11-07-2012, 12:00 PM
Calibrating HART Transmitters
Calibrating HART Transmitters.pdf (Size: 40.41 KB / Downloads: 47)
Abstract
In order to take advantage of the digital capabilities of HART transmitters, especially for
reporting process data values, it is essential that they be calibrated correctly. This paper outlines
the differences between calibrating a conventional and a HART transmitter, and gives
recommendations for calibration practices.
Introduction
This paper assumes that you want to make use of the digital process values that are available from
a HART transmitter. However, before examining calibration requirements, we must first establish
a basic understanding in two areas: calibration concepts, and the operation of a HART
instrument.
Calibration Concepts
The ISA Instrument Calibration Series[6] defines calibration as "Determination of the
experimental relationship between the quantity being measured and the output of the device which
measures it; where the quantity measured is obtained through a recognized standard of
measurement." There are two fundamental operations involved in calibrating any instrument:
· Testing the instrument to determine its performance,
· Adjusting the instrument to perform within specification.
Testing the instrument requires collecting sufficient data to calculate the instrument's operating
errors. This is typically accomplished by performing a multiple point test procedure that includes
the following steps.
1. Using a process variable simulator that matches the input type of the instrument, set a known
input to the instrument.
2. Using an accurate calibrator, read the actual (or reference) value of this input.
3. Read the instrument's interpretation of the value by using an accurate calibrator to measure
the instrument output.
By repeating this process for a series of different input values, you can collect sufficient data to
determine the instrument's accuracy. Depending upon the intended calibration goals and the error
calculations desired, the test procedure may require from 5 to 21 input points.
The first test that is conducted on an instrument before any adjustments are made is called the As-
Found test. If the accuracy calculations from the As-Found data are not within the specifications
for the instrument, then it must be adjusted.
Adjustment is the process of manipulating some part of the instrument so that its input to output
relationship is within specification. For conventional instruments, this may be zero and span
screws. For HART instruments, this normally requires the use of a communicator to convey
specific information to the instrument. Often you will see the term calibrate used as a synonym
for adjust.
After adjusting the instrument, a second multiple point test is required to characterize the
instrument and verify that it is within specification over the defined operating range. This is called
the As-Left test.
It is absolutely essential that the accuracy of the calibration equipment be matched to the
instrument being calibrated. Years ago, a safe rule of thumb stated that the calibrator should be
an order of magnitude (10 times) more accurate than the instrument being calibrated. As the
accuracy of field instruments increased, the recommendation dropped to a ratio of 4 to 1. Many
common calibrators in use today do not even meet this ratio when compared to the rated accuracy
of HART instruments.
Error Calculations
Error calculations are the principal analysis performed on the As-Found and As-Left test data.
There are several different types of error calculations, most of which are defined in the publication
"Process Instrumentation Terminology".[5] All of the ones discussed here are usually expressed in
terms of the percent of ideal span which is defined as:
% span =(reading - low range) / (high range - low range) * 100.0
The first step in the data analysis is to convert the engineering unit values for input and output
into percent of span. Then for each point, calculate the error, which is the deviation of the actual
output from the expected output. Table 1 gives a set of example data and error calculations,
while Figure 1 graphically illustrates the resulting errors.