Page 15 - Instrumentation Reference Book 3E
P. 15
Introduction xv
If we assume that the instrument output is 3 Environment
exactly proportional to a quantity, and we find
discrepancies, this is called “non-linearity error.” Instrument engineers must select their devices
Non-linearity error is the maximum departure of based on the environment where they will be
the true input/output curve from the idealized installed. In plants there will be extremes of tem-
straight line approximating it. perature, vibration, dust, chemicals, and abuse.
It may be noted that this does not cover Instruments for use in plants are very different
changes in “incremental gain,” the term used from those that are designed for laboratory use.
for the local slope of the input/output curve. Two kinds of ill effects arise from badly
Special cases of the accuracy of conversion selected instruments: false readings from excep-
from digital to analog signals and vice versa tional values of influence quantities, and the irre-
are discussed in Sections 29.3.1 and 29.4.5 of versible failure of the instrument itself.
Part 4. Calibration at sufficient intermediate Sometimes manufacturers specify limits to
points in the range of an instrument can cover working conditions. Sometimes instrument engin-
systematic non-linearity. eers must make their own judgments. When
Microprocessor-based instrumentation has working close to the limits of the working condi-
reduced the problem of systematic non-linearity tions of the equipment, a wise engineer derates
to a simple issue. Most modern instruments have the performance of the system or designs environ-
the internal processing capability to do at least a mental mitigation.
multipoirit breakpoint linearization. Many can Because instrumentation engineering is a prac-
even host and process complex linearization tical discipline, a key feature of any system design
equations of third order or higher. must be the reliability of the equipment. Reliabil-
Special terms used in the discussion above are ity is the likelihood of the instrument: or the
defined ir, BS 5233; several ANSI standards, and system, continuing to work satisfactorily over
in the ISA Dictionary of Instrumentation, along long periods. We will discass reliability deeply in
with numerous others. Part 4. It must always be taken into account when
The general approach to errors that we have selecting instruments and designing systems for
outlined follows a statistical approach to a static any application.
situation.
Communications theory emphasizes working
frequencies and time available, and this approach 4 Units
to error is gaining importance in instrumentation
technology as instruments become more intelli- The introductory chapters to some books have
gent. Sensors connected to digital electronics discussed the theme of what system of units is
have little or no error from electronic noise, but used there. Fortunately the question is becoming
most accurate results can still be expected from obsolete because SI units are adopted nearly
longer measurement times. everywhere, and certainly in this book. In the
Instrument engineers must be very wary of United States, and a few other areas, where other
measuring the wrong thing! Even a highly accur- units still have some usage, we have listed the
ate measurement of the wrong quantity may relations for the benefit of those who are still
cause serious process upsets. Significantly for more at home with the older expressions.
instruments used for control, Heisenberg’s Law
applies on the macro level as well as on the sub- 5 References
atomic. The operation of measurement can often
disturb the quantity measured. British Standards Institution. Glossary of terms used in
This can happen in most fields: a flowmeter can Metrology, BS 5233 (1975)
obstruct flow and reduce the velocity to be meas- Dietrich, D. F., Uncertainty, Calibration and Probabil-
ured, an ‘over-large temperature seiisor can cool ity: the Statistics of Scientific and Industrial Measure-
the material studied, or a low-impedance volt- ment, Adam Hilger, London (1973)
meter can reduce the potential it is monitoring. Instrumentation, Systems and Automation Society
Part of th’e instrument engineer’s task is to foresee (ISA) The ISA Comprehensive Dictionary of Measure-
ment and ControMrd edition. online edition
and avoid errors resulting from the effect his http:llwww. isa. org
instrument has on the system it is being used to Topping. J., Errors of Observation and their Treatment.
study. Chapman and Hall. London (1972)