Calibration & the Challenges of Choice

Calibration & the Challenges of Choice

A poor or incomplete calibration can lead to significant errors in your measurements. The time to develop a product may exceed your original schedules and production yields may be too low if the test instruments are not performing to specification. If your test instruments are not measured during calibration to the specifications you depend on, the devices you test are at risk. That can be costly. A detailed calibration report is the only way to verify for sure your instruments are performing to specification. Lower your risk and sleep better by reading this eBook.

Why calibrate?

The warranted specifications of electronic instruments often rely on the ability to fine-tune the hardware under microprocessor control, for example amplifier gain, filter corners, DC offsets, and other functions are all adjusted during the production process. Over time, depending on the measurement environment and component aging, drift can occur. Sometimes this drift is linear over time as shown in Figure 1. Other times sudden shifts happen. Similarly, torque wrenches or precision thermometers can change based on how they are used (or abused). The whole point of sending your instrument for calibration is to accurately measure the current performance to address the questions “Is it still fit for use? Can I continue to count on the warranted specifications?” Your measurements matter. You make pass and fail decisions based on the measurements you make. Proper periodic calibrations identify any out-of-tolerance test equipment drift and correct for it in a timely manner, ensuring that the accuracy of the test equipment you rely on to make your pass/fail or design troubleshooting decisions is the least of your worries.

A simplified definition

Calibration is periodically verifying instrument specifications by measuring actual performance using lab standards that in turn have better performance and are traceable to the International System of Units (SI) through standards maintained by national metrology institutes (NIST, NPL, PTB, NMIJ, etc.) Depending on the service provider and the type of device, when an instrument is observed out-of-specification the calibration service may include adjustments. Electronic instruments usually require proprietary knowledge of the internal architecture so generally only the OEM can adjust them. Dimensional standards such as gauge blocks can’t be adjusted, only replaced if sufficient wear has occurred. Mechanical torque wrenches can usually be adjusted if needed.

What happens when an instrument is out of specification?

Many, if not most, people assume that when they submit an instrument for calibration, the service includes adjustments “to the middle of the tolerance interval”. That is not true! See the calibration and repair process flow in figure 2. Calibration laboratories run a series of performance tests. If your instrument is observed in-specification “As received”, you’ll receive a single report indicating that it “Passed”. When an instrument is observed out-of-specification (or out-of-tolerance) your provider may be able to perform adjustments. Note for many modern electronic instruments, adjustments require complex iterative external measurements and calculations to characterize the instrument while changing internal DAC settings.

When your provider does perform adjustments you should expect to receive the measured results both of the initial performance when you submitted the instrument and after adjustments are performed. This way you can determine the impact of your Pass/Fail decisions on your own products while using an out-of-tolerance instrument. The second measurement report, “As completed” is proof of being fit for continued use.

What is Measurement Uncertainty (MU)?

There are no perfect measurements, as they involve errors and imperfections. Engineers carefully review the accuracy specifications of instruments before they put them into their test systems. System accuracy is usually an engineering estimate of the combination of all potential measurement error sources. In calibration, international standards2 require identification of the different error and inaccuracy contributors involved in the measurement of your instrument. These are then to be accounted for as the so-called “Measurement Uncertainty” (MU). So, while MU may sound unfamiliar, think of it as the standard deviation estimate of the combined error sources during calibration. In plain English, it means the measurement inaccuracy. Large calibration uncertainty increases the risk that your instrument is operating out-of-specification, risking measurement errors that impact the quality or performance of your design. As the calibration MU of your test equipment contributes to your overall test system inaccuracy, knowledge of what it is and keeping it small will save you cost and schedule time. As seen in figure 3 the red area under the curve shows that smaller MUs reduce your out-of-specification risk. Look at what appears to be equivalent test results from two different calibration vendors, Lab A and Lab B. Lab A has a relatively small MU represented by the vertical bar. Lab B has larger MU as seen by the longer vertical bar. The risk that your instrument is operating out-of-specification is represented by the red areas. Clearly, Lab A offers less risk than Lab B.

What is traceability?

Traceable measurements ensure uniformity of manufactured goods and industrial processes. Common measurement references are critical to the world-wide exchange and acceptance of products, services and technology. Traceability is one of the challenges faced in design and manufacturing. Linkage to common standards is critical to measurement correlation, particularly for teams in different geographic locations.

A traceable measurement is one that is linked to recognised standards having known better accuracy through an unbroken chain of comparisons. Records are required to link a product or service event to a process, status or time. If the process is measurement, detail about the equipment, procedure, uncertainty, test conditions and personnel are all relevant to demonstrating traceability.