proper light due to recent cloudy nights.

 

Decided to test M45 to see how camera can cope with wide brightness range.

M45_QHY163M_1200.jpg

 

120 secs at gain 12 saturated stars badly but the shorter exposures compensated that. I think I could have taken just 10 x 15s as shorter ones and still end up same looking image.


Edited by Herra Kuulapaa, 27 December 2016 - 04:29 AM.

  • Campos, artem2 and dvalid like this

#44 Campos

Campos

    Apollo

  • *****
  • Posts: 1264
  • Joined: 26 Dec 2012
  • Loc: Portugal

Posted 29 December 2016 - 12:14 PM

Hi Herra,

 

Nice first light!

Yeah, I also found that 12 gain is a bit too much for broadband, even with RGB filters I get saturated stars very quickly, should use maybe around gain 5 for this, still trying to set my gain for NB though, I'm using now gain 13 and offset 70 but I'm not sure about my callibration frames as I'm getting a "cluster" of black pixels on my bias, so I'm guessing this is due to incorrect offset vallue...have you noticed anything like this with yours?

 

Cheers,

 

Luís


Edited by Campos, 29 December 2016 - 12:14 PM.


#45 Herra Kuulapaa

Herra Kuulapaa

    Viking 1

  • -----
  • Posts: 509
  • Joined: 10 Dec 2014
  • Loc: Vantaa, Finland

Posted 29 December 2016 - 02:15 PM

Hi Herra,

 

Nice first light!

Yeah, I also found that 12 gain is a bit too much for broadband, even with RGB filters I get saturated stars very quickly, should use maybe around gain 5 for this, still trying to set my gain for NB though, I'm using now gain 13 and offset 70 but I'm not sure about my callibration frames as I'm getting a "cluster" of black pixels on my bias, so I'm guessing this is due to incorrect offset vallue...have you noticed anything like this with yours?

 

Cheers,

 

Luís

No, haven't noticed. There is a Bias frame (jpg) above, how does it correlate to yours?

Could you post a sample of the pixels?


  • Campos likes this

#46 hfjacinto

hfjacinto

    I think he's got it!

  • *****
  • Posts: 17657
  • Joined: 12 Jan 2009
  • Loc: Land of clouds and LP

Posted 29 December 2016 - 02:18 PM

Herra,

 

Are you using SGP with the QHY?



#47 Herra Kuulapaa

Herra Kuulapaa

    Viking 1

  • -----
  • Posts: 509
  • Joined: 10 Dec 2014
  • Loc: Vantaa, Finland

Posted 29 December 2016 - 02:19 PM

Just a technical test with 10 x 120s Ha exposures with my f/6.5 TS65Q. Auto screen transfer to histogram, RGB conversion, size reduction and subtle contrast fix, no other editing.

 

Gain 20, Offset 75 (Full well 1918e-; RN 1,28e-; Gain 0,47 e-/ADU)

G20OS75_120s_10kpl_1200.jpg

 

hfjacinto: Yes, SGP


Edited by Herra Kuulapaa, 29 December 2016 - 02:21 PM.

  • hfjacinto, Aenima, artem2 and 1 other like this

#48 hfjacinto

hfjacinto

    I think he's got it!

  • *****
  • Posts: 17657
  • Joined: 12 Jan 2009
  • Loc: Land of clouds and LP

Posted 29 December 2016 - 02:22 PM

Herra,

 

I wanted to thank you for making me spend money. After seeing your images with the QHY 163M, I made up my mind to get one. These are truly great!!! YOU ARE THE MAN!!!!!



#49 Herra Kuulapaa

Herra Kuulapaa

    Viking 1

  • -----
  • Posts: 509
  • Joined: 10 Dec 2014
  • Loc: Vantaa, Finland

Posted 29 December 2016 - 02:28 PM

Herra,

 

I wanted to thank you for making me spend money. After seeing your images with the QHY 163M, I made up my mind to get one. These are truly great!!! YOU ARE THE MAN!!!!!

Hah, so it goes!

 

This is a 30 x 120s exposure (Please note that it is edited). Some of the frames had quite weak SNR due to upper clouds though.

 

Gain 20, Offset 75 (Full well 1918e-; RN 1,28e-; Gain 0,47 e-/ADU)

G20OS75_120s_30kpl_1200.jpg


Edited by Herra Kuulapaa, 29 December 2016 - 02:42 PM.

  • Robin Lee, hfjacinto, Alfredo Beltran and 1 other like this

#50 Campos

Campos

    Apollo

  • *****
  • Posts: 1264
  • Joined: 26 Dec 2012
  • Loc: Portugal

Posted 01 January 2017 - 02:32 PM

 

Hi Herra,

 

Nice first light!

Yeah, I also found that 12 gain is a bit too much for broadband, even with RGB filters I get saturated stars very quickly, should use maybe around gain 5 for this, still trying to set my gain for NB though, I'm using now gain 13 and offset 70 but I'm not sure about my callibration frames as I'm getting a "cluster" of black pixels on my bias, so I'm guessing this is due to incorrect offset vallue...have you noticed anything like this with yours?

 

Cheers,

 

Luís

No, haven't noticed. There is a Bias frame (jpg) above, how does it correlate to yours?

Could you post a sample of the pixels?

 

My bias frame is totally diferent, I don't see the classical vertical lines on a streched bias, mine is smooth with a black pixels cluster near thye center, been doing another bias session just to check this out, I'm noticing something weird with the aquisition though, when checking the histogram I see many fluctuations during capture, it's almost like the camera is changing the gain/offset vallues on it's own during aquisition, I've noticed this on my test images also during capture on several objects but as was imaging through thin clouds I atributed this to transparency fluctuations...now I'm getting pretty convinved my camera has some problem...this sucks :(


Edited by Campos, 01 January 2017 - 02:33 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