Flats, what's going wrong?

[DaveS]

I took some flats to help with processing my Stephan and Deer image but the result has been a disaster. I used the white T-shirt method, aiming to put the peak of the histo at around the 24000 to 26000 ADU, and taking 20 subs of each L, R, G, B. In AstroArt I made a master Flat using Median stacking with no alignment. Here are the PNGs so you can see. I'll upload the FITS if needed.

Master Flat

Luminance stack with Flat.

Luminance stack without flat.

What the heck is going on? I'm obviously doing something stupid, but what I don't know. The last image I posted in the Deep Sky Imaging section obviously needs flats as I can see the holes in the background.

[Jessun]

Very odd indeed! The flat looks inversed and way to blotchy without showing a single dust bunny on any optical surface, and I don't get the sharp lines in the stack either...

The luminance with no flat looks amazingly flat as it is.

I have no idea what's going on here!

/Jesper

[DaveS]

Thanks. Jesper

I pulled the PNG into GIMP 2.9, inverted it then saved it back. In AA I pulled the PNG in, converted to FITS and saved out. I then used the inverted flat in a pre-process, and teh result is looking much better.

Now, how the [lots of removed words] did my flat, or stack get inverted in the first place?

I don't have enough hair left to be pulling any of it out!

[Jessun]

There were still some odd lines in there, and what looked like rotation. (as if the software tried to align the flat - well flats don't move...)

/Jesper

[DaveS]

There were alignment edges on the lights as they were taken on several nights, some before and some after meridian flip.

[sharkmelley]

Very weird indeed!

Which scope is it?

Have you taken flats on previous occasions?  Are they always T-shirt flats? What might have been different this time?

What was the light source?

Mark

[DaveS]

It was the 130 f/7 with the TS 2.5" flattener, 1.25" filters and Trius 694.

I've always used T-shirt flats taken against the north sky.

TBH I have seen flats doing this but this is the first time I've bothered persuing it.

[DaveS]

I'm still no closer to solving this problem.

Here is a "master" red flat after stacking in AA, saving as a PNG, doing a linear inversion in GIMP 2.9 and bringing it back into AA. I did just enough histogram stretch to make the gradients visible.

Not only is the gradient not circular, as I'd expect, but there is also that black ring, which can't be there on teh light frames, as putting the unstretched, inverted flat into the preprocessing stack leaves a bright ring on the resultant image.

What's going on? Do I need to have the camera looked at by SX?

[Shibby]

Is it possible you've got light leaking in to your imaging train somewhere?

[DaveS]

Hadn't thought of that.

I'll try taking a "Dark" with the lens capped but in daylight to see if anything shows up.

[wimvb]

Plastic lens caps are surprisingly translucent. I always wrap the capped lens in two layers of aluminium foil.

Your red flat almost looks like you've focused on the t-shirt that you used. Also, the central part doesn't look circular but rather elliptic. Or is that just the uploaded image?

[DaveS]

My 'scope has a very solid metal lens cap, and yes, that gradient *is* oval, why I don't know. The 'scope was still focussed at infinity and the T-shirt was right up against the lens.

Here is an unprocessed stack of 20 red flats, median stacking, bias calibration and hot pixel removal. Hope you can download it. I had to save as  a PNG. The FITS wouldn't load for some reason

You can just see a bright ring around the central dark oval.

 

[sharkmelley]

I'm becoming confused.  The first master flat you posted was a bright oval.  Now you're posting a dark oval!

Can you upload the master flat somewhere, without any processing applied.

To be honest I've always been sceptical of T-shirt flats.  T-shirts have a weave and light can pass directly through a weave depending on direction.  What does a twilight flat look like?

Mark

[DaveS]

OK

Sorry, I've been confusing myself as well.

The PNG above is as the flats came off the stacking, complete with dark oval. The flat with the bright oval has been inverted to try and use it in processing the lights, and given a stretch to make the gradients more obvious. Please ignore it for the moment.

[wimvb]

When I started taking flats, I experimented with aiming the camera/scope at a neutral coloured wall, and not covering it. Just made sure that the scope didn't produce a shadow in front of it. Maybe you could try something similar, just to isolate the problem. As Mark noted, the t-shirt may be part of the cause, and it now comes to the stage where you need to isolate possible causes.

[Thalestris24]

2 hours ago, wimvb said:

When I started taking flats, I experimented with aiming the camera/scope at a neutral coloured wall, and not covering it. Just made sure that the scope didn't produce a shadow in front of it. Maybe you could try something similar, just to isolate the problem. As Mark noted, the t-shirt may be part of the cause, and it now comes to the stage where you need to isolate possible causes.

That's what I do! I take them after dark and by artificial room light. It's even, reproducible and lets me take long enough exposures

Louise

[DaveS]

I should add that last afternoon I tried taking Twilight Flats, but found the same dark oval with the bright border.

I'll have a look for light leaks later today.

What I don't understand is why the central area is *oval* rather than circular as I'd expect. It almost looks like something going on with the sensor.

[sharkmelley]

Very weird.  I've run out of ideas

Mark

[DaveS]

I'll check for light leaks as soon as the weather allows me to get the laptop out and uncover the 'scope.

Beyond that I'll have to get on to SX to look at the camera.

[RayD]

12 minutes ago, DaveS said:

I'll check for light leaks as soon as the weather allows me to get the laptop out and uncover the 'scope.

Beyond that I'll have to get on to SX to look at the camera.

If your Trius has the built in tilt adjuster on the front, there is a small O ring that sits between that and the camera housing to prevent light leakage.  This can fail, or can need the tilt adjuster tightening to compress the ring.

I think this is probably a job for SX if it is this as they have a little laser jig for realigning the adjuster.

Hope you get it sorted, very frustrating.

[DaveS]

Thanks for the pointer Ray, yes my trius does have the adjuster. It could probably do with resetting too, as I think it may have a tilt.

[RayD]

4 minutes ago, DaveS said:

Thanks for the pointer Ray, yes my trius does have the adjuster. It could probably do with resetting too, as I think it may have a tilt.

Sounds like it could be that Dave.  I don't know the costs, but I wouldn't think it prohibitive to have SX change the O ring and adjust it.  At least then you would also know it's all tight and aligned.

[DaveS]

Time tobget on to it ASAP. I'll check for light leaks but also email SX to see what they have to say.

Bother, looks like I'll be without my Trius over the hols. Fortunately I still have the ASI1600.

[RayD]

7 minutes ago, DaveS said:

Time tobget on to it ASAP. I'll check for light leaks but also email SX to see what they have to say.

 

Good idea.  Perhaps try taking some darks in a lit room (or daytime) with the OTA covered and only the camera open to the light, and this would tell if there is light leaking to the extent that it is showing up in flats.

[DaveS]

Now I'm totally flummoxed

This is a 30 sec Dark through the L filter. Well, not really *through* it, but with it in place. Just slightly stretched to make the gradient obvious, else it wouldn't show up on the JPEG.

 

There does look to be a light leak but it's off to one corner (Unless it's amp-glow, but this is a cooled CCD). No trace of the bright ring I was expecting. Will have to contact SX now I think.