A very small, wide-field, guided imaging rig Part 3. Second tests

The equipment in operation

This time a Mars-C II camera was used which has an optical window with very good, multi-layer anti-reflection coatings.

There was no moonlight during this session but there is a streetlight close to the imaging spot that has to be shielded by the occultation board visible on the right.

B142-43 dark nebulae centred on the star HD185898, captured by AstroDMx Capture (pre-release-version3) through a GS-100 PAPO quadruplet ED f/4 (F=100mm, A=25mm) mini scope with a Player One Mars-C II camera (with a SONY IMX662 1/2.8" CMOS colour sensor). 60 minutes of 60s exposures with a Pegasus Astro L-UV-IR cut filter. Captured by AstroDMx Capture pre-release version 3. Processed in PixInsight, GraXpert, SASPro and Gimp3.

B142-43 dark nebulae

Also the North America and Pelican nebulae centred on the star HD199479, captured through the same scope and camera. 60 minutes with L-UV-IR cut, 30 minutes with quadband, 30 minutes with HaO3 and 15 minutes with S2O3 filters. Captured by AstroDMx Capture pre-release version 3. Processed in PixInsight, GraXpert, SASPro, Gimp3 and CS2, blending the various filter results.

North America and Pelican nebula

The stars are from the Pegasus Astro L-UV-IR cut filter because they are true RGB stars and did not produce halos.

During this session we were able to observe that different filters introduce various gradients across an astronomical image? some produce large gradients that have to be removed whilst others produce no gradient on the same object at essentially the same time.

Screenshot of AstroDMx Capture saving 5 minute exposures of the North America and Pelican nebulae through the Pegasus Astro L-UV-IR cut filter,

 

It can be seen that there is a huge gradient across the image.

Screenshot of AstroDMx Capture saving 5 minute exposures of the North America and Pelican nebulae through the Altair quadband filter

It can be seen that the gradient has gone but some of the brighter stars now have halos. A similar result was obtained when using an Askar  HaO3 filter, and also with an Askar S2O3 filter.

Screenshot of AstroDMx Capture saving 5 minute exposures of the North America and Pelican nebulae through an Asker S2O3 filter

Animation showing the gradient produced by the Pegasus Astro L-UV-IR cut filter vs the absence of the gradient produced by the Altair quadband filter but the presence of star halos

Different astronomical filters introduce vastly different gradients because of the type of light they let pass and how they interact with wide angles of incoming light. When shooting the exact same target at the exact same time, a broadband filter like the Pegasus L-UV-IR cut filter can show a massive sky gradient, while a narrowband filter like the Altairquadband filter can appear completely gradient-free. This is driven by two main factors: the nature of environmental sky glow and the physics of interference filters. 

The primary driver of gradients is sky glow, which includes artificial light pollution, moonlight, and natural atmospheric airglow. During these tests there was no moonlight.

Broadband filters pass wide windows of light (often 100nm or more). Because light pollution (especially from LEDs) and moonlight scatter across the entire visible spectrum, they create a thick background glow. Because the telescope points at an angle through the atmosphere, one side of the image frame is always closer to the horizon or a light source than the other. This creates a steep, highly noticeable brightness gradient across the sensor. 

Narrowband filters isolate an incredibly tiny sliver of light (usually between 3nm and 7nm) centred precisely on the emission lines of nebulae. They block up to 99% of all scattered sky glow.

H (656nm) and SII (672nm): Moonlight and urban light pollution contain very little energy at these deep-red wavelengths. Thus, the background stays completely dark and gradient-free across the whole sensor.

OIII (501nm): Unlike H, Oxygen-III sits in the blue-green spectrum where moonlight and LED scattering are very strong. Even with a narrow 3nm filter, if the Moon is out, OIII will often suffer from severe gradients while an H-alpha frame taken at the exact same moment remains completely clean. As we had no moonlight and the fluorescent streetlight nearby, the S2O3 filter didn't produce a gradient.

Some gradients are not in the sky at all; they are generated inside the imaging equipment itself. Modern astrophotography filters are interference filters, meaning they use microscopic chemical layers to reflect unwanted light and pass specific wavelengths.

When light rays hit an interference filter at an angle (which happens frequently at the edges of a wide-field telescope or fast f-number optics), the physical distance the light travels through the filter layers increases. This shifts the filter's passband toward shorter wavelengths. With broadband filters, this shift doesn't matter because the window is wide enough. But with ultra-narrowband or complex multi-band filters, a bandpass shift means the edges of the sensor stops capturing the target's light and starts capturing background light noise, creating colour gradients.

Some filters have poor off-band blocking performance. When a filter fails to block out-of-band light completely, that light enters the filter glass and creates intra-filter reflections. The light bounces back and forth between the front and back surfaces of the filter glass before finally escaping toward the camera sensor. This reflected, slightly offset pool of light is what produces a halo around bright stars. The halos we observed are most likely due to this cause.

In conclusion the GS-100 mini scope can be used successfully as a wide-field imaging scope in conjunction with the appropriate camera and filters.

We have used these tests to simultaneously test pre-release version 3 of AstroDMx Capture.

A very small, wide-field, guided imaging rig Part 2. First tests

 

The equipment

The scope and guide-scope mounted on the AVX mount

Although a number of components can be added to the rig, such as a filter drawer or filter wheel there are very limited degrees of freedom to develop a setup that is truly flexible. For example, there is such limited back-focus with probably any camera, that it is impossible to fit even a very thin camera rotator to facilitate the framing of astronomical targets. There is room to fit a very small number of spacer shims so by varying the number and thickness of these and tightening up the camera, it may be possible to approximate the desired framing, or not!

It should be possible to build in a camera rotation system into the scope’s design without reducing the back focus. This is of no consequence if the scope is used as a guide scope, but is of huge importance for an imaging device for which it is heavily marketed and hopefully Touptek will consider this.

Considering the components that we used for the first test, a number of important lessons can be learned.

The GS-100 is a fast scope that may cause problems with larger sensors due to the angle of incidence of some of the rays hitting the sensor. The faster the scope, the steeper these can be. This can increase crosstalk between pixels resulting in poorer colour fidelity and contrast.

Probably the biggest problem with using the ZWO ASI585MC camera concerns reflections back and forth between the sensor and the filter. The camera used does not have anti-reflection coatings on the sensor window glass which means that the reflected light is not inhibited and passes back and forth between the sensor and the filter wheel, passing through the quadruplet lenses on its journey. The shifting of the originally parallel light rays leads to the development of Newton’s rings interference patterns which are particularly noticeable towards the ends of the long axis of a rectangular sensor. There are limited options for changing the distance between the filter and the sensor, which would probably have little effect anyway. There is no tilt plate built into this camera so no way of using tilt to mitigate the problem. The filter itself has inadequate anti-reflection coatings which could have reduced the impact of the interference patterns. After consideration, the quadband filter may not have been a good choice. I am not sure whether ours is V1 or V2. If it is V1 then this filter was renouned for poor anti reflection coatings and producing bad halos around stars. We clearly had reflection problems and some stars did have bad halos which were unsatisfactorily cosmetically treated during processing. A high quality UV/IR filter may have been a better choice.

Unlinked channels stretched image showing the concentric interference bands

The interference bands are more evident at the left hand side of the image where the star density and nebulosity are lower.

Careful post processing can reduce these artefacts but not completely eliminate them

North America and Pelican nebulae test

Sadr region test

The images were improved by cropping out the central area where the interference was less evident. However, this partly defeats the object of using the ZWO ASI585MC camera.

Cropped and reoriented images

North America and Pelican nebulae

Sadr region

For a future test our Player One Mars-C II IMX662 Colour Astronomy Camera will be used. Learning from our previous experience, this camera seems to have a number of advantages over the ZWO ASI585MC. It has Diameter 21 * 1.1 mm High Quality AR Plus (Anti Reflection) Multi-Layer Coating. Instead of a single layer of anti-reflective material, this optical element has multiple microscopic layers cooked onto the surface. Multi-layer coatings are superior because they cancel out reflections across a much wider spectrum of light wavelengths (colours), rather than just one. It also has a built in front tilting plate that allows for adjustments to avoid interference patterns if required.

Further camera details:

SONY IMX662 1/2.8" CMOS sensor (color). 2.1 Mega Pixels.   Maximum Resolution  1936×1100

2.9μm square pixels.   Sensor Size  5.6mm×3.2mm

If we centre on the star HD198597 the field of view is such that much of the Cygnus Loop can be framed.

Stellarium simulation of the field of view framed by the GS-100 with the Player One Mars-C II camera. The framed area is within the red frame lines.

The Player One Mars-C II camera will be used in conjunction with a Pegasus Astro L/UV-IR cut filter which offers a transmission range of >95% between 410–680nm. This blocks the ultraviolet and infrared bloat while letting the crucial visible spectrum (including Hydrogen-Alpha at 656nm) pass cleanly through to the sensor. It intentionally leaves a gap between the green and red bands. This blocks the primary Sodium light pollution line at 589nm, resulting in cleaner colour separation. 

The glass  features anti-reflection and anti-halo coatings which is crucial for UV/IR filters. It has a 1/4 wavefront polish.

The Player One Mars-C II camera on the test rig

It is hoped that this setup will allow capture without interference banding and star halos.

Again, the testing will be done with a pre-release version 3 of AstroDMX Capture, so we will be testing not only the hardware but also the software.

The Soap Bubble nebula (PN G75.5+1.7)

The Soap Bubble nebula (PN G75.5+1.7) from 1 hour of narrowband data (Ha and O3). 5 minute exposures (30 minutes through each filter) captured by AstroDMx Capture through a William Optics 81mm APO refractor with a QHY Minicam 8 camera/filter wheel. Processed in PI, SASPro and GIMP3. Rendered as HOO. 'RGB' stars were constructed in Siril from narrowband data using the Standard Continuum Mapping pixelmath procedure. 

If we include the 30 minute worth S2 data in a SHO/NBN palette (giving 1.5 hours total exposure), the Soap Bubble is much less pronounced. This is because most of the information on the Soap Bubble is in the Ha and O3 channels. A blend of HOO and SHO gives a reasonable compromise, with the Soap Bubble being visible along with some of the hues associated with the SHO.

Click on any image to get a closer view

HOO NBN

HOO NBN SHO blend

SHO NBN

This session was revisiting an object that we imaged in September 6, 2023 in narrowband with the same scope but using a 14 bit SVBONY SV605MC using a total of 5 hours of accumulated exposure.

The QHY Minicam 8 camera/filter wheel camera used here is a 12 bit camera and a considerably shorter total exposure time was used. Nevertheless, we were able to make a good capture. The best image was the HOO image with just 1 hour of total exposure time.

Steve Wainwright and Nicola Mackin

A very small, wide-field, guided imaging rig Part 1. The equipment

The imaging scope is the latest edition to the Touptek family of guide scopes; the Touptek GS PAPO Guide scope. It is a 4 element self-flattening Planar Apochromatic ED refractor. It is truly tiny: F = 100mm, D = 25mm, f/4. it has a 1“ imaging circle and gives good stars to the corners of a sensor such as the Sony IMX 533.

There is very little back focus so it is not possible to use a camera rotator to facilitate the framing of subjects. One slight concession to this is that I have fitted two 1mm spacer shims, as each one is added the camera can be rotated to a different position to tighten it up. It remains to see if they will have to be discarded when we attempt to focus on a star using a Bahtinov mask.

We have mounted a 2“ Altair filter drawer on the front of the scope and have added a 30mm M48 extension tube after that to act as a dew/ light shield.

This is an unconventional setup and it will be interesting to see how it performs.

View of the equipment

A view from the front

The guide-scope to be used is the  SVBONY SV165 Guide-scope with F = 120mm paired with a QHY-5II-M guide camera. This harks back to the days when manual guiding used a guide-scope with a longer focal length than the imaging scope, creating forgiving guiding for the image being captured with the imaging scope. A small movement in the guide-scope is an even smaller movement in the imaging scope.

The camera we are using initially is the uncooled ZWO ASI585MC. The ZWO ASI585MC uses an IMX585 sensor which is BSI, and built on Sony's STARVIS 2 technology. It has a full well depth of 40k to 47k electrons and a Read noise of 0.7 to 5.5 electrons depending on gain. It has a high QE of 91% at 500nm and 80.9% at 656nm.

Simulating the field of view of the GS-100 and the IMX585 in Stellarium, if one centres the image on the star HD198626, the whole of the Cygnus Loop can be framed:

Framing of the Cygnus Loop

Even though it is called the GS-100 Guide scope, it is heavily marketed by Touptk as a mini, very wide field imaging scope because of its high quality optics. They show that is is a good match for the new innovative cooled/uncooled camera the Touptek Astro AE676C as well as other cooled or uncooled cameras. They also show that filter drawers, electronic filter wheels and filters can be attached to the front of the scope and that it is directly compatible with the ToupTek Astro AAF electronic focuser.

Whilst as I said earlier, we are using an unconventional configuration, we will be, by no means using the scope out of spec.

Cable clips placed at strategic positions to facilitate cable management

A Bahtinov mask was contructed by using epoxy resin to attach a 3D printed Bahtinov mask intended for a Seestar S50 onto a 10mm M48 extension tube so that is can be screwed onto the 30mm M48 extension tube dewshield.

The attached Bahtinov mask

This will be needed to get the stars well focused.

It can be stored conveniently in a small plastic box of the type we use to store 2“ filters.

An opportunity arose to test the focus of the rig with a Pegasus Astro UV/IR cut filter in the filter drawer. A pre-release version 3 of AstroDMx Capture was used. The field of view of the GS-100 was plate-solved and the scope was sent to the star Vega. The Bahtinov mask was then used to bring the star to focus.

Screenshot of the focused star on the preview screen of AstroDMx Capture

Clouds prevented further testing but this test was a success!

In Part 2 this wide-field imaging rig will be tested with the latest pre-release of AstroDMx Capture version 3 

Getting HOO narrowband images from a Dwarf Mini

The Dwarf mini has a dual-band filter that effectively separates the O3 into the green and blue channels and the H-alpha into the red channel of the colour image it produces. 

We are going to take the 16 bit PNG produced by Stellar Studio and re-process it into a real HOO image. The image we shall use is a mosaic image of NGC7000, the North America Nebula.

We shall convert:

RGB

and convert it into:

HOO

Let us look at one method of doing this:

Click on any image to see it much closer

We shall use GIMP in the main.

There are various ways of doing this. The first thing to do is to save the 16 bit PNG as a 16 bit TIFF. The reason for doing this is that some software only treats PNG files as 8 bit files and converts them after reading them. We want to retain the 16 bits until the very end of the processing.

We shall start with a starless version of the original image plus the stars image. These could have been obtained from the original image in a number of programs so for that part we shall not be prescriptive.

Starless RGB image

We have to split the RGB image into it's three monochrome channels; R,G and B.

Colours

    Components

        Decompose     (NOT to Layers; Colour model RGB)

Three monochrome images will appear:

Red channel: H-alpha, H

Green channel: O3

Blue channel: O3

We next have to combine the Green and Blue channels as they contain the O3 data.

Select the Green channel (the middle one at the top).

Edit

    Copy Visible

Select the Blue Channel (the right hand channe at the top

Edit

    Paste as

        Single layer             (depending on the version of the Gimp you are using, it might say New layer)

At the right hand side, select Mode Addition.

Then:

Image

    Flatten image

This has created the O3 monochrome image.

Then delete the Green channel as it is no longer needed.

You might then need to use levels to bring up the brightness of the new O3 image to a level similar to that of the H image.

The adjustments can be seen in the Levels dialogue in the screenshot below.

Note that is has not just been made brighter, but the dark areas have been preserved by bringing in the left hand slider a little. 

This process of making the H and O3 images of similar brightness distributions is called normalisation.

(This and the next step can be repeated until you have the required amount of blue in the HOO image)

The H and O3 monochrome images now have to be combined into an HOO false colour image.

The H is placed in the Red channel and the O3 is placed in both the Green and Blue channels of the colour image that will be created.

Colours

    Components

        Compose

A dialogue like the one in the screenshot below will appear.

Take your time and make sure that you assign red to red

blue to green and blue to blue

When you click on OK the HOO image will be created.

Save the image with a sensible name like: NANEB_HOO.tif

At this stage it is still 16 bit.

Then load The HOO image into GraXpert

DON'T extract the background because this will already have been done by Stellar Studio right at the start.

Denoise the image.

When Denoising is complete:

Select Saving

Set to 16 bit Tiff

Click on Save Processed:

GraXpert will give a new filename by adding Graxpert to the end of it

The name we can see that has been generated is NANEB_HOO_GraXpert.tiff

Click on Save



You can then load this image into GIMP

If you need to, you can select:

Colours

    Hue-Saturation

        and increase the Saturation a LITTLE as shown in the dialogue in the screenshot below

Then load the Stars image into GIMP

Select the Stars image

Edit

    Copy Visible

Select the HOO image

Edit

    Paste as

        Single layer                                  (or New layer in a different version of GIMP)

On the right hand side

Select Mode Screen

This will screen the stars back into the image.

Export the image as NANEB_HOO_starsback.tif

If you then Export the image as NANEB_HOO_starsback.jpg

It's bit depth will be reduced to 8 bits as a JPG image suitable for sharing.

The original image and the HOO image animation

Test of a pre-release of Version 3 of AstroDMx Capture

An Askar 71F quadruplet apochromatic astrograph refractor was used, paired with an SVBONY SC571CC 16 bit OSC cooled camera. The scope was fitted with an iOptron iEAF motor focuser and an Altair V2 magnetic 2” filter holder fitted with an Askar Colour Magic dualband HaO3 filter. AstroDMx Capture pre-release Version 3 for Linux x86_64 running on an Ubuntu Linux mini computer was used to capture the image data as FITS images with  matching darks, flats and dark-flats. 3 hour's worth of 5 minute subs of the SH2-54 region of nebulosity were captured. The data were stacked and part processed in PixInsight and further processed in GraXpert, SetiAstroSuitePro and Gimp3 and presented as RGB and HOO and a blend of the two.

Screenshot of version 3 of AstroDMx Capture capturing RAW FITS data on the SH2-54 region of nebulosity

Negative preview

SH2-54 region of nebulosity RGB

HOO

RGB-HOO blend

Each test of pre-release version 3 of AstroDMx Capture searches for bugs and anomalies and brings the release of version 3, code-named Serenitatis, ever closer.