ollypenrice

A valid point about setup time and clutter for portable rigs. I regard ambulant astrophotographers as heroic in every case! However, we've just been doing a faint Ha nebula in the very fast RASA and the emission signal is not very strong, this with an ASI2600 OSC. The fact is that the OSC is blind to Ha on three quarters of its pixels. Olly

Very good, clearly. What you don't have with this setup, though, is star colour. The L-extreme (and similar) filters do open up the CMOS OSC possibilities enormously. But, yes, a good image done very quickly. Olly

I agree with you on the matter of focus change between colours. I don't find any change to come from the filters, rather it comes from the optics - which means you can correct for it in mono and can't in OSC. That said, I never do. I just focus in green and scroll RGB, RGB, RGB and then refocus as the temperature changes require. Another non-issue. However, I disagree over a perceived need for multi-night with mono or with any suggestion that OSC is faster. It cannot be faster and isn't. OSC shoots (broadly speaking) with two out of three colours blocked all the time. Luminance shoots with the full visible spectrum being recorded. On top of that, narrowband is captured much faster in mono and can be used to replace luminance, though not by using it in the luminance channel. By way of evidence, here's a Heart Nebula shot in two hours with a 4 inch F5 refractor and old school CCD camera with a QE of 50%. It's Ha RGB. I don't believe this signal could be matched by an OSC, even a modern CMOS one, in the same scope. Mono is fast but does bite you if you're clouded out without one of your layers. (I do have a much better Heart, by the way. 🤣) Olly

No. The filters are not usually the source of the bunnies. These are produced by dust closer to the chip, which is to say on the chip window. I won't say this applies literally every single time but it proves true at least 19 times out of 20 for me. I'm not alone in working this way: a number of our experienced imaging guests do likewise. What is more, the LRGB image's light levels will be determined by the luminance layer so, if that has been properly flattened, it will have a significant tendency to flatten the RGB layer as well, according to its own light. Assuming you have clean filters, I can only suggest that you try it. It works for me. Olly

It isn't 4x the work for mono over OSC. It might add 10% to the processing time but not more. Some people take flats per filter, for instance, but I don't, I use the lum flat for everything on nearly every occasion and very, very rarely have a problem with this. I stack the three colour files, which maybe takes four minutes longer than making a single stack, I open the three stacks, and, in AstroArt, need two clicks to turn that into RGB. A minute? In terms of exposure time, LRGB gets equivalent signal in less time because luminance collects R,G abnd B at the same time. I've just started working with an OSC camera again. It's a CMOS and is much better than the OSC CCDs I've used in the past. There's a certain charm to the simplicity but it's wiped out by the moonlight unless you add a filter and, well, then you've added a filter! It gives great results on reflection nebulae but, unfiltered, doesn't go deep on emission - even in the RASA 8 at F2. Our decision to go for OSC was driven by the RASA, which cannot accommodate a filterwheel. For me the real choice would be between OSC with optional dual band filter and mono. OSC on its own is quite restrictive. Olly

In that case how about posting a crop at the full resolution? A downsampled image doesn't support your argument in favour of 0.39"PP over a much coarser sampling rate. If you really can resolve at that sampling rate I'll be the first to congratulate you (sincerely) but, when I imaged at about 0.62"PP, I found I could not present images at anywhere near full size. I was in a world of 'empty resolution,' if you like. I since went from a FL of 2.4 metres down to one of 1.0 metre using a camera giving just below 1"PP and got results I could crop and view at full size. Even then, though, Vlaiv provided a convincing demonstration that I was still over-sampled: he resampled the image downwards and then back up again. The result was the same as the original. That's a very crisp image indeed, for sure, but the debate is about whether or not you really needed 0.39"PP to achieve it. Some galaxy images from a 1m focal length: Basically I'm not convinced that, with present-day amateur cameras, we still need very long focal lengths. Olly

Yes, it becomes like a RASA, though the RASA is purpose-designed from the off and performs better. The problem is that, arguably in this world of small pixels, the Edge has too much focal length, more than can be exploited on deep sky. You can bin down to a more workable resolution and gain speed but at a high cost in field of view. I agree with your main points but If the 515mm is done with a small refractor then the optical resolution (Dawes limit) will be less than that from a larger refractor or Newt at 1000mm. The Tak 105/TEC140 Dawes limits are 0.83/1.09 arcsecs respectively. What's your view on the relationship between real image resolution and optical resolution? Or, to put it another way, how much optical resolution do you think is needed to support, say, the 1.5 arcsecs you mention? Olly

I'm afraid this is wrong. The Hyperstar advertizing is disgracefully misleading because it gives the impression that, in going from F10 to F2, you get the same image in a fraction of the time. You don't. You get a tiny image in a fraction of the time if you're shooting a small target. This is because the focal length is reduced in order to speed up the F ratio. To get the same image in a fraction of the time you would need to keep the focal length as it is and increase the aperture. This discussion comes under the famously controversial title of The F Ratio Myth. Go there at your peril! Galaxy imaging: first you need to decide on an image scale in arcsecs per pixel. Say 1.2 or 1.5 or whatever. This is determined by FL and pixel size. Then, with both of those fixed, you can think about aperture. If you only think about F ratio you'll go up a gum tree! lly

My experience says that you need more than 500mm and not more than 1000mm now that pixels are so small. You also need excellent resolution, so forget the RASA. I use one and like it very much but not for small galaxies. Provided collimation is good I don't see the 1000mm Newt doing a bad job. Personally I use a TEC 140 for 1000mm but that's an expensive choice and requires long exposure times. Olly

No, I just threw it into the clamshell conversation for the sake of diversity. I like the clamshell for visual because it's quick and neat. Olly

You can also use a flexible teather (eg string!) but under no circumstances would I settle for a system which could slide out of the dovetail. Olly

Or in both ends, since, when fiddling about with telescopes, we can end up with them pointing downwards. (Bitter experience!) Imagers do not love the clamshell. It's a likely source of differential flexure since it holds the tube only in the middle. It can help to put the guide scope on the main OTA but tube rings are the real solution. Olly

The trouble is that, if you want to image at F5 without diffraction spikes and with a full frame chip, where do you go? As far as I know, it's a good FSQ106 or nowt. How do you buy a good one? Buy second hand and ask to see results before you do. Or buy from a very good retailer. Olly

I, too, think the sensor is badly damaged. The flat has patches of black which presumably have no signal, or very little, so there is little or no information there for the calibration to adjust. We often see this when there is dust on the sensor: there is nothing the flats can do to correct for that. Olly

Hardly. I've done literally thousands of hours of happy imaging with Taks but the fact remains that they struggle for consistency with the FSQ range. Olly

I try not to worship brands. The truth is, for example, that Takahashi (yes!) turn out a lot of junk, largely in the form of hopelessly mis-collimated FSQs. But they do very well on their other products. I've used three FSQs and they've all been good but, in my line of work, I've met a good number of others including a significant number of lemons. I've also had a stream of personal messages from people struggling with problematic FSQs. So I don't worship Takahashi but I certainly like the ones I've had myself. Their visual scopes get a consistently good press but I don't know much about those. On the other hand I've been accused by retailers of endorsing what they considered to lemons supplied by competitors. Tough. I described the instruments as they were sent to me. Briefly, I don't think brands give a reliable over-view of what you can expect when you order. What matters more is the mentality of the supplier and that's where retailers like FLO stand out. If the customer isn't happy they'll accept that. There is one brand I wouldn't touch with a ten foot pole but that will be strictly by PM. Olly

That's a good Witch Head but can you provide a link to the original? All the details would be nice to know. Olly

If this is the case then Astronomy Tools is bonkers. True. I do live on a mountain top and, though it's not a desert, it's pretty arrid and I never get anywhere near to resolving at 0.67". Double it at best. I've imaged at about this sampling rate with a 14 inch scope (which had the resolution to support it) and with a mount tracking with an RMS of about 0.3 arcsecs. Even so I resampled every image downwards by a large margin. (The camera did not work properly when hardware binned for reasons unknown.) Olly

That's a really good answer from Daf1983 with an image to prove it. Olly

Sorry Wim!!! 👹 Perhaps I should settle for Wimran or Gorim! (When I was at school I had two inseparable classmates called Unsworth and Drinkwater. One teacher found their inseparability intolerable and called both of them Unswater. He took this to an extreme when he wrote only one end-of-year report between them - describing Unswater's progress that year.) Olly

Yes, collimating the entire lens cell as one (which just means making it point light straight down the tube) is something I've done as well and found it very easy using an illuminated ballbearing as a point source. This method was given to me by a TeleVue optician over the phone, not from uncle Mick on Facebook... 😁lly

But they can be round, as I explained earlier: in order to elongate a star, the guide error has to be more than half the image scale, or thereabouts. If pixels were dimensionless points this would not be true but pixels have a dimension capable of absorbing a small but significant degree of error. This tolerance diminishes as the image scale increases in resolution, which is why you need better guiding to image at finer pixels scales. I'd also want to think about the shape of the elongation. It is spindle-shaped, ie pointed at each end and thicker in the middle. In many years of imaging I have never seen guide error produce that shape and I'm struggling to think of a guiding error scenario which could produce it. The killer indicator, though, is that the elongations are not parallel across the chip. That absolutely rules out tracking error. Olly Edit: I 'crossed in the post' with Wim who has put a numerical value of 0.8 arcsecs on the RMS below which guiding will have no effect and with Stuart whose experiences match my own.

Spacing issues usually worsen away from the optical axis. Like all these things, 'usually' applies to an ideal world rather than this one! What is certainly possible is that there is more than one origin to the problem, though, so it could be spacing plus collimation. One test which should certainly be performed here is the classic star test. Place a star precisely in the centre, set an exposure which does not saturate it, then take a number of images of very short duration in focus and outside and inside focus by differing amounts. The out of focus images should be round. My money says they won't be. The short exposures eliminate tracking error. Olly

See the post above. Nicolàs has clearly demonstrated a systematic error in the image which bears no relation to a guiding error. How can a guiding error produce a quite different angle of elongation on different parts of the chip? The entire chip is subjected to exactly the same displacement from the guide star centroid when guiding fails to hold the centroid in position. Nor do I think it is it coma because coma is an off axis aberration producing fan-like extensions which worsen as we look away from the optical axis. We see nothing like that: the stars are best near the centre of the elongated line, not at the side closest to the optical axis. As Stuart says, it is perfectly possible to have the full resolution of which the system is capable (which is more demanding than having round stars) with an error of zero on one axis and an error of half the image scale in "/PP on the other. Provided the error lies below about half the image scale it won't show. Besides, the elongation is so extreme that, if halved and applied to the other axis, it would surely be visible. Olly

In my opinion this is not a guiding issue. Look at the elongation of the stars. On the left hand side the elongation points down and to the left. On the right hand side it points down and to the right. In the centre it's vertical. Elongation due to guiding error produces parallel elongation and doesn't usually show much in short subs. I think this will be an optical or mechanical error, the lens not being optically aligned with the tube or either the focuser sagging or the chip being tilted. Take a sub, rotate the camera 90 degrees and take another. If the tilt is in the chip it will rotate with the camera. If it's in the optics or the focuser it won't. Olly