Emerging wireless standards bear many implementation challenges. A technology like Wideband Code Division Multiple Access (W-CDMA), for example, is new territory for many designers. Its signals are wider in bandwidth and they are more complex in their modulation and access methods. In addition, W-CDMA signals utilize new handover protocols.
To successfully address such complexities, it is critical to select measurement tools that employ specific techniques for these new standards. Traditional measurement equipment includes radio-frequency (RF) signal generators, base-transceiver-station (BTS) emulators, digital modulation analyzers, and spectrum analyzers. This equipment is chosen with performance, cost, and ease of use in mind. With W-CDMA, engineers are clearly at the beginning of the technology life cycle. The choices that they make now can impact long-term business realities for their products.
To meet these demands, emerging tools are using real-time spectrum analysis. These tools help engineers characterize, troubleshoot, and verify new designs under tight time-to-market deadlines. In addition, real-time approaches to analysis are enabling engineers to speed up common measurements and analysis requirements. By offering increased flexibility, they also help engineers find faults.
This article looks at some of the latest toolset advances employing real-time, multidomain analysis. These improvements can help designers develop network elements and W-CDMA user equipment (particularly for mobile stations). The application examples illustrate such features in action, including the acquisition and analysis of a W-CDMA call-setup procedure and a cell-to-cell handover.
Complex W-CDMA signals call for tools that simultaneously look at larger time segments and more frequency ranges. Changes can occur quickly in the signal. Often, they are changes that designers can't afford to miss, such as PRACH signals and handover protocols.
The traditional solution for RF analysis uses a swept spectrum analyzer. This method accumulates a spectrum display by tuning the path to successive frequency steps. It captures the respective amplitudes at each step. The displayed result represents compiled single-frequency acquisitions that were performed sequentially. A steady, uninterrupted signal is required to produce a meaningful measurement, however. For instance, say a transient event—intentional or otherwise—occurs in the seventh increment after the analyzer has moved on to the eighth increment. The event is then lost.
New spectrum-analysis techniques take an entirely different approach. Rather than acquiring one frequency step at a time, they capture a block of frequencies all at once during a user-specified time frame. These frames repeat continuously with a full acquisition every frame. Instead of waiting for each discrete frequency step to be measured, the instrument constantly samples the full frames. As a result, events that occur anywhere in the spectrum are far more likely to be captured.
Figure 1 shows the real-time spectrum-analysis architectural concept. Instead of having one resolution filter like a swept analyzer, it boasts hundreds of filters in parallel. All of the hundreds of frequency components within the real-time bandwidth are filtered concurrently. They appear on-screen simultaneously. To arrive at the spectrum, a wireless communication analyzer uses fast Fourier transform (FFT) techniques instead of parallel hardware filters. This enabling technology depends on very fast and accurate FFT hardware.
This "all-at-once" capture methodology lies at the heart of real-time spectrum analysis. By using this feature with a vast acquisition memory that is supported by innovative triggering tools, instruments can build up a detailed, deep record. This record is the basis for the new display modes that reveal multidimensional signal details. Such details are shown in a graphically intuitive manner. With codograms, spectrograms, and displays that compile the results of numerous acquisitions, designers can quickly characterize signals, events, and frequencies of interest.
How can these analysis techniques help an engineer who is under tremendous time pressure to complete a project, such as a new prototype handset? One answer lies in the use of dual analyzers. Both analyzers are equipped with real-time modulation and multidomain analysis capabilities. They can be combined to monitor both uplink and downlink simultaneously and in perfect synchronization.
VISUALIZATIONS AID VALIDATION Consider the example of a prototype project in more detail. To prepare for the measurements, the handset is placed in a test stand that connects to a transmission/reception source. This source simulates a BTS. Intervening between the BTS and the handset is a pair of wireless communication analyzers. When the test commences, the apparatus should:- Capture the full W-CDMA call-setup procedure.
- Assist in the interpretation of results in the frequency domain, time domain, code domain, and modulation domain.
The pseudo-BTS acts as a "known-good" transmitter. It receives the call data via the RF link. The wireless communication analyzer then captures the result.
To outline the steps above, realize that a W-CDMA time slot is 666.7 ms. A full sequence lasts 720 ms or one super frame. Using conventional spectrum analyzers, it is impossible to acquire a complete W-CDMA call-setup procedure in one sweep.
In contrast, newer wireless communication analyzers can capture up to 10 seconds of data at a 5-MHz span. In one acquisition, they can capture a total frequency range of up to 15 MHz. The 10 second time interval is sufficient to capture a full W-CDMA super frame with ample pre- and post-trigger time. It is faster and easier to evaluate several super frames that are captured at once than it is to look at a sequence's isolated segments. Taking in the full call-setup procedure allows the engineer to see both spurious events and the intended events that are part of the protocol. Both event types can be viewed in the context of the whole call transaction.
If a full call procedure is captured in one acquisition, all user results can be correlated. This aspect diminishes the need to take multiple measurements, which could introduce uncertainties. Moreover, zooming in on sections of the captured data allows the user to make measurements without having to recapture the signal. When there are hundreds of individual tests to perform, this capability can significantly reduce development time.
The spectrogram and other graphical representations provide the impetus to acquire a deep, full-frequency sweep of the W-CDMA frame. The spectrogram provides at-a-glance details about signal characteristics. It also offers ways to "drill down" into the signal for even more information.
The spectrogram contains three dimensions of information: frequency (the X axis); time (the Y axis); and power. The power axis is expressed by color. If there are any signal deviations, they show up as tell-tale gaps, transients, and color (power) variations. More than any other display format, the spectrogram can reveal the nature of W-CDMA and other spread-spectrum signals. Figure 2 depicts an actual spectrogram screenshot of a PRACH preamble, as well as a PRACH message part.
The spectrogram can provide a huge volume of additional data about the signal. After just one acquisition, this information will reside in the analyzer's memory. Aside from the spectrogram itself, it is relatively easy to generate a myriad of displays for any time-slice on the spectrogram. Such displays include constellation, time versus amplitude, frequency versus amplitude, phase domain, code-domain power, and modulation accuracy. For designers of multi-domain systems who need a variety of information types, these displays offer an improved view (FIG. 3).
NEW TECHNIQUES ARE CRITICAL Whether they are 2G, 2.5G, or 3G, all wireless services must provide for efficient, error-free handovers of calls in progress. These handovers occur as mobile subscribers move from one cell to the next. The handover mechanism involves detection and interaction between the User Equipment (UE)— any device allowing a user acess to network services—and BTS for two adjacent cells, respectively. In a W-CDMA network, these two elements manage the handover process by sending signals in a "compressed," rather than a normal, transmission mode.
The compressed-mode signal has its own class of unpredictable errors. Designers must analyze these errors in the frequency, time, and code domains. Properly functioning handovers allow calls to continue without interruption. At the same time, they maximize resources through capacity sharing.
Poorly executed handovers, on the other hand, can disrupt calls. The net effect is degradation of quality of service (QoS) and a reduction in the cell sites' effective range. Handover problems can impact both revenues and maintenance costs. Of particular concern are the handovers between W-CDMA and Global Systems for Mobile Communications (GSM) or time-division-duplex (TDD) systems. This issue will become especially prominent as 3G network elements get integrated into existing network infrastructures.
The task of analyzing handover behavior is ready-made for a tool that can capture large blocks of spectrum over long periods of time. Using real-time spectrum analysis, a designer can acquire a broad swath of frequencies at one time. If the analyzer has sufficient internal-memory capacity, the designer also can acquire several seconds' worth of signal activity. The resulting data volume can fit well into a three-dimensional approach for measuring the handover process. It also lends itself to simultaneously looking at measurements in the time, frequency, and code domains.
Handovers from one BTS to another are required in several situations:
- Most commonly, they are needed when the user equipment moves from one BTS coverage area to another. It may move between stations within the same radio system or into another system. The W-CDMA standard supports handovers to any GSM TDD network frequency bands that meet the specifications. During a handover to a different BTS, the multi-standard UE may even change its frequency or radio-access mode.
- The user equipment may need a handover if its requested service level exceeds the current cell capacity. For instance, a target cell may not be able to support the combination of bearer services that is being provided by the current serving BTS. Examples of these services include voice, data, multimedia, etc. In this scenario, some or all of those services may be handed over to another BTS.
Using other frequencies and the radio-access systems that it supports, the multi-standard User Equipment continuously monitors for the presence of the BTS. When the network senses the need for a handover, the BTS assesses certain system parameters. It then commands the UE to measure and report.
Where does the BTS find the time to conduct and evaluate all of these measurements? When a handover is needed, it directs the UE to operate in compressed mode. Here, transmissions are turned off for a portion of the 10-ms frame. The resulting "gaps" provide time for the UE and BTS to make needed measurements.
Compressed frames can be set to occur periodically or on demand. The rate and type of compressed frames varies depending on the environment and the measurement requirements. Separate compressed-mode signals must be defined for uplink and downlink paths. They also need to be defined for each mode, radio-access technology, and frequency band that is supported by the UE.
SELF-MEASUREMENTS In compressed mode, a transmission-gap pattern sequence is requested by higher-layer BTS protocols. These parameters are then passed along to the UE by the BTS. The UE conducts only one set of measurements for each transmission-gap pattern sequence.These measurements, which are more like self-tests, are designed to smooth the handover process. They are not intended to characterize performance in the design lab. That job falls under the province of an external measurement tool, such as a wireless communication analyzer.
The UE and BTS measure the OSI Model's Layer 1 (Physical Layer) protocol to determine and report the status of intra-frequency, inter-frequency, inter-system handovers, traffic volume, and QoS levels. The BTS transmits a "measurement control message" to the UE. It then conveys the measurement ID along with the type of measurement to initiate.
When the measurements are completed, the UE sends a "measurement reporting message" to the BTS. This message contains the measurement ID along with the results. The measurement control message is broadcast in idle mode within the system.
When the UE monitors the BTS at other frequencies, modes, and radio-access technologies, the BTS must direct the specific measurement that is needed to fulfill the requested handover. In W-CDMA, the Layer 1 measurements are reported to higher layers of the protocol. In GSM, they are only reported to the GSM terminal.
As long as everything works correctly, the wireless system's internally conducted measurements are designed to ensure clean handovers. Remember that the handover process is designed to check various status parameters—not look for errors. The process can go through the motions of notification, measurement, and report. However, it may still fail because an error has occurred.
Error conditions that arise during compressed-mode measurements and handovers can be brief and unpredictable. To catch these intermittent problems, it's best to monitor power levels, frequency, and modulation information before, during, and after the compressed-mode events. Data must be captured in order to preserve signal characteristics and reveal error sources. The wireless communication analyzer and its graphical views—spectrograms and codograms—address these needs.
As designers discover its effectiveness, the codogram is steadily coming into wider use. This tool provides a view that targets the visualization needs of code-division multiplexing technologies like CDMA and W-CDMA. It is a 3D display of the orthogonal variable spreading factor (OVSF)—that is, channelization code versus time slot versus code power. In Figure 4, the vertical axis is the time slot and the horizontal axis is OVSF. Code power is represented by color.
To address new W-CDMA measurement challenges, solutions must keep pace with long, complex procedures. Such solutions must capture voluminous information while producing easily understood and visualized results.
Real-time analysis invites new acquisition architectures to confront this problem. Such architectures acquire enough data in one sweep to support information-rich, time-correlated displays. These displays are essential in helping designers analyze complex processes like W-CDMA call-setup procedures and compressed-mode handovers.