Advanced Telescope Supplies

Australia's Premier CCD and Astro-Imaging Experts

CCD Imaging. 

Charge Coupled Devices or CCD's are now the tool of choice for most professional observatories and an ever increasing number of amateur astronomers. 

A CCD equipped 130mm aperture telescope is literally  more sensitive than a 1 metre class telescope equipped with a standard photographic camera.

Advanced Telescope Supplies are Australia's most experienced dealers for SBIG CCD cameras. Take a moment to look at our latest CCD Image Gallery.  Then do a search of the web for CCD images taken using cameras other than SBIG. Ask some critical questions. Are the images noisy?  how well guided are they?  how deep do they go? Was a massive telescope used to take the image?

SBIG  offer a comprehensive range of cameras that allow you to capture images using very modest equipment that will rival or better those taken by professional observatories just a decade ago using film. A skilled amateur equipped with a CCD camera can also perform accurate astrometric and photometric studies, spectroscopy and do real science with their camera.

Before you attempt any CCD imaging you will however need the following:

You DO NOT need

Be wary of

 

Note the intensity plot above. It compares an intensity plot of the same star in a pair of images of the same object taken within a few minutes of each other, first through  a 9.25' SCT then through a premium 6" APO refractor.  The smaller instrument shows a higher and sharper peak! While mass produced telescopes may be good value for money, the performance of a superb yet smaller instrument allows users to detect fainter stars with shorter exposures!

What can cause problems

Additionally CCD images can be greatly enhanced by the use of

Which brand of CCD should I choose?

The latest developments in the use of microlens technology has vastly improved the performance of the latest black and white interline and full frame CCD's. When used with interline CCD's their performance can rival that of conventional full frame CCD's. When coupled with full frame CCD's microlensed CCD's such as the Kodak ME series, the Quantum Efficiency, being the ratio of detected photons compared to the total number falling on the detector, peaks at around 85%. Due their structure The best interline devices can achieve is around 20% down from that figure.

A novel method of guiding with an interline device is to use each alternate row of sensors for readout and guiding during imaging, but this effectively reduces the sensors efficiency by 50% and introduces a significant amplifier readout glow in the corner of the image frame, further reducing the Signal to Noise ratio of the sensor. But, why would you want to choose a detector that is virtually blind to nearly half of the the starlight you are trying to capture? Self guiding technology patented by SBIG places no such restrictions on camera performance. But using two ( an optional external remote guide head makes this total of three) independent sensors SBIG cameras can guide using the telescopes main optics ( or in the case of the remote guide head, a separate guidescope. e-finder or off-axis guider) to seamessly guide the telescope will no loss of information going to the main imaging sensor.

 

A little more about CCD's......

Dark current

CCD's glow in the dark all by themselves. This glow is called Dark current. It is caused by the generation of electrons by the heat energy within the CCD itself. According to Janesick (2001) There is only one solution to removing dark current: cool the sensor.

Thermal noise can be reduced by a factor of 2x for each 6 degrees of cooling applied.

Dark current is generated at various locations within the CCD structure, accordingly there is dark noise from the gate surface, depletion zone, diffusion zone, backside and substrate of the CCD. Claims by some manufacturers that dark current is localised to one region are simply false.

The quallity of the silicon and manufacturing process used a CCD can have a great bearing on the overall dark noise of a CCD, and can vary within production runs from the same manufacturer.

Some dark current can be removed by employing a Multi-Pinned Phased (MPP) technology (Thompson, Kodak, SITe, Phillips etc) that moves unwanted charge within the CCD to the gate surface to reduce dark noise or by adding a charged surface (doping) layer to the pixel (SONY HAD sensors). While the doping method is very effective at removing unwanted electrons but has two difficulties. It reduces the surface area of the pixel (sometimes called fill factor) and can trap genuine signal.

A technique called "micro-lensing" which places a small collecting lens on the surface of each pixel, can be used to effectively enlarge the collecting area of a pixel and improve fill factors. SONY used this technque with their HAD sensors, and used even larger microlenses with their latest "super-HAD" sensors. It is also being used now by nearly all CCD manufacturers to improve the sensitivity (quantum efficiency) of their CCD's.
Linearity

The make useful star brightness or photometric measurements, the signal read from a CCD should increase in direct proportion to the amplitude of the original source. Some CCD's are fitted with charge bleed structures that will at some point destroy this linear relationship (eg interline devices and anti-blooming gates or ABG's). Non ABG frame transfer CCD's have excellent linearity in all resolution modes right up to their saturation levels (full capacities). Interline and ABG CCD sensors can be used for photometry, but the usable range of these detectors is reduced and care must be taken to ensure the device is not being used outside its linear range or binning modes have to be employed to achieve reasonable linearity. Full frame non ABG sensors typically have linearities within 1% with no safeguards other than to make sure the pixel has not saturated. Other devices are typically linear to around 3% with appropriate limited range or binning methods employed.

Pictured Left, SBIG CCD image of Eta Carina

CCD's have some rather nice properties for Astroimaging. They have excellent quantum efficiency, linear sensitivity, high signal noise ratios and digital output. In short, image data can be efficiently collected processed and displayed in a number of ways. Despite the high initial cost of ownership (though, prices are still falling) CCD's are very cheap to run, certainly much less expensive than film. CCD's themselves come in various forms such as interline transfer CCD's, linear CCD's and frame transfer CCD's. A standard black and white front illuminated array CCD is seen by many as the best choice for amateur Astronomical applications, as they are easy to produce hence are affordable and have high resolution. For photometric work chips should not be fitted with an anti-blooming gate (ABG). These cause a loss of linearity of the detector beyond approximately 50% full well capacity. ABG's do however give massive overexposure protection and are an excellent choice if astro-imaging or astrometry is the primary application. Back illuminated CCD's perform well toward the blue end of the spectrum, and should be considered for UV and tri-colour work.

When purchasing a CCD camera things to consider are: detector size, D/A resolution, signal to noise ratios, quantum efficiency, pixel size and cooling of the detector.

To effectively employ a CCD detector to take astronomical images that aesthetically rival (or exceed)  the quality film one needs to adopt a fairly terse regime of techniques and equipment. Telescope mountings need to be rigid and exceptionally smooth in their tracking abilities. Telescope optics need not be overly wonderful, though fast F-ratios do help. An often overlooked fact is that the physical size of the airy disc in any optical system is purely due to the F-ratio of that system. It is independent of aperture!  The downside of this equation is, to obtain any sort of image scale, long focal lengths are required. A standard Celestron 14 has a focal length of around 4 metres at its prime focus of F11. That is an airy disc that is 10 times larger that that delivered by the F1.1 focal ratio of the 4 metre aperture Anglo Australian Telescope.

It is true however, to say a Celestron 11 at  F6.3 still may still deliver star images that are larger than say a short focus 4 inch F7 refractor.


Image scale is the key, as our C-11 is looking at a much smaller angular patch of sky, where atmospheric turbulence, drive tracking errors, wind buffeting and a host of other factors conspire to smear incoming starlight over a much greater area on the CCD detector, despite the physically smaller airy disc in the F6.3 instrument. Scattering of light over the detector is also more pronounced as aperture increases. What was a tiny pinprick of light in a small aperture, is often revealed as a blazing sun in an instrument of significant aperture....hence the bad run that images from commercial SCT's often (unfairly) receive, as they are often being compared to images of vastly different scales, F-ratios and intensity levels.

If there is a single rule of CCD imaging it probably would be: Focus Accurately.

Many CCD images show poor attention to focus. Use a Hartmann mask or Diffraction Focuser if you are unsure. Linked with focus is tracking and seeing  which also smear light over area rather than a point. Short of launching our telescope into orbit, we can not do much about the seeing,  also there is little worth in attempting a high resolution (read: long focal length) image when the seeing is bad. Use a shorter focal length or wait for a better night.

Good tracking is intrinsic in the qualtiy and design of a telescope mount. Things to look for are: large gears, large thrust surfaces and ground worms and large numbers of worm wheel teeth. Worm gears should be bearing ( not spring) mounted. Declination tangent drives give more immediate and desirable drive reversal than a worm and wheel. Machined components are usually more rigid than cast components.  Spur gears often have large instantaneous tracking errors and should be avoided. Three manufacturers spring to mind. Losmandy, Astro-Physics and Byers. The heavier Takahashi mounts are also very good, but very expensive for their payload ratings. Commercial fork mounts are generally not as rigid as the former, they also do not allow bulky equipment (ie CCD's) to swing between the fork arms when used at high declinations hence should be avoided.

Glossary of CCD Imaging Terms.

Anti-blooming.
Should a pixel of a CCD saturate with too much light, its charge well will overflow into the adjacent pixel, which may also overflow etc. This gives a streak in the image (pictured at left). This effect may be reduced by the addition of a "anti-blooming gate" or ABG to the surface of the CCD (the gate is a thin metal film, which effectively bleeds off any excess charge). ABG's can give over 100x overexposure protection, however they typically reduce the sensitivity of the CCD by around 20-30%, cause a reduction in full well capacity and slight reduction in resolution of the array.. With applications such as photometry, non-ABG is the CCD detector of choice, as ABG chips also lose linearity as they start to exceed 50% full well capacity. For Astrometry and imaging ABG chips work well.
 
 
 


Astroimaging
Astrophotography by another name, with the subtle difference being the type of detector used to take a picture of the heavens, in this case a CCD.

Back Illuminated CCD's
Some CCD's are designed to be illuminated through their substrate layers. The substrate needs to be "thinned" which is a major cause for chip failure during manufacture, hence they are costly. Such "back illuminated" CCD's however have excellent quantum efficiencies, in the order of 90-97%%. and perform extremely well at the blue end of the spectrum.

 CCD
Charge coupled device. A an electronic light detector which uses the photo-electric effect (vis: light from a photon causing a an electron to be emitted from a metal substrate) to detect incoming packet or quanta of light.
 

D/A resolution
The Digital to Analogue resolution is a measure of how discreetly a signal can be measured. Most systems are 16 bit
that is 2 to the power 16 or approximately 65,000 discreet brightness levels can be recorded.

Diffraction Focuser
A Diffraction Focuser may be simply constructed from two pieces of timber or dowel which are placed in parallel over the aperture end of a telescope. Out of focus stellar images will reveal a pair of diffraction spikes, which will merge into a single spike when best focus is achieved.
 
 
 

 


Fast Guiding
A tip-tilt optical system which can be used to rapidly correct the position of a star, which may be in error due to atmospheric scintilation or telescope drive errors.

Fill Factor

The fractional area of a pixel that is sensitive to light.


 Frame transfer CCD's
A frame transfer CCD consists of two arrays, one of which is masked. Output from the unmasked array can be transfered rapidly to the masked area, which is then read at a slower rate. The ability to rapidly transfer data from the unmasked array allows for very short exposure times, which are ideal for planetary imaging.

Front Illuminated CCD
Most CCD's detect light through their top surface, where their anti-blooming gates and charge transfer surfaces are also located. Incoming photons are absorbed by these surfaces, which in turn do not generate a charge, hence there is a  loss in the efficiency of the detector, particularly at the blue end of the spectrum. Many manufacturers now use transparent gate structures that significantly improve the QE of CCD's.

Full Well Capacity
 The total number of electrons that can be stored in a pixel of a CCD before it becomes saturated. Chips such as the Kodak KAF series with 9 micron pixels can store approximately 85,000 electrons before becoming saturated or "full". Chips such as the SITe 512x 512 back illuminated series have larger 24 micron pixels and can absorb 350,000 electrons before saturating.

 

FWHM.

Short for Full Width Half Maximum. Plotting the intensity curve of a star, then measuring radius (n pixels) from the centre to where the intensity value that half that of the peak gives the FWHM value. This value is useful for determining the point spread radius star. This radius can be affected by a number of factors, but most commonly telescope focus and atmospheric turbulence.


F-Ratio
The focal length of an optical system, divided by its aperture. Hence a telescope of 200mm aperture with a 2000mm focal length with have a F10 focal ratio. "Fast" telescopes are around F4.5 or lower.

Hartmann mask
A device which may be used to accurately test the focus a telescope. In its simplest form it could be described as a lens cap with two symmetrical holes cut through it. The mask is placed over the aperture end of the telescope prior to focusing. If the telescope is not at best focus the airy disk of a star is broken into two components, which merge as best focus is achieved. Unfortunately, atmospheric turbulence can make this point of best focus shift around so it is best to assess the focus over a range of positions before committing to a deep CCD exposure (with the mask removed! )

 


Interline transfer chips (CCD's)

CCD chip manufacturers also make CCD's specialized for television applications. Standard NTSC or PAL video images are built up by displaying alternate field lines of an image in quick succession. Unfortunately such devices lack the resolution of an array CCD.


Linear CCD's
The pixels of a linear CCD are arranged in a single line, and are used in as fax machines, scanners and applications such as spectroscopy

Pixel
A picture element or pixel is the smallest part of a digital image or a single light detecting element of a CCD detector.

Detectors such as the Kodak KAF 0400 consist of an array of 765 by 512 pixels, each being 9 microns square. Detectors such as the Texas Instruments 245 chip have 23 x 27 micron pixels.

 By virtue of their large collecting area, larger pixels have higher sensitivities and full well capacities. Despite the improved sensitivity of  large individual pixels their sampling, hence detecting ability for point sources at a given focal length is not as good as smaller pixel arrays.

The images at left were taken with identical optical systems (C-11 at F6.3) and clearly show the superiority of small pixels when matched to this focal length in detecting faint stars. Extended sources of light such as nebulosity are more clearly revealed with the large pixel ST6 chip.

Current wisdom states that optical systems should be matched to give approimately 2 arc seconds per pixel.


The formula s= (205 x pixel width)/focal length (mm) will give the sky coverage of each pixel for a given focal length.

 

 

 


 
 
 
 
 
 

 


Quantum Efficiency
The quantum efficiency of a CCD detector (or QE) is a measure of how well it converts incoming light into a detectable signal. A QE of 100% means that all every single photon of light coming from an object is converted into a detectable charge.

Back illuminated detectors have the best QE to date with efficiencies close to 98%.

 Detectors such as the Kodak KAF "e" series have peak efficiencies around 65%, dropping to 20% at the blue end of the spectrum. Most manufacturers now employ micro-lenses over each pixel to further increase the sensitivity of their CCD's.  The Kodak KAF3200 series boasts peak QE's of around 87%

Seeing
Seeing is way of describing the amount of turbulence or distortion the earths atmosphere imparts to the image seen through a telescope. Good seeing usually allows a star's diffraction pattern to be clearly seen, whereas really bad seeing makes it almost impossible to focus the telescope.


Signal to noise ratio
A measure of the detection limit of a CCD system is how well it can separate starlight (signal) from skyglow, thermal electrons and other forms of noise. The S/N ratio can improved by cooling the CCD detector to remove thermal sources of noise. Most manufacturers use thermo-electric pieltier elements and or cooling fans to cool the CCD from 30 to 50 degrees Celsius below ambient.

Tracking
When performing CCD (or photographic) imaging a telescope needs to follow the apparent motion of a star across the night sky. This motion  (very apparent when using long focal length instruments) is 15 arc seconds per second of time and needs to be compensated for as accurately as possible. Minor drive motor rate variations,atmospheric refraction and flexure of components make guiding mandatory . Auto and Self guiding will reduce these errors, however, small  erratic errors due to: drive gear machining and indexing errors, bearing flat spots, grit or metal fragments embeded in worm gear lubricants and wind buffeting all cause a telescope drive to mis-track for brief periods. The earth's atmosphere also causes small, rapid variations in the position of a star, which more than swamp any errors due to an optical system's resolving power. At best one can expect around a 2-3 arc seeing disk within a CCD image (circle in diagram left) despite the optics (airy disk within the circle, left) of a telescope being easily capable of resolving well below this limit. It is for these reasons that a smooth, accurate, rigid, well designed telescope mount will help enormously in improving the resolving power of a CCD imaging system. Fast guiding technology also helps.

Tri-Colour Process.
By taking filtered red, green and blue exposures, and them combining them, one can create a full colour image of an object. Helmholtz is best known for first describing this "additive" colour process. For example, say we have a blue object. When viewed through red or green filters it will appear dark. A blue filter will however transmit most of the object's light through to the detector. Combining our exposures our object has, no red, no green, yet strong blue intensity or brightness levels. Hence our combined image of the object appears blue.

Back to homepage