What to change?

[Synchronicity]

I have a Sky-Watcher Esprit 80 ED with matching field flattener which will sit on my newly acquired, mint condition used NEQ6 pro (thanks Michael).  I recently purchased a 2nd hand QHY-10 OSC camera - I have to admit it was a spur of the moment decision without investigating camera /scope matching.

The CCD Suitability Calculator at https://astronomy.tools/calculators/ccd_suitability  tells me it wasn't a great choice and that I will be significantly under-sampling (see below).

As I see it my options are:

1. Do nothing.  I'm still learning and can use this setup to improve my knowledge and skills at aquiring images and processing them.  At some point the under-sampling will be the limiting factor and I can decide then whether I want to change camera, scope or both.
2. Change the camera for something around the same cost 2nd hand.  If I do that does anyone have any suggestions or recommendations?
3. Change the telescope to suit the camera.  The tool says it is a good match for the Esprit 120 but a good used one is probably outside my budget zone so I'd be looking for recommendations for a good 120mm refractor under £1000.  The smaller field of view would annoy me at times but I'll probably go there eventually.
4. Change both.  I like the simplicity of the refractor but I could investigate different scope types.  I probably like this option least because I'm just reaching the point I'll be able to set everything up, polar align, get on a target and guide well, enabling me to get good amounts of data for processing.  Up to now every night has been more experimentation, fixing problems or just getting everything running sweetly and the clouds drift over to watch me🤯🤯

Any thoughts, experience or advice would be most welcome.

Michael

 

 

[haitch]

Well they say don't dither but in this case that may be the best approach

http://www.astrosurf.com/buil/us/spe9/lrgb22.htm

[pete_l]

Since you already own the equipment you lose nothing by trying it out and relying on your own judgement as to whether it gives pleasing images.

So at this time, I'd give it  try.

[vlaiv]

Ok, that calculator is misleading.

First of all - 3"/px is definitively not gross under sampling, it is just wider field imaging, you don't have to worry one bit about being under sampled.

This is actually very good sampling rate to be when you have 80mm scope. It is much better to see what sort of FOV you'll be getting on particular targets and to choose targets that suit said FOV.

For example - M31 will barely fit into FOV with this combination unless you take care to orient it properly:

Pleiades will be framed nicely ...

So will Orion's nebula:

Heart nebula as well (here not rendered very nicely, but you can check it in Stellarium)

In any case, this is very nice combination for large number of objects and it will give you many imaging opportunities and will be nice skill honing platform - don't rush to change anything yet until you had a bit of fun with this combo.

[Synchronicity]

Thanks everyone, especially vlaiv for making me feel less of a fool!  I guess one of the pitfalls of reading lots but having little experience is sometimes not having good perspective on how significant an issue is, or is not.   40 years in electronic engineering, where indicators close to the right and in the red make me very uncomfortable probably didn't help 😁😁.

All the best

Michael

[iapa]

On 21/09/2020 at 18:43, vlaiv said:

Ok, that calculator is misleading.

Hi, I’ve trawling all sorts of stuff while deciding what equipment to put in my new observatory - for once this hobby I am really taking time to look at everything I can, and consider option other than JFDI.

Would be able to explain why you suggest that the CCD suitability calculator is misleading?

Please do this in words with as few syllables as possible for an idiot like me

thanks in advance

[iapa]

On 21/09/2020 at 15:40, Synchronicity said:

I have a Sky-Watcher Esprit 80 ED with matching field flattener which will sit on my newly acquired, mint condition used NEQ6 pro (thanks Michael).  I recently purchased a 2nd hand QHY-10 OSC camera - I have to admit it was a spur of the moment decision without investigating camera /scope matching.

 

OK, a bit late to the party.

Stick with it until you determine what you don’t like/doesn’t work for you, then consider change

[vlaiv]

33 minutes ago, iapa said:

Would be able to explain why you suggest that the CCD suitability calculator is misleading?

There is a thread that I started some time go about all the details, but in simple terms - because it wrongly represents when you over/under sample and does not take all important parameters into account.

Above is copied from Astronomy tools CCD suitability calculator.

There is mathematical relationship between star FWHM in the image and ideal sampling that I could show you (but it is quite a bit complex and I don't want to overload you, and I've written about it as well) that goes like this: star FWHM  / 1.6 = close to optimum sampling.

Now, if seeing is in range of 2-4", then star FWHM in the image will certainly be larger value than that, because seeing FWHM is only part of the blur that impacts star FWHM. Other bits are - tracking / guiding precision and aperture size. Smaller scopes will have larger star FWHM then larger aperture scopes for same seeing conditions. In any case - all these things add up to increase FWHM.

Even if we did not take that into account and took 2-4" FWHM at face value - optimal sampling would still be in 2" / 1.6 - 4" / 1.6 range or 1.25"/px - 2.5"/px

That is quite a bit different result than suggested by CCD suitability calculator.

I can show you step by step how above is derived, and also show you where in CCD suitability calculator explanation things are wrong, for example this:

Quote

In the 1920s Harold Nyquist developed a theorem for digital sampling of analog signals. Nyquist’s formula suggests the sampling rate should be double the frequency of the analog signal. So, if OK seeing is between 2-4” FWHM then the sampling rate, according to Nyquist, should be 1-2”.

That is taken from CCD calculator website as rationale for their calculations. However - it is wrong on several accounts.

First - it does not properly specify what Nyquist criteria is, and that is: For band limited signal, in order to completely and faithfully restore it - one needs to sample at twice maximum frequency component of that signal. So we are not talking about "any frequency of analog signal". Then it goes to equate FWHM with that frequency and use half of that - which simply can't be done as FWHM is not directly related to max frequency component of signal. It is related in non trivial way and Fourier transforms must be used as well as certain approximations to be able to arrive to particular number (which is x1.6 that I mentioned above if one uses 10% energy cutoff in frequency domain because there is no clear cutoff for Gaussian approximation to actual PSF - that is serious Math stuff that we can touch upon if you wish).

Then it later goes to say - its better to use 1/3 instead of 1/2 - for no particular reason and giving only vague explanation of pixels covering the star image - which are simply wrong as Nyquist gives full explanation and no further work is needed except properly understanding and applying it.

Overall - it is very unlikely that any amateur setup will need to sample at 1"/px or less (I personally consider 1"/px to be useful limit for 99% of imaging) - yet CCD suitability will recommend combinations of camera / FL that go higher than that without warning nor mentioning that in such circumstances one should bin their data.

[iapa]

30 minutes ago, vlaiv said:

There is a thread that I started some time go about all the details, but in simple terms - because it wrongly represents when you over/under sample and does not take all important parameters into account.

Above is copied from Astronomy tools CCD suitability calculator.

There is mathematical relationship between star FWHM in the image and ideal sampling that I could show you (but it is quite a bit complex and I don't want to overload you, and I've written about it as well) that goes like this: star FWHM  / 1.6 = close to optimum sampling.

Now, if seeing is in range of 2-4", then star FWHM in the image will certainly be larger value than that, because seeing FWHM is only part of the blur that impacts star FWHM. Other bits are - tracking / guiding precision and aperture size. Smaller scopes will have larger star FWHM then larger aperture scopes for same seeing conditions. In any case - all these things add up to increase FWHM.

Even if we did not take that into account and took 2-4" FWHM at face value - optimal sampling would still be in 2" / 1.6 - 4" / 1.6 range or 1.25"/px - 2.5"/px

That is quite a bit different result than suggested by CCD suitability calculator.

I can show you step by step how above is derived, and also show you where in CCD suitability calculator explanation things are wrong, for example this:

That is taken from CCD calculator website as rationale for their calculations. However - it is wrong on several accounts.

First - it does not properly specify what Nyquist criteria is, and that is: For band limited signal, in order to completely and faithfully restore it - one needs to sample at twice maximum frequency component of that signal. So we are not talking about "any frequency of analog signal". Then it goes to equate FWHM with that frequency and use half of that - which simply can't be done as FWHM is not directly related to max frequency component of signal. It is related in non trivial way and Fourier transforms must be used as well as certain approximations to be able to arrive to particular number (which is x1.6 that I mentioned above if one uses 10% energy cutoff in frequency domain because there is no clear cutoff for Gaussian approximation to actual PSF - that is serious Math stuff that we can touch upon if you wish).

Then it later goes to say - its better to use 1/3 instead of 1/2 - for no particular reason and giving only vague explanation of pixels covering the star image - which are simply wrong as Nyquist gives full explanation and no further work is needed except properly understanding and applying it.

Overall - it is very unlikely that any amateur setup will need to sample at 1"/px or less (I personally consider 1"/px to be useful limit for 99% of imaging) - yet CCD suitability will recommend combinations of camera / FL that go higher than that without warning nor mentioning that in such circumstances one should bin their data.

I’m vaguely familiar with Niquist’s sampling ideas, seem to recall that became regarded as a little outdated as it became possible to sample at much higher frequencies in later years.

Agree translating this to FWHM isn’t automatic with out explanation.

Can you provide the link to your thread mentioned please? FFTs used to make my head hurt

thanks for the help.

[vlaiv]

47 minutes ago, iapa said:

I’m vaguely familiar with Niquist’s sampling ideas, seem to recall that became regarded as a little outdated as it became possible to sample at much higher frequencies in later years.

It's still pretty much proven mathematical theorem (https://en.wikipedia.org/wiki/Nyquist–Shannon_sampling_theorem) and as such - can't really become outdated. Sure you can sample at much higher frequencies - but what is the point if you have band limited signal? You can perfectly reconstruct it with sampling at Nyquist frequency - no need for higher frequencies.

With imaging, sampling at higher rate actually hurts your SNR as light is spread over more pixels and each pixel gets less signal as a consequence.

Here is thread that I started:

 

[iapa]

I’ve got a raging headache - time to take 2 aspirin and call you in the morning

 

[vlaiv]

1 minute ago, iapa said:

I’ve got a raging headache - time to take 2 aspirin and call you in the morning

 

Well, it boils down to this:

1.2"/px - 1"/px : use large scope (8"+) of good figure on premium mount in good seeing conditions to fully exploit that resolution

1.5"/px - 1.2"/px : use largish telescope (5"-8") of a good figure on vg to premium mount in good seeing conditions

1.8"/px - 1.5"/px : this is "standard" high resolution range for most people. Use 4"+ aperture on good enough mount <1" RMS guiding (preferably below ~0.7" RMS total)

1.8"/px - 3"/px is normal range of resolution where you don't chase close up / detail but rather determine resolution based on focal length, pixel size, sensor size .. all the usual stuff, most mounts will work as well as scopes from 60mm upwards (but one should really move towards 2"/px+ with imprecise mounts and smaller scopes).

3"/px and above is wide field stuff. Can be said that it is over sampling in some cases - but you are far away from seeing aliasing issues (or blocky stars for that matter - those don't exist really) and you can happily over sample. In fact - at lower resolutions you hit the issue of pixel size and focal length and you can only image with lenses rather than scopes (below 300mm of FL) - which brings in another thing - lenses are not like scopes, they are not diffraction limited and have blur of their own, so you again stop being over sampled because of sharpness of optics - or rather lack of it.

[iapa]

1 hour ago, vlaiv said:

Well, it boils down to this:

1.2"/px - 1"/px : use large scope (8"+) of good figure on premium mount in good seeing conditions to fully exploit that resolution

1.5"/px - 1.2"/px : use largish telescope (5"-8") of a good figure on vg to premium mount in good seeing conditions

1.8"/px - 1.5"/px : this is "standard" high resolution range for most people. Use 4"+ aperture on good enough mount <1" RMS guiding (preferably below ~0.7" RMS total)

1.8"/px - 3"/px is normal range of resolution where you don't chase close up / detail but rather determine resolution based on focal length, pixel size, sensor size .. all the usual stuff, most mounts will work as well as scopes from 60mm upwards (but one should really move towards 2"/px+ with imprecise mounts and smaller scopes).

3"/px and above is wide field stuff. Can be said that it is over sampling in some cases - but you are far away from seeing aliasing issues (or blocky stars for that matter - those don't exist really) and you can happily over sample. In fact - at lower resolutions you hit the issue of pixel size and focal length and you can only image with lenses rather than scopes (below 300mm of FL) - which brings in another thing - lenses are not like scopes, they are not diffraction limited and have blur of their own, so you again stop being over sampled because of sharpness of optics - or rather lack of it.

I was planning 10” F4, 1000mm FL, Newtonian on CGX-L (0.78”/px). Seems I will need to bin 2x2  to get 1.57”/pixel in purportedly Bortle 3-4 skies?

[vlaiv]

5 hours ago, iapa said:

I was planning 10” F4, 1000mm FL, Newtonian on CGX-L (0.78”/px). Seems I will need to bin 2x2  to get 1.57”/pixel in purportedly Bortle 3-4 skies?

Bortle class does not play into this.

Idea of binning 2x2 is a sound one and I support that line of reasoning.

I also think that it will be good working resolution given your setup.  What is usual total guide RMS you achieve with CGX-L? If it is around 0.5"-0.7" RMS, then you'll be optimally (or close to optimally) matched in your setup.

Just make sure you use good coma corrector - one that does not add spherical aberration and can correct over needed field at F/4. Poorer coma corrector will make image softer than it needs to be - robbing you of some resolution.

[ollypenrice]

The old combination of Tak FSQ106 and Kodak 11 meg chip was a consistent collector of NASA APODS for years. 3.5 arcsecs per pixel.

I'm not into competitions but here are some examples of 3.5 arcsecs per pixel:

Rosette, Ha OIII LRGB 25 hours. ( Olly Penrice ) - AstroBin

NGC7000 widefield with Paul Kummer ( Olly Penrice ) - AstroBin

Breaking Wave around Gamma Cass from Les Granges. ( Olly Penrice ) - AstroBin

Sh2 126 Two panel. ( Olly Penrice ) - AstroBin

Don't fall foul of obsessing over the numbers. It's all about taking pictures.

Olly

[iapa]

46 minutes ago, vlaiv said:

Bortle class does not play into this.

I was more saying I got better skies nya nya nya thhhhppt

😈

 

[iapa]

1 hour ago, vlaiv said:

Bortle class does not play into this.

Idea of binning 2x2 is a sound one and I support that line of reasoning.

I also think that it will be good working resolution given your setup.  What is usual total guide RMS you achieve with CGX-L? If it is around 0.5"-0.7" RMS, then you'll be optimally (or close to optimally) matched in your setup.

Just make sure you use good coma corrector - one that does not add spherical aberration and can correct over needed field at F/4. Poorer coma corrector will make image softer than it needs to be - robbing you of some resolution.

I’ll be starting with f4 aplanic corrector, purportedly non-reducing - i’ll be checking focus both with and without to confirm.

Will then investigation any spherical aberration - if this is possible in a back yard set up with no magical optic measuring gear.

Edited June 21, 2023 by iapa

[iapa]

On 21/09/2020 at 15:40, Synchronicity said:

I have a Sky-Watcher Esprit 80 ED with matching field flattener which will sit on my newly acquired, mint condition used NEQ6 pro (thanks Michael).  I recently purchased a 2nd hand QHY-10 OSC camera - I have to admit it was a spur of the moment decision without investigating camera /scope matching.

The CCD Suitability Calculator at https://astronomy.tools/calculators/ccd_suitability  tells me it wasn't a great choice and that I will be significantly under-sampling (see below).

As I see it my options are:

1. Do nothing.  I'm still learning and can use this setup to improve my knowledge and skills at aquiring images and processing them.  At some point the under-sampling will be the limiting factor and I can decide then whether I want to change camera, scope or both.
2. Change the camera for something around the same cost 2nd hand.  If I do that does anyone have any suggestions or recommendations?
3. Change the telescope to suit the camera.  The tool says it is a good match for the Esprit 120 but a good used one is probably outside my budget zone so I'd be looking for recommendations for a good 120mm refractor under £1000.  The smaller field of view would annoy me at times but I'll probably go there eventually.
4. Change both.  I like the simplicity of the refractor but I could investigate different scope types.  I probably like this option least because I'm just reaching the point I'll be able to set everything up, polar align, get on a target and guide well, enabling me to get good amounts of data for processing.  Up to now every night has been more experimentation, fixing problems or just getting everything running sweetly and the clouds drift over to watch me🤯🤯

Any thoughts, experience or advice would be most welcome.

Michael

 

 

Going back to the original question, based on the commentary subsequent to your query, if your resolution sits between the lower or upper seeing limits divided by 1.6, e.g. it seems you will have a reasonable setup.

you may need to bin the images to fit within the mofified seeing limits, 1.5 and 2.5 for good seeing.

[vlaiv]

23 minutes ago, iapa said:

Will then investigation any spherical aberration - if this is possible in a back yard set up with no magical optic measuring gear.

I don't know of any simple method except comparing stars in center of the field with and without coma corrector.

If they look the same - then you are ok.

Here are a few interesting links comparing various coma correctors:

https://www.astrofotoblog.eu/?p=856

and this one:

https://www.cloudynights.com/topic/554686-coma-corrector-compariosn/

Note on the second link - there are a few comparison images that clearly show star bloat / softness of the image in center of the field with simple 2 element coma correctors.

 

[iapa]

Thanks, 

had a look at the first, and its given me ideas on how test. 

Doubt that the use of a CCD v the CMOS will make a difference.

I’ll publish results for my corrector in due course