Page 224 - Mechanical Engineers' Handbook (Volume 2)

3 Process Data Acquisition 213

scanning must be fast enough to capture faster events and determine the state of other
variables relative to those events. Similarly, if there are related analog data, the scan rate
may have to be fast enough to even characterize the curve of the analog data (remember
Fig. 1).
The capacity of the target system must also be taken into account. Storing large volumes
of data is becoming more feasible as systems increase in speed and power, but the retrieval
and organization of those data may become a time-consuming, overly complex task with too
much data or poorly organized data. Consequently, even though storage itself is less of an
issue, other factors impact how much data are retained and how organized for later retrieval.
Later sections examine approaches for organizing and retrieving data.

3.2 Accuracy and Precision of Data
Accuracy and precision are dependent on the sampling interval as well as the resolution of
the system (Ref. 2; Ref. 3, pp. 78–80; and Chapter 1 in this volume). When dealing with
the A/D conversion process, the step size or number of bits used is critical when determining
the system precision and accuracy (Ref. 1, pp. 78–81). Figure 2 illustrates the difference
between accuracy and precision of data. Table 1 illustrates the effect the number of bits has
on the precision.
This also interacts with range, which will be discussed later, since having an accurate
number over a small percent of the desired range would not allow the ability to fully char-
acterize the process. For example, highly accurate readings with 1% moisture resolution over
a range from 10 to 20% moisture content would be inadequate if one were attempting to
measure moisture over a 5–40% range. When selecting transducers, it is necessary that they
have both the accuracy and range needed for the process being observed. When selecting
converters, one should be aware of the settling time (governs how often readings can be
obtained), resolution of the converter (affects range and detail of measurements), and accu-
racy of sensor.
Chapter 1 includes some characteristics of transducers, including calibration, and the
sampling of data.
One should be aware that an event that has been stabilized in a data collection system
may be offset in time, resulting in a potential discrepancy between events or values when
values are compared from different sources or from multiplexed data. It should be verified

Precise but
inaccurate
measurement

Imprecise but
accurate
Actual signal
measurement
(accurate and Figure 2 Difference between accuracy and pre-
precise) cision.