Pondering an ideal telescope - or a "winter fantasy"
Posted: Thu Jun 13, 2024 1:06 am
I am hoping you’ll indulge my wet/cloudy weather, winter ramblings about an ideal telescope upgrade. My interests lie in deep space objects, which are limited by photons and resolution. I am drawn to far off galaxies or closer planetary nebulae, with similarly small apparent size, such as these examples.
C74 (NGC 3132, mag 9.8, 2000 ly) Eight Burst PNe with very hot central white dwarf (100,000 C - we see less than 1% radiation in the visible). C59 (NGC 3242, mag 7.7, 4800 ly) Ghost of Jupiter PNe. PGC 19481 (mag 14, 300 Mly) Blue Ring galaxy, showing a ring of hot blue young stars believed to be triggered by a head-on galaxy collision.
These PNe are quite bright. But PGC 19481 is magnitude 14 and 300 Mly distant. Objects like it strain resolution and are starved for photons. I feel a rush of ‘aperture fever’ coming on …
Photons argue for more aperture, that’s inescapable, at least up to practical limits imposed by weight and cash. Resolution is largely a question of seeing, especially in my area. Discussions online about ideal image-scale tend to converge somewhere around 0.75 arc-seconds per pixel, with only ‘mountain top’ conditions warranting significantly smaller image scale. My best data (Centaurus A) used a PixInsight FWHM cut-off of 2.62 pixels and an average of 2.55, which is 1.64 arc-seconds. Nyqvist sampling (subdivide by 2 or more) would suggest an ideal image scale for this data of 0.82 or less. The data used 0.64 (EdgeHD 8” with 0.7x reducer), which worked well, but may have still been slightly over-sampled. However, guiding down at 0.3-0.4 arc-seconds RMS total is rare for me (2-3 days per year), and larger FWHM due to seeing is far more common. It seems that target image scale should be no smaller than 0.8 arc-seconds per pixel in my case. Hence, one goal of this exercise is to target ideal image-scale in order to maximise SNR. Put another way, I don’t want to squander SNR with an image scale that is too small.
Of course, I’d also like to target those juicy nebulae that we are blessed with in the Southern Skies, which becomes more a question of field of view, aggravated I suppose by going for long focal length to target DSO. Hence, I would also like to increase my FOV well beyond my current 0.75° x 0.5°. This argues for a full-frame camera and a OTA image train with a well corrected flat field large enough. My Celestron EdgeHD 8” has been great. But the field is restricted when using the 0.7x reducer to about 28 mm. That has been a good match to the ZWO 294MC Pro, which gave me an economical entry into astronomy. But it seems the right time to go to a larger camera.
This quest for the right image scale is complicated by the availability of camera pixel sizes, when considering only the larger cameras. Basically, we only really have a choice between 3.76 um (Sony IMX 455, e.g. ZWO 6200) and 5.94 um (Sony IMX 410, e.g. ZWO 2400). This in turn means, for an image scale of 0.8 arc-second per pixel, that we are focussing on focal lengths around 970 mm (3.76 um pixels) or 1530 mm (5.94 um pixels). Of course, we could bin 2x2 the 3.76 um case to use 1940 mm (7.52 um pixels effective) at smaller FOV. These mostly argue for fast focal ratios as our aperture increases.
This figure endeavours to capture what’s out there in the way of suitable OTAs, although many are way outside my budget. It plots each as focal length versus image scale, with curves for 5 cameras with pixel sizes of 2.9, 3.76, 4.63, 5.94 and 10 um. The latter is too expensive, so not really an option. But it is representative of larger pixels, e.g. in CCDs. And there are no full-frame cameras with 2.9 and 4.63 um pixels. There are some APS-C sized cameras with around 4.78 um pixels (IMX 071, e.g. ZWO 071), which are not shown. Hence, most OTAs are plotted along the 3.76 and 5.94 um curves, with tick marks showing major axis FOV. The star symbols are grey for the smaller OTAs, green for 14” (350 mm) and red for 16” (400 mm) OTAs, the largest considered. The size of each symbol is proportional to pixel signal, including the effects of aperture and image scale.
The fast Newtonian Astrographs stand out, with some noteworthy exceptions, such as the ASA H400 F2.4, a very capable (and super heavy and expensive) platform with nearly 2.2° FOV. The Newts have small linear obstruction values and so lose fewer photons. A few cluster around the target 0.8 image scale, such as the 16” and 14” Hubbles and 16” Dreamscope and have sensitivity of 5-6x what I’m used to with the EdgeHD8. The table below summarises the vital statistics of the OTAs. Amongst the 14” scopes, the Hubble HNA14 stands out with a sensitivity of 5.1 compared to 2.3-3.8 for the rest. It’s nice to think of collecting the same photons per pixel in one hour that previously took 5 hours.
C74 (NGC 3132, mag 9.8, 2000 ly) Eight Burst PNe with very hot central white dwarf (100,000 C - we see less than 1% radiation in the visible). C59 (NGC 3242, mag 7.7, 4800 ly) Ghost of Jupiter PNe. PGC 19481 (mag 14, 300 Mly) Blue Ring galaxy, showing a ring of hot blue young stars believed to be triggered by a head-on galaxy collision.
These PNe are quite bright. But PGC 19481 is magnitude 14 and 300 Mly distant. Objects like it strain resolution and are starved for photons. I feel a rush of ‘aperture fever’ coming on …
Photons argue for more aperture, that’s inescapable, at least up to practical limits imposed by weight and cash. Resolution is largely a question of seeing, especially in my area. Discussions online about ideal image-scale tend to converge somewhere around 0.75 arc-seconds per pixel, with only ‘mountain top’ conditions warranting significantly smaller image scale. My best data (Centaurus A) used a PixInsight FWHM cut-off of 2.62 pixels and an average of 2.55, which is 1.64 arc-seconds. Nyqvist sampling (subdivide by 2 or more) would suggest an ideal image scale for this data of 0.82 or less. The data used 0.64 (EdgeHD 8” with 0.7x reducer), which worked well, but may have still been slightly over-sampled. However, guiding down at 0.3-0.4 arc-seconds RMS total is rare for me (2-3 days per year), and larger FWHM due to seeing is far more common. It seems that target image scale should be no smaller than 0.8 arc-seconds per pixel in my case. Hence, one goal of this exercise is to target ideal image-scale in order to maximise SNR. Put another way, I don’t want to squander SNR with an image scale that is too small.
Of course, I’d also like to target those juicy nebulae that we are blessed with in the Southern Skies, which becomes more a question of field of view, aggravated I suppose by going for long focal length to target DSO. Hence, I would also like to increase my FOV well beyond my current 0.75° x 0.5°. This argues for a full-frame camera and a OTA image train with a well corrected flat field large enough. My Celestron EdgeHD 8” has been great. But the field is restricted when using the 0.7x reducer to about 28 mm. That has been a good match to the ZWO 294MC Pro, which gave me an economical entry into astronomy. But it seems the right time to go to a larger camera.
This quest for the right image scale is complicated by the availability of camera pixel sizes, when considering only the larger cameras. Basically, we only really have a choice between 3.76 um (Sony IMX 455, e.g. ZWO 6200) and 5.94 um (Sony IMX 410, e.g. ZWO 2400). This in turn means, for an image scale of 0.8 arc-second per pixel, that we are focussing on focal lengths around 970 mm (3.76 um pixels) or 1530 mm (5.94 um pixels). Of course, we could bin 2x2 the 3.76 um case to use 1940 mm (7.52 um pixels effective) at smaller FOV. These mostly argue for fast focal ratios as our aperture increases.
This figure endeavours to capture what’s out there in the way of suitable OTAs, although many are way outside my budget. It plots each as focal length versus image scale, with curves for 5 cameras with pixel sizes of 2.9, 3.76, 4.63, 5.94 and 10 um. The latter is too expensive, so not really an option. But it is representative of larger pixels, e.g. in CCDs. And there are no full-frame cameras with 2.9 and 4.63 um pixels. There are some APS-C sized cameras with around 4.78 um pixels (IMX 071, e.g. ZWO 071), which are not shown. Hence, most OTAs are plotted along the 3.76 and 5.94 um curves, with tick marks showing major axis FOV. The star symbols are grey for the smaller OTAs, green for 14” (350 mm) and red for 16” (400 mm) OTAs, the largest considered. The size of each symbol is proportional to pixel signal, including the effects of aperture and image scale.
The fast Newtonian Astrographs stand out, with some noteworthy exceptions, such as the ASA H400 F2.4, a very capable (and super heavy and expensive) platform with nearly 2.2° FOV. The Newts have small linear obstruction values and so lose fewer photons. A few cluster around the target 0.8 image scale, such as the 16” and 14” Hubbles and 16” Dreamscope and have sensitivity of 5-6x what I’m used to with the EdgeHD8. The table below summarises the vital statistics of the OTAs. Amongst the 14” scopes, the Hubble HNA14 stands out with a sensitivity of 5.1 compared to 2.3-3.8 for the rest. It’s nice to think of collecting the same photons per pixel in one hour that previously took 5 hours.