How to focus manually a lens for which it was not designed.

[fivemack]

Originally Posted by fivemack

the point at which the focus ring on the lens stops rotating is a couple of degrees beyond the infinity focus point

I was thinking of a different lens for that; on this lens, the focus ring can be turned as far as you like in either direction, though it feels slightly less free-moving once you're about ten degrees beyond infinity focus. Turning the focus ring changes the focus even when the camera is turned off, so there must be some mechanical coupling there.

Full image at http://www.chiark.greenend.org.uk/~t...pella.full.jpg

Full image in Nikon semi-proprietary lossless-compressed format (I think there exists free software that can read it):

http://www.chiark.greenend.org.uk/~t...pella.full.nef


There is nothing terribly interesting in this field; I was using Capella as a point source to focus the camera, and thought I might as well take a longer exposure to see what the limiting magnitude was like - it's greater than 10, and I don't have a good enough star chart to figure out how *much* greater than ten. Looking at it closely, there's quite obvious chromatic aberration at the edges of the field.

[xilman]

Originally Posted by fivemack

There is nothing terribly interesting in this field; I was using Capella as a point source to focus the camera, and thought I might as well take a longer exposure to see what the limiting magnitude was like - it's greater than 10, and I don't have a good enough star chart to figure out how *much* greater than ten. Looking at it closely, there's quite obvious chromatic aberration at the edges of the field.

Ok, I now know the field scale, orientation and approximate limiting magnitude. There is something which, if not terribly interesting, is at least not entirely boring.

Not visible on the central portion posted earlier is a curved line of three stars. In the full jpg Capella is located at about (2377,1715) according to GIMP. The central fainter star, at about (2387,1834) is "Capella H" a gravitationally bound companion to Capella. This star is actually a red dwarf binary, separation 2.7" (or half a pixel on your image) with magnitudes 10.2 and 13.7. The limiting magnitude is clearly well under 10 but as I don't yet have good photometry for the field I can't do much better than guess it's somewhere between 11 and 12. Of course, if the light wasn't spread over a wide area, contrast would be higher and much fainter stars would be easily visible. Perhaps it's time I investigated the Gaia preliminary data set which was release recently. The Hubble guide star catalogue might also prove useful.

It's just possible, with well collimated optics and a somewhat greater stack of exposures, your image data may be good enough to be able to deconvolve the binary at least to the point where it's clearly elongated if not fully resolved. That would be an interesting project; I'll see what I can do with the image as-is.

[fivemack]

The Gaia archive is slightly irritating because it outputs data either as very wide HTML tables 20 rows at a time, or as .VOT files with huge XML headers followed by a binary data block.

I've extracted the stars around Capella down to G=10 (Capella is of course too bright to show up in the Gaia processing); the plot is rotated by an arbitrary angle, and the magnitudes are in the Gaia G filter which goes from 400 to 1000nm and so makes red stars appear much brighter.

 plot "foo.csv" using 1:2:(10*0.8**$3) with points ps variable pt 7

I think the star at (2239,1729) in the image is at (79.45,46.12) in the plot and the one at (2335,1635) at (79.13,46.14). There are stars in the plot that don't appear on the image and vice versa, and obviously the image is a bit squashed.

[xilman]

Originally Posted by xilman

The limiting magnitude is clearly well under 10 but as I don't yet have good photometry for the field I can't do much better than guess it's somewhere between 11 and 12. Of course, if the light wasn't spread over a wide area, contrast would be higher and much fainter stars would be easily visible.

Take a look at

http://simbad.u-strasbg.fr/simbad/si...ound&Radius=20

AFAICT your limiting magnitude is at least 11.5

[xilman]

Originally Posted by fivemack

Full image in Nikon semi-proprietary lossless-compressed format (I think there exists free software that can read it)

GIMP can read your file but not in any useful sense.

Personally I'd prefer conversion to FITS. Any takers?


Since discovered http://www.xnview.com/en/xnconvert/ --- now to see whether it does what it says on the tin.

[fivemack]

Have a look in http://www.chiark.greenend.org.uk/~t...matic-capella/

I tried running them through pnmtofits, but it crashed while trying to print an invalid-looking FP number to stdout.

[xilman]

Originally Posted by fivemack

Have a look in http://www.chiark.greenend.org.uk/~t...matic-capella/

I tried running them through pnmtofits, but it crashed while trying to print an invalid-looking FP number to stdout.

The red dwarf pair is very obviously redder than its neighbours. The left-most looks bluish to me, the other perhaps yellow.

[chalsall]

Originally Posted by xilman

The red dwarf pair is very obviously redder than its neighbours. The left-most looks bluish to me, the other perhaps yellow.

Are you taking into account the sensor wavelength bais?

[kladner]

This is a DSLR (Nikon D90). The stars are by definition at infinity focus, but the point at which the focus ring on the lens stops rotating is a couple of degrees beyond the infinity focus point - I would not be surprised were the focus ring just input to an angle-encoder, with a motor inside the lens moving the glass around.

I know brands differ in their approaches, but on my Canon lenses the manual focus ring is pretty definitely mechanical. On my long zoom there is a warning to switch the lens to manual before using the ring, to avoid damage.

On the wide zoom, there is no warning. However, the Auto-focus will take over when the shutter release is pressed. I have to hold the release halfway and then use the ring...or set it to manual.
It may be that the lens manufacturer in your situation would realign the elements, and generally tighten things up, for a fee, of course.

[xilman]

Originally Posted by chalsall

Are you taking into account the sensor wavelength bais?

Nope.

[xilman]

Originally Posted by xilman

Since discovered http://www.xnview.com/en/xnconvert/ --- now to see whether it does what it says on the tin.

It sort of does so. It may be my incompetence at driving the program but so far I've only been able to extract 8 bits per pixel, and I've no idea whether that is one of the RGB channels or some average of two or more.

A 32x32 image centered on a moderately bright star gives a nice estimate for the PSF though close examination suggests that there may be one or two barely detected stars off to the right and lower right of the image in which case I'll either have to estimate the dark field in those areas or crop to a 16x16 before deconvolving.