RF Record and Playback Application

This paper describes the technologies that enable RF recording and playback, as well as practical considerations for streaming RF systems.

1. Introduction
2. Application
3. Hardware
4. Software
5. Systems
6. Related Links

Introduction

Numerous applications cannot sustain a sufficient sampling rate for lengthy acquisitions or generations. In these situations, you must compromise by using a slow enough rate for data to be transferred over the bus, or by sampling at the necessary high speeds for the short periods of time that onboard instrument memory allows. Neither sacrifice is desirable. High-speed streaming addresses this need by helping you transform data to or from an instrument at a rate high enough to sustain continuous acquisition or generation. High-speed data streaming of RF signals is a common application because data collection is often desired at high rates for an extended period of time. You can accomplish this with a bus that has sufficient bandwidth for overall data throughput and a system that allows you to store the entire acquisition or generation waveform.

Application

With the growing adoption of wireless technology and an increasing number of protocols competing for limited bandwidth, traditional RF instrumentation is not able to capture the intricacies of today’s wireless transmissions. As channels become more heavily used, it is important that they remain free of transmitters that interfere with legitimate communications. Additionally, it is becoming common to test wireless products in “real-world” environments to ensure performance. These factors motivate the use of test equipment that can stream RF signals for an extended period of time.

Adding a high-speed data bus and a large amount of storage removes the memory limitations of traditional instruments to enable continuous RF acquisition. Such systems are able to acquire and measure longer transmissions, also reducing the need for tightly synchronized signaling and triggering. Long playback times are essential for improving the statistical significance of many communications measurements and displays including bit error rate, CCDF, trellis plots, and constellation plots. Postprocessing the acquired data allows custom measurements, which are either not available on the instrument or impossible in real time. This means it is possible to reveal transients that would have been either not displayed or never acquired on instruments with a limited amount of memory.

Locating unauthorized transmissions on an RF channel or otherwise monitoring that channel for all traffic requires a continuous acquisition and retention of the channel’s entire bandwidth. This application is commonly referred to as signal intelligence, or SIGINT. Streaming is an obvious choice for this type of application. In some cases, you also may need to reproduce the traffic on this channel by generating the acquired data.

Consumer demand for wireless technology is creating a growing number of RF products that must be manufactured and tested. Though you can individually test subcomponents of these products to meet certain specifications, testing integration in a complex system requires significantly more complicated test patterns. Rather than manually generating these patterns, it is far simpler and more realistic to record a real-world signal and play it back to test the device. This also allows for the introduction of natural impairments. For example, rather than simulate multipath effects on signals designed to test terrestrial or satellite radio receivers, recording the signal in a known “dead” area in a major city (where multipath is a common problem) can generate a much better assessment of the product’s real-world performance. This process is known as channel emulation.

Because of their commonalities, extended-duration acquisitions, spectral monitoring, and channel emulation are collectively referred to as RF Recording and Playback throughout this paper. There are other applications that fall under this umbrella, but the aforementioned examples benefit most from streaming technology. The remainder of this document describes the National Instruments hardware and software that you can use to create RF recording and playback systems.

Hardware

Enabling Technologies

NI RF hardware features several technologies to efficiently stream data. The first is the onboard signal processing (OSP) functionality of NI vector signal generators (VSGs) and vector signal analyzers (VSAs). Regarding high-speed streaming, the main OSP function is data reduction. The design of NI RF hardware requires you to use a certain intermediate frequency (IF) before analog upconversion or after analog downconversion. Rather than use host processor resources to upconvert or downconvert digital representations of analog waveforms to IF, you can implement this operation on field-programmable gate arrays (FPGAs) located on NI VSGs and VSAs. This not only reduces the processing burden on the host CPU but also significantly decreases the amount of raw data that you must transfer to or from these modular instruments. For instance, generating an RF signal with only 100 kHz of bandwidth still requires you to digitally upconvert the 100 kHz signal to an IF of 25 MHz before you can further upconvert it to the desired radio frequency. This process is called digital upconversion (DUC). Performing DUC on the host processor requires you to transfer more than 50 MS/s of data to the VSG. Moving this conversion to VSG hardware requires only 200 kS/s of data throughput, which is one of the advantages this process presents to streaming applications. The reverse of this process, known as digital downconversion (DDC), also reduces the waveform transfer rate from a VSA to its host controller in much the same way. For both of these cases, the transferred data is 100 kS/s in-phase (I) data and 100 kS/s quadrature (Q) data. For more information on I and Q data, please see the “What Is I/Q Data?” tutorial.

In addition to data reduction, you can use OSP to perform interpolation, decimation, amplitude modulation, function generation, and the addition of frequency conversion impairments. Learn more about OSP specific to NI signal generators and the benefits of using DDCs on signal analyzers.

Data Formats

Another important aspect of RF streaming is the digital data format you use to represent an analog signal. NI VSGs and VSAs use a process called quadrature digital upconversion or downconversion, respectively, which facilitates the conversion process. Instead of representing an analog signal with only real data, quadrature digital conversion uses both real and imaginary data in the form of I and Q pairs. Because the sample rates of the resulting I and Q data are each half the sample rate of the baseband, time-domain signal, the resulting data rate is the same. It is important to note that you must perform the conversion process from time-domain, baseband data to IQ data on the host processor and not on the instrument.

If an application requires direct control of an IF signal, you can disable the DDC or DUC to greatly increase data transmission requirements, which limits streaming performance. Instead of sending baseband I and Q data to or from the instrument, this process requires purely real-time domain data at a rate more than twice the intermediate frequency. If possible, it is best to allow the DUC and DDC to generate and receive IF signals, freeing streaming resources.

NI Hardware for Recording (Acquisition)

For streaming acquisitions, National Instruments recommends the NI PXI-5661 vector signal analyzer with digital downconversion. It combines a 2.7 GHz analog downconverter with a 14-bit 100 MS/s digitizer to provide 20 MHz of real-time bandwidth, which you can stream to disk at a rate of 100 MB/s. Without a DDC on the digitizer, using an IF of 15 MHz, streaming from the digitizer to disk would require more than 120 MB/s data transfer rates, which approaches the practical limit of the PXI bus.

NI Hardware for Playback (Generation)

For streaming generation, National Instruments recommends the NI PXI-5671 vector signal generator with digital upconversion. It combines a 16-bit 100 MS/s arbitrary waveform generator with a 2.7 GHz analog upconverter. With the DUC on the signal generator, this pairing provides 6.6 MHz of real-time streaming bandwidth, or streaming from disk at a rate of 33.3 MB/s. By disabling the DUC, using an IF of 25 MHz would require more than a 140 MB/s data transfer rate, which is above that of the PXI bus.

Software

LabVIEW for RF Streaming

National Instruments LabVIEW is an ideal programming language for streaming applications because it features intuitive support of file I/O and efficient design patterns as well as inherent parallelism. Because of the time-critical nature of communicating with instruments and reading from or writing this data to disk at high rates, it is not possible to use a traditional single-loop design pattern. The limitation of this type of architecture is that instrument and disk access must occur in lockstep. If either experiences any type of communication delay, it negatively impacts the performance of the other. By using a producer/consumer design pattern and eliminating the sequential, alternating access of instrument and disk, each can operate at its own rate, with NI LabVIEW providing buffer overflow or underflow protection through the use of additional buffering in host PC memory. LabVIEW makes this process incredibly easy through its native support of queue operations. Additionally, with its default ability to create multiple threads, LabVIEW can independently execute producer and consumer loops. With multiprocessor computers and multicore processors, streaming experiences a significant performance advantage when using this type of program architecture.

Specific to RF streaming, LabVIEW offers the NI Modulation Toolkit as well as the NI Spectral Measurements Toolkit. These libraries of common RF communication and measurement functions offer much shorter development times. Learn more about the Modulation Toolkit and the Spectral Measurements Toolkit.

Systems

There are several possible configurations that enable RF streaming systems. They are outlined in the “Streaming Options for PXI” tutorial. In the case of signal analysis, any system that has the ability to simultaneously read data at a rate of 100 MB/s from a digitizer and write this data to a hard drive at that same rate suffices. This precludes systems that use a single PCI or PXI bus for data transfer in both directions (instrument to host memory, host memory to disk) because the aggregate rate would be above the PCI bus bandwidth limit of 133 MB/s. For signal generation, any system with the ability to stream data from disk at a rate of 33.3 MB/s has the performance to continuously generate 6.6 MHz of real-time bandwidth. This rate is just beyond the abilities of National Instruments embedded controllers, necessitating an external hard disk solution. For simultaneous acquisition and generation at 6.6 MHz, you need 66.6 MB/s total bus bandwidth.

Related Links

RF Homepage
What Is I/Q Data?
Onboard Signal Processing (OSP) on National Instruments Signal Generators
Use DDCs for High-Speed RF/IF Streaming
Streaming Options for PXI
NI PXI-5661
NI PXI-5671
NI Modulation Toolkit
NI Spectral Measurements Toolkit

0 Ratings | 0.00 out of 5

Read in  English French |  Print |  PDF

Reader Comments | Submit a comment »