Page 225 - Mechanical Engineers' Handbook (Volume 2)

214 Data Acquisition and Display Systems

Table 1 Relationship between Number of Bits and
Precision
Resolution
Number on 5-V Percent
of Bits Steps Measurement Resolution
8 256 0.01950 0.3900
10 1024 0.00488 0.0980
12 4096 0.00122 0.0240
16 65,536 0.00008 0.0015

that the transducer is collecting the data fast enough to allow one to have relevant times of
collection in the data acquisition system. Also one should ensure that the potential relation-
ship between events from different sources and their intended use is understood when con-
sidering the speed and accuracy of transducers.

3.3 Time-Based versus Event-Driven Collection

There are two major approaches when collecting data with a general-purpose data acquisition
system. In one approach data are collected on a regular frequency based on time, such as
once per second. This is easy to institute and it is relatively easy to analyze the data and
their relationships after the fact. This approach tends to require more data storage and can
make it difficult to identify events or the interactions with manufacturing attribute data. The
other approach is event-based acquisition. An event is identified, such as when a package is
dropped onto a platform, the time of that event is recorded, and the values of related variables
are collected for that time. The sampling rate of the transducers to acquire the other variables
may be important, as their values may become irrelevant if too long a time interval has
passed after the event has occurred when the related variables are sampled. Batch processes,
such as mixing a tankful of chemicals, often have some data collected only at the start and
end of the process. Other data may be recorded at fixed time intervals during the batch
process. Depending on the needs, the data during the actual reaction may be of great or of
little use. The time between sampled events may be several minutes, hours, or even days in
length, but the time of the event may be critical, as well as detailed data at the time of the
event, resulting in a common tactic of using high-speed scanning to detect the occurrence
of an infrequent event. Approaches for combining and analyzing data will be covered in a
later section.

4 DATA CONDITIONING

Often the data obtained from a process are not in the form or units desired. This section
describes several methods of transforming data to produce proper units, reduce storage quan-
tity, and reduce noise.
There are many reasons why process measurements might need to be transformed in
order to be useful. Usually the signals obtained will be values whose units (e.g., voltage,
current) are other than the desired units (e.g., temperature, pressure). For example, the mea-
surement from a pressure transducer may be in the range 4–20 mA. To use this as a pressure
measurement in pounds per square inch (PSI), one would need to convert it using some
equation. As environmental conditions change, the performance characteristics of many sen-