1767 replies to this topic

[FirstC8]

Please educate me, from Jon's stretched biases and darks, am I correct the ASI1600 almost has no amp glow nor dark current at 300s exposures?

Also Sam, are you saying it cannot do more than 2000s exposures? There are certain NB exposures that will certainly benefit from longer exposures, especially when the chip has such low noise.

[Jon Rista]

My crops are only the center 512x512, so I wouldn't use them to determine amp glow. I did see a very low level of it, but they were only 300s subs. I am curious to see 2000s subs.

[wenjha]

Please educate me, from Jon's stretched biases and darks, am I correct the ASI1600 almost has no amp glow nor dark current at 300s exposures?

Also Sam, are you saying it cannot do more than 2000s exposures? There are certain NB exposures that will certainly benefit from longer exposures, especially when the chip has such low noise.

the amp-glow is very very small

and yes, there is 2000s limit

 

I do think NB imaging don't need to exposure as longer as these CCDs

because the read noise is lower 

[David Ault]

How low is the read noise?  My QSI 6120 has a read noise of 1.8e- RMS and I still need to take 3600s exposures (or longer) to get the background high enough that read noise is only about 5% of the total noise when using my Astrodon 3nm Ha filter (QE is around 70% at 656.28nm with that sensor).  That's at F/6.7 with a 90mm scope.

 

Regards,

David

[Jon Rista]

How low is the read noise?  My QSI 6120 has a read noise of 1.8e- RMS and I still need to take 3600s exposures (or longer) to get the background high enough that read noise is only about 5% of the total noise when using my Astrodon 3nm Ha filter (QE is around 70% at 656.28nm with that sensor).  That's at F/6.7 with a 90mm scope.

 

Regards,

David

 

Read noise ranges from ~3.5e- at low gain to 1.2e- at high gain. You get about 12.5 stops DR at low gain, maybe 9.5-9.7 stops at high gain. At least, based on the data from the ZWO site.

 

I am pretty amazed that you are able to reach skyfog limited subs at all, though. Not surprising it takes 3600 seconds. 

 

I haven't seen any official spec for quantum efficiency with the ASI1600 sensor yet, although someone rumored it was ~60% (don't know at what band).

[David Ault]

Based on measurements of non-nebulous regions I should probably be taking exposures of somewhere around 5400s.  However, typically when I am imaging with the 3nm Ha there is some faint Ha signal in the entire field pushing me above the read noise sooner.

 

Regards,

David

[wenjha]

How low is the read noise?  My QSI 6120 has a read noise of 1.8e- RMS and I still need to take 3600s exposures (or longer) to get the background high enough that read noise is only about 5% of the total noise when using my Astrodon 3nm Ha filter (QE is around 70% at 656.28nm with that sensor).  That's at F/6.7 with a 90mm scope.

 

Regards,

David

that's one hour exposure, amazing!

It's my first time to hear NB imaging need so long exposure

 

but why do you choose so long exposure? will 3600s equal to 360s*10 subs?

PS: your images are very nice!David

[David Ault]

Thanks Sam!

 

360s * 10 subs will be roughly equivalent to a single 3600s sub provided the primary source of noise is photon shot noise and you don't factor in download times.  With extremely narrowband filters it can take a significant amount of time before the background level is far enough above the read noise to truly be photon shot noise limited.  For my particular setup I need very long exposures to achieve this.  That isn't to say I couldn't get equivalent results by taking more subs but it would take more time to reach an equivalent SNR with the longer subs.  There's a lot in the details here.  Just how much longer would it take for me to reach an the same target SNR if I took only 2000s subs?  Maybe it is only 5% longer which could be tolerable.

 

Regards,

David

[bilgebay]

Tolga, this camera looks very promising indeed!

 

Sam, below is a target that needs very long exposures.

 

This is 5x 3600s at f/5

 

[rkayakr]

 

 

I haven't seen any official spec for quantum efficiency with the ASI1600 sensor yet, although someone rumored it was ~60% (don't know at what band).

 

Sam confirmed that the sensor is the same as in the Olympus OM-D-E-M1.

Sensorgen rates that color sensor QE as 48%.

[bilgebay]

And the price is $ 1180 till the end of May, right ?

[tolgagumus]

Here is a 900 sec dark frame

 

https://www.dropbox....fvJXfOk7sa?dl=0

 

I added it to the calibration files folder in this link

Attached Thumbnails

• 

[FiremanDan]

Yikes! That's a lot of amp glow. 

[tolgagumus]

Yikes! That's a lot of amp glow. 

1800 and 2000 second darks are coming up

[terry59]

Yikes! That's a lot of amp glow. 

 

That was my first thought and it evoked memories of my time imaging with the Nikon D50

[FiremanDan]

That's a shame. I was really looking forward to this camera. I have never seen that kind of glow on my DSLR. This makes me really nervous. 

[Jon Rista]

I have amp glow at least that bad with my 5D III. For RGB imaging, once it was calibrated out, it wasn't really an issue. The difficulty with a DSLR was matching the darks to the lights properly in order to accurately remove the entire dark current profile. With a regulated sensor temperature, that shouldn't be an issue, so removing the amp glow with dark calibration should be fine. You might experience slightly different noise in the areas affected by the flow after calibration. I've never found that to be a problem for RGB imaging. I don't know how it will pan out for NB imaging, but I guess I'll find out in a few weeks here.

[syscore]

Why do CMOS cameras suffer amp glow like this (although this is nothing compared to the other ASI cameras)? Is it just not on the chip maker's list of priorities, since the vast majority of the chips are used in regular imaging, not long exposure astro photography?

 

It will be hard for CMOS to replace CCD if they don't make CMOS chips dedicated to the same roles as CCD.

[tolgagumus]

1800 second dark

 

Again the fits file is in this folder

 

https://www.dropbox....fvJXfOk7sa?dl=0

Attached Thumbnails

• 

[Jon Rista]

Why do CMOS cameras suffer amp glow like this (although this is nothing compared to the other ASI cameras)? Is it just not on the chip maker's list of priorities, since the vast majority of the chips are used in regular imaging, not long exposure astro photography?
 
It will be hard for CMOS to replace CCD if they don't make CMOS chips dedicated to the same roles as CCD.

Not all CMOS cameras do. Glows are usually heat related. These cameras are quite compact, and often involve additional off-die processing units, like a high speed DSP. Those off-die chips often operate at pretty high frequencies, so they generate heat, which if that heat is able to reach the sensor, could increase the temperature of some pixels non-uniformly across the sensor.

Since most of the ASI cameras are also video cameras, and very very high speed video cameras at that, I suspect other processing chips probably operate at a very high frequency, and probably produce a good deal of heat.

CCD cameras usually operate with much lower frequency readout components, often in the hundreds of kilohertz range, vs. the hundreds of megahertz or higher range. But, that comes with a tradeoff as well. The G3-16200 I'm interested in has a 25 second download time...that is a pretty long time in the grand scheme of things, considering my 5D III downloads in a couple of seconds, and these ASI cameras could read out within a fraction of a second.

I think if a company like QSI or FLI started using CMOS sensors, they would likely operate the readout at a lower frequency, reducing heat, and improving quality. They may even employ some kind of active cooling on those other electronics as well.

[syscore]

It actually looks worse than it is. When I download it and stretch it side by side with the 300 sec dark, you can hardly see the amp glow. As long as it is consistent, I can't see this being detectible in a calibrated image. The "lightest" parts are only lighter by 100 e-.

 

Actually, i guess 100 ADU, not e-. I don't know what the gain is at.

Edited by syscore, 29 April 2016 - 01:27 PM.

[Jon Rista]

I dunno. I measured the difference in level between the center of the frame and the brighter parts of the glow at ~27%. Now, that is self-relative. I am expecting to be doing a lot of mosaicing with this camera...I do wonder how the glow with 2000 second subs might affect noise along the borders, and how that might affect the quality of seams in a mosaic.

[fetoma]

That amp glow is the deal breaker for me. I don't think I'll be ordering one because of it. What a pity, this camera was looking pretty good.

[syscore]

"I dunno. I measured the difference in level between the center of the frame and the brighter parts of the glow at ~27%" 

 

That's about right.

[tolgagumus]

That amp glow is the deal breaker for me. I don't think I'll be ordering one because of it. What a pity, this camera was looking pretty good.

I don't think this camera is a replacement for a CCD, but I do think paired with a small fast refractor(small pixels), small mount like an AVX or a Sirius, it will make a nice imaging package for a reasonable price.