Revisiting M57 and M13 lucky imaging. After the CEM70 was fixed

[timh]

For about a year now I have been interested in trying to push the boundaries of angular resolution (low FWHM) in images of some of the brighter deep sky objects via semi-lucky imaging using short (<~3s) frames at high gain (low read noise).

With an 8 inch telescope at an image scale of about 0.5 arcsec/pixel the technique does work but there have been practical problems to solve along the way. In my case - always using fast Newtonians - some key if obvious general lessons were ..

1) Maintain near perfect collimation/ alignment (Hotech laser helped) and focus
2) Loosen off mirror clamps after purchase
3) Go for bright - or bright parts of objects nearly overhead

But it took a long time and many wild goose chases to understand that the mount (an Ioptron CEM70) was also presenting a subtle problem that limited resolution (and star shape). viewtopic.php?t=5467

Anyway, to cut a long detective story short - I presented the evidence to the vendor - First Light Optics - who were superb. They fully engaged with the data and offered a replacement. So how much better does lucky imaging work with the mount now fixed?

SW200PDS Newtonian (f = 1000mm, F5.0) Baader MkIII coma corrector, CEM70 Ioptron mount, Baader steeltrack focuser, Pegasus Cube2 focus controller, PHD2 multistar guiding using an ASI 120 mm guide camera and 80 mm SW startravel refractor at f = 400 mm.

ZWO AS1294 MC camera for RGB captures, 4.63 um pixels (0.953 arcsec/pixel) at - 5 C
ZWO ASI294MM mono camera for LUM, 2.315 um pixels (0.477 arcsec/pixel) at - 5 C

ZWO IR/UV cut filter

M13. 110 x 20s RGB exposures at gain 124 and 546 x 2s MONO exposures at gain 285.

M57. 137 x 20s RGB exposures at gain 124 and 926 x 2s MONO exposures at gain 285.

All frames captured under Bortle 6 skies - moon low in south. All pre-selected for quality using the FWHM and brightness filter within Sharpcap, darks and grayscale master flats (no bias) prepared using Sharpcap.

Preprocessing and processing in PixInsight. The linear stacked 2s mono images were both relatively sharp (FWHM ~ 1.65 and 1.48 for M13 and M57 respectively) and after background subtraction and noise reduction (MLT) were each used as luminance for the lower resolution RGB images. The images were then adjusted with curves and in photoscapeX. In the case of M57 the 2s mono image was also deconvolved (50 iterations) in PixInsight which bought out yet further detail.

Overall with the new mount it all seems to have worked very well. A definite improvement - enabling the generation of short frame images with 2s exposures which can be stacked to provide not only good sharpness but also quite acceptable levels of SNR for these brighter objects. I was particularly pleased with the level of detail now evident in M57 and the fact that the resolution was even high enough to split a very difficult (FWHM ~ 1.5) line of sight triple star which sits just above M57.

Tim

[admin]

Hi Tim,

good to hear that you got the problems with the mount resolved (although in a way it would have been interesting to find out just *what* was causing the high frequency wobble). Certainly really good sharp images with the new kit

cheers,

Robin

[timh]

Thanks Robin,

A difference between the early and later CEM70s is apparently in the RA and DEC stepper motor controller circuitry. It may be to do with that rather than anything physical but I am not sure.

Anyway for sure very pleased that the set up now works so well and also still surprised that it is possible to push deep sky resolution quite so far with exposure times that are really only quite modestly short?

Tim

[turfpit]

Nice images Tim, I particularly like the M57. Thanks again for a detailed write-up of the processing.

Dave

[timh]

One thing that I didn't include in this thread that Dave (turfpit) suggested might be useful are pictures at different stages of the process so that it is clearer where the main improvements were.

Using M57 as the example ...all images at ~ 0.477 arcsec/ pixel

A is the 137 x 20s OSC image (X2 drizzled from 0.953 arcsec/ pixel)

B is the 926 x 2 s MONO image

C is the 926 x 2s MONO image after deconvolution


The starting OSC image, A, is reasonable (FWHM ~ 2.3) but the stars are mis-shaped as if there had been some movement at some point. I didn't worry too much about this though since it was only being used for chrominance

The 2s mono image, B, is much sharper (FWHM ~ 1.5) than A and the triple star is resolved

Deconvolution of B produced a further big improvement resulting in image C

In the final image the luminance of C is transferred to A. Thereafter there are endless possible variations for altering the stretch using Curves, contrast, colour saturation etc - according to taste really.

Tim

[ChrisR Oz]

Inspirational result Tim. And such good final resolution for such a small (3 arc-min) object.
Cheers, Chris.

[turfpit]

Thanks Tim for following up the suggestion. The stars within M57 of the A+C image really demonstrate how the processing steps have improved the final result. The B -> C comparison shows the benefit of deconvolution. Well worth the pain and time that has gone into the learning I think.

Dave

[timh]

Thanks Chris and Dave. The CMOS cameras and processing tools - especially within PI - that are now available are amazing. Especially deconvolution which has been a bit of a revelation. Tim