Settings for ASI183MC?

[AstroViking]

Hi all,

I've had the rig out the last two nights figuring out the magics of a dedicated OSC camera and an EAF. It's been interesting!

I did a lot of reading on various forums and web sites and they all said that a gain of 120 was 'Unity' gain for the '183, and to avoid using a gain under 50 because you'll get terrible colour banding. Also to take dark frames as there is amp-glow.

When I stacked the first set of images, it was very vignetted - far worse than what I got with my DSLR. (Just realised I should have posted an image. Will do so tonight when I get home from work.)

Telescope: EvoStar 72ED + 0.85 flattener (357mm focal length)
Camera: ZWO ASI183MC-Pro

I have 55mm of backfocus between the flattener and the camera sensor, as per the ZWO instructions.

I took dark frames as well, as I read that the '183 needs them. Still struggling to get flats sorted out (both the light source and the camera settings to avoid over-exposing everything).

Camera settings:
Gain: 100
Temperature: 0 degrees C
Offset: 40 (Ekos default setting)
Exposure: 30 seconds

Environment:
Suburban Melbourne - Bortle 5/6
No moon

DSLR specs:
Nikon D7100 (6000x4000 pixels, 23.5x15.6mm sensor, 3.9um pixel size)

ASI183 specs:
5496x3672 pixels, 13.x8.8mm sensor, 2.4um pixel size


My questions are as follows:

1) What settings should I be using for the camera to avoid over-exposing all the subs?

2) How do I take bias frames? What are the recommended camera settings for taking flats?

3) Is there anything I can do to get rid of the vignetting? I had to crop the outer quarters off the image, and even then corners were still a lot darker than the centre.

4) How much different will my post-processing workflow be when compared to the DSLR? I am using SiriL to stack and do the base processing, then Affinity Photo for the 'beautification' steps.

Many thanks!
V

[oska]

I'll try :)

1) I've found 30s to be fairly universal although the bright targets like Orion or Ol' Horsey 10-15s.

2) All calibration is taken at the same gain, offset and temperature as your lights. Bias - short as possible but 0.001s should work. Flats - 1-3s

3) Calibration. I think the flats do it specifically.

4) Not a lot if at all, DSLR is osc technically(ish) ;)

I use -10C, the lower the better noise reduction BUT it can cause dew on the sensor glass. I tried at -30 and -25 and had troubles. After a bit of research it seems below -10 the noise reduction returns are diminishing.

Flats can be taken with a white tshirt and pointed at the dawn sky: "sky flats". Or a tablet with a white screen, any browser/pic viewer that works for you will do. Or a led tracing panel, you will need a few white a4's to mute the brightness enough probably.

Welcome to the cool club :cheers:

[Bodon]

GDay


I have an ED 72, SW flattener and the 183 colour as well. I am not a guru as I have only just set this up as well. My pictures turn out fine. On another forum the gurus recommended gain 111 for this camera... So I use Gain 111 @ 60 sec -10 - with no amp glow. I also get nice pics at no gain @120 sec again no amp glow. I have experimented at gain @111 and 300 sec but when pics are stacked amp glow is on the right side of the pic and easily sorted with some dark frames. Im not guru and only used it minimally due to rain. When I experimented with darks they were patchy in color....worked that out pretty quick, focuser lets in light. Easily fixed with a black sock and velcro which still allows me to focus. Velcro is handy when I coil my USB cables to the focuser. I use the ASIAIR for darks and just make sure that temp and gain are the same. Focus is the same as it sort of set (I do check with a mask though). I haven't used flats yet as my pics seem ok to me. I have tried flats with a led tracing panel and they are far from an even colour. Not a pixel hunter.

[Drac0]

Hi Steve, are you using a UV/IR filter? The 183 needs one.

1. Exposure length depends a lot on the target - something bright like Orion will only want short exposures - I do between 10 & 90 secs, blending to bring the core out. But with a dimmer target 5 min or longer is quite doable.


Now, the other settings.
Gain: Tricky question, maybe someone with 183 can answer better. But unlike newer cameras, the 183 doesn't have a HCG mode where a specific setting causes a large drop in read noise & a return on things like DR & full well. For you, using 0 gain won't be much different to using 120 gain except in brightness. You will get a little more noise at 0 gain, but by far MOST noise is from light pollution!
Temp: I use 0° for a couple of reasons - (a) Diminishing returns as you push the temp lower, working the cooling harder for not much advantage. (b) It is much easier to reach 0° throughout the year, allowing consistent usage. This is important to me, especially in summer where night temps can be over 30°. (c) Cooling doesn't need to work so hard, especially over long sessions. (d) Thermal/read/sensor noise is really nothing compared to the noise caused by light pollution. Dr Robin Glover from SharpCap goes through this in detail in a lecture that I can link if you like.


2. Bias: I don't use bias with ZWO cameras & even they say DON'T use bias frames.
Flats: These are really individual things, depending on the method you use to obtain them. I use a white cloth + tablet or tracing pad. With APT, it has a tool for calculating the flats, but you can use the histogram to get decent flats - you want an unstreched flat to peak between about 1/3 to 1/2 from the left. You may also want to do dark flats, which are exactly the same settings as flats but with the lens covered.

Darks: Good thing with cooled cameras you can do these at any time. You just need to make sure all the camera settings are the same. I take my camera off, cap it & place it inside a box & let it do it thing.


3. Vignetting: This is strange. With the sensor size of the 183 you shouldn't be getting vignetting even if using 1.25" connections. All I can suggest is going through the setup & ensuring everything is connected properly & there is nothing that could be blocking the light path. Weird.


4. You're post processing will change very little really. As John says, your 183 & DSLR are really both OSC cameras.

Hope this helps, cheers,
Mark

[AstroViking]

Evening all,

Thanks very much for the helpful replies - much to ponder and chew over there. I also found a number of threads on another forum to digest as well.

To answer Mark's question, there are no filters present. I suspect I will need to invest in some, as well as a rotator. The camera is about 90 degrees off vertical when fully threaded onto the scope. :shrug: I tried packing it with thin circles of card, but that pushed me out to ~56mm of backfocus.

The attached image is a heavily stretched (but with no flats) output from Sunday night's work. It's supposed to be NGC290 and a good chunk of the SMC... I have a feeling it's also a bit out of focus - well, it was the first time the EAF got used, so with experience it will get better.

I spent time today taking flats and darks to match yesterday's settings. Hopefully I'll be able to get a decent image of 47 Tuc out of it all.

Interestingly enough, the flats showed heavy vignetting when I turned on the 'auto-stretch' function in Ekos' FITS viewer window. But looking through the telescope gave me no idea as to why it's vignetted so badly.

Now to stack all those images of 47Tuc and see what comes out the other end.

Cheers,
V

[oska]

Give it a quick re-stack with the cal data and see :)

Can you get ekos to do the focusing for you? Mathematically at least, better than my eyes can do (except by fluke)

[AstroViking]

I just did my 47Tuc data and the end result from Siril was very, very green. Trying the inbuilt 'green removal' turned it totally trippy and psychedelic. A lot of work in Affinity got me an outcome that I'd call acceptable. I'll post it later tonight in a proper location.

I'll have another go at the NGC290 data now and see what I come up with. I suspect that I totally over-stretched it the first time around and that really didn't work...

[oska]

Maybe try an unlinked stretch? I don't know siril so just a guess.

[AstroViking]

Siril has quite the learning 'cliff' to get over. I know enough to be dangerous and eventually drag out an image I can work with.

The asinh transformations require some care, and the histogram transformation even more so.



However, this hasn't sorted out why I'm getting such serious vignetting... I'd love to get to the bottom of that. It wouldn't be caused by the flattener, would it? Now I need to go back and check my DSLR images and see if they have any vignetting present.


BTW - I did use Ekos' focus module. It took some getting used to, and I still have to spend time to get the routine sorted out. But it's cool to watch the image sharpen up as the software does it's thing. And frustrating to watch it get confused and wander off into the weeds. :screwy:

[oska]

I can't say for sure but afaik flattener issues result in elongated stars in the corners and yours seem good.

If you get even wispy clouds during a focus run or have a travel issue either cable snag related or run out of travel then it can all go to pot. A rerun can sort it if it's clouds.

[oska]

Are you using whatever the equivalent of "Star HFR method" and "Hyperbolic curve fitting strategy"?

[Bodon]

Can a light leak give you that effect in the pic you posted? Like I mentioned my focusor did on my ED72.

[AstroViking]

Hi all,



Ekos uses HFR (Half Flux Radius) and offers multiple ways of getting there - Polynomial, Iterative, and at least 2 others. The documentation on indilib.org seems to be out of date on this.

I'll have a look at what I'm using when I get a chance, but it worked pretty well on Monday night when I was imaging 47Tuc. The stars in the final image look good - perhaps a little blurred - which some tuning of Ekos should resolve.



Maybe. Although I'm not certain it's a light leak - there is only about 5mm of focuser tube exposed and all the fittings are screwed together. (M54-M54 adapter / SW flattener / M48-M42 spacer / M42 spacer / M42 spacer / camera)

I suspect that what I'm seeing as vignetting is an artifact of a hyper-aggressive asinh-stretch operation. Why do I say this? I went back and restacked the NGC290 data with flats, darks, bias and lights and got out a much better looking image after a gentle asinh-stretch.

Similarly, the 47Tuc image stack came out quite clean using the same method.

Cheers,
V

[Camelopardalis]

I’m going to go out on a limb and suggest that it’s light pollution.

Also, 30 seconds is likely too short. What’s the median value of one of your (uncalibrated) light frames and one of your dark frames?

[AstroViking]

Hi Dunk,

You might be right with the light pollution. It can be bad here, depending on where I'm aiming.

The image I posted above was massively over-stretched, so probably a very poor example to work with. I think we can ignore it for now.

I'm going to do more research on what the 'best' camera settings are and work from there.

Thanks for all your help, folks.

[AstroViking]

Hi all,

I've been doing a bit more reading and researching. The values below are the ones that crop up the most.

Gain: between 100 and 120 (Unity gain is 120)
Offset: default is supposedly 40, but 10 is used a lot
Temperature: -10 seems to be the most common.

I've seen people reporting that a gain of 200 works fine. The cost to dynamic range would be pretty high, so I can't see the point of going that high when imaging DSOs.

Offset appears to be related to the black level and should not be zero.

Temperature is a personal choice based on the conditions of the night.

Now we get to the strange pair of red and blue white-balance.

WB_R: Default = 52
WB_B: Default = 95

I've seen a couple of recommendations to make them both 50. What would you gain by messing with the white balance so much?

There is a great post (https://www.iceinspace.com.au/forum/showthread.php?t=166567) by Shiraz with a series of tables of gain / scope f-ratio / Bortle score that seem like an excellent place to start.

Cheers,
V

[Drac0]

For me it's not the cost of the DR, it's that you lose so much of your already restrictive full well capacity which in turn restricts your exposure time. At 120 gain, you lose over 2/3rds, dropping to ~4k.

Never moved mine off the default.

This is true. The 183 seems to have better cooling than the 294 so perhaps you won't have a problem getting down to -10 in summer. Just remember if you change the temp (or gain, exposure length, etc) you will need a new set of darks.

As I said previously, I'm not overly concerned with thermal noise as by far most noise will come from light pollution. Unless you get away from light pollution to good dark skies, worrying about pushing the temps down too much isn't needed if it's only a small percentage of the overall noise.
Here's the lecture by Robin Glover discussing DSO imaging with a CMOS camera - you might find it interesting. https://www.youtube.com/watch?v=3RH93UvP358

Can't help with the white balance issue, it's mono for me. :D

Hope you're enjoying your learning experience. Cheers,
Mark

[Camelopardalis]

There are practical considerations at work too. If you dial up the gain, your exposure time will be reduced and you’ll end up with a larger number of output files. This increases the amount of time, and storage required, to process through the calibration/registration/integration workflow.

Ray’s posts on exposure times for CMOS cameras should be a must-read. The key take home message is to expose for the suggested value above the bias/dark frames. This will ensure there is enough signal in each frame, and allows you to methodically judge your exposure time at a given gain very easily.

In typical urban/suburban environments, the light pollution is by far the largest source of noise. Even at low gain, the read noise is not large, and neither is the thermal noise.

[AstroViking]

Hey everyone,

Just a quick update on where I am with the hunt for usable settings with the '183MC-Pro that I am using.

Settings that work for me:
Gain: 50
Offset: 10
Temp: 0 degrees C

You MUST also use calibration frames in addition to your light frames. This means darks, flats and dark-flats. (The dark-flat do the same job as bias frames, so don't panic about not having them.)

If you use a low gain (under 50) then you are highly likely to get some serious colour banding in your stacked image. I used a gain of 25 for a while and suffered through this.

Note that I am getting saturation in the blue channel for some reason. I suspect this is light pollution from the LED street lights. Once I get my dual-band filter installed and working, I will add more info to this thread.

Cheers,
V

[Crater101]

Do keep the information coming young man, you're pushing me along as you go! :prey2:

[AstroViking]

Hi again,

Another interesting 'quirk' of the '183MC-Pro is the default (in camera) white balance settings.

There are 2 settings - WB_R and WB_B, which default to 52 and 95, respectively. WB_R is the RED balance, whilst WB_B is the BLUE balance.

With the default values in place, can anyone guess why all my images had a massive blue tint to them? :P

I've since dropped the WB_B value to 50, and the images look a lot, lot better. I saw a forum post that gave some reasonable values for WB_R and WB_B which I will post below.

Of course, your software may be over-riding this setting, so if you are not getting horrible blue images, then you can ignore this update!

Cheers,
V

[Drac0]

Interesting. So with OSC either change the settings to bring it back to about a 1:1:1 RGB ratio, or use the ASCOM driver. Will have to remember that if I get an OSC, which is still a possibility later.

Thanks Steve for the info.

[Crater101]

You know, after years of owning digital cameras and getting used to their varying oddities, it never occurred to me to check the white balance settings on the 183MC. :confuse3: Thank You for the jog to my obviously useless brain. At this rate I'll be able to post something that doesn't look like something from a kindergarten painting class shortly!

[AstroViking]

Mark and Warren - you're very welcome! I'm glad I could help in some small way.

[ZL4PLM]

You should def watch this

https://youtu.be/nl_LQDmkg2o

the comments here are relevant for the 183 as much as the 1600/183/294 etc

anything with amp glow this process will help with

thanks

Simon

[SB]

Hi Simon,
Thanks for drawing attention to this video. Very helpful.

Chris

[ZL4PLM]

no problem- hope it helps

thanks
Simon

[Drac0]

Good video & the CN post is an interesting read - especially interesting to me was the idea of putting a short exposure between long exposures when doing darks. Got me wondering now whether that would also make a difference with normal imaging, not just darks, the write up seems to suggest it does...

Cheers,
Mark

[AstroViking]

Hi Simon,

Thanks very much for the link to the video. Aside from the "keep your exposures over 3 seconds", most of your tips/suggestions I am already doing. Moving from bias to dark-flats was a major improvement.

I just need to get a proper light source for taking the flats now.

Cheers,
V

[ZL4PLM]

The flat panel is a must - It's been one of the most used items in my kit - I use it so often for my own gear and testing others. It's a def need to have IMHO