This discusses the technique for acquiring various characteristics of an unknown signal using the DT-2805 board. The DT-2805 board is used to set the period between data point acquisition and the number of data points collected.
A program (not available here) was written to control then rate of data acquisition, and to calculate the following signal characteristics: mean, variance, standard deviation, skewness, and kurtosis. The program also displays a scatter plot of the signal. The values at a specific data point can be obtained from the scatter plot by moving a cursor box along the waveform. The probability density function of the signal is displayed using a histogram. The frequency of a given range can be obtained by choosing a rectangle of the histogram.
The following is a list of equipment used in the experiment.
See
Figure 1 for a suggested experimental
setup.
(1) IBM PS/2 Model 30 computer with an installed HP-IB interface.
(2) Data Translation Model DT 2805 Data Acquisition Board installed in the computer.
(3) Data Translation Model DT 707T screw terminal connection panel.
(4) HP 8904a Multifunction Synthesizer
(5) Tektronix 2235 analog oscilloscope
(6) HP-IB interface cable and coaxial signal cables.
(7) QuickBASIC 4.5 program.
The definition for the various characteristics of a waveform are as follows:
(1) Mean: the arithmetic average of the sampled data.
(2) Variance: the expected value of its squared deviations from the mean.
(3) Standard Deviation: the measure of the precision in a sample mean. Note that the standard variation is the square root of the variance.
(4) Skewness: a shift to the right or left in a frequency distribution.
(5) Kurtosis: a measure of a curve describing the statistical frequency distribution in the region about its mode.
The above terms are also known as orders of moments with the mean as the first moment, the variance as the second moment, the skewness as the third moment, and the kurtosis as the fourth moment.
This has been omitted.
This program has three menus (file input and output, wave acquisition, statistical) which are user friendly. There are two graphic screens which allow the user to obtain values for a specific data point or data group.
The program was tested using several different wave forms. Shown in Figure#2 is a sample from the CRT display, of a noise wave generated from a HP8904A with an amplitude of 2 volts. One thousand points of data are shown in this plot. The clock setting for the DT-2805 board was about 1400 Hz . The gain of the board was set to one.
A scatter plot of data captured from the noise waveform is shown
in
Figure #3. The period and voltage
of each data point acquired for the wave can be save and retrieved
from the file menu. The lowest possible graphic screen was used
for the scatter plot to allow this program to be compatible with
the lower resolution CGA screens. The values at a specific data
point can be obtained from the scatter plot by moving a cursor
box along the waveform, using the left or right arrow keys and
the Ctrl-left or Ctrl-right arrow keys.
Furthermore, the frequency distribution for this noise wave is
shown in
Figure #4. The histogram's
distribution nearly resembles a "Bell Shaped Curve"
which is what was expected. The number of bins selected was 25.
The program will allow the number of bins to range from 2 to 100.
In addition, frequency data can also be written to a file if one
desires to plot the histogram with another software package. The
lowest possible graphic screen was used for the histogram to allow
this program to be compatible with the lower resolution CGA screens.
The frequency of a given range can be obtained by choosing a rectangle
of the histogram via the left and right arrow keys. The needed
information will be displayed at the bottom of the screen.
Using this program provides a quick and easy way to acquire statistical information about waveforms. The program was written to calculate the following signal characteristics: mean, variance, standard deviation, skewness, kurtosis, and to displayed a scatter plot and histogram. The histogram's distribution shown in Figure#4 nearly resembles a "Bell Shaped Curve" which is what was expected for the wave sampled in this experiment.