ASI294MM Dark frames quality measurement

[barbulo]

Hi all, 

I am building the darks library for my brand new 294MM and, as I don't think I'll have the chance to take any lights due to the preceptive cloudy weeks that follow the purchase (you know), I am wondering if there is any way to know the quality of the dark frames. I don't know if I have to measure the ADU, Read noise... All those concepts still seem a bit blurry to me.

Any clue would be much appreciated.

I've attached a single frame and the Master stacked with SiriL using the average method with Winsorized Sigma Clipping rejection, in case you want to have a look

Master Dark_G120_300s__-20C-Enero 2022.fitD_Library_20-12-17_G120_300s__-20C_0506.fit

Edited February 1, 2022 by barbulo

[AstroMuni]

Thats quite a few amp glow areas on the Master. Generally I have seen only one and its fuzzy not having starry lines. Perhaps someone with more experience of your camera can comment

[barbulo]

Tank you @AstroMuni,

Yes, this sensor suffers from serious amp glow. That’s why I believe good dark frames are so important. But I don’t know how to measure their quality before calibrating the lights. Hence this topic. 

[geeklee]

1 hour ago, barbulo said:

But I don’t know how to measure their quality before calibrating the lights. Hence this topic. 

If you've temperature, gain and offset matched and they don't show any abnormalities then they should be fine - just try calibrating a Light sub with them.  You can always retake darks.

I had a look but unfortunately I can't directly compare to mine as you're using offset 8 and I'm using 30.  Your median values look inline with the offset.  Your starburst certainly looks like mine   Just make sure your lights are offset 8 as well, if you're using that.

Hope this helps.

Edited February 1, 2022 by geeklee

[barbulo]

58 minutes ago, geeklee said:

I had a look but unfortunately I can't directly compare to mine as you're using offset 8

I suspected that. May I ask how did you find out? I took another set but I'm not sure if it is at offset 8 or 30.

[geeklee]

12 minutes ago, barbulo said:

May I ask how did you find out? I took another set but I'm not sure if it is at offset 8 or 30.

I had a look at the FITS header for the individual dark frame.  Example (I've removed some of the FITS info like your latitude and longitude)

I use Apt for my calibration frames and use the ASCOM driver - the defaults for unity gain (120 - which you're also using) give me 30 as the offset.  Perhaps the native ZWO driver is giving you 8.

Edited February 1, 2022 by geeklee

[barbulo]

Thank you @geeklee

The second set I took has no OFFSET in the FITS header, so I am not sure if I took the frames with Offset=30 as I intended. I'll take another set tomorrow. 

2 hours ago, geeklee said:

just try calibrating a Light sub with them

First clear night. Some day. Maybe in a few weeks. 😢

[geeklee]

21 minutes ago, barbulo said:

The second set I took has no OFFSET in the FITS header, so I am not sure if I took the frames with Offset=30 as I intended. I'll take another set tomorrow.

Using Statistics (16-bit [0,65535]) in PixInsight or the Pixel Aid tool in Apt will give you an idea (For the Pixel Aid tool, ensure it shows "FITS Mode".  You get this by double clicking the image - e.g. dark frame - on screen so it zooms in).  ~1920 should be offset 30, ~512 should be offset 8.

The Pixel Aid below is on a shortish Dark Flat sub as it's all I had on that machine that Apt could read.  My 300s Master Dark in PixInsight shows similar numbers for average/median.  

Apt:   PI: 

Your file earlier was around 512 (offset 8).

 

Edited February 1, 2022 by geeklee

[wimvb]

You used an exposure time of 300 s at gain 120. For narrowband that would be ok, but for LRGB I think you won't need this much. I image with 4 minutes subs at gain 0 and offset 8 (RGB) or 120 - 180 s for L. Narrowband: gain 200 offset 30, 4 minutes exposures.

Make sure that your offset is such that a single sub has a minimum value different from 0, but not much higher. Decrease offset if the minimum value is high, increase if minimum = 0.

[The Lazy Astronomer]

11 minutes ago, wimvb said:

You used an exposure time of 300 s at gain 120. For narrowband that would be ok, but for LRGB I think you won't need this much. I image with 4 minutes subs at gain 0 and offset 8 (RGB) or 120 - 180 s for L. Narrowband: gain 200 offset 30, 4 minutes exposures.

Make sure that your offset is such that a single sub has a minimum value different from 0, but not much higher. Decrease offset if the minimum value is high, increase if minimum = 0.

Just curious: what made you settle on gain 0 vs unity (or higher) for rgb only?

[wimvb]

27 minutes ago, The Lazy Astronomer said:

Just curious: what made you settle on gain 0 vs unity (or higher) for rgb only?

There were reports a while ago that the point where the camera switches from low conversion gain to high conversion gain could show some instability, and that a gain value close to this point better be avoided. Those reports were never really confirmed. But more importantly, at 0 gain the dynamic range (full well / read noise) is largest, so you can have weak signal and still avoid blown out stars. At low gain, the camera works more like a classical ccd camera. Instead of doing a lot of short exposures to increase dynamic range in the final image, you can do fewer but longer exposures. Mind you, cmos still works better with many and shorter exposures than CCD. I find that low gain works for me.

[barbulo]

Thank you @geeklee @wimvb and @The Lazy Astronomer

I've just created a new library at G120 and Offset 30 for 30, 60, 90, 120, 180, 240 and 300 seconds.

I'll initially stick to gain 120. When I have plenty of clear nights (LOL) I'll experiment with other values.

[AstroMuni]

On 01/02/2022 at 17:33, barbulo said:

Yes, this sensor suffers from serious amp glow.

I have searched on the internet for examples of amp glow from ASI294 and all show just one spot from which amp glow is emanating. So worth checking that the rest are from camera or stray light.

[barbulo]

40 minutes ago, AstroMuni said:

So worth checking that the rest are from camera or stray light.

Thank you @AstroMuni, but those dark frames were taken with the camera in the fridge and the cap on, so I don’t think there were any light leaks.

[scotty38]

It’s normal  for the 294mc to have some glow bottom right and on the left hand side along with the usual top right. Not 100% sure if the mm is the same but I think it is. 

[barbulo]

I've run all the Dark frames through the PI Batch Statistics tool and put some of the measured parameters into graphs, just to see if the data is coherent. I've chosen those parameters that showed significant change across the sample. And showed like this:

Sample comprises 50x subs per exposure time sorted from lower to higher (30s, 60s... 300s). Values are shown without normalization and unclipped.

First and last graphs show some discontinuities because some frames were re-taken the day after (10 of 30s, 15 of 120s, 10 of 180s and 20 of 240s).

However, my concern is the value of the maximum: 65532 for all the 300s subs and some 240s. Does it mean there are some saturated pixels? Will that affect the calibration effectiveness? I can't wait to test them with some lights, but the weather's being awful.

Thanks in advance.

Edit: attached a 300s frame

D_Library_2022-02-02_22-42-35_G120_300s__-20C_OFS30_Bin2x2_1617.fit

Edited February 3, 2022 by barbulo