first generation of commercial VNAs are no longer relevant.
The second portion of this chapter describes the wide range of measurements and characteristics that can be derived from the basic measurements. In Section 2.3, the basic functionality for making measurements is described along with real‐world issues and errors that affect these measurements. Particularly in VNA‐based measurements, many of these errors can be characterized during a calibration process, and error correction can be applied to the results to remove, to a great extent, the effects of these errors. The calibration and error correction process will be described in detail in Chapter 3. Detailed descriptions of measurements of particular devices are covered in subsequent chapters: linear devices (Chapter 5), amplifiers (Chapter 6), mixers (Chapter 7), and balanced devices (Chapter 10).
Author's notes on the second edition: The evolution of VNA architecture has dramatically increased since the first edition was written in 2011. Almost all VNAs on the market today have a full reflectometer on each port, so the discussion the three‐receiver architecture presented in the first edition of this book (principally the architecture of the HP 8753) has been greatly reduced to make room for more interesting VNA enhancements such as true multiport.
2.2 VNA Block Diagrams
The basic block diagram for a component test system is a stimulus source, which is applied to the input of the device‐under‐test (DUT), and a response receiver at the output of the DUT. For S‐parameter measurements, the inputs consist of incident waves at all the ports, and the outputs consist of scattered waves at all the ports, so in general one would require a stimulus and two receivers at each port. In addition, there must be signal separation devices at each port to isolate the incident and scattered waves.
Early systems measured only transmission and/or reflection response, in only one direction, and thus consisted of at most a directional device (bridge or coupler) at the input and a receiver at the output. These systems were classified as transmission/reflection (TR) systems and were most commonly found as scalar network analyzers, although lower‐cost VNAs sometimes had TR test sets as well. The advantage of a vector TR analyzer is that the errors in the directional device could be removed with calibration and error correction.
Figure 2.1 shows the block diagram of a TR system. For simplicity sake the reference receiver will be normally at port 1, measuring the a1 wave, and the test receivers were limited to the two ports as well; normally the test receiver at port 1 is the reflection receiver (b1), and the test receiver at port 2 is the transmission receiver (b2). The source is typically split using a two‐resistor power splitter or a coupler to create a reference signal that is proportional to the incident signal on the DUT, followed by a directional‐coupler or directional bridge, whose coupled arm goes to the reflection test receiver, measuring b1. After the DUT, the transmission test receiver measures b2. Because of advances in VNAs and integration of dual reflectometers and receivers, this architecture is seldom seen in modern VNAs.
Figure 2.1 A TR network analyzer block diagram.
A full S‐parameter system extends the block diagram of the TR system by adding a reflectometer at each port and provides a source at each port. Older systems would switch the source between the ports using a test port switch. Two such block diagrams are shown in Figure 2.2. There are two distinct versions that use either one or two reference receivers.
Figure 2.2 S‐parameter block diagrams for a three‐receiver and four‐receiver VNA.
The three‐receiver version (upper diagram) was common in lower‐cost or RF network analyzers in the past but has largely been replaced with four receiver versions. Having individual receivers for all the reference and test port channels provides for more and better calibration choices, as will be discussed in Chapter 3.
Older analyzers such as the HP‐8510 used separate external sources and switched the source between the ports; others had internal sources, but the cost of the source was a major portion of the instrument cost, so a single switched‐source was used in these integrated analyzers. Often, the reference channel splitter was integrated into this switch as well. This provided a compact switch‐splitter assembly and allowed a lower‐cost alternative to individual splitters or directional‐couplers.
Modern network analyzers make use of a hybrid approach with two or more internal sources, such that more than one port at a time can have an output signal, as shown in Figure 2.3.
Figure 2.3 Multiple sources in a single VNA.
While there is no requirement for having more than one port active in traditional S‐parameter measurements, advanced measurements such as two‐tone intermodulation distortion (IMD), active‐load, or differential‐device test can make good use of these extra sources. In these systems it is common to use a directional‐coupler in both the reference and test arms for lower loss from the source to the test port, allowing higher maximum test port power. One synthesizer (that is, the frequency generation unit) may be shared between two ports (of a 4‐port VNA) with an individual output amplifier and leveling circuits available at each of the ports. This system can provide outputs at any of the ports at the same time or provide two different frequencies out of pairs of ports, which is useful for mixer test applications.
2.2.1 VNA Source
The VNA source provides the stimulus for the S‐parameter measurement. In the original VNAs these were open‐loop sweepers, but since about 1985 the use of frequency synthesizers has become the norm. Sweepers used open‐loop swept‐frequency oscillators to produce the stimulus signals; synthesizers replaced the open‐loop control with fractional‐N or multi‐loop signal generation where the output signal is digitally derived from a 10 MHz reference oscillator to a resolution of less than 1 Hz. Early sources were routed directly to the test sets of VNAs, which were also external, stand‐alone instruments, often with the first converter assembly inside.
More recently the source is provided internally to the VNA. In the first integrated VNAs, the signal quality of the internal source was less than that of external instrumentation sources but had the advantage of being much faster in sweeping frequency. For applications such as filter tuning and test, fast sweep times across wide frequency ranges were required. Common to VNA sources are the ability to vary power level
