Could someone please check this flat?

[BrendanC]

Hi all,

I've been enjoying my ASI533MC Pro, but I've recently noticed vignetting and gradients at the edges that I'm finding hard to remove. The flats should be fixing this, I think, but they're not.

So, could someone take a very quick look at the attached flat, give it a bit of a stretch, and tell me whether it's OK? Cos it looks like my sensor might be way out of centre, and I need to collimate.

This is through a 130PDS with a 0.9x coma corrector, and a 2-inch SVBONY UV/IR cut filter, created in APT using the flats aid to get to 19,000 ADU, and a Lacerta flat panel.

I've been stacking with just bias, no darks cos I tested this and there's no discernible difference (I also stacked with darks and got the same result). If it helps, I've attached a bias and image file too.

I've checked previous flats, with this and another dualband filter, and get the same pattern, but I'm sure I didn't have these issues. So, it might be something else. I don't know.

I've collimated the heck out of this thing but if I'm too far out, I might just have to give it (yet) another go.

Thanks, Brendan

F_26103_0.28379s__-10C.fit

L_26136_60s__-10C.fit B_22587_1s__-10C.fit

Edited January 19, 2023 by BrendanC

[BrendanC]

[originally posted another file here but not needed]

 

Edited January 19, 2023 by BrendanC

[Jonk]

A quick check for a flat is to calibrate itself with itself. I've just done that and it appears to behave.

However, your flat doesn't have a great deal of change in terms of values from middle to edge. In other words, it's too 'flat'.

When you take your flats, do you aim for 50% or slightly higher than your well depth? Is that where you get the 19,000ADU from? To me it doesn't look enough.

 

[Jonk]

Pixinsight shows the flat to look like this (using the contour plot script), so my previous comment about 19,000 ADU not being enough may have been wrong. I would check alignment as you say.

Turn the camera 180degrees, see if the issue follows the camera or not, this will give you a better idea of where to look.

[BrendanC]

Excellent, thank you so much for this. As I say, other images have calibrated out fine, so maybe I need to look at something else. I haven't had the camera that long so it could just be light pollution from shooting in a different area of sky. I just needed a check on the flats as a first step. Thanks again. 

[Jonk]

No problem - I haven't imaged using a newtonian for a while, but the sensor being centre to the secondary mirror is important.

The sensor being off centre in the camera wouldn't be a huge surprise, I've opened a handful of ZWO cameras and they're just sensor boards screwed down, very easy to unscrew and move.

Having said that, the flat should sort out the light frame as is, as long as nothing's moved. There may be some gradients, you need to stack and check the final image, splitting the RGB into grayscale and dealing with each one alone, then combining, and dealing with any residual on the combined image.

Your issue doesn't look like collimation of the optics per se, I would guess that the camera is off axis with respect to the secondary mirror. It could be the secondary mirror is not true to the focuser drawtube.

Can you try another camera at all, changing nothing else to see if the result is the same or not?

Edited January 19, 2023 by Jonk
Missed a bit!

[ONIKKINEN]

Light frame looks decent, tilt measurements show nothing really and the frame looks good so i think your primary to camera collimation is plenty good enough. The flat frame does look off-center as shown above in a previous comment, which would suggest your secondary mirror is not perfectly centered under your focuser, but this wont stop the subs from being calibrated as long as they match the lights. Tried a bit of pixel math by first subtracting the bias frame from the light frame and the flat frame, and then dividing the calibrated light with the calibrated flat and it works out just fine to my eyes and it looks like your lights and flats match well and correct the off-center illumination, result in ultra stretch mode below:

There is a gradient but it looks quite linear to my eyes indicating a normal sky gradient from light pollution. Doubt this was the difficult gradient you mentioned?

But since you mention using no darks, does this mean you do no dark calibration to your light frames at all? You need to remove offset somehow, and for this purpose a bias frame can be used as a dark frame. But calibration where you dont remove offset at all from your lights will always result in a completely failed flats calibration, like below where i subtracted the biasframe value from the flatframe, but not the light frame (like you would in stacking if you were to not designate the biasframes as darkframes):

If your stacked result looks like this, then the solution is to start using dark frames, or just use your bias for both bias and dark frames (works fine with your camera).

As for the secondary not centered issue, im not sure i would bother fixing that because i am lazy, but in case you want to, a concenter eyepiece makes the job as easy as it can be made in case you had not used one already.

[BrendanC]

Hey, thanks both for your help! 

@ONIKKINEN, I really appreciate you doing all this. I was just using bias, but I did a test with darks yesterday and it did show an improvement. I always used to use them, I'm just a bit lazy too! It's just that I'd noticed very little difference, if any, with just using bias because of the 533's zero amp glow. I guess that for some images it does make a difference. So, I'm going back to 'proper' calibration.

I also stacked another image last night and it came out fine, so I think this was light pollution, which again is a lot clearer to see for some reason when I stack with darks. It's just that horrible feeling when you think something is wrong, but you don't know which of the many possible things, or combinations of things, it might be.

So, all good, and I'm just going to carry on. Totally get what you're saying about the sensor not being exactly central, but if what I'm getting through is workable, I'm not going to start making micro adjustments here and there.

Thanks again both of you. :)

[ONIKKINEN]

2 hours ago, BrendanC said:

Hey, thanks both for your help! 

@ONIKKINEN, I really appreciate you doing all this. I was just using bias, but I did a test with darks yesterday and it did show an improvement. I always used to use them, I'm just a bit lazy too! It's just that I'd noticed very little difference, if any, with just using bias because of the 533's zero amp glow. I guess that for some images it does make a difference. So, I'm going back to 'proper' calibration.

I also stacked another image last night and it came out fine, so I think this was light pollution, which again is a lot clearer to see for some reason when I stack with darks. It's just that horrible feeling when you think something is wrong, but you don't know which of the many possible things, or combinations of things, it might be.

So, all good, and I'm just going to carry on. Totally get what you're saying about the sensor not being exactly central, but if what I'm getting through is workable, I'm not going to start making micro adjustments here and there.

Thanks again both of you.

If you have light leaks or a different offset (or both) in your darks, that would also create a funny looking gradient that is difficult to remove. Might have been why gradients are worse with them?

Darks should ideally be taken with the camera off the scope and completely plugged so that not a single photon gets to enter the camera. I take mine in the fridge now, helps with cooling too.

Edited January 20, 2023 by ONIKKINEN

[BrendanC]

It's actually better with the darks. Believe me, I've been through all those hoops with creating dark libraries, light leakage and all! Thanks again.