Quality 101: The Basics of Making a Measurement
Whenever a person makes a measurement, they are really operating a calibration procedure. What is a calibration? A calibration can be defined as the determination of the amount of some physical property by comparison of the number of units of a measure of that property to a standard of known quantity. This may sound rather complicated at first, but to illustrate its simplicity let’s examine the elementary measurement involved when you purchase tomatoes in a supermarket.
The weight of the tomatoes (unknown physical property being measured) must be compared to some known quantity using some sort of measurement apparatus. The measurement apparatus in our example is a scale, which has a dial marked in the appropriate units of measure (pounds or kilograms) and an internal standard of known quantity. The person weighing the tomatoes is actually operating a simple calibration procedure that consists of putting the tomatoes on the scale, reading the value displayed by the scale, and transferring the indicated weight from the dial to the tomatoes. As in any calibration, the accuracy of the weight of the tomatoes is highly dependent on the accuracy of the standard used to weight the tomatoes, e.g. the scale. A metrologist may refer to this as a calibration by comparison, or a transfer measurement.
This process is no different from one used to weight an ingot of gold, but the precision and accuracy to which it is weighed is much greater than that required for the purchase of tomatoes. This analogy is extreme, but true because the measurements themselves involve the same general considerations. Now let’s focus on what goes into performing a calibration.
The International Vocabulary of Metrology (VIM) defines calibration as: “The closeness of agreement between the result of a measurement and the true value of the measurand.” It is the process of calibration that is aimed at determining just how close the measurement result is to the true value. Once a set of expectation parameters are established, calibration verifies that the measured result is close enough to be acceptable-within the desired parameters. The calibration is founded on an acceptable procedure that is meant to yield results that is determined mathematically to be acceptable, or not. All measurements have certain common characteristics: preparation, performance, and reporting the measurement results. Let’s delve into these characteristics from a calibration perspective in a little more detail.
Preparation: In preparation, the requirements are reviewed, the person performing the calibration selects a satisfactory procedure, establishes a data analysis technique, notes environmental constraints, and selects the reference standards to be used.
Performance: The calibration procedure provides all the elements of calibration: preliminary operations, measurement operations, and data recording. Preliminary steps for both the standards and the device being calibrated are given in the procedure. Measurement operations are presented in a step-by-step sequence. Finally, provisions are made for recording data and performing the necessary data processing.
Report of Results: The report of results of a calibration delineates the condition of the device calibrated. The calibration procedure typically provides tolerance values (accuracy) for all data to be recorded so that an in-tolerance or out-of-tolerance judgment can be made. This judgment is reported by use of calibration certificates, labels and tags. These are identified and approved by an organization’s quality program. They convey all information that is required for a good report.
Measurements are a part of daily life whether you are driving a car, buying tomatoes, or calibrating a pipette that is going to be used in an operating room to save someone’s life. The units of measure in these examples may be different, and the level of accuracy might be widely different, but the fundamental measurement process will be the same, i.e. the determination of the amount of a physical property by comparison of the number of units of a measure of that property to a standard of known quantity.