Star Adventurer mount: small round stars at a low plate scale

[Andy Lucy]

The Skywatcher Star Adventurer equatorial mount is a lightweight portable tracker that is generally used for wide field astrophotography. I have been trying to see how far it can be pushed for astrophotography of deep space objects. My aim has been to get small, well resolved, round stars while using a camera and lens system that has a plate scale of as low as 1.6 arcseconds / pixel [Nikon J5 camera with a 20MP 1-inch sensor, 300mm PF f/4 lens (7.5cm max aperture)]. To put 1.6 arcseconds / pixel in context: atmospheric turbulence usually blurs objects by about 2 arcseconds; the diffraction Airy disc for a 7.5cm aperture is 3 arcseconds. The target was M13, since the tight cluster of stars is a real challenge to resolve.

To maximise the chances of getting round stars I have a system that minimises the load on the tracker. All the equipment on the Star Adventurer mount comes to just 3.5kg (camera, lens and hood, red dot finder, guide camera and scope, ball head, declination assembly, counterweight). The Star Adventurer wedge is mounted on a Berlebach Report 112 photo tripod.

Guiding was with a ZWO ASI 120MM Mini USB 2.0 Mono camera and a ZWO mini guide scope (30mm diameter, 120mm focal length).

Polar alignment was done initially with the polar scope on the Star Adventurer, then refined with the drift alignment utility on PHD2 guiding software. Once aligned, the declination drift rate averaged 1.4 arcseconds/minute, i.e. less than 0.5 pixel for a 30s exposure on the imaging camera.

During 2 hours of guiding PHD2 reported a typical RA error of 2 arcseconds. 187 30s images were captured.

Raw files were cropped and converted to tiff files. The best 15 files were stacked in deepskystacker. Final processing was done in lightroom.


M13 - Hercules globular cluster. Field of View = 1.1 degrees. Best viewed at 100%.

To my eyes, the faint stars look round (with a minimum diameter of about 3 pixels or 5 arcseconds), and those which are 7 arcseconds or more apart seem readily distinguishable. Bright stars show some slight distortions / edge effects. Overall, I’m very pleased with the result. My conclusion is that the Star Adventurer can do a fine job even with a system with a plate scale of 1.6 arcseconds / pixel.


Andy

 

[BlackgumNate]

Impressive, I think you achieved your goal.

I believe M13 is more yellow and red, and less blue stars. Accurate color would probably be even more pleasing as it would differentiate the stars further.

Nate

 

[Rudy Pohl]

Hi Andy,

It certainly sounds very good and looks good too.

Just to help me understand better since I'm not much of a techie, is what you did with your 1-inch sensor and 300mm lens the same as shooting a 600mm lens on a crop sensor Nikon camera like my D5500?

Thanks,
Rudy

 

[OllieN]

Very nice, gives me confidence in what I'm hoping to achieve.

Rudy, his is a crop factor of 2.7x compared to the 1.5x of our D5500. So (300mm x 2.7) / 1.5 = 540mm so if my method is right it's equivalent of 540mm for us.

Ollie

 

[Rudy Pohl]

Very nice, gives me confidence in what I'm hoping to achieve.

Rudy, his is a crop factor of 2.7x compared to the 1.5x of our D5500. So (300mm x 2.7) / 1.5 = 540mm so if my method is right it's equivalent of 540mm for us.

Ollie

Got it Ollie, thanks! That is impressive for sure!

Rudy

 

[Andy Lucy]

Nate,

Thanks for your comments.

Although I was pleased with the resolution in the image of M13, I would have liked better colours. The lack of good colour may be due to my lack of experience in post-processing (using Lightroom).

However, the J5 sensor has such small pixels (only 2.4 micron pixel pitch) that it has relatively poor dynamic range, tonal range and colour sensitivity (according to measurements by Dxomark). This may be the reason why the colours aren't very good.

I plan to do some tests with a camera with larger pixels (a Nikon D7100) to see if this makes a difference to the colours.


Andy

 

[Rudy Pohl]

Hi Andy, it's me again....

I've been reading through your text one more time and I need you to confirm something else for me to make sure that I understand your description.

Are you saying the following...
- that you shot 187 separate images
- that each image has a 30-second exposure time
- that from these 187 images you selected the best 15 images and stacked them
- that you processed this stack into the above final image of M13?

If all this is correct, then could I ask you why you would only use 8% of your total image pool to produce a final image? Is this because only 8% had stars that were round enough to yield a stacked image with acceptably round stars?

If you were to use 50-75% of your total images would your final stack still have round stars, and if not, how elongated would the stars be?

The reason I ask these questions is because I don't know if it's reasonable to conclude performance levels of any mount based on only a sample of 8% of the total images from a given session. I say this because I suspect that very few people would be willing to throw out over 90% of their images from a night of imaging, precious as these rare nights are, just for the sake of ending up with a handful of images with round stars.

Sorry if I'm misunderstanding something here.

Rudy

--
https://www.flickr.com/photos/rudypohl/

 

[1llusive]

Bright stars show some slight distortions / edge effects.

First, good work with a modest mount.

Second, a question: do you think the above is due to periodic error, lens aberrations, or the fresnel element?

 

[Andy Lucy]

Rudy,

Good questions. Yes to all of:

Are you saying the following...
- that you shot 187 separate images
- that each image has a 30-second exposure time
- that from these 187 images you selected the best 15 images and stacked them
- that you processed this stack into the above final image of M13?

I was taking the photos from my back garden (Bortle zone 5), and the target sank over the 2 hours of the session from an elevation of 48 to 30 degrees. Consequently, there was significantly more light pollution for the latter images and adding these images didn't help improve the stack.

Using the deepskystacker image scores to rank the images, I've attached the best image (before stacking), taken at 20.52, and the one ranked at the 90th percentile mark (i.e. nearly the worst) taken at 22.31. Both images were processed and cropped in lightroom identically. You should be able to see that the stars are still round for the latter, but the background is much lighter.

The worst images (90th to 100th percentile) were mostly when a tree got in the field of the guide scope.

Incidentally, comparing the images you can also see confirmation that the dec drift over the 99 minutes between the two images was low.


Best image (1st percentile).


90th percentile image

Regards,

Andy

 

[Andy Lucy]

I'm only guessing, but...

The effects are only noticeable for bright stars, which could be because of the Fresnel design of the lens

and

The effects are asymmetric, which might be because of some spherical aberrations

Regards,


Andy

 

[swimswithtrout]

Hi Andy, it's me again....

I've been reading through your text one more time and I need you to confirm something else for me to make sure that I understand your description.

Are you saying the following...
- that you shot 187 separate images
- that each image has a 30-second exposure time
- that from these 187 images you selected the best 15 images and stacked them
- that you processed this stack into the above final image of M13?

If all this is correct, then could I ask you why you would only use 8% of your total image pool to produce a final image? Is this because only 8% had stars that were round enough to yield a stacked image with acceptably round stars?

If you were to use 50-75% of your total images would your final stack still have round stars, and if not, how elongated would the stars be?

The reason I ask these questions is because I don't know if it's reasonable to conclude performance levels of any mount based on only a sample of 8% of the total images from a given session. I say this because I suspect that very few people would be willing to throw out over 90% of their images from a night of imaging, precious as these rare nights are, just for the sake of ending up with a handful of images with round stars.

Sorry if I'm misunderstanding something here.

Rudy

I'm with Rudy. I saw the OP's original photo and declined to comment on it due the horribly bad star shapes.


I can shoot 30 sec subs @ 300mm with an APS-c t 'ill the cows come home , no auto guider, on my little iOptron SkyTracker and not have to toss a single sub. I can shoot 75 sec subs using a 180mm lens, and not lose a thing, all perfectly round


30 sec subs is not a test. If you had to toss any, even without autoguiding, you have a seriously defective mount !!!

If you want to prove your point, shoot 30, 180 sec subs keeping every one of them and then post the results.

 

[Rutgerbus]

How do you like your mini guidescope with ZWO ASI120MM guidecam....does it rock! Easy to find a guidestar etc etc.

I would like to buy the same together with the New ASI Air to get a standalone autoguider for in the field!

 

[Astrophotographer 10]

Impressive, I think you achieved your goal.

I believe M13 is more yellow and red, and less blue stars. Accurate color would probably be even more pleasing as it would differentiate the stars further.

Nate

Rob Gendler is an icon in astrophotography and has the best sense of colour processing of anyone I know.

Here is one of his rendtions:

http://robgendlerastropics.com/M13NM.html

It may be that its a tad too blue as globular clusters are typically older smallish stars that are yellowish and often portrayed as more bluish white. Certainly a lot of globs are yellowish but the brighter ones like M13 may appear blue/white.

The largest glob is Omega Centauri - a southern hemisphere object and super impressive in a visual telescope look. 11 million stars and one theory is its an old dwarf galaxy swallowed up by the Milky Way (which has several companion smaller galaxies revolving around it).

I think this version I did of Omega Centauri is probably pretty close to the truth with regards to star colour - more yellow stars and less blue.

http://www.pbase.com/gregbradley/image/86746460/large

When I first saw this object through a telescope I couldn't believe it. Wow, what a sight. It fills the entire view of a decent telescope.

Greg.

 

[Astrophotographer 10]

Hi Andy, it's me again....

I've been reading through your text one more time and I need you to confirm something else for me to make sure that I understand your description.

Are you saying the following...
- that you shot 187 separate images
- that each image has a 30-second exposure time
- that from these 187 images you selected the best 15 images and stacked them
- that you processed this stack into the above final image of M13?

If all this is correct, then could I ask you why you would only use 8% of your total image pool to produce a final image? Is this because only 8% had stars that were round enough to yield a stacked image with acceptably round stars?

If you were to use 50-75% of your total images would your final stack still have round stars, and if not, how elongated would the stars be?

The reason I ask these questions is because I don't know if it's reasonable to conclude performance levels of any mount based on only a sample of 8% of the total images from a given session. I say this because I suspect that very few people would be willing to throw out over 90% of their images from a night of imaging, precious as these rare nights are, just for the sake of ending up with a handful of images with round stars.

Sorry if I'm misunderstanding something here.

Rudy

I'm with Rudy. I saw the OP's original photo and declined to comment on it due the horribly bad star shapes.

I can shoot 30 sec subs @ 300mm with an APS-c t 'ill the cows come home , no auto guider, on my little iOptron SkyTracker and not have to toss a single sub. I can shoot 75 sec subs using a 180mm lens, and not lose a thing, all perfectly round

30 sec subs is not a test. If you had to toss any, even without autoguiding, you have a seriously defective mount !!!

If you want to prove your point, shoot 30, 180 sec subs keeping every one of them and then post the results.

Well I think you are being a bit harsh. 30 seconds at 300mm round stars is a test.

In my experience 30 seconds with even a shortish lens can be way off quite easily despite "tracking" or despite a reasonably expensive mount.

I don't think I could get round stars routinely with 30 seconds 300mm APSc on my Vixen Polarie. Maybe if alignment is precise and balance is taken well care of, but even then - no.

Chinese mounts are improving but they can be a handful to tame. I have never used one but there are endless posts about upgrading them to be more workable.

For serious deep sky astro my first mission is to achieve round stars in 10 minute exposure and that can be a challenge no matter how high end your mount is.

For DSLR imaging that test could be scaled back but round stars at 60 seconds at 300mm might be a nice workable test.

Greg.

 

[Astrophotographer 10]

Bright stars show some slight distortions / edge effects.

First, good work with a modest mount.

Second, a question: do you think the above is due to periodic error, lens aberrations, or the fresnel element?

It might also have to do with poor seeing conditions and the small pixel size. Although 300mm would take a pretty small pixel. That setup even though only 300mm has a tiny pixel so its going to need decent seeing conditions (seeing is a term describing how calm or turbulent the atmosphere is and how much it prevents getting a sharp image). There is a reason why professional observatories are on mountain tops!

Greg.

 

[rnclark]

Impressive, I think you achieved your goal.

I believe M13 is more yellow and red, and less blue stars. Accurate color would probably be even more pleasing as it would differentiate the stars further.

Nate

Rob Gendler is an icon in astrophotography and has the best sense of colour processing of anyone I know.

Here is one of his rendtions:

http://robgendlerastropics.com/M13NM.html

It may be that its a tad too blue as globular clusters are typically older smallish stars that are yellowish and often portrayed as more bluish white. Certainly a lot of globs are yellowish but the brighter ones like M13 may appear blue/white.

I have communicated with Gendler about his color. I document the exchange here:

Loading…

www.clarkvision.com

"Hi Roger,
Thank you for the kind feedback about my images. I have no pretense about there being perfectly true color in my images. People should not hold up my images as highly precise representations of true color. Its such a controversial topic especially among amateur astrophotographers. Interesting that even visible light Hubble images are not precise true color images. For many of those images the red channel is 814nm which is actually near IR. That's why many of the galaxies imaged by HST are devoid of HII regions unless they specifically acquire 656nm data." - Robert Gendler

Again, the B-V data for an object is the best indicator of color we have. See this 1955 Astrophysical Journal article: http://adsabs.harvard.edu/full/1955ApJ...122..171A

You can do "print this article" and get the full paper. Figure 1 shows the color magnitude diagram. There are a few blue and white stars, but the majority of stars are redder than solar (B-V ~ 0.6). So M13 in natural color would look predominantly yellow. Reddest stars are B-V around +1.6.

I have observed M13 though a 31-inch telescope and it did look yellow with an impressive sprinkling of very red stars. I do not recall any blue stars.

The largest glob is Omega Centauri - a southern hemisphere object and super impressive in a visual telescope look. 11 million stars and one theory is its an old dwarf galaxy swallowed up by the Milky Way (which has several companion smaller galaxies revolving around it).

I think this version I did of Omega Centauri is probably pretty close to the truth with regards to star colour - more yellow stars and less blue.

http://www.pbase.com/gregbradley/image/86746460/large

When I first saw this object through a telescope I couldn't believe it. Wow, what a sight. It fills the entire view of a decent telescope.

Again the color magnitude diagram is key. It is less yellow than M13, more yellow-white.

Here is my attempt at natural color:

Loading…

www.clarkvision.com

Note that latest research indicates that Omega Centauri is the core of a galaxy captured by the Milky Way and not a globular cluster.

I agree, it is an amazing sight.

Roger

 

[Andy Lucy]

I found that the ZWO mini guidescope / mini ASI 120MM camera worked flawlessly for guiding - except when there was dew around. I fixed this with some tape and a sheet of insulating foam.... and it is now very dew-resistant, even if it's good looks are sadly impaired.







Andy

 

[Andy Lucy]

I have re-looked at the image which you described as having “horribly bad star shapes”. After a bit of experimentation I found that if I reduced the severity of some of my post-processing parameters I could get star shapes that, to my eyes, looked better. To kill two birds with one stone, rather than stack just a few subs I also stacked 90% of my subs (not 100% because for a period guiding was interrupted) and used the revised post-processing parameters to make the image below. What do you think of the star shapes in the result? – if the verdict is “still horrible”, please would you say what specific aspect of the appearance is the problem.


M13

To clarify the discussion below on tracking accuracy, I’d like to emphasise that pixel pitch of the sensor, along with the lens focal length, is an essential element in considering tracking difficulty. The critical factor is the plate scale, expressed in arcseconds / pixel. Consider the following four cameras each with a 300mm lens:

• Sony A7S (12 MP full frame sensor, 8.4 micron pixel pitch): 5.8 arcseconds /pixel
• Nikon D40 (6MP APS-C sensor, 7.8 micron pixel pitch): 5.4 arcseconds / pixel
• Nikon D5300 (24MP APS-C sensor, 3.9 micron pixel pitch): 2.7 arcseconds / pixel
• Nikon J5 (20MP 1-inch sensor, 1.6 micron pixel pitch) : 1.6 arcseconds / pixel

In these examples, the difficulty of achieving adequate tracking accuracy is higher for the Nikon J5 system than for the APS-C and FF systems.

Is my mount defective? There is plenty of data available to show that the typical tracking error of the Star Adventurer is 25 – 50 arcseconds, improving to roughly 2 - 5 arcseconds when guided. This is just what I find with my mount, so it appears – from the guiding data - to be working properly.

You have previously noted that the ioptron skytracker has worse performance than the Star Adventurer, with up to 100 arcseconds peak-peak periodic error, and that with a 300mm lens 30 seconds was the longest feasible exposure setting. Similarly, another report noted that it had “a maximum tracking error of 20 arc-seconds per minute of time”: http://www.covingtoninnovations.com/michael/blog/1302/index.html on the ioptron skytracker

When you say that you don’t “have to toss a single sub”, do you mean that you aren’t seeing errors of this sort of size (i.e. around10 arcseconds in a 30s exposure), or that you accept this size of error in your keepers? On a J5 sensor, a 10 arcsecond error would be around 6 pixels, which I would regard as unacceptable.

I’d like to take on your challenge of using longer exposures but unfortunately this isn’t possible with the J5 camera. Maximum timed exposure setting on the camera is 30s, there is no way to attach an external intervalometer and I can’t even use bulb mode since this is disabled when the FT1 adapter (needed to attach the 300mm PF lens) is used.

Regards,

Andy

 

[Alen K]

Seems to be an improvement but the funny assymetrical tails are still there (just fainter). I recently also shot M13 with a setup that yields a somewhat smaller image scale: 762mm focal length S-N telescope at f/5, 3.9 micron pixels with an APS-C DSLR, yielding 1.06 seconds per pixel. I also did 30-second subs but unguided and the scope was mounted on a Vixen GP-DX without PEC engaged, which may not actually track as well as a guided Star Adventurer. Here's the result with 14 subs (I only shot 16), for comparison.

PS. It's not great but we're only looking at star roundness, right?


M13, 14 x 30s unguided

 

[Andy Lucy]

Alen,

Thanks for the comments. You stars in M13 definitely look more round than mine.

I find that the trouble with dealing with non-round stars is that there are several things that can cause a problem.

Elongated stars because of tracking errors is an issue. At least this one is easy to diagnose.

The appearance of odd shaped stars can be mitigated by post-processing, but the question is really what causes the odd shapes in the first place.

I've just discovered that stopping down my lens from f/4 (wide open) to f/5.6 drastically changes the appearance of the stars. The lack of roundness is greatly reduced. For the brighter stars, where the appearance was worst, any remaining lack of roundness was partly masked by the appearance of sunstars that appear when stopped down.


Corner crop, f/4, Pleiades


Corner crop, f/5.6, Pleiades

I'm going to investigate a bit further how much the stars change as the f-number is increased. There is also the option of looking into stopping down with a step-down ring fitted to the lens rather than with the shutter blades.

Andy