Beginner help required!

[Adrian J]

Hello all,

I've recently become interested in astrophotography again and pulled out my WO Zenithstar 80ED, purchased a ZWO ASIAir Pro controller, ZWO ASI 294MC-PRO camera and  ZWO ASI120MM guider and it’s all mounted on an HEQ5 Pro. I have the belt mod kit but not installed it yet.

I’ve starting imaging on M31 and reasonably pleased with the results but I’ve seen so much better.

Before purchasing a new scope, possibly a triplet, I thought I would ask for some much valued feedback from peeps who know much more than myself.

My main concern is that stars appear to be bloated and I dont know where to start trying to improve things.

I had good p.a., a stable base, nothing lose anywhere and I think I had the focus pretty much spot on.

I would greatly appreciate any feedback but I’m out of my depth when it comes to a lot of the technical jargon and learning all the time.

Here’s a link to my full size image taken with 32 flats x 180 sec’s and processed in APP.  I’m reasonably good with Photoshop and I cam minimise the stars in PS but that doesn’t really help against the bright core in M31 …

https://www.flickr.com/photos/41279068@N08/50552621392/in/photostream/

 

 

[vlaiv]

That is very nice looking image.

Yes, stars seem a bit bloated, and there are a few possible reasons for that.

First would be use of 80ED scope. This is ED doublet scope so color correction will not be the best. Fact that you are using OSC camera is going to make things a bit worse (people with mono camera and RGB filters can refocus between filters which helps a bit). In order to see how bad things really are due to this point - just look at R, G and B channels separately.

If possible - measure star FWHM for each channel on same stars and see if you can spot the pattern. If blue channel is the worst, followed by red channel and then green channel (this is the order I would expect from an ED doublet) then telescope might be the culprit for bloated stars.

Solution to this one is to either switch to triplet scope with good color correction or possibly use additional UV/IR cut filter that is a bit more restrictive - like Astronomik L3 filter for example.

It cuts away most offending parts of the spectrum.

Second reason can be related to guiding / tracking and working resolution. What is your guide performance like in arc seconds? You really need to be below 1" RMS to get tight stars. Fitting belt mod will certainly help things there.

Third possible cause is light scatter due to high humidity. This creates "wings" around stars more than anything else. It looks more like some sort of glow around tight star cores than bloated star, but if you stretch your data aggressively (and we always tend to do that when processing image for faint detail) - your stars will end up looking large.

You'll see if this is in effect by checking for actual FWHM in the image and remembering what sort of weather was on particular night. This high humidity happens in fall (at least in my climate) when weather is rather nice during the day - high pressure, sunny and no wind. Seeing on those nights is usually good but due to high pressure and no wind - there is going to be a bit of mist/moisture in the air. Things dew up rather easily in these conditions.

I just looked up your scope and it is FPL-51 F/6.8 doublet scope.

My money is on the first cause. There is a way to process image in special way as to minimize this effect. Works good with galaxies but not as good - or rather poorly with emission type nebulae. Use green channel as luminance layer. Green channel is often the sharpest of the three if there are issues with residual chromatic aberration. Green channel will have tightest stars. Another good thing is - with OSC cameras you have twice as many pixels capturing green then blue and red and green is the most sensitive region of camera response - it will have best SNR for broad band targets.

You should also consider using super pixel mode debayering. I was not able to view your image at 1:1 because flickr offers bunch of resolutions but no indication of which one is native. In any case, super pixel debayer mode will make smaller image in pixel width/height but it will look sharper on 1:1 inspection.

Another little thing - background looks too dark / clipped. Maybe try not to clip the histogram as much?

 

[Adrian J]

I very much appreciate your reply Vlaiv

There are some things where I need to educate myself such as FWHM and super pixel mode debayering but I'm understanding most of what you're saying.

I've viewed many images looking for equipment used and settings and was coming to the conclusion that my doublet might be the cause of bloated stars and also noticed that as the scope gets closer to the zenith my images appear softer which I put down to the mount and / or balancing and will fit the EQ5 belt modification shortly.

I believe there must have been high humidity when I took my flats, all looked good to the naked eye in bortle 4 but my equipment was soaking wet at the end of the session.  I have dew heaters and know that the lenses werent misted up but that's a very interesting point.

Unfortunately I dont know what the RMS was, I was watching the graph but wasnt understanding what it meant in terms of star sharpness so that's another one where I need to understand what's happening.

Using APP there are many different settings but the defaults seem to work okay, the image I posted on Flickr was processed using the basic software tools and I bumped up the contrast to give a black background showing the stars and probably far too much.

Also knowing that more light should give better results I've recently been considering a reflector (150mm or 200mm) but I'm still in early days and looking for advice before spending money and regretting my decision.  I think I'll stick with a refractor though for weight and portability, maybe something like the WO GT-81 IV and working backwards to eleminate similar problems in the future.

I've uploaded the stretched tiff image from APP in native form prior to any colour calibration to drop-box  https://www.dropbox.com/s/wa4sdb68gt12hdj/M31_25-9-20-RGB-session_1-St.tiff?dl=0

Many thanks for your comments.

[vlaiv]

FWHM is full width at half maximum. Basically a measure of how "fat" the star is by specifying "waist" diameter

It was noticed that regardless of the star intensity - most stars in the image share this value (or are very close in value). Stacking software often calculates this for you, or you can use software like AstroImageJ (java based open source image processing software based on ImageJ - which was made for scientific analysis - primarily for microscope images) to find this values.

It will be expressed in pixel units, but if you know your sampling rate - you can convert to arc seconds. Very high resolution images have this value at about 1.5". Professional observatories on mountain peaks have this value often below 1". For the equipment that you are using, you should be happy with 3-4" FWHM stars (star FWHM depends on aperture size as well among other things).

Super pixel debayering is just a type of debayering often used in stacking software. Since Bayer matrix on OSC sensors has every other pixel "colored" differently - there are pixels that capture red, green and blue light and are interleaved, regular debayering algorithms interpolate missing values. This makes image soft and does not correspond well to actual sampling rate of OSC cameras. Super pixel mode uses different approach - it takes 2x2 group of pixels that have different colors (one red, one blue and two green) and creates one single pixel with three colors out of them.

It effectively reduces image size by 2 in both height and width and acts very similar to binning of mono sensors in that regard.

Result is smaller (in terms of resolution / pixel count) image that is sharper when viewed at 100% zoom.

RMS is short for "root mean square" and is just that - when you have bunch of values it calculates square root of mean of their squares. This is sort of average for errors and such - when you have both positive and negative values that represent some error and would cancel out in regular average - way to calculate their average without them canceling.

It shows how good your guiding is as it represents average error of where your mount is pointing (as opposed to where it should point).

If you use PHD2 for guiding it will be displayed to you in stats window:

Values below 0.3-0.4" represent exceptional guiding results. values that are in 0.4-0.6" represent very good guiding. 0.6-0.8" are good guiding, 0.8"-1.0" average and above 1" is below average guiding.

Mind you, even if you guide and have round stars - it does not mean your guiding is good. How tight your stars are - that is true measure of guide performance. Another important point -  above scale is just rough guide and actual guide performance depends on working resolution. Good rule of thumb is that you want your guide RMS to be at or below half of imaging resolution.

This means that if you image with lens at ~100mm and resolution of let's say 6-8"/px, then 1" RMS guiding will be excellent guiding with respect to impact on the image. Same RMS guiding value when imaging at 1"/px is going to be very poor result.

Do you by any chance have image that is still linear, prior to any stretching? Any measurement or inspection of the data is best done when data is still linear. You can't measure FWHM on stretched image as stretching changes the "shape" of the stars.

[Adrian J]

Hi Vlaiv, thanks again, I can see I have a way to go for a full understanding of AP, a long way ...

I've uploaded the unstretched image to drop box https://www.dropbox.com/s/t1dzprhif5dnb3j/M31_25-9-20-RGB-session_1-NoSt.tiff?dl=0

[vlaiv]

Well, I'm not sure what it is that I'm seeing here.

I suspect that is maybe a bit off all of there cases that I listed above. First optics.

It definitively has some residual color issues, here are examples:

These are stars in the center of the image (color balanced to show issue / blue boosted):

This is red, green and blue channel separated and montage made out of them. Red has larger stars, green looks good, and blue has some color fringing around stars - unfocused light.

When color combined, here what stars look like:

If you look carefully you'll be able to spot that red / bluish edge on some of the stars.

Second - you need some sort of field flattener. This scope has rather curved field and it shows in corners. Again, next to astigmatism, there are color issues as well:

Again, red, green and blue channels - all three showing astigmatism "cross". When color balanced and color combined, here is what it looks like:

Again, rather colorful as R, G and B don't really match in bloat level.

Stars are rather big even at this sampling rate - 3px FWHM at center and going even up to 7px-8px in the corners (thanks to field curvature). This data will benefit from being debayered with super pixel mode, or there is another way to do it - binned in software x2.

 

[Adrian J]

Hi Vlaiv, last week I ordered a flattener / reducer (WO adjustable Flat6A III 0.8x) as I could see the elongated stars in the corners also knowing it would be compatible with the WO GT-81 IV if I go in that direction.

I need to think more about your answers to fully understand them but I think the conclusion is the ED 80 doublet optics are not going to give me the ultimate sharpness I'm looking for and probably time to retire my scope and look for something better.

The alternatives I'm considering are a 150mm reflector e.g. Sky-Watcher Explorer 150PL (I was considering a 200mm but the weight is probably too much for the HEQ5) or the WO GT-81 IV.  I dont want to upgrade to the EQ6 as the weight will be a problem for me carrying it from place to place.

In my position, would you think the WO GT-81 IV would be the best way forward?  I dont want to spend more than I need to but willing to pay for something which I wont need to upgrade in the foreseeable future.

I really appeciate the time you've spent helping me with your amazing knowledge

Many thanks again.

[vlaiv]

I think that 150PL is going to be a bit much in focal length - it is 1200mm fL if I'm not mistaken. It is also rather large scope and will have big arm momentum. I used 8" F/6 scope on heq5 and it was no picnick.

If you want to go that route - 6" and above meter of focal length - take look at this scope:

https://www.firstlightoptics.com/stellalyra-telescopes/stellalyra-6-f9-m-crf-ritchey-chrtien-telescope-ota.html

I have no idea what the WO GT-81 is like, but if it's anything like TS Photoline 80mm f/6 apo (often these scopes share same source only different branding and accessories) - it will be very good. I have later and I'm very pleased with it.

[Adrian J]

The StellaLyra 6" F/9 looks very nice indeed and flexible options with x0.5 and x0.33 reducer, I've been looking through the topic on SGL but havent had time to read all 12 pages yet.

I take your point about the momentum and F/L with the 150mm, I'll give that one a miss.

Are you thinking of buying a StellaLyra 6" or happy enough with the TS 80mm?  FLO dont sell the TS otherwise I would buy one, the only other options are to buy from outside the UK but I'm always cautious about transportation.  Probably just as risky being sent by courier in the UK though!

Based on what you say and I dont think I've ever seen any criticism of WO, I think I'll probably go with the GT-81 IV.

Just for information, the WO ZS 80ED I have is serial number 100001.  I didnt know when I bought it second hand and an email from the big cheese William himself confirmed it was the first off the production line in 2010.  Quite amusing really

[Adrian J]

Hi Vlaiv, no need to reply I'm sure you're really busy.  I've ordered the GT81 from FLO.

Thanks for helping me on my AP journey

 

 

[vlaiv]

4 hours ago, Adrian J said:

Are you thinking of buying a StellaLyra 6" or happy enough with the TS 80mm?  FLO dont sell the TS otherwise I would buy one, the only other options are to buy from outside the UK but I'm always cautious about transportation.  Probably just as risky being sent by courier in the UK though!

I have 8" version from GSO, so I use that when I want long focal length.

[Adrian J]

Which mount do you use for the 8"?

[vlaiv]

2 minutes ago, Adrian J said:

Which mount do you use for the 8"?

HEQ5

It's been modded quite a bit - belt mod, saddle plate and puck replaced, Berlebach planet tripod ...

I now guide it at about 0.5" RMS on a good night.

[Adrian J]

Wow, that's really interesting.  You've adapted the HEQ5 'head' to mount on a Berlebach planet tripod?

If so, the belt drive will cope with the additional weight and it's the flex / instability in the tripod that limits the performance?

Sorry to be a real pain asking all these additional questions and I'm so grateful for your replies.

[Adrian J]

Ah, I've found the info on the saddle and puck replacement, beginning to undertsand now ...

[vlaiv]

2 minutes ago, Adrian J said:

You've adapted the HEQ5 'head' to mount on a Berlebach planet tripod?

There are ready made tripod adapters by Berlebach, so I got one to go with my tripod:

https://www.teleskop-express.de/shop/product_info.php/info/p8483

They provide these for almost any mount type so its worth getting.

It is extremely stable platform and when imaging on the ground, you can push it into the ground (has spikes) for additional stability. I did this to improve guiding performance.

Mount itself is capable of handling 8" RC scope weight wise without any issues. Scope and camera and OAG and balance weights - total at probably less than 11Kg. I used almost 15Kg of gear on Heq5 and although it worked - I was not feeling comfortable with that. This is short scope so 11Kg is quite ok for imaging.

Saddle plate is combined Losmandy / Vixen. Larger connection adds stability. Belt mod of course helps with guiding.

[Adrian J]

Brilliant, extremely useful information, I will leave you in peace now

[Adrian J]

Hi Vlaiv

Apologies for getting in contact again but when I encounter a problem my brain goes into overdrive trying to find solutions and working through the valued information you gave me I've found relevant data in APP which is only seen by scrolling right on the information window.  I'm still learning APP the hard way as instructions / videos are a bit scarce

RMS (Registration RMS) and FWHM values for 30 light frames shown below but I cant understand how these RMS values relate to your comment that below 0.3-0.4" represents exceptional guiding and above 1" is below average guiding when worst value shown is 0.13.  Does this make sense to you?

[vlaiv]

1 minute ago, Adrian J said:

RMS (Registration RMS) and FWHM values for 30 light frames shown below but I cant understand how these RMS values relate to your comment that below 0.3-0.4" represents exceptional guiding and above 1" is below average guiding when worst value shown is 0.13.  Does this make sense to you?

RMS can be thought as average for collection of errors. I've written about it above - it is square root of average of squares as regular average would lead to 0 error if there was balance between negative and positive values.

Important part is not how RMS is calculated - important part is what it represents - average error.

When guiding, RMS value will be average error of pointing of the mount - where it should point in comparison to where it is pointing. This is calculated by examining position of a guide star on sensor over period of time. Ideally - guide star should stay still - in the same place. This would mean perfect guiding.

In reality what happens is following - software captures image of the guide star, calculates its position - compares to previous/expected position and if finds error - adds to running average and issues correction. Later being important part for guiding while former just giving us some indication of how well our mount is tracking. Remember, guiding corrects for error after it has already happened - larger that error - more impact to the image.

Registration RMS is something entirely different. Again, it is average error of some sorts, but this time it represents measure of how well two subs match after alignment.

In order to align two subs, software analyzes each sub for stars - finds the stars and their centers (and also other statistics on the stars like their FWHM values). After it has found stars in each image - software tries to find best matching transform in order to align stars in both images.

It calculates something like - shift 3.23px to the right, and 1.12px up and rotate by 0.1 degrees or similar. It is actually represented by matrix (the transform, and whole process is done by least squares method).

Stars won't align perfectly in two images. There are many factors for this:

- there is noise in both subs and due to noise it is impossible to accurately determine star center

- there is finite precision in sampling of both images (finite number of pixels) - again that limits precision in finding star centers

- there is certain level of "lens" distortion and other effects in the image that impact different stars differently (as you have seen on your images - stars in center are round but those at the edge - astigmatic). As you dither (either purposely or "natural" dither due to shift in the system of some sort) stars end up in different parts of the frame and have different intensity of such effect on them. This changes their "center" somewhat (there are symmetric aberrations and asymmetric aberrations - later change star center) and this leads to error in alignment.

- due to seeing and star "jumping around" somewhat, actual star center can change position between frames (although seeing averages out in long exposure - there is always some residual error, small but still present).

In any case, after registration is performed - there will be some error in alignment of star centers - that is registration RMS.

In your case - this error being as low as 0.1 or there about just means when that particular frame was transformed, average error between where star should have ended up and where it really ended up is about 1/10 of a pixel. For all intents and purposes that can't be visually perceived.

These two RMS values are not related (they have one common component - seeing, but impact is so small that there is virtually no correlation between guide RMS and registration RMS).

Part of guiding RMS is due to seeing and not due to mount tracking error, but we try to minimize that - otherwise we end up chasing the seeing instead of correcting for mount tracking error. If you suspect that your guide RMS is higher because of poor seeing - just increase guide exposure as that averages seeing out a bit more. Although many guide at 1s intervals - I've found that I like to use at least 3-4 seconds of guide exposure, but in order to do so - you need smooth(er) running mount (belt mod ).

In the end FWHM value is related in part to guide RMS. With "perfect" scope, we can sort of write down very decent approximation of expected FWHM value:  FWHM = sqrt(guide RMS^2 + airy_disk_size^2 + seeing_blur_size^2)

(one needs to use corresponding values for each of these and relation between RMS and FWHM for gaussian is x2.355, also one needs gaussian approximation for airy disk as well).

 

 

[Adrian J]

Thanks Vlaiv for your detailed explanation, I'm learning all the time and very appreciative.