ASI Air flat frame issues

[Todo43]

Hi all,

Another problem to do with the ASI Air I think. I have finally managed to get guiding going and I now understand the problem (it's a whole thing). So last night, I took my flats after my imaging session, after watching the focus positions (ZWO really needs to make it an option so that you can see your previous focus points). I averaged the positions, took the exposures, came inside, pre-processed the images, and got a result. It is attached. I've only made the background better and stretched it so that you can all see. I've also attached an image of a single flat frame that was produced from the ASI Air.

I'm wondering if these flats are over compensating? Or under compensating. I don't know how to get rid of the weird vignetting thing at all and I don't think it has happened to me before.

I'm using a ASI533MC Pro with a 2.5" Flattener on a Sharpstar AL-90R. I don't think that it is normal vingetting because the sensor is way smaller than the image field.

Any ideas anyone? Thanks!

[RyanJones]

Hi Lachlan,

It does look like your flats don’t match and have over compensated as you have said. I normally don’t touch my focuser after my imaging session before taking flats. There’s just to much risk of something moving or introducing human error. I simply use the last position of the focuser for my flats. In fairness I also take my flats the next day so due to expansion they’re never going to be absolutely perfect. I don’t see any issues when I do it this way though either. I use the ASIair too.

[Todo43]

Hi Ryan,

I didn't touch the focuser at all. Went outside, took my flat panel, placed it on, took the flats, came inside. Is it worth trying again and taking the flats from the original focuser position? I have an EAF so I can move it easier

[Nikolas]

Flats should be at 25-30000 adu and 2 second exposures then dark flats at the same exposure is the recommended advice for these and the 294mc that is the sweet spot for flats

[Zuts]

You can select auto flats with the ASIAir and it will automatically control the exposure to hit the mid ADU range.

Also you said you could reset the EAF to the original position, but also said you never touched the focuser? The EAF/focuser/image train should not be altered from what you had when you took the lights.

Cheers
Paul

[RyanJones]

That is what I read too paul, that the positions had been averaged and set at a point to take the flats.

I also use the auto flats option on the air for the exposure.

[Todo43]

So what I meant by the original focus position was the first auto focus position. After an hour, it ran auto focus again and changed position.

Lachlan

[RyanJones]

So you did move the focuser ? Back to its first auto focus position ? I’m sorry if it seems like I/we are going over the same thing but it’s this detail that, at least I think, could be the source of the issue. Is it possible that you made a mistake in recording the original focuser position ? Another thought that crossed my mind ( that also requires moving the focuser ) is your scope uses a rack and pinion focuser. Is there any backlash what so ever in the rack ? And I mean ANY ? It would take so little to not to reliably return back to the exact position despite what your AIR is reporting on the EAF. To prove this you could move it and move it back to where you think you had it and see if the flat changes again ?

[The_bluester]

So long as the focuser is not moved by much it should not impact flats unduly. In general you should not need to worry about averaging the position of the focuser over several focus runs, I just use the position from the last focus run of the night.

What other calibration frames are you using? I find with my ASI2600 (Either of them, I have a colour and an OSC one) that lights shot at low gain can be pretty nicely calibrated with a master bias and a master flat (Which itself is normally calibrated with either a master bias or master dark-flat) but if I shoot at high gain, the corners are over corrected unless I calibrate the lights with a master flat and a master dark for the relevant sub length and temperature. I had been being lazy as I have recently found myself shooting a sub length at high gain that I did not shoot darks for.

What I am fairly sure is happening is that the high gain setting, while it reduces the READ noise of the sensor to a lower level than low gain, effectively increases the impact of dark current. The low gain is about 0.8e/ADU and gain 100 is just over 0.2e/ADU so nearly four times the gain.

At a dark current that equates to about 0.2e in a 300 second exposure that comes out at about 1.5ADU, which you would not recon would make a difference in even a stretched image, but the images I did not have darks for (That I mentioned above) had over corrected corners, I shot matching darks overnight last night and applied them to one of the problem integrations this morning, perfect!

[ChrisV]

I'm wondering similar to Nik.

The single flat image looks very noisy.
So I'm assuming the flats were taken at a light level that is too low??

Chris

[Todo43]

Hi Chris,

I just used the ASI Air flat plan auto exposure. Should I use a different exposure?

[Nikolas]

As stated earlier Average ADU between 25-30000 2-3 second exposures and dark flats same exposure as the flats this is the magic formula for these zwo one shot cameras.

[The_bluester]

I don't know about the 533 as I have never used one but that is not the case with my pigeon pair of 2600MM and 2600MC. I shoot dawn sky flats and aim for 32K ADU (With a 15% tolerance to avoid it chasing exposure times and repeating the same flat) and the exposure times vary from around .2 seconds to 30 seconds or more depending on how close to sunrise it is and the filter being shot. I calibrate the flats with a master bias.

The biggest impacts I have found on flats is that a 32 bit master bias is better, and the lights need to also be calibrated with a master dark matching the exposure and temp, preferably a 32 bit one. I recently completed some unintentional testing where I was having issues with over corrected corners and in terms of quality of flat field correction it was (From worst to best) light frames calibrated by:

*Master flat (32 bit) which was calibrated by an older 16 bit master bias. Lights calibrated by the master flat and by the same master bias. Very visible brightening in the corners that required a significant crop to avoid.

*Master flat (32 bit) calibrated by a 32 bit master bias, lights calibrated by the master flat and a 16 bit master darks. A big improvement in the corners.

*Master flat (32 bit) calibrated by a 32 bit master bias, lights calibrated by the master flat and 32 bit master darks. A smaller improvement in the corners compared to the one above, good enough that I only cropped out the area where the effect of dithering was visible.

All of the above used the same flats and lights.

[Startrek]

My Calibration process ......
I use the ZWO 2600MC and through extensive testing last year in my Newts found the following Calibration works best for an even flat field
Capture software is APT using Flats Aid Tool
Flats ADU range 8500 to 32000
Without filters exposures usually 1 to 2 sec
With L Extreme filter usually 5 to 15 sec
Generally capture 40 Flats
Flats are calibrated with Bias ( exposure time 0.000032 sec ) I usually capture 50 to 100
Dark flats don’t work
Darks don’t work
Lights , Flats and Bias stacked in DSS
I don’t use Master Bias , I find new Bias frames for each session work better in DSS ( they take no time to capture just a minute or so )