Cosmic perspectives

(move your mouse over the picture to see the Baidu M1 Rocket trail on a single 60s shot)
The other night, I took some pictures of the famous Leo triplet, Messier 65, Messier 66 and NGC 3628. The atmosphere was amazingly stable and so I went on to try to record 1 hour of 60sec exposures. It turned out that only 27 of the 60 pictures had good enough tracking to be used for the final picture, but result turned out pretty good (mouse over the picture to see the difference between one 60 sec shot and the stacked processed result).

One interesting thing I discovered while reviewing the raw shots is that on 2 of them, I had captured a trail. After quick measurement, it was something moving at about 26' per minute. Clearly too slow for a plane or a low altitude satellite. Calsky.com, the excellent satellite viewing prediction website, proved to be remarkebly useful in finding what it was. It turns out that this trail was the Baidu M1 rocket (orbital details), launched in 2007, and orbiting on a very elliptic orbit (523km x 21046km). At the time of passage, it was distant by 6707km.


Exposure Data

• Lens: Newtown Perl-Vixen 130/720
• F/stop: f/5.5 (focal)
• Mount: Perl Vixen Super Polaris mount

• Exposure: 27 min total exposure, composed of:
• 27 x 60 sec (120 min) unfiltered RGB @ ISO 1600 Canon 500D
• Mode: Raw
• White Balance: Auto
• Calibration: Flat, Darks and Offset in DeepSkyStacker
• Processing: Standard deep-sky adjustment in Photoline

• Date Start 2011-Apr-09 02h02 BST, 2011-Apr-09 03h13 BST
• Temp: 9C, 8C
• Location: West Witterings, UK

Image Scale

There are a few astro bloggers that I think do a fantastic job of sharing their expertise in the field of astrophoto. Jerry Lodriguss is certainly one of them. His blog post "Catching the Light" is a fantastic source of info to understand more about how to take great pictures.

In a recent post on Image Scale, he describes how to calculate the image scale on your camera, that is, what field of view each pixel is representing.

So an approximation of the formula (converted in metric system) goes like:

P = (206 * S) / (FL)
Where:

P is the image scale per pixel in arc seconds
S is the size of the pixel in microns
FL is the focal length in millimeters

(the real formula is P=3600*arctan(S/FL), but since S is very small compared to FL, the simplification will be good enough for astro use)

Example: Calculate the image scale per pixel for a Canon 500D camera with 4.7 micron pixels when used on a Perl Vixen 130/720 telescope (720 of focal length).

P = (206 * S) / (FL)
P = (206 * 4.7) / (720)
P = 1.3 arc seconds per pixel

So each pixel of the 500D on the scope will cover 1.3" of the sky.

Jerry goes on with saying that this is very useful to estimate the smallest details you hope to record. You typically would like your image scale per pixel to be 2x to 3x smaller than these smallest details (eg 0.3" to 0.5" to record 1" details on Jupiter).

How do you do this? Increase your focal length with either a barlow, eyepiece projection or lens on your camera held up to the eyepiece of the scope.

Something else to consider as Jerry points out is the seeing. Let's say the seeing is 1". If you are doing deep sky astrophoto, then with the additional tracking errors and other atmospheric variations over the time of the picture, you're unlikely to reach the seeing true limit, so you're probably correctly sampled at 1.3"

But in planetary, where exposure times are very short, it's worth considering the 2x or 3x oversampling as the seeing is less of a real limit.

About Canon 500D noise

DPreview.com is one of those sites that do an amazing job of dissecting cameras down to a level which can become useful for astrophoto.

For example their measure of noise levels for the 500D is something I think might be useful to estimate the optimal ISO / # pics required for deep sky imaging and stacking.

ISO Std Dev

100 3

200 4

400 5.8

800 8.3

1600 11.3

3200 15.8

6400 24

12800 35.7

However, if you have a clue on how to do it, drop me a line