Help with vignetting, TaK 106N and Atik 383l+

[nucdoc]

 

I am using a TaK 106N, with Tak spacers to attach the telescope to my Starlight Express filter wheel Using 2 inch astronomik  Deep sky RGB CCD filters.  The camera is an Atik 383L +.

 

This is only one hour (5 minute subs) for each channel, I was hoping to get some questions answered

 

(1) When I shoot luminance I use an Astronomic CLS light pollution filter for this channel.  Do you need a light pollution filter in the imaging train for RGB? (i.e. two filters)

(2) I worked out the spacing with the help of Tak USA, is it usual to get this much vignetting? or have I made an error with the spacing from the back of the scope to the camera?

(3) Flats do not entirely remove the vignetting, I used the Flats wizard in SGP, stipulating 42000 +/_ 500 ADU, with 3 second minimum to avoid shutter artifact.  I have a Flats panel, and given the speed of the TAK I needed 3 pieces of paper between the panel and the scope.  Less saturated flats ~30,000 ADU did a much worse job of cleaning up the image

 

All processing in PI

 

 

One hour each of RGB

 

 

Single Red Sub

 

 

Flat for Red Filter

 

[swag72]

I use a LP filter as my luminance filter and so the RGB gets no LP protection at all. 

Regarding your flats here's a couple of observations. 

I set my ADU at 35500 as per the QSI recommendation for their 683 (using the Kodak chip) - Your ADU seems very high to me. Then there's something that I've not yet had the opportunity to try out but may be worth checking....... In SGP it MAY help if your lights are taken at a slower readout speed than your flats. I say this as Steve Richards discovered this issue when using Maxim. Whether it's the same for SGP I'm not sure as I've not tested it yet.

Can you perhaps explain your calibration procedure so the people in the know can ascertain that the flats, bias and darks are being subtracted at the right point.

[ollypenrice]

There is no spacing to worry about with this system, you just focus - which clearly you can do so that's all good. (Spacers come into the story with focal reducers but the FSQ106N cannot take the focal reducer. Being a Petzval it has an inherently flat field.)

In your train it is important to have the FW right next to the camera. Any spacer or adapter separating these two will push the filter up into a wider part of the light cone, so try to mount the camera bang up against the FW if you can.

You should have no problem at all with vignetting on your outfit. We have a pair of these scopes on a dual rig and are using the much bigger Kodak full format chips, one with 2 inch mounted filters and one with 2 inch unmounted. The unmounted filters vignette less but both vignette, though well within the power of flats to correct. The centre to corner drop-off is about 4000 ADU with the mounted filters. None of this is the fault of the scope, it is our fault for not spending thousands more on bigger filters! However, we image very happily with this setup and it's a very well known scope-chip combination.

Rather than looking hard at a stretched flat (because the purpose of stretching is to exaggerate contrasts) just check your ADU values in the middle and the corners to check the drop off. You shouldn't be anywhere near the 4000 that we have.

Flats. Are you subtracting a master bias (in effect a dark-for-flats) as you make your master flat? This is important or your flats will probably over-correct. You could take 'real' darks for flats but at short exposures they will be no different from bias and take forever.

What we are seeing here is not just vignetting but colour-specific vignetting. Markarians seems to have much more vignetting in red. Could this be fogging of the camera window rather than vignetting? Or was there anything very different about the red flat? Maybe try using a green flat as a flat for your reds to see if that works better? (I've often used L flats for everything without any difficulty whatever.)

Olly

[steppenwolf]

Although I don't think it is related to your particular issue, just for clarification:-

Quote

Then there's something that I've not yet had the opportunity to try out but may be worth checking....... In SGP it MAY help if your lights are taken at a slower readout speed than your flats. I say this as Steve Richards discovered this issue when using Maxim.

I had an over-correction issue with Flats on my QSI 683 (same sensor as yours) and carried out some comprehensive research on the issue and resolved it. The summary of which was:-

1. Lights captured in fast readout mode are over-corrected by Flats captured in slow readout mode

2. Lights captured in fast readout mode remain over-corrected by Flats captured in matching fast readout mode

3. The underlying deciding factor is the readout mode for the Lights but to ensure the best correction, it makes sense to use slow readout mode for both Lights and ALL your calibration files

[Uranium235]

The OP states that he is using a 3 second sub to avoid the shutter. With a 383, IIRC (from the manual) the recommended minimum time is six seconds - and I usually aim for 10 just to be on the safe side. Also, aim for a much lower level - about 22,000 (maximum, not average) so the panel will need to be dimmed further.

It may go against the recommendations of others, but with this camera on an f5 flatfield refractor - it works for me.

[nucdoc]

Thanks Guys for the insights, very useful.

 

Sara - I use a pixinsight tutorial that I found on the 'Trapped Photons' blog.  I have a bias master made from integrating 25 bias frames, and I use this to calibrate each flat sub, they are then integrated into a flat master.  I calibrate each light sub with the aforementioned flat master, as well as the appropriate Dark Master.  The lights are then registered and integrated.

 

I stated with 30,000 ADU like I has used on my SX-9, but this did not work very well, I read a thread in one of the forms where the op used 42000 ADU on the Atik 383, can't recall right now where that was.  I will take the advice of Red Dwarf and shoot new flats at 22,000 and with 10 second duration and repeat the processing

 

Olly, you are right, serious Red vignetting much worse than Blue or Green, or even Lum.  Not sure why, the filter looks pristine and is brand new.  Perhaps it is my high ADU flats?  I will try a 22,000 ADU and see if the issue with RED flat persists.  Yes, the camera is screwed right next to the filter wheel.    The current flats have a drop of of 4500-5000 ADU, so I will reshoot them as I mentioned above.

 

Again, thanks for the suggestions I will put them to work and see if things get better

 

Mark

[nucdoc]

 

So, I created new flats, using and ADU of 22,000 and an exposure time of 10 seconds, the drop off from centre to edge is 3000 or less, I still have a tinge of red, but much less, and I think this looks much better.  I will try to use my green flat to calibrate the red subs and see if it looks better

 

Thanks again

 

Mark

[carastro]

When I used an Atik383 I always aimed for around 25,000 ADU and found this worked just fine.  

Of course always the problem of the shutter, I found that 3 seconds was enough for that and dimmed the light down accordingly.  Different filters require different length flats, so I varied the length of the flats/number of sheets of paper depending on which filters I was using, NB requiring longer time.  

I must confess I have never used a LP filter with my CCD camera even at home in my LP location.  I probably should, but always forget.  

Carole 

[ollypenrice]

Mark, do you use Dynamic Background Extraction in Pixinsight? You want to fix your flats of course, but what you have there is now well within the power of DBE to correct perfectly.

Olly

[nucdoc]

Yes, I perform DBE, for this setup the division function produced better results than the subtraction option, I have a street light about 50 m away, below the part of the sky I am imaging.

 

I usually perform it once on the final calibrated, registered, integrated light for each channel and then again after RGB combination.  In the attached image I used the new flats, obtained at 22,000 ADU, and used the green master flat to calibrate the red subs, I then did a 3rd DBE on the final image, I think it finally looks ok

 

Thanks again for your help

 

Mark