Really need help with flats

[BrendanC]

Hi all,

So, another day, another problem.

I've been trying out a 1.25 inch UV/IR filter, to see whether I can use smaller filters with my ASI533MC Pro and 130PDS with 0.9x coma corrector.

Initially I had it attached incorrectly, but now it is most definitely attached OK, using an adapter, right up against the sensor. It should be fine, but I'm experiencing calibration errors with my flats.

This is what I'm getting - pronounced dust motes top left and right:

I'm using flats, darks, and flat darks, all taken with the same setup, all at the same temperature, gain and offset. I don't use bias frames, because they make no difference, and I've tested bias with this data and I get the same result.

The flats were taken with a Lacerta flat field light, using APT's flats wizard to get to 30K.

I've not had a problem like this with my 2 inch filter, and I could just go back to it, but that's kind of academic: there's clearly an issue here, and I want to understand it before I progress. It's probably been in other images too but not visible until I shoot a largely dark star field. Going down to 1.25 inches really shouldn't be a problem with this sensor.

The most likely candidate is that something is moving, but I don't know what. My secondary mirror and focuser seem solid. The primary is held by clips that don't actually touch the mirror, but tilt is fine as is collimation, so I don't suspect mirror slop either.

Or is this over-correction? If so, how does that happen, and what's the fix? Or is it under-correction? Etc.

My head is spinning, frankly. After nearly four years of doing this - two with a DSLR and perfect calibration every time, a few months with an ASI1600 that I really didn't get along with, and now with this camera - I really should be moving along nicely but keep coming up against weird, random issues.

Rather than make suggestions - because, however helpful, they do tend to stack up and leave me no closer to solving the problem - would anyone be willing please actually to look at my data and have a stab at what could be going wrong?

If so, it's here (1.23GB download, includes master dark, flat dark and optional bias)...
Google Drive: https://drive.google.com/drive/folders/1jaZB2V8LRoRhs5Roz8zLrGhYD9pZTXrb?usp=drive_link
OneDrive: https://1drv.ms/f/s!AqovBuVZMwj3mqlsJBNWnsGxgkBghA?e=zqcXl8

PLEASE only make suggestions if you're pretty confident about the fix. I've had a ton of problems recently and I'm starting to lose my mojo.

Thanks
Brendan

 

[Xilman]

It is not clear from your description whether you are taking separate flats for each filter you use. I hope you are, because each filter will have its own distribution of crud on its surfaces.

Ideally you should take flats before each observing session so that changes in dust distribution are accounted for. Twilight flats are generally used for this approach, though dome flats are also a possibility. In my experience, very few people are this fastidious --- I am certainly not, but I have a very stable set up with a fully enclosed optical train mounted in an observatory so I can get away with re-using flats for a period of a few weeks.

 

[BrendanC]

It's a one-shot camera so just the one set of flats, taken before the shoot, using the Lacerta box.

[BrendanC]

Quick bump in case anyone else can help?

[ONIKKINEN]

I'll take a look at the data later today and see if there is something that comes to mind.

I think you are right to assume that something is moving, but figuring out what is not easy. One test you can run to confirm this hunch is to take flats at different tube orientations and then divide one flat with another using pixel math.

For example take one flat with the tube pointed at the zenith, then one with the tube horizontal. Divide the zenith flat with the horizontal flat and see what you get. If everything is rock solid you get an even image with no gradients or blotches of any type as the pixel math result. If things are moving you get some kind of pattern as a result and the suspicion is confirmed.

Edited December 21, 2023 by ONIKKINEN

[BrendanC]

Interesting test, I'll give that a go when I get the chance.

 

[ONIKKINEN]

I cant see any issues in the data itself, darks and darkflats look good and match in offset to lights so not a data issue i would say. The only thing that an analysis of the fits headers reveals is that you had very slightly different focus positions between the flats and lights, 54 steps to be exact. Not sure how much movement that is with your focus motor and focuser, but doesn't sound like a lot and is probably not at all an issue. After all focus moves in and out as the night progresses and your tube contracts as it cools and this would almost certainly be a larger movement than whatever the length of 54 steps is (and flats still should work regardless of small thermal expansion during night - at least never had issues with that myself).

Give the pixel math flat test a try, at least you'll get a definitive answer on whether its scope stability related or not.

[BrendanC]

Absolute superstar - thank you so much for taking the time to do this, as ever I really appreciate it.

As you say, the difference in focus positions shouldn't be drastic, as focus changes throughout the night. I refocus every hour.

It's a strange one. Here's a super stretch of another cluster I took the same night. I think I can just about make out similar artefacts in this one, but they're very indistinct.

So clearly something happened between that one, and this one.

I'm wondering if it's a combination of factors:

• The M67 data was collected at the very end of astro night (and in fact going about 15 minutes beyond), and at that time, it was pointing at a house nearby. I think they might have had a light on all night, without curtains, so the scope was pointing almost directly at that light. This could have affected things in some way, with lots of light bouncing around. There were certainly strong gradients in the subs, as you probably noticed.
• I recently removed the flocking from my scope because it was coming away. So, if there was lots of light bouncing around in the OTA, the lack of flocking might have exacerbated it.
• There may have been dew. It wasn't forecast, but when I brought the scope in later that day, the primary was dewy, and I had to discard almost half of the subs but I assumed that was because of cloud, not dew. So, the subs may not have matched the flats because of this.

So, I wonder whether these, combined, have just created a unique set of circumstances. Thing is, wouldn't you know it, they happened on the same night I was testing a new filter.

Anyway I will definitely do your test. When I do the pixel math, will it really just be a simple expression of zenith flat / horizontal flat? And it should yield a flat image with no details at all?

Thanks again, you've already been very helpful.

 

[Elp]

You may still see dust etc on the flats, I don't see details for flats taken at F3 or faster. But you'd also see the same dust etc on your lights.

[ONIKKINEN]

10 hours ago, BrendanC said:

Anyway I will definitely do your test. When I do the pixel math, will it really just be a simple expression of zenith flat / horizontal flat? And it should yield a flat image with no details at all?

Yes, just simple division of image one / image two.

Here are some examples of what the results look like, with the false colour rendering mode and histogram view mode enabled. First a known good flat divided by another known good flat of the same image run:

Then i cropped 8 pixels off the left edge of flat 1 and 8 pixels off the right edge of flat 2, effectively shifting the center of the optics by 8 pixels, or 30,08 microns.

Now there is a little something, maybe. Edges of some dust motes may be visible, and calibration could fail with this one. Only 30 microns of movement!

Then as the last one, i cropped 16 pixels from both from opposite edges and now the flat is completely ruined with dust spots readily available:

 

[BrendanC]

Wow. Perfect demonstration there. Thank you again.

[ollypenrice]

Firstly, can we be sure we are talking about the same thing? The aberrations I see from your flats are arrowed below.

1) These don't look like any dust motes, or any over-correction, that I've ever seen.

2) They do look very like each other. In both cases we see a segment from a circle (from about 3.00 o'clock to about 6.00 o'clock) and, at the circle's centre, we see a dark spot. They are so similar to each other that we have to consider the possibility that they come from the same source. Since they are dark in the flats-calibrated image, they must have been bright in the source. An educated guess would tell us that, if they were not cropped by the edge of the chip, they might look like a full circle with a spot in the middle. What telescope are you using?

Could the bright flats-equivalent of the dark patches on the calibrated image be reflections created by some part of the light path illuminated by the panel? Another argument in favour of this comes from the fact that, the artifacts apart, the image posted here is very well flattened.

I would bet on a reflection.

Olly

Edited December 22, 2023 by ollypenrice
typo

[ONIKKINEN]

14 minutes ago, ollypenrice said:

Firstly, can we be sure we are talking about the same thing? The aberrations I see from your flats are arrowed below.

1) These don't look like any dust motes, or any over-correction, that I've ever seen.

2) They do look very like each other. In both cases we see a segment from a circle (from about 3.00 o'clock to about 6.00 o'clock) and, at the circle's centre, we see a dark spot. They are so similar to each other that we have to consider the possibility that they come from the same source. Since they are dark in the flats-calibrated image, they must have been bright in the source. An educated guess would tell us that, if they were not cropped by the edge of the chip, they might look like a full circle with a spot in the middle. What telescope are you using?

Could the bright flats-equivalent of the dark patches on the calibrated image be reflections created by some part of the light path illuminated by the panel? Another argument in favour of this comes from the fact that, the artifacts apart, the image posted here is very well flattened.

I would bet on a reflection.

Olly

They look like newtonian dust motes with how the central obstruction also produces its own shadow, but i do agree they are unusual.

Using this tool might give an answer on where in the imaging train this came from: https://astronomy.tools/calculators/dust_reflection_calculator

Lack of flocking might make things worse, although never did see it quite this bad with my newtonian before flocking. Lights from a nearby house could also make things worse. Lots of "could" options here, testing is required to single out the issue, or a multitude of issues that all work together in exaggerating the motes/reflections.

[BrendanC]

Thank you for the input Olly.

I've tried reading this several times and I still don't quite understand it: "Could the bright flats-equivalent of the dark patches on the calibrated image be reflections created by some part of the light path illuminated by the panel?"

I'm wondering whether it's reflections too.

This is from a 130PDS Newtonian.

[ollypenrice]

23 minutes ago, BrendanC said:

Thank you for the input Olly.

I've tried reading this several times and I still don't quite understand it: "Could the bright flats-equivalent of the dark patches on the calibrated image be reflections created by some part of the light path illuminated by the panel?"

I'm wondering whether it's reflections too.

This is from a 130PDS Newtonian.

My English was pretty garbled - sorry! I was called away mid-post and was rushing.

You have artifacts shaped like a circle with a with a dot in the middle. This is pretty much what you have when you look down (or up) a Newtonian... This might be a coincidence but it might not.

What is dark on your calibrated image will have been bright on the flats in use. So you are looking for something which created a dark bright ring with a bright dot in the middle of it.  That's why I suspect a reflection, made even more likely by the fact that there are two identical artifacts offset from each other, perhaps by a combination of tilt and refraction.

Could you post the stretched flat used to calibrate the image we are discussing?

Olly

Edit: I haven't used Newt flats so wasn't aware of what ONIKKINEN said about their appearance.

Edited December 22, 2023 by ollypenrice

[BrendanC]

Sure, here you go - really stretched master flat (actually taken from the GHS preview cos I think it displays in fewer bits and shows off the artefacts better).

I can see what you're saying (literally) - the two circles do look somewhat like what you'd expect to see when looking down the OTA, complete with a bit of secondary spider for good measure. and therefore could be reflections.

[ollypenrice]

1 hour ago, BrendanC said:

Sure, here you go - really stretched master flat (actually taken from the GHS preview cos I think it displays in fewer bits and shows off the artefacts better).

I can see what you're saying (literally) - the two circles do look somewhat like what you'd expect to see when looking down the OTA, complete with a bit of secondary spider for good measure. and therefore could be reflections.

Oh yes, that secondary spider vane is a dead give-away. That is not a dust bunny, it's your telescope.

Olly

 

[BrendanC]

That's no moon... 

Right. So, is this a light leak? 

[ONIKKINEN]

I think your flat looks normal, ghost views of the central obstruction included. All my flats have those and always have, so i dont think there is something to worry about with that.

I'll post one of my own flats in a bit, which also has those dark spots resembling the secondary.

[BrendanC]

I've just been going through my old flats and they all have it too. I recently installed the Backyard Universe secondary spider, and before and after that, I can still see ghost views of the central obstruction.

I guess that no matter what is being captured in a flat, it should be calibrating out what's in a light.

Which leads me back to your test for things shifting when taking flats.

[ONIKKINEN]

3 hours ago, BrendanC said:

Sure, here you go - really stretched master flat (actually taken from the GHS preview cos I think it displays in fewer bits and shows off the artefacts better).

I can see what you're saying (literally) - the two circles do look somewhat like what you'd expect to see when looking down the OTA, complete with a bit of secondary spider for good measure. and therefore could be reflections.

Here is one of mine:

I think these are from the sensor window as they are quite small in size. The next glass element in my scope is in the coma corrector so probably too far away to make noticeable shadows, unlike with yours since you have that filter in between.

[BrendanC]

Interesting. Thank you both again for your help. I'll try the flats test and I'll give the 1.25 inch filter another go but at the end of the day, since I had no problems with my 2 inch filter, if that just works then I'll go back to that, keep calm, and carry on.