ICM7226A_01 Datasheet, PDF (12/19 Pages) Intersil Corporation – 8-Digit, Multi-Function, Frequency Counter/Timer

ICM7226A, ICM7226B
Display Considerations
The display is multiplexed at a 500Hz rate with a digit time of
244µs. An interdigit blanking time of 6µs is used to prevent
display ghosting (faint display of data from previous digit
superimposed on the next digit). Leading zero blanking is
provided, which blanks the left hand zeroes after decimal
point or any non zero digits. Digits to the right of the decimal
point are always displayed. The leading zero blanking will be
disabled when the Main Counter overflows.
The lCM7226A is designed to drive common anode LED dis-
plays at peak current of 25mA/segment, using displays with
V_F = 1.8V at 25mA. The average DC current will be greater
than 3mA under these conditions. The lCM7226B is designed
to drive common cathode displays at peak current of
15mA/segment using displays with V_F = 1.8V at 15mA. Resis-
tors can be added in series with the segment drivers to limit
the display current, if required. The Typical Performance
Curves show the digit and segment currents as a function of
output voltage for common anode and common cathode
drivers.
To increase the light output from the displays, V_DD may be
increased to 6.0V. However, care should be taken to see that
maximum power and current ratings are not exceeded.
The SEGment and Digit outputs in both the ICM7226A and
ICM7226B are not directly compatible with either TTL or
CMOS logic. Therefore, level shifting with discrete transis-
tors may be required to use these outputs as logic signals.
External latching should be down on the leading edge of the
digit signal.
Accuracy
In a Universal Counter, crystal drift and quantization errors
cause errors. In frequency, period and time interval
modes, a signal derived from the oscillator is used in either
the Reference Counter or Main Counter, and in these
modes, an error in the oscillator frequency will cause an
identical error in the measurement. For instance, an oscilla-
tor temperature coefficient of 20ppm/^oC will cause a mea-
surement error of 20ppm/^oC.
In addition, there is a quantization error inherent in any digi-
tal measurement of ±1 count. Clearly this error is reduced by
displaying more digits. In the frequency mode maximum
accuracy is obtained with high frequency inputs and in
period mode maximum accuracy is obtained with low fre-
quency inputs. As can be seen in Figure 16. In time interval
measurements there can be an error of 1 count per interval.
As a result there is the same inherent accuracy in all ranges
as shown in Figure 17. In frequency ratio measurement
can be more accurately obtained by averaging over more
cycles of INPUT B as shown in Figure 18.
0
FREQUENCY MEASURE
0.01s
2 0.1s
1s
10s
4
1 CYCLE
10 CYCLES
10² CYCLES
6 ^{103 CYCLES}
PERIOD MEASURE
f_OSC = 10MHz
8
1
10
10³
10⁵
10⁷
FREQUENCY (Hz)
FIGURE 16. MAXIMUM ACCURACY OF FREQUENCY AND
PERIOD MEASUREMENTS DUE TO LIMITATIONS
OF QUANTIZATION ERRORS
0
1
2
MAXIMUM TIME INTERVAL
3
FOR 10³ INTERVALS
4
MAXIMUM TIME
INTERVAL FOR
5
10² INTERVALS
6
MAXIMUM TIME INTERVAL
7
FOR 10 INTERVALS
8
1
10
10² 10³ 10⁴ 10⁵ 10⁶ 10⁷ 10⁸
TIME INTERVAL (µs)
FIGURE 17. MAXIMUM ACCURACY OF TIME INTERVAL
MEASUREMENT DUE TO LIMITATIONS OF
QUANTIZATION ERRORS
0
1
RANGE
2
3
1 CYCLE
10 CYCLES
4
10² CYCLES
10³ CYCLES
5
6
7
8
1
10 10² 10³ 10⁴ 10⁵ 10⁶ 10⁷ 10⁸
f_A/f_B
FIGURE 18. MAXIMUM ACCURACY FOR FREQUENCY RATIO MEASUREMENT DUE TO LIMITATION OF QUANTIZATION ERRORS
12