I have a decision to make, it is a nice decision to have to make but still for me a substantial investment is about to be made for the next year or two in this hobby and I want to choose wisely.
I have had a great time with my IR Cut Modified Canon 600D and to be honest there is still plenty of learning I could do with this camera. However, for some time I have wondered how much better my images would be if I had a camera that was specifically designed for astro imaging in mind. To that end I have slowly been putting a little cash away while researching the hell out of the subject, you know what it is like, the research is as much fun as the actual purchase!
Like most of us aspiring Astrophotographers I have debated with myself a number of topics related to purchase of an Astro specific camera. Let me break this down with some of my thoughts that are driving my thinking.
NOTE - this will not turn into a technical comparison of each feature - you can do your own reading into the data sheet, although I may make reference to some of the more important (IMO) specifications when comparing certain cameras.
One Shot Colour or Mono?
OSC because I do not have the weather that provides me with enough hours of imaging to pull together light through 3 or 4 filters before I have what I would consider to be a complete image. One good full clear night and I would generally have a good image at the end (maybe even two!). I do understand that through automation of the imaging session with N.I.N.A. Advanced Sequencing that it is possible to gather 30 minutes through each filter and cycle therefore potentially ending up with a couple of hours through each filter and therefore a solid image at the end. Even so, plenty of opportunity for sessions that are cut short end up with less data in a given channel than one would like.
OSC is also feasible at a lower price point, I reckon on an additional 75%-100% of budget on top of the price of an OSC camera to go Mono; Mono version of the OSC will be 10-20% more expensive, cost of a filter wheel plus the costs of filters - admittedly 2-3 to start with but still a substantial further investment.
There are other factors that I understand exist - telescope quality does not have to be as solid when imaging Mono as the Narrowband (NB) filters mean chromatic aberration is not an issue as it is with OSC, and therefore a cheaper scope is something you can get away with. This is interesting, but of course I do not want the change of camera to introduce a wholesale change in my rig.
Also, the NB does mean possibility of imaging during bright moonlit nights when even with a Dual Band filter in the OSC setup it would not be worth doing.
In the end Mono is something that I will perhaps make my next step, to be honest the final decision is as much financial as it is to do with the complexity and time taken for image capture.
OSC it is! (but you can tell there is a tinge of mono regret here ;-) )
Cooling
Yes.
A lengthier answer being that without cooling a large part of the advantage of an Astro camera is lost. Being able to set the temperature of operation either to build a Dark / Dark Flat library or simply to reduce the thermal noise during the night is a massive advantage. In fact going further, the cooling needs to use Thermoelectric Cooling (TEC). This allows absolute placement of the temperature rather than just being able to move the temperature a few degrees below ambient. I am thinking this must be one of the major improvements over the DSLR where thermal management is simply not possible forcing you to rely on processing techniques (Siril Dark Optimisation for example).
Brand
Admittedly most of my decisions on the manufacturer are related to availability in the UK as I want someone to take any hassle out of returns or support rather than having to package and return items halfway round the world with the associated delays. This pre-requisite still leaves the big players, ZWO, QHY but also some other less obvious options such as Player One and Altair. Difficult here as QHY/ZWO are well understood entities with many bloggers / YouTubers on in their camp. Player One is one that has only just caught my eye, mainly because of a lone YouTuber extolling their virtues, now of course I am not so naive as to believe that their is some advantage in terms of the reviews that are given to the various cameras. I am not suggesting here of direct payments but possibly there is some hesitation to give bad reviews as the channel in question may not be given products in the future.
Each of the companies basically take the latest and greatest CMOS sensor and package that up with their electronics. I will assume that similar to the TV market, the panel that displays the picture is surrounded by electronics by specific TV manufactures and I don't think you can say that there is no difference between say a Sony vs a unnamed Supermarket brand. Therefore I am of the view that there is some difference, what is harder to ascertain is are these differences to some degree superficial and do they really make a difference in the final image, after all that is what counts.
I think for me given there is a pretty penny involved some comfort in sticking with a well established brand is important. This probably means ZWO for me, even though there seems to be a price premium compared to say the same sensor wrapped in a Altair package.
Form Factor
So when it comes to form factor there is for me financial considerations, of course a medium format or full frame sensor would be great, the multiple thousands £'s less so! As with many things in life there are compromises to be made. Backing away from these next level cameras as ASP-C format camera is something I would aspire to, this would match the Canon 600d I currently use in terms of aspect ratio. However, as a first timer of a dedicated camera even the cost of an ASP-C is perhaps just out of reach although their is a wide range of prices across different manufactures. The next and perhaps more aligned to my budget are those 'micro' form factors. The smaller 4:3 and square sensors these fall around the £1000± 200. There is an advantage to these smaller chips, you get a free zoom from the ASP-C sensor.
It was important to me that the subjects I am likely to make a success of with my William Optics ZenithStar II's 360mm focal length fit well within the new Field of View provided by the chosen camera sensor. In some ways the square format factor that is provided by the Sony IMX 533 sensor could work really well, many subjects are fairly compact balls of gas that a lightly held together in globes makes them symmetrical. Also a square sensor means rotation for correct framing is less important.
On the other hand the 4:3 format akin to what I am used to with ASP-C in my DSLR and in some cases this provides for some artistic choices of subject placement that square sensor seems to lack. The modern 4:3 micro format sensor is the Sony IMX 294.
Pixel sizes with my DSLR in there to provide a comparison:-
Canon 600D ASP-C - 4.3μm
IMX 294 - 4.63μm
IMX 533 - 3.76μm
The importance of Pixel Size is in the combination with the focal length of the optics to produce a number that indicates whether the final image under good seeing will look under-sampled or over-sampled. Pixel Size / Focus Length * 206 =X (arc second / pixel); X should be around 2" - this is a very rough guide and is dependent on how good your 'seeing' is. I'm taking as a rough guide this score between 1.5" to 3" as a good guide. Over-sampling can be compensated for by using Deconvolution in software to reduce bloating and improve star shapes, Under-sampling isn't something that can be recovered from, therefore it is better to be over rather than under sampled.
Under-sampled - images can make details (like star shapes) look blocky and square
Over-sampled - provides softer edges making stars look more round and perhaps a little blotchy
Telescope | Sensor | Pixel(μm) / Focal Length (mm) * 206 | Arcsec / Pixel |
William Optics Z61 II | Canon 600d | 4.3 / 360 * 206 | 2.4 |
William Optics Z61 II | IMX 294 | 4.63 / 360 * 206 | 2.64 |
William Optics Z61 II | IMX 533 | 3.76 / 360 * 206 | 2.15 |
William Optics GT81 | IMX 294 | 4.63 / 478 * 206 | 1.99 |
William Optics GT81 | IMX 533 | 3.76 / 467 * 206 | 1.62 |
I have thrown in another scope (that might be on my wish list) that has a little extra reach in focal length as it would be nice to know that the camera I choose would also work well with any future upgrade. The above table does show that I am in the sweet spot with all these combinations, if anything the 533 matches very well with my current scope but given the variation in seeing conditions I won't be making my decision based solely on this data.
Field of View
So size of the sensor and the form factor come together to dictate the field of view that the sensor and optics will have on the sky, it is really important for me that the targets I am familiar with fit nicely in frame. To that end I have used N.I.N.A's great Framing Assistance feature with the 294 and the 533 specifications along with my current WO Z61 scope to show the difference on a wide range of targets.
Scroll through the pairs of images, the blue rectangle is showing the field of view, same object will show the 4:3 image followed by the 533. These are not my images, they are taken from the N.I.N.A Framing Assistant which uses a Digital Sky Survey as its source. I am showing here purely for framing examples.
In some cases I have deliberately rotated the frame to highlight where an object would fit, in some I did not bother but with a small amount of imagination it's possible to envisage whether an object would fit or not in a different orientation.
Some stand-out comparisons for me, with a little scoring in brackets:-
Pleiades for me looks better with some room around it, so although the 533 sensor still manages to hold the cluster it is tight. In fact as a matter of opinion the 533 manages to focus the attention more on the details of the object rather than creating an overall composition. What works best is obviously a personal choice. (294 = +1)
Elephant Trunk nebula is a bit tight to show the whole of the nebula, however this could work to the advantage of getting closer to the actual trunk. (Draw)
Flaming Star Nebula using the 294 sensor still allows the Tadpole Nebula in the same field of view, I do like that composition. 533 would show off the Flaming Star in one shot and actually it's tight enough that Tadpoles in their own shot would show a good degree of detail, so again what it lost with composition is gained with subject detail. (294= +0.5)
California Nebula I think is one of the images that or me is definitely better in the 294 framing. (294 = +1)
Veil has a similar comment o the Flaming Star, I like getting the whole supernova remnant in one FoV, but the 533 would provide two very good images. (Draw)
Pacman is well framed in the 533 and it's narrow FoV I think would help here. (533= +1)
Heart and Soul, I'll bring both of these shot together here are the same comment applies. The 533 framing just manages to enclose both of the objects and undoubtedly the tighter crop would highlight the individual objects well. In the 294 framing I could imaging that a 2 or may be 4 panel mosaic would provide and fabulous view of both objects, using the 533 I think 4 maybe 8 panels maybe required making it a more challenging proposition. (294 = +1)
Bodes Galaxy, Leo Triplet, Hercules Globular Cluster, Iris Nebula and the Triangulum Galaxy would all benefit form the 533 sensor closer crop. (533 = +2.5, challenging light domes and Moonless nights means I have not weighted them quite as much)
Andromeda looks a little too tight in the 533 so a win for the 294. (294 = +0.5)
One thing I would note out of all the objects is that the broadband targets; galaxies, clusters and reflection nebulae, are targets that I do not often shoot myself due to light pollution and susceptibility to moon light destroying the signal, hence weighting them down a little.
Scores in terms of the framing give a difference of 0.5 in favour of the 294 sensor, interestingly this is not as much as I thought while I was putting these images together in N.I.N.A.. In any case this is close between the two sensors.
Specification Comparison (a few main features)
Dark current the 294 sensor is not blessed with the newest circuitry present on the the 533 where efforts have been made to reduce thermal noise. For the record:-
294 @ -10ºc = 0.0465 (e/s/pixel)
533@ -10ºc = 0.00049 (e/s/pixel)
An order of magnitude better meaning that Dark calibration frames may be something you can not perform at all. On the other hand, generating a Dark Library at -10 ºc for instance and having the master dark used in calibration is not such a big deal.
Full Well depth is the number of electrons that a site can hold before it bleeds into surrounding pixels, in effect it is the point at which a site is saturated and therefore depending on the target could be the limiting factor in exposure times:-
294 = 63.7 ke
533 = 50 ke
ADC (Analogue to Digital Converter):-
294 = 14 bit
533 = 14 bit
Optimal gain setting with read noise:-
294 @ gain 120 read noise of 1.7
533 @ gain 100 read noise of 1.5
Conclusions
Sensor wise I doubt I would be disappointed with either sensor, the 294 based gives more artistic creativity vs the 533 truly focussing on the subject.
Under vs Over sampling, on pure numbers the 533 is closer to the magic 2 arcseconds / pixel but given the variability of seeing and where > 3" would be where on average seeing we would start to see evidence of under-sampling and < 1" oversampling then either of these cameras are going to be fine.
Specification wise the 533 has the edge, the biggest area is in the dark current noise - which is practically not existent, however with one fairly minor piece of work to generate a Dark Library at the temperature you are going to run at and perform a Dark calibration process (which I have automated anyway) this is not a massive advantage.
Manufacturer wise, the obvious choice is to go with ZWO, they are the manufacturer that appears to have the most market share and therefore the most stability in terms of drivers and support. They are not the cheapest, Altair Astro at the time of this blog entry £950 vs £1050 for the 294 based sensors.
After all the above, so many pros and cons, I finally did a search on astrobin for the sensor and the same aperture (60/61). There are lots of good example images using ZWO 533MC Pro and ZWO Asi 294MC Pro cameras, but on balance the image quality I saw has edged me to the 533 based camera, this is actually slightly against where I thought I might end up. The one item in the mind about the artistic framing I will solve by getting into Mosaics.
So purchase button pressed! ZWO Asi 533mc Pro - Done. Watch this space for some earlier images and and no doubt many updates from acquisition changes through to processing modifications