Just bought my first monochrome camera, help me understand some things!

[Gumminess8083]

Hello everyone, as a result of my mirrorless camera dying on me, I have decided to purchase a dedicated monochrome camera, ZWO ASI294MM Pro, a filter wheel, and an lrgb filter set.

 

While the camera has not been delivered yet, I'd like to start asking some questions, so I can set everything up quicker when it does finally get here.

 

1- Gain and offset:

I have read a bit about how gain and offset work, so I was wondering if I got it right. What I plan to do is to set my gain at unity (175) and keep my offset as low as possible. If I can, I will just take longer frames to avoid having to raise the offset above 0, provided I can manage that wothout light pollution getting in the way.

 

2- Binning:

I read that binning is not as important on CMOS sensors than it is on CCD, so I should just leave it at 1x1. Does this sound correct?

 

3- Temperature:

What temperature should I set my camera to? I've read that -15 C° is a good temperature, but can I just set it to reach 15 at all times, regardless of the atmospheric conditions and outdoors temperature?

 

4- Sensor tilt:

I've read that some cameras have sensor tilt problems. How would I go about diagnosing this, and is this even something I should be worried about?

 

5- Filter and imaging order:

Suppose I'd like to image a target over multiple nights, would it be better to shoot with all 4 filters on every night, or would it be better to just do maybe l + r on one night, and g + b on the next?

 

I think I may have other questions, but at the moment these are all the ones that come to mind.

 

Thanks in advance for your replies!

[vlaiv]

1) Offset should be raised - it serves a purpose.

Best way to get good offset is to do following:

- set offset to some value

- take dozen or so bias subs

- stack them using minimum function (instead of average) and find minimum pixel in resulting stack.

- If minimum pixel is larger than 0 (or other minimal reported value for that camera), then you are done, else increase offset a bit and return to second point in this list

I like using unity gain and advise it, but as far as I know color model of that camera had some non linearity issues and people often used higher gain value. Not sure if that is also true for mono model (as far as I know they are slightly different sensors in mono and color version so mono might not have this issue). Do use unity gain but maybe first do a search if nonlinearity issue is only for color model or for mono as well. If you run into issues with flats - it could be because of this - then raise gain to 200 and something (do search on topic - you'll find recommendation), or we could ask @Adam J?

2) No it's not correct. Binning is as important on CMOS as on CCD - it is only done differently and it slightly differs in result.

If you need to bin - then bin, and if you don't need to bin - then don't. Benefit of software binning is that you can decide to bin after you do recording - you don't have to decide in advance like with CCD (in order to have that small benefit that CCD binning has over CMOS one).

3) Set temperature as low as you can manage given your ambient conditions. Cameras have deltaT - how much they can cool below ambient. If for example deltaT is 35C and in the evening temperature drops below 15C (as does in winter) - then feel free to cool to -20C, camera will be able to do it.

It is not unusual to have different target temperatures in winter and in summer because ambient conditions change. I use -20C in winter and -15C in summer, and I even had issues with -15C on some summer nights. If this happens - then go with temperature that you can achieve on a given night - you can always shoot matching set of darks afterwards to calibrate your data (that is the point of camera cooling really - to be able to recreate target temperature for dark calibration).

4) You will see it in your image. If you happen to have issues with corner stars - like stars being elongated or out of focus in one or two corners while others are fine - then you might have a tilt.

Tilt can be due to camera or due to focuser. If you rotate camera and shoot same target and elongated stars also rotate in FOV - then it is sensor / camera related tilt, but if star issues remain in the same place with respect to actual target in the image - then you have focuser tilt.

In any case - don't worry about it until you need to worry about it - and you will know if it comes to that.

5) That really depends.

It is best to shoot L when seeing is good as it carries the most of detail information / sharpness. Color part does not carry nearly as much detail (and people often bin color to capture it faster, because of this). However - you never know if first or second night will have better seeing.

Similarly - you want to do L and B on night of good transparency - because again L will bring in detail and SNR (most noise also comes from luminance so you want good luminance data). Blue is often weakest of the colors - both because attenuation / scatter is greatest in those wavelengths (that is why sky is blue) but also sensor sensitivity is lowest in blue (with few camera exceptions that have high QE in blue part of spectrum).

If you shoot part of subs on one evening and part of them on other - you run a risk of having subs with different SNR in your stack. That is not ideal, and if anything try using stacking algorithm that assigns weights to each sub to compensate at least partially for this.

 

[Adam J]

2 hours ago, vlaiv said:

1) Offset should be raised - it serves a purpose.

Best way to get good offset is to do following:

- set offset to some value

- take dozen or so bias subs

- stack them using minimum function (instead of average) and find minimum pixel in resulting stack.

- If minimum pixel is larger than 0 (or other minimal reported value for that camera), then you are done, else increase offset a bit and return to second point in this list

I like using unity gain and advise it, but as far as I know color model of that camera had some non linearity issues and people often used higher gain value. Not sure if that is also true for mono model (as far as I know they are slightly different sensors in mono and color version so mono might not have this issue). Do use unity gain but maybe first do a search if nonlinearity issue is only for color model or for mono as well. If you run into issues with flats - it could be because of this - then raise gain to 200 and something (do search on topic - you'll find recommendation), or we could ask @Adam J?

2) No it's not correct. Binning is as important on CMOS as on CCD - it is only done differently and it slightly differs in result.

If you need to bin - then bin, and if you don't need to bin - then don't. Benefit of software binning is that you can decide to bin after you do recording - you don't have to decide in advance like with CCD (in order to have that small benefit that CCD binning has over CMOS one).

3) Set temperature as low as you can manage given your ambient conditions. Cameras have deltaT - how much they can cool below ambient. If for example deltaT is 35C and in the evening temperature drops below 15C (as does in winter) - then feel free to cool to -20C, camera will be able to do it.

It is not unusual to have different target temperatures in winter and in summer because ambient conditions change. I use -20C in winter and -15C in summer, and I even had issues with -15C on some summer nights. If this happens - then go with temperature that you can achieve on a given night - you can always shoot matching set of darks afterwards to calibrate your data (that is the point of camera cooling really - to be able to recreate target temperature for dark calibration).

4) You will see it in your image. If you happen to have issues with corner stars - like stars being elongated or out of focus in one or two corners while others are fine - then you might have a tilt.

Tilt can be due to camera or due to focuser. If you rotate camera and shoot same target and elongated stars also rotate in FOV - then it is sensor / camera related tilt, but if star issues remain in the same place with respect to actual target in the image - then you have focuser tilt.

In any case - don't worry about it until you need to worry about it - and you will know if it comes to that.

5) That really depends.

It is best to shoot L when seeing is good as it carries the most of detail information / sharpness. Color part does not carry nearly as much detail (and people often bin color to capture it faster, because of this). However - you never know if first or second night will have better seeing.

Similarly - you want to do L and B on night of good transparency - because again L will bring in detail and SNR (most noise also comes from luminance so you want good luminance data). Blue is often weakest of the colors - both because attenuation / scatter is greatest in those wavelengths (that is why sky is blue) but also sensor sensitivity is lowest in blue (with few camera exceptions that have high QE in blue part of spectrum).

If you shoot part of subs on one evening and part of them on other - you run a risk of having subs with different SNR in your stack. That is not ideal, and if anything try using stacking algorithm that assigns weights to each sub to compensate at least partially for this.

 

So the interesting thing is that what people often think of as the mono version of the IMX294 mostly because it's called the ASI294mm pro is actually somewhat uniquely, as far as I know without going through the entire sony sensor catalogue (although I am reasonably familiar with it) the only instance when Sony have given what is apparently a mono version of the sensor a different IMX designation. In this case the IMX492, which just happens to be 294 backwards. So you could conclude that something more changed in moving to a mono version than simply removing the Bayer matrix and updating the firmware. This is likely because of the need to read the sub pixels of the quad Bayer individually to enable the full resolution bin 1 as opposed to in pairs or groups of four as in the OSC dependent on read mode. My personal feeling with this sensor is that much of the widely reported issues with the 294 in terms of calibration result from the way it reads the quad baryer matrix or rather the circuit that switches the sensor from quad baryer mode to HDR mode. 

One thing seems clear and that is that the pattern visible in imx294 flats is compounded by the apparent non linearity in the chip at some gain values and that these two issues together create the perfect storm.  Many explanations have been proposed for this pattern and it is still present in the mono chip but even so I don't see thread after thread of people complaining about calibration of the IMX492 in the way you do with the IMX294 however, a quick Google search still reveals a unsettling pattern. Furthermore testing undertaken by owners confirms the odd behaviours present in the 294 still persist in the 492. The issue you have with both chips is that the pattern in changing from one sample to the next creates and issue for some and not for others. Hence you also find people with good copies insisting that the problem doesn't exist and everyone else is just rubbish at calibration. 

Putting all that together I would treat them the same way and use gain 200 which has been shown to result in more predictable behaviour, you will still have plenty of dynamic range at that gain. But for me it's the reason I'll be sticking with my ASI1600mm pro until a more forgiving 4/3 mono sensor replacement is available in the future. This stuff is difficult enough as it is without a lottery on chip quality and the resulting potential that you will need to play nurse to the calibration process. 

Of course the flip side is that you may get a good copy or one that calibrates with a little knowledge, in which case you have a very sensitive camera indeed, even more so than the imx571 notably in the deep red for Ha and SII. 

 

Adam 

 

Edited October 6, 2022 by Adam J

[scotty38]

I have the QHY version but keep it simple to start with and put offset to about 30, gain at unity as Vlaiv says and leave it in 11M mode. Then you can work on getting decent flats. I'm assuming you bought the ZWO package that comes with the filter masks so you should have minimal reflections and be pretty much good to go.

[Clarkey]

10 hours ago, vlaiv said:

Set temperature as low as you can manage given your ambient conditions.

I would add that this is true for running on a mains supply. If you are running on a battery supply, you might want to increase this to reduce the power consumed.

[vlaiv]

36 minutes ago, Clarkey said:

I would add that this is true for running on a mains supply. If you are running on a battery supply, you might want to increase this to reduce the power consumed.

Very good point - never considered it, but yes - there is power consumption indicator in software (cooler running at XX percent) and I bet it will be much less if temperature is not set as low as possible.

[scotty38]

Oh yes forgot that bit but I've never found going less than -5 is needed but that could well be because I'm not fussy 🙂

[vlaiv]

34 minutes ago, scotty38 said:

Oh yes forgot that bit but I've never found going less than -5 is needed but that could well be because I'm not fussy 🙂

It does not have much to do with being fussy - there is really a point where going cooler simply does not matter as much.

That depends on use case though, and is different for different setups and conditions.

Dark current doubles roughly at every 6C difference. If you are using -5 and someone is using -20, then that is (let's round up) three times doubling, or x8 more dark current.

Now, by itself dark current is no biggie - it is dark current noise that is sometimes issue and it is equal to square root of dark current, so we increase dark current noise by factor of about x2.83.

If for exposure time you are using, read noise is x10 larger than dark current noise at -20C, then you won't see much if any difference if you cool to only -5C as read noise will swamp dark current noise by factor of x3 and a bit.

This is also related to LP noise. One should aim for exposures that swamp read noise with LP noise. Even if at -5C dark current noise is comparable to read noise - it will also be swamped by LP noise.

According to this:

-5C has dark current of about 0.012 or something like that?

In order for dark current noise to be the same as read noise (around 2e) - dark current must be 4e and exposure time therefore around 4/0.012 = 333.333s

If you are swamping read noise with LP noise with exposures below 5 minutes - yes, with this camera model, you can even cool to -5C, it won't make significant difference (yes, there will be improvement by going cooler, but barely detectable from measurement and certainly imperceptible by eye).

[barbulo]

The original resolution of this camera was 11Mpx - 4144x2822 pixels (bin2). ZWO later “unlocked” the bin1 mode: 44Mpx - 8288x5644 pixels. It might be appealing, but you have to bear in mind that in bin1, the pixel size is 2,31x2,31 microns so the light gathering capability decreases dramatically. Just look at the charts in the manual and you’ll realize that the bin1 mode is much less effective. However, you always have it there in case you might need want to use it (perhaps with scopes FL<300mm).

As stated by @vlaiv, good luminance data is paramount. Add as much integration time as you can (you’ll never know what night is going to be better) and finally add some RGB to color the luminance layer. I personally take between 0,5-1,0 hours per color channel.

I use gain 120, offset 30 and -20ºC. YMMV though.

I haven’t experienced any issues with my flats. Following some advice I read somewhere, I use exposure times over 1s: from 2s for luminance to 30s for SII.

[scotty38]

45 minutes ago, vlaiv said:

It does not have much to do with being fussy - there is really a point where going cooler simply does not matter as much.

That depends on use case though, and is different for different setups and conditions.

Dark current doubles roughly at every 6C difference. If you are using -5 and someone is using -20, then that is (let's round up) three times doubling, or x8 more dark current.

Now, by itself dark current is no biggie - it is dark current noise that is sometimes issue and it is equal to square root of dark current, so we increase dark current noise by factor of about x2.83.

If for exposure time you are using, read noise is x10 larger than dark current noise at -20C, then you won't see much if any difference if you cool to only -5C as read noise will swamp dark current noise by factor of x3 and a bit.

This is also related to LP noise. One should aim for exposures that swamp read noise with LP noise. Even if at -5C dark current noise is comparable to read noise - it will also be swamped by LP noise.

According to this:

-5C has dark current of about 0.012 or something like that?

In order for dark current noise to be the same as read noise (around 2e) - dark current must be 4e and exposure time therefore around 4/0.012 = 333.333s

If you are swamping read noise with LP noise with exposures below 5 minutes - yes, with this camera model, you can even cool to -5C, it won't make significant difference (yes, there will be improvement by going cooler, but barely detectable from measurement and certainly imperceptible by eye).

Thanks Vlaiv, very interesting as usual. For my much more simple brain and in my Bortle 4/3 and my "usual" 300s narrowband exposures does this mean that, all things being equal, I "should" be cooling lower than -5 then?

Edited October 6, 2022 by scotty38

[vlaiv]

3 minutes ago, scotty38 said:

Thanks Vlaiv, very interesting as usual. For my much more simple brain and in my Bortle 4/3 and my "usual" 300s narrowband exposures does this mean that, all things being equal, I "should" be cooling lower than -5 then?

Yes, that is one example where going with lower temperature can show actual benefit. Exposure is long enough for dark current to build up to significant level and filter removes LP so LP noise is not most significant factor any more.

If you can, in that particular case cool down more.

[scotty38]

Thanks and will do. There is nothing stopping me doing so I just used -5 as it was easier on the kit more than anything else to be fair.

[Gumminess8083]

21 hours ago, vlaiv said:

1) Offset should be raised - it serves a purpose.

Best way to get good offset is to do following:

- set offset to some value

- take dozen or so bias subs

- stack them using minimum function (instead of average) and find minimum pixel in resulting stack.

- If minimum pixel is larger than 0 (or other minimal reported value for that camera), then you are done, else increase offset a bit and return to second point in this list

I like using unity gain and advise it, but as far as I know color model of that camera had some non linearity issues and people often used higher gain value. Not sure if that is also true for mono model (as far as I know they are slightly different sensors in mono and color version so mono might not have this issue). Do use unity gain but maybe first do a search if nonlinearity issue is only for color model or for mono as well. If you run into issues with flats - it could be because of this - then raise gain to 200 and something (do search on topic - you'll find recommendation), or we could ask @Adam J?

2) No it's not correct. Binning is as important on CMOS as on CCD - it is only done differently and it slightly differs in result.

If you need to bin - then bin, and if you don't need to bin - then don't. Benefit of software binning is that you can decide to bin after you do recording - you don't have to decide in advance like with CCD (in order to have that small benefit that CCD binning has over CMOS one).

3) Set temperature as low as you can manage given your ambient conditions. Cameras have deltaT - how much they can cool below ambient. If for example deltaT is 35C and in the evening temperature drops below 15C (as does in winter) - then feel free to cool to -20C, camera will be able to do it.

It is not unusual to have different target temperatures in winter and in summer because ambient conditions change. I use -20C in winter and -15C in summer, and I even had issues with -15C on some summer nights. If this happens - then go with temperature that you can achieve on a given night - you can always shoot matching set of darks afterwards to calibrate your data (that is the point of camera cooling really - to be able to recreate target temperature for dark calibration).

4) You will see it in your image. If you happen to have issues with corner stars - like stars being elongated or out of focus in one or two corners while others are fine - then you might have a tilt.

Tilt can be due to camera or due to focuser. If you rotate camera and shoot same target and elongated stars also rotate in FOV - then it is sensor / camera related tilt, but if star issues remain in the same place with respect to actual target in the image - then you have focuser tilt.

In any case - don't worry about it until you need to worry about it - and you will know if it comes to that.

5) That really depends.

It is best to shoot L when seeing is good as it carries the most of detail information / sharpness. Color part does not carry nearly as much detail (and people often bin color to capture it faster, because of this). However - you never know if first or second night will have better seeing.

Similarly - you want to do L and B on night of good transparency - because again L will bring in detail and SNR (most noise also comes from luminance so you want good luminance data). Blue is often weakest of the colors - both because attenuation / scatter is greatest in those wavelengths (that is why sky is blue) but also sensor sensitivity is lowest in blue (with few camera exceptions that have high QE in blue part of spectrum).

If you shoot part of subs on one evening and part of them on other - you run a risk of having subs with different SNR in your stack. That is not ideal, and if anything try using stacking algorithm that assigns weights to each sub to compensate at least partially for this.

 

Thanks for your reply!

 

1- Gotcha about the offset. I'll try using that method with Pixinsight and see what I can achieve. As for bias, I'll be following Adam's suggestions

 

2- Good to know that I can bin in post and not care about it too much while shooting. I'm still learning about it, so I feel safer knowing that I can just shoot withou risking messing it up.

 

3- Temps here range from 0-10 in the winter to 30-40 in the summer, so I'll see what I can do to keep the temperature the same between shots and aim as low as I can go!

 

4- I'll hold off and try not to worry too much about this for the moment, then.

 

5- Thanks for explaining in detail how this all works! I'll keep this in mind when I get to shoot my first sequence.

[Gumminess8083]

19 hours ago, Adam J said:

So the interesting thing is that what people often think of as the mono version of the IMX294 mostly because it's called the ASI294mm pro is actually somewhat uniquely, as far as I know without going through the entire sony sensor catalogue (although I am reasonably familiar with it) the only instance when Sony have given what is apparently a mono version of the sensor a different IMX designation. In this case the IMX492, which just happens to be 294 backwards. So you could conclude that something more changed in moving to a mono version than simply removing the Bayer matrix and updating the firmware. This is likely because of the need to read the sub pixels of the quad Bayer individually to enable the full resolution bin 1 as opposed to in pairs or groups of four as in the OSC dependent on read mode. My personal feeling with this sensor is that much of the widely reported issues with the 294 in terms of calibration result from the way it reads the quad baryer matrix or rather the circuit that switches the sensor from quad baryer mode to HDR mode. 

One thing seems clear and that is that the pattern visible in imx294 flats is compounded by the apparent non linearity in the chip at some gain values and that these two issues together create the perfect storm.  Many explanations have been proposed for this pattern and it is still present in the mono chip but even so I don't see thread after thread of people complaining about calibration of the IMX492 in the way you do with the IMX294 however, a quick Google search still reveals a unsettling pattern. Furthermore testing undertaken by owners confirms the odd behaviours present in the 294 still persist in the 492. The issue you have with both chips is that the pattern in changing from one sample to the next creates and issue for some and not for others. Hence you also find people with good copies insisting that the problem doesn't exist and everyone else is just rubbish at calibration. 

Putting all that together I would treat them the same way and use gain 200 which has been shown to result in more predictable behaviour, you will still have plenty of dynamic range at that gain. But for me it's the reason I'll be sticking with my ASI1600mm pro until a more forgiving 4/3 mono sensor replacement is available in the future. This stuff is difficult enough as it is without a lottery on chip quality and the resulting potential that you will need to play nurse to the calibration process. 

Of course the flip side is that you may get a good copy or one that calibrates with a little knowledge, in which case you have a very sensitive camera indeed, even more so than the imx571 notably in the deep red for Ha and SII. 

 

Adam 

 

Thanks for the detailed explanation! Even though I'm still a novice and not very familiar with all the intricacies of how cameras work, I think I get the gist of what you're saying and will stick with gain 200 for the forst few sequences.

 

In the future, I may test to see if I can go lower without encountering any problems, but I don't think I'm yet experienced enough to notice those probles, even if I do come across them...

[Gumminess8083]

12 hours ago, scotty38 said:

I have the QHY version but keep it simple to start with and put offset to about 30, gain at unity as Vlaiv says and leave it in 11M mode. Then you can work on getting decent flats. I'm assuming you bought the ZWO package that comes with the filter masks so you should have minimal reflections and be pretty much good to go.

Thanks for your reply!

I bought everything from a third party reseller, since import fees would have been too high otherwise...

 

I got the standalone camera + filter wheel separately, and then the lrgb filters directly from Astronomik's website, so I'm not sure if they come with masks, or if I'll even need them, since they are 1.25" pre-mounted

[Gumminess8083]

9 hours ago, barbulo said:

The original resolution of this camera was 11Mpx - 4144x2822 pixels (bin2). ZWO later “unlocked” the bin1 mode: 44Mpx - 8288x5644 pixels. It might be appealing, but you have to bear in mind that in bin1, the pixel size is 2,31x2,31 microns so the light gathering capability decreases dramatically. Just look at the charts in the manual and you’ll realize that the bin1 mode is much less effective. However, you always have it there in case you might need want to use it (perhaps with scopes FL<300mm).

As stated by @vlaiv, good luminance data is paramount. Add as much integration time as you can (you’ll never know what night is going to be better) and finally add some RGB to color the luminance layer. I personally take between 0,5-1,0 hours per color channel.

I use gain 120, offset 30 and -20ºC. YMMV though.

I haven’t experienced any issues with my flats. Following some advice I read somewhere, I use exposure times over 1s: from 2s for luminance to 30s for SII.

If I'm understanding your post correctly, since I am actually using a 250mm redcat, I should probably consider going with bin1 rather than higher values, correct?

[scotty38]

2 hours ago, Gumminess8083 said:

Thanks for your reply!

I bought everything from a third party reseller, since import fees would have been too high otherwise...

 

I got the standalone camera + filter wheel separately, and then the lrgb filters directly from Astronomik's website, so I'm not sure if they come with masks, or if I'll even need them, since they are 1.25" pre-mounted

Ah ok not an issue then with the mounted filters.

[wimvb]

I have the ASI294MM Pro with zwo filter wheel and oag. This camera has much better specs at bin 2 than bin 1, and 4x smaller file size. Just check the ZWO product page. Smaller file size means faster download, and a lot faster image processing during stacking. In my experience, cmos cameras give best results if you stack many subs, so having smaller files does make a difference in processing speed and storage. I believe that ZWO reduce amp glow by powering down unnecessary circuitry on the chip and in the camera during exposure. This circuitry is then only turned on during read out. The camera has star burst amp glow, which needs matching darks to calibrate out. I have found that a low temperature isn't a must and run it at -10 C during autumn and at -20 C during winter. The most important thing is to keep the temperature stable and below ambient, or calibration may not work properly.

For LRGB I use the camera at 0 gain, offset 8, and for narrow band 200 gain, offset 30. Exposure times at f/5.3 are 180s for L, 300s for RGB and 240s for Ha (7nm) at a Bortle 4/5 site. The exposures are long enough to drown fixed pattern "noise", while also short enough to avoid overexposing any but the brightest stars.

Hope this helps.

[wimvb]

On 05/10/2022 at 23:23, Gumminess8083 said:

5- Filter and imaging order:

Suppose I'd like to image a target over multiple nights, would it be better to shoot with all 4 filters on every night, or would it be better to just do maybe l + r on one night, and g + b on the next?

Use the filter wheel unidirectionally (can be set in the driver) and alternate between l, r, g, and b during the night. At least if you don't need to refocus between filters. That way, if the weather doesn't cooperate, you at least have a full colour set every night. And also, if for some reason one night's subs are bad, you can still create an image with what you have. At my site, fwhm may drift during one night, but most often, there is more variation between nights than within the same night. Collecting all colours during a night gives me more consistent results.

Edited October 6, 2022 by wimvb

[barbulo]

10 hours ago, Gumminess8083 said:

If I'm understanding your post correctly, since I am actually using a 250mm redcat, I should probably consider going with bin1 rather than higher values, correct?

In terms of arcsecs/px resolution yes (aorund 2 is a good value), but with lower DR and FW than with bin1.

[Gumminess8083]

On 05/10/2022 at 23:48, vlaiv said:

1) Offset should be raised - it serves a purpose.

Best way to get good offset is to do following:

- set offset to some value

- take dozen or so bias subs

- stack them using minimum function (instead of average) and find minimum pixel in resulting stack.

- If minimum pixel is larger than 0 (or other minimal reported value for that camera), then you are done, else increase offset a bit and return to second point in this list

I like using unity gain and advise it, but as far as I know color model of that camera had some non linearity issues and people often used higher gain value. Not sure if that is also true for mono model (as far as I know they are slightly different sensors in mono and color version so mono might not have this issue). Do use unity gain but maybe first do a search if nonlinearity issue is only for color model or for mono as well. If you run into issues with flats - it could be because of this - then raise gain to 200 and something (do search on topic - you'll find recommendation), or we could ask @Adam J?

2) No it's not correct. Binning is as important on CMOS as on CCD - it is only done differently and it slightly differs in result.

If you need to bin - then bin, and if you don't need to bin - then don't. Benefit of software binning is that you can decide to bin after you do recording - you don't have to decide in advance like with CCD (in order to have that small benefit that CCD binning has over CMOS one).

3) Set temperature as low as you can manage given your ambient conditions. Cameras have deltaT - how much they can cool below ambient. If for example deltaT is 35C and in the evening temperature drops below 15C (as does in winter) - then feel free to cool to -20C, camera will be able to do it.

It is not unusual to have different target temperatures in winter and in summer because ambient conditions change. I use -20C in winter and -15C in summer, and I even had issues with -15C on some summer nights. If this happens - then go with temperature that you can achieve on a given night - you can always shoot matching set of darks afterwards to calibrate your data (that is the point of camera cooling really - to be able to recreate target temperature for dark calibration).

4) You will see it in your image. If you happen to have issues with corner stars - like stars being elongated or out of focus in one or two corners while others are fine - then you might have a tilt.

Tilt can be due to camera or due to focuser. If you rotate camera and shoot same target and elongated stars also rotate in FOV - then it is sensor / camera related tilt, but if star issues remain in the same place with respect to actual target in the image - then you have focuser tilt.

In any case - don't worry about it until you need to worry about it - and you will know if it comes to that.

5) That really depends.

It is best to shoot L when seeing is good as it carries the most of detail information / sharpness. Color part does not carry nearly as much detail (and people often bin color to capture it faster, because of this). However - you never know if first or second night will have better seeing.

Similarly - you want to do L and B on night of good transparency - because again L will bring in detail and SNR (most noise also comes from luminance so you want good luminance data). Blue is often weakest of the colors - both because attenuation / scatter is greatest in those wavelengths (that is why sky is blue) but also sensor sensitivity is lowest in blue (with few camera exceptions that have high QE in blue part of spectrum).

If you shoot part of subs on one evening and part of them on other - you run a risk of having subs with different SNR in your stack. That is not ideal, and if anything try using stacking algorithm that assigns weights to each sub to compensate at least partially for this.

 

Just got my camera and ran some tests!

Apparently, with gain at 200, as Adam suggested, it looks like offset can be kept pretty low.

This is what a stack of 13 bias frames with 0 offset look like, so I could probably get away with just bumping my offset up to 1 or 2

As for the chip itself, might I bother you, @Adam J , for another question?

You mention some paatterns being evident in the flats when taken at lower gain values, so I was wondering what these would look like, so I could try and run some tests to see if I can, by any chance, lower this value without running into any trouble!

[Adam J]

8 minutes ago, Gumminess8083 said:

Just got my camera and ran some tests!

Apparently, with gain at 200, as Adam suggested, it looks like offset can be kept pretty low.

This is what a stack of 13 bias frames with 0 offset look like, so I could probably get away with just bumping my offset up to 1 or 2

As for the chip itself, might I bother you, @Adam J , for another question?

You mention some paatterns being evident in the flats when taken at lower gain values, so I was wondering what these would look like, so I could try and run some tests to see if I can, by any chance, lower this value without running into any trouble!

Patterns are present irrespective of gain if you have them. 

https://www.cloudynights.com/topic/779735-asi294mm-flats-lets-compare-notes/

Some images and other info in this thread. 

Adam 

[Gumminess8083]

On 15/10/2022 at 22:43, Adam J said:

Patterns are present irrespective of gain if you have them. 

https://www.cloudynights.com/topic/779735-asi294mm-flats-lets-compare-notes/

Some images and other info in this thread. 

Adam 

Hello Adam, sorry to resurrect this old thread!

I finally got around to imaging some targets and I am indeed suffering from some of the problems you mentioned regarding the flat frames (similar to those mentioned in the linked thread).

I tried taking them under various conditions (sky flats, phone screen, laptop screen, tablet, ...) with various gain and offset settings, but they alway come out looking weird (not all filters present the same patterns. For example the OIII filter seems to be perfectly even without any artifacts).

I tried changing the settings around and following some suggestions I saw online, but the results always comes out looking like this...
I'd be grateful if you had any recommendations of things I can try out!

[geeklee]

48 minutes ago, Gumminess8083 said:

flat frames

Hi @Gumminess8083 Please see this thread where I pasted an image of my narrowband flats with the 294MM at Gain 120 (unity).

The image above mine was another example from a 294MM user.

These flats fully calibrate (and you can see the same pattern in my flats and in my Ha lights - for example - when stretched hard).

Edited November 2, 2022 by geeklee

[Gumminess8083]

4 hours ago, geeklee said:

Hi @Gumminess8083 Please see this thread where I pasted an image of my narrowband flats with the 294MM at Gain 120 (unity).

The image above mine was another example from a 294MM user.

These flats fully calibrate (and you can see the same pattern in my flats and in my Ha lights - for example - when stretched hard).

Thanks for your reply and for sharing your experinece!

Unfortunately, it seems like this doesn't hold true for my flats...

Here you can see that the integrated and calibrated images seem to present the same pattern as the flat frames, almost as if it was introduced by the flats themselves (keep in mind these images have been locally normalized, hence the gradient going from right to left—I used an O frame as the reference—but the problem is still present, even when I skip that phase).

If it's not asking for too much, I have uploaded some of my images to the following link: https://drive.google.com/drive/folders/1veeZFJm2dMvpqpiGXjTFDKqWJePYP2CA?usp=sharing
 

Is there any chance you could have a look to see if anything strikes you as different from your data?

Just in case it changes something, I've been working following this tutorial/workflow for narrowband data: https://www.chaoticnebula.com/workflows/

Edited November 2, 2022 by Gumminess8083