THE KAF-0400 CCD

Several families of CCDs are on this photography. The KAF-0400 is in the foreground.

The dimension of the sensitive surface of the KAF-0400 is 6.9 X 4.6 millimeters: it is much less than the surface of a photographic film. However, as the CCD is very sensitive, the field is smaller but the depth of the image is bigger and you can thus see very weak objects.

The technology MPP (Multi-Pined Phase) is a significant characteristic of the KAF-0400: that makes it possible to obtain an extremely low dark current. It is important for a low-cost camera: even if cooling is weak, you can have a good performance in deep-sky images with exposure times of several minutes.

Finally the KAF-0400 has very good cosmetic and optoelectronic qualities for a component costing approximately 2,000 to 4,000 FF including all taxes (depending on the level of quality).

The main characteristics of the KAF-0400 are:

• Number of pixels: 768 X 512
• Dimensions of the pixel: 9 X 9 microns
• Dimensions of the sensitive surface: 6.9 X 4.6 mm
• Fill factor of the pixel: 100 %
• Dark current: less than 10 pA/cm2 (MPP technology)
• Sensitivity of the output stage: 10 µV/e-
• Average quantum efficiency: 35%
• Load capacity: 80,000 electrons
• Type of clock: 2 phases
• Number of clocks: 5

If you multiply the number of pixels by their dimensions, you get the sensitive surface. It is approximately 32 mm2. That is sufficient to quickly point the telescope without too many problems. But that is a little small to take part effectively in survey-type programs (to discover comets, novae, supernovas...). The sensitive surface of a 35 mm film is 864 mm2 (27 times larger).

With the KAF-0400, all the surface of the pixel remains photosensitive. The fill factor is 100%. There is a version of the KAF-0400 with anti-blooming. That makes it possible to limit the trails of stars sufficiently bright to saturate the detector. But with anti-blooming, the sensitive surface of the pixel and the sensitivity are weaker by 30%. We recommend you use the version without anti-blooming in order to optimize the sensitivity of the detector.

The KAF-0400 uses the MPP technology so that the dark current is the weakest possible. According to Kodak, 1,000 seconds at 25°C are necessary to get saturation because of the thermal signal. This result shows that the KAF-0400 is usable almost without cooling for astronomical time exposures of 2 or 3 minutes at least. In practice, this performance is not obtained because there are pixels with a thermal signal largely higher than the average value. After a few tens of seconds of integration at ambient temperature, an image acquired in the darkness reveals brilliant points due to these pixels. These pixels are called hot-pixels. The most intense remain visible even if you cool the detector to -10°C with an exposure time of 2 or 3 minutes. But these pixels are always on the same locations in the image and you can thus easily remove them using a preprocessing. However, these pixels quickly become very numerous if the temperature of the CCD is 25°C and if the exposure time is higher than one minute. Under these conditions, the mathematical processing of the image reaches its limits and that can leave residues. In fact, the maximum exposure time at ambient temperature is one minute approximately. That is sufficient for applications such as autoguiding or planetary imagery. At -10°C, the KAF-0400 produces an average dark current of 0.1 electron/pixel/second. After 2 minutes, only 12 thermal electrons on average. It is a very weak signal compared to the signal coming from the sky background in most cases. The cooling system integrated in the Audine camera makes it possible to cool the CCD to -15°C when the ambient temperature is 20°C.

The sensitivity of the output stage of the KAF-0400 is very high (conversion rate of the electrons into volts). This performance makes it possible to install a low-gain amplifier at the output of the CCD. This amplifier will produce a very low noise and it will have a broad band for applications where the CCD must be read very quickly.

The quantum efficiency is normal for a thick CCD, but it is lower compared to much more expensive CCDs (thinned CCDs).

The clocks have two phases and there are 5 clocks: two clocks for the transfer of the charges of the image area into the horizontal register, two clocks to control the horizontal register and a clock to reset the output stage capacity on each arrival of a new pixel. The two clocks of the horizontal register are symmetrical: it is thus only necessary to generate only one command.

The KAF-0400 has some problems:

• Dynamics is relatively weak. Since a charge package contains only 80,000 electrons before saturation and since the reading noise is approximately 18 electrons, effective dynamics is 80,000/18 = 4400. It is relatively weak for the astronomical observation. Audine makes analog/digital conversion to 15 bits but Audine could operate with a 13-bit converter for its dynamics. The 2 overhead bits of the electronic part make the quantification noise of minor significance.

•  It is not possible to make an electronic shutter because there is no memory area for the intermediate storage of the image. And there are no specific electrodes for this function in the CCD structure. A mechanical shutter is thus necessary at the time of reading. You can always make a manual shutter and mask the opening of the telescope with an opaque object. For a much higher budget, you can choose the electromechanical shutter. This electromechanical shutter is placed before the CCD surface. An intermediate solution consists in using a mechanical iris shutter controlled by a cable release. Finally, for particular applications (such as drift-scanning or planetary observations in half-frame mode), the shutter is not necessary. These techniques of reading of the image are implemented in the software provided with the Audine camera. If you make deep-sky images, there is an algorithm to process images which makes it possible to largely reduce the smearing effect (trails left by the objects which continue to illuminate the CCD when it is in reading mode).

 
• The price of the component is relatively high if you choose a "high quality" version. This is an important point and we will now analyze it.

An electromechanical shutter (Uniblitz here) is a relatively expensive mechanism. Installing this equipment in Audine would have much increased its cost (approximately by 1,000 F, without the mechanical and electronic part for control purposes). There are simpler alternative solutions.

The KAF-0400 is available according to several quality grades. Class 0 means a component without defects. Class 2 means worse quality. Of course, the price increases if the number of defects falls.

To define the class of a CCD, the manufacturer makes a sorting under conditions of use which are not the conditions of astronomy. For example, the tests are not made at low temperature. Certain defects disappear in such conditions. Sometimes, new users of CCD cameras complain because there are 3 or 4 defective pixels in the image. That is not significant: a KAF-0400 includes approximately 400,000 pixels and that affects only 0.001% of the sensitive surface. A defective column in an awkward location, i.e. in the centre of the image, is a bit of a problem but you can easily mask this defect by processing the image. To sum up, class 0 is not really justified. Class 1 is largely sufficient. It is not very risky to use class 2 in an astronomical camera (note: whatever the class, the number of hot pixels will be roughly the same). In conclusion: buy class 2 for approximately 2,500 F including all taxes.