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

#1 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 102
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 03 October 2016 - 03:30 AM

Hello,

 

I received a few days ago one of QHY163 mono camera for testing. This is the one http://www.qhyccd.com/QHY163.html . So far I had a chance to install it, made some desk tests and made first sky shots at star party in Zatom, Poland. I have taken with me only limited setup withtout guiding, so I shot short exposures only and stacked them.

All images come from SCT8" with Alan Gee Mark II telecompressor and QHY163 camera put on CG5 GT mount without guiding, which gaves 0.55"px scale. No filter has been used. Single exposure time was always 5 seconds, and camera settings were: gain=20, offset=100, sensor temperature -25*C. Collected, calibrated and stacked with MaxIm 6.13, then resize 50%.

 

NGC891, 80x5 seconds: https://goo.gl/XNKBzl

M1, 200x5 seconds: https://goo.gl/1UZa5N

M15, 100x5 seconds: https://goo.gl/LjwtX7

M76, 100x5 seconds: https://goo.gl/a2BRDf

 

Here is two thumbnails:

_891_DBE_resize50_1_1200_1.jpg

M1_DBE_resize50_2_1200_2.jpg


Edited by drjolo, 04 October 2016 - 01:39 AM.

  • mikewayne3, DuncanM, Cristina Cellini and 3 others like this

#2 Alfredo Beltran

Alfredo Beltran

    Viking 1

  • *****
  • Posts: 741
  • Joined: 01 Aug 2010
  • Loc: Bogota, Colombia

Posted 03 October 2016 - 07:02 AM

Impressive results for individual short exposure times at that focal length.

 

This seems a very promising camera.

 

Alfredo



#3 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 102
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 03 October 2016 - 07:10 AM

During desktop tests I have measured linearity and other parameters. So linearity looks very good, basically up to 60000 ADU the response is perfectly linear:

linearity1.jpg

linearity2.jpg

 

Then basing on this method http://www.phy.cuhk....Measurement.pdf I have determined camera real gain (for 16bit resolution). It is g=0.26e -/ADU. Basing on this I can estimate pixel capacity to 17100 electrons assuming linear response till the range end. But the response is getting lower at high end, so I assumed capacity 18000 electrons for further calculations. Having this it was pretty straightforward to calculate following data:

2016-10-03_134948.png

 

First column is gain set in the camera driver settings. Second column is measured signal for 1s exposure. Then third columns is real gain in dB, and in the next column real gain in electrons / ADU. Then we have camera read noise, saturation and dynamic range both in EV and dB. Last row is Atik383 values for comparison.
Read noise is pretty impressive in my opinion. 

QHY163_5.jpg


Edited by drjolo, 03 October 2016 - 08:20 AM.

  • BenKolt, keithlt and Jon Rista like this

#4 siriusarcher

siriusarcher

    Ranger 4

  • *****
  • Posts: 318
  • Joined: 25 Feb 2008
  • Loc: Canada

Posted 03 October 2016 - 10:57 PM

Results very impressive also for the fact that you used the CG-5 GT. I've been investigating this camera, seems like a good choice for a mount where weight must be kept to a minimum. Thanks for the useful measurements on camera performance which are not provided at the QHY web site description of this camera.

 

Dave



#5 XS_Man

XS_Man

    Explorer 1

  • -----
  • Posts: 93
  • Joined: 20 Jul 2016
  • Loc: FRANCE

Posted 05 October 2016 - 02:52 PM

Good first images !

 

I see you are working with short exposures (5 sec).

As you are using offset value = 100, I guess that it's to remove the remains of 'amp glow' near the edges of the sensor ?

 

For my QHY174 M the Anti-amp glow circuit is activated for exposure greater than 7 seconds.

 

Albéric


Edited by XS_Man, 05 October 2016 - 02:53 PM.


#6 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 102
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 06 October 2016 - 03:11 AM

Thanks Albéric! I haven't noticed AG as an issue for QHY163. It is barely visible on 300s dark, you can see 600s H alpha NGC6888 picture attached. AG signal is so small, that its not visible in real narrowband images. 

When I raise gain parameter I need also to raise offset to move bias histogram to the right, so no pixels will be close or equal to zero - that produces artifacts. For QHY163 I found these settings good for me:

GAIN      OFFSET

0          40

10         70

20         100

30         150

>=40       256

 

I have no information if there is any AG system in QHY163 (probably there is) and when it is activated :( 

Two more shots attached, this time with 130mm triplet and 0.79x FF/FR on EQ6 mount. Calibrated with 100xbias and hot pixels removed:

 

NGC6888 single 600s H alpha shot, guided on EQ6, gain=0, offset=40, t=-30*C. Full size here: https://goo.gl/ySIaSp (large)

6888HaHP.jpg

 

Andromeda triplet, 180x10 seconds unguided on EQ6, gain=10, offset=80, t=-25*C. 50% resize here: https://goo.gl/i5jgZN

M110.jpg


Edited by drjolo, 06 October 2016 - 06:34 AM.

  • DesertRat, Cyclop_si, mikewayne3 and 6 others like this

#7 JohnPlayerSpecial

JohnPlayerSpecial

    Lift Off

  • -----
  • Posts: 1
  • Joined: 23 Sep 2016

Posted 09 October 2016 - 11:55 AM

Hi drjolo,

 

How do you have made this messurements?

I should receive the color version soon. 

 

Kind regards 

 

Kevin



#8 Mark Sibole

Mark Sibole

    Mercury-Atlas

  • *****
  • Posts: 2853
  • Joined: 30 Nov 2007
  • Loc: Fife Lake Mi.

Posted 09 October 2016 - 05:49 PM

Dont know if its the short exposures or the processing but IMO there is no detail in M1 NGC 891 ans M31.

It may be the charactoristics of the imager or the short exposures.But with that being said not to bad of a job.



#9 QHYCCD

QHYCCD

    Vostok 1

  • -----
  • Vendors
  • Posts: 127
  • Joined: 06 Aug 2014

Posted 09 October 2016 - 11:56 PM

Dont know if its the short exposures or the processing but IMO there is no detail in M1 NGC 891 ans M31.

It may be the charactoristics of the imager or the short exposures.But with that being said not to bad of a job.

Looks the NGC6888 and M31 has good details when looking at  the big size image. 



#10 DesertRat

DesertRat

    Fly Me to the Moon

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

Posted 10 October 2016 - 11:47 AM

drjolo,

 

Those images look good and given the integration times show great promise for this camera. :waytogo: :waytogo:

 

As far as your characterization of measures I think it preferable to express gain in terms referencing back to the resolution of the ADC for this camera which is 12 bits.  Hence the gain at the zero setting would be 0.26*16 or ~4 e-/ADU.  So unity gain would be somewhere above a setting of 10.   The noise figures look great!

 

Glenn



#11 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 102
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 10 October 2016 - 12:31 PM

Hi drjolo,

 

How do you have made this messurements?

I should receive the color version soon. 

Hi Kevin,

 

So linearity I measured just with stable light source pointed to sensor and with increasing exposure time. Then using Maxim DL I made from each calibrated frame 100x100px crop at the center and measured average signal minus offset added by Maxim (its value is in the FITS header). That gave me the linearity plots. Then using the document I linked to measure real gain. And then (data in table) I made some stable light pictures again with different gain and offset settings, measured noise in Maxim DL, and after calibration again - the signal in center 100x100px crop. I can describe process with more details - just let me know, however I am not sure if you can use linked document to measure gain of color camera - probably not :( But anyway the gain and bias noise should be the same, as far as I know it is the same Panasonic sensor only with RGB bayer matrix added. 

 

Dont know if its the short exposures or the processing but IMO there is no detail in M1 NGC 891 ans M31.

It may be the charactoristics of the imager or the short exposures.But with that being said not to bad of a job.

It's just a single stretch run with histogram transformation in Pixinsight. I haven't processed images more, so anyone can take 80x5s shots or single 400s shot and easily compare without being bothered with processing skills. The point is that you can take many 5 or 10 seconds subframes without guiding on entry level mounts like EQ6 or CG5GT or even dobson with tracking platform and get decent results even for large scale. That short exposures are much less suitable for CCD sensors due to its large read noise.

 

drjolo,

 

Those images look good and given the integration times show great promise for this camera. :waytogo: :waytogo:

 

As far as your characterization of measures I think it preferable to express gain in terms referencing back to the resolution of the ADC for this camera which is 12 bits.  Hence the gain at the zero setting would be 0.26*16 or ~4 e-/ADU.  So unity gain would be somewhere above a setting of 10.   The noise figures look great!

 

Glenn

Hi Glenn,

 

You're right, it makes sense from ADC point of view, so for gain setting of 10 we have 0.08125*16 = 1.3 e-/ADU, and as you wrote unity gain will be little above 10. When we plot this real ADC gain versus camera gain setting we got this log plot:

2016-10-10_192357.png

So the unity gain seems to happen at gain setting between 12 and 13. 

 

Cheers

Lucas


Edited by drjolo, 10 October 2016 - 12:47 PM.

  • DesertRat, Alfredo Beltran and keithlt like this

#12 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 102
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 11 October 2016 - 02:24 PM

There is still "calibration frames" weather in my location, so small update - I measured current consuption of QHY163M. Camera works without 12V DC connected, both on USB2.0 and USB3.0 ports, all works fine, only cooling is disabled. After connecting to 12V DC and with TEC off the camera drains 0.15A (fan is working). Increasing TEC power (using slider in EZCAP) current consumption looks like this:

2016-10-11_194333.png



#13 entilza

entilza

    Mercury-Atlas

  • *****
  • Posts: 2989
  • Joined: 06 Oct 2014
  • Loc: Oakville, ON, Canada

Posted 11 October 2016 - 03:50 PM

Great Crescent and Andromeda, Looks like some solid initial tests!  



#14 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 102
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 14 October 2016 - 02:39 AM

Thanks! Weather is still crappy at my location. Yesterday there was a short window, but yeah, here is current sky hero :) 
 

qhy163-moon.jpg

 

Pictured of course with QHY163M and my deep sky setup - TS130/910 0.79x, Baader G, 10x0.006s stack. Here is full res https://goo.gl/BL8NM8


Edited by drjolo, 14 October 2016 - 02:40 AM.

  • mikewayne3 and dvalid like this

#15 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 102
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 27 October 2016 - 03:58 AM

Weather drives me nuts... indian summer resists to come at my location :( A few days ago there was a short window in clouds, so I collected 5x10 minutes of H alpha Western Veil through high clouds and 99% humidity. I am almost there with optimum distance to flattener, next attempt should be good :) 

 

VeilW-fin.jpg


  • mikewayne3, Aenima, dvalid and 1 other like this

#16 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 102
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 30 October 2016 - 03:49 AM

Last night I spent some time on adjusting tilt and distance to flattener. During this process I collected some pictures, still considered to be test pictures. Conditions were bad (NELM 4.5 and 99% humidity) but eventually I managed to adjust setup, and these are probably the last test pictures. Time for "real work" :) :)

Shot with QHY163M, Photoline 130 refractor, EQ6 mount.

 

NGC6940 open cluster, 30 minutes L, 3x15 minutes RGB https://goo.gl/D1xhIE

NGC7635 Bubble nebula 12x10 minutes Ha full frame https://goo.gl/JRLd3e, and enlarged crop https://goo.gl/nosGRs 

 

N7635-30.jpg

 

Cygnus Wall 6x10 minutes Ha  https://goo.gl/tEBn52

 

WallH-starless.jpg

 

Cygnus Wall 6x10 minutes Ha, 6x10 minutes Oiii, 3x16 minutes RGB  https://goo.gl/tOaYNE

 

WallHORGB-fin.jpg

 

Some more info at 

http://astrojolo.com...ne-tuning-haze/  

http://astrojolo.com...g-side-effects/


Edited by drjolo, 30 October 2016 - 03:54 AM.

  • nytecam, mikewayne3, BenKolt and 7 others like this

#17 dvalid

dvalid

    Vostok 1

  • -----
  • Posts: 122
  • Joined: 02 Apr 2014

Posted 30 October 2016 - 06:51 AM

Lukas, your test images are very encouraging! I'm still waiting for a better skies.. 



#18 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 102
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 30 October 2016 - 03:35 PM

Thanks David, me as well :) This was as far as I remember the worst October ever... 


  • dvalid likes this

#19 keithlt

keithlt

    Apollo

  • *****
  • Posts: 1304
  • Joined: 21 Dec 2013
  • Loc: Arizona©

Posted 01 November 2016 - 08:17 AM

really nice images, could you share the gain and offset ?  I hope to start working with mine again.

(edit I just found the gain and offset info above)


Edited by keithlt, 01 November 2016 - 08:20 AM.


#20 Alfredo Beltran

Alfredo Beltran

    Viking 1

  • *****
  • Posts: 741
  • Joined: 01 Aug 2010
  • Loc: Bogota, Colombia

Posted 01 November 2016 - 11:10 AM

It would be interesting to see a read noise file and its FFT to see how random the read noise is.

Could any of the beta testers provide it?

#21 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 102
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 01 November 2016 - 02:07 PM

Alfredo, here are the links to bias and 300s dark frame (gain=0, offset=40, t=-25*C):

bias: https://drive.google...iew?usp=sharing

dark: https://drive.google...iew?usp=sharing

 

The bias noise is not 100% random, but making  50 or 100 or even more bias frames is a matter of minute with USB3.0, and calibrating with such amount of frames gives very good result. And actual value of bias noise compared to light frame is almost none.

Also there are some AG visible on dark frames, but again - when compared to light frames AG signal is invisible in LRGB frames, and barely affects narrowband images. Calibrating with 10-20 darks completely removes AG and does not poison light frames with any significant calibration noise. I will post tomorrow some super stretched narrowband images before and after calibration, so you can see.



#22 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 102
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 01 November 2016 - 02:37 PM

Actually I found some raw images at my local machine, so here it is :)

 

This is single 600s H alpha exposure, high stretch in Pix. Raw image left, calibrated at right. You can see some AG at the right border and two wide horizontal bands at 1/3 and 2/3 frame height. That AG has been removed after calibration with darks:

ff-comp.jpg

This AG is not visible at all at photos with LRGB filters (at least at photos made at my location).

 

And here is 1:1 crop of the same frame. There is a few hot pixels visible at raw image that has been removed after calibration, and there is no visible calibration noise added after calibration:

crop-comp.jpg

 

Lucas

PS - uh, forum engine ruined full frame compare image, here is link to original one https://drive.google...iew?usp=sharing


Edited by drjolo, 01 November 2016 - 02:42 PM.

  • artem2 and Mebelek like this

#23 Alfredo Beltran

Alfredo Beltran

    Viking 1

  • *****
  • Posts: 741
  • Joined: 01 Aug 2010
  • Loc: Bogota, Colombia

Posted 01 November 2016 - 02:58 PM

Thank you very much Lukas.

 

Do you have a read noise frame generated by subtracting the stack of the master bias from a single bias? I'd like to see the FFT to see if there's any fixed pattern noise.

 

Rrgards,

 

Alfredo



#24 drjolo

drjolo

    Vostok 1

  • -----
  • topic starter
  • Posts: 102
  • Joined: 02 Apr 2010
  • Loc: Nieborowice, Poland

Posted 01 November 2016 - 03:23 PM

Ah, ok, now I understand what you wanted :) Ok, here it is - single bias frame minus master plus 500ADU constant https://drive.google...iew?usp=sharing


Edited by drjolo, 01 November 2016 - 03:23 PM.


#25 Alfredo Beltran

Alfredo Beltran

    Viking 1

  • *****
  • Posts: 741
  • Joined: 01 Aug 2010
  • Loc: Bogota, Colombia

Posted 01 November 2016 - 09:57 PM

Thank you very much Lukas

 

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

 

Read 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


Edited by Alfredo Beltran, 01 November 2016 - 09:59 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.







Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics