by BEN MAXSON, Head Applications Engineer, Copper Mountain Technologies

Here are a few basic things to know about how to use vector network analyzers that measure complex impedance in RF circuits.

Engineers who work with signals at relatively low frequencies don’t have to worry about conducting power through circuits. Simple wiring suffices for handling frequencies measured in the tens of kilohertz or less. The resistance of the wire is relatively low and it doesn’t much impact the shape of the signal it carries.

Of course, this situation changes drastically at radio frequencies and higher. Here, the signal wavelengths involved can be comparable to or smaller than the length of the conductors in the circuit. Design engineers must think of circuit components in terms of how they interact with traveling waves. Circuits become transmission lines, which, when terminated in their characteristic impedance, transfer maximum power to the load. When the termination impedance doesn’t equal the characteristic impedance, the part of the signal not absorbed by the load reflects back to the source.

These fundamental facts of RF transmission are becoming of interest to a large number of engineers who are working on projects with wireless technology. The Internet of Things concept has brought forth wireless products ranging from Bluetooth in bathroom scales to WiFi in refrigerators. When designers add wireless capabilities to their products, the results must be tested to confirm optimal function and performance.

The most efficient and accurate instruments for testing RF hardware are vector network analyzers (VNAs). A VNA stimulates an electronic device by sending it a signal, then measures how much of the signal passes through the device or is reflected back. These instruments evaluate the performance of components ranging in complexity from a single cable to highly integrated RF circuits and devices. VNAs record both the magnitude and phase of the signals received. The measurements are expressed as scattering or “S” parameters that can be mathematically manipulated and displayed in many different ways.

A VNA is comprised of a signal generator, a test set, one or more receivers, a display, and the computing power needed to manage the measurement procedure and subsequently process the results. During operation, the test set sends a signal using one of the analyzer’s ports through a cable to the device under test (DUT). The amount of signal that passes through the device and continues to a second port on the VNA is expressed as a transmission parameter.

Similarly, the degree to which the device blocks or reflects the signal is recorded as a reflection parameter. The measured reflection or transmission can be anywhere from 99.99999 to 0.000001% of the signal that was sent. For example, if the VNA sends a signal of a certain power and 98% of what was sent returns, reflection is considered to be nearly full-scale or complete.

When selecting a VNA, the first consideration is the normal operating frequency of the DUT. For instance, Bluetooth devices operate in a range of 2.4 to 2.5 GHz. Thus, a VNA testing a Bluetooth device must generate signals below and somewhat above 2.4 GHz. Typical operating frequency ranges for other applications include broadcast TV equipment below 1 GHz and cable TV amplifiers and components below 1.3 GHz. The majority of cellular equipment operates below 2 GHz. Thus, for tests of cellphone antennas, base stations, cabling or amplifiers, a VNA must generate test signals at 2 GHz and below. A widely used frequency for WiFi is approximately 5.8 to 5.9 GHz. To measure the performance of devices using that band, a VNA must generate signals in a range below and above approximately 6 GHz.

In some situations, however, simply matching the VNA to the frequency of the DUT will not provide sufficient information. For testing in an R&D environment, especially for active devices, users should consider a VNA that can generate frequencies at least three times higher than those employed by the device being tested. The added range lets engineers detect unexpected electronic behavior outside the device’s specified operating range. It also helps characterize nonlinearities and allows for future expansions of bandwidth requirements. On the other hand, a VNA used exclusively to confirm specs on a manufacturing line might not need an extended measuring frequency range.

The number of ports on a VNA also defines its capabilities. As when choosing a VNA frequency range, the characteristics of the devices being tested determine the number of necessary ports. A simple cable or antenna requires an analyzer with one port. Devices with both an input and an output, or two inputs, need a two-port analyzer. Devices with numerous inputs can be analyzed with a four-port analyzer or multiple two-port units. In some cases, multiple ports may be unnecessary. For example, there are lower-cost, single-port VNAs that are candidates for designers adding Bluetooth to their product who simply want to know if the antenna works well and need to investigate ways to optimize its matching circuit.

In the VNA measurement process, there is a fundamental compromise between speed and measurement quality. VNA software contains a setting called intermediate frequency (IF) bandwidth. Increasing the IF bandwidth shortens the measurement sweep time, yet increases the amount of measurement noise. Conversely, decreasing the IF bandwidth slows down the pace of measurement and improves the dynamic range, thereby lowering measurement noise. A suitable VNA will be able to make a measurement of a required accuracy in what the user considers to be a reasonable amount of time.

Because VNAs differ, prospective buyers should consult with the instrument vendor or the specific instrument’s data sheet to find the typical measurement time at the needed dynamic range. For example, if some tests need 110 dB of dynamic range, the vendor or data sheet should be able to provide the typical sweep time at that range. The buyer may also look for the instrument’s dynamic range at the lowest tolerable measurement speed.

VNA instruments typically come in three basic configurations. A traditional VNA is set up in the form of big-box test equipment with a large, self-contained chassis that houses the signal generator, the test sets, the receivers and a computer. Another approach is a VNA card or module that plugs into a slot in a general-application equipment chassis. A third format represents purely modular systems.

In modular arrangements, a USB- or Ethernet-controlled analyzer separates the measurement module from the processing module. The measurement hardware consists of a more compact unit that connects to an external PC. The VNA software offloads the measurement results formatting and display to the PC. The user interface runs on the PC app that enables quicker mouse point-and-click navigation through the instrument menu pages. PC hardware can be updated independently from the measurement hardware, and the data generated can be saved directly onto the computer. This eliminates the need to transfer information by a thumb drive or by other means to generate reports and/or presentations.

Used and refurbished VNAs are available at relatively low prices. But depending on the viability of the company that produced them, service and support may no longer be available. As with any electronic equipment, VNA technology is continually upgraded. The performance of older units, especially their built-in computers, quickly becomes outdated.

Sometimes the VNA serves in a role where it repeatedly tests the same device over a period of time. Here, the instrument should have automation capacity that consists of a programming interface and software designed to support autonomous control of the measurement process.

Regardless of a VNA’s configuration, it must be accurately calibrated along with the components with which it works. Calibration eliminates repeatable or “systematic” measurement errors. The presence of imperfections in the cables, adapters and other components between the measuring instrument and the device being tested will cause such measurement errors. Collectively, those connectors and components are called the test fixture.

For example, any length of cable will produce a delay in signal propagation. The delay may be a matter of picoseconds, but even that amount can be significant in certain measurement scenarios. Because systemic errors remain constant, calibration allows them to be quantified and mathematically accounted for. Calibration produces measurements nearly as accurate as would be obtained from a system without any imperfections.

Traditional calibration involves the use of calibration kits with standards that provide known measurement characteristics as the baseline. These characteristics are provided in advance to the VNA, which measures these standards using the fixture that will be employed in testing the actual device. The VNA then compares the measured results to the entered characteristics. Based on the difference, the VNA mathematically determines corrections required to remove the effects of imperfections in the fixture. Operators must always handle cables and adapters with care, as out-of-spec or damaged fixture components can affect accuracy and even damage the components to which they are connected.

It’s important to run calibrations often enough to ensure accurate measurements. The nature of the measurements and the environmental conditions determine the frequency of calibration. A basic pass/fail evaluation may not need acute precision; depending on the volume of work, monthly calibration may be okay, especially if the elements of the fixture remain unchanged.

However, the situation is different when an application requires critical measurements that are near the limits of the test instrument’s precision. Here, daily or even hourly recalibration is necessary to compensate for changing temperatures of the environment and other factors. Environmental temperature changes of 1 or 2° C can significantly affect measurement results.

Calibration kits can be purchased or rented from numerous vendors. The cost of a typical top-level calibration kit can be $10,000, and rental fees will reflect the provider’s investment. More economical but slightly less precise kits are also available. Calibration kits generally contain a variety of standards that are connected to and disconnected from the instrument individually.

A relatively new development in calibration technology is the Automatic Calibration Module (ACM). It is a separate physical device from the VNA, but is supported by the VNA software. When connected to the fixture, it internally switches among a number of different standards, stores the results and sends them to the VNA software so adjustments can take place. Some ACMs work on an exclusive basis with the VNAs of their manufacturer, while others can be applied universally. ACMs are generally faster and easier to use, less expensive and more accurate than traditional mechanical standards.

As digital communication technologies continually develop, the need to analyze and optimize the hardware involved becomes crucial. Both development and production operations will benefit from fast, flexible and comprehensive evaluation of simple and complex components. VNAs provide cost-effective RF performance measurement and verification that can be customized for specific research, development and manufacturing process situations.

Copper Mountain Technologies
www.coppermountaintech.com