[Prev][Next][Index][Thread]

Re: (ATMoB:Discuss) astrophotography with ordinary digital cameras



At 06:38 AM 12/5/2001 -0500, Michael Carnes wrote:
Bruce asked me where I found the camera reviews. ETC.

Mike, I know this has been a long thread about using conventional CCD cameras, but there are still some missing facts.

First, the Nyquist theorem requires 2 pixels within the Airy disk. Some enhancement techniques can give a slightly better resolution. Sony has released a $999 CCD with a 4 megapixel chip that produces slightly over 5.3  megapixel image file.

The Airy Disk is only determined by the spectrum of wavelengths recorded and the F/No.


D = 2.44 x  Lambda x F/No.


Note that the F/No. = F.L./Aperture so it has a dependence on the focal length and therefore the image scale.
Since we most frequently use  green light at 5500 Angstroms, this is .55 microns and the equation becomes much simpler:

D = 1.2 x F/No.

Most camera lenses cannot reach the diffraction limit because of residual aberrations such as chromatic and spherical in order to minimize field curvature and astigmatism. This complicates the issue because of the great variety of lens designs, especially zooms, where many compromises have to be made.
For a CCD with 25 micron pixel spacing, we would want airy disks of 12 micron diameter and need to keep turbulence and tracking errors within that size. Hence, we would want an F/10 system. Faster for blue light photography and slightly slower for red light.

Most CCD camera lenses operate at F/2.5 which would create severe oversampling. However, they likely to have enough residual errors to reduce the mismatch and vary F/No. with zoom position. This greatly complicates the issue and makes field testing important.
For a CCD with 9 micron pixels, the Airy disk should be 4.5 microns, and the perfect lens should be about F/4. Precise tracking accuracy becomes a severe requirement.

It is important to note that the recording of stellar (point) images is strictly determined by the aperture that is gathering photons. The bigger the lens, the better.
Now, to confuse everybody, the recording of extended objects such as nebulosity and unresolved galaxies is dependent only on the F/No. ! Hence, a fast, tiny lens can record faint comets but at a small image scale with little resolution.

Last month, just before the Leonid shower, I put my Meade CMOS Electronic Eyepiece on the 13" C.A. Schupmann at Stellafane, which is about F/10, and fed the signal into my Sony Palmcorder.
The camera has 9 micron pixels, 340 X 280. I got pretty nice B&W pictures of Jupiter with festoons, and Saturn, including the Cassini division.
I recorded four of the stars in the M42 Trapezium ( 5th magnitude, I think)( I could see six visually) but absolutely no trace of the nebulosity!
 I even scanned across the nebula and could not tell the difference from the sky background. In each frame, the stars have a slightly different shape, due to air turbulence.
This is where the focal reducers I described in Feb '73 S&T come in handy. They are faster than the standard F/6.3 design that is on the market, but suffer field distortion and vignetting.    


 Another question you had was about focusing.

Rather than build a knife edge or Ronchi screen, John Martin places two sticks across the telescope aperture, as recommended in Berry's AIP book.
They form two shadows in the out-of-focus image that merge into a single bright diffraction line at best focus, then diverge again. The same thing should work well with threads across a small aperture lens. This assumes you can see the images in near real time. Remove them of course, before taking an exposure.

Several few months ago, Deep Sky Dave had a problem with a bright line radiating from the planets and bright stars in his scope. The 45 deg. viewer ahead of the EP had an Amici roof prism inside. The hairline intersection at the roof caused a strong diffraction line. This would not be present in a standard 90deg. prism or mirror diagonal.

Clear, Dark Skies,
Paul

Follow-Ups: References: