QHY268M has arrived

[Spongey]

On 22/04/2021 at 14:34, tooth_dr said:

No I was just wondering about you saying about the need for pixel rejection.  I thought maybe you had a very small number of frames to work with.  But even at 600s, you would still have 25 subs, I would have assumed that was plenty for the rejection algorithm to work effectively.

Ah I see. Yes I opted for shorter subs for that reason as well as others, like less reliance on tracking / guiding etc.

With bright stars in the field I do clip star cores with 300s subs, but it's very minor. From a pure SNR standpoint I calculated (using this spreadsheet - for the ASI2600 but it's the same sensor characteristics) that I'm best off using ~9 min subs under my skies with my specific gear setup. However, the hit to read noise representation in the final image from using 5 min subs is only ~3% so I opted to use those instead, for the above reasons.

Cheers 

Edited April 23, 2021 by Spongey

[tooth_dr]

1 hour ago, Spongey said:

Ah I see. Yes I opted for shorter subs for that reason as well as others, like less reliance on tracking / guiding etc.

With bright stars in the field I do clip star cores with 300s subs, but it's very minor. From a pure SNR standpoint I calculated (using this spreadsheet - for the ASI2600 but it's the same sensor characteristics) that I'm best off using ~9 min subs under my skies with my specific gear setup. However, the hit to SNR from using 5 min subs is only ~5-10% so I opted to use those instead, for the above reasons.

Cheers 

I’ve never thought of calculating the correct exposure time properly.   Thanks very much for that link 

[teoria_del_big_bang]

3 hours ago, Spongey said:

Mode 1, Gain 56, Offset 10, 300s subs is what I use for narrowband, and it's been giving me pretty good results. 600s subs would probably be better but I'd rather take the small hit to noise and have more subs for pixel rejection to play with. I'm using 3nm filters for reference.

I'm not super experienced when it comes to NB imaging (this is my first mono camera), but an integration of a few hours has been enough for me to work with.

As I'm currently scope-less (although hopefully not for too much longer!), I've been using my Samyang 135mm at f/2.8 for imaging, but I did try the same settings at f/5.5 on my Esprit 100 before it was returned to good effect. I'm working on a widefield narrowband cygnus mosaic at the moment and I'm very happy with the results so far.

Cheers 

Pretty much the same for me, Mode 1, Gain 56, offset 30, ended up mainly using 600S subs, I did try dome 1200S but honestly saw little gain  I can upload some lights when I power up my RPi on the rig.

Steve

[tooth_dr]

3 hours ago, Spongey said:

Ah I see. Yes I opted for shorter subs for that reason as well as others, like less reliance on tracking / guiding etc.

With bright stars in the field I do clip star cores with 300s subs, but it's very minor. From a pure SNR standpoint I calculated (using this spreadsheet - for the ASI2600 but it's the same sensor characteristics) that I'm best off using ~9 min subs under my skies with my specific gear setup. However, the hit to SNR from using 5 min subs is only ~5-10% so I opted to use those instead, for the above reasons.

Cheers 

Sorry to bother you again!  I'm trying to figure out the spread sheet.  The 2600 high low gain conversion is 100, and 56 for the QHY.  But then I read below and it says when using mode #0, the gain conversion is 26? **EDITED - I  am now not sure what mode I've been using lol*

 

What effect is there on the figures in the spreadsheet for the QHY camera if the gain conversion point is different to the ZWO?

 

"

#0 Photographic Mode

This is the standard “default” capture mode. There is a high/low conversion gain cutoff point between gain 25 and 26 where you will see the read noise drop from about 5.2e- at gain 25 to around 2.5e- at gain 26. The read noise response is quite constant from gain 0 to gain 25, and again from gain 26 to about 60. Due to this, most users will probably want to use either gain 0 or gain 26 and nothing in between, as you would only be losing fullwell capacity and dynamic range in the middle. I mainly image at fast focal ratios of f/2.2 so for me it makes more sense to use gain 0. Users of slower focal ratio telescopes, or in dark skies, will most likely want to use gain 26 to take advantage of the lowered read noise. Increased exposure time is going to be recommended over increasing the gain into the “middle ranges.”

#1 – High Gain Mode

This mode features a consistently lower read noise throughout its response, with the tradeoff of a slightly lowered fullwell capacity. Like #0 Photographic Mode, there is high/low conversion gain cutoff point. For this mode the cutoff is between gain 55 and gain 56, where the read noise drops from approximately 3e- to 1.5e-. I like this mode the most for narrowband imaging as you can take advantage of the extremely low read noise at gain 56 without having to worry much about fullwell capacity. For most of my standard broadband imaging, I also prefer the lower read noise of gain 0 and still relatively high fullwell of 60ke-. Unless I am shooting a field with very bright stars in it where the higher fullwell capacity of Photographic Mode may be a better choice, this mode is my main choice for imaging with my RASA. This mode also features the best dynamic range at 14.26 stops.

"

Edited April 22, 2021 by tooth_dr

[Spongey]

20 hours ago, tooth_dr said:

Sorry to bother you again!  I'm trying to figure out the spread sheet.  The 2600 high low gain conversion is 100, and 56 for the QHY.  But then I read below and it says when using mode #0, the gain conversion is 26? **EDITED - I  am now not sure what mode I've been using lol*

 

What effect is there on the figures in the spreadsheet for the QHY camera if the gain conversion point is different to the ZWO?

 

"

#0 Photographic Mode

This is the standard “default” capture mode. There is a high/low conversion gain cutoff point between gain 25 and 26 where you will see the read noise drop from about 5.2e- at gain 25 to around 2.5e- at gain 26. The read noise response is quite constant from gain 0 to gain 25, and again from gain 26 to about 60. Due to this, most users will probably want to use either gain 0 or gain 26 and nothing in between, as you would only be losing fullwell capacity and dynamic range in the middle. I mainly image at fast focal ratios of f/2.2 so for me it makes more sense to use gain 0. Users of slower focal ratio telescopes, or in dark skies, will most likely want to use gain 26 to take advantage of the lowered read noise. Increased exposure time is going to be recommended over increasing the gain into the “middle ranges.”

#1 – High Gain Mode

This mode features a consistently lower read noise throughout its response, with the tradeoff of a slightly lowered fullwell capacity. Like #0 Photographic Mode, there is high/low conversion gain cutoff point. For this mode the cutoff is between gain 55 and gain 56, where the read noise drops from approximately 3e- to 1.5e-. I like this mode the most for narrowband imaging as you can take advantage of the extremely low read noise at gain 56 without having to worry much about fullwell capacity. For most of my standard broadband imaging, I also prefer the lower read noise of gain 0 and still relatively high fullwell of 60ke-. Unless I am shooting a field with very bright stars in it where the higher fullwell capacity of Photographic Mode may be a better choice, this mode is my main choice for imaging with my RASA. This mode also features the best dynamic range at 14.26 stops.

"

The only two useful gains imo (for DSO imaging) for the ASI2600 are gain 0 and gain 100. Everything else you just sacrifice dynamic range for minor gains elsewhere.

Gain 0 for ZWO = Gain 0, mode 1 (High gain mode) on QHY.

Gain 100 for ZWO = Gain 56, mode 1 (High gain mode) on QHY.

The other modes for QHY are useful in some fringe cases, but the majority of applications are best suited to one of the two gain / mode settings above. Offset you need to determine for your own chip, but something like 20 is a safe bet if you can't be bothered to figure it out.

 

The spreadsheet can be operated from the first two sheets only, though only the first sheet is really needed for general calculations. Input your gear and sky conditions here: 

I have this set up for my Samyang 135mm, and 3nm Ha filter. For luminance imaging enter ~300 in the filter bandwidth cell, RGB enter ~100, and narrowband enter the bandpass of your filter. The LRGB values can be tailored to your specific filter set.

A transmission efficiency of 60% is used as a rough estimate of light transmission through your optical train, sensor QE etc.

Note that a desired swamp factor of 20 is excessive; a value of 10 is typically suggested (10 will leave only 5% read noise contribution in your final image).

The camera parameters located here can be changed to your specific sensor if you want, but the default ones are fine as sample to sample variation in these chips is very low. This section can be useful to calculate your specific skyflux and magnitude, if you wish.

The important output is this section:

This calculates, for your gear and sky conditions, what exposure length you need to achieve the 'desired swamp factor' from the first input section. The units are in minutes.

I have highlighted the only two important rows (imo), which refer to gain 0 and 100 (or their QHY equivalents).

The table also gives you the number of subs it would take for you to achieve the 'total integration time' from the first section, the corresponding dynamic range in bits, and the SNR. 

As shown, for 3nm narrowband in my skies with my setup, 9.3 minute subs are 'optimal' at gain 100 / 56 (for QHY).

Note that this calculator is ONLY useful from a perspective of determining sufficient exposure length to swamp read noise. Other factors, such as those mentioned above, could lead you to choose a different exposure time. For LRGB imaging you will undoubtedly want to shoot longer subs than the calculator suggests, or you will end up with a vast amount of files.

Hope this helps!

 

Edited April 23, 2021 by Spongey

[tooth_dr]

28 minutes ago, Spongey said:

The only two useful gains imo (for DSO imaging) for the ASI2600 are gain 0 and gain 100. Everything else you just sacrifice dynamic range for minor gains elsewhere.

Gain 0 for ZWO = Gain 0, mode 1 (High gain mode) on QHY.

Gain 50 for ZWO = Gain 56, mode 1 (High gain mode) on QHY.

The other modes for QHY are useful in some fringe cases, but the majority of applications are best suited to one of the two gain / mode settings above. Offset you need to determine for your own chip, but something like 20 is a safe bet if you can't be bothered to figure it out.

 

The spreadsheet can be operated from the first two sheets only, though only the first sheet is really needed for general calculations. Input your gear and sky conditions here: 

I have this set up for my Samyang 135mm, and 3nm Ha filter. For luminance imaging enter ~300 in the filter bandwidth cell, RGB enter ~100, and narrowband enter the bandpass of your filter. The LRGB values can be tailored to your specific filter set.

A transmission efficiency of 60% is used as a rough estimate of light transmission through your optical train, sensor QE etc.

Note that a desired swamp factor of 20 is excessive; a value of 10 is typically suggested (10 will leave only 5% read noise contribution in your final image).

The camera parameters located here can be changed to your specific sensor if you want, but the default ones are fine as sample to sample variation in these chips is very low. This section can be useful to calculate your specific skyflux and magnitude, if you wish.

The important output is this section:

This calculates, for your gear and sky conditions, what exposure length you need to achieve the 'desired swamp factor' from the first input section. The units are in minutes.

I have highlighted the only two important rows (imo), which refer to gain 0 and 100 (or their QHY equivalents).

The table also gives you the number of subs it would take for you to achieve the 'total integration time' from the first section, the corresponding dynamic range in bits, and the SNR. 

As shown, for 3nm narrowband in my skies with my setup, 9.3 minute subs are 'optimal' at gain 100 / 56 (for QHY).

Note that this calculator is ONLY useful from a perspective of determining sufficient exposure length to swamp read noise. Other factors, such as those mentioned above, could lead you to choose a different exposure time. For LRGB imaging you will undoubtedly want to shoot longer subs than the calculator suggests, or you will end up with a vast amount of files.

Hope this helps!

 

Thanks for that, for taking time to make it understandable 🙏. All makes sense now

[teoria_del_big_bang]

10 hours ago, tooth_dr said:

Hi guys

Any suggestions for gain/mode/exposure time for H-alpha imaging. 
 

I grabbed a handful of H-alpha subs last night 300s/gain 56/photographic mode and thought they looked a bit noisy.  Are these the right settings for narrowband?  I’ve attached an uncalibrated stack of 11x300s

 

Thank you!

Adam

 

This is a stack of 5 Ha subs from last night, same as you they are totally uncalibrated and I am now a bit worried about the amount of hot pixels I have.
I probably have not noticed these so far due to pre-processing my previous images that will have masked these.

You can clearly see many hot pixels that are made cleared because of the dithering. Does this look normal or should a new camera be better than this ?

 

Steve

 

 

Edited April 22, 2021 by teoria_del_big_bang
wrong image

[Spongey]

Just now, teoria_del_big_bang said:

This is a stack of 5 Ha subs from last night, same as you they are totally uncalibrated and I am now a bit worried about the amount of hot pixels I have.
I probably have not noticed these so far due to pre-processing my previous images that will have masked these.

You can clearly see many hot pixels that are made cleared because of the dithering. Does this look normal or should a new camera be better than this ?

Steve

 

I personally wouldn't worry about hot pixels in uncalibrated images. If you have a few hot pixels remaining in a calibrated sub then that is fine, and is what dithering and pixel rejection algorithms are for.

[Xiga]

19 hours ago, Spongey said:

The only two useful gains imo (for DSO imaging) for the ASI2600 are gain 0 and gain 100. Everything else you just sacrifice dynamic range for minor gains elsewhere.

Gain 0 for ZWO = Gain 0, mode 1 (High gain mode) on QHY.

Gain 50 for ZWO = Gain 56, mode 1 (High gain mode) on QHY.

Thanks for the spreadsheet, very kind of you to share that with us 🙏

Can i just check something please - should the 2nd line above actually read: Gain 100 for ZWO = Gain 56, mode 1 (High gain mode) on QHY ?  

[Spongey]

33 minutes ago, Xiga said:

Thanks for the spreadsheet, very kind of you to share that with us 🙏

Can i just check something please - should the 2nd line above actually read: Gain 100 for ZWO = Gain 56, mode 1 (High gain mode) on QHY ?  

No problem at all! I didn't make this spreadsheet, someone on CN did so thank them instead!

Oops yes, good point. I will edit the original post

[MarsG76]

Excellent write up.... thanks for sharing.

 

[MarsG76]

On 02/02/2021 at 20:29, smr said:

I've just seen a 10 minute sub of the Rosette Nebula with the QHY268M, from Bortle 9, which packs more detail than 6 hours integration with my DSLR in a Bortle 5. 

😮

This is the sort of expectation I have when comparing a DSLR to QHY268M/ZWO2600MM...

[SyedT]

Update: My QHY268M's cooling board broke down randomly. Contacted QHY via Bernard at Modern Astronomy, and he did a wonderful job of securing a new camera for me as it's still well within warranty.

As it's now galaxy season, I've hooked up the 268M to my Edge HD 8. I initially thought of using it with the 0.7x reducer, but it wasn't quite compatible with my setup in terms of significant vignetting and oddly shaped stars (likely due to tilt caused by the extra length that comes with the reducer). I was also having a nightmare with collimation with the reducer, not so much now. Stars aren't perfect, but I'm not messing with collimation any further for now. Minimal vignetting compared to the reducer. I'm imaging at f/10 so still binning 2x2, and I do like the results so far as I'm able to get that extra reach (albeit at the expense of longer imaging time + the requirement for reasonable guiding performance).

Single uncalibrated 60s sub, Gain 56 & Offset 10 (yes I know the sensor needs a quick puff of air!). Single sub time was short due to a nearly full moon:

Calibrated (flats + bias, no darks):

Integration of 43x60s subs:

[rsarwar]

finally got confirmation that my 2inch LRGB filter are on the way from FLO. but the weather has gone bad.... so all i have is a bunch of Ha subs for several galaxies.

 

Is it worth while putting a IDAS P2 filter in front of the LRGB filters? not sure how to address the light pollution for LRGB subs - first time using a mono cam 

[SyedT]

11 minutes ago, rsarwar said:

finally got confirmation that my 2inch LRGB filter are on the way from FLO. but the weather has gone bad.... so all i have is a bunch of Ha subs for several galaxies.

 

Is it worth while putting a IDAS P2 filter in front of the LRGB filters? not sure how to address the light pollution for LRGB subs - first time using a mono cam 

I wouldn't add any more glass into your imaging train; IMO the less material between the sensor and the scope the better for image quality. Most streetlights nowadays are LED which are broadband (as opposed to the old sodium vapour lights which could be filtered out), and so can't be blocked out as easily without having a knock on effect on your image quality. One way to tackle LP is using shorter subs, so-called "lucky imaging". It has the added benefit of requiring not-so-good guiding, and is exactly where CMOS cameras excel. I took the below from my back garden which is Bortle 5/6 facing towards a city centre with plenty of air glow, and 2 street lights right next to it. You can see that I was able to bring out a lot of faint dust (which surprised me!). If you have a streetlight next to you which spills into your property, you can make a request to the council to add shades to direct light on to the pavement.

Edited April 29, 2021 by SyedT

[teoria_del_big_bang]

1 hour ago, rsarwar said:

Is it worth while putting a IDAS P2 filter in front of the LRGB filters? not sure how to address the light pollution for LRGB subs

 

1 hour ago, SyedT said:

wouldn't add any more glass into your imaging train; IMO the less material between the sensor and the scope the better for image quality. Most streetlights nowadays are LED which are broadband (as opposed to the old sodium vapour lights which could be filtered out), and so can't be blocked out as easily without having a knock on effect on your image quality. One way to tackle LP is using shorter subs, so-called "lucky imaging". It has the added benefit of requiring not-so-good guiding, and is exactly where CMOS cameras excel.

These two comments have got me thinking now.

Most of my recent imaging has been NB and for a long while now I have always had a 2" IDAS P2 permanantly in the image train (inside the flattener).
My thinking that it will not affect the NB imaging in anyway as the utra NB filters wavelength is outside that the IDAS filter removes anyway and that it would only help the small amount of LRGB imaging I do (only in current galaxy season when moon is not near full).

But maybe it is not helping my LRGB ?
I still assume it has little effect if any on the NB stuff.
I always intended to do some testing with it in and with it out but never got on with that.

Steve
 

[rsarwar]

2 minutes ago, teoria_del_big_bang said:

 

These two comments have got me thinking now.

Most of my recent imaging has been NB and for a long while now I have always had a 2" IDAS P2 permanantly in the image train (inside the flattener).
My thinking that it will not affect the NB imaging in anyway as the utra NB filters wavelength is outside that the IDAS filter removes anyway and that it would only help the small amount of LRGB imaging I do (only in current galaxy season when moon is not near full).

But maybe it is not helping my LRGB ?
I still assume it has little effect if any on the NB stuff.
I always intended to do some testing with it in and with it out but never got on with that.

@SyedT makes a very interesting point though. i will definately try short exposures and see how it goes. although i am not sure what to do with "

The P2 filter will definately reduce your NB photons due to transmission loses. so it depends if you want to count every possible photon or happy with a 5-8 % loss.

 

[rsarwar]

1 hour ago, SyedT said:

I wouldn't add any more glass into your imaging train; IMO the less material between the sensor and the scope the better for image quality. Most streetlights nowadays are LED which are broadband (as opposed to the old sodium vapour lights which could be filtered out), and so can't be blocked out as easily without having a knock on effect on your image quality. One way to tackle LP is using shorter subs, so-called "lucky imaging". It has the added benefit of requiring not-so-good guiding, and is exactly where CMOS cameras excel. I took the below from my back garden which is Bortle 5/6 facing towards a city centre with plenty of air glow, and 2 street lights right next to it. You can see that I was able to bring out a lot of faint dust (which surprised me!). If you have a streetlight next to you which spills into your property, you can make a request to the council to add shades to direct light on to the pavement.

 

Interesting point you make. May i you to elaborate on what you mean by "lucky imaging". light pollution is constant, unless there is a power loss

so why would level of light pollution vary from one sub to another? only reason why it would vary that i can of is if the target is moving towards or away from a distant light source, maybe a big city in the horizon.

[teoria_del_big_bang]

1 minute ago, rsarwar said:

The P2 filter will definately reduce your NB photons due to transmission loses. so it depends if you want to count every possible photon or happy with a 5-8 % loss.

I must admit I always intended to test it with and without and maybe it is not doing me any favours.
Unfortunately not a lot of NB targets about now and astro darkness fading but definitely need to try without when I can.

Steve

[SyedT]

9 minutes ago, teoria_del_big_bang said:

 

These two comments have got me thinking now.

Most of my recent imaging has been NB and for a long while now I have always had a 2" IDAS P2 permanantly in the image train (inside the flattener).
My thinking that it will not affect the NB imaging in anyway as the utra NB filters wavelength is outside that the IDAS filter removes anyway and that it would only help the small amount of LRGB imaging I do (only in current galaxy season when moon is not near full).

But maybe it is not helping my LRGB ?
I still assume it has little effect if any on the NB stuff.
I always intended to do some testing with it in and with it out but never got on with that.

Steve
 

Steve, if it ain't broke don't try to fix it! If you find that your narrowband images look reasonable then I'd leave the LP filter alone, unless you notice a significant degradation in your LRGB subs. I've not routinely heard of anyone having particularly better results with the filter in place, and it could potentially cause some issues with internal reflections; if you're not having this then I wouldn't adjust the setup. That said, it's hard to know what the difference will be unless you try it!

2 minutes ago, rsarwar said:

Interesting point you make. May i you to elaborate on what you mean by "lucky imaging". light pollution is constant, unless there is a power loss

so why would level of light pollution vary from one sub to another? only reason why it would vary that i can of is if the target is moving towards or away from a distant light source, maybe a big city in the horizon.

I suppose I'm using the term lucky imaging liberally, hence the commas. It usually refers to taking hundreds to thousands of very short exposures and stacking a small selection of these to overcome the effects of atmospheric blurring. Light pollution is constant, but atmospheric disturbances and the position of the target in the sky can have an additive effect. In order to "beat" these disturbances, a greater number of shorter exposures can be taken, and a selection of these stacked. Compared to a smaller number of longer exposures, I've found that my background skyglow in the integrated image can be a lot better. It has the added benefit of allowing guiding to be more forgiving. I end up discarding way fewer subs at the end of an imaging session.