Design of Simulcast Paging Systems using the Infostream Cypher Document Number 95-1003. Revsion B 2005 Infostream Pty Ltd. All rights reserved
1 INTRODUCTION 2 2 TRANSMITTER FREQUENCY CONTROL 3 2.1 Introduction 3 2.2 High stability oscillators 4 2.3 GPS disciplined oscillators 4 2.4 Transmitter interface 4 3 FREQUENCY SETTING 4 3.1 Fixed frequency offsets 5 3.2 True or pseudo random noise offsets 5 4 SIMULCAST EQUALIZATION AND OFFSETS 6 4.1 Overview 6 4.2 Path length equalization 7 4.3 GPS synchronization 7 4.4 Non-symmetrical overlap zones 8 5 CYPHER SETUP 8 6 SUMMARY 9 1 Introduction This paper discusses some of the issues that must be considered when designing a simulcast paging system, particularly a system that is based on Infostream s CYPHER paging encoder system. A simulcast paging system is a system that uses multiple paging transmitters to provide coverage over a greater area and with greater reliability than could be achieved using a single transmitter. The system is described as simulcast because the transmitters all operate simultaneously to transmit identical signal to the air in unison. This is distinct from a sequential paging system where the transmitters operate one after another in turn so as not to interfere with each other. A simulcast system has many advantages over sequential systems including much greater channel capacity and avoidance of repeated messages on the pager however simulcast systems are rather more difficult to design and setup. With the advent of the GPS system and with Infostream s CYPHER encoder system, simulcast systems are now nearly as easy to design and operate as sequential systems. Whitepaper - Simulcast Network Design Using Cypher. Doc No 95-1003 Page 2 of 9
The greatest difficulty in implementing simulcast-paging systems is to solve the problem of coverage overlap. When multiple transmitters are operating at the same time, the signals from adjacent transmitters can interfere with each other in a way which, if not properly designed, would cause the pager to miss messages while operating in the overlap zone. There are two reasons for this. The first has to do with the mixing of transmitter RF carriers in the overlap zone. When transmitters overlaps with similar RF field strengths, the RF energy can combine in the overlapping zone to cancel the signal out. Various techniques have been developed to overcome this problem. The second reason relates to the propagation speed of the RF signal, which is limited to the speed of light. While fast, light is not infinite and with the relatively high data speeds and long distances involved in paging systems, the speed of light must be taken into consideration when designing a simulcast system. For proper operation the paging encoders at each transmitter site must act in exact unison to ensure that the modulation in the overlap zones adds in a constructive way. It is important to distinguish these two problems from each other. One problem arises from the paging transmitter RF carriers interfering with each other. The other is a function of the modulation time alignment and the distances between the transmitters. 2 Transmitter frequency control 2.1 Introduction In a simulcast paging system, there will be zones in the desired coverage area where the contribution to the overall RF field from two or more transmitters will be sufficiently close that several transmitters will affect the paging receiver. In order to control the performance of the network in these zones of overlap, it is important that good control of transmitter frequency is achieved. Although this discussion paper deals mostly with two level paging modulation schemes such as POCSAG, even tighter requirements exist for four level modulation schemes such as the ERMES protocol and the proprietary FLEX protocol. Exactly how stable a transmitter must be depends on a number of factors and is the subject of considerably controversy and debate. Clearly, the first requirement is that the transmitter must meet local legislative requirements for frequency stability. Furthermore, for ease of network maintenance and long-term reliability, the transmitter should be both short term and long term stable. This means that the frequency of the transmitter must not exhibit significant drift for periods of several years, ideally for the operational life of the network. The nominal deviation requirement for a POCSAG modulation system is +/- 4.5 khz. In practice, it makes little difference to the performance of the pager if the modulation level varies by as much at 500 Hz however. From this it can be concluded that the frequency stability for a POCSAG paging transmitter, at least when operated independently (i.e. not part of a simulcast system) is relatively low. For a four level FLEX or ERMES system, the situation is considerably different. For these systems, the pager must discriminate between four levels, not just two and the FLEX specification explicitly requires that deviation should be accurate to better than +/- 10 Hz. In addition to the requirements of the paging protocol, the use of a transmitter as part of a simulcast network imposes additional constraints on frequency stability. These requirements are discussed in detail in the next section. However before proceeding with this discussion, the two common methods for transmitter frequency control are discussed. Whitepaper - Simulcast Network Design Using Cypher. Doc No 95-1003 Page 3 of 9
2.2 High stability oscillators A high stability oscillator is an oscillator where special precautions have been taken to ensure that the frequency of oscillation is very tightly controlled. Typically this is achieved through a variety of analogue and digital circuit techniques aimed at overcoming drift in the crystal that typically forms the heart of the oscillator. Either temperature compensation (adjusting the circuit to compensate for variations in ambient temperature) or temperature stabilization (in the form of a constant temperature oven ) is used to compensate for temperature variations that might affect oscillator frequency. Sometimes, both techniques are used. Normally a high stability oscillator is sold as an option to the paging transmitter and is used to replace the usually low-grade oscillator used for voice or non-data applications. 2.3 GPS disciplined oscillators With the advent of GPS, a new highly accurate frequency reference has become available. The GPS system, while primarily designed to provide a positioning and navigation service has, as an inherent side effect, provided a very accurate and long-term stable frequency reference. Certain GPS receivers have been developed whose primary function is to provide not position, but frequency and time information with a high degree of accuracy and stability. The Trimble Thunderbolt, as an example of this type of GPS receiver is in fact described not as a GPS receiver but rather as a GPS disciplined oscillator. A GSP disciplined oscillator is a device that combines a highly stable oscillator with the GPS receiver. The stable oscillator is designed to be very accurate and stable over the short to medium term and to exhibit very low phase noise. This oscillator is then controlled by the frequency reference derived from the GPS system. By combining the long-term stability of the GPS system with the high short-term stability of the built in oscillator, a GPS disciplined oscillator can achieve very high frequency accuracy indefinitely. By further combining temperature and ovenized crystal control with GPS disciplining, a good GPS disciplined oscillator can operate for long periods of time independently of the GPS network, allowing for short term GPS outages without affecting oscillator performance. With a CYPHER based system, a GPS time reference is required to synchronize the transmitter modulator to achieve the required simulcast data alignment. Consequently it is attractive to consider the use of a GPS disciplined oscillator for transmitter frequency stability in CYPHER based simulcast systems. 2.4 Transmitter interface Whether a high stability free running oscillator is used or a GPS disciplined oscillator, the oscillator output is typically at a fixed frequency below the paging transmitter operating frequency. Typically a 10 MHz output is provided to match the frequency reference input to the transmitter. Phase locked loops and other circuits are used to multiply the reference frequency to the desired paging frequency. 3 Frequency setting Assuming that either via high stability oscillators or GPS disciplined oscillators, we have two highly frequency accurate paging transmitters. Whenever these transmitters are operated so that their coverage zones overlap the RF energy from the two fields generated by the transmitters will combine. How they combine depends on the frequency and phase of the signals from the two transmitters. Whitepaper - Simulcast Network Design Using Cypher. Doc No 95-1003 Page 4 of 9
If the two transmitters are operating exactly at the same frequency, there will be a point approximately mid way between the transmitters (assuming that they are of the same power) where the two signals are of identical strengths. At that particular point, these two RF fields will constructively combine to increase the field strength seen by a paging receiver located at that point. Similarly there will be a point where the two signal combine destructively so as to reduce the RF field to something approaching zero. A pager placed at that point of destructive combination will see no RF signal at all and will not be able to receive paging messages. The distance between these two points will be exactly half a wavelength of the RF frequency which for a 150 MHz VHF band paging transmitter will correspond to a distance of about one metre. For as long at the two transmitters are of roughly equal power, there will be a large number of points mid way between them where the pager will at intervals of roughly one metre see either a good signal (constructive interference) or no signal at all (destructive interference). For a simulcast network, this is an unsatisfactory situation. It will result in areas of very extreme fading of the RF signal for a paging moving within this area or no signal at all when the pager is stationary at a null point. To overcome this problem, a number of strategies have been developed and deployed. 3.1 Fixed frequency offsets The first technique is to use fixed frequency offsets. Unlike the situation described above where the transmitters are operating on an identical frequency, the transmitters can be set up to be deliberately some fixed frequency apart. If we look at the situation described above where a pager is placed mid way between two transmitters of roughly equal power, we will see a completely different situation when frequency offsets are used. Instead of having areas of nulls and peaks separated by intervals of approximately ½ a wavelength, we will find an oscillating pattern of peaks and nulls moving around in space at frequency corresponding to the difference between the two paging transmitter frequencies. From a pager s point of view in such a field, the situation is very much more satisfactory. The speed with which the pattern of peaks and nulls moves is generally much higher that the speed with which a pager might move through the field (assuming that the pager might be in a car for instance) and a stationary pager cannot be placed in a location where a fixed signal null occurs. The pager will experience nulls in the RF field of course, but the duration of these nulls will typically be extremely short. By virtue of the filtering and error correcting capabilities of the pager, the effect of these moving nulls is very nearly negligible. Typically a null will be of a duration rather less than a single bit. The situation with three or more transmitters, all with different frequency offsets is very similar although the pattern of nulls and peaks becomes increasingly more complex. 3.2 True or pseudo random noise offsets Some paging transmitter manufacturers have developed systems that introduce variable frequency offsets. In these systems, rather than using a fixed frequency offset, a random or pseudo random frequency offset is applied to the paging frequency carrier. In practice this technique will lead to a very complex pattern in the overlap zone of two transmitters but the effect is similar to the frequency offset system described above a stationary or relatively slowing moving pager in the overlap field will experience very short duration nulls that will not affect the overall reliability of the paging signal. Whitepaper - Simulcast Network Design Using Cypher. Doc No 95-1003 Page 5 of 9
These pseudo random frequency systems are sometimes called black noise systems. It is a matter of debate as to which of these two systems is more effective and there are plus and minus points for both. One advantage of the black noise systems is that they do not require specially planning on the part of the network operator to devise a frequency offset plan as would be required for a fixed frequency offset system. Each transmitter generates it s own random sequence of offsets so that the combined effect of any number of transmitters will itself be a random field. In either case however, it can be seen that at least for a two level modulation system such as POCSAG, the actual frequency of the transmitter is not that critical at all. In a situation where either significant (say 500 Hz) or random frequency offsets are used, relatively small changes in transmitter frequency as a result of oscillator drift are of little consequence. In fact it was common, until relatively recently to simply allow the random frequency variation between adjacent paging transmitters to serve the purpose of the frequency offset scheme. Although fairly ad-hoc, simulcast systems of this type worked surprisingly well. 4 Simulcast equalization and offsets 4.1 Overview A simulcast paging system necessarily involves the use of at least two and possibly hundreds of transmitters depending on the required coverage area. In order to provide continuous coverage over the entire area it is necessary to arrange for the transmitters in the system to have sufficient power and to ensure that they are sufficiently close to each other to create overlap regions where two or more transmitters provide sufficient field strength for the pager to reliably receive a message. An overlap is created when the field strength contribution of two or more transmitters are within a relatively narrow range. Typically a figure of 6 db is considered to constitute an overlap zone. Anything above this level and the pager will lock on to a single transmitter and the effects of the contribution of the lower field strengths can be ignored. If we assume that two transmitters of equal power and with equivalent antenna systems are operating in a system covering essentially flat terrain, the overlap zone will normally occur in the area approximately mid-way between the two systems. In this case, the inverse square power law will attenuate the signals from the two transmitters equally at the mid-way point. For a certain distance either side of the mid-way point, the signal level from each transmitter will be within 6 db of the other, creating an overlap zone. Because the speed of light (and radio propagation) is not infinite, the signals from the transmitters will arrive at the overlap zone slightly after they were transmitted according to the following formula: Delay = 3.3 microseconds / kilometre from the transmitter As a rule of thumb, the modulation from each transmitter covering the overlap zone should be arranged so that the delay from each of the contributing transmitters is no more than approximately ¼ of a bit different in the zone. Above this figure there will be a measurable drop in pager performance (sensitivity). For a 1200 baud POCSAG paging system, this is roughly 200 microseconds which, according to the formula above represents a path length from transmitter to overlap zone of about 63 kilometres. From the above, it can be seen that for any simulcast paging system, covering a region equivalent to the size of a reasonable city or more, it is critical that the modulation timing of each transmitter is very accurately controlled. Whitepaper - Simulcast Network Design Using Cypher. Doc No 95-1003 Page 6 of 9
There are two basic approaches in use to ensure that modulation of each transmitter is tightly controlled. 4.2 Path length equalization Path length equalization has been the traditional approach for paging transmitter timing control for at least the last twenty years or more. This approach relies on taking into account the entire path from the encoder to transmitter for each of the transmitters in the simulcast system and introducing artificial delays (through programmable delay lines) in each path to ensure that the path delay is the same for each transmitter. For a paging system utilizing a point to multipoint RF distribution system, this is a relatively easy thing to arrange since the path can be readily calculated and compensated for. For simulcast distribution systems involving land lines, satellite links and other distribution methods, this can be much more complicated because often the path is not known beforehand since it is usually provided by an independent party such as the local telephone carrier and is not guaranteed. In such cases the delay must usually be measured during installation, rather than calculated and then compensated for. The situation is further complicated for path length equalization systems because backup systems, which are becoming an increasingly necessary part of sophisticated paging systems, may involve switching one or more transmitters to alternative paths which may (and almost certainly will) have different delays. 4.3 GPS synchronization The more modern approach, and the approach which has been employed by virtually all simulcast systems in the last few years is to use GPS timing references at each of the transmitter sites for modulation timing. The basic approach is similar in all systems using this method and involves sending the paging data (in some form or other) to each transmitter site a few seconds ahead of time and arranging for a device at the transmitter site to time and clock out the data based on the GPS clock. If correctly done, this can result in very accurate timing offsets (less than one microsecond) between all transmitters in the simulcast system. One great advantage of this approach, apart from the generally high timing accuracy achieved is that the timing of the transmitters is completely independent of the path length from encoder to transmitter. Systems involving RF, landline, satellite and IP based links can all be used and mixed in a system to provide flexibility and redundancy without affecting the simulcast network performance. The Infostream CYPHER simulcast encoder system is an excellent example of a modern GPS based simulcast system. The Infostream CYPHER system utilizes GPS timing, producing modulation timing accuracy of better than one microsecond, improving on the required timing accuracy for a 1200 baud system by a factor of nearly one thousand. In the CYPHER system, encoding takes place once at a central distribution point. The encoding process is an entirely software function and does not rely on dedicated or specialized hardware to operate. As such, redundancy and backup of the encoding process can be performed entirely in software making this part of the system extremely reliable. The intermediate data format generated by the CYPHER encoder is then distributed along with timing information to each of the paging transmitter sites. The timing of these signals is encoded to include instructions to the transmitter as to when the encoded data is to be transmitted relative to a GPS timing reference. This implies that both the encoder and Whitepaper - Simulcast Network Design Using Cypher. Doc No 95-1003 Page 7 of 9
transmitter base station must both be in accurate synchronization with each other. Fortunately with GPS, this is very easily accomplished. Because of the low level intermediate format utilized between the CYPHER encoder and the transmitter controller, the data is in the form of a continuous stream of small packets. This allows the timing or forward link delay between the encoder and transmitter controller to be very finely controlled. Only sufficient forward link delay need be programmed into the system to ensure that the worst-case network distribution delay is taken into account. For terrestrial links of various kinds, delays of less than one second are feasible. For satellite distribution systems, up to 1.5 seconds might be required. The CYPHER system utilizes forward error correction over the distribution link to ensure high reliability of the paging stream. At the base station, the intermediate data format is transferred to the CYPHER transmitter controller where the GPS timing reference is used to clock the data stream into the transmitter with an accuracy of better than 1 microsecond. Offsets in microsecond steps can be specified to take account of non-symmetrical overlap zones. 4.4 Non-symmetrical overlap zones The situation described above deals only with transmitters of equal power covering generally flat terrain. The situation is considerably complicated if transmitters of unequal power are used, or where an overlap from more than two transmitters exists or where geographic effects create varying propagation characteristics from each transmitter. In such cases, the 6 db overlap region will not fall exactly mid way between two transmitters but somewhere closer to one transmitter or another. In such cases it may be necessary introduce deliberate offsets in the modulation of each transmitter so that modulation of the contributing transmitters in the overlap zone are within the required ¼ bit range. This can be done either by introducing a delay in the transmitter, or by adjusting the encoder. The Infostream CYPHER encoder provides a direct method for specifying transmitter modulation offset in microsecond steps, providing a simple and convenient way of setting modulation offset. Delays can be specified up to +/- 1000 microseconds. The delay is adjusted so that the modulation arriving from two adjacent transmitters is within ¼ of a bit for the entire 6 db overlap region. Using the CYPHER and with typical transmitter power up to a few hundred watts, considerably less than ¼ bit offset can be achieved even for the highest POCSAG speed of 2400 baud. 5 CYPHER setup The CYPHER encoding system is an extremely easy to setup system. The CYPHER transmitter interface, which is installed at each transmitter site is completely adjustment free. It requires only external power (a few milliamps at 12 volts), a one pulse per second timing reference from the GPS receiver, a serial connection to the base station controller and connections to the transmitter. The transmitter connections comprise data and push-to-talk signals. Four lights on the CYPHER transmitter controller quickly indicate it s operational status and are described in detail in the CYPHER user and operation manual. There are absolutely no settings whatsoever on the CYPHER device. Simulcast timing offset (the only adjustable parameter) is configured in the base station control CPU and is remotely configurable. Whitepaper - Simulcast Network Design Using Cypher. Doc No 95-1003 Page 8 of 9
All other parameters relating to the transmitter are controlled at the central encoding site. These parameters include transmitter key up and key down delay times, forward link timing offset etc. 6 Summary Simulcast paging systems offer the significant advantages of greater channel capacity and freedom from repeat messages when compared to sequential paging systems. By utilizing multiple transmitters, larger areas can be provided with reliable paging service than could be achieved using just a single transmitter. In designing and operating simulcast-paging systems, two important factors must be considered in the RF design: paging frequency control and offset, and simulcast equalization. Paging frequency control can now be easily achieved using readily available high stability oscillators or GPS disciplined oscillators. Using either frequency offsets or noise modulation schemes, the problem of RF overlap can be readily overcome although some care and testing is required to achieve optimum results. Using the CYPHER encoder with GPS synchronization, simulcast equalization is almost trivial, particularly if designing systems with many equal power transmitters where overlap is largely symmetrical. The CYPHER system completely takes care of unequal distribution path delays through the use the use of GPS timing at the encoder and transmitter sites. GPS timed simulcast system and the CYPHER in particular which is IP based are particularly well suited to the implementation of multicast distribution over IP links with redundancy for reliability. Whitepaper - Simulcast Network Design Using Cypher. Doc No 95-1003 Page 9 of 9