Using Calibration to Optimize Performance in Crucial Measurements

Keysight Technologies

Using Calibration to Optimize Performance in Crucial Measurements

Article Reprint

Using Calibration to Optimize Performance in Crucial Measurements

Dipti Chheda

Keysight Technologies, Santa Rosa, Calif

Every engineer responsible for a test system is also responsible for the accuracy and repeatability of the measurements it makes. Repeatability, perhaps more than pure accuracy, is often the key to success in design, manufacturing and ongoing operations. In a test system, repeatability is also the foundation of the warranted performance of the included instruments. This is especially true for crucial equipment such as network analyzers, signal analyzers, power meters, oscilloscopes and signal generators. If any specified parameter is out of tolerance, measurement results can be negatively affected.

An accurate, professional and accredited calibration is the bedrock that ensures reliable and repeatable results. Calibration and metrology are a specialized subset of engineering, and relatively few engineers have been trained in these topics. Fortunately, developing familiarity with a few fundamental concepts will improve measurement performance, enhance the interpretation of results and, ultimately, reduce the risks associated with every decision that is based on measured results.

Meeting Measure Requriements

A test system supports a test plan, and the essential first step is to identify the crucial specifications that characterize the performance of the device under test (DUT). Each specification will have an associated set of tests, tolerances and accuracy requirements. The development of the test plan includes the selection of hardware elements that provide the necessary features and functions. For an engineer, the natural response is to thoroughly understand the choices and tradeoffs in the various hardware alternatives.

Typically, less time is spent considering the calibration and repair services needed to sustain the warranted specifications of each instrument. It’s easy to assume that periodic calibration is all that’s needed to ensure measurement integrity over the long term. In reality, test equipment ages and drifts, and sometimes it breaks. What’s more, calibration is not a generic commodity, and the process of ensuring long-term measurement repeatability is not as simple as “set it and forget it.” Taking a proactive stance can have a significant impact on the ongoing accuracy and repeatability of the test system, not only reducing the risk of out-of-tolerance measurements, but actually improving the system’s effective accuracy. This can help ensure the performance of the DUT and enhance overall productivity in manufacturing.

Using Calibration to Improve Spurious Measurements

An example focused on the pursuit of spurious signals using a signal analyzer will show how to ensure greater confidence in results. 

Collectively, spurs are the source of many potential problems. In a radar system, spurs may obscure the system’s ability to see small return signals, which can affect the believability of what’s on the screen. For those performing sensitive fi eld operations, self-generated spurs emanating from a receiving antenna may betray their presence and location. Thus, when making a measurement, the key question is when a spur appears, is it real?

not known in advance, the process starts with a wideband spectrum measurement. The best setting for input attenuation depends on the magnitude of the largest signal in the widest span. With this combination of wide span and the likely presence of larger signals, many low-level signals will be missed due to insuffi cient frequency resolution and a higher-than-desired effective noise fl oor. To increase the available dynamic range, input attenuation should be minimized while remaining suffi cient to prevent analyzer generated signals, such as harmonics and intermodulation, from interfering with the measurement. The resolution bandwidth (RBW) should be just narrow enough to reduce the effective analyzer noise floor and resolve closely spaced spurs while providing sufficient measurement speed.

A useful example is the verifi cation of spurious-free dynamic range (SFDR) in a radar exciter. The carrier fundamental is at 10 GHz. The exciter’s SFDR must be 80 dB below the carrier (-80 dBc), and this equates to -65 dBm relative to an exciter with a +15 dBm output level. These are the key specs for the DUT. Characterizing those parameters depends on the signal analyzer’s dynamic range, and that depends on specifi cations related to noise and spurs. Suppose a signal analyzer has a specifi ed displayed average noise level (DANL) of -148 dBm. Because DANL is typically normalized to a 1 Hz RBW, the actual specifi - cation is -108 dBm when using a 10 kHz RBW. Residual responses are specifi ed to have a level of -100 dBm or less. Related to this, third-order intermodulation (TOI) is specified to be -90 dBm. Understanding the trade-off between expected DANL (not a hard specifi cation) and TOI is important when setting input attenuation and mixer level for a spurious measurement. Beyond the generic specifi cations, it would also be helpful to know the actual performance of an individual analyzer. Is it below spec, at spec or better than spec? If better than spec, how much better is it? This information is essential to enhancing the ability to interpret the actual measurement results from the analyzer.

Conclusion

Opting for the most dependable calibration provider is the best way to ensure that test equipment continues to provide the performance that led to the purchase decision. In a commercial setting, this often translates into better throughput, margin and yield. In the aerospace and defense environment, it increases the likelihood of mission success. In any setting, reliable calibration ensures consistent results that make it easier to pinpoint product or design problems thereby minimizing delays in development and manufacturing.