Sigh* Newtonian flats

[david_taurus83]

The bane of many a newt imager, I'm sure! I shoot flats every session so as to eliminate mirror movement. On LRGB I get a weird dark patch to the right of centre. Ignore the dark shadow on the left hand side, that's from my OAG prism and I can crop it out. I thought I was doing everything correctly. Tube is flocked, I have a 12" dew shield, I autofocus my chosen filter at the start of the session as well before flats to get the position correct. Not sure exactly what it could be? Below is 2 targets, flats applied to the left stacks, flats and stacks without flats.

 

To rule out Pixinsight i tried DSS (below, left) and get a similar result. To throw a spanner into the mix, narrowband is completely unaffected and calibrates just fine! The Ha flat looks horrible compared to the broadband filters but works a charm.

 

Any ideas? 

 

[pipnina]

Is this on the 200P?

I didn't have any issues with my 130-PDS in this reagard, however I struggled endlessly to get my mirrors to stay put on my TS-PHOTON 8"

My conclusion has been that smaller mirrors can be held securely far more easily, and even the simple step up to 8" can make a world of difference to the rigidity required.

Out of interest, have you tried using a laser collimator, and turning the scope in the mount and seeing if the return point of the laser moves as the scope flips around? This revealed a lot of things in my photon.

[david_taurus83]

7 minutes ago, pipnina said:

Is this on the 200P?

I didn't have any issues with my 130-PDS in this reagard, however I struggled endlessly to get my mirrors to stay put on my TS-PHOTON 8"

My conclusion has been that smaller mirrors can be held securely far more easily, and even the simple step up to 8" can make a world of difference to the rigidity required.

Out of interest, have you tried using a laser collimator, and turning the scope in the mount and seeing if the return point of the laser moves as the scope flips around? This revealed a lot of things in my photon.

Sorry, I probably wasn't clear. The mirror doesn't move, I meant movement of sensor relative to the mirror.  Colimation isnt perfect but i have round stars with your old GPU coma corrector so i leave well alone! It's the flats calibration that is giving me trouble.

[ONIKKINEN]

You sure its not just a gradient?

Looks like a fairly normal linear sky gradient to my eyes and not good old newtonian flat headaches. Narrowband Ha would be largely unaffected because the sky is much darker at Ha wavelenghts so less gradient overall.

The uncalibrated stacks are a poor comparison when autostretched, because in those cases the shadow reference used by the autostretch would be from the vignetted corners instead of the sky gradients. Vignetting likely being significantly more severe than the gradient leads to the gradient being invisible on an uncalibrated autostretched image. Maybe try cropping the images so that only the fully illuminated center is visible and see if they still look different?

[david_taurus83]

57 minutes ago, ONIKKINEN said:

You sure its not just a gradient?

Looks like a fairly normal linear sky gradient to my eyes and not good old newtonian flat headaches. Narrowband Ha would be largely unaffected because the sky is much darker at Ha wavelenghts so less gradient overall.

The uncalibrated stacks are a poor comparison when autostretched, because in those cases the shadow reference used by the autostretch would be from the vignetted corners instead of the sky gradients. Vignetting likely being significantly more severe than the gradient leads to the gradient being invisible on an uncalibrated autostretched image. Maybe try cropping the images so that only the fully illuminated center is visible and see if they still look different?

I'm not so sure. I do get gradients depending on where I'm pointing and they are your typical brighter on one side to less so on the other. This looks too localised to me. And almost identical on 2 different targets taken months apart with different position/declination in the sky? Could a light leak cause issues? 

[ONIKKINEN]

27 minutes ago, david_taurus83 said:

I'm not so sure. I do get gradients depending on where I'm pointing and they are your typical brighter on one side to less so on the other. This looks too localised to me. And almost identical on 2 different targets taken months apart with different position/declination in the sky? Could a light leak cause issues? 

The reason i think calibration has worked ok is that the extremes of illumination in your flat and light appear to be corrected without anything funny going on in the corners. Below a quick example with the STF in histogram mode in Siril and rainbow false color rendering:

I would expect the extremes to have something wrong with them in the calibrated image if flats were up to no good, but it looks fairly even to me and this method of extreme stretch and a false color rendering reveals the tiniest issue possible in gradients and flats. These are JPEGs of course but it works ok for this comparison purpose since the image is stretched. I suppose it could be just a coincidence that the 2 images were oriented the same way with up, in terms of towards the zenith, being oriented similarly in both cases?

But light leaks could throw a wrench in the works too. I sometimes have a light leak from the back of the scope by choice, because i want to have ventilation on the mirror if its still cooling so i choose the lesser of 2 evils by allowing light leaks in but keeping the mirror ventilated. Most of the time with light leaks from the rear of the scope the effect is similar to a slight overcorrection of flats where corners appear brighter than they should (something missing from your M63 for example). If you dont have any kind of cover behind the mirror keeping light out, that might be a good idea to try next time.

Hard to say from these images, could be an issue could be a coincidence. If the gradient is removable with DBE/something else (Seems to work ok on the JPEG) then i would probably be inclined to ignore the issue. You could have a light leak in your dark too by the way, but i am assuming you have checked that already?

[pipnina]

if the sensor is moving, then that implies the collimation is being thrown off at the focuser via slopping perhaps?

If that's rigid, I dunno why your flats would be unbalanced...

It is possible though that the flat panel being close to the aperture is allowing light into the focuser, this could potentially cause a visible issue...

However the only other possible cause might be that the flats are overcorrecting. If your flats and lights are collected with an offset, try calibrating with input pedestal set to match the ADU value of the offset (the median pixel value in a bias frame)