VNA Reflections

Accurately Analyzing Nonlinear Active Devices

November 28, 2012

Design activity for digital communication RF components such as power amplifiers (PA) has never been higher, due in large part to the proliferation of smart phones. At the top of the list of concerns for PA design engineers is power added efficiency (PAE). High PAE in a power amplifier provides:

• Increased battery life for mobile devices

• Maximum antenna coverage for base stations

• Lower electricity bills for network providers

• The opportunity for manufacturers to charge higher prices for better performing devices and consequently, greater revenue return per wafer.

PA design engineers must also consider the impact of amplifying signals with high Peak-to-Average Ratios (PAR) on maintaining acceptable Error Vector Magnitude (EVM) levels and hence low bit error rates. A linear vector network analyzer (VNA) provides essential information regarding the performance of small signal amplifiers operating under linear conditions. However, when high-power amplifiers are designed for compressed nonlinear operation, the VNA must provide additional information to help the engineer optimize the PA design. The VectorStar™ nonlinear VNA system from Anritsu provides essential data to help develop optimized nonlinear devices.

When measuring a device operating nonlinearly, the VNA must account for the fact that the device is generating multiple harmonics in addition to the fundamental frequency. A critical aspect of a nonlinear VNA is its ability to measure harmonic content as well as the fundamental signal, and to provide the performance data to facilitate optimum nonlinear PA design.

Load Pull

A critical element of nonlinear analysis is load pull measurement, especially harmonic load pull, which is necessary to properly maximize the active device performance. To optimize performance of the nonlinear device, the match presented to the device must be optimized for both the fundamental and harmonic frequencies.

Design of a nonlinear power amplifier begins at the wafer level. The on-wafer device ultimately will be embedded into a 50-ohm system. However, the input impedance of on-wafer active devices is not 50 ohms and most high-power devices normally operate in the 1- to 2-ohm output impedance range. A matching network is necessary to transition the input and output of the on-wafer device to a 50-ohm system (Figure 1). Since the device is rich in harmonics, the matching network must include the optimum match for fundamental and harmonic components. Providing information on the device performance relative to the source and load impedances is the primary objective of a nonlinear VNA measurement system.

Figure 1

A unique method of providing load pull measurements for nonlinear device characterization can be seen in Figure 2. The primary difference between the system shown and traditional load pull systems is the location of the tuners and monitoring couplers.

Figure 2

Load pull analysis provides measured data of an operating device under real conditions. The measured data can be used to design a matching network for the specific device measured. Assuming the device represents typical performance for all devices on the same wafer (and subsequent wafers), the data can then be used to design the optimum network for the device product. Alternatively, the data can be exported to an EDA program, such as Microwave Office™ from AWR, to create, run, or improve models that represent the device. This is important for running simulations of multiple networks and circuits.

A complete app note on analyzing nonlinear active devices can be found here.

Simplifying IMD Measurements at mm-wave Frequencies

November 20, 2012

At microwave frequencies, intermodulation distortion (IMD) has been a valuable measurement tool for specification conformance testing, predicting system performance based on component behavior, and general device and subsystem modeling. As demanding applications proliferate at millimeter (mm-wave) frequencies, this type of characterization information is increasingly needed but has sometimes not been obtained because of difficulties in performing the measurement at those higher frequencies.

Engineers must face challenges at the higher frequencies, such as:

• Using two sources with enough power control and offset frequency control, and combined in such a way as to produce low residual IMD products that can interfere with the measurement

• A sufficiently linear receiver so that receiver IMD products do not dominate those of the DUT

• A simple, yet flexible user interface that allows an engineer to perform the needed measurement without too much complexity.

Previous broadband and mm-wave systems have had issues with these challenges, including high expense in getting the second tone and the inability to control it well; lack of adequate power control; re-modulation of sources producing spurious products; relatively poor linearity of the receivers; and user interfaces that were too limiting.

Addressing the Challenges

A basic intermodulation distortion setup consists of sources providing two tones, a combining network, and a receiver. In the VectorStar™ ME7838A system from Anritsu, small modules form the final interface for the sources and receivers so that a compact setup is possible. A second source provides the second tone and an external synthesizer under GPIB control can fulfill that role in conjunction with a third module, as shown in Fig. 1.

 

Another major item is the combining network. A variety of options are available, including Wilkinson combiners, Lange couplers, directional couplers, and resistive splitters. Any of these can work at the appropriate bandwidth, although isolation between the source ports should sometimes be considered in very high dynamic range (very low IMD product) scenarios. The usual issue is that the leveling loops of the two sources can interact and cause spurious IMD products at the combination stage but a highly buffered leveling system will eliminate the need for high isolation down to –100 dBm product levels or lower.

Sometimes additional padding at the receiver input may be needed if the DUT output power is high. Highly linear receivers lessen the level requirements yet still make it possible to drive them hard enough that residual IMD products are generated.

Calibration

In IMD measurements, there are two primary calibrations of interest: power calibrations to get the tone power accurate and receiver calibrations to accurately measure the received power.

Power calibrations use power meters connected to the reference plane of interest. The internal power settings are then dithered until the power is at the requested level at the appropriate reference plane. In the case of IMD measurements, the reference plane of interest is after the combiner and at the plane where the DUT input will be connected. When performing the internal source calibration, the external source power should be reduced to a minimum value or disconnected so that it does not interfere with the measurement, thus power calibrating the two tones individually.

Receiver calibration can be done with the uncombined signal or combined signal port after a power calibration. The purpose is to transfer a known power level to the receiver so that readings can be interpreted as dBm. This calibration should be done with the receiver tuned to the internal source frequency or in regular mode and a thru should be connected from the power calibrated plane to the desired receiver reference plane.

The foregoing addresses the setup and calibration. Once set up, it is possible to perform a number of measurements, including swept measurements of lower and upper intermodulation products as shown in figure 2. The figure is an example of a swept offset measurement for both the upper product (right) and lower product (left).

 

The techniques outlined here address the challenges associated with conducting IMD measurements at mm-wave frequencies. If you’d like more detail of the set-up, calibration, measurements and measurement uncertainties, please read our application note on the topic, which can be found here.