Remote Operation

SLR and mirror-less cameras are both capable for use in astrophotography. One quickly realizes that red sensitivity is another important attribute and many photographic cameras block deep red wavelengths with their infrared blocking filter. I discovered my Fuji X-Pro 1 camera could detect the IR emitter on a TV remote control and figured it might be excellent for use on a telescope. The only problem was I had to use a cable release and hang around with a stopwatch. Some of the later models in the range have an electronic remote release, which can also accept an electronic interval timer or connected to their own dedicated image tethering software. Even so, it has to be said that this is not the easiest solution for two reasons: seamless remote operation and focusing. Focusing is an unavoidable requirement in any setup. Even with 10x image magnification on its LCD screen, it is very difficult to judge the precise focus on any star, even with a focusing aid such as a Bahtinov Mask. With the right software, however, and tethered operation over a USB connection, things become much tidier; imaging software takes repeated images whilst you adjust the focus and plots the star size, using Full Width Half Max or Half Flux Diameter (FWHM or HFD) for each exposure. The best focus is achieved when the width is at its lowest value. Although several camera brands supply general software utilities for remote capture, it is the third-party apps and full astrophotography programs that have the tools to assess focus accuracy and these work best when they are fully integrated.

At present, the best-supported SLR brand by astronomy programs is Canon EOS although some of the high-end astrophotography programs like Maxim DL and others increasingly support others. Nikon has released an astrophotography specific version of the D810 to compete. Modern cameras operate via their USB connections but not all support long exposures (greater than 30 seconds) directly. Those models and the older pre-USB models can be triggered with a modified electric cable release and an interface circuit to a PC. One popular source of adaptors is Shoestring Astronomy. Long exposures consume batteries and a full-nights imaging requires several changes. The process of changing batteries can disturb the imaging alignment and for uninterrupted operation, a battery adaptor powered from 12 volts is a better option. (The battery adaptors supplied by the OEMs or third-party retailers comprise a battery adaptor and a small DC power supply, powered by mains.) With a little ingenuity a small variable DC-DC converter module can be housed in a small plastic box and adjusted to provide the exact same DC voltage at the camera end, from a DC voltage. This is a safer and reliable alternative and it can share the mount’s power source.

fig.3 A color CMOS sensor is made up of a grid of light sensitive diodes, overlaid with a color Bayer filter array and a micro-lens matrix on top, to improve optical efficiency.

Image Quality

Camera shake from a SLR mirror and shutter is not an issue during long exposures, as any vibration fades in milliseconds. There are a few settings on a camera, however, that you need to take care of to extract the best quality. All digital cameras have a long-exposure noise-reduction option, which takes and subtracts a single dark frame exposure from the image. This is not sufficient for high-quality astrophotography and should be disabled. It is also really important to use unmolested RAW files in the highest possible bit depth and without any interference from any in-camera processing. For best results, images are stored in the camera’s native RAW file format and to reduce thermal noise, any toasty “Live View” option is turned off whenever possible. RAW files are not always what they seem; keen-eyed amateurs noted that some older Nikon cameras process RAW files and mistakenly treated faint stars as hot-pixels and remove them. Additionally RAW files are not quite as unadulterated as we are led to believe. All the cameras I have tried have to some extent manipulated the RAW data in camera. This undisclosed manipulation is often detected in weird dark current results and requires special attention during image calibration prior to stacking.

fig.4 A monochrome CCD sensor is made of chains of sensor elements, in this case, overlaid with a micro-lens matrix. The CCD is a more linear device for measuring photons than a CMOS sensor and is ideal for scientific study. In both the CCD and CMOS sensors, the photo-active regions do not cover 100% of the sensor surface, as they have to make room for the interface electronics. The design of the micro-lens ensures that most of the incident light is focused onto active regions of silicon, improving detection efficiency.

Image processing, especially with deep sky images, severely distorts the tonal range to show faint detail and requires the highest tonal resolution it can muster. Most digital cameras have 12- or 14-bit resolution in their sensor electronics. These produce 2¹² (4,096) to 2¹⁴ (16,384) light levels for red, green and blue light, stored in a 16-bit file format. Image processing on the combined image files averages between exposures and creates a higher tonal resolution that ideally uses 32-bits per channel. JPEG files on the other hand have just 8-bits per channel resolution and this is insufficient to withstand extreme image processing without displaying abrupt tone changes (posterization). Choosing a RAW file format also bypasses any in-camera high-ISO noise reduction modes that generally mess things up. In-camera noise reduction typically blurs the image to reduce the apparent noise but in doing so, destroys fine detail. For a static subject there are far more effective means to reduce shadow noise and astrophotographers use multiple exposures combined with statistical techniques to reduce image noise in dim images. This important subject has a full explanation in later chapters.

It is easy to imagine the benefits of remote operation: In addition to the welcome convenience during a long exposure sequence there are sometimes less obvious benefits to imaging quality too. A third, lesser-known benefit of remote operation occurs during the conversion from the separate RAW sensor element values to a RGB color image. Camera RAW files require processing (DeBayering) to create a color image from the individually filtered sensor values. The standard RAW converters used by photographers and the standard RAW converters apply some averaging (interpolation) between adjacent sensor elements for pictorial smoothness. (Some cameras additionally have an anti-alias filter, a sort of mild diffuser, in front of the sensor that accomplishes the same thing.) In general photography, an anti-alias filter trades resolution for smoothness and is most needed to depict straight lines without jagged edges or strange colored banding. While this is important for normal pictures it is all but irrelevant to astrophotography as there are no straight lines. Several astrophotography programs use their own optimized algorithms (especially for star-fields) that preserve and later convert the individual RGB information into a 16- or 32-bit RGB FITS or TIFF image file. If you can, choose remote tethered image capture, using an astrophotography program, rather than store on the memory card. Photography and astrophotography have different visual needs and the specialist capture programs are optimized for the purpose.

fig.5 This Canon EOS 60Da is ready for action. It is fitted with a T-ring adaptor and T-thread to 2-inch nosepiece with a Hutech® IDAS light-pollution filter screwed into the front. In practice the EOS is set in manual exposure mode to allow remote operation from its USB port. This combination works well with a short, fast focal length scope. If camera optics are used, it is best to use them at full aperture to avoid strange diffraction patterns from the polygon shaped aperture (see below). Some other makes of camera also require a “shoot without lens” option enabled, to work with a T-adaptor.

Color Sensitivity

The individual sensor elements (some texts refer to these as photosites) in a CMOS or CCD have a broad color sensitivity that extends from ultraviolet (UV), through visible and includes infrared (IR) light. Video camcorder “night-shot” modes make good use of the extended infrared sensitivity but this is not a desirable feature in either general photography and astronomy. Infrared light will focus to a different point through any refractive optic and even color-corrected compound elements will refract visible and infrared light to a different extent. The outcome is a blurred image. (Mirrors reflect all light by the same amount but any glass-based correction optic may introduce the problem just before the sensor.) The answer is to block IR (and UV) light from the sensor. The filters used in a color sensor’s Bayer array are not sufficient and an additional IR blocking filter is required to stop infrared light reaching the sensor.

fig.6 This cropped image was taken with a Canon EOS, fitted with a 300 mm f /4 L lens, accidentally set to f/5.6. After image processing, the faint diffraction patterns from the non-circular aperture become obvious and detract from the image. I should have used full aperture and sharpened the softer image.

In general, the efficiency of a photographic camera is less than that of a filtered monochrome CCD and in some models, the added IR filter further reduces the light intensity of deep red light. (The primary color of emission nebula is a deep red from ionized hydrogen (Hα) at 656 nm and an even deeper red from ionized sulfur (SII) at 672 nm.) In these instances, a longer exposure is required to detect the faint glow, with the added risk of higher image noise and over-exposed stars in the region.

The standard Canon EOS camera bodies currently have the best third-party software support but at the same time their IR filters reduce the intensity of these important wavelengths by 80%. Nikon, Canon and Fuji all market specially filtered bodies ideal for astrophotography (the Nikon D810A, EOS 60Da and Fuji X-T1 IR). These cameras are not cheap and many use a consumer model or a used body and the services of a third-party company to remove or replace the infrared blocking filter. This modification improves the deep red sensitivity and depending on whether the IR filter is simply removed or replaced, may affect color balance and autofocus.

Light Pollution

A camera’s color sensor is also sensitive to visible light-pollution. Although light-pollution filters block the principal street-lamp colors and reduce the background light intensity, they also reduce star intensity at the same time and require color correction to remove their characteristic blue or green color cast. These filters mount in the camera lens throat (Canon EOS) or on the end of a T-thread to telescope adaptor (1.25- or 2-inch). Light pollution filters can perhaps be more accurately described as nebula filters, since they pass the specific nebulae emission wavelengths, which thankfully are not the same as those found in urban light pollution. The result is that image contrast is improved and requires less manipulation to tease out faint details. Whilst on the subject of light pollution, the intense reflected light from a full Moon can wreck a night’s imaging. No light pollution filter can remove this broad-band sky illumination but narrow-band filters are still effective for imaging emission nebulae.

Thermal Noise

Thermal noise adds random signal to each sensor photosite, that accumulate over time and at a rate that increases with temperature. This noise consists of a general background rate, called the dark current, and a random noise element too, which increases with the background rate. The dark current typically doubles for every 6–8°C rise and luckily, subtracting a matched dark frame from each image virtually eliminates its effect. With each doubling of the dark current, however, the random noise increases by 1.4x. A simple dark frame subtraction will not reduce random noise (it actually increases it) and the only way to reduce its effect is to combine many images. It requires twice as many averaged exposures to return the signal to noise ratio to its former (cooler) value for a 6–8°C rise.

Manufacturers have not yet resorted to cooling their sensors and unfortunately they gradually warm up with extended use, especially if the LCD screen is on. In addition to general thermal noise, some cameras (especially the older models) exhibit “amplifier glow”, which increases with exposure time and temperature and shows up as a general fogging of the image around a corner or edge. The most recent consumer EOS bodies have improved sensor and micro-lens efficiency and those with the lowest pixel density currently have the best dark noise performance. (At a point in time, the consumer EOS 1100D /Rebel T3 model has a 14-bit RAW file output, remote control capability and exhibits lower image noise than its higher pixel-count APS-C cousins.)

For the best performance with a conventional camera, choose a modern camera with a low pixel count, take many averaged exposures with via tethered control and download the images, avoiding using the rear LCD panel for extended periods of time and hot conditions. Easy!

Choosing a CCD camera

CCD cameras are available in a number of configurations; with integrated cooling, in-built guider, off-axis guider and integrated filter wheel to name the most common. I have made a few changes over the years and can relate to the dilemma of product selection. The final choice is a balance of performance and size. The largest sensors are not necessarily the best in terms of sensitivity and noise. I initially used a planetarium program to show the field of view with my various combinations of telescope and field flatteners, with respect to the objects I wish to image. My original KAF8300 sensor was too big for the smaller objects, even with my longest scope, but too small for the largest nebulas and galaxies with my shorter scope. I chose a smaller, more sensitive Sony monochrome sensor to suit the smaller deep sky objects and that works, without cropping, using a 1.25” diameter filters. I used L, RGB and Ha filters in a filter wheel with an off-axis guider tube. Later on, I realized the field of view was too small for wide field objects and I replaced it with another KAF8300-based model, with an integrated 8-filter wheel, an off-axis guider and bought a long and short focal length refractor to cover the range of magnifications.

As you can tell, there is no perfect camera for all purposes; the sensor choice alone is a compromise between size, pitch, well depth, read noise, thermal noise, sensitivity, efficiency and gain. The specifications for a range of popular sensors are shown in fig.7 with their computed dynamic range and effective bit depth. There can be small differences in the final camera specification between camera manufacturers that use the same chip and sometimes they use different quality grades too.

fig.13 The size of the sensor and the focal length choice may be too big or small for the deep sky object. This graph shows how many Messier objects are smaller than a certain angular width, in comparison to four common sensors for a focal length range of 500–1,000 mm. Only 2 Messier objects are wider than 100 arc minutes.

Sensor Spacing

It is worth mentioning a little on the precise optical path length at this point. The filter wheel inserts between the field-flattener or coma corrector and the camera. The optical design of a field-flattener has an optimum distance from the coupling to the sensor. A common spacing is 55–57 mm, similar to the T-mount standard. The combined distance from the sensor plane to the face of the T-thread should be within ±1 mm for best results. (For the same reason, T-thread adaptors for different photographic cameras are different lengths, so that the overall T-thread to sensor distance is 55 mm.) Dedicated CCDs often have matching filter wheels or integrate with them to accomplish the same coupling to sensor distance. Hopefully the distance is spot on or a little less than required, in which case T-thread extension tubes and plastic Delrin® shims in 1 and 0.5 mm thicknesses are combined to fine-tune the separation. The physical separation is shorter than the optical distance, since the sensor cover glass and filter can be quite thick, normally around 2–3 mm and generally increases the optical path by 1–1.5 mm. (The optical path of a medium is its thickness multiplied by its refractive index; in the case of glass, about 1.5x)

There is another less obvious advantage to exposing through individual filters over a one-shot color CCD or conventional photographic camera that improves image quality through refractor telescopes. When imaging through individual or narrowband filters, it is possible to fine-tune the focus for each wavelength. As it is not a requirement to have all light colors focusing at the same point at the same time, it is possible to use a cheaper achromatic rather than apochromatic refractor. (There are limits – if the focal length is significantly different for different wavelengths, the image scale will be slightly different too and will require careful registration to form a color image.) Individual focus positions become a practical proposition with a motorized focuser, since you can work out the optimum focus position for each color and store the offsets in the filter wheel driver software (or alternatively autofocus after a filter change). When the imaging software finishes an exposure, it commands the filter wheel to change, reads the focus offset for the new filter and passes it to the focuser driver before starting the next exposure. I found that even with an APO refractor, the combination of the apochromatic triplet and a two-element field-flattener adds a tiny amount of focus shift that can produce slightly soft stars in the outer field. It also helps to choose all your filters from the same range; not everyone owns a motorized focuser, and an LRGB and narrowband filter set is approximately parfocal (focus at the same point) and are designed to work together to achieve good color balance.

Guide Cameras

Thankfully guiding does not require a large expensive sensor and it actually helps if it is small and sensitive with a fast download speed. Some guide sensors are integrated in up-market imaging CCD modules or otherwise bought as a separate camera. Ideally, this is a small separate monochrome CCD still camera, often 1.25 inches in diameter or a webcam (that is able to take long exposures). It can also give an extra lease of life to one of the early CCD imagers, like the original Meade DSI range. Others choose a sensitive CCD marketed specifically for guiding and which include a ST4 guide output (effectively four opto-isolated switch outputs for N, S, E & W control). The Starlight Xpress Lodestar® is a popular choice and equivalent models are offered by QHY, SBIG, ATIK and Skywatcher to name a few. The Lodestar slides neatly into a 1.25-inch eyepiece holder or can screw into a C-mount adaptor for a more secure and repeatable assembly. It is possible to guide with a webcam too, though a color CCD is less sensitive and will require a brighter star for guiding. Since guider corrections are not predictive but reactive, the delay or latency of the correction should be as small as possible but not so rapid to try and guide out seeing conditions. To speed up image downloads through the USB, the better CCD models have the ability to output a sub-frame centered on the guide star.

Guider Optics

The best position for a guide camera is to take a sneak peek through the imaging telescope, focuser and reducer, adjacent to the main imaging sensor. In this position the guide camera has precisely the same alignment as the imaging camera and will automatically correct for any flex in the mount, telescope and focus tube assembly. One way to achieve this in practice is to reflect a part of the peripheral image to a sensor before the light passes through the filter system, using an off-axis guider. You can see the small mirror in fig.20 that splits some of the image off to an externally mounted guide camera. The two adjustments on the periscope fix the mirror intrusion and the focus position of the camera. Since the off-axis guider moves with the imaging camera it tracks any focus changes. Using the screw coupling in fig.21, this is a one-time calibration.

fig.24 This piggy-back partnership of a refractor and a Meade LX200 SCT can be configured so that either will act at as the guide camera for the other. The refractor is fixed to a Losmandy plate that is screwed firmly to the SCT. The counterweights can be slid along a dovetail bar and extended, balancing both axes.

It is not always possible to use an off-axis guider, either because you do not own one or do not have the space to insert it in-between the field-flattener and the sensor. This is the case with a T-coupled digital SLR. In this situation a separate telescope or guide scope, mounted alongside or directly to the imaging scope is required. This can be effective but any difference in the flex between the imaging systems will show up as a guiding error. (This is a particular issue with those SCT designs with moveable mirrors, as the guiding system cannot correct for any mirror shift that occurs during an exposure.) One of the interesting things is the guide scope does not have to be the same focal length as the imaging scope but can be considerably shorter and still deliver accurate guiding information. The ability to detect a guiding error is determined by the pixel pitch, focal length of the guider optics and the ability to detect the center of a star.

fig.25 This screen grab from PHD shows the DEC and RA tracking error. It can show the error or the correction values along the time axis.

Guider software calculates the exact center of a star not only from the bright pixel(s) in the middle but the dimmer ones at the edge too. It can calculate the center extremely accurately, to about 1/10th of a pixel or better. When we take into consideration that astronomical seeing often limits the image resolution, say 1–3 arc seconds, the accuracy of the guiding should only need to be practically compatible at a system level. The practical upshot is that you use a guide scope with a focal length that is a fraction of the imaging scope. Since longer guider exposures even out seeing, the aim is to have a guider system with an angular resolution about 2x finer than the seeing conditions (the overall performance is a combination of the angular resolution imposed by tracking, optics, seeing and imaging pixel size). As a guide, the minimum focal length in mm can be determined by the following formula:

In the case of the Starlight Xpress Lodestar, with a pixel pitch of 8.4 µm, seeing of 2 arc seconds and a guider resolution of 1/10th pixel, the minimum guider focal length would be 173 mm. In practice, the root mean squared (RMS) error for the finder scope system shown in fig.23 was about 2.5 arc seconds peak-to-peak on a summer’s night and influenced by the seeing conditions during the short, 1-second exposures.

fig.26 This screen grab from Maxim DL shows the image capture and guider images, together with the tracking graph, which displays the positional error every second or so. The graph is a useful confirmation that everything is going smoothly. There are several options for showing errors (in pixels or arc seconds) and the mount corrections.

Guider Control

Guider control will need a computer somewhere along the line. There are some stand-alone systems, the granddaddy of which is the SBIG ST-4, that gives its name to the popular mount control interface. More recently the SBIG SG-4 and the Celestron NexGuide guiding systems automatically calibrate, lock onto the brightest star and issue guiding pulses to an ST4 guider port on a telescope mount. These systems comprise a guide camera with an integrated micro controller and a simple interface. These are ideal for setups without a PC and using a photographic camera. The alternative is to use a guide camera and autoguider software on a PC (in this case meaning any personal computer; Windows and Apple OSX operating systems are both capable for image capture and guiding). Modern autoguider software, after performing an initial calibration that measures the orientation and scale of the guider image, takes an exposure and identifies a bright star. It then takes repeated exposures, calculates the positional error and issues the necessary mount adjustment after each exposure. Some of the better systems rapidly download a portion of the image and also adjust for backlash on the declination axis, either automatically or as a manual adjustment.

Autoguider software may be a stand-alone program or incorporated into the imaging software, of which Maxim DL® and TheSkyX® are the best known. PC control offers more options than stand-alone systems, including full control over guide aggressiveness; dither, anti-backlash settings as well as the physical means of moving the telescope (pulse guiding and ST4). This last choice is not without some controversy: An ST4 guider port accepts direct simple N, S, E & W control signals into the mount, whereas pulse guiding takes the required correction and accounts for any periodic error correction and issues corrections through the handset or PC serial link. To my engineering mind, pulse guiding is more intelligent, as potentially contrary commands can be combined in software rather than fight each other at the motor. (For instance, the RA motor is always moving at the tracking rate and guiding should never have to be that severe that the RA motor stops moving or reverses direction, it merely has to speed up or slow down a little.) Having said that, I have successfully used both on a SkyWatcher EQ6 mount but your experience may differ with another mount and its unique motor control setup. I would try pulse guiding if you are also using a PC to correct for periodic error or ST4 control if not.

The most popular (and free) stand-alone autoguiding program is PHD2 (“push here dummy”) for Apple and Windows platforms and is compatible with both still cameras and webcams. Several image capture programs, including Nebulosity and Sequence Generator Pro, interface with PHD2 so that it temporarily stops guiding (and potentially hogging computer and USB resources) during image download. Guiding can sometimes be inexplicably stubborn and some users prefer two other free alternatives, GuideMaster and GuideDog, both of which favor webcams as guide cameras.

Focus Mechanisms

Many telescopes are shipped with a focus mechanism designed for the light duty demands of visual use and these mechanics struggle with the combined weight of a camera and filter wheel. Some manufacturers offer a heavy-duty alternative for imaging, or an adaptor to one of the popular third-party designs. There are three main types of focuser design; the Crayford, rack and pinion (R&P), and the helicoid mechanism. Although the helicoid is extensively used on photographic lenses, it is only practical proposition for short focal length telescopes that do not require a large focus travel or motor control. Crayford and R&P designs make up the majority of the product offerings and as before, although the design architecture influences performance so too does its execution.

The Crayford focusing mechanism was originally designed for amateur astronomers as an economical alternative to the rack and pinion design. Opposing roller or Teflon® bearings support a focusing tube and push it against a sprung loaded metal roller. This roller is fitted to the axis of the focusing control and grips the focusing tube by friction alone. This system has no backlash but can slip under heavy load. (Unlike photography where one is typically pointing the lens horizontally, in the case of astrophotography, you are often lifting and lowering the camera mass with the focus mechanism.) Mechanical tolerances play an important part of the overall performance: I had two Crayford mechanisms that had good grip at one end of their travel but became progressively weaker towards the other end. I could not find a tension setting that allowed the focus tube to move smoothly and without slipping across the entire focus range. In my own setup, I quickly replaced the original Crayford mechanisms with a quality rack and pinion focuser.

fig.27 This underside view of a focus mechanism shows the “rack” of the rack and pinion mechanism. These groves are slanted to improve lateral and longitudinal stability. Other designs are orthogonal and although cheaper to make, can suffer from lateral play.

fig.28 A Crayford design simply has a smooth metal surface on the underside of the focus tube. This can be a milled surface or a metal bar. For optimum grip it should be kept clean and free of lubricants. This view of the underside of my original telescope’s Crayford focuser shows the polished metal bar that is clamped by friction alone to the metal rod passing between the focus knobs. The tension adjuster is in the middle of the black anodized supporting block. The multiple holes are for aligning the bearings that run along the focus tube, so it runs true. The focus adjuster on the right hand side has a reduction drive within it for precise manual operation.

fig.29 This view shows the stepper motor, assembled onto the end of the focuser shaft. In this case the knobs are removed and the motor’s own gear reducer scales down the stepper motor resolution to ~4 microns/step. Usefully, it has a temperature sensor mounted on the connector, that quickly tracks ambient changes. Some designs employ the focuser’s gear reducer and have a simpler motor mechanism.

fig.30 This is the control module for the popular Lakeside focus system. It houses the PC stepper motor interface and you can use it to manually change the focus position and compensate for temperature effects in the focus position. A USB port enables it to be controlled by a computer; essential for autofocus operation. It connects via a ribbon cable to the motor. I made a custom motor cable, so that it was easier to route through my telescope mount’s body.

Rack and pinion focus mechanisms replace the friction drive with a toothed gear-train. Gears do not slip but suffer from backlash. In this case a change in the focus knob direction does not immediately translate into a change in focus. Backlash is easily overcome by approaching the focus position from one direction only and if the focuser is motorized, this direction can be automated. The implementation again is crucial; I purchased a new telescope with its latest rack and pinion focus mechanism, (replacing a previous lightweight Crayford design). There was no slippage in travel but the tube wiggled from side to side and up and down unless the focus lock was tight. This made manual adjustment impractical and was unusable as part of a motor-driven system. I have learned to be more cautious; I always test a new focus mechanism for smooth operation over the entire focusing range and I make sure it can lift the imaging equipment and hold it without slipping. I also check for lateral play as well, which can ruin an image from movement during an exposure or tilt. This may be detected by hand or by looking through a medium eyepiece and noting the image shift as you adjust the focus position. (Even a high-quality focuser may have a slight image shift between focus positions.)

Motorized Focusing

Both Crayford and R&P focusers normally have a geared reduction drive on the focusing knob for fine control and where there are gears, there is backlash. Motor drives are available in a number of configurations, some of which are more suitable for remote operation and autofocus programs. The motors themselves couple to the focusing mechanism by varied ingenious solutions. Some motors directly couple to the focus shaft and require removal of a focus knob. Others use toothed belts and pulleys around the focus knob. The DC servomotor or stepper motor is normally held in position by a bracket fastened to the focus mechanism. DC motors combined with Crayford focus mechanisms offer an economical way of hands-free focusing, using a small, wired control-paddle. For computer control, stepper motor varieties offer precise, repeatable absolute positioning, especially when attached to a rack and pinion focuser. Any movement is precisely defined by a number of steps rather than an analog voltage and duration on a DC servomotor.

Even though R&P focusers will inevitably exhibit backlash, the better control programs drive to the final focus position from one direction only, normally against gravity, eliminating its effect. Microtouch systems are designed for Feather Touch focusers but there are many other motor control systems that can be used with Feather Touch and other focus mechanisms, via an array of ingenious brackets, for example those from Robofocus, Lakeside Astro, Rigel Systems and Shoestring Astronomy.

fig.31 This screen grab from Maxim DL shows an automated focusing sequence. The graph shows the width (half flux width in this case) of the star as the focus tube is slowly stepped in. It then works out the optimum position and moves the focus tube back to that point. The whole process takes a few minutes.

A highly regarded manufacturer of R&P focusers is Starlight Instruments, who manufacture the Feather Touch® range and also a matching direct-coupled stepper motor drive system. Their focusers come in a number of sizes and compatible with many refractor and reflector designs. I used a 3-inch and 3.5-inch Feather Touch focuser on my refractors, fitted with Micro Touch® motors that are controlled by a small module. It was not possible to fit all my telescopes with these motors and I changed over to Lakeside Motor units and a single embedded module in my interface box. These modules remember the focuser position when it is switched off and I return the focuser to a reference position on the scope during my shutdown process. In that way, the absolute focus position is known if I switch scopes. A good focuser is a precision assembly and expensive to manufacture. The control modules, have a simple button interface, remote PC control and often feature automatic temperature compensation. This last feature is an interesting one; when a telescope changes temperature, it expands or contracts and the focus position changes. In some cases the focus shift is sufficient to degrade an image and needs to be compensated for. By logging the precise focus position for a range of ambient conditions, a focus steps/degree compensation value may be determined, although the elements of an optical assembly can heat and cool at different rates and the steady state and transient focus position may be different. The focus motor control box monitors the ambient temperature and it (or the PC) issues a small adjustment to the motor position. To avoid image shift during an exposure it is better for the PC to decide when to adjust the focus, normally between exposures. Motorized focusers are brilliant but they have a drawback; the motor and gearbox lock when un-powered and without a clutch mechanism prevent any manual focus adjustment using the focus knob. You need to use the module’s buttons and provide a power source for visual work.

Other high-quality after-market focusers include those from Baader, APM telescopes and MoonLite Telescope Accessories, who specialize in Crayford style designs, offer many adaptors for refractor and reflector applications, as their own DC and stepper motor control systems.

Interface Speed

At one time, beguiled by promising reviews on various forums, I swapped a power hungry laptop for a tiny netbook with a 12-hour battery life. Whilst the image capture, guiding and simple planetarium programs ran concurrently, I soon realized that the images had strange stripes across them. The forums suggested it was a USB speed issue but after upgrading the hard drive to a solid-state model and streamlining the system, although it dramatically improved overall performance, the stripes persisted. I then discovered that some CCDs need to download their images over USB without any interruption, since any delay affects the image data in the sensor. Although this banding effect is small, it is exaggerated by subsequent image processing. In the end, swapping back to a laptop fixed the problem. Although the laptop processor speed was only twice as fast, the laptop had additional graphics and interface electronics that reduced the burden on the microprocessor. Whichever interface solution you choose, it should be able to cope with fast image downloads and in the case of planetary imaging, up to a 60 fps video stream without dropping frames. Reliability is the key to success.

Computing Hardware

Laptops are the obvious choice for portability or for a quick exit from a damp shed. My MacBook Pro is an expensive computer and although I ruggedized it to protect it from accidental abuse, I could have chosen a less expensive Windows laptop for the same purpose. Battery life is an obvious priority for remote locations and an external lithium battery pack is an effective, if expensive, means to supplement the computer’s internal battery to deliver a long night’s use. For those with an observatory, assuming it is weather-proof, you can permanently install an old desktop PC or seal a low-power miniature PC into a weather-proof box. The demands placed on the computer hardware are not as extreme as that needed for modern PC games or video processing and it is possible that a retired PC from an upgrade may be ideal for the purpose. There are some gotcha’s – as I mentioned before, a netbook may appear to be sufficiently powerful and inexpensive but a 1.6 GHz, 1 GB netbook had insufficient processing resources for image capture from an 8 megapixel CCD camera. Any computer will require several USB 2.0 ports and possibly FireWire, Serial and Ethernet too. Having a network option allows for later remote operation. Backup is essential and a high capacity external hard drive is essential to store the gigabytes of image and calibration data that astrophotography quickly acquires. After each night’s imaging, I copy the image data over to my external drive and keep that safe. It pays to keep the original data since, as your processing skills improve, another go at an old set of image files may produce a better end result.

Operating Systems

Alternative operating systems are a favorite punch bag for Internet forums. After writing my last book, I favored Apple Mac OSX rather than Windows XP. I use Macs and PCs daily and although I prefer the OSX experience, there are simply more astronomy programs available for the Windows platform. Having said that, one only needs a working system so a choice of say 10 rather than 3 planetarium applications is not a big deal. A more important issue though is hardware support: In OSX, the astrophotographer is reliant on the application (or operating system) directly supporting your hardware. That can also apply to Windows applications but usefully, many hardware manufacturers support ASCOM, a vendor independent group of plug and play device drivers that provides extensive hardware and software compatibility. ASCOM only works in a Windows environment and although Mac software will improve with time, presently the image capture, focuser and planetarium applications do not support all available astronomy hardware.

I started down the Mac road for astronomy and was able to put together a system, including a planetarium (Equinox Pro / Starry Night Pro / SkySafari), image capture (Nebulosity), autoguiding (PHD) and image processing (Nebulosity and Photoshop). I produced several pleasing images with this combination and could have continued quite happily on a MacBook Pro with its 9-hour battery life. As my technique improved, I became more aware of alignment and focusing issues and I eventually abandoned the otherwise excellent Nebulosity and PHD for Maxim DL, MaxPoint and FocusMax on a Windows platform. (The only application that offers full control in OSX is TheSkyX with its add-ons.) Things move on and my system has evolved to TheSkyX, PHD2 and Sequence Generator Pro in Windows.

Image processing software is increasingly sophisticated and a modern computer will process images quickly. Most current astronomy applications are 32-bit but some (for example PixInsight) only work in a 64-bit environment. A purchased version of Windows 7 has two install DVDs; a 32-bit and 64-bit version. The real advantage of 64-bit Windows 7 is that it can access more than 4 GB of memory to support multiple applications. A few useful utilities will only run in 32-bit windows (for example PERecorder) but over time these will become the exceptions. Windows platforms come and go; I quickly moved from XP to Windows 7, skipping Vista, resisted the tablet temptation of Windows 8 and finally moved over to Windows 10. I still use my MacBook Pro and by using Boot Camp, I can run both Windows for control, capture and image processing and OSX 10.11 for publishing. I am fortunate to have a foot in both camps (more by luck than judgement) and the best of both worlds!

Software Choices

Astronomy software packages offer a dizzy range of capabilities: Some do a single task very well; others take on several of the major roles of telescope control, acquisition and image processing. There are two major integrated packages; Maxim DL and TheSkyX. Neither is a budget option (~$600) but they are able to display, control, capture, autofocus, guide, align and process images (to varying degrees). Maxim DL includes their own drivers for many astronomy peripherals and connect to other hardware using ASCOM. TheSkyX is developed by the same company that manufactures the exquisite Paramount equatorial mounts and not surprisingly, their software has close ties to their own equipment, SBIG cameras and additionally promote their own interface standard X2 for other vendors. Recently they have expanded TheSkyX hardware compatibility with native support of ASCOM and Maxim DL drivers. No application is perfect and many users selectively augment their capabilities with the additional features of FocusMax or the unique abilities of specialist image processing and manipulation applications such as Deep Sky Stacker, AstroArt and PixInsight to name a few. Usefully, the major applications are scriptable for automated control and have the ability for enhancement through plug-ins and remote control.

fig.33 I thought it would be a good idea to trawl the Internet and create an extensive list of software titles used in astrophotography. It is a daunting task to list them all, let alone document the capabilities of each. These two tables then are an indicator of the choice available in 2017 of the more popular programs and utilities, their operating system compatibility and general functions. The choice, pricing and features will change over time but even so it serves as a useful comparator. In addition to these are numerous others on mobile platforms (iOS and others).

fig.33 (continued) Some packages, marked**, have guider and plate solving capability by linking to freeware applications, for example PHD, PHD2, astro tortilla and Elbrus. Most software is distributed through the Internet rather than by CD / DVD. An Internet search and browse of each software title will find its website and its latest features and pricing.

Deciding which applications to choose is a difficult and personal process. You will undoubtedly get there after a few false starts and as your experience grows you will likely move away from the simpler applications to the heavyweight ones. Thankfully, many companies offer a full-featured trial period with which to evaluate their product. This may give adequate time to check for hardware compatibility and basic performance (providing you have clear skies).

It is quite tempting to continually change applications and workflows, as a result of forum suggestions and early experiences. Something similar occurred in photography with the allure of magic film and developer combinations and it is perhaps better to stick to one system for a while and become familiar with it, before making an informed decision to change to something else. Choosing applications is not easy and I think it helps to consider how each application meets the individual needs of the principal functions:

Planetarium and Telescope Control

There are many applications covering these functions, with the principal difference between programs being their hardware compatibility and ergonomics. Some are very graphical, others less so, with correspondingly lower demand on computer resources. For imaging purposes, they all have sufficient data and precision to plan and point most mounts to the target object. I own the pretty and educational PC/Mac fully featured programs but often use a simpler iPad application SkySafari, for image planning or simply enter the target object into the Maxim catalog tab. Some fully featured packages, like Starry Night Pro, acquire images too, either directly or through an interface to a separate application. C2A is fast and free. It has a simple quick interface yet extends its capabilities through links to other programs such as Maxim and PinPoint for alignment, imaging and plate solving. Once the imaging sequence is under way, the planetarium is largely redundant. For this purpose, I rate the planetarium applications by their ease of navigating around the sky; searching for an object, displaying its information, zooming in and displaying the camera’s field of view for any particular time in relation to the horizon and meridian. These are basic requirements but there is an amazing difference in the usability of available programs.

fig.34 CDC and the lesser known C2A (above) planetariums are both free. They are able to reference multiple catalogs and have object filters to just display the information you need without being obscured by millions of stars and labels. They can also interface to image capture programs and a telescope mount, to direct it to the object on the display and perform basic sync functions. The simpler graphics on this application use less computing power than the more expensive image based planetariums.

If the pictorial aspects of a planetarium are not required but one simply requires a planning tool to identify promising targets,AstroPlanner and Skytools 3 offer an alternative database selection approach, selecting promising objects using a set of user-entered parameters. These may be a combination of position, size, brightness and so on. These also interface to mount control and alignment programs as well as direct target input into image acquisition programs. Most mount manufacturers define an interface control protocol that allows PC/Mac/mobile control through a serial or USB port. Physically, these either couple directly to the scope or via the mount control handset. Some of these drivers are fully-fledged applications too, emulating handset controls and settings on a computer. In the case of the popular SkyWatcher EQ mounts, an independent utility, EQMOD (free), largely replaces the handset and allows direct PC to mount control, including PEC, modelling, horizon and mount limits, gamepad control and pulse-guiding as an ASCOM compatible device.

Several mount interface utilities have a database of prominent guide stars and in the instance of EQMOD and MaxPoint can calculate a pointing model from a series of synchronized alignments across the sky. A pointing model can account for the mechanical tolerances of the mount, sky refraction and polar alignment and generally improves the pointing accuracy, which leads nicely onto the subject of Astrometry.

fig.35 A typical plate-solve result from Maxim DL. Using the selected image and an approximate image scale and position, it accurately calculates the image scale, rotation and center position. This result can be used in a variety of ways; including building an alignment model for the telescope, synchronizing the telescope position to the planetarium application or a prior image and for aligning and combining images during processing.

Astrometry

Slipping in between camera and mount control is plate solving and astrometry. This function correlates an image with a star database and accurately calculates its position, scale and rotation. There are many stand-alone applications that include: PinPoint (a light edition is provided with some premium versions of Maxim DL), Elbrus, PlateSolve2, AstroTortilla and Astrometry.net (all free), the last of which is web-based. (There is a locally served version of Astrometry.net server too.) Premium versions of TheSkyX also have plate solving capabilities. Used in conjunction with camera and telescope control programs, a pointing model for the sky is quickly established. It is a wonderful thing to behold, as the telescope automatically slews to a sequence of stars, exposes, correlates to the star database and updates its accuracy. It is not necessary to manually center a guide star each time; the program just needs to know generally where it is pointing and the approximate image scale and does the rest. This feature improves the general pointing accuracy or more simply, for a single object alignment, a plate-solve nearby and a sync with the telescope controller quickly determines the pointing error and necessary adjustment.

Plate solving additionally enables precise alignment to a prior image. This is a very useful facility for an imaging session that spans several nights and that requires the same precise image center. For these purposes a real-time plate-solve program is required for quick positional feedback. These programs also provide the muscle to identify a supernova, replacing the age-old technique of flicking between two photographs.

Camera Control and Acquisition

The thing that sets the acquisition applications apart is their hardware compatibility and automation. The actual function is straightforward but there are many more camera interfaces than there are mount protocols. Many photographers start astrophotography with a handy SLR and most applications will download RAW files from Canon, Nikon and some of the other major brands. Some of these applications support dedicated CCD cameras too. The fully-featured applications include sequencing, filter wheel and focuser control as well as the ability to pause autoguiding during image download. Before choosing an application, double-check it reliably supports your hardware or indeed, future upgrade path, via its support web page or from a forum search. Bad news travels fast and it is often apparent how long it takes to fix a bug or update a driver. Nebulosity is a cross platform favorite that also includes image-processing capabilities. For Windows, Sequence Generator Pro is critically acclaimed and worth every cent. In addition there are the heavyweights, Maxim DL and TheSkyX, both of which have a subscription-based upgrade path.

Guiding and Focusing

The program choice for guiding software is considerably easier. There are only a few standalone applications, of which PHD2 is far and away the most popular. You can also use the guiding functions within a package like Maxim DL, TheSkyX or AstroArt. When choosing the application the compatibility and ease of use are the primary considerations. Most applications support guiding via a ST4 port but not all can pulse-guide. Some applications only can use a webcam as a guide camera and require a bright star for tracking.

A good focus application should be able to acquire images, measure star profiles and control the focus position. The focus module suppliers normally provide a stand-alone focus control application and a driver (often ASCOM compliant), to run with the focus applications. Again, the big packages include integrated autofocus modules and interestingly both TheSkyX and Maxim DL acknowledge and promote FocusMax (originally free but now marketed by CCDWare) as an enhanced tool to augment their own autofocus module. FocusMax expands the functionality; once it has characterized a setup, it obtains precise focus in less than 60 seconds, remotely. If the full version of PinPoint is installed, it can also directly interface to a telescope and perform an automatic slew to a star of the right magnitude, focus and return. Utilities such as these keep me in the warm and away from insects.

fig.36 Laptops are great but the small screen can become quickly cluttered when multiple applications are running at the same time. In particular, FocusMax quickly proliferates windows across the screen. When I am at home, I use a second monitor so that I can place all the windows next to each other, just like Houston control!

On the subject of staying in the warm and getting some sleep, I recently invested in a cloud detector (fig.37). This little device detects temperature, light, rain and cloud and with an add-on, wind speed. Using a RS232 data link, a small program on the PC (or Mac) shows the environment status. It also has an alarm state that will execute a script. I originally set mine to cause an alarm on my iPhone using an iOS App called the “Good Night System”. The unit also has a relay output to indicate bad conditions and can also execute a script for more advanced control options. More recently, ASCOM safety device drivers use its information to provide “safe to open” and “safe for imaging” status to the roof and imaging control applications.

fig.37 A useful acquisition for a permanent installation is a cloud detector. This AAG CloudWatcher uses an IR and temperature sensor. A heated rain detector is included too, as is an anemometer input.

Software Automation (Scripting)

Automation is a specialist requirement that performs a sequence of actions to run on the computer, rather like a macro. Some imaging programs (for example Sequence Generator Pro) have considerable automation functionality built in, other mainstream packages are script-enabled. A script is a programmed sequence of actions; for instance, the startup, focus, alignment, imaging sequence and shutdown for unmanned operation. Scripts look intimidating at first and there are two methods to avoid an abrupt learning curve; the first is to find an existing script that can be simply modified, the second is to use an external program, like ACP, CCDAutoPilot or CCD Commander that provide an accessible way of creating an instruction sequence. Most of the practical chapters use Sequence Generator Pro; this modern software may not have the overall expansion capabilities of Maxim DL (via scripting) but the package offers accessible and reliable automation that covers all the bases and at a modest price.

Image Processing

Image processing is a huge arena but essentially can be thought of as a sequence of distinct steps, starting with calibration, alignment and stacking to more advanced processing and enhancement, by both mathematical algorithms and user choice. Some applications offer a full suite, (for instance AstroArt, Maxim DL, Nebulosity and PixInsight), others specialize in processing video files, (RegiStax, AutoStakkert and Keith’s Image Stacker) or in a particular aspect of image processing, like DeepSkyStacker that calibrates, aligns and stacks exposures. The program choice is particularly tricky, since there are few hard and fast rules in image processing and the applications are constantly being refined. The practical chapters use a number of different programs to give you an idea of the software capabilities.

Image processing skills develop gradually and at a later date, you will almost certainly be able to produce a better final image from your original exposures, either due to processing experience or the tools at your disposal. (Did you remember that suggestion to duplicate image files on a large external storage drive?) Some of the image-processing applications are overwhelming at first. The trick is to keep things simple at the start, develop your own style and identify where additional tools might be of assistance. These tools exist because the otherwise heavyweight Adobe Photoshop is not purposed for astrophotography and its highly specialized needs. One reason is that many editing functions are limited to 16-bit processing and the initial severe manipulations required in astrophotography are more effective in 32-bits. Photoshop or similar programs may usefully serve a purpose after the core image processing, for photographic manipulation and preparation for publishing. That is not to say that image enhancement is impossible with these imaging programs but it requires a good knowledge of layers, masking and blending techniques. These sequences can often be stored and used again at a later stage. A number of useful astronomy Adobe “actions” are bundled and sold by astrophotographers for image manipulation; a popular example being the astronomy tools by Noel Carboni. Other devious techniques using multiple layers and blending options are to be found scattered, like the stars, across countless websites. I have learned more about sensors and digital manipulation in 2 years of astrophotography than in 10 years of digital photography.

Utilities

The most useful of the utilities improve the reliability of your polar alignment, exposure, optical alignment and mount control. There are perhaps more polar alignment utilities than there are decent imaging nights in a UK year. I can see the appeal if one has an observatory and can dedicate an entire night to polar alignment but for a mobile setup, I find a polar scope and autoguiding are sufficient. A polar misalignment of 5 arc minutes gives a worse case drift-rate of about 1.3 arc seconds per minute. The recent PoleMaster from QHY is a camera-based polar scope that achieves sub 30 arc seconds alignment in about 5 minutes!

Exposure calculation is particular interesting. It is not a trivial subject and there are no hard and fast rules. The optimum exposure depends upon many factors that are unique to your equipment and sky conditions. Exposure utilities are often plug-ins to image capture programs and compute an optimum exposure based on a target signal to noise ratio, sky illumination, number of sub exposures and sensor noise characteristics. There is considerable science behind these utilities but they are not as reliable as normal photographic tools and an exploratory image is often a better approach.

Image analysis tools, for example CCDInspector, can be quite useful to check the optical properties of an image. Among its tools, it can derive field tilt and curvature from individual star shapes throughout the image and measure image vignetting from the background intensity fall-off. These assist in setting the right field-flattener distance and confirming the couplings and sensor are orthogonal to the optical path.

There are numerous other utilities, many free, which provide useful information; Polaris hour angle, alternative times systems, GPS location, compass, electronic level, meteor shower calendar to name a few. Some reside on my iPad, others make their way to Windows. Remote control has enabled several companies to set up dedicated remote-site operations that are rented out to astronomers all over the world. Their locations are chosen carefully, in areas of frequent clear skies and low light pollution. The details of two typical operations can be found at w­ww.­ite­les­cop­e.n­et and w­ww.­lig­htb­uck­ets­.c­om. It is like being a grandparent with grandchildren; you do not need to own a telescope to do deep sky imaging and you can hand it back when it goes wrong!