Flats problem - help please. Atik 16200

[Skipper Billy]

New camera on existing scope, filter wheel and filters. No problem with Atik 460 - the only thing that has changed is the camera - it is now an Atik 16200 - mechanical shutter. Not been used yet so I am taking advantage of cloud and rain to build a flats library.

Attached are 3 flats - one Lum binned 1x1 0.33 seconds, 1 x Red binned 1x1 1.67 seconds and 1 x Ha binned 1x1 42.07 seconds - all yield an ADU value of ~30000. The images have been stretched and converted to Jpeg.

To me the Lum shows evidence of 'shutter shadow' (although Atik suggest no shorter exp than 0.2 seconds - this is 0.33) whilst the Red and Ha show evidence of uneven light source and/or vignetting. The latter two - particularly the Ha show vertical lines / crosshatch pattern.

I am using an Atik EFW3 with 2" filters so I wasn't expecting vignetting.

The light source is a Huion flat panel with the voltage dropped down to 9v to make it dimmer and even then it was too bright so I have put 4 sheets of typing paper over the scope dew shield.

Any ideas ???  Maybe the paper is causing the issue?  Maybe the light panel doesn't like being dimmed?? Maybe I am worrying about nothing but they are horrible compared to the flats from my Atik 460!!

Any suggestions most gratefully received !!

[vlaiv]

I doubt that pattern is from paper - it's unfocused light after all and nothing that close to objective lens, or aperture of the scope will be anywhere near focus do display any features.

It might as well be "feature" of the sensor - manufacturing issue with micro lens - small variation in QE that displays like pattern. I've seen it on my camera as well - but that is CMOS - much more likely to happen since CMOS has integrated circuits next to each pixel.

One way to be certain is to do flat calibration and see if it calibrates out ok.

As for mechanical shutter - way of dealing with that, as far as I know is to use longer exposures for flats - that again means dimming your flat panel further. Ha one certainly does not show shutter issue since it's ~42 seconds long. I guess you need at least couple of seconds for each flat sub to minimize impact of mechanical shutter.

[Laurin Dave]

Hi David..  they look very similar to mine from my SX-46 with the 16200 sensor..  I do use longer subs though 5-10 secs for LRGB and Ha to ensure there’s no shutter shadow. .  involves lots of sheets of paper and fiddling about with a  Huion. I also use Pixinsight superbias..  it all calibrates out nicely though, had me worried too as I did the same as you waiting for the skies to clear 

Dave

[Skipper Billy]

Thanks guys

I have been playing - taking the Huion panel voltage back up to 12v made no difference

Taking the paper out of the equation got rid of the cross hatch pattern though!! Very weird - its right in front of the objective and cannot be anywhere near in focus!

I am just messing about with the panel brightness and alternative light sources.

I am now finding that 4 seconds with Lum equates to 10 mins with Ha  - that cant be right!!???

I have ordered some proper 4 stop neutral density film and I am off to bed - work tomorrow ?

Another day - fresh eyes and brain etc ......

 

[vlaiv]

10 minutes ago, Skipper Billy said:

I am now finding that 4 seconds with Lum equates to 10 mins with Ha  - that cant be right!!???

No, you don't have to do everything proportionally - just get longer Lum and RGB flats, leave Ha as is, or maybe even shorter exposure if you remove all paper. Aim to get dimming of flat box so that your flat exposures for each filter are 5s or more. You want the shutter closing time to be very small compared to total exposure time (1/10000 of total exposure time of less).

This means that you will have to dim flat panel more when taking Lum and RGB than when doing Ha.

10 minutes ago, Skipper Billy said:

Taking the paper out of the equation got rid of the cross hatch pattern though!! Very weird - its right in front of the objective and cannot be anywhere near in focus!

Pattern is there, but seeing it depends on how even illumination is on your flat - depends on stretch. In first post, you don't see it in your Lum flat, but it's there. You don't see it because variation in brightness of pattern is less than 1%, and the rest of flat varies more considerably due to shutter issue, so pattern is very low contrast. Here is Lum sub from the first post stretched without regard for top part of frame and to bring out flat features - dust shadows and also pattern:

[ollypenrice]

I'm very surprised by Atik's recommending a minimum of 0.2 secs. Could this be a typo on someone's part for 2.0 secs which would strike me as more likely for a shutter camera?

I don't think the flats look particularly bad. There is a huge difference in chip size, meaning more opportunity for vignetting. The cross hatching cannot be from the paper because it is orientated along the axes of the chip so must surely be electronic in origin. Vlaiv has demonstrated that it's there in all cases.

The first thing I'd do is see if they work. If you've no lights to try them on it might be instructive to take a set of flats at different exposure time and then apply these to the earlier set to see if 'flats flatten your flat.' If they do then the flats will probably work.

Olly

[Skipper Billy]

Thanks all.

I have ordered some 4 stop neutral density film and I am going to start again from fresh.

The 4 stop film is grain free even though that shouldn't make any difference.

The flat panel is back at 12v in case the drop to 9v is causing any flickering and I am going to adjust the panel so its absolutely square to the scope to try and get rid of the gradient.

I think I will make two scope masks - one with enough sheets of ND filter to give 4 - 5 secs Lum and use it for the other broadband filters and one with fewer (if any) sheets for the narrowband filters.

Meantime - seeing if the flats flatten the flats is a great idea !

The shared collective wisdom is very much appreciated !!

 

 

 

[Laurin Dave]

Hi

This is what one of my flats looks like, very similar to yours and they work fine. Well, at least they seem to. 

HTH

Dave

 

[davew]

I've seen this before and think it's a property of a Full Frame chip ( As in KAF not DSLR !) that showed up on my and other's KAF 8300 chips. There was some discussion about it at the time but they calibrate out just fine. 

I used to use an exposure time of around three seconds minimum.

The hatching I had can't be seen in the final image,

Dave.

Well well, I found an example ?