September 2008

Keep it valid: Calibrate

When interested in obtaining valid test data, calibrate all equipment used in the test measurement system. Calibration involves applying a series of known inputs to the equipment and recording the resulting output values. The calibration process may also involve adjusting the equipment to eliminate or minimize the errors that can occur in these output values.

Calibrating all equipment in the test measurement system means calibrating every piece of instrumentation installed to set test conditions or to make resulting test measurements. We must use calibrated instrumentation to make any measurement we cite in the test results or use to make test conclusions.

Calibration is an essential step because it is the only thing that can guarantee validity of measurements and test conclusions. Calibration allows us to assign valid numbers to measurements and to state possible errors in those measurements. Without calibration, we could only report relative results such as, "There was an increase in vibration as the flow increased."

When we say calibration involves applying a series of known inputs, we mean instruments are calibrated over the full range of measurements we take. Any measurements we take outside the calibrated range we cannot consider valid since there is no guarantee the instrument would have produced those specific results. We also need a series of inputs to increase our confidence in the interpolated values that fall between each known input. We must calibrate at a sufficient number of steps in order to determine instrument linearity or conformance.

Using known inputs means we know the value of the source or signal we use in the calibration is true within some specified tolerance. We use calibration standards to serve as a reference when producing these known inputs. These standards have a substantially smaller error than the test instruments and have themselves been calibrated by more accurate standards.

The recording of the outputs during calibration means we use documentation to specify the values of the output for the series of known inputs. This documented information allows us to determine the degree or error in our measurements and serves as proof the calibration took place.

Since calibration is a process, it is a planned activity that uses a prescribed procedure. The procedure specifies how, when, and under what conditions the calibration took place. The calibration process is often a requirement of a manufacturer's quality assurance program, which will document the activities to carry out.

We may adjust the output of the instrument we are calibrating if such adjustments are available. A device such as an RTD does not have adjustments, and the calibration involves recording the output resistance for a series of known input temperatures. A pressure transducer with integral amplifier will likely have zero and span adjustments available. We can use these adjustments to set the output as close as possible to a nominal output such as 0 to 10 volts DC when we apply a series of known pressures to the input of the pressure transducer.

Each test laboratory and manufacturer will have to decide whether to carry out calibrations in house or hire an outside calibration service contractor. This decision is based on the cost of each choice and an evaluation of any factors that do not have a direct cost but do impact the operation of the facility. The final decision may be a combination of the two choices, where some equipment is calibrated in house and some is contracted out.

Component calibration

Component calibration involves separately calibrating each component in the measurement system, usually in a calibration laboratory. The separate components could consist of the transducer, the signal conditioner, and the display device.

The components making up a temperature measurement system might consist of an RTD, a linearizing amplifier, and a digital meter. In this case, you would calibrate the RTD in a temperature bath with a standard RTC as a reference and use a resistance bridge or high-accuracy resistance meter to measure the RTD resistance. Calibrate the linearizing amplifier with a decade resistance box or RTD simulator connected to the input and a high-accuracy digital voltmeter connected to the input. Record the resulting digital display as the output reading. During these calibrations, adjust the output of any component with zero, and span adjustments to be as close as possible to the ideal nominal output. Calculate the overall system accuracy from the errors found in each separate component calibration. After calibrating the components, install and wire them in the test rig or plant.

System calibration

In a system calibration, the individual components of a measurement system are wired together as they would be in the final field installation. We then calibrate the assembled system, usually in a calibration laboratory. For the temperature measurement system described above, we would insert the RTD into a temperature bath along with a standard RTD. During the calibration, vary the bath temperature, and note the standard RTD reading and the digital display reading. Advantages include less time, fewer standards, and known system error rather than calculated. Disadvantages include a more complicated procedure, recalibrating individual components if the system is dismantled, and new system calibration if any one component is replaced.

SOURCE: Fundamentals of Test measurement Instrumentation, by Keith Cheatle, ISA Press, 2006. To order this book, visit www.isa.org/books.