Sun (21 March 2025)
Full disk image of the Sun in H𝛼, on 21 March 2025. Ha
Close-up of a large sun flare on the surface of the Sun
Introduction
Photographing the Sun has been in the works now for a long time. Back in 2022 I wrote a blog about all the elements that went into choosing a solar imaging setup. In that blog I mention the decision to go for the Solarscope double-stacked front mounted filters. Those filters were indeed ordered, but after two years of promises, delays and no/poor communication, I decided to abandon the plan and look for an alternative. That was found in a very similar Lunt LS60FHa filter. Initially as a single filter, with the option to later double stack if desired to do so. The already purchased Takahashi FOA-60Q was equipped with the necessary accessories. This included a mounting ring for the Ha filter, adapters for the blocking filter and camera, a sol-searcher for polar alignment and a fine-focuser. During a few test sessions, some first attempts towards solar imaging were made. Based on these test sessions, some modifications were made to rig and workflow. While still not optimal, it is now ready for actual imaging. The current report describes the first full-blown imaging session with this new solar imaging rig.
Conditions
Images were collected on 21 March 2025 from the backyard observatory in Groningen, The Netherlands. From this location, the sun only barely reaches above the house and trees in the garden. Early afternoon is usually the best time to image the sun. Images presented here were collected around 14:00h.
Equipment
The rig consisted of a Takahashi FOA-60Q, with a Lunt LS60FHa front-mounted H𝛼 filter and B1200 blocking filter. Full disk images were recorded using a Player One Saturn-M SQR camera directly behind the blocking filter. The close-up images were recorded using a QHY5III568M camera in combination with a Takahashi 2x Ortho extender. The telescope was mounted on a Rainbow Astro RST-135E mount. SER videos were recorded using FireCapture software. Polar alignment during day-time is particularly challenging. In a first attempt, the Polar Scope Align Pro app for iOS was used. This app has a module for daytime alignment. While the alignment was not too far off, apparently the compass and gyroscopes in an iPhone are not precise enough to do a real good alignment, so some drift of the sun remained apparent. As the Sol-searcher was properly aligned with the sun centered in the image, a much better approach for polar alignment appeared to be to do a GoTo to the Sun with the mount and than adjust Altitude and Azimuth knobs to center the Sun into the center of the Sol-searcher.
Telescope
Mount
Camera
Filters
Guiding
Accessoires
Software
Takahashi FOA-60Q, Lunt LS60Ha front mounted H𝛼 filter, Takahashi Ortho Extender 2x
Rainbow Astro RST-135E, EuroEMC S130 pier
Player One Saturn-M SQR , QHY5III568M, both uncooled
Lunt B1200 blocking filter
TeleVue Sol-Searcher, Unguided
Apple MacBook Pro M2 Max
MacOS 15.3.2, FireCapture 2.7.15, AutoStakkert!4, PixInsight 1.9.3, Photoshop 24.0.1, Final Cut Pro 11.0.1
Imaging
Two types of images were made. First a full disk image and second a detailed recording of a very large solar flare. The IMX533 sensor in the Player One Saturn-M SQR gives enough field of view on this telescope in its native focal length to comfortably have a full-disk image, without any optical aides. Focusing was done manually. This can be a bit challenging and probably requires just more experience to feel more comfortable with it. 30s video recordings in SER-format were made using 7ms exposures and gain 100. This camera provides an average of 36 fps, collecting around 1000 frames in those 30s. The Autorun feature in FireCapture was used to collect a 30s recording each minute. A total of 20 of these recordings were made.
Then the system was modified to record some detail of the solar flare. The recently released Takahashi Ortho Extender 2x was introduced in the imaging train. As camera, the QHY5III568M with smaller 2.7 micron pixels and much higher frame rate was used. In similar fashion, 20 recordings were made over 20 minutes. With the higher frame rate, each recording now lasted 15s and collected similarly roughly 1000 frames
Full Disk
Resolution
Focal length
Pixel size
Field of View
Detail
Resolution
Focal length
Pixel size
Field of View
3008 × 3008 px (9 MP)
900 mm @ f/15
3.78 µm
43' x 43'
2472x2064 px (5MP)
1800 mm @ f/30
2.7 µm
25’ x 21’
Takahashi FOA-60Q with Lunt FS60Ha filter, a 60mm f/15 solar rig
Processing
The initial processing for both the full disk and detail images was the same. All SER-videos were analysed in AutoStakkert!. The 200 best frames of roughly 1000 total frames were combined into a single stack and exported as TIF file. The rest of the processing was done in PixInsight. In the Solar Toolbox script proper exposures were set. This included black- and whitepoint, fine-tuning of the rim and boosting the prominences. The Full disk images were inverted at this stage. Contrast enhancement was done using BlurXTerminator. While this tool is not particularly designed for solar images, the type of atmospheric aberrations that BXT is correcting for, is similar to both deep sky objects and the solar surface. With a bit of trial and error, some proper values could be found for PSF diameter and the amount of sharpening, without creating artefacts. Colorising the image was done using CurvesTransformation. This can also be done in the Solar Toolbox, but the colours come out very saturated, while the CurvesTransformation tool gives a little bit more control over colour intensities. I’m also more used to CurvesTransformation, so that probably also plays a part. Some extra sharpening was applied using LocalHistogramEqualization.
For both the full disk and the detail image, 20 SER-videos were recorded. So that produces a total of 40 stacked files to process. This can only work if there is a large portion of automation possible. Autostakkert! can batch-process as many SER-videos as you like. When opening a file, just open the whole batch of files. You will be guided through the workflow on the first video, and once that is done, all the others follow automatically with the same settings. In PixInsight, batch processing can be achieved using Image- and Process Containers. When the settings on Solar Toolbox, CurvesTransformation, BXT and LHE have been established, they can be dragged to the desktop and from there being dragged into a Process Container. All stacked files can be added to an Image Container. From there, it is a matter of dragging the triangle bottom left over the bottom of the Process Container to let the batch process begin.
When all 20 stacks of both the full disk image and the detail image had been processed, the next step was to create a video, showing the dynamics of the solar surface. For that to happen, all images had to be aligned. The alignment process in PixInsight is relying on star field and cannot be used. But there is a Fourier-based image registration script, called FFRegistration, that can be used for this. The full disk images were aligned this way and combined into a video using FinalCut Pro. Unfortunately the FFRegistration script altered the geometry of the prominence images, making it hard to combine them into a video. Instead, the images were loaded into Photoshop as a layer stack and converted into a video using the Timeline feature in Photoshop. Both videos contain 20 individual images and run at 5fps.
The detail images of the prominence show some serious imaging artefacts. First of all there are the so-called Newton rings. They are an interference pattern that can be caused by light reflected between flat glass/filter surfaces. This is a common phenomenon in solar imaging. The solution is to tilt the camera a bit relative to the light-path. This can be done using regular tilt adjustment tools, or special tools designed for solar imaging. This is an improvement that still needs to be made to the rig. The other artefact is a lot of dirt that is present on the QHY5III568M camera sensor. Attempts have been made to clear this up, but that appears to be more difficult than expected. The best way to eliminate this is to shoot flat frames. This is something that needs to be worked out how best to do and added to the workflow.
Processing workflow (click to enlarge)
In conclusion, this first full solar imaging session with the new setup was fairly successful. Some nice images and videos were produced. There is still a lot of room for improvement though. The processing can probably be enhanced quite a bit. More experience can be built with the currently using tools, enhancing the outcome. But there are more tools available out there. For example the software ImPGG seems to do a specifically good job in sharpening solar images. Also regular photo editing tools, such as Affinity Photo and Topaz Photo AI have been used in solar image processing with good success. Also in the imaging itself improvements can be made, such as the elimination of newton rings and the inclusion of flat frames in the overall workflow.
On March 29, 2025 there is a partial solar eclipse, and the next goal is to create a timelapse of that using the current methods described above.
This image has been published on Astrobin