Help ! I struggle with my flats!! - Forums

Topic starter

Jan 22, 2022

Dear Fellow Astronomers,

I am currently struggling with my flats... I use an RC8 scope, an ASI533MC camera (but I had the same issue with the ASI1600MM) and a generic A4 LED panel.

Initially, I thought the issue was with the automated flats using the ASIAIR Pro. So I tried to get some flats using the tool in APT.
When I use this APT tool with the same camera but different scope (Redcat 51) I don't have the same issue.
The automated tool will define the right exposure time. Of course the gain setup is the same on the flat and the image (Gain 100).

See below a streched version of my master flat.

When I apply this flat to an image taken from Orion, it will overexpose the flat, resulting in the dust appearing in white. You can weel see these dust marks on the picture below.
I use DSS to register and stack my images.

Any idea what I am doing wrong with these flats?

Thanks,
Axel

Well Written Helpful Engaging

D. Jung

Jan 22, 2022 · Edited: Jan 22, 2022

A few things:

1. You seem to have blemish in the flat that is not present in your light frames. This is not an issue per-se, but maybe something you want to correct.

2. Are you using dark flats? If not, that can easily explain the wrong flat correction. Also, are you using bias frames? I don't use any bias frames for my 533.

3. Check that your flat panel is actually flat. I often do flats in cloudy nights as if I was taking images to make sure I don't get any stray light effects in the flats

Helpful Concise

Doversole83

Topic starter

Jan 22, 2022

Thanks a lot for the suggestions.

You are right, there were no dark flats, I became too lazy. Problem is it didn't change anything...

So I tried outside, using a cloudy day as a source of light for the flats. I still add my usual white tee shirt.
the Masterflat looks about similar:

The result unfortunately looks very much the same:

For information, I used the ASIAIR this time to generate my flat images. I took 60 flats, 30ms each. Dark Flats don't seem to affect the result.

What could I try next?

Thanks!

D. Jung

Jan 22, 2022

I assume when you generate your flats you make sure that all parameters are exactly identical between your flat darks and flats:
- offset
- exposure
- gain
- camera cooling

And when you do your light calibration you match the dark and light frames in the same way?

Doversole83

Topic starter

Jan 22, 2022

Yes I used the exact same parameters.

Lynn K

Jan 22, 2022 · Edited: Jan 22, 2022

Did you De-Bayer the flats? It looks like the Bayer grid.

Disclaimer: I am mainly a mono narrow band imager.

Lynn K.

andrea tasselli

Frequent contributor

Jan 22, 2022

Try a different program, such as Siril or APP and see what happens.

Well Written

Brian Boyle

Jan 23, 2022 · Edited: Jan 23, 2022

I also own an RC8 and ZWO cameras (both OSC and Mono). I use both APP and PI. I dont take flat darks, which in my view just add noise (CMOS cameras now have such low dark current and FF exposures should be so short, that a bias frame alone should be sufficient).

I have never had a problem with flat fielding as long as I use twilight sky flats. Usually about 15-20 taken with the ZWO ASIAir “auto” method on a patch of sky 90degs from the sun.

Taking multiple flats and then stacking also gets rid of passing cloud, birds, seeds etc. i also try
to image when the sky is sufficiently bright to get to half well in under a sec. Around sunset for OSC/RGB and an hour before sunset for NB.

In contrast I could never get light panel LED flats to work (either as an amateur or professional astronomer). I always believed this was dome flats are not plane parellel rays. Unlike those from the sky.

CS

Brian

Helpful Insightful Engaging Supportive

Michel Makhlouta

Jan 23, 2022

This might be related to the design of the baffle tube, there is a video about it here: https://www.youtube.com/watch?v=nHuLrYe-oik

I am not sure if this is your problem or not, but worth checking. I bought my RC8 late last year, and it comes with an baffle tube extension installed (I remove it for primary collimation):

I also don't know if this was a fix added later on after the video brought the issue up, but my flats seems to be working.

Adrian Criss

Jan 23, 2022 · Edited: Jan 23, 2022

if you Check the dark VALUE of your image in PI and the VALUE of your dark master, you will see big diference.
I have had the same problem, solution: renew the darks

Ian McIntyre

Jan 23, 2022

I have the same OTA and have only had a couple clear nights to work with it. I ran into the same issue using the flat technique you do with the light panel. The light does not fill the FOV evenly. I believe it's simply a matter of the camera sensor being more sensitive to the position of the light source in the panel as the diameter of the tube is so wide, thus the flat is brighter toward the center. (Never have this problem with the 300mm lens.

I like Brian's solution of using daylight. Can't believe I never thought of that. Though I would be concerned the mirrors/glass and sensor would respond to the temp increase after sun up and through calibration off.

The other thing that occurred to me just today, is lining a dew shield with foil. As long as the white cloth is placed over the end of the OTA and the light panel placed atop the foil lined dew shield, the light reflected off the foil should fill the field more evenly.

Or maybe not. I'm just spit ballin' here.

Helpful Insightful Respectful Engaging

Doversole83

Topic starter

Jan 23, 2022

Thanks a lot to all of you, lots of great suggestions.

I tried a few things more:
- I took a series of flats outside, yesterday was overcast. Using the automatic flats with the ASIAIR.
- I took a series of flats and dark flats inside, using the LED pannel and making sure the average ADU wasn't too high. I tried 22500 ADU and even 19500 ADU. I also added some more T-shirts layers to increase the exposure time around 500ms.

The result is still the same using DSS with different settings. Below is a stretched version of M101 I took last Thursday. You can clearly see the "inverse" vignetting and some dust. Frustrating.

@D.Jung kindly reprocessed my images with PI and doesnt seem to see anything wrong.
So I guess I need to give it a try! I will report further here once I get the initial results with PI.

Thanks again to you all.

Respectful

Tim Hawkes

Jan 23, 2022 · Edited: Jan 23, 2022

Just to note that I had what looks like an identical issue a year or so ago. I thought that it was a flat problem (because the flat features were superimposed on the image) but it actually turned out to be to do with the darks, Simple resolution…

I was also using a ZWO camera — and was using Sharpcap to prepare the master flats and darks which works very well.

However a peculiarity of ZWO cameras is that in they do not use the term 'offset' but rather 'brightness' which appears under the image controls in SC.

When SC saves master darks for you it nicely sorts out the subfolder structure - exposure time, temperature, gain etc so that you can pull out the right dark to match to your lights. However - caveat emptor - there is one additional parameter that you have to match for yourself and this is 'brightness' . I was making the mistake of not understanding the huge significance of the brightness setting . It has to be spot on the same – a difference of 1 or 2 ruins the image— it is vital to add this parameter also to your dark library.

After fixing that the problem which appeared to be a flat problem (because the flat pattern appeared in the image) went away.

One other problem I had at one time with matching flats was to save master flats with a bias correction when I did not need to .
hope this helps
best wishes
Tim

Helpful Respectful Supportive

Doversole83

Topic starter

Jan 28, 2022

So a bit of update.

Here is M101 stretched processed with a flat series I took with the LED panel, average 19300 ADU (Gain 100, Temp -20C, 500ms). Dark Flats same parameters.

Now, same M101 using flats more exposed (1s, average ADU 33500). Difference is not huge. Overcorrection still clearly visible.

Now processed with no flats, you obviously see the reverse effect. You see some dust in black now.

When mixing the 2 using photoshop, I get something more reasonnable.

In summary:
- I tried with the LED panel or outside using clouds in daylight
- I tried different software (DSS, Siril and PixInsight, although I am a real beginner with PI...)
- I tried different ADUs, from low 19500 to high 33000.

Same result each time. Flats overcorrect.

I am a bit lost there...

Helpful Engaging

Björn Arnold

Jan 28, 2022 · Edited: Jan 28, 2022

Hi Axel,

Excuse me for this seemingly stupid question: when exactly are you applying the flats?
A Masterflat is applied to calibrate each light frame. I somehow have the impression that you apply the flat frame after integration?

Sorry, if I am mistaking something!

The image above, where you don't apply the Flat, shows clearly why Flats are taken. The dust spots remain and would need cumbersome post-processing. It also helps with vignetting.

CS
Björn

Helpful Respectful

Doversole83

Topic starter

Jan 28, 2022

Björn Arnold:
Hi Axel,

Excuse me for this seemingly stupid question: when exactly are you applying the flats?
A Masterflat is applied to calibrate each light frame. I somehow have the impression that you apply the flat frame after integration?

Sorry, if I am mistaking something!

The image above, where you don't apply the Flat, shows clearly why Flats are taken. The dust spots remain and would need cumbersome post-processing. It also helps with vignetting.

CS
Björn

Thanks Bjorn.
You seem to use PI. I recognise the vocabulary!!
In the examples of M101 just above, I have used DSS.
You just specify you lights, flats, dark flats and the software will automatically register, create the masterflat, callibrate and stack.
From my very early experience in PI last WE, I saw indeed you do all these steps manually. Not in DSS.

In PI. I may have it done wrong last WE ...

The "hybrid" image I created was using the 2 stacked images (one with flats, one without) and mixing them in PS.

Hope that's clear. I am getting a headache myself haha.

Jeffbax Velocicaptor

IOTD Staff

Jan 28, 2022

You just specify you lights, flats, dark flats and the software will automatically register, create the masterflat, callibrate and stack.

*Hi Axel. And of course you use darks (not specified) ?

JF

Doversole83

Topic starter

Jan 28, 2022

Jeffbax Velocicaptor:
You just specify you lights, flats, dark flats and the software will automatically register, create the masterflat, callibrate and stack.

*Hi Axel. And of course you use darks (not specified) ?

JF

Not in this case.
I used Darks in the past but they seem to have very little effect on the ASI533MC (almost no amp glow). I use them regularly with my ASI1600MM though.

Axel

Tim Hawkes

Jan 28, 2022 · Edited: Jan 28, 2022

Judged by your images I had what looks like exactly the same problem in the past with an ASI294MC and it did turn out to be a mismatch in the Darks. It looked exactly like your pictures. The problem was entirely down to me having taken lights at a different offset value (a.k.a. brightness for ZWO cameras) from the value I had used for creating the master darks. It makes sense because calibration calculates (Lights - Darks)/ Flats and - numerically - the effect of the flat can appear to be either exaggerated or undercooked according to whether the dark value is offset to be higher or lower than it should be. Anyway fixing this solved the problem in my particular case.

Helpful Respectful

Jeffbax Velocicaptor

IOTD Staff

Jan 28, 2022 · Edited: Jan 28, 2022

Not in this case.
I used Darks in the past but they seem to have very little effect on the ASI533MC (almost no amp glow). I use them regularly with my ASI1600MM though.

Axel

*I dont know if this is the solution. But IMO if you don't use darks, you should not use flats darks either.

The darks contain the thermal signal, but Also the offset signal. In your case, you substract this offset from you flats using flat darks but not from your lights.

Might be interesting to try with darks, or at least with an offset substraction from the lights.

Just my 2 cts.

JF

Doversole83

Topic starter

Jan 28, 2022

Update.

@D.Jung kindly guided me to use PI.
No more overcorrected flats.
Thank you all for your inputs. I guess this has convinced me to follow the PI route!

D. Jung

Jan 28, 2022

All hail to pixinsight