An analysis of golden unit versus parametric testing of digital radios

Golden modems, transmitters and receivers are commonly used in production test of digital radio systems. Golden standards can provide an inexpensive reference for functional test of major subassemblies. However these simplistic custom built test fixtures often provide little insight into the performance level of the device under test, and if used improperly can lead to disastrous conclusions, particularly when bit error rates are of interest. The use of commercially available test equipment enables standards based parametric tests that are independent of any golden unit and traceable to international standards. Unfortunately parametric testing is often viewed as more expensive for complex digital modulations, leading many test professionals to choose a golden unit and a bit error rate test set as a substitute. This paper looks at the common fallacies of golden unit testing with respect to digital data links and the alternatives that are capable of providing more accurate performance characterization. This analysis, designed for the test professional, provides guidelines for the use of golden units versus parametric testing, as well as the basic physics of phase noise and nonlinearity behind choosing the appropriate test solution.