u2402b ETC-unknow, u2402b Datasheet - Page 8

• Manufacturer Part Number: u2402b
• Description: Fast Charge Controller Nicd/nimh Batteries
• Manufacturer: ETC-unknow

Analog-Digital-Converter (ADC),
Test Sequence
A special analog-digital-converter consists of a five-bit
coarse and a five-bit fine converter . It operates by a linear
count method which can digitalize a battery voltage of
4 V at Pin 10 in 6.5 mV steps of sensitivity.
In a duty cycle, T, of 20.48 s, the converter executes the
measurement from a standard oscillation frequency of
f
charge break time of 2.56 s (see figure 6), i.e., no-load
voltage (or currentless phase). Therefore it has optimum
measurement accuracy because all interferences are
cut-off during this period (e.g., terminal resistances or
dynamic load current fluctuations).
After a delay of 1.28 s the actual measurement phase of
1.28 s follows. During this idle interval of cut-off
conditions, battery voltage is stabilized and hence
measurement is possible.
An output pulse of 10 ms appears at Pin 9 during charge
break after a delay of 40 ms. The output signal can be used
in a variety of way, e.g., synchronising the test control
(reference measurement).
U2402B
8 (17)
osc
= 800 Hz. The voltage measurement is during the
conversion
(internal)
ADC
time
output
break
charge
Status
1.28 s 1.28 s
Charge break
Figure 6. Operating sequence of voltage measurements
10 ms
2.56 s
40 ms
T= 20.48 s
Charge
Plausibility for Charge Break
There are two criterian considered for charge break
plausibility:
–
When the signal at Pin 10 of the DA converter is 12 mV
below the actual value, the comparator identifies it as a
voltage drop of – dV. The validity of – dV cutt-off is
considered only if the actual value is below 12 mV for
three consective cycles of measurement.
d
A four bit forward/ backward counter is used to register
the slope change (d
clocked by each tracking phase of the fine AD-counter.
Beginning from its initial value, the counter counts the
first eight cycles in forward direction and the next eight
cycles in reverse direction. At the end of 16 cycles, the
actual value is compared with the initial value. If there is
a difference of more than two LSB-bit (13.5 mV) from the
actual counter value, then there is an identification of
slope change which leads to normal charge cut-off. A
second counter in the same configuration is operating in
parallel with eight clock cycles delay, to reduce the total
cut-off delay, from 16 test cycles to eight test cycles.
2
D
V/dt
V Cut-Off
2
Cut-Off
2
V/dt
TELEFUNKEN Semiconductors
2
, V
Batt
– slope). This counter is
t
t
t
94 8693
Rev. A3, 14-Nov-96