Overcorrecting flats

[tooth_dr]

I've taken a few sets of flats now with this set up - QHY9 and Nikon 180mm lens, using SGPro flats wizard, to expose to 30000ADUs (middle of histogram).

All attempts to stack the data has resulted in over compensation.  I am using the same filter for my flats as the lights (Ha)

 

Any input appreciated.

1) calibrated 2) uncalibrated 3) single flat 4) master flat

 

[matt_baker]

Just wondering if the master flat has been calibrated with flat darks?

[tooth_dr]

4 hours ago, matt_baker said:

Just wondering if the master flat has been calibrated with flat darks?

I calibrated it with a master bias.  I could try with dark flats correctly exposed to the duration of the flat (it was around 12.33s)

[vlaiv]

This is for lens and not scope?

It does not look quite like over correction to me. Well - it does for vignetting but not for dust.

In fact, I can't seem to find dust on any of the two subs. Sub without flat correction does not seem to suffer from vignetting.

Is there any chance that you might have changed aperture settings on the lens between lights and flats (by accident or on purpose)?.

[tooth_dr]

14 hours ago, vlaiv said:

Is there any chance that you might have changed aperture settings on the lens between lights and flats (by accident or on purpose)?.

Hi Vlad. Yes I’m sure the aperture wasn’t adjusted. It’s done manually on the lens ring itself and I took flats just after lights without adjusting anything. 

14 hours ago, vlaiv said:

This is for lens and not scope?

Yes this a 180mm lens connected to an EFW then QHY9. 

[andrew s]

Assuming the images were correctly calibrated (dark, bias etc.) then if the flat does not correct correctly  it is probably due to one or more of: light leakage on darks,  mismatch in light path illumination between target and flat or something changed e.g. filter or focus.

Just to check the process and software apply the master flat to a single  flat calibrated with bark or bias to see if it works on that. 

Regards Andrew 

[matt_baker]

12 seconds seems a bit long for a flat frame. What bandpass is your ha filter?

[tooth_dr]

8 minutes ago, andrew s said:

Assuming the images were correctly calibrated (dark, bias etc.) then if the flat does not correct correctly  it is probably due to one or more of: light leakage on darks,  mismatch in light path illumination between target and flat or something changed e.g. filter or focus.

Just to check the process and software apply the master flat to a single  flat calibrated with bark or bias to see if it works on that. 

Regards Andrew 

Thanks Andrew, I'll try that calibration process later on the master flat itself

[andrew s]

1 minute ago, tooth_dr said:

Thanks Andrew, I'll try that calibration process later on the master flat itself

No, apply the master flat to a single flat frame.  Regards Andrew 

[tooth_dr]

Just now, matt_baker said:

12 seconds seems a bit long for a flat frame. What bandpass is your ha filter?

It's only 7nm, but I'm using a camera with a mechanical shutter, so my flats always have to be longer than 3.5s.  The panel had a couple of sheets of perspex on it for lum flats, so at it's brightest 12s was the length of exposured to give me a correctly exposed flat.

[tooth_dr]

2 minutes ago, andrew s said:

No, apply the master flat to a single flat frame.  Regards Andrew 

Sorry that's what I meant!  Typing to fast!

[tooth_dr]

26 minutes ago, andrew s said:

Assuming the images were correctly calibrated (dark, bias etc.) then if the flat does not correct correctly  it is probably due to one or more of: light leakage on darks,  mismatch in light path illumination between target and flat or something changed e.g. filter or focus.

Just to check the process and software apply the master flat to a single  flat calibrated with bark or bias to see if it works on that. 

Regards Andrew 

Andrew I just took a few minutes to do this now.  Doesnt the second one look like my image - maybe there is an issue with the bias calibration?

1) Single Flat

2) Single Flat with MF

3) Single Flat with MB and MF

 

 

[matt_baker]

Definitely looks like something to do with the master bias. Potential light leak?

[ollypenrice]

I suppose 12 seconds might be long enough to contain enough thermal noise to prevent the master bias from correctly calibrating the flats.

On one setup I was plagued by over-correcting flats. I tried everything I could think of including shorter and longer flat subs, bias as flat dark, dedicated flat darks, rigorous flushing between flats, etc etc etc. In the end I shot flats using a different capture program (AstroArt instead of Nebulosity) and, glory be, they worked (on lights captured in Nebulosity.) Well, they worked for a long time and then one fine morning they started over correcting again. Nightmare. I never did get any further with this so I'm not being much help. However, I'm pretty convinced that, in my case, the mismatch was arising  from within in the capture program. Is there a 'capture mode' or 'download mode/speed' in your capture software which might have varied. The other candidate was the flushing process. Have you tried leaving a longer gap between flats? I don't remember the details but there was some talk of getting rid of the RBI (residual bulk image) when I was fighting this problem.

I evolved a messy workaround in which I stacked with and without flats and averaged the two stacks.

Olly

[tooth_dr]

9 minutes ago, matt_baker said:

Definitely looks like something to do with the master bias. Potential light leak?

Possible. I do take all my calibration data at night to avoid this, and at the shortest exposure for bias light should not be an issue.  If above, I can show that the single flat is calibrating correctly, then perhaps I have an issue with my lights?

[andrew s]

3) looks fine. I would expect 2) to be wrong if the bias has not been subtracted .

Check the back ground in your image before and after bias subtraction just to see if it was done correctly. 

Not sure what else to suggest but you original image does not obviously show the vignetting your flat does. Can you check this ?

Final thoughts, I assume they were taken at the same temperature?  Is there a reason you don't use darks and flat darks.

.

Regards Andrew 

[tooth_dr]

1 hour ago, andrew s said:

3) looks fine. I would expect 2) to be wrong if the bias has not been subtracted .

Check the back ground in your image before and after bias subtraction just to see if it was done correctly. 

Not sure what else to suggest but you original image does not obviously show the vignetting your flat does. Can you check this ?

Final thoughts, I assume they were taken at the same temperature?  Is there a reason you don't use darks and flat darks.

.

Regards Andrew 

Andrew, some more images:

1) single light

2) single light with MB calibration

3) single light with MB + MF calibration

4) single light with MB + MDF* + MF calibration

5) single light with MB + MDF* + MF + MD calibration

6) BPM

7) MD

*MDF is a MB

I do normally use matching exposure time dark flats but read somewhere that bias is ok too.  I do use darks, 5) below is a single light calibrated with MB MF MDF MD

 

 

 

 

 

 

 

 

[tooth_dr]

All taken at same temperature of minus 15 deg C

[tooth_dr]

3 hours ago, ollypenrice said:

Is there a 'capture mode' or 'download mode/speed' in your capture software which might have varied

Hi Olly

I have the QHY9 set to slow download speed (as opposed to normal or fast), same for both lights and calibration frames.  I can take some 12s darks this evening to see if that helps, nothing to lose.

[andrew s]

Hi @tooth_dr sorry don't know what more to suggest.  Just looking isn't giving me any inspiration.  Sorry.

Regards Andrew 

[tooth_dr]

23 minutes ago, andrew s said:

Hi @tooth_dr sorry don't know what more to suggest.  Just looking isn't giving me any inspiration.  Sorry.

Regards Andrew 

No problem.

[wimvb]

Your master dark shows amp glow. What does a proper master flat dark look like?

On 09/11/2020 at 17:13, tooth_dr said:

Andrew, some more images

7) MD

 

[vlaiv]

10 hours ago, wimvb said:

Your master dark shows amp glow. What does a proper master flat dark look like?

I don't think that is amp glow - I think it is a light leak. @tooth_dr could you take your camera with lens detached and do set of darks with same settings (exposure and all) - but making sure there is no light leak - use aluminum foil and place camera face down on heavy desk or something.

In the end - compare new MD with above MD to see if that strange looking halo around the edges is gone. If yes - then MD / light leak is the problem.

[tooth_dr]

11 hours ago, wimvb said:

Your master dark shows amp glow. What does a proper master flat dark look like?

 

 

58 minutes ago, vlaiv said:

I don't think that is amp glow - I think it is a light leak. @tooth_dr could you take your camera with lens detached and do set of darks with same settings (exposure and all) - but making sure there is no light leak - use aluminum foil and place camera face down on heavy desk or something.

In the end - compare new MD with above MD to see if that strange looking halo around the edges is gone. If yes - then MD / light leak is the problem.

Thanks guys. I’ve been really busy with work as well as filming for a tv advert (!) but I’ll try exactly this.