The EQ3 DSO Challenge

[Art Gecko]

On 05/06/2017 at 21:48, bobro said:

As previously mentioned, I wondered what was the actual f# of my scope, what with the obstruction due to the secondary mirror, scope shortening to achieve focus and the addition of a coma corrector. Here are the results of using the camera with a 135mm lens at f# from 5.6 to 16, comparing with the camera on the scope at a nominal 5. Same exposure time and ISO for all images.

What you've found there Bob is a T-stop. The focal ratio hasn't changed, but the transmission of light has... Here's an extract from wikipedia that explains it better than me...

T-stop

A T-stop (for transmission stops, by convention written with capital letter T) is an f-number adjusted to account for light transmission efficiency (transmittance). A lens with a T-stop of N projects an image of the same brightness as an ideal lens with 100% transmittance and an f-number of N. A particular lens' T-stop, T, is given by dividing the f-number by the square root of the transmittance of that lens:

 

T = f / (transmission square root)

For example, an f/2.0 lens with transmittance of 75% has a T-stop of 2.3:

         T = 2.0/ (0.75 square root) = 2.309...

Since real lenses have transmittances of less than 100%, a lens's T-stop is always greater than its f-number.^[7]

Lens transmittances of 60%–90% are typical,^[8] so T-stops are sometimes used instead of f-numbers to more accurately determine exposure, particularly when using external light meters.^[9] T-stops are often used in cinematography, where many images are seen in rapid succession and even small changes in exposure will be noticeable. Cinema camera lenses are typically calibrated in T-stops instead of f-numbers. In still photography, without the need for rigorous consistency of all lenses and cameras used, slight differences in exposure are less important, however, T-stops are still used in some kinds of special-purpose lenses such as Smooth Trans Focus lenses by Minolta and Sony.

 

Had to ditch a couple of diagrams as they were coming out black with black background, so I wrote in the equations.

Art

Edited June 7, 2017 by Art Gecko

[bobro]

Thanks for looking into this Art. I hadn't come across the concept of a T-stop before - there's always something new to understand, making AP a great subject. No doubt you are aware that scopes such as the 130P-DS have the primary mirror moved a bit closer to the secondary mirror to help with DSLR focus. I've just read the 130P-DS has a slightly larger secondary mirror than the 130P to ensure light is not lost from the closer primary. With my scope the secondary is unchanged so there will presumably be some inefficiency there.

I wonder what the T-stop of my scope is......

[rotatux]

On 05/06/2017 at 22:48, bobro said:

Here are the results of using the camera with a 135mm lens at f# from 5.6 to 16, comparing with the camera on the scope at a nominal 5. Same exposure time and ISO for all images.

Very interesting. Though, for the sake of comparison, I'd like to see the scope's result without CC, to account for just the mirrors' transmission, and starting the lens at f/4.0.

Given that I've heard pixel individual illumination should be proportional to F ratio (all other conditions being equal), it's quite strange to observe your scope's F:5 result to be between the lens F:8 and F:11. That's a kind of flat images you made, but was your light source for them constant in intensity and distance ?

BTW I don't believe in the T-stop explanation, as that would make the result even worse, i.e. the lens results around the scope's would not be F:8 and F:11 but more something like F:9 and F:12 (assuming 80% transmission). Don't forget that your 2 newton's mirrors each have typically 95% reflectivity so overall ~90% transmission.

[alacant]

1 hour ago, Art Gecko said:

results of using the camera with a 135mm lens at f# from 5.6 to 16, comparing with the camera on the scope at a nominal 5

Hi. I'm -almost- convinced that comparing f# between optics of different aperture is misleading. Surely, a 150mm mirror collects more light than a 50mm diameter lens? Or maybe the latter collects more light because it grabs it from a larger area, hence the lens is brighter. Ahgghh, not this again!

[Art Gecko]

5 minutes ago, alacant said:

Hi. I'm -almost- convinced that comparing f# between optics of different aperture is misleading.

I agree, I quoted the T-stop explanation to demonstrate that Bob's f# hadn't changed.... the modifications he's made will have only altered the transmission of the light. The f# is still focal length/apeture and is governed by the primary mirror.

Why there is such a big difference between his f5 scope and his f5.6 lens..... I have no idea!!?

[alacant]

14 minutes ago, Art Gecko said:

such a big difference between his f5 scope and his f5.6 lens

I think it's because the lens is collecting light from a much larger area but concentrating it onto the same size sensor.

[Art Gecko]

4 minutes ago, alacant said:

I think it's because the lens is collecting light from a much larger area but concentrating it onto the same size sensor.

Agreed... From what I've read, that does seem to be the case.

[mikey2000]

https://en.wikipedia.org/wiki/F-number

A bit of light reading, if you will excuse the pun.

 

The key thing here is that "F Number" is simply a ratio:  focal length divided by aperture size.      Knowing the F Number is most helpful in determining the correct exposure for a non-astro photo.  I'm more from a photographic background and it took me a *long* time to realise that astronomers talk about their optics quite differently to photographers.  My telescope is a 150mm F5.  Using the astronomers terms, that means I can calculate the focal length at 750mm as the aperture is actually 150mm wide (main mirror diameter).    For my camera, I have a 135mm F4 lens.  But here, the 135mm refers to the focal length.  If I wanted to, I could calculate the physical diameter of the lens aperture as 135/4 cm.    If I had a 150mm F5 camera lens (in photographers' terms) it would be about the size of a beer can.  A 150mm F5 telescope is quite a beast by comparison.

 

For a telescope, we talk about the absolute size of the mirror (aperture) as that is more relevant to telling us how big the 'light bucket' is, catching extra photons and concentrating them down to a little tube of light that fits nicely into our pupil.

 

In photographic terms, I suppose the reason the *ratio* is important to exposure calculations is that the light is more 'stretched' by longer focal lengths.   Look at a field-side sports pro photographer with 600/2.8 lens.  It's massive!  a 24/2.8 lens isn't much bigger than a puck.  The same shutter speed for the same exposure though...

[bobro]

I don't know if you are old enough to remember the days of separate exposure meters and interchangeable lenses, but once a light reading was taken, film sensitivity and exposure time was set in the meter, the f# was the result. It then didn't matter what lens was put on the camera - any lens set to the same f# delivered the same amount of light to the film plane (within reason) and hence the resulting exposure was correct. That's why I wanted to measure the 'true' f# on my scope setup - basically comparing it with a camera lens which doesn't have the same obstructions and light loss. Art enlightened me on the concept of T-stop, which is a more appropriate way of measuring what I was looking for.

Using Excel, I previously did a basic calculation on my Newtonian scope, taking into account only the light loss due to the central obstruction - the result was about f5.6 (in non T-stop terms).

My thinking about exposure and f numbers comes from many years of photography with cameras, so I tend to look at things that way. It does make me see a refractor scope as just a big camera lens and I would expect an f# measurement to turn out just as would a lens. A reflector scope doesn't seem quite as efficient - especially after my mods! I'll try and re-measure without the CC as rotatux suggested.

[wimvb]

Haven't read all the posts in this topic, but be carefull when comparing "brightness". This is a term left over from film photography, but can be misleading in digital photography.

If you change the f-stop by changing the focal length, you also change the "resolution", as in how much of an extended object is covered by one pixel. A fair comparison of optical systems in digital photography has to take pixel size into account. Or try to normalise to a 1 um sized pixel (as Stan Moore tries to do in "Lessons from the Masters").

(I realise that this is throwing the discussion right into the jaws of the "F-ratio Myth", but I'm not going to discuss that. That poor horse has been beaten to death many times over already.)

I'm leaning more to using resolution ("/pixel) multiplied by aperture diameter, as a measure for light gathering power of an imaging system. This may not be entirely correct, but at least it takes into account pixel size.

In the discussion of images, here and in other online communities, the f-ratio isn't mentioned often, but aperture and total exposure time are. As you may be aware of, @gorann and I have been processing data from the Liverpool telescope recently. For this telescope (which operates at a slow f/10), total integration time is measured in minutes, rather than hours. But on the other hand the mirror is 2 m in diameter, and the pixel resolution is 0.3 "/pixel (binned) at 15 um pixel size. Apparantly, f-ratio is less important than aperture, and the f/10 system isn't that slow after all.

(Just tossing my € 0.02 worth into the discussion)

[bobro]

Here is a comparison of images taken with and without the (DIY) Coma Corrector on my Meade 130EQ scope. There doesn't seem to be too much light loss (though this is subjective) due to the addition of the CC. Commercial CCs are usually 2" aperture, where my DIY CC is 1.25". I don't know if that aperture difference will make any difference in light loss - possibly the light loss is mainly down to the CC lens glass in the optical path.

Edited June 15, 2017 by bobro

[Stub Mandrel]

2 hours ago, bobro said:

Here is a comparison of images taken with and without the (DIY) Coma Corrector on my Meade 130EQ scope. There doesn't seem to be too much light loss (though this is subjective) due to the addition of the CC. Commercial CCs are usually 2" aperture, where my DIY CC is 1.25". I don't know if that aperture difference will make any difference in light loss - possibly the light loss is mainly down to the CC lens glass in the optical path.

Interesting, most people assume that because the 0.9 coma corrector makes the f-ratio smaller it makes the image brighter.

[rotatux]

19 hours ago, bobro said:

Here is a comparison of images taken with and without the (DIY) Coma Corrector on my Meade 130EQ scope.

It appears your CC introduces more vignetting as there was initially. Meaning you will *need* flats to correct your images taken with it

The global level is also very different : 202 vs 217 (on the full-width bottom area below the label) which is about 7%. This is difficult to see (and hereforth quantify) with the eye+screen, so I used Gimp to measure (though I wanted the median, it only offered the mean, but that's another story). I think that result is not bad per itself (90-95% Tx factor is usual with good glass assemblies), but if you're comparing brightness variations between apertures, you should remember to apply that value as transmission factor of the CC.

Such an interesting subject that I'm now planning myself to check exposure levels from all my lenses/scopes x apertures, just to try to clarify whether focal+transmission ratio play a role alone, or must be combined with focal (have seen both theories defendable). Just need to build a DIY light lab to keep things constant :-P

[Peco4321]

M57, 30 x 30sec ISO 400 as the sky was bright this time of year, 20 darks and 30 bias, stacked in DSS and stretched in GIMP. Quite pleased as getting a bit more colour to this lovely planetary Nebula. Is the dot really the left over white dwarf or another star nearby? 

[wimvb]

Nice catch, at the very limit of what can be done with this setup, it seems. There really is a white dwarf in the center if this gem, and you seem to have caught it. When nights get darker again, you can try with different settings and maybe resolve it a bit clearer.

[Peco4321]

2 hours ago, wimvb said:

at the very limit of what can be done with this setup

I love taking my set up to the limits, with better polar alignment and darker skies, I hope to do a bit better another time. 

[wimvb]

That's the fun in this game that is AP. Whether you have a barndoor tracker or a MESU 200, getting the most out of your setup and having fun while doing so is what it's all about. IMO.

But of course, at some time you'll outgrow the barndoor tracker, but probably not the MESU 200. That's when you have to throw (a serious amount of) money at this hobby.

[bobro]

I liked Peter's M57, so had to have a go too, stretching my setup perhaps a bit too far...19 subs @ 120sec.

[Stub Mandrel]

6 hours ago, bobro said:

I liked Peter's M57, so had to have a go too, stretching my setup perhaps a bit too far...19 subs @ 120sec.

M57 responds really well to lots of subs + drizzling.

[Stub Mandrel]

First results with using Ha:

The Ha stacked at 50% resolution using DSS super pixel mode, was very noisy (it was hot night!) so denoised in Astra which took out some of the fine detail. 60 1 minute subs and 30 2-minute subs.

Mixed in a 50% lum with an ordinary DSLR image from last summer (but with different framing):

The original image without Ha:

 

[bobro]

4 hours ago, Stub Mandrel said:

First results with using Ha

It's certainly made a worthwhile improvement to the earlier image. What's the plan for imaging with the Ha filter - B&W, false colour, other filters, improving earlier images? Now I'm typing this I wonder if the answer is 'yes' to everything! 

[Stub Mandrel]

Initially I will revisit  ha-rich targets, although I may have a go a the Eagle nebula or with my 135mm lens on the crescent.

 

[wimvb]

Nice result, Neil. Did you also mix Ha with the red channel, or just as lum in the dslr data?

[Stub Mandrel]

So far I just mixed it 50/50 with the 'synthetic' LUM

[bobro]

Another attempt at M57 - as suggested by Neil: more subs + drizzling. Plus better focus and used the centre of the scope's image without the CC in place in case the CC degraded the image. I think that's about the best I can manage at the moment. 32 subs @ 60 sec.

Edited June 19, 2017 by bobro