Analyze the importance of calibration work in test and measurement

　　Calibration always seems to be scheduled at the most critical time in a project, such as when a team is busy preparing for the annual industry trade show. As an example of the impact of calibration on a project, let’s assume that a piece of test equipment has a calibration cycle of 6 months. In the 5th month, the design engineer starts a test project that will last for two months. If the instrument is recalibrated during the test, the accumulated drift or error in the first 5 months will be large, which will require retesting. Would it be better to calibrate the equipment before starting the large project that requires it? Or is it better to postpone calibration until it is safe? Like any other event, calibration should be planned in the project planning software and should be on the critical path. If calibration is neglected, it will cause project delays.

　　What is calibration?

　　Some people consider two instruments (such as an oscilloscope and a multimeter) that read the same to be "calibrated." However, this approach is problematic, or at least completely unscientific. It is like asking an accountant to audit his own books. There are three obvious situations that cannot be explained if calibration is performed in this way: First, if one instrument is correct and the other is wrong, who is right and who is wrong? Second, if the two instruments fail in opposite ways, how can an engineer say that neither is correct? Finally, if the two instruments fail in the same way, the result is wrong and the engineer has no idea. Without a true traceable external standard, it is impossible to say that an instrument is correct.

　　Calibration standards (circled) are critical to the mission. Image courtesy of NASA/JPL-Caltec.

　　The accuracy of the calibration standard must be much higher than the accuracy of the instrument being measured. Don't forget that standards have tolerances too. If the tolerance range of the device under test (DUT) overlaps the tolerance range of the standard, a complete calibration cannot be achieved. This is why it is usually required that the accuracy of the standard be at least 10 times higher than the accuracy of the DUT during calibration. If the tolerance of the standard is well defined and small enough, the DUT can be adjusted to prevent it from reading out of specification due to normal drift between calibrations. When calibrating instruments using the latest technology, it is not possible to have a standard that is 10 times more accurate. As a rule of thumb, a standard that is 4 times more accurate can be used, but with a more complex process that includes cross-checking with other standards.

　　Metrology Laboratory

　　Serious engineers buy instruments from top brands and expect high accuracy from these electronic test instruments. As time goes by, some parameters drift, so engineers send the instruments to metrology labs for calibration. But what is the real situation in metrology labs?

　　Most metrology labs are good and effective. Unfortunately, some are substandard. Some companies choose the lowest-priced lab without investigating its credentials, leaving it up to the engineer to verify the lab's authenticity. Good labs are happy to have you visit and are proud of the traceability of their standards. You can sit down with the lab personnel with the calibration manual for your instrument and confirm that the lab has the necessary instruments to perform high-precision calibrations.

　　You must also ensure that the lab’s instruments are properly calibrated and traceable. Most countries have their own national standards laboratory, such as the National Institute of Standards and Technology (NIST) in the United States1. A good metrology laboratory will have records showing that its instruments are properly compared to a chain of standards that are traceable to master standards maintained by NIST or other national standards.

　　The internal parts of a test instrument, such as voltage references, input dividers, amplifier gain, and offset, drift over time. Good calibration ensures that typical small drifts do not affect measurements. Calibration is about finding and correcting for this drift. But sometimes things go wrong: an instrument may be dropped or an operator may slip and probe a higher voltage. For example, a digital multimeter (DMM) may be overloaded, causing large errors. Because the DMM's inputs are protected by fuses or circuit breakers, some people mistakenly assume that this situation will not cause it to exceed the rating. However, the high voltage may bypass the input protection device or damage the circuit in the split second before the protection device takes effect.

　　When we send an instrument out for calibration, we expect the metrology lab to return the instrument to a calibrated state. Engineers should also receive a calibration report that lists the instrument's deviations before and after calibration. If the calibration report shows that the instrument has a large calibration error, it may be necessary to re-run the test items that have been completed with the instrument and make new measurements.

　　What is the frequency requirement for calibration?

　　This question cannot be generalized because it varies with the instrument, environment, and application. Test instrument manufacturers will give recommended calibration intervals under typical conditions; extreme conditions and critical measurements may require more frequent calibration. Here are some general rules for calibration intervals:

　　1. Routine calibration may be required as specified by user contracts, quality standards organizations, military specifications, or other industry requirements. Before testing, applicable requirements must be considered to ensure that the calibration or verification of the test equipment is met.

　　2. Before and after key measurement projects. For example, after completing the trial run of a new product, the design engineer will characterize the product to ensure that the technical requirements are met and optimize the test steps. Final test adjustments made at this time will fundamentally shorten the test time and affect the revenue. Complete and reliable testing requires verification of the status of the instrument before and after the test cycle.

　　3. When you suspect a measurement error, or when the instrument is overloaded or dropped, it is very important to check and confirm the calibration and safety status (for example, if the instrument drops and causes a short circuit between the wire and the housing).

　　High-precision calibration is not a luxury that can be ignored - calibration ensures the reliability of test instruments and even the safety of personnel. For example, before using a device, an engineer may measure the voltage of a meter to ensure it is safe. If the meter is damaged or provides inaccurate information, it can cause injury or even death. In addition, calibration is a guarantee of quality, ensuring that the final test results of the product are accurate and meet the technical specifications when it is delivered to the customer.