![](http://datasheet.mmic.net.cn/230000/LM9811CCVF_datasheet_15593399/LM9811CCVF_31.png)
Applications Information
(Continued)
Since this is a CCD sensor, the Signal Polarity is set to a 1
(inverted) to match the CCD’s output signal. The number of
dummy pixels and optical black reference pixels are given in
the CCD’s datasheet. The dummy pixel register should be
programmed with the number of dummy pixels in the CCD +
1 (for example, if the CCD has 16 dummy pixels then the
register should contain 17). The optical black reference reg-
ister should be programmed with the number of optical black
pixels in the CCD.
The PGA gain coefficient register and PGA Gain Source bit
are used during calibration (see Section 5.0). The Power
Down bit should be set to 0 for normal operation. The Offset
Add bit is also programmed during calibration.
The VGA and Offset DAC bits are programmed during cali-
bration (Section 5.0). The Test Mode bits should always be
set to “0”.
9.0 TYPICAL CIS APPLICATION
Many CIS sensors (such as those made by Dyna Image Cor-
poration) have only one clock input, a transfer signal, and an
output signal that is referred to ground (Figure 37). Figure 38
shows the analog and digital circuitry required to connect a
typical Dyna CIS to the LM9811.
Because the CIS requires only one clock with a duty cycle of
less than 50%, the LM9811’s RS output is used as the CIS’s
CLK source.
φ
1 and
φ
2 are not used. The 74HC74 D flip-flop
is used to lengthen the transfer pulse (SI, or “Shift In” on the
CIS) so that it overlaps the first RS pulse and meets the tim-
ing requirement of the CIS (see Figure 39).
The final “trick” required to interface a CIS to the LM9811 is
the generation of optical black pixels for the LM9811 to
clamp to at the beginning of a line. Unlike CCDs, CIS de-
vices do not have a sequence of optical black pixels at the
beginning of a line— the first pixel out of a CIS is valid image
data. There are several ways to create black pixels for the
LM9811 to clamp to.
The simplest solution is to physically place a light shield
(black plastic, tape or metal) over the first 10 or so pixels.
This reduces the voltage output of the CIS to nearly 0V,
which is adequate for the LM9811 to clamp to. This has the
side effect of slightly reducing the number of active pixels
available for image capture.
A second option is to artifically generate “black” pixels by
holding the CLOCK input high for 10 or so RS pulses (Figure
40). This forces the output voltage to zero for the time that
the CLOCK input is high, and only one active image pixel is
lost. The BLACK signal could be generated by the ASIC/
external logic that generates a pulse on the first rising edge
of RS after the TR pulse.
Suggested timing for CIS devices is:
Mode = 0 (Standard Mode)
*
RS Pulse Width = 0 (2 MCLKs)
DS012813-45
FIGURE 37. CIS Waveforms
DS012813-46
FIGURE 38. Minimum CIS Interface
DS012813-47
FIGURE 39. CIS Interface Digital Timing
DS012813-48
FIGURE 40. Generating Artificial Black Pixels
www.national.com
31