THE ART AND SCIENCE OF SIMULCASTING REDUX

By DENNIS CAMERON, TELCOM TECHNOLOGIES ASSOCIATES

It has been almost 25 years since I published my first simulcast paper. While it was state-of-the-art at the time, a lot of technology has come along to make it easier to implement and improve the quality of simulcast systems. This paper will revisit the history and theory, and discuss the new technologies. [2/2/07]
[Reformatted 5/19/10]

History

While PCS has intruded on some of the basic functions of paging there are still a lot of paging/voice paging/messaging systems around. The need for simulcasting has remained constant, to provide paging/messaging over a wide area and/or increase signal level within a given area. Also, some hardy souls may be around whom are still simulcasting voice.

Wide area paging, at least in the US, started with “Ma Bell”. These were bulky receivers that used two-tone (then three-tone) signaling and simply beeped (thus the term “beeper”) when signaled. Bell used Hi-band VHF and started with a single transmitter. In some cases they sequenced a second transmitter, and in a few cases they “simulcast” non-overlapping transmitters in conjunction with a sequence with other non-overlapping transmitters.

Bell had a different philosophy about paging, they were the Phone Company, and saw paging as just another way of generating phone calls. They applied a version of the Erlang charts and determined the max number of pagers they could handle on a channel was 500! If they needed more capacity, they would apply for a new channel. Because they were most interested in generating phone calls, they had little desire to use anything but tone only paging. Because they were “Ma Bell” they did not worry too much about capacity, if they ran out of capacity on a channel (according to the Erlang charts) and ran out of channels they would simply create a waiting list.

At the same time the FCC allocated paging frequencies for the phone companies (wireline carriers), they also allocated another group of frequencies for non-wireline carriers. Because regulators viewed this allocation as common carriage, its use also came under state control and a certificate of convenience and necessity.

Most of the early non-wireline carriers were answering services. These operators were looking for additional income and saw paging as a direct revenue generator. For the most part these folks loaded channels to the max, and they would basically continue loading until disconnects equaled connects! They constantly looked for ways to increase channel capacity. Back in those days only 6 channels (4 lo-band and 2 hi-band VHF) existed, and in the larger markets it was difficult to come by a channel, especially with the protected areas associated with lo-band.

The need for wide area coverage further drove the need for improved channel capacity. Even with the advent of high speed two-tone and five-tone paging formats the carriers were running out of capacity because the only effective way to go wide areas was to sequence the transmitters or use the combination of simulcasting non-overlapping transmitters in sequence with other non-overlapping transmitters (see figs. 1A, B). In addition to capacity, these methods still left a big problem in most major markets, building penetration.

The carriers were unable to get a signal into large buildings, especially with lo-band. The problem came in two forms; small apertures and reflective glass. Aperture has to do with the windows on older buildings. In RF terms, (this part will interest engineers and “teckies”) “aperture” is an opening that an RF signal can pass through. The optimum minimum aperture is ½ λ (wavelength). A 35 MHz signal’s wavelength is about 28 ft (8.8 meters) long, which requires an aperture of 14’ (or 4.4 meters). Not too many buildings have windows this large, so signals from these frequencies had difficulty in penetrating into the interiors of the buildings. Buildings that do have large windows often (especially in warmer climes) are all glass exteriors but the glass has a metallic content reflective surface. While this design is ok for reflecting the sun and heat, it also reflects RF signals creating the same problem as small apertures.

Carriers tried to solve the penetration problem by installing “fill” transmitters. In the larger markets this practice could require 3 or 4 fill transmitters further complicating the coverage vs capacity issue. If they sequenced the transmissions, then capacity was sacrificed. However, if they tried simulcasting they would have large areas of interference and their system would get clogged with re-calls. This not only affected capacity but required more phone lines to handle the calls (Ma Bell watched the lines and required common carriers to have only so many busies on a line).

With all these factors in play the carriers started asking the vendors for solutions, and a few hardy carriers started looking for their own solutions.

Early attempts at simulcasting proved to be problematic at best and completely useless at its worst. Most of these early systems were attempted using wireline and, in a few cases, microwave. Suffice it to say many man-years were spent trying to make these systems work (to little or no avail). When radio links were first tried it appeared to solve the problem but as faster paging formats came along (and voice paging was attempted) it was back to the drawing board! It wasn't until about 1980 that the first simulcast system that was designed from the ground up as a fully coherent simulcast “system”, was simulcasting truly successful.

The basics

First, the definition of simulcast (as used in the Land Mobile industry): Simulcasting is the simultaneous transmission of the same data (digital, analog or voice) through two or more transmitters within the same geographical area. Another description for simulcast is controlled multipath (we will look at that later). Diagrams 2A, and 2B are examples of overlap areas. An overlap area has been defined as an area where two or more RF signals have signal strengths within 6 dB of each other. This definition is only partially correct.

This rule-of-thumb came about because of the differences between AM and FM radios. Amplitude modulation, as most of you know can be very noisy. The intelligence (modulation) causes the amplitude of the signal to vary. The noise (static, lightning, etc., rides along with the amplitude peaks of the signal. Even in strong signal conditions, noise can sometimes be heard. With FM the intelligence (modulation) causes the frequency or phase of the signal to change. Noise still rides on the amplitude peaks of the signal but a FM receiver has a circuit called a limiter that cuts off the amplitude peaks. Because the intelligence is on the frequency or phase differences the limiter does not affect it but does greatly reduce or eliminate the noise. Because the limiter kicks in at about 6dB above the minimum signal level needed to hear a signal this parameter is known as the “capture ratio”, or, the receiver’s ability to capture signal over noise.

Designers then thought that if the overlap signal had a difference of 6 dB, no simulcast effect would exist. The problem with this conclusion is that the interference caused by overlap signals consists of both amplitude and phase noises. While the limiter could deal with some of the amplitude interference, it can do nothing with the phase noises. Unfortunately, half or more of the overlap interference components are phase noises. In addition, a large portion of the amplitude noise is caused by the RF signals beating together.

(WARNING: Teckie session ahead). The reason some of the amplitude noise is a problem has to do with the RF beat note. In almost all simulcast systems (past and present) each transmitter must generate the carrier signal using an oscillator. Because it is difficult to synchronize, oscillators. Even today, they are essentially free running (in relation to each other) devices. In any overlap area the signals generated by these oscillators will go in and out of phase with each other. While in phase (assuming near equal signal strength), the RF signals will add together giving a stronger signal. As these signals start to go out of phase they eventually reach 180° out of phase. This is known as a Zero Crossing. As the phase difference approaches 180°, the signals will start subtracting from each other until there is no signal left (still assuming near equal signal strength). At this point there will be nothing but noise. Rising and then lowering of the noise will occur on either side of the zero crossing point. So, from approximately 110° to 250° there will be a noise pulse that contains both amplitude and phase noise. If you are listening to this signal on a regular FM receiver it will take on the properties of a beat note.

The frequency of the beat note will be determined by the relative offset and the stability of the oscillators. The strength or amplitude of the beat note is determined by the relative signal strength of the two signals in the overlap areas.

---

Getting overlap areas under control was nearly impossible until the advent of ultra-high-stability oscillators. Although these devices are very expensive , they were far cheaper than any other method of achieving high stability. Even today, they are still the best way of achieving stability.

The second part of simulcasting is the distribution of the data to and through the transmitters. Long after the advent of “hi-stab” oscillators carriers still could not get simulcast to work properly. Two reasons accounted for this problem: (1) distribution of the base band signal to the transmitter and (2) the transmitter itself.

In the beginning, almost all carriers used phone lines for distributing signal to the transmitters. While this approach was a good method of distributing to a single transmitter, it became a disaster when used with simulcast. Without going into mind numbing detail, the problems with phone lines are many. First, line length is a major problem. On a phone line, length equates to time; therefore, the more length, the more time (See Fig 3).

Because Phone Companies cannot guarantee particular path it was always an unknown as to how long the line would be. As the length grew, so did the delay. If one line is twice as long as the other the signal will be delayed by a time that is equal to the length of the line. Being twice as long the signals will be 180° out of phase and assuming about equal signal level, they will cancel each other. Various lengths would produce various phase differences in the overlap areas, which in turn caused varying levels of distortion. Although this can be corrected, there are other factors such as frequency response, envelope delay and just the nature of the copper wires themselves. To make a long story short, phone lines never worked out well for high speed paging formats or voice paging.

Some of the carriers tried light-route microwave but this solution also has its own set of problems. First, it can be very expensive if the transmitter sites do not match the microwave drops, second, most of the multiplex (MUX) in use for "light route" microwave (almost the only type available to use for simulcast distribution) uses a multiplex method known as single-sideband, suppressed carrier (SSBSC). With SSBSC multiplex, absolute phase can be controlled from one end of a channel to the other. However, there usually is no way to control the phase from one channel to the next just as there is no way to control phase from one line to the next with the Phone Company

Finally, some hardy souls tried radio links and things seemed to get better. With radio links there is usually just one distribution transmitter per geographic system. Because most links were phase modulated each receiver was locked to the transmitter which reduced phase jitter. It was a great solution for two-tone paging. However, when five-tone, POCSAG, GOLAY and voice paging was tried the problems started all over again.

The problems had to do with the radio links and the audio circuits of the paging transmitters. For various technical and regulatory reasons these radios had circuits in them known as pre-emphasis and de-emphasis. Again, without going into detail, these circuits had wide, loose frequency and phase responses and could not give good, consistent performance for the higher speed paging and voice paging systems.

In the early ‘80s Quintron (a paging transmitter manufacturer) took a long look at the problems facing simulcast and decided to correct the problems. First we (yes, I worked there) approached simulcast as a system instead of individual pieces of equipment. We then took each piece of the system and matched the electrical characteristics to each other, removing or redesigning any circuits that could cause a problem. We did this for wireline systems and radio links. We then worked out procedures to optimize a system for any type of area (urban, suburban, open country, etc.). The rest is history. Wide area paging systems started popping up all over the country (and a good part of the world). There were even two-way voice systems (some very large). There was such a demand for paging (and messaging) higher speeds were required to handle the capacity.

Today, most of the early type systems have been replaced with newer technologies. While high-stability oscillators are still in use in the transmitters, the rest of the modern simulcast system is new. The analog paging transmitters have been replaced with precision digital transmitters. GPS has allowed systems to use store-and-forward methods that have superior phase and delay characteristics that are far better than any analog system.

Store-and-forward uses GPS to provide a precise timing pulse to synchronize various circuits starting with the NOC system controller and the individual transmitter simulcast controllers. The timing pulses are corrected for geographic position by the GPS system. Data generated in the NOC forms the paging codes and message data and assigns a time slot (a given number of clock pulses after the transmitter controllers receive the data) to transmit that particular information. Each transmitter transmits the page/message based on what the NOC timing has indicated, plus or minus any time offsets programmed into each controller. This method removes any time and phase differences introduced by the distribution medium.

With the advent of store-and-forward it is again possible to use phone lines or microwave or satellite systems, as well as radio links, to distribute the signal to transmitters. Systems can now cover large areas (nationwide, regional) with little or no problems (but setup and maintenance procedures are still required). Although some smaller systems and voice systems still use analog transmitters and link distribution, most have switched to pure digital.

With the advent of modern cellular and its attendant messaging features paging has lost some of its luster. However, hundreds of small systems, as well as some larger systems, continue to provide service to thousands of customers who demand the best possible coverage and assured message delivery.

What started as a simple adjunct service has, for almost 50 years, continued to deliver what the customers want and need. I don’t think paging will ever go away!

Anyone wishing a more in-depth engineering paper on this subject can request it by e-mailing me at dcameron@ttagroup.com I will email you the paper.

Dennis Cameron, along with Bill Hays, own Telcom Technologies Associates, a consulting firm specializing in RF communications. He has extensive experience in high-speed paging, satellite communications, two-way communications, IP distribution, microwave and communications control systems. Most of the last 35 years has been spent in engineering management with most of the time being "hands-on" management. In addition, Cameron has had multiple patents issued in the field of radio communications and has done advanced communications research with the University of Mississippi. Cameron was one of the prime developers of modern simulcasting and has published several papers and articles on the subject. He has designed and implemented many one-way and two-way simulcast systems.