CCD FROM THE GROUND UP

by Ted Saker, Jr.

© 2000, 2010 Ted Saker, Jr. All Rights Reserved.

Getting Started: Exploring the Brave New World of CCD (1998)

I decided to make the investment in a new telescope in late 1991 or early 1992. My ambition at the time was to discover a way to use computers and computer technology to take deep sky images from my backyard. I quickly discovered at the time that the equipment was either unavailable or cost-prohibitive to obtain. I purchased a Celestron Super Polaris C8, the Schmidt-Cassegrain OTA on a German equatorial mount made by Vixen. I began my astrophotographical adventures with photoemulsion planetary imaging with eyepiece projection. As the years went by and costs came down, a variety of cameras in many price ranges entered the marketplace.

CCD cameras suitable for deep sky astrophotography cost anywhere from $2,000 to $8,000. The costs are mainly related to the size and sensitivity of the imaging chip. The current state of the art is the Kodak KAF1600 chip used in the SBIG ST-8 camera. Cameras suitable for planetary or other types of applications cost much less, anywhere in the $500 to $1,500 range.

Computer hardware is required in virtually all cases and may add extra costs. One question must be answered: laptop or desktop? Desktop equipment is more durable and cheaper, but if mobility is important, hauling around a case, keyboard and monitor will quickly pall. Laptops are much easier to transport, but are more expensive and may be less tolerant of cold and moisture found in the field. Computer software for camera control and image processing must also be obtained. Camera manufacturers usually provide their own proprietary control programs; however, there is at least one commercially available unbundled package that claims to control the camera and process the image. It has received a favorable review in a major astronomical magazine. The other major question is how much processing power is needed to run the control and image processing software. The answer to this question can impact the costs as well.

Finally, as with any computer based technology, unplanned obsolescence can overtake the CCD imager. Today’s state of the art imaging rig may well be tomorrow’s Astromart fodder.

My Vision: Joining the CCD Revolution on a Shoestring Budget

I wanted to obtain deep sky images from my backyard under extremely light polluted conditions. The key was a small investment; therefore, using the existing telescope and mount, a Celestron Super Polaris C-8, purchased in 1992, formed the optical base for the system. It has a good light grasp and compact design. The mount is obsolete, having been succeeded by the Great Polaris series of German equatorial mounts which in turn have been supplanted by the new G8 series. I also learned about the book "The CCD Camera Cookbook" available through Willman-Bell publishing. To buy or build, that was the question. The answer was directly related to the amount of funding and the amount of time. The question was decided as I had much of the latter and little of the former, plus a spirit of adventure in abundance.

Choosing a Camera: The Cookbook 245 (1999-2003)

I chose to build the Cookbook 245 instead of buying a commercially designed and manufactured model. I figured I could handle a soldering iron and power tools. The total investment for a deep sky CCD camera was under $500, including the chip. The diagnostic and camera control software was provided with the book.

The heart of the Cookbook 245 is the Texas Instruments TC 245 chip. It has a 378 x 245 pixel array, at 19 x 25 microns per pixel. The imaging area of the chip is only slightly smaller than the KAF-400 chip used in the SBIG ST-7 camera. The CB 245 is both liquid and thermoelectrically cooled. The cost of the chip was around $100. Many other electrical parts were available through local Radio Shack stores. Some parts not carried by the store had to be ordered through RS Unlimited. Shipping very prompt. University Optics of Ann Arbor, Michigan, carries a kit that contains many of the harder to find parts as well as the machined aluminum pieces for the camera head and a peltier module for thermoelectric cooling of the chip. I scrounged other parts from several sources. The total construction time was approximately three and a half months of evening and weekend work.

Cookbook Camera Quirks

Having had no experience with CCD imaging, I found a steep learning curve existed with learning to use the camera and software. Cooling the chip is critical to reduce dark current and provide a clean image free of graininess. Heat is the enemy. The CB 245's peltier module provides thermoelectric cooling while a water/alcohol mix for the liquid cooling system conducts the waste heat to a coil immersed in an ice water bath. The version of the control software provided with the book runs under MS-DOS, and even a lowly 80286 (AT) computer can handle it. The unit is self-warranted: repairs can be performed “in house”.

Integration with Telescope

Finding, focusing and placement of the object on the chip's image plane is essential. The trick is that the image area of the TC 245 is offset from optical axis unlike film. Focusing is especially crucial. If the object is out of focus even the slightest bit, the camera cannot obtain an image. Further, a time lag exists between the point when the camera obtains an image and the point when the computer displays the image. This download lag makes focusing on the fly difficult at best. The real trick is coordinating the image plane and observing plane to find and position the object on the chip. A flip mirror system is highly recommended: I selected the Taurus Tracker III for finding and focusing on deep sky objects. The Tracker III has a finding/focusing port as well as a guiding port. The parfocalization process places the image plane of the camera at the same distance from the flip mirror as the image plane of a high powered eyepiece. Thus, once the focused object is placed in the proper location in the eyepiece field of view, the camera will obtain a proper image of the object. The Taurus III accepted my Celestron f 6.3 focal reducer to provide a wider field of view for easier acquisition of objects and shorter integration times. However, I soon encountered a backfocus problem: the chip was too far to the rear of the image plane. A member of the CAS came to my rescue by milling an aluminum adapter that placed the camera head at the correct distance from the flip mirror.

CCD imaging has several advantages over photoemulsion. CCD chips are much more sensitive than film; thus, exposure times are much shorter. In many instances, manual guiding is not necessary. For long integrations needed for objects having a low surface brightness, autoguiding is a real advantage. The CB245 is adaptable for autoguiding. Unfortunately, however, my Super Polaris mount cannot accommodate autoguiding without expensive and radical modifications. One supplier suggested buying a new telescope if I wanted autoguiding and/or "go to" capability.

Even with the advantage of working with a CCD chip instead of photographic film, the problem of finding dimmer objects in lower than less than ideal conditions still posed a serious problem. "Go to" capability would solve much of the problem since a computer can position the telescope more accurately than a person can using setting circles or by star-hopping. Once again, the obsolescence of the Super Polaris mount caused a problem. It is not designed for encoders that are needed to communicate positional angles to the computer, nor has any after market manufacturer saw fit to design, manufacture or market variable speed motors for the SP mount to slew to objects and track them.

After reading an article on the "computer-telescope interface" in a major astronomical magazine in early 1999, I surmised it might be possible to modify the SP mount for use with a computerized star atlas, but it would never be a true "go to" mount. The best I could hope for was to install a computer-telescope interface that would in essence turn the telescope into a computer mouse with the direction it points at displayed as a large cross-hairs on the computer's monitor. I selected Brian Kidwell's Deep Space Navigator for the computer-telescope interface and David Chandler's Deep Space computer program for the star atlas. I would have to manually slew the telescope to the correct position, then find, focus and guide. This system has proven to be reliable and cost-effective even if it isn't the latest and greatest.

Imaging and Image Processing

Knowing how CCDs work is necessary. Film uses photographic grains, CCD chips have picture elements, known as pixels. Pixels are far more sensitive than grains; however, both suffer from the effects of limiting magnitudes imposed by light pollution that limit integration or exposure lengths. Non anti-blooming gate (NABG) CCD chips generally suffer from "blooming", the appearance of bright streaks when the camera is pointed at a bright object, usually a star. Anti-blooming gate (ABG) chips, like the KAF 1600L chip (more on that below) are designed to eliminate blooming at the cost of reduced sensitivity.

The astrophotographer who uses film is enslaved to the photodeveloper unless he has his own darkroom. The developer has an enormous amount of discretion in the color balancing of the image. Many of the films in use today give the night sky a greenish hue unless the developer takes great care during the development process. If the astrophotographer desires to post the images on the internet, the images have to be scanned as well. Scanning offers some advantages. If the negative is scanned instead of the print, the astrophotographer can balance the colors using a computer program that normally comes with the scanner, thereby avoiding the tyranny of the developer. Also, there's the time the developer requires for developing the negatives, plus the costs. All that nickel and diming adds up over time.

The CCD camera downloads the image directly to the computer with no "middleman". Shooting dark frames is absolutely required in order for the image processing software to cancel out the "dark noise" present in every image. A dark frame is an image taken with the camera covered. The concern I had was that the CB 245 format is proprietary. None of the commercially available packages I am aware of support that format. Luckily, one of the designers of the CB 245 wrote software that converts the CB 245 format to a standard FTS file, allowing further processing of the image with other commercially available packages. It also has a very excellent ability to subtract the dark frame from the image, yielding an image that is much higher in real data, lower in noise and is easier to process.

Light pollution tends to affect CCD cameras less than photoemulsion (PE). However, the chip is not immune from the effects of light pollution. Light pollution will wash out an image if the integration time is too long. A CCD image may be viewed immediately after acquisition in sharp contrast to PE. Unless you have your own darkroom, PE requires a trip to the developer, a wait for the pictures, paying for the development and a dependence on the technician for proper color balance. Unless one knows a developer who is willing to spend time to color balance the image correctly, the print will not display the image accurately.

On the other hand, CCD images are mainly limited to viewing on a computer screen unless one owns a high resolution printer. Also, quality of the image display is dependent upon the resolution of the monitor.

Decisions, Decisions

Whether or not to invest in a CCD imaging system depends on whether a person is an astrophotographer or an observer. Is CCD really necessary? The answer is a definite yes if one lives in the city and has any ambition of imaging deep sky objects. Photographic film is still viable if the deep sky object is fairly large and bright, like the Great Orion Nebula, but it just won't cut it with other objects having lower surface brightnesses.

Buying a camera carries certain pros and cons. The pros: the buyer gets a company backing the hardware and software with warranty guarantees; if one's budget allows, state of the art is within reach, and many manufacturers work with software developers to provide seamless programs. The cons: in the event of defects or breakage, the buyer can lose the camera for a long time if repairs necessary; the cost of buying a camera suitable for deep sky work increases depending on the chip and the bells and whistles built in. In other words, the cash investment increases the closer one gets to state of the art. As newer technologies enter the market, today's state of the art may become obsolete rather quickly.

Building a camera likewise brings pros and cons to the table. The pros: one knows the design and hence, can repair it. I once had to perform a quick field repair on the camera head. One can minimize the amount of cash invested while enjoying all the advantages of learning how to conduct CCD imaging. The experience of building a camera can also increase one's knowledge of electronics that can be applied in other projects. The cons: It's not state of the art in hardware or software. It can be challenging and frustrating to gather everything necessary to make the whole rig work.

Recommendations

• Build the Cookbook 245 if budget is a major factor and/or one has the skills and perseverance to complete a project.

• Follow the book’ s instructions and one can hardly go wrong.

• If one runs into the unexpected, the designer and co-author will answer questions if one asks really nicely.

• If one buys a camera, the more one spends, the more one gets. LX200 and ST7 or above owners have the advantage of hardware and software vendors building and writing products especially for that format.

• In particular, autoguiders, adaptive optics are entering the market specifically designed for that telescope model.

Whether one builds or buys, the CCD revolution is within reach for any astrophotographer.

Some images taken with the Cookbook 245 camera (dates and exposure data has been misplaced):

Epilogue: September 15, 2010

This is the eve of my older son's 18^th birthday. This story began shortly before he was born and much has happened since the day I first took delivery of my SP-C8. CCD cameras have come a long way since the original SBIG ST-4.

In summary, I learned that imaging from sites in and around a major city is a difficult undertaking through much trial and error (mostly error). Emulsion photography is pretty much limited to lunar and planetary subjects. I imaged with the Cookbook successfully through early 2002 when the opportunity to purchase the Genesis 16 camera kit and a KAF-1600 ABG chip. That chip was the same one used in the first generation ST-8 that was the flagship camera model of the Santa Barbara Instruments Group.

The Genesis camera saw first light at the 2003 Texas Star Party. I imaged with it for three years, but never really achieved the results I had hoped for. Once again, progress intervened. Technical support for the Genesis camera ended in 2004 as prices for comparable commercial cameras continued to plummet. By the time the Genesis camera died at the 2005 Texas Star Party, commercially built CCD cameras had come down so far in price that I could afford a commercial model which had been very, very pricey only a few short years before. The march of technology sent the self-built camera market into extinction.

I was not ready to give up on the Genesis camera because although the imaging chip was obsolete, it was still serviceable. A consultant friend of mine who had previously owned a Genesis camera and was familiar with its architecture attempted to revive mine. After some months of work, the Genesis camera was still not ready, so I had to press the Cookbook back into service for the 2006 Winter Star Party. Unfortunately, it suffered some kind of tropical meltdown. It's currently sitting on a shelf awaiting a rebuild. Later that spring, my friend offered to purchase the camera and I used the proceeds to purchase my first commercially built camera, a previously owned ST-8 that was equipped with the same class and vintage sensor I used in the Genesis camera. The ST-8 saw first light at the 2006 Texas Star Party.

The generally unsatisfactory results I obtained from ABG chips I used from 2002 to 2008 led me to believe that the greater sensitivity of NABG chips would yield better signal to noise ratio (SNR). The most challenging aspect of urban CCD imaging is obtaining a high enough SNR in the face of severe to extreme light pollution. ABG chips are less sensitive than NABG chips, and the difference affects the ability to obtain a good SN ratio especially in tricolor imaging. One thing I suspect but haven't yet been able to prove is that the NABG chip is superior to the ABG chip. Comparing the Cookbook's NABG chip to the Genesis and ST-8's ABG chips, the Cookbook camera's chip yielded better SNR in my opinion. I thought that the Cookbook's TC 245 chip with its larger pixels and lack of a blooming gate may have been the reason for the better results. Time and experience bore this out in part.

Nonetheless, SNR is and always will be the key to good results. Thousands of low pressure sodium exterior light fixtures create a decidedly greenish background that makes image processing difficult. Obtaining a good, black background and a proper representation of the object is still a significant challenge. The problem isn't limited to tricolor imaging, either. It's very similar to the problem with emulsion photography: light pollution would “fog” the film out before getting a decent image. CCD cameras yield better results than photo emulsion, but could be still better if people turned out a few lights. Even though my goal was to conduct monochrome and tricolor imaging from my urban home, I soon learned that the quality of the image and ease of processing still depends enormously on a dark sky site.

In order to test my theories regarding ABG vs. NABG chips, in early July, 2008, I sold the ST-8 (ABG) camera and acquired an SBIG ST-8E (NABG) camera. It was equipped with the Kodak KAF-1602E, the second generation of the 1600 series sensor. I put my chief theory to the test: whether the NABG chip is superior to its ABG predecessors. The comparison gallery shows the difference in the images obtained with both types of sensors. The second generation chip in the 8E was more sensitive in the blue spectrum than its predecessor.

In the summer of 2010, I sold the 8E and acquired the third generation SBIG ST-8XME. Instead of the parallel port interface I had been working with up until now, the 8XME has a USB interface. The extra bonus is the ability to utilize SBIG's proprietary I2C port which enables the user to control the CFW without separate interface and power cables. Coupled with the replacement of the heavier parallel cable with the lighter USB cable, the wiring harness is much lighter. Losing the extra weight of the cables reduces the load on the OTA, thereby making autoguiding easier. I believe this is reflected in the recent images I acquired with the new imaging rig.

In researching imaging processing techniques led me to adjust how I acquire images with the camera. I believe that these techniques and a NABG chip have led to a vast improvement in the quality of my images. All of the imaging gear I use now is commercially designed and manufactured, but I do miss the challenging days of making my own equipment and learning how to use it. The only significant challenge I face today is effective image processing.

Back to Ted's Astropix CCD page