Image Scale Question - Newbie
Posted 21 October 2012 - 09:12 AM
I want to image double stars in order to measure their separation and Position Angle for eventual publication in the Journal of Double Star Observations.
I will be using either a 6" f/8 refractor (resolving power around 0.75") or an 8" f/8 RC (resolving power around 0.5")
I don't expect to successfully measure below 3 or 4 times the resolution limit of the scope. Unless you all tell me it is possible.......
What I'd like to know is what is the best "arcseconds per pixel" for this type of imaging as it pertains to the resolving power of the telescope.
I calculated the 6", at 1200mm will have 0.74 arcsec per pixel. I can barlow this up to a higher level with a barlow, I suppose.....
If I purchase the 8" I'll get 0.55 arcsec per pixel with no barlow and up to 0.11 arcsec per pixel with a 5x barlow.
All I need is enough image scale to get good centroids and enough separation for the software to utilize.
Your thoughts will be appreciated.
Posted 21 October 2012 - 08:25 PM
I see your point about the great detail being captured but double stars will present a problem not seen in planetary imaging. They will be, I assume, fainter on average, necessitating longer exposures. Longer exposures make you more vulnerable to bad seeing. Also, a lot of the detail in planetary images is capture in the R or IR channels. I suppose you can use an R or IR channel for the stars - not sure of that. But B and even G channels are often much, much fuzzier.
Also, the most detailed images are coming from larger apertures - 11-14" generally (the C11 and C14 being among the most common instruments in this and other forums).
I use my Obsession 15 - you might consider using your Teeter. My guess is it will get you much better resolution than a 6 or 8 inch refractor and you'll get more light, keeping exposures down.
But these are some in-expert thoughts. I hope others will chime in if they have a more quantitative idea.
Posted 22 October 2012 - 10:25 PM
Does anyone have any data on this topic?
Thanks, in advance
Posted 22 October 2012 - 11:07 PM
Now to business...
Although interested in binaries for some time I have not done much in imaging them. I have imaged some just for fun, as well as used them to calculate with some precision the plate scale. Problem there is getting up to date information on the latest measures as most of the interesting close ones vary fast enough, and even then sources differ on the measures in some cases.
Ideally you want enough measures to do statistics, so I think its a lot of work. But hey, thats who we are!
As far as plate scale I think you want to image at similar f ratios that many of the planetary imagers do for a given pixel size. So for the common 5.6u pixel in the DMK21 and Flea3 618 type cameras that would be around f/20 to 30. Of course for wide separation you might have to drop down. You will need a standard measure somewhere to calculate the actual plate scale, and for your application that will be key.
If your using your Canon DSLR I might suggest doing it in green light if seeing permits. Half of those pixels are green and its fairly easy to extract the green from the raw images. And those pixels are small so maybe f/15 would be better? Not sure. If using a refractor I would recommend using a green or red filter. I've tested lots of refractors and most of them seem to be best corrected in green light. In good seeing the airy disk of a star in green light will be smaller and hence more accurate for really close doubles. You might begin in IR if you have a mono camera to discover best techniques to track and stack 2 (or 3?) points. Often the seeing is so much better in IR its amazing - but you do take a hit in resolution.
You probably know already that it would be important to expose a trail, drive off or slewed slowly, to get a fix on the orientation of the field for calculating the position angle.
Hope this helps. Since your experimenting anyway you might try several things with these points in mind. Your technique I think would be an important section in your paper.
Edit: For really close binaries, I would'nt hesitate to crank up the scale to f/50 or so. One thing I forgot to mention is that its important not to saturate the star, or the relative intensities of the airy disk and rings will be compromised.
Posted 22 October 2012 - 11:24 PM
I found an IR image of Castor I took a year ago. My scope was not collimated precisely yet, but the image is good enough to do a measure. I calculated a plate scale of 0.069 asec/pixel. Interesting side note - I used a microscope objective to get this projection!
Posted 23 October 2012 - 10:06 AM
I think you will want to oversample just a tad... Nyquist says 2 pixels per resolution element, so you might want to go for 4... which means imaging around F30 or so (usually). Too high and you loose image contrast since the light is spread over more pixels.
I'd start at prime focus with a good known double, then move up to a 2x barlow. If it was me, I don't think I'd go as high as 5x.
Cheers and good luck.
Posted 23 October 2012 - 01:00 PM
The formula is: I = 206 x P/F where
I is image scale in arcseconds per pixel,
P is pixel size in microns
Fis focal length of the scope/lens.
Nyquist states that the minimum sampling is to have two pixels to cover the scope's theoretical resolving power. For my 6" scope, with a resolving power (Dawes) of 0.75 arcsec the image scale for each pixel would be 0.375" as a minimum and, for 2x oversampling as you suggest the image scale per pixel is 0.1875".
My pixel size is 4.2 microns so, crunching the numbers for the minimum we get:
0.375 = 206 x 4.2/F
so F= 2307mm which is f/15.4.
For the 2x oversampling F= 4614mm or f/30.8.
I have a 2x Barlow so I'll start there. The 2x oversampling would require a 4x barlow which I don't have at the moment....
Posted 27 February 2013 - 10:48 PM
I have no idea how this compares. But I thought it may be of interest on this topic, if you're still interested, Dave.
Posted 28 February 2013 - 11:19 AM
If you apply more than 2x the focal will reduce the brightness of the stars and will cost reach of magnitude fainter stars, but the measure will be more easy.
I have a project closed double measuring <1" with a Newton 14" and QHY5LII, using this approach I can not measure stars of magnitude fainter from 10 to 10.5 and helped me a filter that improves seeing 21A much, so I can capture from 200fps to 30 fps depending on the brightness of the star and filter images in Registax.
You find the problem can not be measured very closed doubles when the brightness difference between them is very great, in my case it is impossible if the difference is> 2 magnitudes.
With only 8" i think you can measure about 0.6" with good night seeing and magnitud 8.
I'm capturing double stars 0.5" measurable, others up to 0.4" look like doubles but very complicated to measure without a big mistake.
Posted 28 February 2013 - 02:49 PM
My recollection on image scale was that I was operating around 0.09 arcsec per pixel. I was trying to see if I could detect duplicity in very close doubles (half arcsec or closer - i.e. inside the Dawes limit). As I recall, I was successful at least some of the time. Seeing was everything though, as you might expect. I'll have to dig up a few of the captures I made and look again.
Posted 28 February 2013 - 02:59 PM