Flat issue, is it time for an upgrade?

[Ivor]

I'm having a real challenge getting my flats to be evenly illuminated I've completed some systematic testing, with sky flats, a diy light box and a sketch back light and flat subs ranging from 2 - 15 seconds I see no significant improvement.

I suspect the cause of the problem is my light train rather my approach to capturing the flats.

I'm using a ASI2600MC with a william optics 0.8x flattener AFR-IV on a FLT110, 

WO AFR-IV https://ensoptical.co.uk/index.php?route=product/product&product_id=16035

I bought the ASI2600MC (35x26mm) to replace a SBIG 8300M (18 x 13,5mm) and I think my issue stems from the different in chip size even though the AFR-IV stated is supports a 35mm image circle.

Looking at a light frame it appears there is a small tilt issue however the AFR-IV spacing is right and the frame is fully illuminated.

 

My flats come out looking like this.

 

 

Would I be right in thinking I need to replace AFV-IV?

[Elp]

The diagonal measurement of the 2600 is 42mm (sensor on the drawing is 33 x 26) so it's larger than the illuminated circle hence the vignetting, well what it would seem from your flats, you can also see the loss of contrast in the star image.

But the spec of the flattener doesn't match up, it states suited for 35mm cameras and 50mm clear aperture, so it should be a large enough illuminated circle.

Are you using filters?

Edited 12 hours ago by Elp

[vlaiv]

13 minutes ago, Ivor said:

I bought the ASI2600MC (35x26mm)

 

5 minutes ago, Elp said:

The diagonal measurement of the 2600 is 43.6mm

What am I missing here? ZWO says ASI2600MC pro is APS-C sized sensor.

In case you both meant ASI2400 which is full frame sensor - then issue is simple diagonal of camera is 43mm while flattener has T2 connection which is M42x0.75.

Any sort of T2 adapter will have clear aperture of about 40mm or maybe even a bit less as you need to have some metal in those threads so if you say take 1mm thick adapter that is two sides of 1mm or 2mm of material in total - M42 - 2mm = 40mm or less.

I found the specs that full illuminated circle is 35mm for that flattener:

That could work on APS-C sized sensor that has diagonal of about 28mm but will vignette severely on 43 diagonal of full frame.

[Elp]

2 minutes ago, vlaiv said:

What am I missing here? ZWO says ASI2600MC pro is APS-C sized sensor.

Maybe it is. The mechanical drawing shows otherwise:

https://www.zwoastro.com/product/asi2600/

 

[Elp]

4 minutes ago, vlaiv said:

I found the specs that full illuminated circle is 35mm for that flattener:

I thought that at first, but when I read it again I think it means it suits 35mm format camera sensor, and can illuminate 50mm circle, but that's not what the flats show.

[vlaiv]

1 minute ago, Elp said:

I thought that at first, but when I read it again I think it means it suits 35mm format camera sensor, and can illuminate 50mm circle, but that's not what the flats show.

How can it do that if it has M42x0.75 connection?

In order to illuminate anything close to 50mm circle - you'd need 2" or M48 connection on camera side.

[Elp]

2 minutes ago, vlaiv said:

How can it do that if it has M42x0.75 connection?

In order to illuminate anything close to 50mm circle - you'd need 2" or M48 connection on camera side.

The description is plain confusing. They shouldn't even mention the 50mm at all.

[vlaiv]

1 minute ago, Elp said:

The description is plain confusing. They shouldn't even mention the 50mm at all.

Indeed - even if we understand it as being aperture at the scope side - they are still wrong as it has 2" nosepiece and thus must be less than 50mm (again same logic - 2" is 50.8mm and there must be some material in nosepiece walls and I'm sure it's not only 0.4mm thick metal tube).

 

[Elp]

I would have thought any distance away from the back of the flattener will also make a difference in this case due to the sensor size.

[vlaiv]

1 minute ago, Elp said:

I would have thought any distance away from the back of the flattener will also make a difference in this case due to the sensor size.

Yes - you can sort of think of it this way - entrance is 50mm and it has reduction of 0.8 so "exit" will be at most that wide - or 50mm * 0.8 = 40mm and then you have working distance and converging cone - which will make it even smaller, so 35mm illuminated circle is about right.

It should illuminate APS-C sized camera without issues if there are no filters in the way that might cause vignetting - but it will certainly not illuminate full sensor.

We still don't have any idea what the actual camera used is - is it 2400 and full sized sensor or is it 2600 and APS-C sized sensor (and in that case I'd probably blame filters or something else).

[Steve Ward]

16 minutes ago, Elp said:

Maybe it is. The mechanical drawing shows otherwise:

https://www.zwoastro.com/product/asi2600/

 

Compare the photo of the camera front to the mechanical drawing and you'll see the difference between the opening and the actual sensor.

The description states the sensor size several times.

[symmetal]

The AFR-IV with its 2" nosepiece has a clear aperture of 47.5mm reducing to 45mm about 10mm from the opening. Just measured mine. The adjustable 'back-focus' distance scale on the AFR-IV is only correct when the sensor is placed 55mm from the rear of the AFR-IV.

The 2600 is APS-C. I haven't used the two together. The 35x26mm shown on the drawing is the size of the rectangular opening behind the adjustable tilt plate.

Alan

Edited 12 hours ago by symmetal

[Elp]

10 minutes ago, Steve Ward said:

Compare the photo of the camera front to the mechanical drawing and you'll see the difference between the opening and the actual sensor.

The description states the sensor size several times.

Maybe it's my background, the mechanical drawing is the bible and should be accurate to what's it's supposed to be dimensioning. The image above confirms the sensor is smaller than the aperture dimensioned (what exactly is the point of dimensioning the aperture around the sensor in the first place and sharing it to customers?).

Edited 12 hours ago by Elp

[symmetal]

11 minutes ago, Elp said:

(what exactly is the point of dimensioning the aperture around the sensor in the first place and sharing it to customers?)

I edited my reply after your quote. The aperture dimension is the size of the rectangular opening on the front of the camera.

Alan

[Ivor]

wow I grabbed dinner and my question has exploded, thanks for the repsonses.

I don't use any filters nor a OAG.

light frames get elongated stars very quickly if the back space is off so I'm confident I have the spacing right. I allowed for distance from the front of the asi2600 to sensor when adding in the spacers.

 

 

Edited 11 hours ago by Ivor

[Steve Ward]

20 minutes ago, Elp said:

Maybe it's my background, the mechanical drawing is the bible and should be accurate to what's it's supposed to be dimensioning. The image above confirms the sensor is smaller than the aperture dimensioned (what exactly is the point of dimensioning the aperture around the sensor in the first place and sharing it to customers?).

I guess the 'mechanical' drawing shows the dimensions of the 'mechanical' bits , not the electronic bits ... 😏

[vlaiv]

17 minutes ago, Ivor said:

allowed for distance from the front of the asi2600 to sensor when adding in the spacers.

So it is definitively ASI2600 - meaning APS-C sized chip (IMX571), right?

If so - then it's definitively time to change the flattener as this one simply does not work as specified. It claims 35mm of fully illuminated circle and it's obviously not the case.

[Ivor]

Yes I accidently caused confusion by taking the dimension for the snsor from FLO mechanical diagram https://www.firstlightoptics.com/zwo-cameras/zwo-asi-2600mc-pro-usb-30-cooled-colour-camera.html

Until I replace the AFR is can I edit out the issue in PI?

As the FLT110 is discontinued most flattener don't mention the FLT110, what should I look for in a replacement?
Looking at the WO site you would think my scope didn't exist https://williamoptics.com/products/flat8?pr_prod_strat=e5_desc&pr_rec_id=83916f91f&pr_rec_pid=8513622507737&pr_ref_pid=8678961086681&pr_seq=uniform

Edited 11 hours ago by Ivor

[vlaiv]

4 minutes ago, Ivor said:

Until I replace the AFR is can I edit out the issue in PI?

I took another look at flat you posted - one with measurements. First time I did not look at the scale on the right side.

Can you post one of your raw flat files?

It looks like total vignetting is about 15% at the edge - that should be easily solved with flat field calibration. What issue do you have with flats exactly?

[Ivor]

With short subs like with this image https://www.astrobin.com/psukpw/ I get away with it OKish, however I could (with more skill and practice) pull out more detail if it wasn't for the issues with the flats.

I am currently seeing what I could get on IC 1848 without filters using 300s subs, the image below is a basic STF transferred to HT on the integrated image. I used WBPP to integrate.

2024-09-22_21-49-40_ASI Camera (2) (ASCOM)_28.00_15.25s_0000.xisf

 

Edited 10 hours ago by Ivor

[vlaiv]

3 minutes ago, Ivor said:

I am currently seeing what I could get on IC 1848 without filters using 300s subs, the image below is a basic STF transferred to HT on the integrated image. I used WBPP to integrate.

Well, flats are over correcting - which is sign that something was wrong with calibration.

What sort of calibration data are you using? Darks, dark flats and flats or something else?

[Ivor]

Just darks, flats & Bias

[symmetal]

It appears the initial flat you posted has been stretched by applying an STF in Pixinsight which is why it looks awful. You need to look at the master flat raw with no processing applied to see what it actually looks like.

As vlaiv has said the AFR-IV should be fine with an APS-C sensor. If vignetting is too severe then the final stacked image after calibration will look noisy in the corners compared to the rest of the image as a lot more gain has to be applied to the corner pixels to get an even background brightness.

Your image just posted has over-corrected the corners, as vlaiv has said, but they don't look any noiser than the rest of the image so the actual vignetting is quite modest and easily fixed with the correct flat frame.

You will likely get some tilt with the AFR-IV, as the 2" nosepiece attachment isn't as mechanically stable as a screwed together system. I found this problem with my AFR-IV when I was using it around 8 years ago.

Alan

[vlaiv]

1 minute ago, Ivor said:

Just darks, flats & Bias

Well, I have to ask silly questions like these to try to get to the bottom of things:

1. did you change offset / gain between darks flats and bias files?

2. did you change temperature (darks shot at one temperature but bias without cooling or something like that)

3. Is there any chance of light leak in your system? How did you take your darks? With camera off the scope or did you simply put the scope cover on the scope and shoot darks?

4. How did you take your flats? Is there any chance of some light that is not coming thru the objective lens hitting the sensor?

[vlaiv]

Hm, let me think.

Over correcting flats mean that signal / flat is higher than it should be so either is signal higher than it should be or flat value is lower than it should be.

Flat value lower than it should be is very unlikely - so signal higher than it should be is reasonable explanation (I'm guessing it is not focus change as both vignetting and dust has been equally affected).

Either darks are of mismatching length or mismatching temperature.

Darks taken during the day if it has been hot - might fail to reach set point temperature and are "warmer" then they should be. However - that would not make signal stronger as warmer darks have higher values (and these subtract from signal so signal ends up lower not higher).

Soundest explanation is mismatched dark length (shorter darks than they should be) or light leak when imaging (that also increases signal).