Configuring a Software-Defined WLAN Test System

1. Introduction to WLAN physical layer

WLAN standards are defined and maintained by the IEEE 802.11 working group, which includes chip manufacturers and access point manufacturers. This group has defined several 802.11 standards – from 802.11a to 802.11z. However, the most common protocols for WLAN devices are IEEE 802.11a, b, g, and n.

In 1999, the working group defined the 802.11a and 802.11b standards for WLAN. The IEEE 802.11a standard, which is set to the 5 GHz unlicensed Industrial, Scientific and Medical (ISM) band, can achieve a maximum data rate of 54 Mb/s. In contrast, the IEEE 802.11b standard can achieve a maximum information rate of 11 Mb/s on the 2.4 GHz ISM band. IEEE 802.11g, released in 2003, can also achieve a data rate of up to 54 Mb/s in the 2.4 GHz ISM band. IEEE 802.11n is the latest version, which integrates functions such as multiple input/output (MIMO) and parallel channels, and can achieve a data rate of 300 Mb/s in the 2.4 and 5 GHz bands. The

two basic transmission architectures used by WLAN are direct sequence spread spectrum (DSSS) and orthogonal frequency division multiplexing (OFDM). In addition, its inherent modulation architecture includes CCK, as well as orthogonal architectures such as BPSK and 64-QAM. Table 1 shows the standards that use specific transmission architectures and modulation types.

Table 1. Transmission architectures and modulation types used by various 802.11 versions

 
Unlike the OFDM architecture standards of WiMAX (IEEE 802.16d/e) and 3GPP Long Term Evolution (LTE), all subcarriers in WLAN OFDM signals use the same modulation architecture. Therefore, for IEEE 802.11a/g signals, the modulation architecture can directly affect the maximum transmission rate and the coding rate of a specific signal. Table 2 presents this relationship.

Table 2. Data transmission rate, coding rate, and burst interval time

In Table 2, a high signaling rate such as 54 Mb/s requires the use of a high bit modulation architecture such as 64-QAM. Furthermore, the standard burst interval of 1024 data bits will be significantly higher than a lower bit modulation architecture. When increasing the measurement speed of a test system, it is important to understand the relationship between longer burst intervals and longer measurement times. Generally speaking, when performing error vector magnitude (EVM) measurements on a single burst, the measurement speed can be increased if the instrument can be set to acquire only the required measurement data. For example, when measuring a 64-QAM burst, if the acquisition time length is set to 200 µs, the measurement speed can be higher than that of a time length of 10 ms or more.

2. Overview of RF Virtual Instruments

With NI software-defined WLAN testing, you can choose from a variety of instruments to test WLAN devices. This white paper provides an overview of the architecture of a virtual PXI measurement system to explain the differences between traditional instruments and virtual instruments.

PXI instruments integrate high-performance multi-core controllers, high-speed PCI/PCI Express data buses, and optimized measurement algorithms to achieve industry-leading measurement speeds. The software used for WLAN measurements is the NI WLAN Measurement Suite, which includes the NI WLAN Analysis and WLAN Generation Toolkits. The recommended NI hardware is the NI PXIe-5663 vector signal analyzer and the NI PXIe-5673 vector signal generator. The NI PXIe-5663 can analyze signals from 10 MHz to 6.6 GHz with a maximum instantaneous bandwidth of 50 MHz. The NI PXIe-5673 can generate signals from 85 MHz to 6.6 GHz with a maximum instantaneous bandwidth of 100 MHz. Either set of instruments can be paired with other generators or analyzers to perform phase-coherent measurements. Figure 1 shows a common WLAN device test system setup with a vector signal generator and a vector signal analyzer.

Figure 1. PXI system for WLAN measurements.

Software-defined instruments are particularly suitable for automated test applications. Architecturally, the main difference between PXI modular instruments and traditional instruments is their processing core. Although the two systems use many similar components, the main difference is that PXI systems can use high-performance multi-core central processing units (CPUs). Figure 2 shows traditional and PXI instruments with many of the same core components, including memory, high dynamic range analog-to-digital converters, and high-performance RF front ends.

Figure 2. A user-defined CPU is an essential component of PXI RF instrumentation.

The multi-core CPUs of PXI modular instruments provide excellent signal processing capabilities. As a result, many PXI-based measurement systems have become significantly faster than traditional instruments. In general, CPU performance from chip manufacturers such as Intel and AMD continues to grow in line with Moore's Law. Therefore, when manufacturers release new processors, users only need to upgrade the controller of the PXI system. For existing test systems, measurement speed can be greatly improved at a fraction of the cost of the components. The

second advantage of software-defined instruments is that multiple wireless standards can be tested on a single hardware platform. This advantage is particularly applicable to multi-standard consumer products or system-on-a-chip devices. In the past, engineers had to purchase several dedicated instruments when the device under test included a GPS receiver, a WLAN radio, and an FM radio. With software-defined instruments, all standards can be tested by simply integrating common hardware and using a dedicated software toolkit. Figure 3 shows this concept.

Figure 3. Software-defined instrument architecture

As shown in Figure 3, you can use a common RF front end (either generator or analyzer) with a Windows-based CPU to create a software-defined instrument. With NI’s software-defined RF instruments, you can test WLAN, GPS, GSM/EDGE/WCDMA, WiMAXTM, BluetoothTM, DVB-T/ATSC/ISDB-T, FM/RDS/IBOC, and other wireless standards. 

3. Introduction to NI WLAN Measurement Suite

The software-defined features of existing PXI instruments, such as the combination of the NI WLAN Measurement Suite and related software, are essential components of the measurement system. The WLAN Measurement Suite includes the NI WLAN Creation Toolkit and the NI WLAN Analysis Toolkit. Both toolkits include APIs for LabVIEW, LabWindows™/CVI, and ANSI C/C++, and can be used with PXI RF vector signal generators and analyzers. For high-level operations, the WLAN Creation Toolkit can be used to create IEEE 802.11a/b/g signals. The WLAN Analysis Toolkit can further provide measurement results based on the signals acquired by the vector signal analyzer. Figure 4 is a flowchart of this measurement method.

Figure 4. WLAN test system architecture

Whether using the property node or the programming API, you can set special standards, data rates, burst intervals, carrier frequencies, etc. Figures 5 and 6 show how to adjust common settings using the property node or the programming API.

Figure 5. Setting up WLAN measurements with LabVIEW Property Nodes

Figure 6. Setting up WLAN measurements with the LabVIEW programming API

Figure 6a. Setting up WLAN measurements with the LabWindows™/CVI programming API

The introductory example program is designed for automated measurement applications. If you want to perform more interactive measurements, you can use the LabVIEW or LabWindows™/CVI display panel shown in Figure 7.

Figure 7. LabVIEW display panel for WLAN measurement

Figure 7 shows the basic 802.11g spectrum mask in the frequency domain. Please note that all measurements described in the following sections are performed using this example. 

4. Common WLAN measurements

When characterizing any WLAN component or radio, the specific measurements required often depend on the device under test. For example, to characterize a power amplifier, EVM and third-order intermodulation (IM3) measurements must be combined to characterize nonlinearities. However, carrier offset measurements are less important because they are a function of the RF signal generator. Table 3 lists some common WLAN measurements. As shown in Table 3, many measurements can be performed using the WLAN Analysis Toolkit if the associated auxiliary measurements are mentioned in the following sections.

Table 3. Measurements performed with the WLAN Analysis Toolkit

5. Transmission power

One of the key aspects of WLAN measurements is the transmit power. There are many ways to measure power, and different power measurements require different WLAN standards. When characterizing 802.11a/g transmitters, WLAN measurement systems can generate both peak and average power results. For 802.11b devices, common measurement systems also provide the number of rise and fall times for power. Note that while peak power meters are effective tools for power measurements, RF vector signal analyzers are still the fastest for measuring the average power of a signal. Average power meters only measure power when the transmitter is set to output a continuously modulated carrier.

When measuring power with an RF vector signal analyzer, the result is calculated over the triggered burst. This allows the average power to be measured over a complete burst or a specific portion of a burst. The WLAN Analysis Toolkit allows you to set up gated power measurements where the average power is measured relative to user-defined start and stop times. In addition, the toolkit can be used to return a power versus time trace of an IEEE 802.11a/g signal, as shown in Figure 8.

Figure 8. Training sequence, channel estimate, and data in power versus time trace.

               
The power versus time trace in Figure 8 is often used as a debugging tool to ensure that all parts of the burst – from the training sequence to the OFDM symbols – are being transmitted.

6. Error Vector Magnitude

EVM is one of the most important measurements because it finds errors caused by multiple impairments, including quadrature offset, IQ gain imbalance, phase noise, and nonlinear distortion. For a modulated signal, EVM compares the expected and actual phase/strength of the signal. As shown in Figure 9, the NI WLAN Analysis Toolkit multiplies the error vector |E| by the strength vector |V| to get this value.

 
Figure 9. Graphical representation of EVM Measurement

Generally, users can specify EVM units in percentage (%) or decibel (dB). However, IEEE 802.11a/g measures EVM in decibels, while IEEE 802.11b measures EVM in percentage. Equation 1 shows how to convert between these two units.

Equation 1. Decibel to Percent Conversion

 
For example, 1% EVM is equal to -40 dB, and 5% EVM is equal to -26 dB. When measuring the EVM of a complete burst, instruments often present the EVM result as root mean square (RMS). For OFDM signals, the EVM is taken across all subcarriers and symbols as an RMS result. For DSSS signals, the RMS is taken across all slices.

In many cases, almost all EVM performance can be viewed through a constellation diagram. The constellation diagram shows the phase and strength of each symbol, allowing the user to identify specific orthogonal impairments. Figure 10 shows a constellation diagram for 64-QAM.

Figure 10. Graphical representation of EVM Measurement

As shown in Figure 10, -46 dB EVM is equal to 0.5%. An NI PXIe-5673 RF vector signal generator and an NI PXIe-5663 RF vector signal analyzer were used and set to loopback mode. Both instruments were set to a center frequency of 2.412 GHz and an RF power level of -10 dBm. Therefore, under these settings, both instruments achieved the same -46 dB EVM. Also note that the WLAN Analysis Toolkit in Figure 10 performs all of the time domain measurements in parallel. The composite measurement provides EVM, carrier offset, and carrier leakage, as well as quadrature impairments such as IQ gain imbalance and quadrature skew.

7. Spectral mask measurement

The spectrum mask can describe the nonlinear characteristics of the transmitter. Generally speaking, the spectrum diagram can be used as a diagnostic tool to determine whether the signal under analysis is distorted. Since the spectrum mask measurement is a pass/fail test, its result constitutes the spectrum mask margin; this margin is in dB, which is the power difference between the actual signal measured and the mask. Figure 11 is a spectrum mask measurement of an 802.11b signal.

Figure 11. Spectral mask of an 802.11b signal.

IEEE 802.11b signals and IEEE 802.11a/g signals actually use different spectrum masks. Figure 12 shows the mask of OFDM 802.11a/g signals.

Figure 12. Spectral mask of 802.11a/g signal

Note that the spectrum mask can also describe a variety of signal characteristics. For example, nonlinear characteristics of the transmitter can cause the signal sidebands to reach the limits of the mask. In addition, improperly set sideband signals can also form unwanted sidebands on the DFDM signal.

8. Conclusion

As described in this article, users can set up a variety of WLAN measurements through software toolkits. In fact, the WLAN measurement suite provides generation and analysis capabilities for IEEE 802.11a/b/g measurements. PXI RF vector signal generators and analyzers can be set up through application programming environments such as LabVIEW, LabWindows/CVI, and even .NET to quickly and easily test WLAN products. Although these software-defined instruments can test WLAN and many other wireless standards, one of the main advantages of this approach is its test speed.