New 16mp ZWO camera

[rustigsmed]

Looks very interesting esp for the price.
Qe to be determined still .. US$1180

http://astronomy-imaging-camera.com/products/asi-cooled-cameras/asi1600mm-cool/

Sensor: 4/3″ CMOS
Diagonal: 21.9mm
Resolution: 16Mega Pixels 4656×3520
Pixel Size: 3.8µm
Max FPS at full resolution :23FPS
Exposure Range: 32µs-2000s
Read Noise: 1.2e @30db gain
QE peak: TBD
Full well: 20ke
ADC:12bit
Interface: USB3.0/USB2.0
Adaptor: 2″ / 1.25″ / M42X0.75
Protect window:AR window
Dimensions: 78mm Diameter
Weight: 410g
Back Focus Distance: 6.5mm
Cooling: Regulated Two Stage TEC
Delta T: 40-45 below ambient
Cooler Power consumption: 12V at 2A Max


I see bintel have it at $1899

[Slawomir]

Looks very interesting, thanks Russell.

[Terry B]

I can't comment on this camera but a friend tried the 174 version of the camera and found the dark noise to be very high. Cooling made little difference as CMOS doesn't respond to cooling as well as CCD. He abandoned the camera.

[Shiraz]

thanks Russel. looks like it could be my next camera.
the sample -25C dark frame they provide is very quiet and the read noise is really low. Just hope the QE is pretty good, but if this is a back illuminated chip (which it may be) then that should be OK as well.

[Camelopardalis]

FWIW, the sensor sounds like it's from Panasonic as used in the Olympus OMD E-M10 MkII.

Definitely looks promising...read noise looks very low, not wildly above the Sony 224.

[rmuhlack]

From your previous discussion threads on camera sensitivity calcs, I figure this new camera will have at best 2/3rds of the sensitivity of the ICX694. The removable T2 spacer of this new camera means its minimum backfocus is only 6.5 mm, meaning that it would probably be possible to use camera lenses together with a filter wheel, potentially providing a cooled (high QE?) alternative to a DSLR for very wide field deep sky imaging. Is that where you would see this camera fitting in, or would you be looking to pair it with another scope setup (presumably also short FL) ?

[Shiraz]

Hi Richard. yes, it would be significantly less sensitive than the 694 (which I will definitely still keep for dim galaxy imaging). However, I have felt the need for a bit finer sampling on brighter targets at about this time of year, when seeing can occasionally get below 2 arcsec. The ZWO would provide finer sampling and, with low read noise, it should be usable with ~ 1 minute subs on my 250f4 scope on bright targets, so I would hope to see even better resolution than 2 arcsec using crude lucky imaging.

Then, as you noted, there is Slawomir's idea of using a large-pixel-count /small-pixel chip with a fast camera lens of about 200mm fl for widefield imaging - in theory, should be directly competitive with an 11002/FSQ combo

It also should be very effective for planetary imaging due to the low read noise and high framerates available with sub-image selection.

I had almost decided on a cooled ZWO178 (which is a much more radical change from the 694), but this new camera could well be a better bet. the main downsides would seem to be the reduced sensitivity re the 694 and the thought of processing more than 10x as much data. Regards Ray

[Ross G]

Hello,

This camera is looking good.
I do a lot of my imaging with camera lenses ( 35mm up to 300mm) on a QHY8L ccd.
The stars however can look a bit "blocky" which I think is due to the large pixel size of the ccd. So this camera with it's smaller chip, smaller pixels and 3x the number of pixels seems like a good solution.
Attractive price as well!
However I notice that it has 12 bit ADC as opposed to 16bit on my QHY8L.
Is this going to make a difference to my final image?

Thanks.
Ross.

[Shiraz]

it shouldn't make any difference.

most common cameras have around 11-12bits dynamic range (this camera has very slightly more than 12 bits, so it is better than many). The lower bits in 16 bit cameras are generally dominated by read noise, so they do not contribute anything to an image. 12 bits is sufficient to encode almost all of the available dynamic range from the ZWO camera, so no problems.

http://olympus.magnet.fsu.edu/primer/digitalimaging/concepts/dynamicrange.html
http://www.photometrics.com/resources/whitepapers/dynamic-range.php

[Ross G]

Thanks Ray.


Ross.

[alistairsam]

Im very interested in the OSC version of the camera for RGB data with my WOGT81 and they have a relative output/relative QE graph.
is that different from a QE graph that shows QE at different wavelengths?
http://astronomy-imaging-camera.com/products/asi-cooled-cameras/asi1600mc-cool/

Alistair

http://astronomy-imaging-camera.com/wp-content/uploads/ASI1600MC-QE.jpg

[StuTodd]

Looking forward to the announcement soon.
I see the ZWO website lists 2 adapters for US$35 and $38 (for the T2 to M43).
Nice for narrowband imaging?

AFAIK the 12 bit "inadequacy" can be easily overcome with stacking??

[StuTodd]

The ZWO website Q and A for this camera states now that 2x2 binning will output 10 bit only. Sensor is that of a Olympus DSLM: E-M1.

http://astronomy-imaging-camera.com/products/asi-cooled-cameras/asi1600mm-cool/

[rmuhlack]

deep sky lucky imaging looks very interesting but i have a few questions...how would deep sky lucky imaging work with a camera like this - Would one capture video as an AVI file and then use planetary imaging software (eg registax or autostakkert) for image analysis and initial stacking, or can it be done using more conventional deep sky processing software like pixinsight? what sort of sub times / frame rates would be necessary. Also if the sub times are short (say less than 1s) , then it would seem like there would be a colossal amount of image data to process if total integration is several hours or more... :question:

[Shiraz]

my thought was to use subs of maybe 1 minute or less and apply normal PI processing methods. The gain in resolution (if there is much) would be due to improved rejection of short bursts of poor seeing rather than freezing of the seeing, such as you get at high framerates. However, there are a few very bright targets (eg homunculus), where much higher framerates could be used and it would be interesting to see what AS!2 could do on gigabytes of video at (say) 1 sec framerate. Of course this all depends on the chip having good QE - no data on that yet.

Data quantity and processing time will be a major issue with such a big chip at high framerates - however, some bright targets will not need much overall time and subframe selection + 8 bit digitisation will be appropriate for some. Current planetary image sequences extend up to maybe 20,000 frames with up to 1mpixel frames, so large datasets are manageable. Even so, there will be an order of magnitude step up in processing time and data storage requirements. I hope to have a couple of quad core I7 machines available soon.

[Slawomir]

As long as electronics side of this camera is well designed, then this camera, to my knowledge, could potentially be a very neat CMOS-based cooled astro-dedicated product capable of imaging DSOs in a nice resolution. Hopefully many other designs/manufacturers will follow soon. How much would one need to pay for a camera with similarly-sized CCD sensor?

I will look forward to test reports with great interest :)

[glend]

I am considering ordering the ZWO ASI 1600MM-Cool(mono).
I have been watching the comments on this camera here and on CN, and it looks promising. It would be nice to have a smaller form factor camera after years of cooled dslrs.

[rmuhlack]

very impressive result here from the creator of AS!2 with the ASI1600MM: http://www.cloudynights.com/topic/536494-m51-in-poor-seeing-20001s-asi1600mm-cool/

[h0ughy]

you have to be kidding - WOW what a result (http://www.cloudynights.com/index.php?app=core&module=attach&section=attach&attach_id=693919) thanks for highlighting that link

this certainly changes things, its smooth, detailed yet raw. :eyepop:

and i might be able to get away with doing solar?:question:

[Atmos]

2000x1s subs! That's a pretty good result!

[glend]

The Beta Tester on the Cloudy Nights ASI1600 thread is reporting bad AMP Glow issues on long darks. ZWO sent him a new driver but I can't see that fixing what is usually a firmware issue. I am holding off on my order until I find out what is happening with this and if it is backed up by other testers.

[Camelopardalis]

Glen, many device drivers upload firmware to the device as part of their initialisation process...

[rmuhlack]

40GB of data for 34 minutes of integration...that's going to take data storage requirements to a whole new level !

[rustigsmed]

that is a crazy result for 1 second subs, what would you need for narrowband, 6 seconds??
would like to see what an actual sub looks like.

im guessing M51 is pretty bright but still. With 1 second respectable deep space images are we going to see a time where a large goto dob with 20 second subs will be fine for narrow band?

[glend]

Beta tester Tolgagumus has an image up on CN now, in the Beginners imaging section,

http://www.cloudynights.com/topic/535506-asi1600mmc-beta-test/page-11

it too is very large. He shot 125 x 60" subs, even with having to throw out some the folder was 20 gig. Final image is very nice.
Here is his Astrobin image:

http://www.astrobin.com/full/248375/None/

[glend]

I have gone ahead an ordered a ASI1600MM-Cooled from Bintel. I am told by John that they expect the first shipment from ZWO at the end of this week (today?), and that they hold enough pre-orders to pretty much run out the initial delivery.
Now I am looking for a 36mm filter wheel that suits (as I already have a set of 36mm narrow band filters and I don't wish to sell them). The Starlight Express wheel looks promising. :thumbsup:

[gregbradley]

I look forward to your review Glen. Its potentially a very nice camera.

Greg.

[glend]

Thanks Greg. I bought a QHY 5 position 36mm filter wheel today from Cyclops Optics in Hong Kong to go with the new camera. I have 36mm narrowband filters already so i did not want to downsize. Looks like the camera will screw right onto the filter wheel. I will probably sell my mono cooled 450D down the track if this ASI1600 is everything its cracked up to be.

[Atmos]

Glen, to run the QHY filter wheel via USB will are going to have to basically flip a switch in the clear electronic area so that your computer will read it as a serial unit. Takes 5 minutes and is dead easy.

I run mine via the ASCOM QHYCFW Serial Driver.

[glend]

Thanks Colin, that is just what i need to know, i was wondering about driving it. Cheers

[gregbradley]

Glen,

I wonder about the viability of this lucky imaging and using a camera like that with 60 second exposures etc. That is still a DSLR approach.

I think this camera should be used with a CCD imaging approach where you expose to get a good signal above noise and not overexpose the bright areas of your image.

Is it because the target audience are used to using their DSLRs or planetary cams and think that way because of that?

Luminance 6 x 10mins 1x1 binning and RGB 10 minutes 2x2 binning all at -20C would put it right next to KAF8300 images for a direct comparison.

With 1.2 electron noise it could be great. The 3.8 micron pixels are quite small but with QE of something like 65% then 2x2 binning could be pretty hot on many setups. You get more well depth and sensitivity and lower noise. That would be the go for narrowband. All 10minutes at 2x2 or longer if your tracking can handle it.

Greg.

[glend]

Greg, personally I don't like the term 'Lucky Imaging" as it seems to demean peoples approach to imaging. I am buying this camera as my main imaging camera for the next few years. Yes the pixels are a little small at 3.8 micron compared to my 5.2 micron Canon however, they are fine for imaging off my 1000mm focal length Mak-Newt, and as you say binning will give me a larger effective pixel size for use with my 1250mm focal length 10" newt or even my RC08.

I don't know that the 'target audience' is DSLR people but alot of DSLR people are looking at this camera as a way forward for a few reasons: first is cost, it undercuts entry level CCDs substantially and is really going to change the game so to speak; secondly this camera represents the next generation CMOS technology leading to sCMOS and replacement chips for the coming CCD retirement by major chip makers; finally is it allowing people to change the way they approach imaging (if they choose to)..
Personally my approach to this camera is that it will be used for everything I do. Narrowband it top of my list.
Since you mentioned the venerable KAF-8300 chip I thought I would share this chart I ran across comparing the ASI1600 with the KAF-8300 (a chip that has an entry point $1000 above the ASI1600):

http://i64.tinypic.com/2zf1h6a.jpg

[Shiraz]

that's the point of this camera Greg - there is almost no noise, so subs can go almost as short as you like and even a tiny amount of signal will still be above the read noise (and apart from shot noise in the signal, there is no other noise at all, since these devices have little dark current when cooled).

The "CCD imaging approach" still applies, but it can be vastly modified from the assumptions used in the days of the 10-20e RN CCD dinosaurs, where you needed to soak the things with 10+ minute subs to see anything at all above the read noise. Those older chips can still produce great results, but they cannot do it with short subs - the new CMOS ones can.

low noise CMOS is finally coming into the mainstream and we are nearly at the theoretical limits for detectors - woohoo :thumbsup:.

[glend]

There is an interesting post by mclark over on dpreview dealing with the SNR approach and 'lucky imaging'. His wobbling video examples are very interesting, and the high resolution results impressive.

http://www.dpreview.com/forums/thread/3992021

In my view "lucky imaging" is just as worthy (if not more) an approach as Adaptive Optics, and much cheaper. Why try to move surfaces to stablise an image when you can just align selected high rate exposures to produce a sharp stable result. As mclark at dpreview noted, dithering is a problem in high rate exposures due to the time taken to settle and you certainly could not do it on very short subs. This means to me that the use of flawed sensors that rely on dithering to process out the flaws probably is going to disappear as short sub video capture evolves around the new sensor generation.

[Shiraz]

Dithering may not be such an issue Glen, because you will not need to wait for a guider to settle - you won't need a guider at all if the subs are short enough. Just wait long enough for the mount to get to a new position (maybe a second?) and start imaging.

In fact, if we can reduce the exposures enough, mount quality becomes pretty much immaterial - something good enough to keep the scope off the ground and roughly pointing in the right direction should do. Expect an EQ6 renaissance and a lot of huff and puff from the top end suppliers.

In the words beloved of "consultants" from a decade or two ago, this is the foothills of a paradigm shift.

[glend]

Agreed Ray, it is a new approach, borrowed from the planetary guys but enabled for DSOs by new generation high sensitivity, low noise, sensors. And yeah my NEQ6 is looking more than adequate now. I wonder about the future of the high end CCD manufacturers who have had a gravy train of high margin products for a long time. Many tech changes over the years simply strand users on platforms with no value in the new world. They will hang on for awhile but the future is taking the market elsewhere.:welcome:
Early adopters always have some teething issues but i expect any problems can be managed. The owners of ZWO must be delighted with the market response, and their stratgy of leveraging their expertise in planetary high frame rate cameras coupled with DSO capable sensors is a winner. That business will certainly be growing.

[gregbradley]

Thanks Ray.

I am not sure, is this CMOS sensor considered to be an sCMOS?

My Sony A7r2 is backside illuminated, has copper wiring instead of aluminium, is 42.4mp for full frame size and very high QE.
To get that in perspective the cheapest backside illuminated CCD camera in the FLI range is around US$30K!!!

I read one test where read noise is about .3 electrons.

So yeah read noise in CCD is way high compared to the latest Sony Exmor R sensors.

Having said that, something in the implementation is off because it gets quite bad colour noise in the shadows of high ISO long exposure. Also a general light amp glow in the lower 1/3rd. Both of these are correctable.

The Olympus sensor sounds good. But every user of Micro 4/3rds sensor complains about noise kicking in even at lowish ISO levels. So it will have to prove itself as whilst read noise may be low there can be amp glow which is fine if its stable to dark subtract out. Some of these earlier ASI cameras from what I read did not have stable amp glow (perhaps because the temp regulation is not to a temperature but to an ambient, it needs to be regulated like an Astro CCD camera to within a close range of a target temp).

If you can dark subtract the amp glow and flat field it successfully it should be good.

Perhaps the PixInsight overscan bias technique may be useful with a camera like this.

Exciting times and hopefully it works out well. I personally would prefer a Sony Exmor R sensor like in the Fuji X series (16mp APSc) or 24mp APSc sized (around 25mm x19mm).

QHY list some Sony Exmor CMOS sensors on future models. They have large pixels and are large sensors. As you say Sony are quitting CCD to concentrate on CMOS so that is likely the way of the future. Sony mirrorless sensors have probably surpassed CCDs a while ago.

Greg.

[glend]

Greg, I was using sCMOS as a generic term for the future of CMOS and not claiming the territory for ZWO's ASI1600. I too look forward to the larger chip sizes. I talked to Sam at ZWO about amp glow, as a beta tester on CN had reported it on long subs, it was not too bad from what I saw. The beta tester reported that Sam had provided him with some sort of driver update that cleaned it up substantially. Sam had posted a 300" Dark on the webpage for the camera, before the driver update, and it is almost undetectable. Sure it might be an issue at 600" but I don't really see a need to expose that long with this camera, even in narrowband, and besides I can process it out. If that's the only issue I'll be happy.

[codemonkey]

Definitely an interesting camera. Further to Greg's point, does it have set-point cooling? Edit: Nevermind, I found it down the bottom on the specs, the cooling is indeed regulated. Unregulated cooling would have been a deal breaker.

Additionally, what's the implication of a rolling shutter on shorter exposures? I know naught about this, but I thought I read somewhere that rolling shutters can create a ghosting effect where the previous frame can be somewhat visible in the next frame?

The big FOV, low read noise and small pixels are otherwise pretty attractive to me. With a decent QE I could almost consider replacing my cam... although I'm not sure what the minimum filter size would be on this. 36?

[gregbradley]

Amp glow should subtract out with a good quality dark as long as the temperature is accurate. I suppose you could use scaled darks with a bias and that should also work.

In fact thinking about it if the temperature regulation is not to a setpoint then scaled dark subtracts would be the way to go.

Greg.

[Shiraz]

Pretty sure it is set point cooling.

the rolling shutter issue probably only applies to something that is moving rapidly. Hopefully most of the our targets will stay still.:rofl:

The example dark frame shows no amp glow that I can see

[glend]

The beta testers on CN report that it has set point cooling, however, remember it requires a 12v power supply with adequate amperage to get there. The power supply i believe is not supplied with the camera. This is not an issue as i have a couple of very good 6amp supplies that i use with my cooled DSLRs. Of course the TEC will only draw what it can handle so i know the supplies i have are adequate. In a dark site situation, on battery power, you probably need at least a 40AH battery for a nights imaging run, that ensures not discharging to low.

Now what exactly is their definition of set point cooling; is it set point against ambient or set point on the slider bar. My cooled dslrs are set point against ambient to a delta t of -30. I suspect the ASI1600s are setpoint against the slider bar position because i don't think the camera measures ambient. The result is a bit of mental gymnastics to work out the slider bar set point relationship with ambient results. I suspect that i will put in some form of ambient measurement near the camera to track ambient and i may use my dslr setpoint pwm switch to throttle the power to the TEC to get it to my externally measured ambient baselined setpoint. This is pretty easy to set up, well because i have the gear already, but some people may assume they are running at -25C when in effect the sensor is actually at 0C on a 25C night. Not necessarily a bad thing as long as you can cool to the sensors 'sweet spot', where bias frames equal long darks, which in the Canon 450D is 0C. Going deep below that is not necessary, so having temperature info is important.

[codemonkey]

12v is standard for cooled CCD cameras so that's not an issue. I've already got a three big batteries in my micro obs for running my gear, hooked up to a portable solar panel that I rotate over the batteries.

I wouldn't expect the ASI1600 to measure ambient: it should measure the sensor temperature and if it's regulated, keep it at that specified temperature, so long as it's capable of doing so. If it can't do that, I'm not sure I'm interested. I've never tried dark scaling, maybe it's fine, but it does make wary.

[Shiraz]

All their specs are stated at absolute temperatures and they claim regulated cooling (which I take to mean that it is set to a fixed temperature). Cannot think of any reason why they would want to do it differently from everyone else and assume that they put a temperature sensor on the chip and a feedback loop.

[gregbradley]

This question was asked of Sam on a Cloudy Nights thread whether its ambient or a setpoint. He answered ambient.

Greg.

[glend]

Yes Greg i now agree. I asked the question on CN about set point temperature and have had several answers confirming it is adjusting temperature on the driver slider. Apparently Sam has confirmed it there somewhere as well. That is good news.
It appears that individual deliveries are now happening in the US for those that ordered direct from ZWO.

[gregbradley]

Not many sleeps to go now Glen!

Greg.

[codemonkey]

Awesome. I'll definitely be keeping an eye on this one.

[gregbradley]

the rolling shutter issue probably only applies to something that is moving rapidly. Hopefully most of the our targets will stay still.:rofl:

As you say rolling shutter is only a problem for video of fast moving objects.
Perhaps it may impact planetary imagers and that was why it was brought up on Cloudy Nights? Don't planets sometimes flick around with the seeing?

Greg.

[Shiraz]

Back in the days when CMOS was about to take over from CCDs in planetary, there used to be a lot of waffle about rolling shutters being bad for planetary imaging. Then the ASI120 and QHY5L2 cameras came on the market - it turned out to be a total non-issue. Just another red herring from the CCD purveyors (and I guess we will hear a bit from them in the coming months as the CMOS chips show what alternatives might be possible in DSO imaging). :lol:

[glend]

There was a question about fast moving objects like comets, i think it was on the ZWO Q&A page.

[gregbradley]

Well I would be stunned if you got the jello wobbly effect of rolling shutter from something as slow as a comet. Examples of jello wobble usually are of an object like a tripod and then rapidly shaking the camera back and forth to create the bendy effect where the straight lines bend in video not stills.

No way is it going to be a problem. For one thing we do DSO imaging tracked.

Greg.

[Camelopardalis]

The IMX224 has a rolling shutter but you don't hear many complaints from owners of cameras equipped with that sensor ;)

[Eden]

Cameras based on this sensor are going to create a new market segment, particularly because of the sensor size and the projected pricing.

Pros:

Large sensor size
Relatively low cost compared to other cameras with similar sensor size
Smaller form factor and lighter weight than other cameras with similar specs

Cons:

Noisy sensor
Shallow wells
Low sensitivity beyond the visible part of the electromagnetic spectrum

The disadvantages easily explain the low cost at which vendors are able to deliver this product. However, these can be partially overcome with careful integration and studious processing. Moreover, this sensor will make deep-sky work more affordable for the budget-conscious and beginner market.

This sensor has caught the attention of ONAG owners, since it allows on-axis guiding at almost full aperture (depending on the OTA) and without the need to adjust the X/Y staging to find a guide star. This benefit is in some ways offset by the lower sensitivity at this end of the spectrum (compared to the ICX825 it is > 50% less sensitive).

[Shiraz]

these sensors have almost zero noise Brett - they are way less noisy than anything else out there, including all the super expensive hero cameras.

well depth is totally immaterial if you have a thousand frames to integrate - the effective well depth is then 1000x that of a single sub. 20,000,000 electrons should be enough well depth for the average user.

the quantum efficiency has not yet been specified, but this is a modern chip design from a state of the art digital camera - it has to be pretty good to be competitive in that arena.

you cannot really go past the image in Richard's link (#18) - this is definitely not el-cheapo technology.

[glend]

Ha ha, Ray I was not going to take the bait, but you have my support. I don't know where he is getting his information from (maybe QHY). I don't discount the possibility that vendors have agents on these forums (paranoia yeah). He obviously has not followed the long beta test thread on CN for this camera.
http://www.cloudynights.com/topic/535506-asi1600mmc-beta-test/

BTW Quantum Efficiency of the mono version is stated at 60% and likely a little higher (see the CN thread). The QE of the colour version of this Panasonic chip as used in the OM camera is shown by Senorgen to be 48%.

[gregbradley]

Yeah I was going to say something too. QE on this sensor is likely to be high.
Probably higher than CCDs which are quite dinosaur in their level of sophistication compared to Sony Exmor and other modern sensors that are packed with bright ideas.

Well depth I am not sure about but I thought I read somewhere that it was higher than KAF8300.

Low noise though is what this sensor is mostly about. Low noise, high QE like the Sony ICX CCDs. I expect it will be quite a hot camera for the price.

Olympus have many diehard fans out there who own the OMD series cameras.

Greg.

[Eden]

Hi Ray,



Almost zero noise? Are you sure about that, Ray? Looking at the ZWO-supplied 5 minute dark frame, I would hardly call this sensor "almost zero noise". The QHY12 I am using here might not necessarily be of precisely the same specification, nor is it the hero camera you describe, but it most assuredly produces less noisy 5 minute darks than those shown here:

http://astronomy-imaging-camera.com/Images/ASI1600300s.zip



If this is the case Ray, then why do sensor manufacturers bother with varying well depths at all? More importantly, why do discerning astrophotographers prefer sensors with deeper wells? If what you say is true, everyone may as well abandon all other sensors and immediately switch to the new Panasonic sensor.



There are numerous graphs which have been published which show the QE, both on ASI's site and also over at Point Grey.



I never went as far as calling it "el cheapo", but let's face it Ray, if you look at the numbers this is not intended as a professional/scientific sensor.

[Shiraz]

Thanks Brett.

I measured the RMS variability of the ZWO dark and it is ~12 electrons due mainly to the dark current of about 0.5e/s/p (I had to make of few assumptions here - ZWO has not published a dark current number). That is almost exactly equivalent to a 5 minute dark from your QHY12, which also has an RMS variability of 12 electrons due to read noise. Your darks will not have the pedestal of the ZWO dark, but, when that is subtracted, the total noise will be ~identical for the two cameras. For anything shorter than 5 minutes, the ZWO will be better, but the QHY will always have 12 electrons read noise. At a sub length of 5 seconds (say), the ZWO could have under 2 electrons RMS total noise - this is very close to zero noise and I know of no other affordable camera that even gets close.

ZWO is going to have a problem though, because people are still thinking in CCD terms, where you must have long subs to get over the top of the read noise and that means that you must have deep wells to get around saturation on bright stars. The new CMOS chips will work best with short subs (where the older CCDs are hopeless) and that brings all sorts of advantages including drastically reduced mount requirements, no need for guiding, better resolution and almost unlimited well depth. However, it requires a complete rethink of how one uses a sensor - there will be people who will try to soak the new chips with 20 minute subs and f10 telescopes (like they used to do) and they will not get good results - expect a barrage of "this thing has bad dark current and saturates easily" as people drastically misuse the new cameras.

CCDs still have a place - the old methods still work well. The CMOS cameras also will likely not do quite as well at narrow band imaging, where the higher read noise will be more of a problem. They will also possibly do best with relatively fast scopes (haven't done that analysis yet). However, the new chips offer some very exciting possibilities for anyone willing to experiment with radically new ways of doing things (which will not be appealing to everyone). This really is shaping up to be a revolutionary development, as Emil's extraordinary image shows (see post #18):thumbsup:

[gregbradley]

Good post Ray. You are good at summing things up in a concise way.

Greg.

[glend]

Thanks for that post Ray. Which dark did you use? I understand the new ZWO driver update posted by Sam to beta testers helped to reduce the amp glow in the 300" Dark they posted as an example. I think thing will continue to evolve as more people get the production versions and updated drivers are released.

[Shiraz]

the one in the 1600 section on the ZWO website Glen. I thought that dark looked quite nicely uniform anyway -not a lot of "glow" visible and very few bad pixels.

Yes, it is early days yet I guess and these things will only improve, but it is already exciting.

[glend]

I hope to get mine by the end of this week, and will ring Bintel tomorrow to check on the arrival. The QHY 36mm filter wheel ships tomorrow out of Hong Kong via Fedex, i hope to be testing soon.

[Eden]

Thanks Ray.

You're welcome, Ray. I'm glad you took the time to review and edit your original post because I was quite surprised by some of the sweeping statements that you made. Would I be correct in saying that you no longer hold the view that Well Depth is simply a marketing gimmick?

If so, I'd be interested to know what changed your view overnight.

If not, I encourage you to do a quick search on the Cornell University archives (arXiv) (http://arXiv.com), where you will find an abundance of peer-reviewed papers which make heavy use of this important metric. Many of these originate from NASA/JPL. Given their experience in this field, I remain unconvinced that they would fall victim to mere marketing hogwash.

If I'm not mistaken, Ray -- and please by all means tell me if I am -- you made similar comments in the past with regards to pixel size, without having so much as an iota of data to back up your claims.



Yesterday, you made mention of the ZWO dark and how it somehow compared to an SX ICX649-based camera (which is an excellent instrument) running at -10C. I'm not sure what you were trying to achieve there (given the difference in setpoint). Today, you say you've measured the ZWO dark variability and you're comparing it against... what?

Now I realize that comparing the QHY12 to the 1600MM is not an apples-to-apples comparison (forgive me, I don't have a 16Mpx sensor on hand), but a side-by-side visual inspection of 5 minute darks at -25C in no way suggests that the Panasonic sensor has less noise than the former -- much to the contrary. Look closely at the ZWO-supplied darks -- you'll find what many would consider a serious row defect consisting of almost 10 contiguous pixels. In my relatively limited experience, I've never seen a defect of that magnitude on any sensor, even at room temperature.

If you received a camera from nearly any other vendor with that sort of defect, you'd not only get a heartfelt apology, you'd more than likely be promptly provided with a replacement camera (especially if it were a company like SX, whose sensor you made mention of in your previous post).



So you're suggesting that, in an age where we are apparently moving away from CCD, folks can look forward to taking images of DSO's with the help of hundreds of 5 second exposures, just to circumvent the so-called well-depth limitation/marketing ploy and read noise?

Not this camera -- it simply doesn't have the sensitivity for the pipe-dream that you're describing. The saturation is real, Ray -- which is why well depth matters.

Let's look at the maths. The download time for a 16Mpx image, the storage required (admittedly of lesser concern in 2016) and the eventual processing power and thus the time involved is significant. Going by your 5 second suggestion, that's 120 images per minute and 7200 images per hour? Running that through PixInsight or Maxim? For a typical 3-4 hour session of imaging time (28,000 or so images?), you're looking at the better part of a day to fully integrate and stack all of the resulting data, even on an 8 or 12 thread system. No thanks :) It sounds to me as though you're trying to weasel your way out of the fact that this is a noisy CMOS sensor with highly limited wells.

This works on something like an ICX825 -- which many folks here (including myself) have. It might be a smaller sensor than the one in the 1600MM but it is far and away much more sensitive and for most astronomical targets, the ideal sensor for the imaging technique to which you are inferring. That's why it ships with software that does it.



Sorry to bring this up, Ray, but you in your previous post you asserted that the Panasonic sensor used in the 1600MM is a "dedicated astronomy sensor". This is untrue. I'm not sure where you sourced this information from (Cloudy Nights perhaps?) but few CCD manufacturers have developed a sensor specifically for astronomy usage and Panasonic is certainly not one of them, nor have they made any such claims about the intended market for this sensor.

What happened to your claims that the available QE data was useless/unhelpful/unreliable due to the fact that it is expressed in relative terms as opposed to the absolute? Has your stance changed on this point? Most of us here are using Sony sensors and are used to comparing such sensors using the relative data available on their whitepapers. Would you mind explaining how much difference there would be in having the absolute sensitivity characteristics available, when it's quite a simple matter to compare sensors using the available relative data, pixel size, read noise, well depth and other characteristics?

I'm not attempting to discredit cameras based on this sensor, just to be clear. I have one on backorder, but I intend to use it for guiding.

[glend]

Ah Eden you really like to stir the pot don't you? And nothing better to do in the middle of the night? I don't think i have heard such a long essay of sledges in some time. Disguised sledges are a real art form. And claiming you are going to relegate the ASI1600 to a guide camera role is a real hoot. Congratulations on your master piece. Now since you have no interest in this camera can you move on to lurking somewhere else.

[rmuhlack]

The more I look into it, the more i think that CMOS lucky deep sky imaging is where the future innovation will lie. The ability to capture high resolution images with no guiding and far less demanding mount requirements will be a real game changer. All this is possible (as Ray has already explained) because of ultra low read noise. Furthermore, if the absolute QE is also high then there isn't even a compromise in sensitivity (see Ray's thread on sensitivity here (http://www.iceinspace.com.au/forum/showthread.php?t=136008)for more detail). Data storage requirements are a magnitude larger as mentioned here already but fortunately storage is cheap (~$50 per TB), especially compared with a mount upgrade.

One thing I have been wondering about though is automation. Currently I'm using SGPro for automation and image acquisition with my SBIG ST8 and ST10, but this wont work directly for video capture. Does this mean that we will be (at least in the short term) sacrificing some automation convenience in favour of increased resolution...?

A compromise I have thought of is that one could use a more traditional capture camera on a guide scope (or maybe an ONAG). A sequence could be queued up in SGPro with the "guide" scope, giving the convenience of automated pointing and plate solving, meridian flips etc. In parallel, a capture sequence is setup in Firecapture to collect the higher frame rate video. Ideally we will see new software that will be able to combine these activities (eg plate solving and auto meridian flips in Firecapture...), but in the short term this might suffice...?

[gregbradley]

The proof of all this will be in the images presented no matter what approach was taken. Then the better images will show the best way to achieve them. If that is traditional LRGB 5-15 minute subs 1x1 binned and 2x2 binned colour at 5-15 minute subs or very short stacked.

In my experience even reliable 60 second unguided image is difficult even with high end mounts. Star elongation sets in fast at anything above 500mm focal length.

But autoguided 60 second images are relatively "easy" ( I don't know that anything is "easy" in tracking!).

The Sony ICX CCD cameras required a different approach to the usual for optimal results. Dark subtraction often visibly added noise.

Time will tell meanwhile its fun to speculate on what might be. So no need for heated discussions on mere speculations and predictions.

The images will speak for themselves.

Greg.

[glend]

Richard, i saw a post on CNs saying Sam at ZWO was working with SGP on integration of the ASI1600, whatever that involves.

[Shiraz]

eg, some early examples Greg - apologies if these have already been linked to or if you have already seen.

(Richard's link to Emil's image) 2000x1second!!!! subs http://www.astrokraai.nl/dump/20160505_M51_2000x1s_AutoStakkert_A SI1600MM_Emil_Kraaikamp.jpg

15 minute total narrowband comparison 45x20 seconds or 1x 900 seconds http://www.cloudynights.com/index.php?app=core&module=attach&section=attach&attach_id=690427

~ 1 hour total colour - 54x30 seconds bin 2 for each of RGB http://www.cloudynights.com/index.php?app=core&module=attach&section=attach&attach_id=690357

It has already gone past mere speculations and predictions Greg - and it looks like cMOS is poised to make some fundamental changes to our hobby. Unfortunately that is going to be a pretty scary time for some.

[Shiraz]

for anyone considering this camera, the above analysis is wrong by a factor of 10x. The processing will be onerous, but not this bad.

[multiweb]

Expensive EQ mounts manufacturers will probably start to get worried. Big dobs might take over as imaging platforms.

[gregbradley]

The M101 image is particularly impressive. 27 minutes total each for rgb = 81 minutes. How is it so deep with such little total time? I take it M101 is not super bright and equivalent to say Omega Cent or even less bright. If you just showed me that image I would assume it was 12 hours total expoure or more.

The 900 second Ha of the Crescent seems a bit nicer than the shorter stacked image but that is also impressive. The 2000 image M51 is OK but the core is totally overexposed and blown. Perhaps merely a processing error.

I wonder also about how to process 2000 x 1 sec x 32mb files. They are the same size as my Proline 16803 and CCDstack starts to chuck a wobbly if I open more than about 12 of those! Sam mentioned some free software perhaps it handles large files better than CCDstack which is rather poor at that.

A great start.

I agree, this type of imaging will push for large aperture over tracking ability and high quality mounts. So a large dob with some basic tracker that can do 20 seconds would be good enough. Like Alex did with his 25 inch dob a few years back and a modded Sony NEX 5n.

Greg.

[rmuhlack]

The M51 image with 2000x1s subs was processed with Autostakkert. Emil (who captured the M51 image) is the author of the Autostakkert software.

Note that in the Cloudy Nights thread he says that he captured that image with maximum gain, and because the saturation is so low at that gain level he captured using 8bit images only ("as that is enough to represent all the signal") see here (http://www.cloudynights.com/topic/536494-m51-in-poor-seeing-20001s-asi1600mm-cool/?p=7208540)

[Shiraz]

I would hate to be in their position. I guess that there will always be a market for strong accurate mounts, but I really hope they have some strategy for dealing with what seems to be coming in the entry/intermediate market. Same goes for the makers of CCD cameras I guess.

i guess that this is what is called a disruptive technology.

[gregbradley]

Thanks Richard.

So he went 8bit to speed up the processing of the files at the expense of dynamic range. He can't say 8 bit was enough when the core is way overblown - it wasn't enough. It seems strange though as 1 second exposures surely can't blow out the core. But I suppose if he is shooting at the equivalent of ISO6400 or something you could. 8 bit would be sacrificing some dynamic range surely or the if not then the gain was too high.

Its an impressive image for 1 second exposures and 2000 of them (it must be some sort of record!). But basically its still a fail as far as a high quality astro image goes.

Greg.

[glend]

Getting ready, I have just installed Autostakkert (the new beta version 2.6.6) and had a look at it. A new world for sure but it seems logically laid out & includes ZWO Firecapture.

Prepping the laptop for the new camera, and I will be cleaning out alot of old image files that have piled up on the hard drive. Mostly individual RAW images off the old cameras, I will keep many of the final images and archive them off to my external USB3 drive.

Waiting to hear from Bintel.

[glend]

Just spoke to John at Bintel re shipping and was advise that ZWO has held up the shipment and it will not leave China until Thursday (19th) of this week. The reason for the delay is to resolve something with the cameras, don't know what. I do know that some Beta Testers had some AMP Glow concerns on long Darks and Sam (ZWO) had a fix for that, and one tester reported a problem with dessicant pills coming loose in transit, but no other concerns I am aware of right now. So given transit times, customs, etc I would not expect Bintel to get them until the end of next week (say 26th or so). So for those of us waiting, it looks like delivery in the first week of June.

[gregbradley]

Its better ZWO holds up the shipment to enhance the camera rather than rush it out and people have trouble updating the firmware at home and possibly bricking the camera.

Greg.

[Slawomir]

Although it all sounds really promising and there is definitely a sense of fresh breeze in terms of new technologies being made available to astroimagers, but at the same time I remain sceptical, possibly due to my ignorance...

It seems to me that the laws of physics cannot be broken - there are only so many photons hitting a sensor from a DSO in a given time. So if there is one photon per say 2 seconds from a DSO per unit of area/pixel, chances are we will miss a significant percentage of them with 1s exposures. Just speculating here though...and to clarify, with ideally zero read noise, would longer exposures help to extract DSO's signal from noise coming from the skyglow, or is that corrected with number of subs alone:question:

Either way, those brilliantly fast and sensitive sensors certainly open doors to new ways of imaging.

[Atmos]

Suavi, you have just hit the nail on the head as to why professionals are not going to go down the path of short subs (depending on situation obviously), faint stuff! Difficult to detect at the best of times, short subs can at times leave you with more noise detection than real stuff leading to faint bits being rejected.

[Camelopardalis]

Who said the 1600 refuses to capture photons if set to low gain for 10 minutes? :shrug:

[gregbradley]

Exactly. The sample narrowband images of a bunch of short Ha images versus 1 15 minute image showed the 15 minute image as being nicer overall to my eye.

The camera can do both which is the edge over CCDs.

This sensor is really somewhat old in the sense that the current cutting edge Sony sensors have copper wiring (instead of the usual aluminium and what this means is 4X faster read time and probably less noise, plus a thinner stack meaning it can take light at greater acute angles) and backside illuminated with the circuitry on the back gaining an extra 40% imaging area for the pixels.

Backside illuminated sensors are what most new Sony sensors are using.
This is CCDs increases QE up to the 90-95% band. I am not sure what it does for Sony's. I don't think it is that much though. My Sony A7r2 has all this technology in its 42.4mp sensor and its a pretty amazing sensor. Not ideal in its current state for astro though as it has some colour noise in the shadows at high ISO long exposure (long exposure noise reduction gets rid of it but still).

In addition there are lots of subtle tricks like 3 stage ADC where the as the ISO rises at a certain point this kicks in and drops the noise further.

Some Sony sensors have a built in memory buffer chip on the back of the sensor to improve how many images they can hold.

Sony is also working on some exciting new types of CMOS sensors. One is able to control individual exposures per pixel, they have a curved full frame sensor already. They are supposed to be working on a similar sensor to the Foveon sensor with RGB pixels at different depths on the same pixel site.

Global shutters are not far away either.

Greg.

[Slawomir]

I thought the idea was to capture lots of really short subs, otherwise I cannot see where is the revolution...:shrug:...maybe except for a lower price for an astro-camera! :)

[glend]

It can operate in either mode. And yes it is a price leader for what your getting. It is pretty clear that it is an upgrade path for many DSLR users, that cannot make the leap (financially) to the higher end CCD cameras. Let's wait and see what the new owners produce in either mode before making value judgements.

[Shiraz]

the full frame download time should be around 40ms (it runs at up to 23 fps full res), so the chance of photons being missed while the chip is doing something else is quite small, even at 1 second subs.

The big thing though is the low read noise, which allows you to get to sky-limited performance with short subs. One second is a bit extreme (but possible) - maybe 30 seconds would be more mainstream and still provide some significant advantages. You don't have to use short subs, but why not if there is no SNR penalty and you get improved dynamic range and all the mount advantages?

Note that short exposures are not unprecedented in the professional arena. For example, PanSTARRS has a low read noise detector (for larger pixels at 5 electrons) and is sky limited after 15 seconds - typical exposures are 30 seconds. Once you get to sky-limited, longer subs do not help you extract signal in the final image - longer subs will look better individually, but the final stack from a lot of short subs will be the same as that from a few long subs if they are all sky-limited. Longer total exposure improves things, but not longer subs.

[Slawomir]

Thank you for explaining that Ray.

I thought that 1-second subs would be on the extreme end of the spectrum and even low read noise would most likely drown weaker signals from a DSO.
30 or so seconds sounds much more reasonable for faint fuzzies. So accurate guiding is still required meaning expensive mount for top quality data. I was quietly hoping I was wrong and could spend "dream mount" money on a boat... :(

[Camelopardalis]

Cooled mono sensor ~22mm diameter for under $2k? Revolution, pah :lol:

[RickS]

Of course, it takes significantly longer to get sky limited with narrowband, so it might still be a while before you have your magic bullet, Suavi.

I'm sure that low read noise, high QE sensors are the future and this might be another step on the road there... It will be good to see results from production hardware. I'd also like to see a 16803 sized sensor rather than these tiny toys :lol:

Cheers,
Rick.

[Slawomir]

As long as desiccant tablets won't come loose! :lol:

But more seriously, maybe with significantly lower cost one beautiful day we will all be able to enjoy using cameras with sensors as big as 16803 :prey:

[RobF]

Now that would be a manufacturing revolution I'd be happy to partake in...... :)

[gregbradley]

QHY has a few already in planning. Look at the QHY42. That is similarly sized to the KAF16803 but with larger pixels which would suit long focal length (9 microns is a bit small for 3 metres).

Greg.

[Camelopardalis]

:lol: yeah we'll have to keep dreaming for a while longer I reckon :rofl:

[glend]

Jon Rista over on CN has been doing some work on master darks, bias frames, and gain settings for this camera , which it seems he just received yesterday. It is worth checking out the files he has posted, on page 23.

[Slawomir]

Thanks Glen for the heads up.

Bias and Dark frames look good, except for a substantial amp glow even in a 120s dark. So darks at this stage would be essential in calibrating subs; not a big deal really. I must admit I am spoiled with my camera, as I only need to use a SuperBias and Flats :D

[Shiraz]

thanks Glen - gosh that looks like a very tidy DSO camera :thumbsup:.

edit: no it isn't, it's brilliant.

[codemonkey]

+1 ... however, with the lower read noise, smaller pixels (for me, if not you) and bigger FOV it's still very tempting. I reckon I can get more detail yet with a bit more resolution, and it's sometimes nicer (mostly for nebulae) to have a wider FOV... my camera's 674 chip is very small in terms of FOV.

For me to go forward with that acquisition I'd have to be certain that amp glow would calibrate out, and I'd very much like to know the absolute QE.

[Slawomir]

I am tempted myself, I am just being very picky :lol:

It certainly will be a very popular camera. Based on the results from CN, read noise at low gain is only 3.35e, full well circa 20,000e and dynamic range about 6000 steps - outstanding results IMO!

I will wait though for the 2nd generation of this camera...

EDIT: Not trying to detour the thread, but I was curious how my camera compares "on paper", so I have just run the same PI script and the results are attached. ICX 814 has half the well depth and significantly lower dynamic range of about 4300 steps, but read noise in ICX814 is still lower - only 2.4e according to this script. Nonetheless, the specs of the new CMOS sensor look very promising, especially at this price.

[RickS]

Even if it does it will leave noise behind. I'd let someone else be the early adopter. May be a great camera but it will probably get better after the initial teething problems are resolved...

Cheers,
Rick.

[glend]

Some bad news for those of us waiting here in Australia for the ASI1600 camera. I just had an email from Sam at ZWO in reply to my question to him in the ASI1600 Q&As on their website, concerning emerging supply chain issues and information provided by Bintel to me regarding the pre-order shipment. Here is Sam's email in total:

"We arranged the Bintel's shippment on Tuesday, we planned to send it on Thursday, but we didn't send all of the new cameras they ordered, because we don't have enough stock, we are lack of PCB board ,whcih needs long time to produce,and that's also why shippment delayed. It's just supply problem.
We olny sent a few 1600 cameras to our customers who ordered directly from us.
More new cameras will be available at the end of this month or the start of next month."

So it appears that Bintel will be not be getting all their expected pre-ordered ASI1600s in this next pending shipment. And that direct orders are also having to wait. Lack of PCBs is a fairly serious impediment to supply at this launch point, and it's obvious that ZWO may have under estimated the demand they would have to fill.

Hopefully Bintel communicates with the pre-order list and lets them know who is getting cameras this month. I have sent Sam's email on to Bintel.
I expect that pre-order pricing must be honoured by all concerned until all have been supplied.
Obviously I am disappointed at having to wait, assuming I won't get one in this next shipment, but it provides more time for the CN beta testers to tease us.

[h0ughy]

ok to be very clear with this from past experience BINTEL will only honour the special price if you have paid in full - as i said just my previous experience

[glend]

I just received a shipping advice from Bintel, and my credit card has been charged the exact amount of the ASI1600MM-Cool, looks like it is on its way to my door. And the pre-order price is what I was charged.

I know the ZWO shipment arrived on Thursday so they must be getting them out asap, and there seem to be enough in the shipment that I can get mine.
In other good news, the Fedex tracking on the QHY filter wheel shows it has arrived in Australia. I already have any T-thread spacers I might need.
Looks like I might be ready to put this camera to a test by middle of next week! Exciting. Just in time for the June new moon period.:thumbsup:

In other news re this camera, there is a new thread on CN looking at Sub exposure durations for the ASI1600:

http://www.cloudynights.com/topic/537622-sub-exposure-duration-recommendations-for-asi1600-camera/

I need to read that.

[Shiraz]

for interest, there has been a lot of consternation from people visually inspecting the darks from some of the cooled ZWO CMOS cameras and deciding that the relatively high levels of signal = high dark current. An explanation of the mechanism that introduces a pedestal (not dark current and no shot noise) in these cameras has been posted on the ZWO user group forum. http://zwoug.org/viewtopic.php?f=21&t=6204&start=30
" it turns out the effect you have finally identified looks like what is called clock feedthrough: when you retrieve your data from each pixel read transistor swithes and a parasitic charge is transferred from gate to drain which offset the ouput voltage, and it has no shot noise and is about the same over the chip. the way to make sure it is true is to measure the evolution of the shot noise with exposure time. You can infer the dark current from the shot noise as well to know the real dark current. Pierre"

Seems that the pedestal level has some dependence on the integration time, but it appears to be repeatable, so can calibrate out. Not sure how applicable this is to the 1600, but worth keeping in mind, since dark scaling may be a problem.

[glend]

Strange. Bintel is now showing these cameras as in stock on their website. Must have received more than the pre-order list.

[glend]

My ASI1600MM-Cool arrived today an hour after the QHY CFW2-S filter wheel. Now to put them together and start doing some testing. First I have to reconfigure the filter wheel for USB operation, and sort out the thread connections.
Photos attached.

[Slawomir]

Congratulations Glen!

Thank you for the photos, and we look forward to reading your impressions with your new shiny camera+FW :thumbsup:

[Atmos]

If you need some information on doing this (configuring to USB) just let me know and I'll forward you a document :)

[glend]

Thanks guys. Colin they (Cyclops Optics in Hong Kong) have already configured it for usb for me. When i was chatting to their helpline about the order they asked me about the camera i would be using and how i would control. When i opened the plastic housing this afternoon to change the jumper for usb, i found it was already setup by them. Pretty good sevice i think. I have a copy of the manual as well. Those little spacers and screws are confusing but it turned out all i needed was the 2mm screws and its spacer for my Baader filters.

I have the camera and filter wheel joined up now. The distance from the rear of the T-adaptor to the sensor is under 55mm, so my scope focuser setup for the old DSLR will still work perfectly. I don't particularly like the long screws that they supplied with the filter wheel for the clamp but they will do for now. At least with that setup I can separate them easily. I might power it up tonight, or at least see if I can get the wheel to operate.

[Atmos]

I have to admit, I am not a huge fan of those screw things either but at this stage, that is how I have my QHY22 connected to it. I did really like that I could connect the QHY9 directly to the CFW, they would be bolted together to be flexure free, cannot say the same about the QHY22 sadly :/

Looking forward to seeing how that camera performs!

[Camelopardalis]

Woohoo! So is it clear tonight there Glen :D

[glend]

Well yes Dunk it is a wonderful clear night with the Full Moon just rising! I don't think I am ready for imaging just yet, I have not even powered it up yet. I need to work out the drivers, and the wheel operation. Probably do some darks first so I can get a handle on controlling it.
BTW, I have torn the wheel apart to remove the universal camera adaptor QHY put on the camera side, and now have the ASI screwed directly to the wheel (yes there is just enough M42 thread to do that). I have a spacer on the scope side to prevent the T-Adaptor from getting near the filter lock screws, its tight but seems ok. Filters are loaded, test soon.
New config photos below. The new distance from the back of the T-Adaptor t the sensor is under 30mm. I have plenty of backfocus range on all my scopes so its not a problem, and if it was I could stick a t-thread spacer behind the scope adaptor. I prefer the filter to be as close to the sensor as possible, but really vignetting is not a problem with 36mm filters in front of this sensor. The guys on CN running small filters will have to watch it.
New Config below.

[gregbradley]

Congrats Glen. It looks like a beautiful camera and filter wheel combo.

Looking forward to your first light with it.

Greg.

[glend]

Thanks Greg. I have installed all the drivers (I think). Powered the camera up, using my trusty 12V 6AH AC adaptor that I have been using to run my DSLR cooling system. The ASI only needs 2amps but it should not hurt to have more available, it will use what it can handle. I think I had seen a post on CN by Jon Rista claiming that he thought it might need more amps (than 2) in a warm climate to hold a delta T of -40C.

Started up Sharpcap and it raced away stacking AVI frames before I knew what was going on, how do you stop this thing. I finally slowed it down to one every 30 seconds. Turned the cooling controls on and can report that from an ambient room of +21C, it managed to get down to -22C actual sensor temp reported, so a Delta T of -41C which I think is pretty good. It did take awhile to get the last five degrees but that's ok.

Then I tried to get the filter wheel to work, but that's a problem. ASCOM shows it, and I seem to be able to configure it. I have the USB driver, and the wheel comes to life and does what looks like a start up spin cycle. Sharpcap doesn't seem to control wheels, so I tried APT which I have as well, and it says it connects but doesn't do anything.
Anyone have a simple wheel control app that they can recommend - one I don't have to spend money on?

[Atmos]

When you connect to the ASCOM Driver you need to set it up, connect to the filter wheel via a COM link, choose the CFW-2 and then set up the filter positions.

[glend]

Yes Colin, did all that and then nothing.

[Atmos]

Hmmm... The moment the filter wheel gets power it does its spin thing to get to filter position 1. I have two different ASCOM Drivers that work, one is the USB and the other is a RS-232 Serial that also works. Both appear stable, you could give the other one a shot.

[Camelopardalis]

Glen, I don't a usb filter wheel (yet) but I think in APT you can set the filter type for it to shoot when create a plan. You also need to assign the filter labels to the positions in APT Settings.

[glend]

I have both those drivers Colin. Dunk APT never gets that far, it fails to connect to the wheel. I might try deleting the QHY drivers and reinstalling. I would be happy with simple Audrino control, i saw an example online and the code is free.

[glend]

I have everything working now. I decided to go with the recommended control app (from CN) and have installed SGP. Lo and behold it had no problems at all with the filter wheel, which is all setup and cycles through all the filters. I like that fact that the colours displayed by the wheel when I select a NB filter match what I would normally assign to those narrowband filters. So I have been running desk top darks and playing around with temperature control in SGP. I prefer the Temp Control in Sharpcap but can't have everything I suppose, SGP temp control works but a bit confusing compared to the sliders in Sharpcap. But it does have cool down and warm up periods. SGP does it all but in a different way to what I am used to, but I will pick it all up.
Guess I will have to pay for SGP in 45 days but it is not that costly at $99 (USD).
It's ashame that there is not a single app that provides the necessary planetary and DSO capture processing. So Sharpcap for planetary and SGP for DSOs. Just about ready to put this gear on a scope.

[Shiraz]

mine is on the way:) - Bintel had one.

been hoping a really low read noise camera would become available before I expired - and it has woohoo!! - just made it. Now to see if the theory works.

[glend]

Congratulations Ray. :thumbsup:

[Atmos]

With the frame rate that is camera can achieve at full res, it could have a fantastic solar imager too!

[glend]

Had my first try at imaging with the ASI1600 tonight, and had some issues. Nothing really with camera itself but the support infrastructure and software. At first i thought i might have a dead camera but it turned out you can't run the camera through a usb3 powered hub! I was tearing my hair out trying all the different apps and even pointed it at the moon, but nothing - until i plugged the camera usb3 cable directly into my laptop. I recall that Jon R on CN reported issues with USB3 cables. I have a nice Mbeat combined usb2 and usb3 powered hub, which has been serving me well in the dslr world for years. It will not pass through the camera traffic however. Since i only have one usb3 hub on my laptop, if i plug the camera in directly then i can't plug in the guide camera and the filter wheel - both usb2 but they were working off the mbeat usb3 hub, which the camera doesn't like. On top of this, i can't get my ZWO ASI130mm guide camera to work at the same time as the ASI1600, in SGPro, which ever one starts first takes the ASI driver as its own. I am going to have to ask Sam at ZWO about that - unacceptable. Focusing and framing! Why does that have to be so hard? Love the simplicity of dslr focusing in Liveview, or BYEOS. SGPro runs a series of short subs to help focus but the lag is annoying, Sharpcap is better because you can do it in video mode. So more work to do in the next few days. Anyone have any ideas about usb3 management?
Ray better think about how your going to handle the usb3 issue.:shrug:

[Shiraz]

thanks for the tips Glen. Crikey, it's unheard of for anyone to have USB problems :P

I have a QHY guide cam and a pc with 3x USB3 ports, so I guess I should be able to connect directly from the camera to the PC for initial testing. Hopefully we can find a hub that will play with the ZWO - off to Officeworks today to see what they have. Of course, it is also possible that we will be able to ditch the guiders altogether, which would solve your ASI driver clash.

Am planning to do a minimalist changeover by unscrewing the H694, screwing on the 1600 and accepting the vignetting for now - then changing to a larger filter wheel at a later date (want to keep the 694 though for really deep imaging until the full capability of the new camera is understood). I expect that a lot of galaxy targets of interest will be available in sub image regions on the 1600 (if that works as hoped) - using less than the full 16mp will be a sensible way to keep the data load manageable with 10 sec subs.

The CN threads have been interesting, but a lot of the folks are still looking at using the thing at unity (or lower) gain with long subs - seems to me that they will just end up with a 16mp 8300 . I am itching to get some high quality long exposure data at high gain and with very short subs - that's where the totally new capability of the camera lies. And of course, a full image of the moon at super high res will be an interesting thing to try - and the planets at 100+ fps with sub-region selection.....

[glend]

The problems seem to be solved, or at least everything is working at the same time now and I did not have to buy any more equipment.

#1, The problem with the USB connectivity. I have the camera plugged directly into my laptop USB3 connection, and it works that way. The guide camera and filter wheel are running off my mbeat 7 port powered hub, and they are USB2. As the hub USB3 cable is backward compatible to USB2, I have it plugged into the spare USB2 connection on the laptop. The filter wheel and guide camera are working through that. I had to reboot the laptop to reflect the reorganisation but its looking good.

#2, The problem with not being able to run the ASI1600 and the ASI130mm (guide camera) at the same time has been resolved. It is important to start the ASI1600 first so it can be picked in SGPro with the driver and connected there. Then the guide camera can be started and it works and does not impact on the ASI1600. I have written to Sam at ZWO about that as I think people using ZWO guide cams should know about that.

I still need to do a USB3 cable extension test to see if I can get away with the USB3 extension that I was using for the mbeat hub, if that works I can probably just use a USB2 extension off the mbeat hub to the laptop, and can return my workstation to its regular spot.
More to come.

[Camelopardalis]

Glen, interesting about the camera detection, glad to hear you trail blazers are sorting it all out before I can afford one :lol:

I'd connect up the camera requiring the highest throughout to the dedicated link. Alternatively, try different ports on the hub. I've found that my Canon will only connect on certain ports, whereas other devices seem less fussy. I'm sure an electrical engineer would be able to decipher the USB hierarchy and tell us why, but the rest of us just have to go with what works.

Focusing on a dedicated cam is definitely different :D on my CCD I just use binning and short exposures and that works well enough. Sharpcap would seem to be the way to go with the 1600 though.

[RickS]

I'm an adjunct professor of Information Technology and Electrical Engineering and I work with a bunch of smart EEs designing products with USB interfaces and AFAIK nobody can make USB work 100% reliably, especially under Windows :lol:

[gregbradley]

That's right USB problems are unheard of and you should write 100 times on a blackboard I will not overload my crappy USB.

Glad you sorted that. I also find I have a startup procedure that minimises problems with corrections. I always turn my computer on and let it fully boot up before I turn on my astro gear. The other way round sometimes stops gear from being recognised for reasons best known to Microsoft engineers.

Very exciting Glen. Good on you for overcoming what is likely to be the only real obstacle - getting the thing connected. That is often the first and most difficult step with new gear in my experience.

What would be good when you get the chance is a traditional image with 10 min 1x1 subs and unity gain and high gain and shorter subs of the same object. Something like the Trifid would be good as we are all familiar with what that should look like

Greg.

[Slawomir]

Great news with sorting everything out :thumbsup:

[glend]

Thank you guys, not everything sorted out but making progress. I did the USB3 cable extension test today and it just won't work, even though its only a 2m extension cable. I expect i will need an 'active cable', one of the powered ones. I am reorganising the obs so i can get the laptop closer. I heard back from Sam at ZWO re the driver issue i raised, i don't think he understood my question because he just asked if "i visited ASCOM".
Greg the test shot will have to wait, the cloud is back tonight and it going to rain tomorrow. However, i will be ready tomorrow night if it clears by then. With the moon out of the way all the Sag nebulas will be available.

[gregbradley]

I got a powered USB extension from Jaycar and it worked really well whereas non powered USB extension cables failed.

All part of the fun!

Greg.

[luka]

1. Are your powered/unpowered extension USB2 or USB3? Mixing USB2 and USB3 cables may cause issues.

2. Is the camera powered via USB or is there a separate power supply?

[Shiraz]

Another image recently posted - Mars, 8 inch Celestron, ASI1600
http://www.astrobin.com/250254/

[Atmos]

That's pretty awesome!

[ZeroID]

Wow !
What was the scope setup ? F#, Barlow, ROI etc ?

[glend]

Well the rain has cleared away and there is a nice dry wind to dry things out and the sky is clear blue, looks like first images tonight. The waning moon rises around 9pm, and sadly is still close enough to the Lagoon and Trifud to throw some light in the front of the scope even with a light shield, i would prefer to stay 45 degrees away. For this first light i probably would prefer something further away, even though i will be shooting Ha. I am thinking that maybe NGC 6188 would be a good choice, it is a fairly bright (+5.2) emission nebula with good gas cloud definition, and has a bright central star that can give some insight into clipping behaviour, and is high enough by 9pm to be well away from moonlight.
Other suggestions? I prefer easterly targets.

[Shiraz]

don't know any more than what is given in the equipment list on astrobin - an 8 inch celestron edge with a 2.5 powermate and the 1600 camera. It is moderately impressive though.

[gregbradley]

NGC6188 or 6164 (the little flower nebula next to 6188) would be the go.

Vela Supernova Remnant is another IC2944 the Running Chicken is another.

Greg.

[Slawomir]

As Greg suggested, I think NGC 6164-5 would be a great target for testing the camera. It has a relatively strong signal in Ha and there are some faint bits too.

[glend]

It looks doubtful for tonight despite clear skies earlier there is high cloud overhead now. Just my luck. Will check later but it doesn't look promising.:shrug:

[ZeroID]

C8 Edge is f10 native and a 2.5 barlow so f25. Explains the image size and detail. Thanks.

[glend]

Well i managed to get some images tonight, nothing that is ready to share but making progress. No issues with the camera but the support systems are driving me crazy. SG Pro is a learning curve for me, and coming from the dslr/BYEOS world it is very different. I have got focusing working fine now, but initally just worked with luminance through an open filter wheel spot as that easier than doing it through a narrowband filter. Did some test subs on the Lagoon Nebula, with no filters, just to see if the gain setting and sub lengths were right. Shot several different exposure lengths, 30, 60, 90, 120, 180 secs in luminance, to check the histogram and clipping. Shot some more in Ha to compare the histogram, and moving it right off the bias side does take some time. I probably have the camera setting wrong. I was initially using unity gain setting, and tried Jon Ristas suggestion of optimal gain at 60. The images are displayed on the screen after down loading but then when i tried to go into the file to have a look at the images there was nothing there. I have no idea why it is not saving images in the designated Light file. Update, just found out i have to right click on the temporary image to save it, hope there is away to save a sequence of images without having to save each one like that. Any suggestions on that? Is it a limitation of the trial version? It is supposed to be a full working version.
SGPro froze on me once for some reason, but that has been reported on CN as well, some sort of memory issues that ZWO is discussing with the SGPro people.
I may try APT the next session.
Hope someone else gets one of these cameras soon.

[Shiraz]

It will be a lot different to a DSLR, but when you get it going, SGpro does it all very nicely. It has the ability to focus through a broadband filter and then switch back to narrowband, which makes it easy with parfocal filters. If it is working, I guess that you must be able to see stars in the images, so the camera is working. You will be on a huge learning curve if you are beginning to use a new camera, filter based imaging and SGpro all at the same time!! SGpro by itself has a daunting collection of options, many of which only make sense after you have been using it a while, so some simpler software that only does image capture would be a good idea while you check out the camera. FWIW, I intend to use Nebulosity while getting the camera sorted because I know it well and it is the most stable software I have. I also occasionally use FITSliberator for checking FITS files if anything has gone wrong - it is a nice tool for viewing files.

Are you quite sure that there is nothing saved in the files?. the sky data will be pretty close to zero (which I understand defaults to 21 for this camera/software) for short subs, especially if you are using a low gain setting and on a linear display you will see a totally black background with a few stars and hot pixels. You will need to use a huge amount of screen stretch (eg with the sliders in SGpro) to see anything dim. You don't normally have to right click on an image to get it to save properly - when running the sequence it will flash a status display down the bottom of the screen that says "downloading" and then "saving file" (or something like that).

It will all work:thumbsup:, but it is worrying that SGpro froze - it generally seems to be stable and reliable. Do you have the latest ASCOM driver for the camera?

[gregbradley]

I agree with Ray. Learning 3 different things at once is not ideal in my experience. I try to limit the change I am working on to one at a time. Otherwise it can be frustrating and take the fun out of it.

Greg.

[glend]

Thanks for that response Ray. I am sure there is nothing in the SGP files. I went into the subfolders that it created for Darks and Lights, and it is creating subfolders in those by date but there is noting inside the sub folders. I had shot a bunch of darks and only looked at them initially while they were on the sequence capture screen and it never occurred to me that they were not being saved, until I checked the files last night. I can see the sequencer working and yes I can see the download taking place on the bar at the bottom, but it does not save after that.
Re your other question, yes it certainly is working and I can clearly see my target object in the sequence temp display, even down to the gas cloud detail in the Lagoon both in Luminance and Ha. It was when I tired to open one in Photoshop that I discovered nothing in the destination file.
I have downloaded the SGPro Manual pdf (all 200 pages of it) and will be going through that today.
Yes I have the correct (and current) driver for the camera. Re the freeze, maybe that was the incorrect term, it goes into a "Not Responding" mode, and I suspect it maybe due to filter wheel start up routine. It only requried a restart once. I will try to raise a question on the ZWO forum about it but Sam has previously said SGPro had a memory management issue, so perhaps he will just refer me to them. Time will tell. Now that I am getting familiar with SGPro, I don't really want to try another one, if I can just solve the 'save' issue. I do miss the ability to review images in great detail on a Liveview screen.:thanx:

[Atmos]

Might seem a bit silly (its happened to me so I am talking from experience :lol:) but are you sure that it hasn't saved the images somewhere else? I'd suggest searching for "Frame 1" and seeing if anything pops up somewhere.

And with the Filter Wheel, I depending on which driver you are using it does hold up the software (like it has frozen) until the filter wheel has finished doing it twice rotation thing. The normal driver does it, just means that it resets after every connection. The RS232 driver doesn't, just connects and assumes that it is on filter position 1 even when it isn't. After the first move to another position it finds itself but it doesn't pause SGP upon connection.

[Shiraz]

that's a good point - maybe try running it with a camera simulator and see where the files end up.

[glend]

It's my fault, I did not realise that to open an image you have to go to the little folder icon on the tool bar. When I do that I can see my Darks and Light captured images all there. I've just looked through them and I might put up a few jpgs here later to illustrate the various exposures and what they picked up (in luminance and Ha).

I have also just upgrade my ASCOM to Platform 6.2 and the latest release of the ASI1600 driver V1.0.2.5, I had been using V1.0.2.4

I feel like I am making some progress. I will do some daytime testing just to make sure the new ASCOM, drivers, and devices are playing nice, but I am encouraged. Images soon.

[glend]

Of course SGP only produces FITs files, so I have had to go through Fits Liberator before going to Photoshop to get this image. Not sure I want to do that all the time.

So this first image is pretty rough, but as a first attempt with this new equipment and software I am ok with it. It is a single sub of 180" shot at Unity Gain setting (I think), with sensor at -25C. Histogram is way over to the left. No darks, bias, or flats applied, this is just a single sub. No processing in Photoshop other than conversion to jpg for upload to Astrobin:

Detail page here:
http://www.astrobin.com/full/250497/None/

Full size here:
http://www.astrobin.com/full/250497/None/

[Atmos]

This has just had a histogram stretch of the above data.

[Shiraz]

woohoo

hmm, that would qualify as pretty convincing - 3 minutes eh?

but where is all the CMOS noise!!!! What trick have you used to hide it???:lol:

[codemonkey]

Nice first light Glen! By the way, it's normal for the histogram to be over to the left with these sorts of cameras. If it's not, you've overexposed it.

How do you feel about PixInsight?

[glend]

Here is the 60 second test sub (single frame) no stacking, but this one I have stretched. The histogram looks much better.
This is just 60" folks:

http://www.astrobin.com/250499/

Full size here:

http://www.astrobin.com/full/250499/None/

[Shiraz]

:thumbsup::thumbsup: no sign of any fixed pattern noise, "amp glow" or overload streaks on the bright stars - cool.

[glend]

And finally here is the 30" single frame that has been stretched.

http://www.astrobin.com/full/250500/None/

I am thinking the short 30" subs are enough given the way they stretch, and the clipping is not there (see the Astrobin histogram on the detail page here:
http://www.astrobin.com/250500/
).

So a stack of these 30" subs will be my next effort I think. PS: remember the moon was up when these were shot.

[codemonkey]

So... anyone wanna buy a SX-674? ;)

[Atmos]

For 30" all of the brighter stars look very bloated, I do wonder if this has anything to do with the settings that you mentioned you were a bit unsure with.

To get them correct I found following a QHY How To quite useful.
You can do it inside which is handy, you basically start with everything at 0 (Gain and Offset which I assume are the two you can change).

Point the camera at a wall that is reasonably evenly illuminated. Start with fast exposures (0.001) and double the exposure time each frame looking at the average ADU, what you are looking for is when the average ADU stops getting higher (looking for saturation).

Then using that exposure (where it just reaches saturation) you increase the gain until that levels out at 65535 ADU.
To deal with the offset, start taking bias frames and slowly increase the offset until the bias level is 500-1000 ADU.

[Shiraz]

Colin, I think that the 1600 has split amplification (and an on-chip DSP keeping watch?) - it does not behave much like a raw CCD.

ZWO/Panasonic? have already done the calibration (I guess controlled by on-chip registers) and you can dial up any allowable mixture of read noise, dynamic range etc. as per the graphs. And you can control the exposure - should be an interesting optimisation problem. The offset defaults to 21 ADU? in most software.

[Atmos]

I assumed that was Glen was referring to may have been similar to what I needed to do with my QHY9&22 which is set the gain and offset as they don't.

Looks different :)

[glend]

The camera has three standard pick list gain & offset settings to choose from: high dynamic range, unity, and lowest noise. I can't be sure but i think i was using Unity, which happens to have a gain of 139 as shown on the graph below. In addition you can slide the bars to choose your own settings. Jon Rista (CN) seems to think he is going to do everything at a gain setting of 60.

I have to say all the ADU comments below go right over my head.
I beieve, in time, as more cameras get into imagers hands, the settings issues will resolve themselves through user feedback. I am not seeking perfection right now, just to produce some reasonable images with this big change in approach is enough for me for now.

As it is clearing up here i am starting to consider narrowband for tonight. The beauty of this camera is that you can probably do enough data acquistion in a few hours to build a nice narrowband image. What used to take me three nights with my mono dslr is only going to take a couple of hours now, imho.

Looking at the exposure charts on CN it looks like anything above 20" will give good data. That might be true for LRGB but i expect narrowband to take much longer. However, i did do one Ha image last night at 60", it was out of focus because i forgot to refocus, but showed me enough to think maybe that could be a reasonable sub length to start at, something impossible with most cameras.

[Shiraz]

gain 60 looks to me like a good choice to start with - full 12 bit dynamic range and still a very low read noise. It is going to be exciting to push things into the gain300 region though...

[Slawomir]

I will! Is $20 enough? ;)

[codemonkey]

How about $25? ;)

[luka]

I have been following this (and pretty much all the other threads on IIS) and all this narrow band imaging is way over my head. Glen, why will few hours with the new camera work better that few nights with the mono DSLR? Is it just increased sensitivity and lower noise being that much better? Apologies if it is a stupid question but otherwise I will never learn.

[Shiraz]

While waiting patiently for Australia post to move my camera from NSW to SA, have been analysing some dark image data presented on CN to try to understand the "amp glo" that some users are reporting. I had to make a few assumptions about how the camera and software works, but if I got anywhere near right, the glow is just slight variation in dark current across the chip (due either to peltier temperature variations or to the chip itself).

my best guess is that, at the tested -15C, the dark current varies (over the specific chip where data was available) from about 0.013 e/s/pix in the centre to 0.021 in the brightest of the "glow" on one edge. This level of dark current is not as good as the Sony CCDs, but slightly better than the Kodak ones - ie neither the dark current nor noise from the "glow' should be a major issue. However, there may possibly be some difficulty with scaling darks, if the start point varies with position on the chip. Real darks taken at the true temperature may possibly be best with this camera. Of course, the dark current is still very low and it may be possible to get good results by ignoring dark calibration altogether and relying on dithering. That could be very useful if the datasets get big (eg with short subs)

[RickS]

Not a stupid question at all.

The new camera appears to have some excellent incremental improvements but it won't be orders of magnitude better than the old one and won't produce images of the same depth in a small fraction of the integration time.

Cheers,
Rick.

[glend]

I ran some narrow band tests on M8 last night. Shot 10 x 60" Ha subs, 10 x 60" OIII subs, and 10 x 60" SII subs. I also shot Darks and Bias subs to match. I ran this all using SGP with no issues at all. I have upgraded my ASCOM Platform to the most recent release (6.2) and downloaded the newest ASI1600 driver update. The plan was to shoot a complete narrow band test sequence using 60" subs, with camera gain set at 60 (the recommendation of Jon Rista (CN), and sensor temp at -25C. I also shot the same set of narrow band subs at 90". Not surprisingly it does not take long to run this sequence using these short subs, and frankly I was surprised by the outcome.

I stacked the 60" sets and the 90" sets in DSS, with the Darks and Bias frames (no flats at this time).
Took the final Ha images (60" and 90") out to Photoshop and despite the initial dark appearance when stretched they leap out.

I won't say I am happy with the one linked here, the 60" example (the centre is blown out) but it gives you an indication of what this camera can do with such short narrowband subs. My processing skills are probably not up to handling this.
I will be continuing my testing this coming week. I may get to the point of producing a false color combined image of M8, as the OIII and SII data sets are already gathered at 60" and 90". I am tempted to go to 30" subs, in larger numbers, but would welcome any suggestions.

Link to Astrobin M8 Ha detail page:
http://www.astrobin.com/250623/

Full size here:
http://www.astrobin.com/full/250623/None/

[gregbradley]

Great stuff Glen. That's an impressive M8 for only 10 minutes total exposure time.

The central area is very bright. Is the core data still there and can be retrieved with shadows/highlights or is it gone? 60 seconds is such a short exposure time for the core t blow out. Perhaps the gain should be reduced when imaging a neb with a bright core to preserve the highlights.

Greg.

[trent_julie]

Thats impressive! What sized aperture/ focal length is the scope?

Trent

[glend]

The scope used is a Skywatcher MN190 Mak-Newt, f5.3.

Greg i am going to reduce gain and reshoot. The charts put up on CN suggest maximun SNR occurs by the time you get to 30 seconds, and there is no improvement beyond 60 seconds. So in theory more shorter subs to build integration time but the trade off is final file size.
I am sure processing experts could get better results with my data, i am really just voming to grips with it.

[Atmos]

Using the graph that Shiraz posted.
http://www.iceinspace.com.au/forum/showpost.php?p=1252539&postcount=157

It should still take about 6 minutes to become read noise limited, even at ~1.6e-. I based the calculations off of my previous setup was at F/5.7 and a KAF-8300 with a read noise of 8.5e-. It needed 90-120 minutes to reach that point!

Last night I did 10 minute subs with my ICX694 (more sensitive, smaller wells and larger pixels) and it wasn't saturated at the core. Sounds like the gain may be a tad high but you can also throw in some 1-10s subs to tone that down during processing.

[glend]

Thanks Colin for that suggestion. I agree the gain is the probable issue. I have been shooting NGC6188 tonight with the gain turned way down, and initial looks at single subs show no saturation of the core area. Interestingly i was having a look at the camera presets, and one of these is HDR (High Dynamic Range) and the gain in that preset is almost zero. I will post up NGC6188 tomorrow.

[rmuhlack]

how did you determine that from the graph?

[Atmos]

I just used the graph to get the figures, then did some calculations against figures I'd calculated a while back. It's not directly from the graph :)

[Shiraz]

this is one of the two areas of great interest with this camera - if you accept that most NB imagers under reasonable sky will not be sky-limited, the one with lowest read noise will have the deepest images (dark noise and QE being equal) :thumbsup:.

edit: and the other area of interest is the ability to use short subs with broadband and still get comparable SNR results to those from conventional cameras with long subs (sky noise and dark current will still dominate, so there are no SNR advantages, but there are possibly major ancillary advantages from short subs)

[Atmos]

That is true :) I wonder how it would go with combining an image that mostly has a gain of 250 with 5min narrowband subs and then a couple of gain 0 5min subs to collect everything other than the faint background.

[Shiraz]

naah, keep it. The QEs of the Sony CCDs still give them a significant edge in conventional imaging with medium size scopes. The big Kodak chips also still have their place because they are best matched to large, long focal length scopes or widefield with scopes like the FSQ106.

However, the 1600 opens up some entirely new ways of doing things with medium size scopes and even camera lenses (including high resolution DSO, solar and planetary imaging) and doesn't appear to give up much/any performance in conventional long-sub DSO imaging, if you want to use it that way.

Mine just arrived at the local PO, so now begins the process of changing cameras and sorting out all the little issues like focal plane tilt, vignetting, software etc - groan. One nice thing though is that the back focus is almost identical to that of the 694, so, for initial testing, I will just be able to unscrew one camera and screw on the other (that's the theory anyway)

[glend]

I have pulled down my narrowband sub of NGC6188, which was previously posted in this spot, it was rubbish and not worth posting, it is gone from Astrobin as well. The combination of low gain (30), narrowband, and short sub just did not generate enough signal. I will rerun it at some point but with a higher gain setting.

[glend]

Well Ray, good to hear it has arrived. I will be expecting some images tomorrow morning! Have fun.
PS, it would be nice to get your test results on M8 to compare settings, processing, etc.;)

[gregbradley]

I've been following this thread with interest.

It seems to me the following are critical with this camera:

1. Highlights will blow out if the gain is too high. I think gain is the same as ISO on a digital camera so then when you increase the gain you lower the dynamic range. Also typically there is usually a sweet spot in the ISO range of a digital camera. Where the signal is boosted enough to raise shadows but not introducing much noise and not lowering dynamic range too much.

So think in terms of ISO and you would aim for the lowest ISO to get the shot as noise is lowest at lowest ISO and dynamic range is greatest. If you look at dynamic range graphs of digital cameras you'll see its not a straight line drop with increases in ISO. It tends to worsen more quickly the higher you go.

From what I have seen so far this seems to be the most critical point as 2000x 1 second subs at high gain of M51 with a totally blown core and surrounds is never going to cut it against a high quality CCD camera.

2. Again the low read noise is probably down with the low gain as the higher gain probably amplifies all the noise including the read noise.

One way of working out the optimum gain (ISO) for this sensor would be to look at the DXO scoring for the EM1 this sensor came from. They measure what is the highest ISO you can use before image degradation starts. I haven't looked it up but its probably around ISO1200 for a small sensor like this. So trying to relate that to the gain settings you might be able to find a table of gain and ISO from someone who tests various cameras that way.
It would be valuable to know.

Most DSLR imagers use around ISO800 when stacking lots of images so whatever that equates to would be good to know also. Its probably quite low.

It seems that gain control is a potential image wrecker and should be treated with a little is a lot type of approach and less is better.

Greg.

[glend]

Thanks Greg, I used to shoot ISO 800 on my 450D as it was the 'sweet spot' for that sensor, but cooling it allowed me to get up to ISO1600 without too much degradation (and was a necessity for narrowband with that camera).

I am waiting for Shawn (CN) to send me a chart he is making up for me, specific to my scope fl, of SNR performance at various sub lengths (although it can be inferred from his other charts). However, no one seems to be addressing gain/preset influence on performance. Jon Rista has mentioned he thinks 60 is the 'sweet spot' but I would suggest it depends on the target to some degree.
There are still not enough ASI1600Ms getting into the hands of testers (most of the US guys that bought through retailers still have not received any delivery advice, other than the Beta Testers, there are a few that ordered direct that now have them but they are hampered by weather and light pollution so far).

Unfortunately, I am probably the last guy to be an early tester of this camera as I am such a hacker when it comes to capture and processing. I can't wait for Ray to get stuck into it.;)

[Shiraz]

nope, the read noise goes just the opposite way Greg - it reduces with increasing gain and is lowest at highest gain. However, the effective wells are not very big at high gain, so you need short subs - which you can do because of the low read noise.

The graphs posted earlier and available on the ZWO website show exactly how the various parameters vary with gain. ISO is not used as a measure of gain with these cameras, actual amplifier gain is used. This chip has made a complete transition from the world of DSLRs to the world of CCD measurements.

[Slawomir]

I also have been following this thread with a great interest. Regardless of knowing the exact numbers up to 4th decimal place for gain, QE, dynamic range etc, it appears that amateur astrophotographers have indeed a bright future, not only in terms of light pollution, but also in terms of having access to relatively inexpensive and very capable astro-cameras.

If I did not have my current camera, I would be all over this one. But it is good to know that when the day will come that my camera will retire, I will be possibly able to afford a replacement camera that hopefully will have better parameters at a lower cost :-)

I think it is not long before other manufacturers such as Atik, SX and QSI, just to name a few, will also be offering CMOS cameras.

Now, if I only could 3D print an AP mount...

[Shiraz]

thanks Glen. Just unpacked it and fired it up.

Whaaat the!! it just instantly started working. that was nice change - no software fuss at all.

[trent_julie]

Silly Question,
Given the sensitivity of these cameras; Could the colour variants overcome the advantages of monos when considering filters and the like?

Ie could I use a colour camera with a narrowband filter (something which was not recommended in the past).

Trent

[codemonkey]

It's all about trade-offs. I've been wanting better sampling, so the 3.8u pixels are attractive, vs the 4.54u of the 674. I've been wanting a bigger FOV for nebulae. The lower read noise will offset the reduced QE to an extent, but it's obviously impossible to say how well because no one knows what the QE is.

The other thing to remember is that by virtue of having a much bigger sensor, scaling down for presentation has the effect of averaging the image, improving the effective SNR.

Unless something big becomes apparent about these cameras, I'll probably sell the 674 to finance an ASI1600.

[glend]

Lee, John Rista (CN) has published a graph on the CN thread comparing the QE of the KAF8300 and the ASI1600MM. They are very close, in some part of the spectrum ASI is higher, peaking above 60, and in others like the red end the 8300 has a slight edge. Don't ask me how he did it.

[codemonkey]

Cheers Glen. He's plotted the relative QE, which is definitely useful info, but it's still not absolute QE.

My 674 has a peak absolute QE of 77% (from memory), but it's much flatter at the peak, so the shape of the relative QE tells me that I'll be losing out in red and blue, but I still don't know by how much because it doesn't tell me that the peak absolute QE of the 1600's sensor is X%.

It's estimated to be about 60%, but last I looked ZWO still didn't know exactly what it was.

[glend]

Lee, I thought he said he had calculated it. The chart is in Post #1115 here:

http://www.cloudynights.com/topic/535506-asi1600mmc-beta-test/page-45

and on page 47, Post #1159, he explains it all - not that I understand the math.

Honestly the QE is so much better than my DSLR that I won't split hairs on the absolute value. I don't know that the intention was to kill off 674's market.

As to Jon Rista work, I won't call him ' The Oracle' but he is doing some good work on this topic. He has now revised his 'Sweet Spot' recommendation, quoted here in a new post:

"There is definitely more testing needed to find the sweet spot for this camera. My original guess of a gain setting of 60 was close, but not dead on to maximize DR and minimize noise. I tested a gain of 70 last night, with bias offset of 12, and that seems to be closer but still not right on the ball. "
I will try 70 and 12 tonight on various sub lengths, and targets.

And thanks to Jerome for reprocessing my M8 image to demonstrate that the core is not blown out - it was just my hacker processing that caused it.

[codemonkey]

Nah, the graph is the plotted relative QE. The absolute is not published, and I believe it's non-trivial to determine? That's what ZWO have said anyway.

Jon's post on page 47 says If we assume we gather about the same charge per unit area, then we might gather 20.2e- total (object and skyfog) per 5.4 microns square in a short exposure of say 60 seconds.

That assumption essentially says "if we assume they have the same QE". The 1600 is assumed to be pretty much bang on with the kaf-8300, with different relative differences, but it's still assumed, no one actually knows.

Thinking about the whole SNR thing though... read noise adds in quadrature in an image. If we take zero gain on the 1600, we have a read noise of about 3.6e-/s. Squaring this we get 12.96. I measured my 674 at ~5.45e-/s, which gives me a squared value of 29.7

If my understanding is correct, that means the absolute QE would have to be (12.96/29.7 =) less than 0.43 of the 674 for me to be "worse off" in terms of SNR. So I think I'd actually still be ahead here, unless the estimated QE of the 1600 is waaaay off. Keen for someone to correct me if I'm wrong here!

[glend]

Certainly won't be me correcting you Lee.

In other activities, I have managed to reprocess my M8 Ha 10 sub stack to eliminate the core blow out. Here:

http://www.astrobin.com/250826/

Big Screen view here:

http://www.astrobin.com/full/250826/None/

I have to pay more attention to processing.

[Shiraz]

I think that all applies if read noise is the limiting factor. If sub exposures are long enough that shot noise buries the read noise, then QE wins out. The 674 is still a very effective chip, especially for Ha - it just needs much longer subs than the 1600 (sub length varies as RN^2).

[Shiraz]

just in case anyone is interested in analysing data from these cameras, just did an inspection of a FITs sub from Nebulosity and all numbers that came up were divisible by 16 - conclusion: that the lower 4 bits are padded to give 16 bit numbers using a 12 bit ADC. So if trying to work stuff out in electrons, divide the ADU by 16 before taking the gain into account.

[lazjen]

I prefer the previous version - it seemed more silky smooth to the sharp grains on this one. Maybe it's just me though. :)

[glend]

It is not just you Chris, the original is still up on the Astrobin Big Wall.

[gregbradley]

I wasn't really referring to read noise there Ray but rather overall noise. But yes I thought read noise dropped with higher ISO. Not sure why but I have heard that before.
As far as gain and amplifier in digital cameras there is amplifier gain after a certain point in the ISO in a digital camera. For example the Fuji Xpro 2 starts using amplifier gain at ISO1600. Not sure where the gain comes from below that point, I suppose its digital gain in the ADC.

I don't know about a complete transition as the images I have seen so far seem to have similar constraints - blown out highlights.

A typical DSLR image differs from an Astro CCD image mainly in the stars and highlights. Typically DSLR images have blown stars that are all white with no colour data.

Its worth being aware that this sensor may respond similarly and work around that limitation to avoid images with white stars and blown highlights.

Anyway that is just conjecture and your testing will show these things more clearly.

Greg.

[codemonkey]

Yep, I think you're right Ray. I used my trusty SNR spreadsheet and it backs up what you said (not a surprise).

By the way, I'm absolutely in agreement that the 674 is still a very effective chip. Really, it's better than the 1600. The 1600 has, however, just the blend of features that make it especially attractive to me with my specific conditions / gear, so that I think I would be happy to make the switch, as long as the 1600 does perform.

I'll be very interested to see how you find it in comparison to your 694.

[Atmos]

I just want a 694/814 chip the size of the 16803 ;) Perfection! Plus, only worth maybe $40k ;)

[Slawomir]

Plus the cost of getting an expensive telescope with larger imaging circle, 4" sturdy focuser, 50mm filters...that's why I (read: my wallet) like small ICx814 :lol:

[gregbradley]

Yes that would be heaven as long as it had deep wells.

Mind you some implementations of the KAF16200 sensor run at 6 electrons or less red noise. That is much the same as the 694 and "only" double the ASI1600 in its most likely setup.

Greg.

[codemonkey]

Coming from a 674, I'd settle for an 814 at the cost of the 1600 ;-) Seriously, an 814 would be awesome, but alas, the price point is too high for me at this point (i.e. the wife says no).

[RickS]

Put me down for one once price gets to about $10K :thumbsup:

[Slawomir]

Originally I ordered QSI 660 (674 chip), but would have had to wait a few weeks due to a large demand for 660 cameras, so I changed my order to QSI690 (814 chip), and very happy that things turned out this way! :lol:

[codemonkey]

So what you're saying is, you don't want to swap? Sad :(

[RobF]

Holy Multiscale Transform Batman! :eyepop: :lol:
Still a great image Glen. Looking to following your exploits with this new cam.

[Shiraz]

been doing some bias and dark runs. of course it's cloudy.

at high gain, this thing has almost no dark current and a vanishingly small amount of read noise - I have not found any pixels that are stuck on and even the few hot ones are really only warm. There is a little bit of large scale variation in what dark levels there are - looks a bit like wafer cutting/polishing marks maybe. However, the overall dark levels are low enough that I think dark cal can be dispensed with if dither is used.

There is a small amount of very low level impulse noise at high gain - it will be easy to process out and is of no consequence, but more intriguing is what it might be. Some of the noise locations can persist across a couple of frames, suggesting that it might be local ionising radiation dumping charge in some parts of the chip circuitry. Before anyone races out to say that the chip is noisy, please note that the exact opposite is actually the case. The level of the noise impulses is typically only a few electrons and in any other chip, they would be almost entirely hidden by read noise (eg the effect would not be visible in my H694). It is only because the 1600 is so eerily devoid of noise that this minor effect (whatever it is) can even be seen - that really appeals to the inner nerd.

[Atmos]

Sounds akin to the fact that 1% of the static that we USED to see on TV when there was no signal was radiation from the dawn of the universe (Cosmic Microwave Background). Bring on the inner nerd ;)

Looks like the camera could be a good performer and at a very reasonably price.

[glend]

I have done some Dark & Bias comparisons for different Gain and Offset settings, there is some stretched histrogram differences but really not much at least in the 60" lengths for Darks.

For example, using the SGP stretched histogram feature, a series of 10 x 60" Darks (with camera at -25C) all come in almost identical for the same Gain and Offset setting. Using the Unity Gain and Offset preset (which is Gain = 139 and Offset = 21) the histrogram shows B 15, W 612. When the same Dark sequence is run with Gain at 71, and Offset at 12 (which are the Jon Rista recommendations), I get B 15, W 388.

Moving to Bias Frames, using the same sequence setup & temperature, ten bias frames yield:
For Unity Presets (G 139, O 21) I get B15, W 548. For Rista settings (G 71, O 12) I get B 15, W 356.

So you can see that Gain and Offset settings do have some marginal impact on the Dark and Bias frames but I don't know if this warrants efforts to capture and maintain Dark library files for different camera settings.

Opinions?

[Shiraz]

Looks generally similar to what I am finding.

At 100 gain, the bias is about 290 and quite clean looking.
At 200 gain, the bias is about 220 with small flecks of noise (2 pixel spots, I guess due to pixels sharing the amps?) are starting to show. It is only a few electrons in amplitude, but possibly worth getting rid of in some circumstances.

On the basis that the bias changes in average value with gain change (even though it is a very small change - divide it by 16 to get down to the 12 bit data), I think that it might be worth doing a bias run for each power up - at least until we can establish the importance of the bias change. My guess is that a single dark set (bias removed) will be OK for each exposure and temperature, but that we may need to do a separate bias for each power up for the highest sensitivity (bias runs are delightfully quick for this chip).

In any case, it clearly will be best to choose a couple of set points to work at and then forget about varying the gain or temp. For now, I am going to use 100 and 200 gain at a temperature of -15 and see what sort of images that can produce.

good fun:thumbsup: am looking forward to the clouds getting out of the way for a little while so that I can see if the camera changeover was optically OK. We have done enough tests to be reasonably confident that the camera works OK - time for some imaging - you need some competition:lol:

[glend]

Last night was another great clear one, probably the last for awhile, so I ran some image sequences and can share the results here. What I am trying to do with this image series is dial in the camera settings (Gain and Offset) and learn how to use what combinations on various targets. There is similiar work being done on the CN forum but of course we have the better sky.

For the first series I returned to NGC6188, which I had shot a few days ago and was not happy with the lack of data acquired with the lower gain and offset settings. I ran two series of 10 x 60" Ha subs, one series with the current recommendations by Jon Rista on CN, of Gain = 70, and Offset = 12; the other series used the Unity Preset (Gain 139, and Offset = 21). For this target the Unity preset worked best and really brought out the detail.

Note: All these series were stacked with Darks and Bias frames (shot with the same settings as the lights) in DSS. Camera sensor was maintained at -25C throughout. Processing was limited to only auto levels in Photoshop, as I wanted only a basic treatment. Obviously more processing can improve them all, but that's not the purpose here.


You can see it here on Astrobin:

NGC 6188 Detail Page: http://www.astrobin.com/250900/

NGC 6188 Full Screen: http://www.astrobin.com/full/250900/0/

Next in the test series I moved over to the Trifud Nebula (M20) and conducted the same two 10 x 60" series using the same Gain and Offset settings as above. The Trifud is a brighter target and in this case the lower settings seem to work better. It would benefit from some additional shadow and highlights processing.

You can see both Trifuds on Astrobin here:

First the lower gain (71) and offset (12) settings:

Detail page: http://www.astrobin.com/250901/

Full Screen: http://www.astrobin.com/full/250901/0/

Here is the second Trifud using Unity presets:

Detail Page: http://www.astrobin.com/250903/

Full Screen: http://www.astrobin.com/full/250903/0/

I should add that I also ran OIII and SII series for these targets. It amazes me how much work this camera can get through in just a couple of hours. It took me just 30 minutes to run the entire 10 x 60" Ha, OIII and SII series. And folks this is narrowband I'm talking about.

Since the weather is turning bad I am going to turn now to processing the Ha, OIII, and SII series into false colour images of these targets. I will post them up on Astrobin in due course.
Over to you Ray.

[Shiraz]

excellent work Glen - thanks for the update.

[glend]

I have had a stab at the false colour narrowband image of the Trifud Nebula. This image is built from the Gain 71 Offset 12 subs (just 10 x 60" for each filter).

located here:

http://www.astrobin.com/250936/

click on the image to enlarge to full screen size if you wish.

It's clear to me that I need more OIII and SII data. I am tempted to combine the Ha from this G71 sequence with the Unity gain OIII and SII sequence to get a boost in those areas.
Yeah I already know my processing skills are hopeless, this is more about the camera performance.

[Shiraz]

unscrewed the H694, screwed on the 1600 and there were the stars - almost in focus :).

Was going to try it out with 1.25 filters to see how bad the vignetting really is. But before I could get any sort of image, the clouds took over and it looks like they are staying for the foreseeable future.

[Atmos]

Shouldn't you be able to get a pretty good idea on vignetting by taking a flat?

[codemonkey]

Obviously not the same camera, but check out this post on CN demonstrating the impact of turning the fan on with the ASI178 (http://www.cloudynights.com/topic/538704-asi1600-has-its-performance-matched-the-hype/page-3#entry7251185). Be very interesting to see if the 1600 is any better... if not, I'd probably have to pass.

[Shiraz]

yep, I only need one sky flat, but not a cloud flat:mad2: - and clouds are all there has been here has been since the first fleeting image showed some stars. Guess I should make up a flat box, so that I can do some calibrations when it is like this.

[Shiraz]

suspect it is a complete red herring Lee. The test was done at 0.2 arcsec sampling and shows an almost perfect diffraction pattern, so it is based on an artificial star and is waaay magnified. A real test would be on a star through the atmposphere and I suspect that will show nothing measurable on DSO imaging - should be easy to do. Would be an issue for planetary imaging, if it is that bad, but you don't need cooling for planetary.

Any camera with a fan will have fan vibration to some degree - these are the only ones that allow it to be tested because they run fast enough to see it (at video rates) and you can turn off the fan.

[glend]

No vibration from my ASI1600 fan. My bahtinov mask star test shows perfect sharp diffraction pattern, with no observable difference in fan off or fan on. If he thinks he has any vibration in his 178 he should return it under warranty. As far as that test is concerned i agree with Ray, it would be irrelevant on DSOs and in the real world of lucky planetary imaging where any sus frames are discarded from the video stream. If he wants to do anything about it, four small silicon pads under the mounting screws, would futher isolate the fan. There are no reported vibration problems from any of the Beta testers, who have done a good job in real world testing.

Ray i would like to copy your flat box build.

[codemonkey]

Cool, cheers guys. That was my last concern with this camera. My trius is now up for sale.

[Camelopardalis]

Crikey Glen, that Ha data is fantastic for only 10 minutes exposure :thumbsup:

[glend]

Thanks Dunk, it is so sensitive and easy to use. I would need more Oiii and Sii on those recent targets, as you can see from the coloured Trifud. Still learning the gain and offset for various targets. Hopefully Ray will be getting some images soon to compare notes. I know that the images can be improved by more data but i was just trying to cover alot of the learning curve quickly and get some images up for comment.

[Shiraz]

hot off the press - finally got some sky between the clouds. M83 from 84x1minute subs at gain100 and -15C. Have not yet done any sort of calibration and the setup I was using did not support dither, so the low noise of the camera was advantageous. Processing was alignment, stack, deconvolution based on measured PSF, histogram stretch, very mild noise reduction and a crop to select the most evenly illuminated region. Seeing was average and the EQ8 was only managing about 0.9arcsec RMS in the conditions. Transparency was also nothing better than average. I also have some 30 second subs and they look OK as well - haven't used them yet. The galaxy was also readily visible in 5 second focusing images.

http://astrob.in/full/251238/0/?real=&mod=None

overall, the camera worked well. I was babysitting it to see if any issues arose and caught myself regularly saying B*** heck under the breath as the low noise images poured out. No odd behaviour and there was no sign of fan vibration - also seems that 1.25 filters will be usable, but that decision will await some flat data. Will process it properly at some stage, but wanted to get something up on IIS asap. The attached small image shows the core with a more mild stretch - at 1 minute it was just under saturation levels. The theory works :).

So, FWIW, based on Glen's and my images to date, looks to me like:
the camera is very low noise - there is very little dark current, fixed pattern or read noise and there is no issue with "amp glow"
there is no problem with fan vibration
the low read noise really does allow very short subs without loss of final SNR
the effective dynamic range is excellent with a little care
Regards Ray

[gregbradley]

Very impressive Ray. A great start.

What would be great also if later if you could do a usual image with your 694 as a comparison.

Greg.

[Slawomir]

The image looks great Ray.

No doubt this camera is excellent value for money :thumbsup:

[Shiraz]

to go with earlier M83 post, here is a screen shot of a single sub (STF stretch, still with bias) showing vignetting with the 1.25 filters - it doesn't look too bad and the filter wheel can be moved about 10mm closer to the chip, which will help further.

[glend]

Nice M83 Ray. :thumbsup: I assume 1x1bin. I like the disc gas definition. Can you run put up the 30 sec version? With your Gain set at 100, what offset were you using?
I have been reading up on Gain, Offset and Bit Depth and found this article by Carig Stark useful, not saying you need it, but anyone coming from a DSLR background to this camera might find it useful:
http://www.stark-labs.com/help/blog/files/GainAndOffset.php

East coast low here so I am out of action probably until Tuesday at the earliest.:shrug:

[Shiraz]

offset 21, 1x1 bin. Craig's article is great, but it deals with getting the best dynamic range in CCD subs. The 1600 allows you to trade off some dynamic range if you wish to use very short subs or get vanishingly small read noise for NB - something CCDs cannot readily do.

[Atmos]

Very nice Ray! Looks like the camera is performing really well. I am with Greg, would love to see a comparison between this and the 694 :P

[glend]

Well if they are going to be compared, then try and show short sub performance comparisons as well as long traditional SNR performance. It would be interesting to see if the 694 can stay with the 1600 below 60".

[Atmos]

The 1600 will be slightly more resolved due to having smaller pixels, in broadband they should be very similar but the 1600 will be better and picking up very faint things due to the lower read noise. I imagine that most of the galaxy Ray showed above would look the same as they're "brighter" regions.

[bugeater]

This camera is looking very convincing. I'll be very interested to see how it goes with different sized filters. Should I ever try narrow band, I'd want to use 3nm filters, but damn they are expensive.

[glend]

Marty, Jon Rista on Cloudy Nights (post #1404 in the Beta Testing thread) offered this observation about narrowband:

Based on the stacking efficiency charts by Shawnxc, I actually wonder if 300s is more than enough for an ASI with a 7-12nm bandpass. I think for my 3nm filters, I might need up to 600s for OIII an SII. I would actually like to see the same charts that Shawn has produced for the ASI produced for the KAF-8300. So far, I am a believer that 20-60 second subs, depending on sky darkness, are all that is needed for LRGB imaging, so I think the charts model reality well enough to be a good gauge for how long we should be exposing for. It would tell us the sweet spot in exposure time for both cameras.

I use the Baader Narrowband filters (Ha 7nm, OIII 8.5m, and SII 8nm) and they are fine on this camera. Obviously it depends on the target but I find brighter Ha targets like M8, M20, etc are enough with just 60" subs, but agree that OIII and SII can require more time, again depending on the target.

Re filters, most of the beta testers on CN are using 1.25" filters, and they have to be pretty careful about getting the sensor as close to the filter as possible to avoid vignetting. However, I use 36mm filters and there are no concerns.

[glend]

Just another comment on the comparison with the 694. It is pretty obvious that just about all of the Beta Testing has been done on DSOs, and the planetary capabilities of the 1600 are for the most part ignored. This is, in part, due to the excitement about the low noise, high sensitivity nature of the sensor and system, and the background of the CN testers. Even here on IIS the discussion has be mostly related to DSO performance. This camera, with its very high frame rate video capabilties, pushed through Sharpcap and AutoStakkert is suppose to be an excellent planetary and solar camera as well. Can the 694 play in that space as well?

"One camera to rule them all" maybe a long bow to draw, but for people on a budget looking to try all aspects of imaging, it's not a bad assumption.

[Camelopardalis]

Glen, I wouldn't take too much stock in what those over the big pond are doing. Our skies are better :D (when it isn't raining!)

Maybe it's a good time to re-read Ray's thread on exposing past the read noise? While the read noise is low with the 1600, it's not zero. You'll want to expose your background beyond the total noise in your subs, for best results.

Personally, I'd be looking at one of these as a less expensive, large, cooled sensor for conventional DSO imaging. In which case, I'd only be interested in using unity gain FWIW.

[Atmos]

For planetary the 694 is not in the same league simply for the download speed. the 1600 has a high frame rate where as the 694 has a 2s download. it can be considerably sped up with subframing but not to the same extent.

What the 694 has is a higher QE. I am considering moving over to the 1600 (or at least testing between the two) because the high UV QE of the 694 is playing havoc with my refractor. As a rough guess, the 694 may be 1.5-2x as sensitive to OIII as the 1600. For me the difference between the 694 & 1600 is 250s & 60s (as Ray just showed). Or if run at gain 0, 150s. I personally would run the 1600 at gain 0 as it makes it a camera with 20K well depth and 3.6e- read noise. I prefer having less images to deal with. Could you imagine Rolf having done is 150 hour mammoth runs with 30-60s exposures!

[Camelopardalis]

I mean no disrespect, Colin, but running the 1600 at less than unity gain doesn't make a lot of sense to me...you can't capture additional dynamic range, whereas with your 694 you get full 16-bit (well, depending on the gain setting). The only thing I can conceive as an advantage is not blowing out any bright stars in the FOV. But that's just me, I'm keen to be educated ;)

If anyone wants to donate their 694 (or 674, I'm not fussy :lol:) to a good home, PM me and I'll send you my address :D

[glend]

I am sure Ray remembers the thread he started back in March 2013 about where the 694 would fit in, and comparisons made at that time with the 8300. Once again the 8300 and 694 are benchmarks for the 1600. So Ray, care to comment on the 1600 visa vie the two venerables?

[Atmos]

The only reason I would consider changing from the 694 is because of its blue response not playing well with refractors (I get large blue halos from UV detection).

After doing a bit of reading random threads on CN, ZWO mentioned that the sensor in this comes from an OMD-D-E-M1.
http://www.sensorgen.info/OlympusOM-D-E-M1.html
As Olympus use this sensor in a few of their cameras, the M10 (http://www.sensorgen.info/OlympusOM-D-E-M10.html) has a reported higher QE so I am going to assume that this is more correct for the ASI1600. 60% peak.

Eyeballing the charts, that gives it a QE of ~42% in SII, ~46% Ha and 54% in OIII.
694 (eyeballing chart for QSI660) is 65% in Ha, 75% in OIII and 61% in SII.

[Camelopardalis]

Yeah the allure of something new and shiny is overwhelming, but a 694 is a fantastic (formidable?) imaging sensor FWIW.

[Shiraz]

Glad to if you think it would be useful, but it will take a bit of thinking and that seems to be getting harderer :P - this camera requires some new assumptions - some of which are not obvious. In the meantime, a few thoughts:

If you get to sky noise limited conditions, the 694 will win because of it's high QE and the 8300 and 1600 will come second. However, for broadband, the 1600 needs maybe 1 minute subs, the 694 needs 5 minutes and the 8300 about 20 minute subs to get to sky limited performance with typical gear under typical sky, so there are very significant ancillary equipment/mount implications in favour of the 1600 - a low end mount might do well enough, plus you can sell your OAG,ONAG,large pixel guider etc and get a cheap guidescope with a QHY5L2.

With 3nm narrowband imaging, the 1600 might get to sky limited in 10-20 minutes, but neither of the other two will ever get there (practically) under dark sky (from memory, I think it required about 6 hour subs for the 8300). That gives the 1600 a big advantage if the sub lengths are limited to maybe 30 minutes, since the 1600 read noise will not intrude into an image - there will only be shot noise - but the others will have read noise as well (lots of it in the case of the 8300). Images from the 1600 should be deeper than those from the 694 and much deeper than those from the 8300. In heavy light pollution, all of the cameras should make it to sky-noise limited performance with practical sub lengths, so the 694 will win (ie require less total integration time) due to it's better QE. But the 1600 will still have the advantage of needing vastly shorter subs than the other two.

The 1600 is not magic. Sky glow, shot noise and dark noise will still ultimately limit what it can do, just like the other cameras. However, it has such low read noise that it can operate in modes that are simply not possible with the other two cameras - and that opens up lots of new opportunities. But if you are conservative by nature, you can still use it the old way - long subs will just give you a higher res version of the 8300.

[gregbradley]

Hi Ray,

I am not sure how you concluded that from Craig's article. The gain and offset had a procedure to be set to keep the resulting images within the 65535 levels that 16 bit gives. Its not dependent on exposure length but setting what the sensor will record without loss of data.

So you are saying shorter exposures at higher gain may clip dynamic range but you get better read noise with the higher gain settings?

Just out of interest what are the read noise values for this camera at different gain settings. is that known? I have read figures from 1.6 to 3.6.
Which by the way in CMOS world is quite high. I should look for it but I did read a test of the Sony A7r2 BSI 42.4mp full frame sensor read noise at some ISO levels as .3 electrons. This makes sense as OMD cameras are often reported by owners as noisy even at low ISO. The point here is not to take anything away from the ASI1600 as its a new paradigm and way cleaner than CCDs but more imagine one of those Sony full frame BSI babies in a cooled body like the ASI? Wow.

But this camera is 12 bit? Or is it 12 bit output but 16bit internal processing? (For example Sony A7 cameras output is 12 bit sometimes depending on features activated but the processing pipeline is 14 bit).

How many levels is 12 bit is 4096 levels of brightness.14 bit is 16384 levels of brightness.

Greg.

[Shiraz]

Yep, takes a bit to get the head around, but the read noise decreases as the gain increases. With a CCD the signal to read noise ratio stays much the same as the gain changes - CMOS are completely different beasts in that there is no single "optimum" setting for the 1600, whereas there is for a CCD.

It is all fully documented http://www.iceinspace.com.au/forum/attachment_browse.php?a=199985 - the read noise can be as low as ~1.2 electrons, but you have only 9 bits dynamic range at that gain setting. Sony chips will do the same thing - pity there are no large mono Sony CMOS chips. I think I read somewhere that Sony make about 800,000 image sensors a day, so they may not be all that interested in changing a production line to do a run of a few hundred mono DSLR sensors - I think that it is amazing that ZWO managed to get Panasonic to do it.


The camera has a 12 bit ADC. You end up with 16 bit numbers in the images, so I assume that the software pads out the bottom 4 bits. This is the same as for the 694, except that the 694 has random noise in the bottom 4 bits rather than zeros. In terms of image quality, no practical difference at all - since random noise and zeros are both useless.

[glend]

Ray that is enough of a synopsis for me, thank you. I was not fishing for a thesis.

[Slawomir]

Just to add another perspective to this discussion, here is a recent reply from Atik to my enquiry about implementation of the latest CMOS in their cameras:

At this point, we've found that CMOS sensors still don't compete with CCDs in the realm of long exposure, cooled astrophotography. That said, they are improving all the time and we always keep an eye on new technologies that we could use in our cameras - it has the potential to be an exciting few years!

[glend]

Lets see what happen to their sales figures, while they are waiting.;) isn't QHY working on some new CMOS cameras?

[gregbradley]

Stan Moore the author of CCDstack said long exposures are still an issue for CMOS with banding noise. I thought banding was really only an issue with Canon sensors and even then a thing of the past.


Perhaps that is what Atik are referring to or maybe they simply couldn't get CMOS sensors!

Greg.

[Shiraz]

Looks like they have some really nice CMOS cameras, including the cost leading minicam5 with a cooled Aptina CMOS and four others with cooled Sony CMOS chips up to 6mp.

[gregbradley]

Yeah but they are awfully slow. Their 16200 camera has been under development for what seems an excessive amount of time. Its still not on the market.

Greg.

[bert]

Has anybody done any testing of the ZWO in binned operation?

Was looking at this camera for dual purpose camera. Both 400mm and around 2.8meters (bin 3). Thoughts?

[Slawomir]

I'm sure the latest camera from ZWO represents awesome value and that this camera will be/is a hit among amateur astrophotographers, but there must be reasons why other manufacturers such as Atik, SX, QSI, Moravian, SBIG and FLI are not rushing, to my knowledge, with realising CMOS-based astro cameras :question:

[gregbradley]

It may be they just stick with what they know and what their established markets want. When you consider warranties and problems with software, firmware it would be easier to let someone else do all the beta testing then walk in with a better basic camera platform that takes just about any sensor once the market is established and the beta testing has shown what is needed to be practical.

Its much like Canon and Nikon not rushing in to do a good mirrorless camera leaving Sony, Fuji and Olympus to own that segment and its the only camera market segment that is actually expanding. DSLRs and point and shoots are falling markets due to mobile phones.

Of course SX already does CMOS cameras with their Ultrastar guide/planetary camera.

Greg.