Spring 2024: galaxy season
IN THIS POST
How did I come to this?
Why is spring the galaxy season?
Do you need a big telescope ? Amateur vs. Hubble space telescope
Matching image scale with optical resolution.
Seeing, and the point of diminishing returns
Galaxies I photographed this spring
My plans for improvement
How Did I come to this?
About two years ago, I realised that I wanted to photograph a bit more than landscapes with starry skies. I wanted to make close-ups of deep-sky objects (DSO); to take the challenge of capturing images like those from Hubble … or at least as close as possible. That would bring my passions for photography and astrophysics together. I am a bit of a perfectionist; so, I decided to buy a telescope to perform guided astrophotography.
Plenty of options in the market did not make the buying decision very easy (just a luxury problem). However, there were also a couple of obstacles that narrowed down my options: weather and darkness. We do not get many hours of clear sky here in Switzerland due to cloud coverage and fog. Winter nights are cold, and summer nights are very short. That meant, that in order to enjoy this hobby, I needed an optical system that collects enough light in short time (a fast scope). I also had to decide how much focal length I wanted, which would be determined by the apparent size of the objects that I plan to photograph. After much searching and thinking, I decided to get a RASA11.
The Rowe-Ackermann-Schmidt-Astrograph 11 (RASA11) is a field corrected reflector derived from the pioneering Schmidt camera. It has an aperture diameter of 279 mm (11 inches), and a focal length of 620 mm. This means an f/2.2 aperture to focal ratio, which makes it one of the fastest commercial scopes. It has excellent optics providing clean images with no distortion, which is needed for image stacking.
At 620 mm focal length, it is in a sweet spot for being able to photograph small nebulae, and barely fitting large nebulae in the frame (with a 25 mm wide sensor). However, it falls too short for galaxies (or I thought so!), which are commonly photographed at focal lengths >1000 mm. That was a compromise I had to accept. Longer focal length scopes are normally slower. My decision was also influenced by the fact that I have no easy access to dark sky regions. It seemed more reasonable to photograph targets that are suitable for the use of narrow band filters (emission nebulae) to keep light pollution out of my images. Galaxies, on the contrary, are better captured using wide band red-green-blue filters, or a simple colour camera. So, I sort of gave up initially on the option of photographing galaxies (other than our nearest neighbour Andromeda) .
I got my telescope in spring 2022; but due to a sport (shoulder ) injury, I had to wait about a year to be able to use it. Therefore, my deep-sky-astrophotography journey really took off a year ago. I started, as planned, photographing some large emission nebulae, which you can see here. Eventually spring arrived, and there were no nebulae to aim at, as the spiral arms of our galaxy do not cross the sky at night in this season.
Having nothing else to shoot, driven by curiosity, by the technical challenge, and some rethinking of optical capabilities and processing techniques, I decided to give galaxies try. Little I knew that I would be hooked!
Now, this post will be on the long side, and a bit technical! I am a physicist; and despite being new to astro-imaging, most of the methods, rules, and techniques are rather familiar to me. So, I decided to discuss some of the aspects to take into consideration when choosing your DSO imaging setup.
For the less technical reader, before you jump over the text directly to the images, I can summarise that I am very happy that I started photographing galaxies; and I will continue doing it! Shooting true color RGB images of space is quite fun; results are beautiful, and there is something special about galaxies. The RASA11 is a great tool for people with little imaging time. It may not provide as much magnification as other scopes, but its high resolving power and speed allow us to capture more details from atmospherically sub-optimal locations.
The wide field channel of Hubble provides about 10 times more detail than any telescope from the earth surface.
My mistake, and main learning: it is not all about the scope:
"The single most important item for the imager is the mount. Next comes the mount. Then comes the mount.... Get the picture?"
Roland Christen
why is spring the galaxY season?
Astrophotographers refer to the time between March and May as the galaxy season. But what does spring has to do with galaxies?
Fig 1: Art work illustrating a side view of our galaxy. Wiki Image credit to Pablo Carlos Budassi, CC BY-SA 4.0.
Our galaxy has the shape of a disc with several spiral arms; and our solar system is about half a way out along the Orion arm. The orbit of the earth around the sun is tilted by 60 degrees with respect to the galaxy equator. When we observe the sky at night (away from our sun), we look into different regions of space depending on the season. During summer, autumn and winter, we see some of the arms of our galaxy crossing our sky and we have the chance to photograph many gas nebulae within it.
Fig 2: Art work depicting a top view of our galaxy. ESA/Gaia/DPAC, Stefan Payne-Wardenaar, CC BY-SA 4.0 IGO.
During spring, we are mostly looking out of our galaxy, and we can observe things that lie further away; that is, other galaxies and large clusters. It is not that we cannot see some galaxies during other seasons; but rather that during spring, we mostly see galaxies, and many more of them. Virgo cluster for example, which contains a large number of galaxies, starts appearing in the northern sky at night by late March.
do you need a big telescope? Amateur vs HST.
Well, it depends on what you consider as a large scope, and your expectations. Galaxies are several million light years away from us, which makes them look quite small (with the exception of our nearest neighbour, Andromeda). Normally, one needs scopes with long focal length (> 1000 mm) to get enough magnification to spread the galaxy over the camera sensor. At the same time, you want a good optical resolution. But how much is enough, and where is the point of diminishing returns? Keep reading if you like to know the figures involved when making your own judgement.
There are two primary measurements that determine a scope capabilities; its focal length, and its aperture (diametre).
The optical resolving power of a scope (assuming perfect construction) is limited by diffraction of the incoming light, as it interacts with the scope edges. A point source of light is projected on to our sensor as a diffraction pattern, from which we mainly observe the central lobe. The width of this lobe is called spot size, which limits the smallest detail that our scope can resolve. Accurately speaking, we would need an infinitely wide scope in order to project a point source as a point on the sensor. The wider the aperture, the smaller the spot. For a scope with aperture D, in millimetres, and an incident light with wave-length (lambda) in nanometres, the optical resolving power (OR) in units of arc-sec is
OR = 0.252 * lambda[nm]/D[mm]. (1)
One arc-sec is 1/3600 of a degree. The OR is measured in angles because distances in space (for imaging purposes) are also measured in angles. From our view point, it does not matter how distant the objects are from us, nor the absolute distance between them. Objects that are further away could be larger, and appear smaller to us. From an imaging perspective, it is the apparent angle between two points in space what matters; that is, the angle formed by the two straight lines from us to those points. So, we refer to galaxy sizes in arc-sec, or arc-min, and we refer to optical resolving power in arc-sec. For instance, the sombrero galaxy has an apparent size of 9 arc-min on the long side, and 4 arc-min along the short side. That is about 540 x 240 arc-sec.
Visible light comes in wave-length between 380 and 700 nm. Taking 550 nm as typical wave-length, the RASA11 (with D = 279 mm) has a resolving power of 0.252*550/279 = 0.5 arc-sec. It can resolve 540/0.5 = 1080 points along the sombrero galaxy long edge. The Hubble space telescope (HST) with its 2400 mm primary mirror, has a resolving power of 0.06 arc-sec (9 times more detail!). Resolving power is the reason why professional telescopes are built with such large mirrors.
However, one thing is the optical resolving power, and another is our sampling resolution, or image scale. Image scale is the angular size covered by a single pixel of our sensor, when paired with a scope of a given focal length. In the same way that the angular field of view (FOV) is the ratio of sensor size to focal length, the image scale is the ratio of pixel size to focal length. In arc-sec per pixel, it is calculated as
IS = 206*P[microns]/L[mm] , (2)
where L is the focal length of the scope in millimetres and P, the pixel size in micrometres.
matching image scale with optical resolution
In order to make full use of the optical resolving power of a scope, we want an image scale that satisfies IS ~ OR/2. This comes from the well known Nyquist theorem of sampling, which states that for accurately determining the content of a signal at a certain frequency, we need to sample at least twice as fast (twice as densely in this context of spatial resolution).
Let us consider again my 279 mm scope, with focal length of 620 mm. In order to fully sample the available detail of 0.5 arc-sec, I need an image scale of 0.25 arc-sec/pixel; meaning according to Eq.(2), a pixel size of 0.75 microns. I have never seen such a camera. Smallest pixels I have read about are about 2.4 microns. My astronomy camera, the ASI2600 MM Pro, has 3.76 microns pixels, for an image scale of 1.25 arc-sec/pixel. In this case, the image is being sampled with less density than needed to capture the maximum detail available. In comparison, Hubble achieves an image scale of 0.04 arc-sec/pix with its wide field camera, having rather big 11 micron pixels, thanks to its long focal length of 57.6 m; thus matching its optical resolution of 0.06 arc-sec fairly well.
But how bad is a 1.25 arc-sec/pix image scale? Can we improve it? Does it make any sense to improve it?
In principle 1.25 arc-sec/pix is not bad. It is actually a recommended image scale for normal amateur imaging. Image scales between 1 and 2 arc-sec/pix are recommended due to atmospheric seeing limitations (read further below). But yes, it can be improved, and it could make sense … to a point.
To achieve a smaller IS, one could use cameras with smaller pixels, probably sacrificing in dynamic range. One could also change the scope to something with a similar aperture but larger focal length, at the expenses of speed and cost. There is however a free way to improve effective sampling resolution, combining dithering with drizzling.
Dithering is the procedure of shifting the scope a few pixels in random directions between pictures. This is used to eliminate possible sensor patterns, and improve signal-to-noise ratio (SNR), but it also allows us to capture information that in other frames would be “between” pixels. We can use the information that is spread over several frames, and reconstruct a higher resolution image by oversampling during the integration process. This oversampling algorithm is known as drizzle, and it was developed by NASA for the treatment of some Hubble images. A 2x2 drizzle, reduces the image scale from 1.25 to 0.625 arc-sec/pix. It is also possible to perform (with some good computing power) a 3x3 drizzle, giving for my setup an effective image scale of 0.42 arc-sec/pixel. These figures are excellent although they do not match the ideal 0.25 arc-sec/pix for the chosen scope. Should you hope for better?
I don’t think so. In practice, there are a couple more practical problems that limit our image resolution: atmospheric seeing, and tracking accuracy.
Seeing, and the point of diminishing returns.
The seeing (which has nothing to do with the also undesirable clouds) gives the average resolution that we can perceive. Seeing is limited by the dispersion of light by the content of our atmosphere, and its fluctuations. This is one of the reasons why astronomical observatories are located at high altitudes. Good seeing conditions are about 1 arc-sec, and exceptionally good seeing on earth can reach 0.4 arc-sec. By taking very short exposures, one can get some frames with better than the average seeing. In that way we could slightly reduce blurring by only keeping the sharpest frames (a procedure known as lucky imaging).
Tracking accuracy depends on the quality of your telescope mount. The larger the scope, the more robust your mount has to be to track the sky with a desired accuracy. The maximum accuracy of commercial mounts is about 0.2 arc-sec. For comparison, HST has a pointing accuracy of 0.007 arc-sec.
3 Set-up comparison with ASI2600MM (3.76 micron pixels) and 100 nm color filter
Set-up IS OR Speed ExpT
C11 EdgeHD 0.25 0.5 f/10 200s
170/1200 0.62 0.8 f/7 100s
RASA11 1.25 0.5 f/2.2 10s
Let us compare my RASA11 with two typical choices for galaxy photography. Of the three examples above, the Celestron C11 is without a doubt, the best match of image scale with optical resolution, thanks to its long focal length of 2800 mm. However, with a relative aperture of f/10, it requires exposures above 3 min. The resolution is therefore limited by the average seeing conditions.
A typical refractor used for galaxy photography is a 170/1200. While not as good as the C11 in terms of the Nyquist theorem, it does rather well. In addition, its focal ratio allows one to reduce the exposures to about a half of those required with the C11. Yet, the ideal exposure length of 100 s is on the long side for lucky imaging, meaning that this set up may be limited by the seeing as well.
The RASA11 is clearly the worst in terms of sampling resolution. However, due to its extremely fast focal ratio, optimal exposures of 10 s give some room for lucky imaging (although maximum benefit would require sub-second exposures). For people, like me, living in regions with seeing conditions rarely better than 1 arc-sec, smaller image scales are useful only if one can overcome the seeing, taking very short exposures. For this reason, the apparently worst RASA11 becomes a good choice. It also allows me to capture images in the few hours of clear sky that we commonly have in Switzerland.
Apertures larger than 300 mm, and focal lengths longer than 2 m bring, in most cases, no advantages for galaxy imaging from sea level. For this reason, a popular choice among C11 EdgeHD owners, is to use the native 2800 mm focal length for planetary imaging, but add a 0.7x reducer for galaxies; transforming the C11 in a 279/2000 mm set up.
Plans for improvement (and a Thing to consider when buying a mount.)
What I have not mentioned until now is that the weakest link of my set up is (sadly) the mount. It is a good mount, don’t get me wrong! But it is not the right one for my imaging conditions and scope. Why?
The RASA11 v2 was sold to me by a local telescope shop together with a Celestron CGX mount, as a perfect set for astrophotography. Celestron claims that the CGX can handle a load of 25 kg for this purpose. At the same time, the CGX is considered a semi portable set up (not an observatory grade mount). Unfortunately, my experience does not support their claims. While the CGX seems to be a good mount with no noticeable backlash, it does not have the robustness and rigidity to handle the RASA11 in my balcony. For imaging purposes, the RASA11 must be used with a dew shield to avoid condensation and image contamination by stray light (in light polluted regions ). The surface area of the setup becomes rather large, and catches the softest of the winds. And I really mean, very very soft winds that one cannot even feel. While tracking may be good inside a dome (I wouldn’t know), it becomes impossible to optimize outdoors. On a rare occasion of absolute quietness (helped by some cardboard shielding), I was able to achieve 0.7 arc-sec tracking accuracy for while. Most of the time, my tracking accuracy is worse than 1.5 arc-sec, and I have been able to correlate it with air movements. With the figures above, tracking accuracy is my main limitation for capturing small details. If I were to reduce my image scale by any means, tracking accuracy would have to be improved before, for it to make sense.
Claims of telescope companies concerning the usable payload must be taken with a grain of salt. The required size of the mount does not depend only of the weight of scope plus accessories. It also depends on the total surface area, if it is not wind protected 100%. You probably need a mount capable of handling about twice the weight of your setup. For this reason, although my mount is in perfect working conditions, I am selling it, to replace it with a much heavier mount.
By the time being, I keep imaging when the clouds allow. Here a few galaxies that I manage to photograph this season (though not with the level of details that I would like).
When I bought my scope I did not think I would be photographing galaxies; but I quite like it. In my opinion, the RASA11 turns out to be a very appropriate tool for people living in areas with not ideal seeing.
galaxies that I photographed this spring
Bode’s M81 and Cigar M82 galaxies.
Field of view: 0.752 degrees. Image scale 0.629 arc-sec/pix. Image size 7502 x 4220 pix.
Sombrero galaxy (M104) and Jaws asterism.
Field of view: 0.547 degrees. Image scale 0.629 arc-sec/pix. Image size 4890 x 3912 pix.
Leo Triplet: M65, M66 and NGC 3628 (the hamburger)
Sombrero Galaxy M104.
June has arrived, and the nights are getting too short for imaging (with less than 2h full darkness). Galaxy season is over, but I still hope to get some more imaging time of the circumpolar galaxies in the late summer. Will I have a new mount by then?
Thank you for reading!
