System design considerations for high-speed paging
The advent of modern high-speed paging formats compels wireless messaging carriers to take a closer look at the phenomena affecting system performance in order to optimize their systems effectively and efficiently.
Optimizing system performance has been a continual process for wireless messaging carriers since the introduction of modern high-speed paging protocols. Over the past several years, operators have been designing and operating systems at higher data rates with increasing success.
However, analyses of the operating environment, adjacent- and co-channel interference, simulcast delay spread and frequency offset considerations suggest that further improvements to system performance and reliability are possible.
The mechanisms known to produce increased error rates in modern, high-speed, paging systems are examined. Guidelines are also presented to aid in improving system design, development and performance.
Environmental considerations The first consideration must be the type of environment in which the system will be operating. The surrounding environment can have a dramatic effect on the system performance and required transmitter density. Obviously, we cannot do much to change the surrounding terrain beyond accounting for it in the propagation analyses and system planning.
The surrounding terrain and building density have a significant effect on the severity and frequency of signal fades. Fading is normally caused by a number of multipath components arriving at the receiver out of phase. If a majority of the components are out of phase with each other, they combine destructively to produce a large decrease in the received signal strength. Fading can also be caused when the signals from two or more simulcast transmitters arrive at the receiver out of phase. This is often called zero-beating and can be significantly reduced by employing frequency offsets throughout the system. Urban environments are especially harsh when it comes to fading due to multipath components. A typical received signal strength plot is shown in Figure 1 below. The figure shows the instantaneous signal level fluctuations due to multipath fading in an urban environment.
To compensate for the fading induced by a given environment, the average signal level at the pager must be increased. This is done to ensure that the pager will have sufficient signal strength during all but the most severe fades. It is not practical to compensate for every signal fade, but with the appropriate increases in signal strength, the effects of fading can be made negligible. Recent testing suggests approximate guidelines for the amount of additional power that is required to compensate for multipath fading in various environments. A summary of the results are given in Table 1 on page 30. The values are based on a code-word-error-rate of 1%. This translates to a success rate of about 99% for tone-only pages, 97% for 10-digit numerics, and 90% for 25-character alphanumeric pages. (Success is defined as the reception of the page without errors.)These values do not include any building attenuation factors; they are solely to compensate for the effects of multipath fading.
Adjacent-channel interference Another consideration, adjacent-channel interference, is often difficult to characterize during the system design process. As you would expect, high levels of adjacent-channel interference cause problems with paging reliability. To identify problem areas, consider the instantaneous carrier to adjacent-channel power ratio, not just the ratio of the average powers. If you look only at the average signal levels and see a 30dB differential, you are likely to assume a good performance. This is especially true if the pager specifies 50dB of adjacent-channel rejection. Unfortunately, this is not a good assumption in an environment subject to significant levels of fading.
In a typical urban environment, ratios as small as 25dB can show noticeable reductions in paging reliability because the adjacent channel transmitter is likely to be close to the user, and not fading any appreciable amount; whereas the desired transmitter is some distance away with the possibility of experiencing significant signal level fades. As a general rule, the pager’s adjacent-channel rejection specification should be derated by the typical fading margin for the given environment.
Co-channel interference The designer must also contend with co-channel interference, resulting from intermodulation products, or from a distant transmitter in the same system. When the interference originates from a transmitter in the same system, it is normally a simulcast delay spread issue, which is discussed in more detail below. Co-channel interference, from an uncorrelated source, has the same effect as raising the noise floor in the system. Thus, additional power is required to achieve the same level of performance that would be seen without the interference. It may seem as if a viable solution would be to raise the signal level in the affected area. Unfortunately, things are not that simple. Increasing the signal level in the affected area may cause interference at other sites, presenting the designer with a “Catch-22.”
Normally, if the offending signal is more than 25dB below the level of the desired carrier, it can be safely ignored. Levels in the range of 10dB to 20dB below the desired carrier can have a significant effect on performance, especially as signal levels approach the sensitivity threshold of the pager. If co-channel signals are within 10dB of the desired carrier, they probably come from a distant transmitter in the same system. In this case they are best modeled as a simulcast delay spread problem.
Simulcast delay spread Simulcast delay spread (SDS) is also a form of co-channel interference, but it has some unique qualities. Although the two signals arriving at the pager contain the same data, they can interfere with each other because of the time delay between them. If the time delay is a significant portion of the symbol period, the effect can drastically reduce paging reliability. This is especially true in high-speed paging where the effective symbol width is significantly reduced compared to the older, low-speed formats. This places tighter delay constraints on the system design. A general rule has been to limit the maximum differential delay to one quarter of a symbol period. This is only a rule of thumb and should not be used as a design guideline. We recommend that the designer limit differential delays to the range of 15% to 20% of a symbol period when planning a system. This corresponds to laboratory test results, as well as to actual field measurements, which indicate maximum delays in the range of 50ms to 60ms for Flex 6400. Figure 2 at the right shows the effect of a delayed signal on the error rate as a function of the signal level differential and time delay. As the difference in power between the two signals is decreased, the level of distortion is dramatically increased. The data represent the average response for a group of randomly selected numeric and alphanumeric Flex pagers operating at 6,400bps.
When a propagation analysis tool is used to predict potential SDS areas, a “capture window” of 10dB should be employed. This window size helps to account for both the variability of the signal level predictions made by the analysis tool and for the non-uniform nature of the pager’s antenna pattern when in close proximity to the human body. Wider capture windows can be used, but they tend to overstate the potential SDS problems unless the pager is operating close to its sensitivity threshold.
Frequency offsets When planning a system, the designer should always consider the use of frequency offsets. An offset scheme does not solve every problem in a simulcast situation, but if properly implemented, it can help the designer to squeeze the last few percent out of the system. In laboratory tests, almost any offset frequency scheme outperformed the zero or no offset case. Field tests with frequency offsets have also shown definite performance improvements in a number of situations. There were no cases in which a performance degradation was observed due to the addition of frequency offsets in the system. Therefore, a carefully chosen frequency offset scheme can only help system performance.
The POCSAG protocol performs best with offsets in the range of 1/3 to 1/2 of the symbol rate. Most POCSAG systems incorporate offset values between 200Hz and 800Hz to maximize performance while maintaining compliance with the FCC mask. The Flex family of protocols, on the other hand, performs much better with offsets less than 100Hz. With Flex, as with any interleaved format, the designer must be careful of the exact offsets chosen. In the range between 0Hz and 100Hz, both optimal and sub-optimal offset choices exist. Offset differences that are sub-multiples of 400Hz, (20Hz, 25Hz, 40Hz, 50Hz, 80Hz and 100Hz), tend to perform poorly and should be avoided. On the other hand, offset differences such as 15Hz, 30Hz, 35Hz, 45Hz, 60Hz, 70Hz and 90Hz work well and can reduce the error rate considerably. These offsets are valid for all Flex operating speeds. Figure 3 above provides a set of recommended offset strategies that can be applied to new or existing Flex systems. The offsets have been selected to maintain an optimum frequency differential between all transmitters in the given cell pattern.
The offset values presented have an average bound of 1Hz. This means that an offset frequency within 1Hz of the given value will have almost identical performance. However, if the frequency is shifted two or more Hertz away from the given value, significant changes in the error rate can occur. This effect is most evident when moving away from non-optimum offsets. This implies that transmitters with good frequency stability are necessary for the successful deployment of static frequency offsets.
For larger systems, the complexity associated with implementing a specific offset scheme may be prohibitive. In these circumstances, the use of an automated, random offset technique may be preferred. Glenayre has developed a system that introduces, and periodically updates, a random frequency offset in each transmitter. This approach allows the effective use of offsets without the necessity for extensive bookkeeping.
Conclusion Several mechanisms can produce increased error rates in modern high-speed paging systems. Suggested guidelines have been given for fading margin, allowable adjacent channel and co-channel signal levels, maximum differential delay, and optimal frequency offset strategies. These guidelines will allow the wireless messaging carrier to optimize the design and operation of high-speed paging systems with greater understanding and success.
References Hess, G., Land-Mobile Radio System Engineering, Artech House, Massachusetts, 1993. Rappaport, T., Wireless Communications Principles & Practice, IEEE Press, New Jersey, 1996. Williams, L., “System Integration of the Flex Paging Protocol,” Mobile Radio Technology, June 1996.