Flats not working

[bottletopburly]

I have been trying to calibrate my images so using a zwo533mc ,200pds UV filter ,  so lights,flats and dark flats processing in deepskystacker , flats adu using flats aid in APT set to 25000 panel dimmed to get flats exposure to 4 seconds ,i have tried shorter exposures but i cant seem to get the bunnies to calibrate out ,so can anyone shed any light why this is ?

 

[ONIKKINEN]

I am assuming you have taken the flats after the imaging session, without touching anything on the scope such as removing the camera etc? If so, need to do some educated guesswork. Post a flat, a light frame that matches that flat, a dark frame and a bias/darkflat frame in .fits format as an attachment so the raw data can be seen for possible clues.

I am guessing it will have something to do with light leaks if all the calibration frames are otherwise sound. The rear of the tube will let in light unless you have blocked it somehow for example.

 

[bottletopburly]

5 hours ago, ONIKKINEN said:

I am assuming you have taken the flats after the imaging session, without touching anything on the scope such as removing the camera etc? If so, need to do some educated guesswork. Post a flat, a light frame that matches that flat, a dark frame and a bias/darkflat frame in .fits format as an attachment so the raw data can be seen for possible clues.

I am guessing it will have something to do with light leaks if all the calibration frames are otherwise sound. The rear of the tube will let in light unless you have blocked it somehow for example.

 

hi thanks ,yes imaging train still in situ not moved , i currently have scope covered and just reshot a dark frame and i think you may be correct about light leak from rear as the darks were showing a shadow so i just went out and covered rear with a black bag and that seems to have solved the shadow mark so i have posted a light, a dark flat ,Flat from the other night camera hasnt moved ,and reshot a dark with ota covered, a dark with ota rear not covered, and a Dark flat with rear ota covered  so i need to find a suitable cover for rear of ota images from today the 9th are the reshot ones  . 

Target was the needle galaxy and the moon was bright .

D_Dark uncovered rear OTA_2903_Bin1x1_180s__-10C.fit D_New Dark covered rear ota _2907_Bin1x1_180s__-10C.fit DF_4_NGC 4565_2854_Bin1x1_4.1875s__-10C.fit DF_New dark flat covered rear ota4565_2914_Bin1x1_4.1875s__-10C.fit F_4_NGC 4565_2839_Bin1x1_4.1875s__-10C.fit L_NGC 4565_2683_Bin1x1_180s__-10C.fit

[carastro]

I have come across this problem before with two other new imagers using the APT flats aid and it to me doesn't appear to work properly.  APT is an excellent bit of software otherwise.

On both occasions I got the imagers to use the histogram and simply take a flat exposure without using the APT flats aid.  Make your length of sub short enough so that the histogram is about 1/3 to just under half way across.   

This has resolved the problem in both cases.

I must tell IVO (APT)  about this problem as he is very approachable.

Carole 

Edited April 9, 2023 by carastro

[bottletopburly]

10 minutes ago, carastro said:

I have come across this problem before with two other new imagers using the APT flats aid and it to me doesn't appear to work properly.  APT is an excellent bit of software otherwise.

On both occasions I got the imagers to use the histogram and simply take a flat exposure without using the APT flats aid.  Make your length of sub short enough so that the histogram is about 1/3 to just under half way across.   

This has resolved the problem in both cases.

I must tell IVO (APT)  about this problem as he is very approachable.

Carole 

Thanks carole i will retry later new flats with rear ota covered and see if this solves the problem ,if not i will look at the histogram .

Edited April 9, 2023 by bottletopburly

[carastro]

Might be worth checking the histogram of flats you have previously taken to see where the histogram is on those.

OK it is more faff to do it this way, but if the Flats Aid doesn;t work it is giving you even worse faff.  

Carole 

[carastro]

I just took a look at the Flat you posted earlier and there is definitely vignetting, which means the flat does not work.  

Carole 

[ONIKKINEN]

5 hours ago, bottletopburly said:

hi thanks ,yes imaging train still in situ not moved , i currently have scope covered and just reshot a dark frame and i think you may be correct about light leak from rear as the darks were showing a shadow so i just went out and covered rear with a black bag and that seems to have solved the shadow mark so i have posted a light, a dark flat ,Flat from the other night camera hasnt moved ,and reshot a dark with ota covered, a dark with ota rear not covered, and a Dark flat with rear ota covered  so i need to find a suitable cover for rear of ota images from today the 9th are the reshot ones  . 

Target was the needle galaxy and the moon was bright .

D_Dark uncovered rear OTA_2903_Bin1x1_180s__-10C.fit 17.26 MB · 2 downloads D_New Dark covered rear ota _2907_Bin1x1_180s__-10C.fit 17.26 MB · 1 download DF_4_NGC 4565_2854_Bin1x1_4.1875s__-10C.fit 17.27 MB · 1 download DF_New dark flat covered rear ota4565_2914_Bin1x1_4.1875s__-10C.fit 17.26 MB · 1 download F_4_NGC 4565_2839_Bin1x1_4.1875s__-10C.fit 17.27 MB · 3 downloads L_NGC 4565_2683_Bin1x1_180s__-10C.fit 17.27 MB · 1 download

Just went trough the files, and yes you have significant light leaks that make calibration probably impossible, even with the new files.

First your old dark which is completely ruined by the rear leak, see a rainbow rendered stretch (blues are dark and reds are bright, the light leak is clearly skirting around the primary mirror from the rear of the scope):

But your new one is not great either:

Clear gradient, which shouldn't be there. Median values in the new dark are around 460 when they should be very close to the offset value of 10 which with your camera results in 400 in 16-bit values so you still have almost 60 electrons of light leak per pixel here when just one will make calibration difficult!

In short, light leaks still reign supreme and to fight that i recommend you take the camera off the scope and shoot a library of darks and darkflats in complete darkness. The light leak issue still needs to be solved because they will imprint the weird gradient onto the light frames themselves, which may result in funny looking flats calibration even if the darks were taken in darkness.

On the exposure length and histogram, i think yours are just fine. Unnecessarily long exposures though, there is no need to expose for 4 seconds with your camera but of course also no downside other than it will take a while to capture the flats.

Your flats histogram looks like this:

Mine looks like this with a triband narrowband filter:

Mine calibrate the image just fine even though 1 of the peaks is nowhere near the center. Basically dont worry about the histogram, as long as nothing is clipped to black or white you will have good flats calibration. If you want to nitpick the flat you could push the histogram a little bit further so that all the peaks are centered around the midpoint but i dont think there is a point for doing this other than peace of mind.

In short, round 2: i wouldn't worry about the histogram at all, seems good to me. I think a lot of the flats related talk found online relates specifically to older cameras that can have nonlinear sensor responses to light and/or exposure length and so need to take some specific histogram and a long enough exposure to make the flats work (294 might be one of them). For the new Sony sensors (533,571, the fullframe version + others) you just dont need to do any tricks since the sensors are 99% linear and have no weird instability issues with any exposure lengths. I take my flats at 0.05, 0.2, or 1s depending on what gain and filter i used and never had any weird calibration issues.

[bottletopburly]

3 minutes ago, ONIKKINEN said:

Just went trough the files, and yes you have significant light leaks that make calibration probably impossible, even with the new files.

First your old dark which is completely ruined by the rear leak, see a rainbow rendered stretch (blues are dark and reds are bright, the light leak is clearly skirting around the primary mirror from the rear of the scope):

But your new one is not great either:

Clear gradient, which shouldn't be there. Median values in the new dark are around 460 when they should be very close to the offset value of 10 which with your camera results in 400 in 16-bit values so you still have almost 60 electrons of light leak per pixel here when just one will make calibration difficult!

In short, light leaks still reign supreme and to fight that i recommend you take the camera off the scope and shoot a library of darks and darkflats in complete darkness. The light leak issue still needs to be solved because they will imprint the weird gradient onto the light frames themselves, which may result in funny looking flats calibration even if the darks were taken in darkness.

On the exposure length and histogram, i think yours are just fine. Unnecessarily long exposures though, there is no need to expose for 4 seconds with your camera but of course also no downside other than it will take a while to capture the flats.

Your flats histogram looks like this:

Mine looks like this with a triband narrowband filter:

Mine calibrate the image just fine even though 1 of the peaks is nowhere near the center. Basically dont worry about the histogram, as long as nothing is clipped to black or white you will have good flats calibration. If you want to nitpick the flat you could push the histogram a little bit further so that all the peaks are centered around the midpoint but i dont think there is a point for doing this other than peace of mind.

In short, round 2: i wouldn't worry about the histogram at all, seems good to me. I think a lot of the flats related talk found online relates specifically to older cameras that can have nonlinear sensor responses to light and/or exposure length and so need to take some specific histogram and a long enough exposure to make the flats work (294 might be one of them). For the new Sony sensors (533,571, the fullframe version + others) you just dont need to do any tricks since the sensors are 99% linear and have no weird instability issues with any exposure lengths. I take my flats at 0.05, 0.2, or 1s depending on what gain and filter i used and never had any weird calibration issues.

Thanks for looking at least i can narrow down the issues and resolve hopefully any leak issues i will shoot a dark frame/Dark flats library , i need to take camera off  scope anyway as i have an OAG to fit , the steel track cocuser may also be leaking light around drawtube entry so will investigate that too .

[ONIKKINEN]

16 minutes ago, bottletopburly said:

Thanks for looking at least i can narrow down the issues and resolve hopefully any leak issues i will shoot a dark frame/Dark flats library , i need to take camera off  scope anyway as i have an OAG to fit , the steel track cocuser may also be leaking light around drawtube entry so will investigate that too .

I have a steeltrack on my newtonian too and it does leak a tiny little bit around the drawtube and around the base. The base is easy to seal with some black duct tape or whatever you have on hand. Other than that have found it to be not an issue when shooting lights so have done nothing too drastic about it. My scope is well flocked though so any light that creeps through the focuser to the inside of the tube mostly gets absorbed and not reflected back to the imaging train.

[bottletopburly]

@ONIKKINEN Currently reshooting darks /df , how do these look so far 

DF_NGC 4565_3308_Bin1x1_2s__-10C.fit DF_NGC 4565_3255_Bin1x1_1s__-10C.fit D_NGC 4565_3182_Bin1x1_120s__-10C.fit D_NGC 4565_3206_Bin1x1_180s__-10C.fit

[ONIKKINEN]

5 hours ago, bottletopburly said:

@ONIKKINEN Currently reshooting darks /df , how do these look so far 

DF_NGC 4565_3308_Bin1x1_2s__-10C.fit 17.26 MB · 0 downloads DF_NGC 4565_3255_Bin1x1_1s__-10C.fit 17.26 MB · 0 downloads D_NGC 4565_3182_Bin1x1_120s__-10C.fit 17.26 MB · 0 downloads D_NGC 4565_3206_Bin1x1_180s__-10C.fit 17.26 MB · 0 downloads

Looks very clean, these should work well.

[bottletopburly]

5 hours ago, ONIKKINEN said:

Looks very clean, these should work well.

Phew 👍 thanks .

 

 

Dave