Notes:
1. TINT is the line-scan rate or integration time. It is determined by time interval between
two start pulses, SP. Hence, if SP is generated from a clock count down circuit, it will be
directly proportional to clock frequency. And it will be synchronous with the clock
frequency. The longest integration time is determined by the degree of leakage current
degradation that can be tolerated by the system. A 10ms maximum is a typical rule-of-
thumb. An experienced CIS user can use his discretion and determine the desired
tolerance level for the given system.
2. FREQ is clock frequency, also equal to the signal data rate. It is generally fixed for
many applications for following reasons: One is the exposure time. With a give light
power, the exposure time of the sensor depends on integration time and in most
applications it uses clock count down circuits to generate the SP, shift register start
pulse. The second is the shape of the video output pulse. Because of the output is in
pulse packets of video charges, the signal are processed on the output video line of the
sensors. The signal shape depends greatly upon the amplifier configurations. Please
refer to the referenced PI3039 Data Sheet. It has some brief outline application notes.
Under Note 6 in page 6 there is a discussion on video pulse shapes. On page 8, 9 and
10 there are discussions on the three types of signal output stages.
3. The RESPONSIVITY is the ratio of video signal in volts divided by the unit exposure
(Volts/micro-Joules/cm 2). This exposure was measured with the output level adjusted
to 1.27V. The spread of the measured exposure R-RSP, G-RSP and B-RSP can be
used to compute the user’s desired signal voltage level.
Since this is prototype module, all of the following notes 4, 5, 6, 7 and 8, are on data, based on
engineering scope measurements, the data will be re-measured during pilot production using all
the standard QA practices under the control of ISO 9000 regulations. Furthermore, they will be
taken on fully computerized test systems. If required, these prototype data may be revised.
4. Vpmax = maximum pixel value of Vp(n); Vpmin = minimum pixel value of Vp(n); Vpavg =
∑ Vp(n)/7680; where Vp(n) is the nth pixel in a line scan with the module scanning a
uniform white target. Vp values are measured with a uniform exposure.
5. BRIGHT OUTPUT NONUNIFORMITY: Up(+) = [(Vpmax - Vpavg) / Vpavg] x 100% or
Up(-)= [(Vpavg - Vpmin) / Vpavg] x 100%, whichever polarity with the highest value is
selected. Two further notes, one is that the PI616MC-AS has no requirement for an
optical system, or a light system. The second is that the non-uniformity is dominated by
the LED light bar’s non-uniformity so only the sensor non-uniformity is specified. The
normal standard CIS modules are enclosed in a self-contained module with the
complete optical and LED lighting system. So the light system, usually LED bar, is
included in making the measurement of the optical characteristics. This fixes the optical
geometry for the module and the light source. So the module’s optical characteristics
are simply measured with the module placed on a uniform reflecting target with a know
reflection density. However, the PI616MC-AS is not enclosed with its optical and light
source system. So Up is measured with uniform light source that directly illuminates the
image sensors’ photosite.
6. ADJACENT PIXEL NONUNIFORMITY: Upadj = MAX[ | (Vp(n) - Vp(n+l) | / Vp(n)] x
100%. Upadj is non-uniformity in percentage of Vpavg. It is the maximum difference
amplitude between two neighboring pixels.
7. BRIGHT OUTPUT TOTAL NONUNIFORMITY: Uptotal = [Vpmax -Vpmin]/Vpavg
8. DARK NONUNIFORMITY: Ud = Vdmax – Vdmin: It is measured over the full length of
the array with the light source off and the sensors are placed in the dark. Vdmax is the
maximum pixel value of the video pixel with the exposure off. Vdmin is the minimum
PAGE 6 OF 22 - PI616MC-AS, 12/20/02
AMI [ AMI SEMICONDUCTOR ]
