Periodic checks attest to accuracy

When test equipment returns from a cal lab, you're quite confident that it will perform within specifications. By running periodic performance checks, you can maintain that confidence until it's time for the next calibration. Many test instruments include self-diagnostic and self-calibration capabilities, and you can also use external references to check your equipment's measurement accuracy.

Self-diagnostics let you check an instrument's basic performance. These "first line of defense" checks typically verify the instrument's default settings, test its memory, and confirm that the instrument's internal power supplies operate at the proper levels. In some instruments, self-diagnostics also may measure the units' internal temperature and reference sources and test the communication ports.

Some instruments run their self-diagnostics each time you turn them on. Others run their diagnostics at the push of a button or the click of a mouse. Table 1 lists a few instruments at HP's Houston calibration lab and the self-diagnostics that check each instrument's performance.

When you run an instrument's self-diagnostics, the instrument typically provides a "pass" or "fail" message. If a self-diagnostic fails, the instrument will often display an error code, which you can interpret from the instrument's manual.

Some test instruments go beyond self-diagnostics and include self-calibration features that perform internal checks and adjustments using an internal reference. A DMM, for example, may contain an internal zener voltage source for performing self-calibrations. Oscilloscopes often include a calibrator output signal that lets you check their voltage and timing measurements.

An instrument manufacturer may place a time limit on a self-calibration, after which time the instrument may drift beyond a specified range. Typically, a manufacturer will guarantee an instrument's performance for 24 hrs following a self-calibration.

In some instruments, a manufacturer may specify a 1-hr time limit for when you need the least measurement uncertainty. The time limit depends on the uncertainty you require.

Table 2 shows the uncertainties and their time limits for an Agilent/HP 3458A DMM. The "24 hours" column indicates the instrument's uncertainty 24 hrs after performing a self-calibration. The "90 days" column indicates a greater uncertainty. In this case, it's 24 hrs after a self-calibration performed 90 days after a lab calibration. The "1 year" and "2 years" columns follow the same pattern. As you'd expect, the uncertainty increases the longer it has been since a lab calibration.

Some instruments take self-calibration to another level. The Fluke 5700 multifunction calibrator, for example, measures its own outputs and identifies the output that's closest to the manufacturer's tolerance limits as a percentage of that limit. You should record this information and track it, so you can use it to predict when that parameter might exceed tolerance limits. "Predict the future," below, gives an example of how you can use linear regression to predict when an instrument might require calibration.

In addition to self-diagnostics and self-calibrations, you can use check standards to verify your instrument's performance. Check standards are typically passive artifacts that tend to hold their values over time. Because passive artifacts historically hold their values well, you can use them as "sanity checks" on your equipment. Table 3 shows some types of check standards. Several 50-V RF terminations that you can use as check standards appear in Figure 1.

Figure 1 50-V BNC N-type and APC 3.5 terminations are check standards for RF test equipment.

You should monitor the check standard's value each time you use it to ensure that its value hasn't drastically changed. If the check standard's measured value changes appreciably from its previous values, double check your measurement setup and measure the value again to verify that a poor connection didn't cause the problem.

If practical, keep a second check standard as a type of measurement arbitrator. Using a second check standard will help you verify whether the first check standard's value changed, or whether the equipment that measured the first check standard has a problem. If the first check standard's value differs significantly from previous values, you may need to question the measurements taken with equipment checked with that standard. For more information on check standards and how to use them, see the NIST Engineering Statistical Handbook (Ref. 1).

You should periodically run tests with self-diagnostics, self-calibrations, or check standards. The frequency at which you should check your instruments depends on how often you use each piece of equipment.

Equipment that you use every day may require self-calibrations once a day, or once per production shift, or even once per hour depending on the uncertainty you need to maintain. Equipment that you use infrequently may require a self-check or self-calibration once a month. You can use check standards less frequently than self-diagnostics or self-calibrations, but you should use them whenever you question an instrument's performance following a self-calibration.

Regardless of how frequently you check an instrument, always document the results of your check. Keeping records will let you build a history of your instrument's performance, which you can use to help predict its future performance.