Options for high-resolution imaging?

[Dr Paul Leyland]

My SBIG-8 is a fine though elderly device but has its limitations, principal of which is that the minimum exposure time is 0.11 seconds.  I’m wondering whether I should replace it and would welcome comments and advice.  First some technical specs.

The 9 micron pixels of the SBIG-8 corresponds to 0.72 arcsec on my 0.4m f/6.5 scope, whereas the diffraction limit is about half that.  Although the long term (>= 0.11s) resolution is seeing-limited to perhaps 2as, it would be nice to be able to use lucky imaging and deconvolution to attain effectively diffraction-limited performance.  The Nyquist criterion suggests that a resolution of 0.2as/pixel (ie. diffraction / 2) should be optimal. To achieve this either the focal length needs to be increased (a 4x Barlow, for instance, with 9 micron pixels) or the pixel size reduced to 2.5 micron or so.  That figure is uncomfortably close to the wavelength of light and so a compromise is very likely to be necessary. There are severe trade-offs when small pixel sizes are used.  Either the FOV is seriously curtailed or the detector costs serious money or both.  In either case, the SNR may be severely compromised.

My principle interest at present is VS photometry where a FOV of, say 15 arcmin, enables BAA and AAVSO comparison sequences to be used and where detector dynamic range, noise and (especially) linearity is critical.  A field of 15 arcmin at 0.2as/pixel corresponds to 4500 pixels — a seriously expensive detector with (very probably) lengthy download times.  However, I also want to branch out into high resolution imaging of objects beyond our immediate neighbourhood, here defined as anything further away than Saturn.  A number of people have imaging the inner solar system down to a fine art and I see little purpose in competing with them.  That said, their skills and experience are very relevant to my ambitions.

Finally, and unless you can convince me otherwise, I’ve no use for colour cameras. On the odd occasion that I want to take a pretty picture I’ll combine several images taken through several filters.  A colour camera is just a waste of resolution and/or sensitivity as far as I am concerned.

Any suggestions on what cameras may be an acceptable compromise  (and compromise is clearly necessary)?  I’ve no hang-ups on whether CCD or CMOS detectors are used now that CMOS detectors appear to be getting their noise issues under control.

Perhaps it’s not possible to do both VS observations and high-resolution imaging with the same gear and I’ll have to resign myself to swapping Barlows and/or cameras in and out according to what is to be imaged.  It’s just that retaking flats after each exchange is so tedious…

Thanks in advance.

[Dr Andrew Smith]

Hi Paul, I would advise you look at what the best images use in the categories you mention.

I am not an expert but a few points.

Modern CMOS cameras can quickly download large numbers of pixels with USB 3

Nyquist does not really apply to area detectors like CCD and CMOS cameras and greater oversampling is needed.

If I remember correctly planetary imagers go for about 0.1as.

Regards Andrew 

[Dr Paul Leyland]

The SBIG died unpleasantly and after some deliberation I purchased a Trius 814 from Starlight Xpress.  There were many many teething problems but first light came last night.  Images are still not as pretty as I would like but one, of a beautiful face-on barred spiral in Lyra called IC 1296, appears below. The bright nucleus is over-exposed to bring out the detail in the spiral arms. I’m certain much better images can be obtained after more tweaking the configuration of various bits of hardware.  In particular, I don’t understand what generates the dotty artefacts visible in the image, and there’s some trailing, possibly because the OAG and AO units are not working properly (or at all).

According to a local installation of astrometry.net the plate scale is 0.292 as/pix which matches the theoretical resolution nicely and at 2×2 binning (0.584 as/pix) a typical 2-3as seeing disk is 4-5 pixels across.  Just what I wanted.

(OK, I confess to being mischievous.  The large blobby thing at the lower-left is M57.  I happen to think the 15th magnitude galaxy is prettier.)

Added in edit: I just spotted 2MASX J18530959+3305385, the faint fuzzy just to the left of the top-most star on the right edge of the image.  According to SIMBAD it is 10as across and rather red; I can’t find a V magnitude but guess it’s around 16-17.

[Robin Leadbeater]

Hi Paul,

They look like trails of hot pixels. Is this an aligned sum of several individual exposures? 

(IC1296 brings back memories. It was a test of how deep I could go with my first astro camera, a surveillance  camera which I modified for long exposures in 2002).

http://www.threehillsobservatory.co.uk/astro/1004xcam.htm

The resulting LRGB image of M57 even made it into Astronomy Now. Cutting edge then !

[Dr Paul Leyland]

You are almost certainly correct.  SWarp was used to co-add a number of subs.  As the intention was to acquire an image, any image, no biases, darks, flats, bad pixel masks or anything else were used.  The filter wheel was playing up and I can no longer remember what, if any, filter was used.  The FITS headers claim no filter, which seems plausible, but I wouldn’t swear to it.

[Dr Paul Leyland]

“It was a test of how deep I could go with my first astro camera,”

I’ll take that as a challenge.  The old camera could reach  22.0 Gaia -g, a good approximation to the SBIG’s unfiltered spectral response, with a reasonable exposure time.  When all the niggles have been worked out I’ll try for the H-II regions in IC 1296 and a better view of the more distant galaxy.

[Author]

Posts