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

ZWO ASI294 MC pro

  • Please log in to reply
254 replies to this topic

#126 rockstarbill

rockstarbill

    Fly Me to the Moon

  • *****
  • Posts: 6155
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 29 May 2018 - 07:45 PM

That would be Unity. The boost kicks in there, and that would be the limit to the gain I would want to set it to in order to preserve as much DR as possible.

 

I'll take 10 -5C darks, 10 -10C darks, and 10 -15C darks. I will expose them for 2 mins each. Sound good? That should reveal some info about the change in cooling and how that relates.



#127 jdupton

jdupton

    Surveyor 1

  • *****
  • Posts: 1878
  • Joined: 21 Nov 2010
  • Loc: Central Texas, USA

Posted 29 May 2018 - 09:17 PM

Jon, Andy, and others,

 

   As a followup to my post above, I got to thinking about the LGA packaging of the ICX294CJK sensor. Because of the size of the sensor, only having 248 LGA contacts on the bottom probably means that there are only two to four rows of contacts around the periphery of the sensor's back side. That would leave a relatively large area in the center free of contacts.

 

   If we pair that with the information supplied from ZWO in the other thread, a cold finger could directly contact the back of the sensor but wouldn't necessarily be able to cover the full chip imaging area of the package. That dovetails with the comments about the cold finger being "undersized" compared to the sensor size.

 

   I was wrong in my initial assumption that there might have to be contacts covering most of the back of the chip. A few rows around the periphery is much more likely given the relatively low pin count of the package. It would be nice to find a detailed specification datasheet for the chip to verify the pin-out diagrams and see just how much of the back of the chip cannot be cooled due to pin electrical contact areas of the LGA.

 

   I am still interested in a camera based on the chip but the packaging constraints on cooling do give me pause.

 

 

John


  • Merc and Jon Rista like this

#128 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23600
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 29 May 2018 - 09:38 PM

Jon, Andy, and others,

 

   As a followup to my post above, I got to thinking about the LGA packaging of the ICX294CJK sensor. Because of the size of the sensor, only having 248 LGA contacts on the bottom probably means that there are only two to four rows of contacts around the periphery of the sensor's back side. That would leave a relatively large area in the center free of contacts.

 

   If we pair that with the information supplied from ZWO in the other thread, a cold finger could directly contact the back of the sensor but wouldn't necessarily be able to cover the full chip imaging area of the package. That dovetails with the comments about the cold finger being "undersized" compared to the sensor size.

 

   I was wrong in my initial assumption that there might have to be contacts covering most of the back of the chip. A few rows around the periphery is much more likely given the relatively low pin count of the package. It would be nice to find a detailed specification datasheet for the chip to verify the pin-out diagrams and see just how much of the back of the chip cannot be cooled due to pin electrical contact areas of the LGA.

 

   I am still interested in a camera based on the chip but the packaging constraints on cooling do give me pause.

 

 

John

I think you are on to something there. I would also be curious to know if the sensor itself actually directly contacts the LGA PCB or not. It might only contact at certain points, depending on the kind of pin package the sensor itself has, which might mitigate cooling effectiveness.



#129 Jeff2011

Jeff2011

    Gemini

  • *****
  • Posts: 3494
  • Joined: 01 Jan 2013
  • Loc: Sugar Land, TX

Posted 29 May 2018 - 09:41 PM

I am interested in this camera also but wonder if ZWO will come out with an improved model to solve the cooling issue.  I would rather not take darks the same night as my imaging session even though I intend to use it with very short exposures.  I thought I would never be interested in an OSC camera but this one has my interest.



#130 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23600
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 29 May 2018 - 10:07 PM

I am interested in this camera also but wonder if ZWO will come out with an improved model to solve the cooling issue.  I would rather not take darks the same night as my imaging session even though I intend to use it with very short exposures.  I thought I would never be interested in an OSC camera but this one has my interest.

If the sensor does indeed only come in the LGA package, then ZWO may not be able to do much. I originally thought they might be trying to keep costs low by reusing cooling units designed for other cameras, but if the LGA does indeed have limited area within the center where a TEC could be attached, and even more so if the sensor itself does not fully contact the PCB (would depend on the kind of package the sensor itself is in and how it is soldered to the LGA PCB), then there would be a lot of limitations and hurdles to overcome just to get the sensor to cool at all.

 

I actually noticed in the past when this camera first came onto the market that the dark current levels were higher than other Sony based ASI cameras. It only gets down to about 0.02-0.05e-/s between -15C and -10C, which is around an order of magnitude higher than most Sony sensors. I wondered why at the time, but if the sensor is not being directly cooled, then that would probably explain a lot of it. The Panasonic sensor gets down to ~0.008e-/s @ -15C, ~0.005e-/s @ -20C, the IMX183 gets down to ~0.002e-/s @ -15C, ~0.0035e-/s @ -10C. Those sensors are likely being directly cooled, or much closer to it, though.


  • jdupton likes this

#131 jdupton

jdupton

    Surveyor 1

  • *****
  • Posts: 1878
  • Joined: 21 Nov 2010
  • Loc: Central Texas, USA

Posted 29 May 2018 - 10:12 PM

Jon,

 

I think you are on to something there. I would also be curious to know if the sensor itself actually directly contacts the LGA PCB or not. It might only contact at certain points, depending on the kind of pin package the sensor itself has, which might mitigate cooling effectiveness.

 

   Many LGA packages like the one pictured for this sensor are made with a ceramic base. The leads are wire bonded as usual at the top edges and then routed through the ceramic substrate to the LGA contacts on the back side. I haven't seen one in person, but a current owner of a 294 camera can probably tell whether the substrate is ceramic or FR4-like laminate (or similar) closely examining the sensor under good lighting. (No need to open the camera, of course.)

 

 

John

 

PS: If we can believe the Framos.Com Website, the LGA package is indeed a ceramic substrate.


Edited by jdupton, 29 May 2018 - 10:20 PM.


#132 rockstarbill

rockstarbill

    Fly Me to the Moon

  • *****
  • Posts: 6155
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 29 May 2018 - 11:13 PM

Darks uploading in a zip file here:

 

https://1drv.ms/f/s!...c3mk8mq0GQiKwwQ

 

Should be done in a few mins. 



#133 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23600
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 29 May 2018 - 11:26 PM

Hi Jon,

 

OK, here is a link to a zip file with:
- 4 light frames, 2 before / 2 after m.f. (CFA data, FIT format),
- MasterDark (CFA data, XISF format),
- MasterFlat (CFA data, XISF format),
- 2 integrations, 1 before / 1 after m.f., (RGB data, XISF format).

 

The MasterDark is built of 100 dark frames, the Flat frames are calibrated with a MasterFlat-Dark built of 111 Flat-Darks, the MasterFlat is built of 40 calibrated Flat frames. The first integration is made of 50, the second integration of 57 calibrated, debayered, aligned light frames.

 

https://www.dropbox....ndmann.zip?dl=0

 

Let me know if you see indications of inconsistent calibration.

 

Cheers,

Bernd

I've been fiddling with your data. I am not sure the master dark is properly matched to the lights. Using the following simple pixel math:

 

K: $T + median(MD) - MD

 

To add a pedestal based on the median of the master dark, then subtract the master dark, I get an inversion of the starburst amp glow near the upper right in every light frame. In a proper calibration, the amp glow should be perfectly removed, with no remnant but a barely perceptible change in noise profile. Since there is a remnant glow, the master dark glow seems to be brighter than that in the lights, thus resulting in overcorrection (subtraction of larger values from smaller).

 

The first thing I would like to do, is get a few individual dark frames from you, ideally, sampled from your full range (i.e. a few from the beginning of your 100 darks, a few from say 1/4 in, a few at 1/2, a few at 3/4 and a few from the end of the range). This will allow me to subtract individual darks, as well as subtract darks from darks, to see if there is any kind of drift in dark signal throughout your range of darks. Subtracting a single dark is not ideal, however, if it results in a better correction than a super deep master, then it would actually be preferred over subtracting a super deep master. I suspect there is probably a happy medium in there somewhere, where you subtract a master made from 10, 20, 30 darks...

 

Anyway...based on a preliminary examination, the master is not well matched, so it is not surprising that there would be remnant pattern.



#134 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23600
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 29 May 2018 - 11:41 PM

Darks uploading in a zip file here:

 

https://1drv.ms/f/s!...c3mk8mq0GQiKwwQ

 

Should be done in a few mins. 

Initial review of yours, they seem to calibrate each other fine. I even calibrated a -15C dark with a -10C dark, and it still seemed to work well. I suspect there is probably some remnant pattern there that would show up in a deep integration of lots of dark frames if I did that, and I'll try to integrate a master for each temp range tomorrow and subtract that from each individual dark to see how it goes as well. But, your darks so far, at least on themselves, seem to be effective.

 

I am curious if there may be some kind of signal attenuation once you get light into the frames. If you can get some lights at each of these temps, doesn't even matter if they are night sky, they just need faint signals like a night sky, I can do some more testing to see how effective a single dark and master dark are at calibrating your lights.



#135 rockstarbill

rockstarbill

    Fly Me to the Moon

  • *****
  • Posts: 6155
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 29 May 2018 - 11:49 PM

Yeah when I tested this camera after getting it, the darks seemed to calibrate and work fine. I dont have any light frames with this camera yet. This weekend has a slight chance to be okay, if it is I will see about throwing the camera on the Newt to get a few frames taken. I cant promise the camera will be spaced correctly (it wont be), but I can probably get a few frames.



#136 glend

glend

    Surveyor 1

  • *****
  • Posts: 1875
  • Joined: 04 Feb 2014

Posted 30 May 2018 - 12:06 AM

Jon have you asked Sam (ZWO) about the cooling of this sensor? It might be time to get a response from them. I was interested in this camera but partial cooling is an issue imho.

 

Secondly, reading through this thread I see images that I would regard as very long sub for present day CMOS thinking (almost a return to CCD sub timing), and I get that it is facilitated by the deep wells on this camera, but for many suburban OSC imagers, (that may have Sky Glow issues) they have embraced flexible short sub, low noise cameras like the ASI1600, thus long sub imaging is not an option unless they are going to a dark site a lot, but then they could  take the ASi1600 there as well.  I also question the need for clean 300s + Darks on a OSC, unless they are using it at a dark site to exploit the well depth, yes on a Mono shooting narrowband from anywhere it's important? In that sense, for the suburban imagers, who make a few dark site trips, the ASI1600MC-C is a better option imho.  You just shoot more short subs with the 1600. 

 

Just where is the place for the 294MC in the equipment mix?


Edited by glend, 30 May 2018 - 01:17 AM.


#137 bulrichl

bulrichl

    Vostok 1

  • -----
  • Posts: 139
  • Joined: 27 May 2018
  • Loc: La Palma (Canary Islands)

Posted 30 May 2018 - 01:51 AM

Hi Jon,

 

 

Anyway...based on a preliminary examination, the master is not well matched, so it is not surprising that there would be remnant pattern.

I am very sorry, I just noticed that I made a mistake and uploaded the wrong MasterDark (300 s). This is NOT the MasterDark that I used for the calibration of the lights in the zip file. Within a short time I will upload a new zip file with the corresponding MasterDark (150 s) and a few dark frames from which this MasterDark was built. I apologize for the confusion.

 

Bernd



#138 bulrichl

bulrichl

    Vostok 1

  • -----
  • Posts: 139
  • Joined: 27 May 2018
  • Loc: La Palma (Canary Islands)

Posted 30 May 2018 - 02:44 AM

OK, this is the new link to the correct MasterDark (150 s) and 25 individual dark frames (150 s):

 

https://www.dropbox...._Darks.zip?dl=0

 

Bernd


Edited by bulrichl, 30 May 2018 - 02:45 AM.


#139 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23600
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 30 May 2018 - 06:23 PM

Thanks, Bernd.

 

I played with your data a little more. I still need to fiddle with the new darks, but with the integrations you shared earlier, with a high density DBE and some ACDNR after stretching, the data is still pretty good:

 

UclgIsQ.jpg

 

rloKPR4.jpg

 

The background could definitely be better, and if you have IFN in the frame, the high density DBE technique may not work, but I think it can salvage a lot of images.



#140 bulrichl

bulrichl

    Vostok 1

  • -----
  • Posts: 139
  • Joined: 27 May 2018
  • Loc: La Palma (Canary Islands)

Posted 31 May 2018 - 03:19 AM

I played with your data a little more. I still need to fiddle with the new darks, but with the integrations you shared earlier, with a high density DBE and some ACDNR after stretching, the data is still pretty good:

By the way, you did a great job in noise reduction - I will have to visit your website again and take a deeper look into your NR techniques...

 

 

The background could definitely be better, and if you have IFN in the frame, the high density DBE technique may not work, but I think it can salvage a lot of images.

Yes, at the risk of removing true signal as well you can get rid of this artifact by the background modelization techniques. However, as you stated before (#115): "The thing that is curious to me is, the pattern seems to be quite consistent, and yet it is still not removed with dark calibration. I am having a hard time understanding why that is. If the pattern is so consistent that it shows up the same night after night, even days or weeks apart, why can it not be calibrated out?"

 

My primary concern:

1) Is it possible to take better calibration frames somehow?
2) is it possible to enhance the calibration process in some way?
3) is this remaining background artifact not removable by the calibration process at all?

 

I am curious about your judgement of the 150 s darks.

 

Bernd



#141 glend

glend

    Surveyor 1

  • *****
  • Posts: 1875
  • Joined: 04 Feb 2014

Posted 01 June 2018 - 01:22 PM

The key questions and problem related to background artifact are discussed pretty well on the ZWO forum page for the 294. From my reading they (ZWO, Jose) have no idea what is causing it so the chance of a quick fix is diminished.  I guess with so many new models launched successfully in the last couple of years, they were bound to hit a dud at some point. Too bad i was hoping to buy one, but with uneven, or partial cooling as well, they're not going to fix that with a driver.



#142 John Miele

John Miele

    Gemini

  • *****
  • Posts: 3415
  • Joined: 29 May 2005
  • Loc: North Alabama

Posted 01 June 2018 - 01:28 PM

Any idea if the ASI071 pro has similar problem? I hope not because I just bought an ASI071 Pro but have not had a chance to use it yet!...john



#143 jgraham

jgraham

    ISS

  • *****
  • Posts: 20195
  • Joined: 02 Dec 2004
  • Loc: Miami Valley Astronomical Society

Posted 01 June 2018 - 04:53 PM

It's not a problem until it's a problem. Give your camera a chance and see how it works for you. I have used tne ASI071MC, ASI071MC Pro, and ASI294MC Pro with no problems, but that is with my gear under my sky conditions with my processing.

 

If it t'ain't broke, don't fix it.


  • TeamHawkins and billc2016 like this

#144 tjugo

tjugo

    Apollo

  • *****
  • Posts: 1432
  • Joined: 06 Nov 2007

Posted 01 June 2018 - 07:06 PM

Any idea if the ASI071 pro has similar problem? I hope not because I just bought an ASI071 Pro but have not had a chance to use it yet!...john

AFAIK, no.

 

Cheers,

 

Jose



#145 andysea

andysea

    Soyuz

  • *****
  • topic starter
  • Posts: 3516
  • Joined: 03 Sep 2010
  • Loc: Seattle, WA

Posted 01 June 2018 - 07:20 PM

Even with the background artifact issue, the sensor is still pretty great. Let's not forget how quickly it goes deep, especially for an OSC sensor.

 

It's still unclear to me what is the cause of the artifact and why it isn't calibrated out. However QHY will release their version of the 294 and it is possible that they will capitalize on ZWO's experience is having and try to solve the problem. FLI also has this sensor in their pipeline but no release date yet....and the cost will be much higher.

 

If I were looking at this camera now I would probably wait for the QHY version and try that one.



#146 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23600
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 01 June 2018 - 08:54 PM

By the way, you did a great job in noise reduction - I will have to visit your website again and take a deeper look into your NR techniques...

 

 

Yes, at the risk of removing true signal as well you can get rid of this artifact by the background modelization techniques. However, as you stated before (#115): "The thing that is curious to me is, the pattern seems to be quite consistent, and yet it is still not removed with dark calibration. I am having a hard time understanding why that is. If the pattern is so consistent that it shows up the same night after night, even days or weeks apart, why can it not be calibrated out?"

 

My primary concern:

1) Is it possible to take better calibration frames somehow?
2) is it possible to enhance the calibration process in some way?
3) is this remaining background artifact not removable by the calibration process at all?

 

I am curious about your judgement of the 150 s darks.

 

Bernd

So, your 150s darks seem to work well. They produce a clean master, and I cannot really see the pattern in the master. The glows are fairly strong, though (and, I guess I wasn't aware, there are apparently two starbursts), so that makes it tough to evaluate anything else about the darks. I am currently calibrating each individual dark frame with the master, and I'll then integrate the calibrated darks to see what I get.

 

I did experiment a bit with some of the lights I calibrated with the master dark. I tried VNG, Bilinear, and SuperPixel. I noticed that with VNG and Bilinear, I could see a bit of pattern around the periphery, which is probably the beginnings of the issue being discussed in this thread. With SuperPixel, however, I could barely see the same pattern, and the color of the background was much more uniform.

 

So, I am curious, could you process your entire stack with SuperPixel, and integrate the result, and see what you get?

 

I was also interested in the color differences between VNG and Bilinear...very stark green cast with VNG, much more neutral color with Bilinar. Bilinear is not really the best way to demosaic, but the color differences were interesting.

 

I would like to try a bayer drizzle as well, and see what the outcome is.



#147 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23600
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 01 June 2018 - 09:23 PM

Another update. Are all of you guys using an output pedestal when calibrating? I originally calibrated Bernd's darks without an output pedestal, just to see what would happen. I disabled optimzation and all of that, and the dark frames did NOT calibrate nor integrate properly, and there was definitely remnant pattern (including remnant glow). I added an 800DN output pedestal when calibrating, and then everything calibrated fine. I am not sure if I have enough dark frames here to say for sure there is no remnant pattern outside of horizontal banding, but currently, the master dark looks great with the exception of the horizontal banding.

 

If you guys are not using an output pedestal when calibrating your lights, I highly recommend you try with one. I would be very curious to see what happens if you do.



#148 mikefulb

mikefulb

    Surveyor 1

  • *****
  • Posts: 1672
  • Joined: 17 Apr 2006

Posted 01 June 2018 - 09:38 PM

Here is a copy of BPP with Pedestal I put out awhile ago:

 

https://github.com/m...lb/BPP_Pedestal


  • Jon Rista likes this

#149 jgraham

jgraham

    ISS

  • *****
  • Posts: 20195
  • Joined: 02 Dec 2004
  • Loc: Miami Valley Astronomical Society

Posted 02 June 2018 - 07:21 AM

I went back and reviewed this entire thread just to be sure that I wasn't missing something. Odd colors in the background aren't unusual for any camera when you stretch the data very aggressively. One approach is to separate the luminance and the individual colors, process these, and then put them back together as an LRGB. You'll often find that there is an offset in one of the colors near the black point, but with very strong curves you gotta tread lightly on the processing. However, this is not necessarily an issue specific to the ASI294MC Pro. It is a little less common in monochrome imaging due to the way that the data is processed and you can end up in the same place by applying similar processing to color images.

 

Food for thought.


  • ito304 likes this

#150 bulrichl

bulrichl

    Vostok 1

  • -----
  • Posts: 139
  • Joined: 27 May 2018
  • Loc: La Palma (Canary Islands)

Posted 02 June 2018 - 07:33 AM

So, your 150s darks seem to work well. They produce a clean master, and I cannot really see the pattern in the master. The glows are fairly strong, though (and, I guess I wasn't aware, there are apparently two starbursts), so that makes it tough to evaluate anything else about the darks. I am currently calibrating each individual dark frame with the master, and I'll then integrate the calibrated darks to see what I get.

 

I did experiment a bit with some of the lights I calibrated with the master dark. I tried VNG, Bilinear, and SuperPixel. I noticed that with VNG and Bilinear, I could see a bit of pattern around the periphery, which is probably the beginnings of the issue being discussed in this thread. With SuperPixel, however, I could barely see the same pattern, and the color of the background was much more uniform.

 

So, I am curious, could you process your entire stack with SuperPixel, and integrate the result, and see what you get?

 

I was also interested in the color differences between VNG and Bilinear...very stark green cast with VNG, much more neutral color with Bilinar. Bilinear is not really the best way to demosaic, but the color differences were interesting.

 

I would like to try a bayer drizzle as well, and see what the outcome is.

According to your suggestion I just debayered my calibrated data as well with the Bilinear as with the SuperPixel interpolation method. The result is: It makes no difference in respect to the colored background artifact in the resulting integration.

 

I never tried bayer drizzle, so I have no experience how to apply it to my data.

 

Another update. Are all of you guys using an output pedestal when calibrating? I originally calibrated Bernd's darks without an output pedestal, just to see what would happen. I disabled optimzation and all of that, and the dark frames did NOT calibrate nor integrate properly, and there was definitely remnant pattern (including remnant glow). I added an 800DN output pedestal when calibrating, and then everything calibrated fine. I am not sure if I have enough dark frames here to say for sure there is no remnant pattern outside of horizontal banding, but currently, the master dark looks great with the exception of the horizontal banding.

 

If you guys are not using an output pedestal when calibrating your lights, I highly recommend you try with one. I would be very curious to see what happens if you do.

I don't quite understand this statement. I never used an output pedestal and didn't experience any drawback so far. When my light frames are calibrated with my 150 s MasterDark and the MasterFlat, there is no severe clipping to zero. I have checked this again: with the set of 95 calibrated light frames the average number of pixels that are clipped to zero is 29 pixels. Given that the sensor has 11.7 million pixels, this is negligible. I can only imagine that you are referring to the wrong MasterDark (300 s) that I uploaded by mistake in the first zip archive. The 300 s MasterDark indeed could have produced some clipping when applied to the 150 s light frames.

 

Anyway, I did as you suggested and performed the workflow with an output pedestal of 800 DN applied. Again there was no visual difference in the integration result, and their histograms (workflow with / without pedestal) are barely discernable. So this is also not the cause of the colored background artifact.

 

Bernd

 

 




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