QSI690 Ha flats issue

[strutsinaction]

Hi

Here's a master flat I prepared (from 20 subs) with my QSI690 + Astrodon 5nm Ha filter. Does this look ok to you? I'm a bit concerned by the diagonal banding. Is this typical of narrowband flats? I see something similar with my OIII flats but SII look fine. Also, my Baader LRGB flats are fine.

I used the SGPro Flats Calibration Wizard to determine the exposure setting to give a mean ADU of 30000, a Gerd Neumann EL panel for the illumination and PI for calibration.

Maybe the ADU value is too high?

Thanks for any advice!

Regards

John

[Martin-Devon]

Yes I think these are fine, I also see some degree of banding as well which doesn't alter if I rotate the GN light screen, so it's in the imaging train. I have the 3nM Astrodon's and also see the 'edge' effects like you have as well. If you 2xbin, these edge effects are even more pronounced. Typically I use an ADU of 23,000, but I've seen recently on the QSI forum that they recommend somewhat higher values. 

Martin

[swag72]

Here's a couple of my thoughts.

1) When you calibrate using these flats and your lights can you still see any banding if you push the processing?

2) Have you tried different gain settings and did it make a difference?

[strutsinaction]

Yes I think these are fine, I also see some degree of banding as well which doesn't alter if I rotate the GN light screen, so it's in the imaging train. I have the 3nM Astrodon's and also see the 'edge' effects like you have as well. If you 2xbin, these edge effects are even more pronounced. Typically I use an ADU of 23,000, but I've seen recently on the QSI forum that they recommend somewhat higher values. 

Martin

Hi Martin

Interesting that you have also seen the banding. Any thoughts on what might be causing this effect? Could it be some form of internal reflection? I'd be interested to see an example of one of your Ha flats for comparison if you don't mind posting to this thread.

I'll try a lower ADU and see what effect that has.

Regards

John

[strutsinaction]

Here's a couple of my thoughts.

1) When you calibrate using these flats and your lights can you still see any banding if you push the processing?

2) Have you tried different gain settings and did it make a difference?

Hi Sara

In answer to your questions:

1) No, the calibrated lights (flats + bias) look ok, no hint of banding, although so far I've only processed one NB target (my first, IC443). I'll push the processing a bit further and see if I can make the banding appear.

2) Regarding gain, are you referring to the gain setting in SGPro for various binning settings? This is currently set to 'High' for 1x1 and 'Low' for all other binning cases. I'll try setting this to 'Low' and see if that makes a difference.

I'd be interested to see an example of one of your Ha flats if you don't mind me asking. I can then compare with mine and Martin's (asuming he posts one of his!)

Regards

John

[swag72]

If the banding is not evident in your calibrated images then in my opinion the flats are doing their job and the banding is of little relevance - Or maybe I'm just being too simple with my assertion.

Yes I am referring to the gain setting in SGP - Try something different and see if it makes a differnce ............ but firstly see above........... No issue, no worries

[strutsinaction]

If the banding is not evident in your calibrated images then in my opinion the flats are doing their job and the banding is of little relevance - Or maybe I'm just being too simple with my assertion.

Yes I am referring to the gain setting in SGP - Try something different and see if it makes a differnce ............ but firstly see above........... No issue, no worries

Yes, maybe I like looking for problems that aren't really there 

But, now that I've seen the weird flats, I'd like to understand why they are affecting Ha and SII and not OII, L, R, G and B.

I've opened a ticket with QSI to see if they can shed some light on this (no pun intended).

I'll post my IC433 image later tonight to the 'Imaging - Deep Space' forum. See if you can spot any issues caused by the flats.

Regards

John

[strutsinaction]

I've just posted my calibrated IC443:

http://stargazerslounge.com/topic/237647-ic443-ha-my-first-nb/

No obvious sign of banding I'm pleased to say. I still want to know what causes the banding in the flats though!

Regards

John

[andrewluck]

Hi John

I came upon this thread while researching banding in my flats. From further investigation it appears that these bands are the result of a slight mis-registration of the microlenses on the sensor, resulting in a moire pattern.

Cheers

Andrew

[andrewluck]

I was incorrect in my earlier assumption. An article by SBIG http://www.sbig.com/about-us/blog/flat-fields-the-ugly-truth/ identifies these bands as bond wire glints. This is where light is reflecting off the tiny gold wires that connect the sensor to the package pins. This light then reflects off the cover glass and onto the image.

If you haven't previously seen this article then I recommend a thorough review. It's an eye opener!

Andrew

[ollypenrice]

...but if the flats work, as Sara says, they work. It's a hell of a lot less fun when they don't work! (One of the cameras we have here is a flats nightmare and always has been. I'm constantly struggling to find ways of making accurate flats with it.

Olly

[andrewluck]

The particular issue I have is that the banding is slightly different on the flats and deep sky images. It looks like the solution is to match the flat light source with the local light pollution spectrum. Not a trivial undertaking!

Andrew

[mworion]

Hi John,

Is there any feedback from QSI? I'm using a 690 too an facing the same topic when making flats. LRGB so far ok, but you can see the banding if you look closely. On Ha, OIII, SII I can recognize it easily,

I though on first hand it's a problem of the filters. But even without any optics the camera show that pictures.

Michel

[ChrisLX200]

I've seen it with two entirely different cameras (Atik490 and Moravian G4), and with different NB filters (Baader and Astrodon on the 490EX and Chroma on the Moravian). Don't get it at all with the LRGB filters of any make. I figure it's something perculiar with using NB filters but it is reproducible and flats cancel it out, so some optical effect.

ChrisH

[strutsinaction]

Hi Michael

Nothing conclusive came from Ian King or QSI. I'm inclined to believe Andrew's plausible explanation (see above). In any event, the banding doesn't show up in my calibrated light frames so I'm not overly concerned. I was more concerned about the 'Sony glow' issue which showed up in long exposure NB subframes. I returned the camera to QSI in the US for a fix.

Regards

John

[mworion]

Hi John,

thanks for your response. I also don't see some issues with the calibrated lights. Due to the fact that i own a pretty new produced came (Nov 2015) Sony glow should not be an issue. So I go along with calibration.

Last question: I still using a lancerta lightbox for the flats and try to get to about 30.000 ADU. How are you doing your flats ?

Regards, Michel

[zicklurky]

I get a very similar pattern with my Atik One 9.0 which is also the Sony 814. With a Gerd Neuman flat panel.

I can't see any banding in the calibrated subs so don't think it's anything to worry about.

[strutsinaction]

Last question: I still using a lancerta lightbox for the flats and try to get to about 30.000 ADU. How are you doing your flats ?

Regards, Michel

Hi Michel

I use a Gerd Neumann flat panel and also aim for a mean value of 30k ADU.

Regards

John

[strutsinaction]

On 13/12/2015 at 21:48, andrewluck said:

I was incorrect in my earlier assumption. An article by SBIG http://www.sbig.com/about-us/blog/flat-fields-the-ugly-truth/ identifies these bands as bond wire glints. This is where light is reflecting off the tiny gold wires that connect the sensor to the package pins. This light then reflects off the cover glass and onto the image.

If you haven't previously seen this article then I recommend a thorough review. It's an eye opener!

Andrew

Article now available here:

http://diffractionlimited.com/flat-fields-stray-light-amateur-telescopes/

Regards
John

[gnomus]

Hi John

My QSI 690 flats when stretched using PI's STF look very similar to yours.  My subs seem to calibrate well and I can see no issue whatsoever in the final images.  Of course, the flats are shot at around the 25,000-30,000 ADU mark and the final histogram I get looks pretty 'tight'.  STF (or SGP's autostretch) will widen this histogram in an artificial way, exaggerating very minor differences in contrast.    My unstretched flat looks like a flat grey.  If these are bond wire glints then they are bond wire glints, but they are only calibration files.  All that I am interested in is the final image.  If nothing is showing up in that then I am not concerned.   

[JuergenB]

Hi John,

As many others already wrote here, these artifacts are not specific for QSI. They are related to the CCD chips themselves and are quite common for narrowband filter use. They are also not related to flat boxes or screens, as I can see them also in skyflats. I am using Astrodon filters (5 nm H-alpha, 3 nm [OIII] and 3 nm [SII]) and I get them with all three filters. However, the calibrated lights are fine, no traces of these stripes or border frames visible.

Juergen