Practical Examples
An extensive range of worked examples; with acquisition and processing notes, warts and all.
Following the technical chapters, this section concentrates on some practical examples that illustrate alternative techniques. These are deliberately chosen to use a selection of different capture and processing programs and to cover a range of imaging problems and solutions. In each case the unique or significant aspects are highlighted rather than a full blow-by-blow account. In particular these examples consider techniques to capture and process objects with a high dynamic range, nebulosity, star fields and in narrowband wavelengths. These images were taken with a variety of cameras and telescopes, all on equatorial mounts. These are presented in chronological order and show a deliberate evolution in technique that will resonate with newcomers to the hobby and more experienced practitioners alike.
This is a deliberate journey, the path of which, to quote Rowan Atkinson as Blackadder, “is strewn with cow pats from the devil’s own satanic herd!” To avoid stepping on some, these case studies do not paint a rosy picture of perfection but present a warts-and-all view that highlight issues, mistakes, improvements and lessons learned. Some of these are experiments in alternative techniques and others highlight gotchas that are less well documented. Collectively they are a fascinating insight into the variety of challenges that face the astrophotographer and provide useful information with which to improve your imaging.
General Capture Setup
Polar Alignment
Until recently, I assembled and disassembled my imaging rig each night. This and the uncertainty of the British weather made image capture challenging. To make the most of these brief opportunities, system reliability and quick setup times were essential. In the early days my polar alignment followed the 80–20 rule: to quickly align within 10 arc minutes and accommodate any drift or periodic error using an off-axis autoguider. On the SkyWatcher NEQ6 mount, a calibrated polar scope and the polar alignment routine in EQMOD delivered the results. In this setup, EQMOD moved the mount to a position where Polaris was at transit (6 o’clock in the eyepiece) and then moved to its current hour angle. (Since an EQ6 can rotate full circle, snags and leg-clashes are a possibility and I stood by the mount during these slews.) After upgrading the mount, using ground spikes (detailed in the chapter Summer Projects) close-tolerance mount fixings and a locked azimuth reference I consistently achieved 1 arc minute alignment. A hernia forced a more permanent pier-mounted system that achieves better than 20 arc seconds after using TPoint modelling software. My most recent portable (sub-10 kg) mount is polar aligned using the QHY PoleMaster camera and achieves the same accuracy.
Hardware Evolution
My first system comprised an 8-inch Meade LX200 GPS with a piggy-back refractor as a guide scope, both fitted with early Meade CCD cameras. Whilst very impressive on the driveway, I soon realized it was not best suited to my imaging needs or physical strength. I now use three refractors of different focal lengths with two field-flattener options, to match the subject to the sensor, in addition to a 250 mm f/8 reflector. With the camera at its focus position, the balance point is marked on the dovetail for each assembly. These are color coded according to the field-flattener in use and enable swift repositioning of the dovetail in the clamp. Cables are either routed through a plastic clip positioned close to this mark to reduce cable induced imbalances or through the mount. The autoguider system employs a Starlight Xpress Lodestar fitted to an off-axis guider, parfocal with the main imaging camera. The entire imaging chain is screw-coupled for rigidity; 2- or 1.25-inch eyepiece couplings are banished. The two larger refractors were fitted with a Feather Touch® focuser and MicroTouch motor to improve rigidity and absolute positioning. The MicroTouch® motors and controller have since been replaced by Lakeside units, so that a single module can be used across all my focus mechanisms. The cameras used in these examples include Starlight Xpress and QSI models fitted with the Kodak KAF8300 sensor and the smaller but less noisy Sony ICX694AL. My mount has changed several times: Early images were taken on the popular SkyWatcher NEQ6 running with EQMOD. This was replaced by a 10Micron GM1000HPS and then a Paramount MX. The load capacity of all these mounts is sufficient for my largest telescope but the high-end mounts have considerably less periodic error and backlash, are intrinsically stronger and have better pointing accuracy. The belt-drive systems in the high end mounts crucially have less DEC backlash too, ideal for autoguiding. A previously-owned Avalon mount is used for travelling.
Each system uses remote control; in the early days over a USB extender over Cat 5 module and later, using a miniature host PC, controlled over WiFi with Microsoft’s Remote Desktop application. All USB connections were optimized for lead length and daisy-chain hubs kept to a minimum. The back yard setup used a dual 13-volt linear regulated DC bench power supply (one for the camera and USB system, the other for the dew heater, focuser and mount) carried through 2.5 mm2 copper speaker cables. The observatory system uses permanent high-quality switched mode power supply units mounted in a waterproof enclosure.
Software Evolution
My preferred software solution has equally evolved: After a brief flirtation with Meade’s own image capture software and after switching to the NEQ6 mount I moved to Nebulosity, PHD and Equinox Pro running in Mac OSX. I quickly realized the importance of accurate computer-controlled focus and moved to Starry Night Pro, Maxim DL 5 and FocusMax in Windows 7, using Maxim for both acquisition and processing. This system sufficed for several years but not without issue; the hardware and software system was not robust and difficult to fully diagnose. When I changed camera systems I skipped Maxim DL 6, that had just been released at that time, and decided to try something different. The software market is rapidly evolving with new competitively-priced offerings, delivering intelligent and simplified automation and advanced image-processing capabilities. In particular, two applications radically changed my enjoyment, system’s performance and improved image quality.
The first was Sequence Generator Pro (SGP). This achieved my goal to let the entire system start up, align, focus and run autonomously and reliably without resorting to an external automation program, many of which, with the exception of MaxPilote, are more expensive than SGP. At the same time, the popular guiding program PHD transformed itself into PHD2, adding further refinements and seamless integration with SGP.
The second was PixInsight (PI). The quality improvements brought about by sophisticated image processing tools, including masking and multi-scale processing, addressed the shortcomings of the simpler global manipulations of the earlier systems and bettered complex Photoshop techniques too. The combination of SGP, PHD2 and PI is ideally pitched for my needs, dependable and good value. These core applications are now augmented with my own observatory automation software, controller and drivers.
Most astronomical equipment is nominally rated at 12 volts but often accepts a range of 11.5–15 volts. (If in doubt, consult the device’s specification sheet.) Lead acid cells vary from about 13.8–11.0 volts over a full discharge, depending on load current and temperature. In practice, 11.5 volts is a working minimum, since discharging a battery below that level reduces its life and is for some mounts a minimum requirement too, to guarantee correct motor operation.
Setting Up
In the case of a portable setup, after the physical assembly, I confirm the polar alignment at dusk with a polar scope or QHY PoleMaster. In the case of the Paramount MX, the tripod’s ground spikes and optimized mounting plate have very little play and in most cases this mount requires no further adjustment. When the MX is permanently mounted, I simply home the mount and load the pointing model that corresponds to the equipment configuration. At dusk, I synchronize the PC clock using a NTP server and if required, set the altitude, temperature, pressure and humidity refraction parameters. I then set the focus position to its last used position for that imaging combination. With either setup, and allowing for the system to acclimatize to the ambient conditions, I open AAG CloudWatcher (to supply the ASCOM safety monitor) and run the imaging sequence in SGP, which automatically slews and centers on the target, fine tunes the focus and waits for the camera to cool down or a start time. SGP fires up PHD2 and starts capturing images once the guider calibration has completed and the tracking has settled. If this is part of an imaging run, I ensure the camera orientation is the same as before and reuse a stored calibration for the autoguider (measured near the celestial equator). The 10Micron mount, MaxPoint and its equivalent, TPoint (TSX), are all capable of building a sophisticated pointing model, capable of sub 20-arc second accuracy, from the synchronization of 50–100 data points. Using SGP’s slew and center automation, it is not mandatory to have that level of pointing precision and I employ autoguiding to fix any residual centering or tracking issues. (In Maxim DL5, a similar level of pointing accuracy is achieved by manually pointing, plate-solving, synching and pointing again.)
Exposure Sequencing
Setting the exposure is a juggling act: Too much and colorful stars become white blobs, too short and vital deep sky nebulosity or galaxy periphery is lost in read noise. Typically with LRGB filters I use an exposure range of 3–5 minutes, extending to 10 minutes if the conditions are favorable. Narrowband exposures require and can cope with considerably longer exposures of 10 minutes or more without saturation. The brightest stars will always clip but that does not have to be the case for the more abundant dimmer ones. If the subject is a galaxy, however, I check the maximum pixel value at the core of the galaxy with a test exposure. I typically use the same exposure for each RGB filter and in the early days cycled through LRGB to slowly build up exposure sets. I now complete the exposures one filter at a time, over several nights, to accumulate enough data for higher quality images. (The one exception being when imaging comets.) If the seeing is poor, I will organize the exposures over one night to expose in the order RGBL (red is the least affected by turbulence at low altitude) and move the focuser at each filter change by a predetermined focus offset. After each exposure the ambient temperature is sampled and SGP’s autofocus routine kicks in if there has been a significant change since the last autofocus run (0.5–1 °C).
fig.1 For the beginner, it is rather disconcerting when, even after a 20-minute exposure, the image is just an empty black space with a few white pinpricks corresponding to the very brightest stars. This is not very helpful or encouraging to say the least! All the image capture programs that I have used have the facility to automatically apply a temporary histogram stretch to the image for display purposes only. This confirms the faint details of the galaxy or nebula and a sense of orientation and framing. The example above shows one of the 20-minute Hα of the Heart Nebula, after a strong image stretch, in which the white point is set to just 5% of full-well capacity. In all the imaging examples that follow, the un-stretched images are almost entirely featureless black. Only the very brightest galaxies may reveal some details at this stage.
I used to dither between exposures (using PHD2) to aid rogue-pixel rejection during processing. I rarely do this now since I discovered PixInsight’s CosmeticCorrection tool, which does not require dithered images to eliminate hot pixels during image integration.
The equipment and exposure settings are set and stored in a SGP sequence and equipment profile. Now that I have designed a permanent automated observatory, at the end of the session or, if the weather turns, the mount parks, the roof closes automatically and the sequence is updated, allowing it to be recalled for future use and quickly continue from where it left off, with a couple of button presses.
fig.2 This “mobile” setup, made neater with full module integration within my Mk2 master interface box, has through-the-mount cabling for USB, focus and power. It is fully operational in 20–25 minutes with PHD2 and Sequence Generator Pro. It is now pier-mounted in a roll-off-roof observatory for almost instant operation, and just as important, swift protection from inclement weather.