WIRELESS FUNDAMENTALS

By

David W. Russell

Firelinx Inc.

U.S.A.

J. Larry Mattingly

Entertainment Fireworks, Inc.

U.S.A.

8^th INTERNATIONAL SYMPOSIUM ON FIREWORKS

第8国際花火シンポジウム

April 18 –22

4 月18 日-22 日

Shiga, Japan

滋賀県・日本

WIRELESS FUNDAMENTALS

David W. Russell President, Firelinx Inc. PO Box 8274 Incline Village, NV 89452 USA

J. Larry Mattingly
V.P. Marketing
Entertainment Fireworks, Inc.

ABSTRACT

Wireless firing control systems, while an obvious choice for controlling pyrotechnics, have seldom lived up to the promise of the technology. Wireless systems are a complex mix of variables such as bandwidth, frequency, interference, range, protocol, and spectrum. This paper explains these variables in simple terms and details how they interact in a pyrotechnics environment. This paper will show how standard radio protocols are fundamentally flawed when applied to fireworks, and a new distributed processing approach is outlined and demonstrated that can deliver the speed and reliability required for choreographed pyrotechnics.

INTRODUCTION

Wireless firing control systems, while an obvious choice for controlling pyrotechnics, have seldom lived up to the promise of the technology. These failures have not been due to poor radio design, or insufficient power, or improper protocols; they have failed because standard radio systems simply cannot be made sufficiently reliable for the mainstream pyrotechnics environment. This is because wireless systems work like E-Mail – it doesn’t matter when the data gets to the receiver, as long as it gets there. This is fundamentally unacceptable when dealing with fireworks. This paper will outline the technical aspects of Radio Frequency (RF) communications and explain in simple terms those that impact firing system design. We will also investigate the failure mechanisms that render all but the simplest wireless systems ineffective. Then propose a new (patent pending) distributed processing strategy will be proposed that provides a solution to wireless connectivity that can deliver the speed and reliability required for choreographed pyrotechnics.

FUNDAMENTALS

RF communications are a complex mix of variables such as frequency, bandwidth, spectrum, and protocol. All RF communications are subject to interference. Interference comes from natural sources such as lightning and solar flares, but our concern is primarily for man-made sources, of which there are many. Interference may come from microwave ovens, light switches, other radio sources, marine radar, alarm systems, and virtually any electronic device from a toaster to a cell phone. State and Federal regulatory agencies in most countries restrict the emissions from such devices so that in most cases they do not interfere with critical RF communications devices such as police radio, television, or cellular phones. These devices also have a number of strategies that compensate for interference, and high transmission power levels that allow them to punch through most spurious sources, thus enabling long distance communication without noticeable interference.

For most RF systems that could be used in a pyrotechnics environment, specific frequency ranges and power levels have been allocated. These are often referred to as the ISM bands, for Instrumentation, Scientific, and Medical applications. Power is limited to 1 W or less in most cases. One Watt is about the amount of power it takes to light a small flashlight.

Frequency

Frequency refers to how many cycles of RF energy are being radiated each second (Figure 1). Frequency is determined by the regulatory agencies, and so there are only a few to choose from. It is also important because it determines the size of the antenna that is necessary, and which objects will interfere with, reflect, or scatter the radio signal. The inverse of frequency is called wavelength. As the repeating waves of the RF signal are emitted, they speed away from the source at the speed of light. The distance the signal travels in one repetition is the wavelength. The wavelengths of several standard frequencies are given in Table 1.

Bandwidth

Bandwidth refers to how much information, such as computer data, can be transmitted over a radio channel. This is dependent on how the data is formatted and encoded, but the important fact to remember is that the best a channel can do is to transmit data at ½ the frequency, so a channel that has twice the frequency can transmit twice the data, or transmit the same amount of data in ½ the time. This is important because the amount of time spent transmitting data is directly related to the probability that the data will encounter interference.

Spread Spectrum

The ISM bands are not one frequency, but a band of available frequencies. Like a rainbow, this band is called a spectrum. In World War II, techniques were developed to transfer information over the entire available spectrum, instead of just one frequency, in order to defeat enemy signal jamming. Signal jamming is simply an aggressive form of interference. Most radio systems that transmit digital or computer information use these Spread Spectrum techniques to compensate for interference, but it is important to understand how these techniques work. The simplest technique is called Frequency Hopping Spread Spectrum (FHSS). In this system, the radio transmitter and receiver are preprogrammed with a random sequence of perhaps 50 different frequency channels. Periodically, every 1/10 of a second for example, both the transmitter and receiver simultaneously switch to the next channel in the sequence. If there is interference, from another radio in the same band for example, they simply wait for the next channel change and then try transmitting again. This simple technique allows many radios – each with a different random sequence – to share the spectrum with only moderate interference from each other. The interference exists, but the system sees it only as a delay in getting the data through one of the 50 available channels. The data will get through; it’s only a matter of time. Such a delay, however, is detrimental in a pyrotechnic environment.

Group 802

The IEEE standards committee has a group assigned specifically to wireless standards. This committee is designated as Group 802, and various standards for wireless protocols are therefore denoted as 802.11(b) or 802.15.

ISM Bands

Table 1 shows the three major ISM bands in the US, along with the associated wavelength. Some bands are available world wide, others are available only in certain countries, but otherwise similar bands are available in other countries. The 900 MHz band, for example, is not available in Europe, but an 868 MHz band is.

For best reception, the antennas on both the transmitter and receiver should be equal to the wavelength. Acceptable reception is still possible when the antennas are shortened to ½ or ¼ wavelength. The wavelength also affects how the radio energy is scattered by objects it encounters. In general, if the object is smaller than the wavelength, there is little interference. If the object is larger, it can reflect or scatter the RF energy. This is important when considering the venue where a pyrotechnics display is being set up. In the 900 MHz band, objects larger than 300 mm (one foot) will cause reflections or interference, while with a 2.4 GHz radio, much smaller objects such as support poles and even the mortars themselves can cause reflections and interference. With 5 GHz, objects as small as a wristwatch can cause reflections.

Table 1 - Common ISM Frequency Bands

These reflections, as they bounce between objects, can reinforce or diminish the signal at any given point. This is often referred to as multi-path interference. When multiple copies of the RF signal, after bouncing off various objects, converge on a single point and cancel each other out, the receiver may not be able to pick up the signal at that location. This is why most 2.4 GHz radio systems have two antennas, spaced a few inches apart. This allows for at least one of the antennas to pick up a signal at any given position. Multi-path interference is less of a problem with 900 MHz systems.

The 900 MHz band is only license-free in North America, Australia and parts of South America. In most other countries, a virtually identical band at 868 MHz is available. This band cannot carry as much data as the higher frequency bands, however a 900 MHz signal will travel farther than a 2.4 GHz or 5.8 GHz signal. As a result, 900 MHz products are typically used where low data rates are acceptable but long range is required.

The 2.4 GHz band was opened in the 1980’s and is available globally. Initially the band was lightly used, and had little interference compared to the 900 MHz band. Today there are many devices operating in this frequency band, most notably 802.11 devices and microwave ovens. Most other 900 MHz devices such as wireless telephone handsets have since migrated to the 2.4 and 5.8 GHz bands. Depending on the location, the 2.4 GHz band can now be more congested than then 900 MHz band. The 5.8 GHz band is not yet available without license worldwide, but it provides very high data rates.

RADIO AND PYROTECHNICS

To understand how each of these factors interacts in a pyrotechnic environment, first consider a simple exchange of data between a transmitter and a receiver (see Figure 2). Inside the transmitter, a message is constructed called a packet. The packet often includes information about the transmission source, the address of the intended recipient, the length of the information being transmitted, and the data itself. Once the packet is constructed, it is transmitted. There may be many potential receivers in the field, if the packet was not addressed to it, the receiver ignores the packet data. The receiver that matches the packet address accepts the data and in turn transmits a short packet called an ACK, short for Acknowledge (Figure 3). This packet lets the transmitter know the data was received successfully. The transmitter continues constructing and transmitting the

next packet.

What Could Go Wrong?

Due to a multitude of possible interference sources, it is not uncommon for the transmitter’s data packet to experience interference (Figure 4). This could be anywhere from a single bit of the data being corrupted, to the entire channel being jammed by another transmitter during this time period. In any case, no receiver can accept the corrupted signal, and no ACK packet is returned. The transmitter cannot know that no data was ever received, so it must now wait for a significant period of time to be sure that something went wrong and the packet did not reach its destination. The transmitter then sends the same packet over again, and again waits for the ACK.

It is also possible, however, that the packet was received, but the ACK that was supposed to go back to the transmitter Figure 5 - Acknowledge Interference experienced interference (Figure 5). In this

case, the data did get to the receiver, but the transmitter did not get the acknowledge message, and sends the data again anyway. The receiver now has the added complexity of not only receiving the message, but it has to be able to detect that it is a duplicate message, send the Acknowledge, but ignore the packet.

This is where the bandwidth becomes important. If the frequency of the RF signal is high, the data can be sent in a smaller time period, reducing (but never eliminating) the probability of interference striking just at the moment of transmission. It can never, however, completely avoid the primary source of interference, other transmitters in the same band. These interference sources cause significant delays because the transmitter and receiver may have to delay for several cycles to find a clear channel. In 900 MHz and 2.4 GHz systems, the frequency of the system usually does not significantly affect the length of time spent in one frequency-hopping channel. In 5.8GHz systems, there are more available channels, so the time spent on each may be reduced somewhat.

In a pyrotechnic environment, then, the problem arises when the technician pushes a button or the computer timing the firing sequence that initiates a command to fire. The command could suffer interference, the Acknowledge could be lost, or both, and the frequency hopping system might need to skip several cycles to find a clear channel, resulting in a series of delays, the length of which cannot be predicted. This scheme is shown in Figure 6.

MASTER/SLAVE ARCHITECTURE

From the simplest nail board to the most sophisticated computerized system, most firing controllers are based on what is known as a Master/Slave architecture. In this arrangement, the Master is a person or computer who decides when to fire. When that point is reached, a command is transmitted or a switch is thrown, and the slave reacts by firing. In the simplest system, the technician is the master and the E-Match is the slave. In the most sophisticated systems, the computer is the

master, it transmits coded messages (by wire or radio connection) to a box in the field, and the box switches current to the E-Match to initiate firing.

The master is always in control of what is being fired when. The slave responds to commands it is given. This is a very dependable arrangement and has been used successfully for many years. When adding radio communications to this model, however, a fundamental flaw is revealed. The Master/Slave system is based on the assumption that communications are flawless between the master and the slave. In an RF system, because of the intermittent delays that have been shown to be a fundamental part of radio communications, this is no longer true.

In existing wireless firing systems measures have been taken to mitigate this fundamental flaw. In some systems, mostly manually fired controllers, the problem is ignored – the user simply holds the firing button down and if it takes a second or two longer for the shell to fire. No one cares. In other systems where timing is more important, the solution is usually to go to more expensive, higher bandwidth radios to minimize (but never eliminate) the delays. Unfortunately, the more the radio is improved to minimize the interference, the more expensive it gets.

DISTRIBUTED PROCESSING

The conclusion that must be reached is that it is impossible to build a radio system that is free of interference-caused delays. The solution to this fundamental flaw in radio communications is not to improve the radio, but to build a system that is fundamentally reliable, even if the radio is not. This requires a radical change of the basic architecture of the system, known as Distributed Processing.

In Distributed Processing, any number of equally intelligent units act together to accomplish a common goal. In a firing system, this is accomplished by replacing the Slave field units with any number of independent Master firing units attached directly to the E-Matches. There is no single master giving orders, but a collection of independent units acting in concert to fire the display. To understand the difference, consider a simple example:

Table 2

a few cues, this delay might not be noticeable, but if the command were trying to fire several hundred pieces of product simultaneously in a front, for example, the delay could be noticeable and significant.

In a Distributed Processing system, each of the firing controllers is informed before the show ever starts, of the cues and time it will be firing. Module 1 knows it will fire a cue at 1 second, Module 2 knows it will fire a cue at 2 seconds, etc. Once the clocks on the independent modules are synchronized, they can all fire independently, but the result is as if they were all interconnected. All of the modules can now fire simultaneously – each independent module knows to fire at the same time without commands that could be delayed.

Most importantly, this removes the radio from the critical communications path, resulting in a system that is reliable, even though the radio is not. This has a secondary benefit of allowing slower 900 MHz radios to be utilized, resulting in significantly less expensive radio systems with greater range and less multi-path interference.

SYNCHRONIZATION

The key to Distributed Processing is intelligent controllers and synchronization. Here is how the system works in a pyrotechnic environment:

As the display is being set up, shells are dropped and the E-Matches are wired to the Firing Module – this is an intelligent, independent, master firing controller wired directly to the E-Matches it will fire. There are any number of independent Firing Modules in the display, but they are being wired to a cue list that was designed so that each module has a defined set of cues to which it will be wired.

When the Firing Module is activated, it performs automatic self-test diagnostics, then it establishes communication with a Command Module, held by the pyrotechnician. The Command Module is not the Master controller, but simply a user interface device that simplifies the technician’s command and control of the array of firing modules.

When the firing module establishes contact with the Command module, the Command module automatically downloads that portion of the display cue list that will be fired by that individual unit (Figure 7). In this figure, two clocks are shown on each Firing Module. The topmost is the running time clock, here shown as not running. The lower clock is the firing time that was designed for this cue in the choreography. This time is relative to the start of the show. This is repeated for all of the firing modules in the display. Because this is being done before the show, there is plenty of time for dropped packets, interference, and other delays. Once the cue list is downloaded to the individual units, there’s no longer any need of split-second commands.

A few minutes before the show, after continuity checks and communications checks, the technician issues the Arming command via the Command Module (Figure 8). At this time, all of the independent firing modules synchronize their internal clocks to the Command Module’s clock. In this figure, the clocks are synchronized and started at 00:00. This is a complex process, but the result is clock synchronization between hundreds of independent modules to 1/1000 of a second. This has the added benefit of providing the pyrotechnician with control of the primary clock source, so that he can control it as he sees fit, nudging the timing forwards or backwards to compensate for shell break times, pausing, or adding manually fired cues into the performance.

Once the clocks are synchronized, all firing modules are now ready to act in concert to fire the show. When the technician issues the Firing command (Figure 9), he does so either synchronized to a music system about to start, or with a few seconds of built-in delay. This final delay from automatic start command to firing again allows the Command module to issue the final firing commands and verify receipt from all modules without worrying about dropped packets or delays. This delay allows the Firing command to be non-time critical as demonstrated in Figure 9. Since the clocks have been running since they were armed, the upper clock on all modules now reads 10:05. The delay in this case is set at 56 s. When the Firing command is given, the command includes a timed offset to when in the future, relative to the already synchronized running clock that the show will start. This offset is independent of any delays in transmission. This offset and the current time is added to the cue start time previously loaded, and now the lower clock shows the exact time, relative to the synchronized running clock, that the cue will fire. The current time of 10:05, plus five seconds of delay, plus the firing offset of :01 gives a final firing time of 10:11.

Once the firing command is given, the firing modules are firing their timed cues independently under local control, without the need of any further commands from the Command module. Again, the radio is removed from the critical path. This procedure can be repeated for virtually any number of firing circuits within the module.

The Command module does, however, maintain several levels of communication with the firing modules, to allow the user to adjust timing, issue emergency stop commands, or Operator Presence Detection (OPD or Deadman Switch functions), fire manual cues, etc. All of these functions, however, are less time critical than the split-second timing commands needed to fire the show choreography.

In addition, just because the Firing Modules are designed to automatically fire their programmed cues, it does not mean that pure manual firing is not possible. A purely manually fired display is still possible, although the wireless connection would then still have the unpredictable delays of a Master/Slave system. It should also be noted that RF communication is Line-Of-Sight technology. If a hillside or a steel pillar completely blocks the path from the transmitter to the receiver, it will not receive a signal. Due to these circumstances, and other potential situations where RF communications may not be viable or appropriate, we recommend that a backup wired communication path be implemented in the modules as well. This backup path need not extend all the way back to the Command module, although it could. It should be sufficient that the wired communication link could be extended only between a module that is blocked, and the nearest module that is receiving RF communications, acting as a bridge.

This combination of wired, wireless, automatic, and manual commands offers unparalleled flexibility in the performance art of a display. It could be thought of as that portion of a symphony where a solo artist takes center stage. While the automatic Firing modules are taking care of background shells and pre-staged effects, the operator would now be free to improvise and add his performance by manually firing at the same time. In addition, because moderate distances are now possible, the operator need not be directly under the display unable to see the result, but further back at a safer and more productive distance.

QUESTIONS TO ASK

For anyone contemplating the purchase of a firing system with wireless capability, several questions should be researched to determine if the system will meet you needs:

·What is the frequency band? The higher the frequency, often the more expensive the radio, and the more objects that will reflect and scatter the signal. Is the frequency band appropriate for your country?

·Does the system use a standard protocol like 802.11? Standard protocols are built to

trade time for reliability.

· If I fire a large number of simultaneous cues, will there be any delay?

· Can I fire automatically and manually at the same time?

·What is the power of the transmitter? For most professional systems, the answer will be 1

W, the maximum allowed by regulatory agencies. If the system has less power than this,

it may not be using Spread Spectrum techniques or it’s range may be limited to only a

few meters.

· Is there a limit to the number of field modules that can communicate with the controller?

·Is the radio operating in a Spread Spectrum mode? If not, then the system will be extremely vulnerable to interference, and may not work at all in some areas such as marinas or near airports, military installations, etc.

CONCLUSION

We have shown that traditional Master/Slave architectures are fundamentally flawed when implemented with radio links between the Master and Slave, due to unpredictable delays in the radio communications link. One solution to this problem is to build a firing system based on Distributed Processing, thereby removing the radio link from the critical path. This has the benefits of less expensive radio systems, longer range, virtually unlimited numbers of cues and modules, totally portable equipment, greater safety and control for the operator, and less expense and time spent wiring and setting up the display. The only drawback is a short mandatory delay before automatic firing commences.

HOME