Measuring Output VSWR For An Active Levelled Source

Mismatch error is one of the major contributions to measurement uncertainty in RF & Microwave calibration. When a signal source is used in calibration applications, knowledge of the output impedance (or source VSWR) is necessary to enable users to estimate the mismatch uncertainty contribution when performing their measurement uncertainty calculations. However, measuring the output match of an active levelled source can be difficult. The measurement techniques generally used for passive components cannot be used for active levelled sources. If those methods are used, they are likely to give erroneous misleading results. This paper discusses the output VSWR measurement method chosen for use in manufacturing test and calibrations systems for a new RF & Microwave source instrument. It also describes the steps taken to validate results by comparisons with other methods less well-suited for automation and production line usage. Introduction Most electrical measurement practitioners are aware of the maximum power transfer theorem – usually introduced to students as part of basic circuit theory for the case of a DC voltage source and resistive load. For maximum power transfer to the load, the source and load resistances must be equal. The concept of maximum power transfer appears again when considering transmission line theory, often accompanied by complicated (and confusing) mathematics. The practical implications are very familiar to RF & Microwave metrologists – the need to maintain matching between source and load impedance and any transmission lines (cables, adapters, etc) used to make interconnections in their measurement system. Knowledge of the degree of mismatch can be used to determine measurement uncertainties, and therefore drives the need to specify and measure the impedance of the devices concerned. The majority of ‘load’ or ‘interconnect’ devices used in RF & Microwave measurements such as power sensors, attenuators, adapters, are well specified for their impedance (match) characteristics. However, the same is not usually true for signal sources. The majority of signal sources used in RF & Microwave calibration applications are general purpose signal generators. When a new signal source was developed specifically for RF & Microwave calibration applications it was considered essential to provide detailed output impedance (VSWR) specifications and include measurement data on calibration certificates issued at manufacture and routine recalibrations – not only to confirm specification compliance but also to provide the source impedance information required for users to perform mismatch uncertainty calculations. A number of impedance/match characteristics measurement techniques are available and commonly used, such as reflection bridges and network analysers. The measurement results can be expressed in a variety of terms such as Reflection Coefficient, VSWR (Voltage Standing Wave Ratio) and Return Loss, S-parameters, etc, and a value expressed as one parameter can easily be converted to another. (The references at the end of this paper contain a few of the many sources of detailed explanations of the theory and relevance to RF & Microwave measurements ). The key issue is that measurements are typically made by applying a signal to the impedance under test and measuring the signal reflected, from which the required parameter can be calculated. Unfortunately in the case of an active source, which is itself producing a signal, the result obtained using these methods can be misleading or incorrect. Other techniques for generator output impedance measurement are available, based on reflecting some or all of the generator’s own power back into itself but these tend to be cumbersome, time consuming, and difficult to automate. In the case of the signal source discussed in this paper, the objective was to implement a technique that could easily be implemented in an automated system [3] deployed in the manufacturing plant and service centres, which would not require manual intervention during the measurement process. More onerous procedures could be tolerated for making cross-checks to validate the results, and validation measurements were also sought from independent external laboratories having relevant expertise and experience of generator source match measurement. Signal Source Design Architecture The signal source requiring measurement is the Fluke 9640A RF Reference Source, with a frequency range from 10Hz to 4GHz at amplitudes from +24dBm to -130dBm. It is designed to generate the signals necessary for the most common RF & Microwave calibration applications and provide inherent accuracy without the need to monitor or characterise the output with additional equipment during use. (For example, measuring of the output amplitude with a power splitter and power sensor, etc). In order to facilitate delivery of the output signal direct to the load or Unit Under Test (UUT) input and minimize performance degradation due to cabling and interconnections, the instrument has an external levelling head (see Figure 1). Signals are generated in the mainframe and fed to the levelling head containing the level detector and attenuator circuits. The 9640A output impedance is level dependent. At high output levels, the output impedance is defined by the levelling circuits and at low levels by the attenuators. These attenuators are realised by a series of switched step attenuators allowing the attenuation value to by set in 10dB increments. The attenuators and levelling circuits are controlled from the output amplitude setting. Only two individual output impedance conditions are presented at the output throughout the amplitude range requiring use of the attenuators, and a third is presented for higher amplitudes when no attenuation is required. When the output impedance is defined by the attenuators, the output signal is at sufficiently low level that the typical VSWR measurement methods used for passive devices can be employed without problems. However, at the high output levels when the levelling circuits define the output impedance a different technique must be used. Measurement Methods For Active Outputs A traditional measurement method for an ‘active’ or ‘levelled’ output is to use a short circuit or mismatch to reflect some or all of the generator under test’s own output back into itself. With a perfect match, all of this returned signal would be absorbed by the generator output impedance. However, an imperfect source match will reflect some of the returned signal back out of the generator. The portion of the signal reflecting back from the generator output combines with the original output signal and either adds or subtracts by some amount dependent on its phase relationship. In order to distinguish between the Level Sensing Attenuators