Time domain's role in digital RF world
19 March 2008
Many of today’s RF signals change from one instant to the next. Some hop frequencies, others spike briefly and then disappear.
Many can have complex modulation that can dynamically change in an instant. All, however, prove time cannot be ignored. These activities can also produce their own side effects; random transients, interference and switching anomalies. Time is the common factor in these phenomena. Time is the axis that can no longer be ignored.
The digital RF revolution has introduced an unprecedented number of useful devices while lowering their cost and power consumption. Whole communications systems are integrated in monolithic silicon. The proliferation of information being transmitted over an increasingly scarce spectrum has driven a need for higher data rates per unit bandwidth and for complex communications protocols that allow the peaceful co-existence between RF devices and systems.
Radar model
A key goal of communications protocols is to reliably transmit a packet of data in as little bandwidth as possible over as short a time as possible while minimising interference.
While not designed for communications, radar systems have similar objectives of spectral efficiency and minimum interference with the added goals of security and detection avoidance.
This has led to several classes of RF transmissions, including transmissions that only turn on for the brief time it takes to transmit a unit of data, releasing the spectrum for other use as soon as the data is sent. In many cases, the timing of these short transmissions is both unknown and random.
Other classes of RF transmission are systems that share the same spectrum at the same time as in UWB (ultra-wide-band) and CDMA (code division multiple access). In addition, there are cognitive radios that adjust frequency, modulation and power in response to the spectrum environment at a particular point in time and location and can build upon prior knowledge, and multiple RF devices inside a single shared package. A further class includes RF devices that share the same silicon as digital CPUs with clock rates in the GHz range.
While not designed for communications, radar systems have similar objectives of spectral efficiency and minimum interference with the added goals of security and detection avoidance.
This has led to several classes of RF transmissions, including transmissions that only turn on for the brief time it takes to transmit a unit of data, releasing the spectrum for other use as soon as the data is sent. In many cases, the timing of these short transmissions is both unknown and random.
Other classes of RF transmission are systems that share the same spectrum at the same time as in UWB (ultra-wide-band) and CDMA (code division multiple access). In addition, there are cognitive radios that adjust frequency, modulation and power in response to the spectrum environment at a particular point in time and location and can build upon prior knowledge, and multiple RF devices inside a single shared package. A further class includes RF devices that share the same silicon as digital CPUs with clock rates in the GHz range.
RF measurement challenges
Basic measurement tasks required of RF tools can be shared across different projects. In one form or another, many tasks are common in the various classes of RF transmissions and span applications, ranging from surveillance to physics research. In characterising frequency drift, frequency settling time and response must be characterised to ensure that a device meets functional and operational needs. This requires uninterrupted capture of a signal whose frequency is constantly changing over time.
For detecting interference signals and their sources, it is important to remember that interference signals come and go, often as a result of switching activities inside or outside the system from intentional or unintentional sources. By recording many discrete instances of interference plus the surrounding time, it is possible to localise the offending frequency and infer its source.
When finding and analysing transient signals, remember that transient frequency changes, and whether glitches or intentional transmissions, they can appear unexpectedly amid steadier and even higher level signals. Detection requires a method of distinguishing events of interest from everything else that is going on the observed span.
Up-converting signals
In capturing and analysing channelised signals, baseband signals are often upconverted to a specific channelised band of operation called a passband. These passband signals can be agile and modulated, making it necessary to capture everything that occurs in the frequency band of interest over a period of time. An uninterrupted spectral record is required so the spectrum, time, and modulation characteristics of the signal can be explored.
Adaptive digital modulation is growing increasingly common and complex as bandwidth becomes more precious and security more important. Analysing modulation quality and its relationship to the signal's frequency and time-domain characteristics is a key step in wireless troubleshooting during transitions. Often, testing is required to go beyond standards, especially as implementation is not defined.
It is clear that many emerging applications call for a highly capable RF analysis solution; a tool that captures the dimension of time along with the traditional frequency and amplitude axes.
Currently, there are three types of RF signal analysers available. These are the swept spectrum analyser (SA), the vector signal analyser (VSA), and the real-time spectrum analyser (RTSA).
The traditional SA makes amplitude versus frequency measurements by sweeping a resolution bandwidth (RBW) filter over the frequencies of interest. The disadvantage is that it only records the amplitude data in one frequency at a time. A relatively stable, unchanging input signal is required.
The VSA emerged to address the distinctive requirements of digitally modulated signals and unlike the swept SA, the VSA is optimised for modulation measurements. It captures the whole signal and any digital modulation effects occurring at one instant in time, and stores a record into memory. This process is also a limitation, as the capture cannot be triggered by a frequency-defined event.
As time-varying signals become more common in RF applications, the need for an alternative approach to RF acquisition and analysis becomes more urgent. The RTSA has emerged to solve this tough measurement problem by triggering on a frequency domain event, and then capturing and analysing any passband signal that falls within its real-time bandwidth. Indeed, the RTSA’s digital IF architecture allows a continuous capture of a signal over time. Once triggered on a frequency-defined event, the signal is then stored in the memory as a seamless, continuous record of time. Furthermore, the RTSA is the only RF signal analyser that is optimised to produce a three-dimensional display: frequency, power (amplitude), and time.
Shedding light on problems
Recognising that a problem exists is the first step to resolving it. Problems specific to digital RF are often discovered indirectly. Transients in one device may cause increased bit error rates in another. Radars may occasionally provide inaccurate target information due to self-jamming or susceptibilty to transient interference. Thermal or electrical memory effects in a power amplifier may cause lost data and momentary interference with adjacent channels. The execution of a computationally intensive software sub-routine may cause power supply voltage variations that affect the quality of RF transmissions.
To enable RF designers to discover problems, Tektronix has developed digital phosphor, DPX technology, to emulate variable persistence CRTs. The implementation of digital phosphor spectrum analysis used combines display processing with dedicated DSP hardware that perform frequency transforms several orders of magnitude faster than traditional spectrum analysers. At almost 50,000 trace updates/sec per second, the DPX technology has about 100,000 per cent more spectrum updates than a traditional swept spectrum analyser.
The information from each transform is combined in the DPX engine generating displays at a full motion rate. This engine includes statistical persistence processing that allows full-motion viewing of signal behaviour over time. It also makes weak signals interspersed amid strong ones instantly visible, and highlights infrequent short-duration events. Persistence adjustments allow the user to optimise display characteristics for varying signal conditions, from a live RF view of dynamic signals to the discovery of single occurrences. This sheds light on signal behaviour that was not viewable with traditional spectrum analysers or vector signal analysers. Other RTSA features provide means to trigger on signal behaviour, capture those signals in memory and analyse them in the time, and frequency and modulation domains
As RF signals become more complex and less predictable, designers need to understand time-varying signal behaviour, ranging from frequency hopping to EMI transients. While a number of instruments are available for RF measurements, only the RTSA offers the triggering, capture and analysis features needed for emerging trends in RF design.
JONATHAN MEES is European market development manager, RF, Tektronix
Contact Details and Archive...
Related Articles...
Most Viewed Articles...