Sorry - another M101 - and a Master Flat!

[Adreneline]

I am trying to work my way up the PI learning curve with this latest attempt at M101 - 28x300s subs plus 50 off bias, flat and dark - all at -10deg.

Any advice on colour - too red/too blue ; noise - too much/too little ; sharpness etc. would be much appreciated.

I've been chasing down focus problems for a while which I hope I've now largely resolved; I think softness in this image is down to seeing on the night - and a moon!

I've also attached my Master Flat for comment - it seems too good to be true! Does a flat like this actually achieve anything? I work very hard at keeping my optics spotless. The 50 flats were taken with the scope pointing directly at a daylight illuminated white ceiling through two layers of white tee-shirt - ~30,000 ADU.

Any advice would be much appreciated.

Thank you in advance.

Adrian

SW ED80DS Pro + NEQ6 Pro + 0.85x R/F + UV/IR filter + Atik 414ex osc (fitted with HitecAstro focusing unit and SW motor focus control).

SW 50mm finder/guide scope + IMX224

Sequence Generator Pro + PHD2 + CdC + PI

[Craney]

Your master flat has more detail than some of my better astro-photos  !!   

Great M101  BTW......

[AKB]

All flats will look like that unless the intensity scale is stretched enough.  I can't recall the process you went through to make this flat, but with an OSC you need to take care with exactly when in the operation you debayer and when/how you apply the flat.  Calibration frames, including flats, should be applied before debayering.

[Adreneline]

Thank you for the comments. Regarding the use of the flat(s) I must confess to being lazy! I use the BatchPreProcessing script in PixInsight. I just load up the flats/darks/bias and lights and basically let it get on with it - a cop-out I know but it seems to work for me most of the time - I assume the script applies the debayering at the correct time. The script produces the masters of each type. If I stretch the master flat it looks pretty much exactly the same. I've attached the PI FlatContour script output which shows some variation across the image but I am unsure whether this is good or bad or indifferent - and what impact it has on the final image.

Thanks again.

[Oddsocks]

A couple of observations,

Stretching the flat in PS shows the flat is defective in some way.

A good flat for a conventional refractor should be darker into the corners and brighter towards the centre but yours is inverted with bright corners and a dark centre, (posted below).

In your M101 image the flat is overcorrecting at the edges, the corners are darker than the centre of the image and this is down to the flat.

I am wondering if you are using a IR blocking filter with your OSC camera, this is fairly essential with a refractor because the I.R. component does not come to focus at the same point as the visible wavelengths, although you can't see it the camera does, this may explain the "soft" focus problem and the flat problem. Under I.R. black anodised surfaces become almost white and reflect I.R. strongly. If I.R. is reflecting off the walls of the focuser and camera nosepiece it will be illuminating the edges of the sensor more strongly than the centre and without an I.R. blocking filter fitted the camera will record this.

If you remove the camera, point the telescope at a bright light and look up the focuser draw tube you will probably see it reflects visible light strongly off the internal black anodised surfaces and this becomes even more pronounced with I.R.

If you do not have an I.R. blocking filter fitted try one and see if this helps, if you do have an I.R blocker fitted then look at painting the inside surfaces of the camera nosepiece and focuser draw tube with a pigment based black paint (Rustoleum High Temperature Black Barbecue Paint from B&Q is good for this) to reduce the reflections.

Lastly, take care when making flats indoors under bright lights, stray light will enter the OTA via the gap around the focuser draw tube and this may taint the flats, you need to find a way of operating the telescope in near darkness while taking the flats so that the only light entering the system is from the front objective.

 

 

[Adreneline]

Thank you for the help and advice. I am using an IR/UV filter to try to address the focussing issue. I have used a BM, the SGP focussing routines and the FWHM tool in Artemis to try to ensure I get the lowest possible HFR/FWHM readings on a given night. I am assuming it is 'seeing' that is limiting images of late; I can't remember the last time we had a really clear night

I have also analysed the images in PI and CCD Inspector (trial about to expire!) and noticed a general lack of consistency when comparing results; absolute values cannot be realistically compared although there appears to be some merit in comparing trends in the image analysis data.

I will look at taking a series of flats again noting the advice and comments given. I have had problems with light ingress via the drawtube/focusser in the past but perhaps my assumption that with exposure times of the order of 0.05s there would be no significant effect on the flat was wrong. I have noticed there is a lot of variation in the surface finish of the various links in the optics train between the IR/UV filter and the camera sensor so I will look into trying to eliminate any unwanted reflections along the light path as recommended.

Thanks again for all the advice.  

[Oddsocks]

22 minutes ago, Adreneline said:

I have used a BM, the SGP focussing routines and the FWHM tool in Artemis to try to ensure I get the lowest possible HFR/FWHM readings on a given night. I am assuming it is 'seeing' that is limiting images of late

You could well be right, we haven't had stable seeing down on the south coast for months.

I no longer use an OSC though did start off with a SXH9C many years ago and I do remember that interpreting best focus was very difficult due to the Bayer mask, the colour of the target focus star and its location in the field had a big influence. Choosing a focusing star on one of the thirds grid intersection helped, i.e. not the centre or the edge of the frame but one of the "crossing' points of an imaginary Noughts and Crosses grid.

Some swear by the Bahtinov Mask but I always found it a bit hit and miss trying to decide when the interference spikes were perfectly aligned.

Once set up and imaging remember to check and refocus every forty five minutes or so, as the environment temperature falls during the night the focus point will shift quite a large amount, check your captured image FWHM scores from early in the run to those at the end, they should be the same, any change not attributable to seeing will be down to temperature induced focal shift.

If your focusing motor, focuser and control system is able to drive to a repeatable point then autofocus using V curves works well for OSC since it takes the guess work out of where the best focus point lies but obviously this is useless if the focus motor is unable to return the camera to the exact best focus point on the curve after analysis.

 

[billyharris72]

I think that's a really nice image Adrian, and I like the colour balance, but Oddsocks is dead right about that flat. It's almost like it's back to front. What does the image look like processed but minus the flat? I suspect it may be doing more harm than good. Assuning PixInsight has a gradient / vignetting filter that might give you a better result.

Out of curiosity, what method do you use for actually shooting the flats?

Billy.

[Adreneline]

2 hours ago, billyharris72 said:

Out of curiosity, what method do you use for actually shooting the flats?

Billy.

Hi Billy.

Attached is the configuration I use for flats - iPad running a 'light-pad' app sitting on the end of the OTA. I adjust the exposure in SGP to get a mean of around 30000. I tape some thick black plastic rubble sack around the drawtube (but not the focusser) and fire off 50 flats at a temp of -10deg - typical exposure time per flat is around 0.18 s. I assumed taking them in the daylight was acceptable as there is plenty of advice saying to use the dawn sky, etc. for illumination through a white cloth.

I've just repeated the process and obtained the attached flat - with contour from PI.

Any advice would be most welcome. I think I might be heading off to B&Q!

AP is a tricky old business!

[Oddsocks]

On 04/05/2017 at 18:22, Adreneline said:

I assumed taking them in the daylight was acceptable as there is plenty of advice saying to use the dawn sky, etc. for illumination through a white cloth.

 

On 04/05/2017 at 18:22, Adreneline said:

Any advice would be most welcome. I think I might be heading off to B&Q!

The latest flat is still not right.

The corners are still brighter than the centre.

The brightening is very subtle and it could only be reproduced here in the SGL forum using a very strong gamma curve in PS using a 16bit format. I found it easy to show live on screen using levels but near impossible to save and show here using the native 8bit format your original image is posted under, so you are literally chasing very faint ghosts.

Looking at the length of the coupling between camera and OTA I would suspect internal reflections from the walls of the various adaptors.

Do try sky flats again though just to eliminate the iPad from the equation (though it should be fine, I use an iPad for my mobile setup and they work well as long as a diffuser of some sort is used between the iPad and the OTA).

For sky flats these are taken with or without a diffuser between half an hour and an hour after sunset or before sunrise, the surroundings should be dark while the sky just retains some glow and the stars can not be seen. The diffuser for sky flats does not need to be thick, a single sheet of 80gsm printer paper works well, if using a diffuser the mount can be stationary and should be pointing at the area of sky opposite to the Suns current position below the horizon to avoid gradients. If taking without a diffuser the telescope should be pointed to the same area of sky and moved between exposures to ensure any barely visible stars captured are averaged out in the master flat. The focus position used for flats must be exactly the same as the focus position for the lights otherwise they won't work.

 

[Adreneline]

Thank you for all the advice and the time taken to provide me with some potential ways forward. At the end of the day if I am going to use flats in the calibration process I may as well do my utmost to get them right!

The R/F has a 2" matt black adapter nose-piece courtesy of FLO which holds the 2" IR/UV IDAS filter in the drawtube. Downstream of the R/F is a Baader Varilock which has a matt black internal surface but whether it is matt black enough remains to be seen. The VariLock attaches directly to the 414ex camera body; the 414ex nose has been removed. I have to say it's not obvious where the problem may be as I know the inside of the drawtube is matt black.

I will give the sky flats a go as well and see how I get on.

Many thanks again for all the helpful comments.

Adrian

[Oddsocks]

Have a read of the this linked document, if you have not seen it before it provides some insight, it is a well written explanation of the issues surrounding poor flats.

http://diffractionlimited.com/flat-fields-stray-light-amateur-telescopes/

William.

[Adreneline]

On 5/4/2017 at 15:45, billyharris72 said:

I think that's a really nice image Adrian, and I like the colour balance, but Oddsocks is dead right about that flat. It's almost like it's back to front. What does the image look like processed but minus the flat? I suspect it may be doing more harm than good. Assuning PixInsight has a gradient / vignetting filter that might give you a better result.

Out of curiosity, what method do you use for actually shooting the flats?

Billy.

Hi Billy. In answer to your first question I've attached the image with and without the flat - both processed in PI using exactly the same workflow. I've had numerous further attempts at producing flats and they always come out the same. The feedback I had from a CN forum was "if that the way they are then that's they way they need to be to correct errors in the optics"; it's all a function of sensor size and optics train.

The differences are pretty subtle I feel.