Simulcast design for the Flex paging protocol

The Flex paging protocol, introduced in 1993, is becoming the worldwide standard for paging. More than 40 million Flex-based pagers are in service around the world. The largest U.S. service providers have adopted the protocol, as have service providers in Canada, Latin America, Asia, Africa, Europe and the Middle East.

During the last year, several service providers have begun upgrading their Flex networks to higher speeds to meet demands for increased capacity. Network upgrades have been successfully tested in the real-world RF environment where intermodulation, fading, adjacent channel interference and, most importantly, simulcast distortions can degrade service quality. As a result, carriers realize that upgrading a network from 1,600bps to 6,400bps is complex, yet possible.

The simulcast environment Simulcasting means broadcasting the same message from multiple transmitters operating on the same frequency. Flex uses simulcasting to reliably deliver messages to pagers located anywhere within the coverage area.

Simulcasting boosts the coverage over a wide service area while maintaining a high level of building penetration. Macrodiversity gives simulcast its inherent strengths. Regardless of the pager’s location inside the building, there is always a chance that one of the signals around the building will reach the pager. Moreover, signals from different transmitters can combine constructively, boosting the overall signal level.

Unfortunately, simulcast benefits are inevitably coupled with signal distortion1, which becomes more problematic for the system designer as the channel data rate increases. The quality of signals generated by different transmitters, when received at the pagers, depends on the signals’ magnitude, relative delays and relative frequencies. Viewed from the pager’s physical location, the path lengths to the various transmitting sources generally are not the same. The combination of several paths contributes to distortions known as simulcast delay spread (SDS). SDS is regarded as the most challenging effect that the system designer has to design for. Another important source of simulcast distortion is zero beating, a phase cancellation condition. When signals are received from transmitters at offset frequencies, severe periodic signal cancellations may occur, especially in high-rise building and open-overlap areas.

Simulcast delay spread Simulcast delay spread results from intersymbol interference that occurs when several replicas of the same signal reach the receiver with unequal delays. SDS distortions can be observed in overlap areas where the relative delay between multiple received paths is in excess of the system SDS tolerance2,3. The rule of thumb for admissible SDS performance is to design the system for a 1/4-symbol SDS tolerance. A conservative design goal should consider approximately 15% to 20% of the symbol period. This is about 50ms for Flex systems operating at 6,400bps. (For a two-path static simulcast model with 1dB power differential, 16Hz offset and at 30dB above pager sensitivity, the pager should have acceptable performance when the differential delay between the two paths is around a 1/4 of a symbol.)

Zero beating Zero beating is a phase cancellation condition that occurs when two or more transmitters send the same data at slightly offset frequencies. Two transmitters simulcasting the same carrier with a small frequency offset illustrates the zero beating process. Figure 1 on page 28 shows the cancellation effect in the time domain. Frequency offset and power differential are the key parameters that control this process. Cancellations occur periodically at the frequency offset fo. The smaller the power differential between the two paths becomes, the deeper the cancellations become, as shown in Table 1 above.

For example, a power differential of 1dB causes cancellation depths of about 25dB. As the frequency difference between the two paths becomes smaller, cancellations occur less frequently but with longer duration. Inversely, at high offset frequencies, shorter cancellations occur more frequently. (See Table 2 above.) Frequency offset is the only parameter that controls cancellation duration. It is interesting that for a significant observation interval, the total cumulative period of time during which the signal power goes below a certain threshold is independent of the offset frequency.

This phase cancellation effect is mitigated in Flex protocol systems through error correction and data interleaving.

Simulcast performance The following section provides information about two key pager performance measures: namely, delay capture performance and delay spread performance.

Pager delay capture performance An indication of a pager’s ability to withstand simulcast distortions is its delay capture performance. Delay capture curves reflect the pager’s ability to defeat zero beating and SDS distortions. Figure 2 on page 30 shows an example of static delay capture curves for a Flex 6,400bps system. Two curves illustrate the pager’s performance under different channel conditions (high carrier-to-noise ratio [C/N] versus low C/N). Each curve represents the set of operating points for which the decoded word error rate (WER) equals 0.1%. Capture curves show three operating regions: o low delay operating region or zero beating area. o transition region or delay-limited area. o high delay operating region or capture area.

Low delay operating region-When the pager operates at low C/N, its tolerance to zero beating is diminished. In fact, the available amount of signal power is not great enough to compensate for the periodic cancellations. The pager performance depends primarily on the power differential between the two simulcast paths. Only cancellations with moderate depth are tolerable.

When the signal strength is high, the pager tolerance to zero beating distortion improves. Severe cancellations caused by low differential powers can be filled in partially or completely by the strong received signal. Figure 2 shows that for high C/N values (about 30dB above pager sensitivity) and for a differential delay of 50ms, the pager can tolerate cancellations produced by a power differential as low as 1dB.

Transition region-At delay values of approximately 1/4-symbol in duration, the combined effect of delay spread and zero beating causes additional signal distortions. As the differential delay increases, the pager tolerance to zero beating is significantly reduced. This increase means that only moderate power differentials are tolerated.

High delay operating region-Delay spread is the main cause of signal distortion. For delays greater than 1/2-symbol in duration, the pager performance becomes unacceptable, even at locations with high signal strength.

Only capture can guarantee acceptable pager performance. The pager capture ratio is the power differential, value between the strong and weak signal, at which the pager performs at a given quality level. The ratio is an indication of the receiver’s capability to detect a strong signal in the presence of a weak interfering signal. The capture ratio depends on the received signal strength. Figure 2 shows that at high C/N, the capture ratio is only 3dB, but it exceeds 7dB at low C/N.

Pager delay spread performance Another approach for evaluating pager performance in a simulcast environment uses the delay spread model. The root-mean square (RMS) value of the simulcast path delays, weighted by their respective power levels, represents a reliable parameter for evaluating pager SDS performance1.

This parameter is called simulcast delay spread and noted as Tm: (see issue for formula) Pi and di, respectively, represent the power and delay of the ith signal. The above equation shows that the same Tm value can be produced by a multitude of power and delay combinations. For a fixed Tm value and a fixed C/N level, in a fading environment, the pager performance remains fairly unchanged.

This interesting property makes SDS evaluation an attractive procedure for rating pager performance. Typical delay-spread curves have a “hockey stick” shape as illustrated by Figure 3 at the right. In a low SDS environment, the required C/N that guarantees a certain level of pager performance (for example, 0.1% WER) slightly exceeds pager sensitivity.

Under such conditions, pager performance is only limited by the received signal strength. In the vicinity of the SDS limit, the pager performance degrades severely. Signal cancellations become so severe that even in a high signal strength environment, system performance becomes unacceptable.

Design system for simulcast Pager simulcast performance is a factor in the overall system coverage performance. For example, it is simple to design a simulcast system for large SDS tolerance because performance is mostly limited by the pager sensitivity. The design becomes challenging if SDS tolerance is low. The mission of the system designer is first, to understand the pager simulcast requirements, and second, to ensure that every coverage area meets these requirements.

Design system for simulcast delay spread Traditional low data-rate simulcast systems such as POCSAG or Flex 1,600bps, do not need to address the SDS problem. This is primarily due to the large symbol duration that provides an inherent immunity to delay spread. High-speed simulcast systems are more vulnerable to SDS problems. Paging carriers are therefore compelled to design and optimize their systems methodically to guarantee acceptable coverage performance.

Figure 4 at the left graphs the process of optimizing a simulcast system. Two axes are used to approximate the pager performance. The vert-icle axis reflects the pager sensitivity threshold, while the horizontal axis sets the pager SDS limit. To guarantee acceptable performance for a given location within the coverage area, two conditions must be fulfilled: o The signal strength must be above the sensitivity threshold. o The delay spread distortion must be within the SDS capabilities. The goal of the system designer is to ensure that 95% of the coverage area falls within the SDS limits (assuming that the paging system is designed for 95% area reliability). For example, Figure 4a represents a system design that lacks signal power and has several areas with SDS problems. On the other hand, figure 4d represents a system that is optimized for simulcast.

The authors2,4 have discussed power delay management techniques that can bring the entire coverage area within the pager-imposed limits and can optimize coverage performance.

A simple optimization example using the Motorola propagation tool (MOZAIK) illustrates the coverage performance of each of the scenarios described in Figure 5 on page 33. Each coverage map shown in Figure 5 corresponds to the scenario in Figure 4 with the same label. Figure 5a represents a system where two sites are intentionally placed much higher than the three other sites to cause delay spread problems. The graph shows that the system has SDS and sensitivity problems.

The system illustrated by Figure 5b uses high sites with high transmission power. The simulation shows a severe SDS problem across the entire coverage area. Figure 5c shows a typical example of a sensitivity-limited system. Increased signal strength throughout the coverage area is needed.

Signal strength can be increased by adding sites, increasing the transmit power or increasing the antennas’ heights. Figure 5d shows an optimized system. The sites’ heights and transmission power are increased to meet the sensitivity requirements, while directional antennas and launch delays are used on the external sites to meet the SDS requirements.

Design system for zero beating: frequency offsets An offset plan manages the zero beating process. The goal is to enhance performance in high rise and overlap areas where the power differential between signals is low (in the order of 2dB-3dB or less). Optimum offset frequencies regulate the cancellation duration and the frequency of cancellation to match the coding and interleaving capabilities of the Flex protocol.

Computer simulations, laboratory tests and field measurements show that frequency offsets contribute to a substantial improvement in the coverage performance of zero beating areas. This is true in both static and fading environments. Figure 6 above shows the simulated results of WER vs. frequency offset for Flex 6400 in static (Figure 6a) and 3mph slow fade (Figure 6b) environments.

Figure 7 on page 35 shows the field- measured results of WER vs. frequency offset for Flex 6400 in a static environment. Note that offsets like 1Hz, 40Hz, 50Hz and 100Hz all produce relatively high WER.

The results show that good offset frequencies are found in the range (10Hz to 90Hz) (Figure 6a). The recommended offset frequencies are {245Hz, 230Hz, 215Hz, 0Hz, 15Hz, 30Hz, 45Hz}. Offsets that are too high should be avoided because they result in excessive pager sensitivity loss.

Examples of four-cell and seven-cell offset frequency reuse patterns are proposed in Figure 8 above. With careful offset assignment to each cell, four offset frequencies are sufficient to cover any type of coverage area without adjacent cells sharing the same offset. However, five to seven offset frequencies provide more flexibility, especially for future system expansion.

Conclusion This article provides guidelines for simulcast system design that meet the constraints imposed by upgrading a paging system to higher speeds. Power and delay management techniques can bring the performance of the entire coverage area within the sensitivity and SDS design constraints. Frequency offset reuse plans for Flex 6,400bps are recommended. The plans alleviate zero beating problems in high rise building and overlap areas and improve the coverage performance of the overall system.

References 1. Hess, Garry, Land-mobile Radio System Engineering. Artech House. 1993. 2. Sexton, T., Souissi, S. and Hill, C., “The Performance of High-speed Paging in Simulcast.” IEEE Proceedings of the Wireless Communications Conference, Boulder, CO, August 1997. 3. Williams, Lee, “System Integration of the Flex Paging Protocol,” Mobile Radio Technology, July 1996. 4. Souissi, S, “High-speed Simulcast Messaging: Simulcast Coverage Optimization,” Optimizing Paging Network Design and Management, technical conference, London 24 September 1997 5. Hill, Selwyn, “Identifying the Pitfalls on the Way to High-speed Transmission and the Difficulties of Simultaneously Operating a Low and High-speed Network: The Service Provider Perspective.” Optimizing Paging Network Design and Management, technical conference, London 24 September 1997.

Suissi is principal staff engineer, Sek is senior staff engineer and Hill is director of system technology research for the Motorola Messaging Systems Product Group, Fort Worth, TX.