Two ultrasensitive CCD cameras.
WHEN I BUILT A ROLL-OFF-ROOF observatory for my computer-controlled telescope (March issue, page 8), I was amazed and annoyed when some fellow amateurs asked, "So what are you going to do with it?"--as if stargazing for my own enjoyment weren't enough. But after a few months of visual observing, it dawned on me that perhaps my inquisitors knew something I didn't. Without realizing it, had begun to think what useful projects might undertake.
Maybe it's because of my background as a research scientist. Or maybe it's because I subscribe to all of Sky & Telescope's AstroAlerts, through which professional astronomers almost daily request amateurs' observations of variable stars, asteroid occultations, and other transient phenomena. Whatever the cause, I was soon driven to do some real science at my dark-sky retreat. And that got me thinking about a CCD camera.
The main advantages of CCDs are well known (S&T: September 1987, page 238). These silicon marvels are at least an order of magnitude more sensitive to light than photographic emulsions are. Moreover, their response is linear, so a source that's twice as bright as another produces a signal twice as strong. And because they are digital devices, CCDs generate numerical data that's easy to process with a computer.
Until recently, these advantages were offset--some would say more than offset --by a significant disadvantage: early CCDs were small compared with a frame of 35-millimeter film and positively minuscule next to a photographic plate. Today that's no longer true, thanks to megapixel CCDs for amateurs and mosaic arrays of megapixel CCDs for professionals. In astronomy the photographic plate is all but extinct, and the humble roll of film may not be far behind, especially now that digital point-and-shoot cameras have established a beachhead in the world of astro imaging (August issue, page 128).
Most CCD cameras marketed to amateur astronomers are built around one of several chips manufactured by Texas Instruments or Kodak. The 512-by-512-pixel AP7, introduced by Apogee Instruments five years ago, is different. It sports an SI-502 CCD from Scientific Imaging Technologies (SITe). In a conventional, "front-illuminated" CCD, light passes through control and readout electrodes before depositing its energy into the silicon. Absorption by the electrodes diminishes the camera's response, especially to blue light. In contrast, the SITe chip is illuminated from the back, resulting in a CCD that registers as much as 90 percent of the light falling on it.
This performance comes at a price, however. In order for photons to reach the silicon layer where the signal is recorded, a back-illuminated chip has to be shaved to a thickness of about 10 microns (0.01 mm). Thin chips are difficult to manufacture and are therefore more expensive than their front-illuminated counterparts. With their ultrahigh sensitivity it's no wonder that back-illuminated CCDs have always been the first choice of professionals and the envy of amateurs.
The AP7 has evolved considerably. Early models were priced far beyond the reach of most backyard observers. They also required use of a desktop computer with an ISA-bus interface card, a strike against portability. Then, last year, Apogee introduced the AP7p, which could be plugged into any laptop PC with a standard parallel (printer) port. The camera's price dropped too, and it now costs $5,495--not cheap, but in the same ballpark as top-of-the-line CCD cameras from other manufacturers. And for those wanting the large field of view afforded by a megapixel chip, Apogee unveiled the equally sensitive AP8p, with a 1,024-by-1,024-pixel SI-003, at $12,950.
When you are choosing a CCD camera, the number of pixels is only part of the story: equally important is the size of those pixels. Most popular chips have square pixels that are between about 7 and 24 microns across. All other things being equal, large pixels are more sensitive to light, but smaller ones afford better angular resolution. The "right" size depends on your telescope's focal length, your site's atmospheric seeing, and what you plan to image with your camera.
This simple formula is all you need: pixel size (arcseconds) = 206 x pixel size (microns) / focal length (mm). A useful rule of thumb is to use 1/4- to 1/2-arcsecond pixels for planetary imaging, pixels around 1 arcsecond for deep-sky imaging under extremely steady skies, and pixels 1 1/2 arcseconds or larger for everything else.
Since my Meade 30-centimeter (12-inch) f/10 Schmidt-Cassegrain telescope has a focal length of 3,048 mm, the 24-micron pixels of the SI-502 and SI-003 chips subtend 1.6 arcseconds--squarely in the "everything else" category. Because my New Hampshire skies are rarely rock steady, and because I'm more interested in making scientific measurements than pretty pictures of planets, both Apogee cameras are well suited to my needs. On my telescope, the fields of view of the AP7p and AP8p are approximately 14 and 28 arcminutes square, respectively. For comparison, a frame of 35-mm film covers 24 by 36 arcminutes.
Apogee loaned Sky & Telescope two cameras for evaluation. Senior editor Dennis di Cicco, an experienced CCD imager, used the AP8p with optics ranging from medium-format camera lenses to a Meade 40-cm (16-inch) LX200. Not having touched a CCD camera since my graduate-school days 20 years ago, I took the AP7p with a view toward seeing how quickly I could come (back) up to speed not only on taking images, but also on producing meaningful scientific results.
The camera heads themselves appear identical, except that the AP8p has a larger window and mechanical shutter to accommodate its larger CCD chip. Each unit's aluminum body measures 15 by 15 cm square and 8 cm thick, including external cooling fins and dual fans; each weighs about 1.4 kilograms (3 pounds).
Other supplied hardware includes a standard 2-inch or 1 1/4-inch adapter (depending on your scope's focuser), an AC-to-DC power supply, and a cable set. A 37-pin male connector attaches two shielded cables to the camera. One is a 7 1/2-meter (25-foot) parallel-port cable with the usual 25-pin male D connector at the far end. The other is a 2-meter "pigtail" terminating in a round 5-pin connector; this in turn mates to a cable attached to the power supply.
Apogee does not have its own camera-control and image-processing software. So I used Diffraction Limited's MaxIm DL/CCD for Pentium-based PCs running Windows 95 or higher, since Apogee sells this program for the discounted price of $260 when it's purchased with the camera. The program comes on a single CD-ROM and includes a thick, spiralbound instruction manual in two parts, one for camera control and the other for image processing. I tested version 2.10; the version shipping as of late August was 2.12 (see www.cyanogen.com). An earlier version's image-processing module was reviewed very favorably in Sky & Telescope (July 1999, page 70). Other popular CCD-camera-control programs, most notably Software Bisque's CCDsoft ($249), can also operate Apogee cameras. There is also a camera-control program for Linux-based computers available for $50 from Random Factory (www.randomfactory.com).
The AP7p/AP8p package is rounded out with a 3.5-inch floppy disk containing camera-initialization ("INI") files, as well as a stack of instruction sheets and quick-start guides for both the hardware and, if applicable, the software.
After unpacking everything, I installed MaxIm DL/CCD on my laptop. Diffraction Limited recommends a Pentium-class processor; at least 32 megabytes of RAM; a 16-bit ("high color"), 800-by-600-pixel video display; and 110 megabytes of available hard-disk space. My laptop is relatively new and exceeds all these specifications. MaxIm installed quickly and smoothly.
Hooking up the camera and connecting it to my computer was easy, thanks to a simple diagram and one-paragraph explanation in the instructions. The 2-inch adapter fitted the motorized Crayford focuser that I keep permanently attached to my telescope. As soon as the power supply gets plugged into a wall outlet, the camera's twin cooling fans come on. I found them rather noisy, but not enough to be annoying or distracting, and I never saw any evidence that they produce image-degrading vibrations. It's okay to connect the parallel cable to the computer with the PC either on or off.
Before using the camera the first time, you need to copy the right INI file into MaxIm's program folder. There are instructions telling how to edit the INI file to change various camera settings, but upon reviewing the document I decided that the defaults were best left unchanged.
Ready to take my first image, I consulted the camera's remaining instruction sheets. One, entitled "CCD Camera Operating Instructions," might be better described as "Introduction to Operating a CCD Camera." It lacked specifics but offered a useful overview of cooling, focusing, exposing, calibrating, and shutting down an astronomical CCD camera.
More helpful was the sheet entitled "MaxIm DL CCD Camera-Control Quick Start Guide," a five-page "CliffsNotes" version of MaxIm's camera-control manual. Illustrated with shots of screens you'll encounter as you work with the software, it walks you step by step through loading the data from the INI file (which you have to do only once) and operating the camera. Following its simple instructions, I had my first CCD image (of outoffocus stars) on screen within a few minutes of launching the program.
All interactions with the camera are handled via MaxIm's camera-control window, shown on page 60. First you click on the Setup tab to get the computer and camera talking to each other, turn on the thermoelectric cooler, and set the CCD's operating temperature as low as 50[degrees] to 55[degrees]C below ambient. Next you click on the Focus tab, take a quick exposure of a star field, and choose a small piece (called a subframe) of the resulting image around a single star. Then you let the camera shoot a continuous series of exposures while you adjust the focus between each one. Because you're reading and displaying only the subframe rather than the entire chip (which, for the AP7p, takes about 10 seconds), this potentially tedious process goes quickly. As you get close to focus, you can click on the Inspect tab to examine your star images in detail and make sure they're tack sharp.
If you have a motorized filter wheel supported by MaxIm, you can automatically shoot images through filters one after another in any order, using the Sequence tab. For unfiltered images, this tab functions like a conventional film camera's autowinder, enabling you to shoot a series of identical exposures. This is useful, for example, if you want to combine numerous short, unguided exposures into the equivalent of a single long one. MaxIm includes a convenient align-and-combine routine to make such processing quick and painless.
The heart of the camera-control window is the Expose tab (seen at right). Here is where you choose what type of exposure you'll make and how long it will last. Normal, or light, frames are exposures of celestial targets. The other three types are calibration frames. As explained fully and well in the MaxIm manual, bias frames are exposures of zero duration with the camera shutter closed; they measure the electronic offset of the camera. Dark frames are usually made at the same temperature and with the same exposure time as their corresponding light frames, but with the shutter closed; they measure current that naturally accumulates within the CCD. Finally, flat frames are exposures of a uniformly illuminated field; they reveal pixel-to-pixel differences in sensitivity as well as possible vignetting within your optical system.
In the simplest terms, to calibrate a light frame you subtract a bias and dark frame and divide by a flat frame. MaxIm includes settings to automate this process once you've acquired the necessary calibration exposures, as well as helpful tips on how to obtain the best possible bias, dark, and flat frames in the first place.
Apogee says its back-illuminated chips should take 10 to 15 minutes to cool to 50[degrees]C below ambient and roughly the same time to warm back up. On a night when the air temperature was 13[degrees]C (55[degrees]F), I clocked the cooldown at 17 minutes and the warmup at 20. After a long night of imaging, that last 20 minutes can seem like an eternity, but I resisted the temptation to just turn off the camera and go to bed. Apogee warns that such action could lead to thermal shock of the CCD, degrading its performance for several weeks or more.
It soon became obvious to me that while the performance of a CCD camera depends critically on the hardware, the experience of using a CCD camera depends entirely on the software. Lest this turn into a software review, let me just say that I found MaxIm DL/CCD a delight in all respects. Its interface is so intuitive, and its integrated help menu so thorough, that I rarely had to consult the manual. All basic image-enhancement routines are built in, along with many advanced ones I hadn't encountered before, and all are very straightforward to apply. My only complaint is that the manual could be more concise and better organized; some material is repeated almost verbatim in both the image-processing and camera-control sections.
So what did I actually do with the AP7p? First, of course, I shot images of some of the night sky's showpieces. My uncle, who bought me my first telescope 33 years ago, was visiting from California the night I acquired my first deep-sky image, a 10-second exposure of Messier 51, the Whirlpool Galaxy. We whooped and high-fived when my laptop's screen lit up with the magnificent spiral in all its glory.
It didn't take long to figure out that I could capture a decent image of any Messier object with exposures of 30 seconds or less. Longer exposures would require guiding to prevent smearing images because of the periodic error in my telescope's drive. I could add an autoguider, but I found it much simpler to make a bunch of short exposures and combine them. Using this technique I found that I could produce very impressive monochrome images of star clusters, nebulae, and galaxies with combined exposures totaling just a few minutes. Given that many of the beautiful film and CCD images in this magazine involve exposures of tens of minutes to an hour or more, that's a remarkable statement and a testament to the efficiency of a back-illuminated chip.
I felt comfortable with CCD imaging after only a couple of nights, so I quickly began casting about for interesting projects. Comparing my images of M57, the Ring Nebula, with the chart on page 102 of last September's issue, I found that I could snare stars as faint as 18th magnitude with my 30-second exposures. Since this put me within reach of most of the known asteroids, I decided to begin with positional measurements.
The Minor Planet Center (MPC) of the International Astronomical Union maintains a wonderfully informative Web site at http://cfa-www.harvard.edu/ iau/mpc.html. There I learned how to obtain and report data that would be useful to professionals. I chose two targets. Asteroid 719 Albert was discovered in 1911 and then not seen again until last year (S&T: August 2000, page 22). It was well placed in Pegasus and moderately bright at 16th magnitude.
Minor planet 26760, also known as 2001 K[P.sub.41], is a more recent discovery with only a handful of follow-up observations on the books. It was sailing through Equuleus at 14th magnitude. On several nights in late July I shot images of both asteroids as they moved perceptibly against the starry background.
I grabbed a copy of Richard Berry and James Burnell's new software, Astronomical Image Processing for Windows (AIP4WIN). Using its astrometry utility with reference stars from the Hubble Guide Star Catalog, I spent an evening analyzing my images to derive positions for both targets with an accuracy of just a few tenths of an arcsecond and submitted them to the MPC by e-mail. How good were my results? Within 24 hours I received an encouraging reply bestowing an observatory code, 459, on my New Hampshire hilltop--quite an honor.
When I described my work to a professional-astronomer friend, an interesting project fell in my lap. He's involved in the search for extrasolar planets, in particular those that might signal their presence by partially eclipsing their parent stars. Might I like to try confirming some of their suspected planets? You bet! But the eclipses are very subtle, changing the stars' brightness by at most a few percent. As a challenge and a test, my friend gave me the coordinates of an 11th-magnitude eclipsing binary in Cygnus and the times when upcoming eclipses were predicted to occur.
I cranked up the AP7 on an appropriate night and took sequences of ten 10-second exposures of the star every 15 minutes for five hours. All the images looked exactly the same to my eye. But when I fed them into AIP4WIN's photometry tool, out popped a pleasant surprise: a smooth curve showing the star coming out of a partial eclipse. The total change in brightness was just 0.07 magnitude, and the error bars on each data point were a mere [+ or -] 0.01 magnitude. My friend was duly impressed that such results could be obtained using off-the-shelf equipment readily available to amateurs. So was I!
My experience suggests that Apogee's back-illuminated CCD cameras are excellent choices for anyone interested in the convenience of capturing deep-sky images in the shortest possible times or in making high-quality scientific observations of value to professionals. If your interest lies mainly in imaging, the larger field of view afforded by the AP8p is desirable. If your aim is to do science, the smaller field of the less expensive AP7p should suffice. And here's a bonus: as I was completing my evaluation of the AP7p, it was superseded by the AP7ap, which is based on the new SI-032 CCD. This chip has the same geometry as the SI-502 but offers still higher sensitivity and lower noise. Happily, the price of the AP7ap is the same as that of the AP7p it's replacing.
If you prefer smaller pixels, either because you want higher resolution or because your telescope's focal length is too short to produce a sensible angular scale with 24-micron pixels, you should consider Apogee's new AP47p. It uses a Marconi 1,024-by-1,024-pixel, back-illuminated CCD with 13-micron pixels. On a given telescope the field of view of the AP47p is comparable to that of the AP7ap, but the resolution is twice as good. Apogee is introducing the AP47p at $7,495.
It was a sad day when I had to return my loaner to Apogee. But at least I know what I want for Christmas!
Back-Illuminated CCD Cameras
Apogee AP7p * and AP8p
<ADD> U.S. Prices: AP7p (512 by 512 pixels): $5,495 AP8p (1,024 by 1,024 pixels): $12,950
Apogee Instruments, Inc. 11760 Atwood Rd., Suite 4 Auburn, CA 95603-9075 Phone: 530-888-0500 Fax: 530-888-0540 www.apogee-ccd.com </ADD>
* Now available as AP7ap with higher sensitivity and lower noise at the same price
Editor in chief RICK FIENBERG is glad to be doing science with CCD cameras again after a 20-year hiatus.
|Printer friendly Cite/link Email Feedback|
|Author:||Fienberg, Richard Tresch|
|Publication:||Sky & Telescope|
|Date:||Nov 1, 2001|
|Previous Article:||Colorful echoes of starbirth.|
|Next Article:||Stellar guides for your birthday.|