Flats and dust - help !

[halli]

Hi

It seems to be virtually impossible to keep a CCD optical window clear of dust !  I guess the usual practice is to use flats to get rid of dust doughnuts, as well as helping with gradients etc...

I have an Atik  383l mono which I acquired recently so I embarked on getting a LED panel to take better flats.  After a lot of messing around with printer paper to dim the source enough to take flats at an exposure of around 10 seconds (avoiding shutter shadow which caught me out to start with !) with the correct ADU - ths was with the telescope in doors- but with the imaging train undisturbed from the night's imaging.  I thought I had obtained workable flats.  I was wrong !   The flats had introduced even more dist bunnies as they appeared as light doughnuts on the calibrated images but the original ones were still there !  I wondered whether the reason was that the dust had moved before I took the flats.   

I therefore decided to keep the telescope on the mount and use the traditional method of using a white t shirt the following morning  - except it was two layers of white towelling to get the right ADU and exposure length !  The flats appeared fine and showed up the dust !  However I again had the same result when processing with DSS - existing image dust remained and new dust was introduced by the flats.  I superimposed the flats on the image and it confirmed the dust bunnies didn't align.  Was the dust on the filter wheel and therefore perhaps due to the wheel revolving the dust had moved ?  i did  a calculation based on the diameter of the doughnuts and the distance came out as 10mm ie on the CCD optical window not the filters.   

I am therefore  at a loss to explain the flat and image dust misalignment as I cant see that fresh dust had been introduced to the CCD window before taking flats as I did not disturb the imaging train - am I missing something really  simple here ?  I need to resolve this as the dust is ruining my images !

Any suggestions welcome.........

Thanks in anticipation

[Davey-T]

Check that the flats aren't being flipped or rotated by software, try flipping one manually and see if the doughnuts line up.

Dave

[halli]

Thanks Dave - I think you cracked it !   I had looked for a line of rotational  symmetry when I superimposed the images obviously not close enough though...........

I looked a bit closer after your suggestion and flipped one of the images and looks like that is what  caused the problem !   Ive just realized that I used Artemis for capturing the flats as it gives an easy ADU reading and I used APT for doing the image capture.  This must be the reason for the mirror image between the lights and flats.  I must admit I was doubting my grip on reality.......... 

Brilliant - I can look forward to imaging again instead of thinking of attacking the CCD with a piece of microfibre cloth !

Thanks again  - you have at least achieved one good deed more than me today !!!

Ian

[andrew s]

Ian, in general I would use the same software for all captures, light, darks etc. The software adds a "pedestal" to the CCD output and this can differ between capture packages. This should be correctly removed by the processing software but if programming shortcuts were made it may assume they are all the same. You can see the value in FITS Header of the image.

Regards Andrew

[halli]

Agreed Andrew - I'm sure there must be a good way of reading ADU values from APT.  Only problem  with Artemis is it doesn't seem to support dithering with PHD, apart from that it is very stable and provides what is required

Cheers

ian 

[pyrasanth]

On ‎22‎/‎02‎/‎2016 at 09:27, andrew s said:

Ian, in general I would use the same software for all captures, light, darks etc. The software adds a "pedestal" to the CCD output and this can differ between capture packages. This should be correctly removed by the processing software but if programming shortcuts were made it may assume they are all the same. You can see the value in FITS Header of the image.

Regards Andrew

I completely agree with this. I made a set of calibration frames in CCDstack which worked fine. I then used these with PixInsight & used them in the batch processing script. The results were really bad- just awful calibration. I then made a new set in PixInsight & the results were just fine- so definitely not compatible between applications.