Flats problem

[pi_co100]

I must be doing something daft when I am taking flats. Or it is Pixinsight?  My stacked luminance image + flats (boosted in Pixinsigth) looks like this...

 

The stacked image at first glance made me thing that the flats had not been processed. Having examined the flats (again curves boosted) in more detail, I found that they are actually ruining image... This is a light frame not calibrated with the flats

Here is the master flat frame with appears to be culprit... (STF boosted)

Now, I have been taken flats and light frames with the same telescope, camera and filter and orientation... It is puzzling that the spots only appear on the flats...

To take my flats, I used an EL panel and put some sheets of white paper in front of the lens. I have used the flat calibration wizard in SGP to determine the exposure (the flat has a median of 30000 ADU).

Is it the way that Pixinsight is processing the flats that is causing all this?

Comments/views are very welcome... 

Thank you all.
Piero

[MarsG76]

I think that your flats do not match the lights and so are introducing artifacts... did you take the flats straight after exposing your lights?

 

[pi_co100]

Yes, as soon as the lights were finished... a mistery!!! Thanks...

[MarsG76]

1 hour ago, pi_co100 said:

Yes, as soon as the lights were finished... a mistery!!! Thanks...

That is strange... 

When you created your Master Flat, did you stretch it at all, perhaps there is some clipping in the final Master flat image producing those effects.

[Martin Meredith]

Did you also use a master bias or flat dark? If not, you can get these kind of over-correction effects.

Or: it does look as though the dust may have moved slightly since the artefacts are darker on one side and lighter on the other.

Martin 

[pi_co100]

Thanks to all.

The problem seems confined to the flats taken with the Luminance filter. Here is a comparison between Blue flat/ Blue light calibrated and Luminance flat/ Luminance light calibrated. I have used the functionality "Batch preprocessing" in Pixinsight with bias frames. That generates a master bias. There seems to be no problem with the flats and the processed lights with the blue filter (on the right). I taken the blue flats after the lum flats so my guess that would rule out the dust movement?The process I followed was the same. I am starting to think if this something introduced by the EL panel... I will try using my laptop screen next time; see if that removes the problem.

 

[Oddsocks]

On 06/05/2018 at 08:53, pi_co100 said:

I am starting to think if this something introduced by the EL panel... I will try using my laptop screen next time; see if that removes the problem.

Nothing wrong with your flats panel, this is purely a mechanical issue.

If your flats panel were to blame the effect would be consistent across all the filters.

If you look really carefully at the dust shadows in the luminance frame and the dust shadows in the corresponding luminance-flat frame you will see that the shadows in the flats image are all shifted slightly to the left.

When you calibrate the light with the flat if there has been some mechanical shift in the optical path between taking the light and taking the flat then you will be left with these 'bas-relief' artefacts.

Sometimes this can affect one or two individual dust-donuts in an image if the dust moves about during the time it takes to acquire all the lights and darks but in your image it has affected all the dust artefacts equally so you can rule out 'loose' dust as the reason.

In your image all the 'bas-relief' artefacts have the shadow on the same side, they are not concentric, this indicates that there was a mechanical shift somewhere in the optical train in the 'X' direction, sometime between when the luminance lights were shot and when the flats were taken.

Concentric bas-relief artefacts can be due to a focal distance shift, a change in the 'Z' direction, or a problem with under or over image calibration when the flats are applied to the lights.

Non concentric bas-relief artefacts are a mechanical movement in the 'X' and/or 'Y" directions that leave shadows in the image as if the image were lit strongly from one side.

You are looking for a reason that there was a large mechanical shift in the 'X' direction of the optical train between when you took the luminance light and when you took the luminance flat.

Because the shift is pretty much horizontal, left-right, your other filter flats are all perfect, and provided you can rule out the camera or filter wheel being knocked to one side between the time you took the luminance lights and the luminance flats then my money would be on a mis-registration of the filter wheel carousel. 

Either when the lights were taken or when the flats were taken the wheel did not stop in exactly the same place and the luminance filter rotated a little short or a little long when at the luminance filter position.

If the problem occurs again open the filter wheel and check that the filters are all secure, un-mounted filters can shift a little if not tightened down, mounted filters can foul the filter wheel housing if not screwed down fully in the carousel though in this case unlikely as it would effect all the filters in the sequence. If the carousel is friction driven check the driving wheel is clean and free of oil or grease contamination and if using a rubber 'O' ring as a driving interface that the 'O' ring is in good condition. If the carousel uses a belt drive check that the belt is adjusted correctly and not loose. If the carousel stop positions are determined by I.R. light barriers check the barriers are clean and not dusty. If the carousel uses a mechanical detent roller to 'lock' the wheel in each of the filter stop positions check that the roller is free to move and that it drops into the corresponding notch in the carousel perimeter with a solid engagement at each of the filter positions. A power supply issue to the filter wheel, even very intermittent, can cause the filter carousel to lose it's positional reference and subsequent calls to place the filter in the optical path may lead to a wrongly positioned filter.

Lastly, a sporadic software glitch could also be responsible for the filter carousel not stopping in exactly the same place when the luminance filter was rotated into position for the luminance light and luminance flat. These do occur from time to time with electronic filter wheels, occasionally a carousel will stop in slightly different positions depending which way they are rotating, clockwise or anti-clockwise so the filter stop position can change a little depending which filter was used previous to the luminance filter and which direction the carousel rotated to reach the luminance stop position.

Hope the above is useful, at least, you don't need to worry about the flats panel, this is not the reason for the problem with this image data set.

[pi_co100]

Many thanks for the comprehensive response. That is super helpful.

Filter-wheel mechanical issue - another thing to add to the list of what can go wrong. (and what can go wrong  will go wrong! )

When we look at some incredible astro images that some of the very skillful astrophotographers take probably we do not realise the person who has taken the picture:

- has a great deal of IT skills getting all the ASCOM drivers to work and related acquisition software

- they have a good understanding of optics

- they know all about autoguiding algorithms

- they understand imaging processing very well

- the know phoptography and astronomy

- ...

 but most of all they are great detectives! 

 

[Martin Meredith]

Thanks Oddsocks. I learned a lot too from that explanation.

Martin