USB 802.11 Wireless Adaptor for Stand-Alone NI CompactDAQ and CompactRIO

This paper discusses wireless connectivity options for a Windows Embedded Standard 7 (WES7) based CompactRIO or stand-alone NI CompactDAQ system. We present the most important considerations for selecting a USB 802.11 radio and explain how to conduct in-house testing of alternative USB radios. After outlining a testing framework, we recommend two USB radios for use with such systems.

1. Introduction
2. USB Radio Practical Considerations
3. Benchmarking a USB 802.11 Adapter
4. Recommended 802.11 Adapters 
5. Conclusion 

Introduction

Historically, industrial applications have relied almost exclusively on wired connectivity for communication between systems. Consumers were quick to adopt wireless technology because it enabled flexible user access and freed the PC from being bound to a specific location. Industry was slower to leverage this new technology because of its perceived infancy, environmental ruggedness and reliability concerns, and lack of existing infrastructure.

As wireless technology has matured, it has become a more viable solution for connectivity between systems. Wi-Fi adoption in the embedded space has opened up new opportunities for solving problems in more cost-effective and modular ways than were possible with wired solutions.

Figure 1 – cRIO-9082 with a D-Link DWA-160 802.11 USB adapter for wireless connectivity

USB Radio Practical Considerations

Not all consumer- or industrial-grade USB radios perform equally. National Instruments has evaluated several USB radios and found that performance varied by over 2X for both transmission and reception throughput metrics between radios. Given the large differences in performance between different devices, a system integrator should be thorough when evaluating a particular radio. 

A full exploration of wireless network optimization and device selection criteria is beyond the scope of this paper; however, we will examine some of the top practical considerations below.

Throughput

The throughput of a USB radio is based on a wide range of factors including physical environment, ambient noise level at the frequency of transmission, number of devices on a particular channel, quality of service (QoS), antenna design,  and multiple-input/multiple-output (MIMO) support. Many of these variables can be changed to improve the throughput of a system regardless of the particular radio in use. You can reduce environmental degradation of a wireless signal by reducing the distance between nodes in the system. Additionally, you can improve throughput substantially by adjusting the mounting orientation of a single radio. A latching USB extension cable with integrated mounting (NI PN(s): 152166-02 [2.0 m], 152166-0R5 [0.5 m]) can provide up to 2.0 meters of length, enabling a system designer to adjust the orientation of a radio to better suit the environment.

Latency

Latency itself is highly variable in most wireless networks due to the uncertainty of network traffic and unpredictability of the noise in the environment, which can cause repeat packet transmissions. High levels of latency in a network decrease the maximum data transfer rate and overall performance of a system.

Quality of Service (QoS) 

In networks with higher traffic on available channels, QoS settings often must be adjusted to ensure that higher-priority data transmission requests take precedence over others. One type of application that often warrants such priority adjustment is voice over IP (VoIP); however, QoS modifications are equally applicable to high-throughput applications. An administrator can use policy-based settings on a WES7 system to assign higher network priority to selected applications.

Not all USB radio manufacturers document QoS support, so if the application requires QoS support, the designer or integrator may need to do some testing before system integration. The 802.11 USB radios recommended at the end of this paper both support QoS.

Wireless Standard Support

Many modern USB radios support multiple standards including 802.11a, 802.11b, 802.11g, and 802.11n. 802.11n-enabled devices can support MIMO transfer and the higher 5 GHz frequency band, increasing throughput and improving immunity from environmental sources of 2.4 GHz noise. Be careful in selecting an operating frequency band. 5 GHz is generally less populated and therefore less susceptible to interference. However, higher-frequency signals are more easily attenuated, making them less suitable for some environments.

Security

When IEEE first published the 802.11 standard, Wired Equivalent Privacy (WEP) provided basic security for wireless networks. Shortly after the introduction of WEP, serious security flaws became apparent, and IEEE decided to adopt a new, upgraded standard to address those flaws. The upgraded standard was Wi-Fi Protected Access (WPA), which added several new security features.

WPA was later succeeded by WPA2, which added still more security improvements. Among several new features introduced was the Advanced Encryption Standard (AES).

For applications that involve transfer of sensitive data across a wireless network, NI highly recommends that the system integrator select a USB radio that supports WPA2-PSK and/or a WPA2-EAP security enabled device. National Instruments does not recommend WEP-based security for safeguarding a network.

Availability 

Depending on the country of purchase, there may be regulatory and procurement considerations that make it more challenging to obtain a given radio for system integration. If the recommended USB radios from National Instruments are not available in your region, the selection and testing of a new radio can be aided by the selection criteria and testing procedures found in this paper.

Benchmarking a USB 802.11 Adapter

The following section outlines a Windows GUI-based tool, JPerf, which system integrators can use to predict wireless network performance.

GUI Based Network Performance Test Procedure (Recommended for Most Users)

1. Set up a wireless access point for the first node in the test system, or use existing wireless infrastructure to connect two nodes.
2. Download and install the required drivers for the USB radio on the cRIO-908x or cDAQ-913x.
3. Download and install the latest Java Runtime Engine (JRE) on both the test system and the cRIO-908x or cDAQ-913x.
4. Download and unzip xjperf on both test nodes.
5. In the unzipped jperf-2.0.x folder, double-click the jperf.bat file, which opens the jperf network performance graphical tool shown in the graphs below.
6. For both systems, run the command prompt (All Programs -> Accessories -> Command Prompt) and use the ipconfig command to get the IP address for each system. Ping each system from the other system.
7. Once both nodes have appropriately configured IP addresses, use JPerf to select one node to be the client that points to the IP address of the server (the other node).
8. Configure the other system in JPerf to be the server and select the Application and Transport Layer options that are appropriate for your application (default settings are fine for baseline testing).
9. Click Run Iperf on both the client and server nodes and watch the bandwidth plot update with new values.
10. Repeat the test with varying distances between the wireless nodes and with different orientations of the USB radio in multiple axes. Some recommended distances for testing are 0.5, 3, 10, and 30 meters. At a minimum, the USB radio should be tested in three perpendicular orientations to better orient the adapter antenna for improved throughput.

Recommended 802.11 Adapters 

Based on the selection criteria addressed above in USB 802.11 Adapter Practical Considerations, National Instruments recommends two USB radios, the Edimax 7811UN and D-Link DWA-160, which were the best overall performers of devices that were tested. Both of these devices can be purchased in North America and much of Europe with reasonable ease.

The testing results averaged throughput relative to node distance for the two recommended radios is referenced in the table below. Aggregate testing results were performed at a frequency band of 2.4 GHz in an environment with minimal external interference and line of site visibility between the access point and USB radio.

Conclusion 

While there is no shortage of USB radios available on the market, they differ dramatically in performance and features. Therefore, system designers should take care in selecting a USB radio.

Additionally, wireless network design and implementation are especially important for embedded applications because of the variations that environmental factors can cause. The channel noise level, physical orientation, and distance between two communicating nodes can have large effects on the throughput of a system.

1 Ratings | 5.00 out of 5

 Print

Reader Comments | Submit a comment »