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

Full Frame Mono Camera Coming Soon - QHY600 (IMX455)

  • Please log in to reply
378 replies to this topic

#301 FredOS

FredOS

    Vostok 1

  • -----
  • Posts: 118
  • Joined: 16 Feb 2017

Posted 12 September 2019 - 11:32 AM

Sure but where ?

#302 ManuelJ

ManuelJ

    Apollo

  • -----
  • Posts: 1450
  • Joined: 19 Dec 2005
  • Loc: Madrid, Spain

Posted 12 September 2019 - 11:44 AM

Sure but where ?

https://www.flickr.c...840882/sizes/o/

https://www.flickr.c...880912/sizes/o/

https://www.flickr.c...350926/sizes/o/



#303 rockstarbill

rockstarbill

    Fly Me to the Moon

  • *****
  • Posts: 6315
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 12 September 2019 - 01:23 PM

Bill,
Do you know if the AP130 GTX and Quad works with a 43mm diagonal ? It seems that with the FLI16200 we are close to the limit. There is the option of replacing the Quad with the Prime Focus Field Flattener but I don’t know how good it is (and you have a slower system except if you bin).

The Quad TCC has a 50mm corrected field, so yes it would work fine. The field flattener is 65mm. 

 

About the Sony IMX455 -- does anyone know if this camera supports hardware 2x2 binning? If so, this camera may be added to the herd once ASI releases theirs, and some folks get some  hands on time with it. It would not replace any of my existing cameras, as I have purposes for those, but it would be a good addition.

 

The posted dark frame, shows no amp glow, but does show some overscan area, which the QHY driver should be able to remove.



#304 kingjamez

kingjamez

    Vanguard

  • *****
  • Posts: 2032
  • Joined: 03 Oct 2006
  • Loc: Fairfax, VA

Posted 12 September 2019 - 01:28 PM

The RASA 11 is optimized for a 43.3mm image circle and is advertised to support up to 52mm. Seems like a pretty potent combo with the IMX455.



#305 Morefield

Morefield

    Vostok 1

  • -----
  • Posts: 129
  • Joined: 01 Jun 2014
  • Loc: Portland, OR

Posted 12 September 2019 - 04:24 PM

The RASA 11 is optimized for a 43.3mm image circle and is advertised to support up to 52mm. Seems like a pretty potent combo with the IMX455.

And it seems anyone with a FSQ106 would buy this in a heartbeat.  I mean you've already shown you'll spend $6000 to get top equipment and  with an 88mm corrected circle this is a piece of cake.  

 

I love the idea of running it cropped/sub-framed for galaxy season too.  


  • kingjamez, psandelle and leviathan like this

#306 Gene3

Gene3

    Viking 1

  • -----
  • Posts: 559
  • Joined: 02 Feb 2016
  • Loc: Del Mar, CA

Posted 12 September 2019 - 05:32 PM

I respectfully disagree with this notion. The KAF-16200 is 34.6mm diagonal. Many scopes that amateurs use have 40mm or less in terms of available imaging circle. To fill the full frame 36x24 chip the cost of the optics will be much higher, while the chip is more expensive. 

 

People do realize this right? A $5000 CMOS camera (just the camera) vs a $4500 all in one 16200 solution on the market (camera, wheel, OAG).

 

Now go price OAG's with aperture enough for that chip, wheels for the filters... then come back and we can chat. flowerred.gif

Couple thoughts on this topic:

 

I do see a trace of vignetting with my full frame (QHY367C) camera on both my scopes which have a 42mm image circle, I can live with it. Therefore for my set up the QHY600 will not be any worse.

 

What I see as a drawback is that the QHY600 is not an "all in 1" design like my QHY16200a. The idea of dealing (cost, set up time, additional backfocus requirements, and connectivity), involved with a separate filter  wheel and guider are not something I am willing to go through for the QHY600. 

If and when QHY comes out with an "all in 1" design for the QHY600 then I will seriously consider getting one to replace my QHY16200a. 

 

Best Regards,

Gene


  • psandelle and rockstarbill like this

#307 rockstarbill

rockstarbill

    Fly Me to the Moon

  • *****
  • Posts: 6315
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 12 September 2019 - 07:54 PM

And it seems anyone with a FSQ106 would buy this in a heartbeat.  I mean you've already shown you'll spend $6000 to get top equipment and  with an 88mm corrected circle this is a piece of cake.  

 

I love the idea of running it cropped/sub-framed for galaxy season too.  

The A-P130 GTX with the Quad TCC would outperform the native FSQ with this chip. Oddly enough, a TEC140 with the Quad TCC shows identical performance to the FSQ.

 

Using the standard Performance = aperture^2 * scale^2 * QE formula we see this, with these three configs:

 

IMX_Perf.JPG

 

The FSQ106 Reduced, is really hard to beat though with any chip/config in this measurement. The 44mm illuminated field of the common reducers would be a tight fit, making the 645 Reducer the better bet. At any rate, I wanted to include that. smile.gif

 

IMX_2.JPG


Edited by rockstarbill, 12 September 2019 - 08:22 PM.

  • deepanshu29 and motab like this

#308 rockstarbill

rockstarbill

    Fly Me to the Moon

  • *****
  • Posts: 6315
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 12 September 2019 - 08:18 PM

https://www.onsemi.c...KAF-16200-D.PDF

 

According to this reference it has 69db.

 

But OK, 77db - 12.9bits vs 13.6bits.

The FLI 16200 camera has 6e noise and a 44k full well. How they manage that wizardry is beyond me. lol.gif



#309 bmhjr

bmhjr

    Apollo

  • ****-
  • Posts: 1254
  • Joined: 02 Oct 2015
  • Loc: Colorado

Posted 13 September 2019 - 01:30 PM

Alright, I dropped some frames casually into Pixinsight and Ran Basic CCD parameters:

attachicon.gif Screenshot_5.jpg

 

Seems that the readout noise is lower than the diagrams would anticipate. 

Is this the Dynamic Range that is expected with this camera?  I am curious how those numbers are interpreted.



#310 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23721
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 13 September 2019 - 01:50 PM

Is this the Dynamic Range that is expected with this camera?  I am curious how those numbers are interpreted.

I'm a bit suspicious...that gain of 0.074e-/ADU seems insanely high... That said, 11 stops at a gain as high as that is not bad.


  • bmhjr likes this

#311 cabfl

cabfl

    Lift Off

  • -----
  • Posts: 20
  • Joined: 21 Oct 2018

Posted 16 September 2019 - 02:32 PM

 

Today we publish anothet readout mode of QHY600. In this mode we can get more fullwell of it. Unbelievable!

In this mode the readout noise will higher also. But it give more flexibility for QHY600 for different conditions.

6459b818a94338013e51f4d13ea03a4bo.jpg


Edited by cabfl, 16 September 2019 - 02:33 PM.


#312 cabfl

cabfl

    Lift Off

  • -----
  • Posts: 20
  • Joined: 21 Oct 2018

Posted 16 September 2019 - 05:25 PM

 

Big update from Hongyun Qiu, about the 3 read modes of the qhy600

9738a0a904b1c847d8afdeb490ea2a04o.jpg



#313 deepanshu29

deepanshu29

    Vostok 1

  • -----
  • Posts: 134
  • Joined: 16 Jul 2017

Posted 16 September 2019 - 05:36 PM

I am confused. Both, mode #0 and mode #2 have 80Ke- full well capacity? What exactly is the difference between two?



#314 TxStars

TxStars

    Surveyor 1

  • *****
  • Posts: 1868
  • Joined: 01 Oct 2005
  • Loc: Lost In Space

Posted 16 September 2019 - 10:48 PM

So lower read noise at a lower gain in "Mode 0 vs Mode 2",  but it looks to have a slightly lower dynamic range in Mode 0



#315 Coconuts

Coconuts

    Mariner 2

  • *****
  • Posts: 215
  • Joined: 23 Sep 2012

Posted 18 September 2019 - 06:12 AM

Folks considering cameras based on this sensor (I'm in that bunch) might want to check out Jim Kasson's blog, here:

https://blog.kasson.com/category/a7riv

 

Jim is a retired EE who goes deep into camera testing, and in the above posts he is putting Sony's new a7r IV through his testing methodology.  This camera, which will begin retail shipments any day now, uses the consumer version of the IMX455; it is nearly identical, except for a bunch of added autofocus pixels, and of course Sony's sensor processing firmware.

 

Jim Kasson's blog not only confirmed the star-eating issue for the new a7r IV, but he also tested a LOT of other aspects, many of which are relevant to astrophotography. He goes deep in deriving full well capacity and read noise vs ISO setting, fixed pattern noise, read noise versus self-heating, and a number of other parameters that we care about.  I'm still digging through all of his findings, but it does seem clear that Sony's claim of 15 stops of dynamic range is essentially marketing nonsense, and that the actual dynamic range of the a7r IV is only marginally better than its a7r III predecessor, although both are still quite good in DR compared to other cameras.

 

Direct data from QHY is obviously superior, but some triangulation from Kasson's testing suite can only help as we figure out what to make of Sony's latest generation of BSI CMOS sensors.  By the way, the next size up in the Sony triad of sensors, the IMX461 (100 Mpx, 33mm x 44 mm, with a 55 mm diagonal) will launch this November in Fuji's medium format GFX 100 ($10K).  I have asked both QHY and ZWO if they plan a consumer astrocamera on this larger sensor; both said "no plans at the moment".  QHY sells scientific cameras with both the IMX461 and its even bigger brother, the Sony IMX 411 (67 mm diagonal), but those are big, heavy, and very expensive cameras aimed at the scientific market.

 

All the best,

 

Kevin


Edited by Coconuts, 18 September 2019 - 06:21 AM.


#316 bortle2

bortle2

    Sputnik

  • -----
  • Posts: 25
  • Joined: 18 Sep 2019

Posted 18 September 2019 - 01:17 PM

Folks considering cameras based on this sensor (I'm in that bunch) might want to check out Jim Kasson's blog, here:

https://blog.kasson.com/category/a7riv

 

<...>

 

Jim Kasson's blog not only confirmed the star-eating issue for the new a7r IV, but he also tested a LOT of other aspects, many of which are relevant to astrophotography.

 

<...>

 

Direct data from QHY is obviously superior, but some triangulation from Kasson's testing suite can only help as we figure out what to make of Sony's latest generation of BSI CMOS sensors. <...>

 

Kevin, I agree with everything you say about Jim Kasson: he's indispensable source of the most accurate data on sensors and their performance in the photographic community. I do not remember him posting anything that would later turn out to be not true.

 

That said, I respectfully disagree with your suggestion to do "triangulation" of his and QHY's data... First, at this point we do not even know which sensor (IMX551 or IMX455) was used in A7mkIV; all I was able to find was hearsay, no official data (please correct me if I'm wrong). Second, even if it's IMX455, we do not know which mode (of the above-mentioned three) is used for stills in the A7.

 

Last but by no means least, the main differentiator is always firmware. You mentioned the infamous star eater issue: that's purely a firmware problem, and has nothing to do with how QHY and/or ZWO cameras will perform. Now, to be clear: I'm not saying you implied that it could be an issue. My point is these days firmware has major impact on RAW files, to the point those files can no longer be considered truly RAW, whatever camera manufacturers say.

 

What Jim Kasson is measuring is the output of Sony firmware, not that of the sensor per se, and that "little" fact should be always remembered. Ans it's all too easy to forget. Fun fact: "star eating" of the Sony A7mk3 was decreased (though not eliminated) with a firmware update.

 

Now, why do I bother writing all this? The thing is that once some bogus idea is let loose, it will spread all over the Internet like fire, and one's lifetime won't be enough to debunk all sorts of misconceptions. An example: somebody somewhere said that Sony's compressed RAWs affect "fine" color reproduction, and now it stuck in quite a few people's heads and refuses to die -- despite endless blind tests demonstrating that the *only* noticeable adverse effect is pixel-size artifacts in very high contract areas.

 

Jim's data can be instrumental in, say, figuring out who did better job with essentially the same sensor: Sony with their A7mkI, or Nikon with their D800/810[A]. I wish there was a resource similar to blog.kasson.com dedicated to astrophotography sensors and cameras, but alas... And, since applicability of Jim Kasson's data to QHY/ZWO cameras is IMHO very limited, I'd stick with the official data from Sony/QHY/ZWO/etc., however limited they currently are.


  • psandelle, rockstarbill, ezwheels and 2 others like this

#317 cabfl

cabfl

    Lift Off

  • -----
  • Posts: 20
  • Joined: 21 Oct 2018

Posted 20 September 2019 - 01:42 PM

qhy is analyzing the sensor QE

 

436812a6484567859b8c8b10338fdc4bo.jpg



#318 ezwheels

ezwheels

    Messenger

  • -----
  • Posts: 472
  • Joined: 04 Jul 2017
  • Loc: Oakland, CA.

Posted 20 September 2019 - 03:04 PM

Looks to me like they are basically saying it is way better than we imagined, so we better make sure we are right.



#319 bortle2

bortle2

    Sputnik

  • -----
  • Posts: 25
  • Joined: 18 Sep 2019

Posted 20 September 2019 - 09:00 PM

Looks to me like they are basically saying it is way better than we imagined, so we better make sure we are right.

Well, it's not just "way better", it's also a bit of a mess...

 

They have this Sony chart for "relative QE of IMX455" on their site: https://www.qhyccd.c...26021207472.jpg

 

And now they say that measured QE at 500nm is 94%, while measured QE at Ha (656nm) is 75%. Looking at the Sony chart (on QHY site!), we see that peak QE is at appr. 535nm, while relative QE at 500nm is appr. 97%, and relative QE at 656nm is some 82%. So, measured QE of 94% at 500nm translates to peak QE of 96.91%, while measured QE of 75% at 656nm translates to peak QE of 91.46% -- if Sony chart for relative QE is to be believed (and I see no reason not to; and yes, I know, precision to the hundredth is... optimistic, so round it up or down to your liking).

 

So... Basically, they not only got something that's "way better", but something that does not agree with available Sony specs. No wonder they're going to re-run the tests.

 

Interestingly, they only say "Looks too high in 500nm" (again, translates to peak QE of almost 97%)... Does it mean they're OK with 75% at 656nm? That would mean peak QE of more than 91%, is that what they'd expect? grin.gif

 

Peak QE for IMX455-based cameras is all over the place... So far, we have this:

 

  1. Initial estimate: 80%. That's still current for ASI6200MM, BTW.
     
  2. 87%? From https://m.facebook.c...875215845915023: "So we have a guess that the QHY600 QE should between 84% to 89% and more close to 89%"
     
  3. And now this (^^^). 91%??

 

I hope to live long enough to witness QHY coming up with "final" peak QE number for their QHY600 lol.gif Seriously though, even "just" 80% is great, anything higher would be just amazing, so let's hope these recent measurements are indeed an indication of things to come...



#320 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23721
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 20 September 2019 - 09:17 PM

I am confused. Both, mode #0 and mode #2 have 80Ke- full well capacity? What exactly is the difference between two?

The slope of the FWC plots is very different. While Mode #0 starts out at 80k, it falls off much more quickly as gain is increased than Mode #2. So basically, Mode #2 would deliver a much larger FWC at higher gains than Mode #0. 

 

Also note that the plots the way they are currently represented are kind of misleading if you are not paying attention to all the details. Mode #0 the gain axis caps out at 100, while Mode #2 it caps out at around 140. So they expanded the axis for Mode #2, but kept the physical width of the plot. So that makes the Mode #2 FWC plot slope look steeper than it really is compare to Mode #0. So Mode #0 FWC falls off MUCH more quickly than Mode #2. 

 

EDIT:

 

The image isnt quite large enough to get exact numbers, but it looks like the X-intercept for the Mode #0 FWC LCG plot (where it reaches zero) is around Gain 40. At Gain 40 in Mode #2 the FWC is still ~52k!! Even with the HCG Mode #0 FWC plot it reaches zero at 60k, while Mode #2 is still around 37k. Mode #2 doesn't reach zero until Gain 140, but following the original slope (before it flattens out) Mode #2 would intercept the X axis (Gain) at a bit past 100. So Mode #2 delivers much, much larger FWC at significantly higher  gains than Mode #0. 

 

EDIT 2:

 

Another interesting thing is, despite the high read noise, the dynamic range at Gain 0 for Mode #2 is 13.5 stops, which is higher than the maximum DR for Mode #0 at any gain, which is slightly above 13 stops. DR remains fairly high for Mode #2 well beyond the point where FWC gets very small with Mode #0. For certain applications, the very high DR and very large FWCs of Mode #2 could be quite useful. (Not necessarily pretty pictures, but people doing other astronomy things could benefit a lot from the high FWC mode.) 


Edited by Jon Rista, 20 September 2019 - 09:30 PM.

  • deepanshu29 likes this

#321 rms40

rms40

    Viking 1

  • *****
  • Posts: 643
  • Joined: 15 Apr 2016
  • Loc: NW of Harper, TX

Posted 21 September 2019 - 08:31 AM

Anyone have an idea of how the QHY600, CFW3 - large, and OAG attach? Are they bolted or threaded together? Would you use the medium or large QHY OAG?

 

One thing I really like about the current CMOS sensor offerings is the fast readout and download times. That sure makes focusing easier and allows more frames in a night. Can anyone tell me how long it takes between exposures (readout and download) on the 16803? will the QHY600 be much faster?

 

Thanks, Randall



#322 lukepower_2

lukepower_2

    Lift Off

  • -----
  • Posts: 15
  • Joined: 05 Mar 2019

Posted 21 September 2019 - 08:50 AM

Anyone have an idea of how the QHY600, CFW3 - large, and OAG attach? Are they bolted or threaded together? Would you use the medium or large QHY OAG?

 

One thing I really like about the current CMOS sensor offerings is the fast readout and download times. That sure makes focusing easier and allows more frames in a night. Can anyone tell me how long it takes between exposures (readout and download) on the 16803? will the QHY600 be much faster?

 

Thanks, Randall

Hi Randall,

 

I have the QHY600M with the CFW3L, and they fit. The connection is made with a a nosepiece on the camera-side, which is inserted in a ring on the filter wheel side, and then locked in place with three screws pushing from around the ring towards the camera nosepiece. Not really perfect for my taste, and QHY told me they are working on an alternative where they would use the threaded holes which are also on the filter wheel plate.

 

I have a FLI Proline 16803 which takes about 5 seconds for full-frame download in fast mode (8Mhz) or around 30 seconds in low noise mode (1MhZ), so yes the QHY600 is WAY faster in that regard.

 

Hope this helps

Lukas


  • psandelle likes this

#323 rms40

rms40

    Viking 1

  • *****
  • Posts: 643
  • Joined: 15 Apr 2016
  • Loc: NW of Harper, TX

Posted 21 September 2019 - 01:50 PM

Lukas, That is just what I wanted to know. Thanks.

 

The filter wheel connection sounds like a potential problem. I have one the early QHY EFWs and it has the same type connection with long threaded thumb screws. From what I remember, I was able to use a threaded connection with it instead of the thumb screws. I don't use it anymore because it is only a 5 position wheel.

 

The CWF3 connection sounds like something QHY should improve. A bolted connection with the holes would be my preference.

 

I do believe that the QHY600 will be a popular product. Their OAG looks pretty good too. I just hope they will have a stiff, robust connection all the way from scope to camera. It would be great if they offered a package deal with camera, EFW and OAG.

 

Randall

.


Edited by rms40, 21 September 2019 - 01:51 PM.


#324 cabfl

cabfl

    Lift Off

  • -----
  • Posts: 20
  • Joined: 21 Oct 2018

Posted 22 September 2019 - 12:38 PM

https://www.qhyccd.c...PmsyDvQ2-z14Gb8

 

 

 

(Notificaition: This test is just based on our experiment. The final data looks a little too high, so that we plan to find another sensor to do a test again. The absolutely QE of ICX694 is based on the internet .So we do not know if this QHY22 QE is same with it because each sensor may have a little difference even in the same model. The system gain of this QHY22 is carefully tested, the ADU per second of the light can be trusted because the linear fitting curve is R^2=0.9999. )

The QHY600 uses a SONY IMX455 sensor.  There is an official release of the QE curve for this sensor by Sony.  It is a relative curve and does not indicate the absolute QE.  However, we can determine the actual QE if we can fix one or two points of the curve with reliable absolute data.  Then we can re-map this curve to absolute values.

 

In order to determine the absolute QE data of the QHY600 camera we can compare to a camera with known absolute QE data for a single wavelength.  We use a lens with a narrowband filter to capture an image with each camera against the same place of a uniform flat board, measure the ADUs, then multiply by the system gain to convert the data to photons and then divide by the pixel area of each camera.  In this way we can determine the QE of the QHY600 by comparing to the reference value from the camera with known QE:  Ratio=QE(QHY600)/QE(QHY22) . Then based on the absolute QE curve of the ICX694 sensor we can get the absolute QE of the QHY600 at this wavelength.

In order to avoid differences in any microlens angle effects for wide angle light paths we set the aperture of the lens to F11.

We use a 10nm Ha filter and OIII filter.  The reference camera selected in this experiment is QHY22 which uses the SONY ICX694 CCD sensor.  The QE can be found on various websites. We use the data in http://www.astrosurf...eras/index.html as reference data

 

From this curve we know the absolute QE of ICX694 is 67% at OIII and 58% at Ha.

Test conditions

 

QHY600 Gain=30 (mode 0) System gain=0.37e/ADU Pixel size=3.76um*3.76um Pixel Area=14.1376um^2
QHY22 Gain=0 System gain=0.29e/ADU Pixel size=4.54um*4.54um Pixel Area=20.6116um^2

 

Using a linear fit we determine the response of the QHY600 is 3726.3 ADU per second and the QHY22 is 5434.

 

Multiple by the system gain: The QHY600 is 1378.6e-/sec/pixel and QHY22 is 1576e-/sec/pixel.
And divide by the pixel area: The QHY600 is 97.51e-/sec/um^2 and QHY22 is 76.47e-/sec/um^2.

It shows on the same exposure and area, for the converted electron, the QHY600 has 1.27 times more than QHY22.  Since the QHY22 QE is 58% at Ha, the QE of QHY600 is 58%*1.27=74%

 

OIII Test Results with different exposure times

 

Following the same method as we did for H-alpha we find:

QHY600:  6458.3*0.37/14.1376=169.02e-/sec/um^2
QHY22:   8506.3*0.29/20.6116=119.68e-/sec/um^2

The ratio of QHY600 to QHY22 is 1.41. So, the QE of QHY600 in OIII is 67%*1.41=94.5%


Edited by cabfl, 22 September 2019 - 12:40 PM.


#325 kingjamez

kingjamez

    Vanguard

  • *****
  • Posts: 2032
  • Joined: 03 Oct 2006
  • Loc: Fairfax, VA

Posted 22 September 2019 - 12:50 PM

Their method seems sound. I'd want to do a few samples of both cameras and average out the data.

 

It's looking more and more like this camera will be a game changer. 

 

-Jim




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