Specifications

137Calibration and Automated Calibration
Such environmental factors include
temperature, humidity, vibration,
contamination, and component aging.
These factors are accounted for with a
combination of self-test at power-up
and periodic or continuous testing.
The eld testing can be as simple as
sensing temperature and compensating
accordingly, or it can be more complex.
A simple example of power-up self-
testing is to automatically briey short
the inputs of an amplier together to
set a zero reading point (Figure4).
Doing so allows any changes to input
oset voltage or to downstream circuit
parameters to be calibrated out. Another
example is to electronically swap the
resistive temperature sensor with a
precision xed resistor to enable the
instrument to calibrate the temperature
reading to the expected value. Using two
dierent precision resistors can establish
a line that provides both gain and oset
information. More complex schemes
can be used to adjust for nonlinearities.
Continuous Monitoring and
Readjustment
In some applications, waiting for
periodic calibration at power-up would
occur only very rarely after system
maintenance which can be too costly
if system performance is suering
or safety margins are compromised
from an out of calibration component.
Depending on the impact of a system
not being calibrated, these applications
may need to use continuous monitoring
with subsequent readjustment.
Good examples of applications that
requires continuous monitoring and
calibration are a variety of safety
systems in nuclear power plants.
Continuous calibration consists of
circuitry that self-corrects continuously
or very frequently. This can be
accomplished in a variety of ways either
with techniques similar to periodic
self-testing, just done more frequently,
or with other techniques that allow the
system to continue to operate. In the
former case, very brief interruptions
to the normal signal path may be
made, including making connections
to simulate zero scale and full scale
readings for example. Another use
of these interruptions would be, for
example, to cut a signal path gain in
half and check that the response is
indeed exactly half. If not, an oset
error is indicated and can be corrected
for. In the latter case, where full system
operation needs to be maintained,
out-of-band or noise level techniques
can be used by injecting signals either
above or below the normal signal
frequency range, or signals so small
that they fall within the noise oor of
the system. With proper design, these
signals are detectable by a variety of
methods. These can be used to stimulate
the test and calibration protocol while
standard signal processing continues.
The techniques used are limited only
by the creativity of the engineers. If a
system, during a readjustment, detects
that no further adjustment is possible,
then an alarm condition must be set.
The ability to adjust analog outputs
using digital technology has greatly
enhanced the ability to continuously
monitor and adjust. Digital technology
provides low-cost and nearly error-
free communications for remote
monitoring and subsequent control.
Digital control of analog circuitry using
precision DACs and digital pots allows
economical remote-control processes,
while also ensuring the precision
needed to meet specications.
Circuitry for Electronic
and Automated
Calibration
Electronic calibration is based on
digitally controlled calibration devices:
DACs with voltage or current outputs
can be used to provide temporary inputs
to analog signal chains or to adjust
bias levels. Digital pots with variable
resistances or variable resistance ratios
can provide gain and oset adjustments,
analog switches can select dierent
gain or lter corner setting components,
and potentially any other digital-to-
analog transducer such as a digitally
controlled light source can be used to
stimulate a self-calibration process. All
of these replace mechanical calibration
procedures in factory settings and
within the equipment itself. The digital
approach provides a range of benets:
better reliability, improved employee
safety, increased dependability, and
reduced product liability expense.
In addition, digitally controlled
calibration can be fully automated,
which results in reduced test time and
expense by removing human error.
Solid state solutions such as digital
pots as opposed to mechanical pots
are not susceptible to mechanical
shock and vibration, which can cause
loss of calibration settings and, in the
case of mechanical pots, can cause
momentary wiper contact bounce
which will likely lead to unpredictable
and potentially dangerous behavior.
Analog switches have improved to
the point that their on-resistance is
low enough that they can be used in
high-precision gain setting circuits to
provide a range of precision xed-gain
choices. This capability, combined
with a digital pot for ne adjustments
within a gain range, can provide an
extremely precise calibration capability.
Figure 4. A digital multimeter showing good calibration of the
zero signal level, but is the gain calibrated? This is difficult to
discern without a reference standard to read periodically. Or,
maybe during power-up it reads a precision internal value while
the display is blanked to check for proper gain calibration?