RF engineer, Steve Caldwell walks us through the difficult challenge of coordinating RF at one of the world's largest major events, the Olympic and Paralympic Games.

Foreword

There are few major global events bigger than the Olympic Games held every four years. More than 11,000 athletes took part this year across 306 events in 28 sports — making the Rio games (from a sporting perspective at least) even bigger than London 2012. On top of the main sporting attractions, it's easy to forget how much additional work goes on behind the scenes to help make sure the entire event runs smoothly from a technical and entertainment perspective. The opening and closing ceremonies, for example, require major planning and real skill to execute smoothly. Millions of viewers tune in to watch these events stream live, and with such high expectations, there is little (if not zero) room for error. Few realise just how much RF technology is running behind the scenes at such major events. It's safe to say, much of it is practically invisible — unless, of course, something goes wrong. In such mission-critical applications, correct RF coordination and planning is imperative. To give us a flavour of the task at hand, Steve Caldwell walks us through the process he used to ensure the show at Rio went off without a hitch.

Preparation

Preparation for Spectrum Management at the Rio 2016 Ceremonies began in February 2015, when the Technical Director of Ceremonias Cariocas (CC2016) got in touch. CC2016 were the organising body for Rio 2016 Olympic and Paralympic Opening and Closing Ceremonies. Knowing how challenging the process of managing spectrum at major events can be, CC2016 also approached representatives from ANATEL (the Brazilian spectrum regulator) to help manage this difficult resource. Being involved early in the proceedings (about 18 months out) allowed us to apply for — and ultimately obtain — ideal spectrum for use at the Ceremonies. Also, we were able to determine the best possible equipment — based on the spectrum acquired — long before any equipment tender process. Lastly, and perhaps most importantly, an early start allowed for a relationship to form with the spectrum regulators. A visit to Maracanã Stadium (situated west of Rio de Janeiro city) was arranged during the early stages. This visit allowed spectrum scans to be taken at various positions. Also, it enabled me time to meet with the Spectrum Regulators so we could express our requirements and determine spectrum availability in Brazil. ANATEL were extremely flexible in allowing us access to several bands that aren't normally available. Also at that time, an online spectrum request portal was set up by a third party (a private Brazilian management company brought on to manage auxiliary requirements). This portal would allow all the broadcasters — and anyone else wishing to obtain frequencies during the games — to log spectrum requests. Once all information was collected — including their intended location of operation, equipment capabilities, and time frame — frequencies were assigned and licenced. Even though a request cut-off date was specified, requests were still received and processed by ANATEL right up to the Opening Ceremony. This represents an outstanding commitment by the regulator.

Managing the Spectrum

After a week in Rio — scanning, meeting authorities, getting burnt on Copacabana beach, more scanning — I returned to Sydney and began the process of determining which equipment was best for the available spectrum. From this process, I was able to compile a frequency solution. The systems I was responsible for — both from a spectrum management point, and a compatible frequency standpoint — was essentially anything RF related destined for use in or around the Ceremony stadium and rehearsal site. The rehearsal space was less than a kilometre away from the stadium on the other side of the railway line, as it were. Anyone that visited the rehearsal site will understand the "other side of the tracks" reference. Some of the systems that required spectrum management for use by Ceremonies production alone included:

• 30+ channel pairs of semi-duplex analogue UHF land mobile two-way radio (60+ frequencies)
• 30+ channels of simplex analogue UHF land mobile two-way radio
• 80+ channels of radio microphone (digital and analogue)
• 30+ channels of IEM
• 14 channels of FM broadcast (out of band)
• 3 channels of FM broadcast (in band)
• 10 channels of 1900MHz DECT for digital full duplex mobile communications
• 4 channels of 400MHz telemetry for Pyro firing
• 4 channels of 400MHz telemetry for special effects control
• Telemetry channels for multiple weather reporting systems
• In house TETRA systems
• Countless Wi-Fi devices

This equipment was just for the Opening and Closing Olympics and Paralympics Ceremonies technical production (audio, lighting, comms, etc.). This list does not include the 300 or so Broadcasters, ENG crews, and other interested parties. By the time Ceremony rehearsals began in the venue, there were between 2,500 and 3,000 registered frequencies licenced for use. The spectrum requirements were bigger than London 2012, by far!

Brazilian spectrum

One of the first things I noticed upon landing in Rio (before I'd even so much as touched a scanner) was the huge variation in television broadcast quality from my hotel room. For the first time in many years, I had "snow" on some channels. This could only mean one thing – they still had Analogue TV. After a bit of investigation, I found that I could tune to the same broadcast in both analogue and digital. SHIT! Upon breaking out the analyser, I confirmed what I suspected: Rio (or more likely the whole of Brazil) was still in their analogue to digital transition period. For every analogue TV station that existed, there was probably a digital equivalent transmitting concurrently. Calling the Brazilian Terrestrial Television Band crowded is a massive understatement. As of February, from the available 37 channels in the UHF TV band (US 6MHz raster), only 14 remained unused. The meeting that followed with ANATEL then revealed that of these 14 currently unused channels, 5 were slated for the energising of additional transmitters between now and the beginning of the games. We only had a total of nine 6MHz wide TV channels worth of space between 470 and 700MHz. As ANATEL were very generous in allowing us access to other spectrum, we were also allowed to use the free space between 700MHz and 806MHz, as their digital dividend was not yet finalised. Luckily, it was only radio microphones, IEM, and any UHF style duplex communications devices that would need to use this TV spectrum. And fortunately, because I could specify the equipment and frequency range that would be used for the Ceremonies production, I could move any of this equipment up into 700MHz. To complicate things further, all 42 of these existing — or soon to be energised TV transmitters — were positioned on a hill less than three kilometres away in very plain sight of the stadium. The measured field strength of most of these stations in the centre of Maracanã was more than -60dBm. Brazil utilises the 450-470MHz band for its LMR two-way radio systems, and to say this portion of the spectrum was packed is another huge understatement. Fortunately, the Spectrum Regulators approved our request to use the 403-450MHz portion of the band for our production two-way radios, as we were essentially banned from using the conventional 450-470MHz band. This change presented challenges for the prospective communications contractor when providing the quantity of handheld radios required by Ceremonies, as equipment in this alternate lower band is less common. I also recommended — after providing significant precedence — that the use of digital trunking systems was not the best solution at Ceremonies for several reasons; not least being the latency experienced when using handhelds in earshot of comms panels. All two-way radios had to fit into a space in the spectrum wide enough to accommodate 90+ channels of 12.5kHz bandwidth analogue carriers. As all the transmitters used in the base radio systems for duplex and simplex were cavity aligned; we could not place those frequencies any closer than about 300kHz. The use of analogue carriers also allows for faster fault finding, and resolution of frequency related issues. The FM Broadcast mass cast monitoring system was the next to receive its custom placement. During events of this size (in particular Olympics Ceremonies) there is a system employed called the Mass Cast Monitoring (or Level 3 IEM’s as it’s known in the industry). It is essentially one or more FM Broadcast transmitters set up to provide coverage of the entire performance (and/or rehearsal) spaces. Every single cast member involved in the show wears a low-cost FM receiver and earbuds. This process mainly provides time-zero program to the cast, but can also be used to deliver choreography cues, instructions, or click tracks. In Rio, we used six FM transmitters for the main stadium, and four FM transmitters for each of the two rehearsal spaces. Due to the (once again) very congested 88-108MHz FM Broadcast band in Rio, we requested — and were subsequently authorised to utilize — a portion of spectrum between 76-82MHz (US VHF TV ch5). This allocation allowed us to obtain sufficient spacing between each of the 14 FM frequencies to compensate for the poor adjacent channel rejection exhibited by the low-cost FM receivers. A total of 12,500 of these FM receivers was required, so each receiver had to be as low cost as possible. The only way to ensure adequate performance in this environment was to expose them to as little other RF energy as possible, while maintaining reasonable channel separation. The ability to occupy spectrum so far from the normal Broadcast band was invaluable. It also prevented the Ceremonies rehearsal audio program from being exposed on a public broadcast band. There was also a requirement to provide 3 FM transmitters inside the normal FM 88-108MHz Broadcast band for use by transcription services during the ceremonies. However, as these transmissions would be received on somewhat higher quality receivers (most mobile phones sold in South America are fitted with FM receivers), adjacent channel rejection for these services was not an issue. As such, we managed to shoehorn three FM transmissions into the crowded band. Planning the spectral layout of equipment was next. It was obvious which position the two-way LMR systems would occupy; the same goes for other equipment with restricted frequency options and limited switching bandwidth. Telemetry systems used for various functions such as pyro firing and special effects control were all limited to the 430MHz ISM bands. Wi-Fi devices (however bad an idea) were fixed, as were government and security TETRA systems used in the venue. The DECT systems adopted for the digital full duplex comms systems had some latitude, due to the fact we could arbitrarily choose the DECT bands we would like to occupy. The DECT band varies from country to country somewhat, and most hardware options were available (mainly by programming) in any of the bands. FM broadcast band transmitters were assigned their own private spectrum space, so the only systems left to assign were the radio microphones and IEM systems, for both Ceremonies and Broadcasters. We only had the nine free TV channels between 470 and 700MHz to use, however, some of the 700MHz band was at our disposal. Pictured Above: The Spectrum Landscape from 30 to 950MHz - as seen when entered into Shure WWB 6.12 A decision was made early on to allow 30+ channels of IEM transmitters to occupy the 700MHz spectrum. The one drawback of utilising this spectrum was the recent global trend toward reassigning this band for use by TELCO’s. As such, the availability of equipment in this frequency band is becoming somewhat limited. It's not such a problem if you're only after 10 or 15 IEM receiver packs, but when the specification calls for 370 IEM belt pack receivers fed from 24 channels of transmitter, it's a little trickier. This drawback was outweighed, however, by the advantages of placing the IEM systems up in this higher frequency spectrum, away from the TV transmissions and the majority of radio microphones. With the above all sorted, I had the remaining nine clear 6MHz TV channels between 470-700MHz to position 80+ channels of Ceremonies Production and Broadcasters radio microphones — plus any UHF ‘BTR’ style full duplex comms systems. Later, it was agreed to move the newer digital audio full duplex comms systems to the 1900MHz DECT band, which left more space for the radio microphones.

Calculation, Interference, & Intermodulation

It's now fairly standard practice (through a collective effort I've supported) to serve high priority spectrum users with the frequencies they need — within the limitations outlined by the regulator — before allocating spectrum to other parties. As the Olympics Ceremonies environment was a live global broadcast, live radio microphones and IEM’s are considered the highest priority. Then in a decreasing order of priority; Ceremonies communications, rights holders’ radio microphones and IEM, rights holders’ communications, and then, Broadcasters radio microphones, IEM, and communications. Essentially, this allows the Ceremonies radio microphones and IEM frequency setup to be the most resilient solution possible. All other systems down the priority list are calculated AROUND each of the higher priority systems that precede it. As other systems (such as broadcasters’ radio mics) are considered arbitrary, there is no preference given to its position based on usage. For instance, a broadcaster IEM (or IFB in the relevant nomenclature) is treated the same as a broadcaster radio microphone. At this level, they do not have the different interaction characteristics that form the basis of how we organise higher priority systems. Once a frequency solution for the Ceremonies is calculated (considering all pre-existing contributors and the cautious inclusion of analogue TV), it is submitted to the Spectrum Regulator along with similar calculations for LMR two-way radio systems. From this point, all other frequency requests are then assessed, calculated, and provided to the applicant. Most often, the spectrum regulator will seek advice around the best way to calculate and resolve the applicants’ requests, without compromising the integrity of the Ceremonies RF solution. The process of calculating frequencies required to fulfill an event like this consists of several stages. The actual procedure is somewhat proprietary, and well beyond the scope of this article. However, the process is essential to provide a suitable and stable solution for this hostile environment. Any calculations must place a high emphasis on device usage at the frequency assigned to it. There are several ways to weight these calculations and allow segregation based on their usage. These processes allow for a more efficient use of the spectrum; some of them are outlined below, with advantages:

Spectral segregation	Utilising different frequencies in the same physical location at any one time	The most common and reliable form of segregation. The ability to monitor each carrier independently allows security. Carriers not part of an RF solution can easily be detected.
Spatial segregation	Utilising the same frequencies in different physical locations at any one time	Used most often for solutions that have high numbers of carriers over a large physical area. Cross-contamination of segregated areas more difficult to detect.
Temporal segregation	Utilising the same frequencies, in the same physical location at different times	The least reliable of all three methods, as it requires carriers to be turned on or off at pre-defined times. There is no 2nd level of safety for this form. Contamination is a high risk.

* Standard IMD calculations are used in conjunction with the above weightings The three different types of segregations above carry various factors of reliability, and can be used separately, or together with one-another, depending on the circumstances. Due to the relative proximity of the Maracanã to the other venues involved in the RF solution (such as the rehearsal venues), these calculation procedures required great care. For this event, no fundamental frequencies were re-used at any stage over the period of 4 months — except for the same piece of equipment used during each of the four Ceremonies. Only intermodulation products and their proximity to fundamentals were calculated with the three assessments outlined above; frequency, position, and time. The level at which each piece of RF equipment was calculated — in respect to fundamental spacing requirements and susceptibility to IMD — was very much tailored to each type of device, considering both transmitter based and receiver based IMD. Eventually the 270 odd frequencies for Ceremonies alone were submitted to ANATEL. From this point, the remaining 2000+ frequency requests from other users would be resolved. It should be noted that a percentage of the frequencies requested by a good portion of broadcasters are what are known as ‘land grabs’. Essentially, at the time of request, the broadcasters often don’t know what equipment they will be using at the event. The easiest way to ensure they still have a choice of frequency bands when the event comes about is to make a request for multiple frequencies in all portions of the spectrum. This way, they will have access to licenced frequencies regardless of the equipment they eventually use. Even though there is a good percentage of frequencies licenced that will never be used, at the time of calculation, there is no way to determine which ones, and thus, the situation must be calculated assuming they all will be used. Pictured Above: 'Fat Albert' wideband receiving antenna in London

Monitoring the RF Environment

Monitoring the RF environment, or spectrum ‘landscape’, was the priority upon arriving back in Rio by late May 2016. I took over three large pelican cases of spectrum monitoring equipment and antennas with me on the flight, and this, of course, attracted a bit of attention at Rio international airport customs. Opening the cases prompted the appearance of at least half a dozen customs officers all seeking explanations but lacking the English language skills to comprehend any of my technical explanations. In fact, the more I tried to explain what the equipment did, the more dubious it all sounded. Luckily, I had with me a letter of introduction from the Olympics Ceremonies organisers outlining the conditions of my 6-month work visa. This gave a little credibility to my presence, if not actually explaining what the equipment did. They seemed happy with this, and I spent the next half hour re-packing all my shit. Arriving at the stadium a couple of days later, I began my search for a position in which to set up my analyser antenna. One of the most important things to consider when monitoring spectrum, is not the absolute amplitude of measurements taken, but the relevant changes experienced over time. The one thing that will ruin all your relative measurements is the re-positioning of your measuring antenna. I knew that I would not be able to find an antenna position that would serve me for the following 4 months of stadium set construction, lighting and audio build, rehearsals, ceremonies, and of course periods in between when the entire stadium is handed over to the Olympics sports. For now, any position that would get me through the build period, or at least until my larger, less portable antenna turned up in the sea freight, would be great. I eventually set up the antenna outside a glass control room at the rear of the upper seating in Maracanã. This room was slated for use by the live ground announcers during the ceremonies. This position would allow me to monitor the spectrum over a significant period, and ensure that all the calculations I had done, and all the frequencies submitted and licenced, would still be good to use. For the next couple of hours, my eyes were pinned to the screen. If there was an unexpected and significant change in the spectrum landscape from what I had measured 15 months ago — or at least accounted for based on information from ANATEL — then we could be in trouble. By this time, all the Communications, Audio, and a handful of other relevant contractors had been appointed. All these contractors had provided equipment to fit tender that was in turn determined from the spectrum as known. It was this point where we’d find out how much had changed. Luckily, except for a few analogue TV stations that had shut down early, there was no discrepancy. Thank God. The weeks passed, with very few remarkable changes in the spectrum landscape. Eventually the real monitoring antenna (Fat Albert to those who have been introduced) arrived in the sea freight from Sydney, and I went about finding a more permanent position in the stadium — one that would not be compromised from now until the end of the event. The antenna was placed in the upper rows of seating, directly in front of the Ceremonies control room. This was an excellent position, as it was high enough to be sensitive to any RF activity outside of the stadium, but far enough from all the higher power LMR two-way, IEM, and FM transmitter antenna positions. Nothing upsets a landscape scan more than several high-power carriers mixing in the front end of your wideband receiver.

Equipment and Policing

The equipment that I use to monitor the RF spectrum is nothing particularly exciting, nor expensive. I have found over the years that it’s not how big it is, it’s what you do with it. I have seen Spectrum Authorities bring very expensive AOA and TDOA receivers into these stadiums over the years, without the necessary skills to measure even a single carrier. Almost any reasonable spectrum monitoring device will be able to provide the information required to competently monitor the spectrum landscape. It’s not about absolute measurements, it’s about relative measurements. You don’t need to own a wideband Bi-Conical or Discone antenna with a ruler flat bandwidth from DC to daylight, so long as you use the same antenna, in the same position, you will be able to detect changes in the spectrum. The other thing is getting to know your equipment; is it deaf (has low RF sensitivity) at certain frequency bands? Does it have significant Image Response issues that need to be accounted for? Is the noise floor you are seeing measured or inherent? Once you are familiar with your receiver, you will be better equipped to know what is REALLY going on. My weapons of choice for long term monitoring are not particularly expensive, nor are they difficult to use. I own two devices from the WinRadio stables in Melbourne, Australia. The first is a WinRadio G33WSM, which is probably the most common SDR (Software Defined Radio) receiver in this industry. It is an original Beta test unit sent to me by the guys at WinRadio almost ten years ago, and it’s not missed a beat. The second is a WinRadio G39DDC, which is another SDR unit, but as an entire 16MHz of RF bandwidth is sampled in real time, it’s capable of some amazing tricks. Pictured Above: RF Monitoring screens. WinRadio receivers centre of image One of the best things about both units (especially the G39DDC, as the RF, IF, and audio filtering is completely adjustable in real time), is the ability to monitor the audio that may be modulated on almost any carrier. Over the years I have seen a LOT of people — mainly spectrum police from local authorities — come in with very expensive fixed and portable direction finding devices trying to find a rogue transmitter. The thing about direction finding in a stadium, is the reflections can lead you on a very wild goose chase. Even with advanced Time Difference of Arrival (TDOA) devices. These RDF systems have their place in closer proximity, and I have used them on many occasions. The best procedure I have found for finding a rogue transmitter is not to physically go searching for the RF carrier, but to listen to the carrier. You can find out so much information about the transmitter from the audio demodulated from its carrier, and even the carrier itself. Are their people talking on this carrier? Are they reporting the news? Can you hear the event PA system in the background? Can you hear anything that might give away who they are, and where they are transmitting from? Now look at the carrier itself – does it have side-tones, subcarriers, or audio tone-keys that give away the type of transmitter? Is the carrier on a frequency that suggests a brand or model of transmitter (if the carrier is on a 100kHz frequency step, you have a better than 4:1 chance it’s a Lectrosonics). If there is a 19kHz pilot tone, then chances are it’s an IEM/IFB unit. Most equipment has a signature that gives them away. Even determining the brand of the equipment will go a long way toward who might be using it, and where they might be. Another trick I use often when tracking down rogue transmitters, is to use all the receivers at your disposal. If the rogue transmitter is on a frequency that your radio mic system is capable of (there is a very good chance your system is, as this is probably how you detected its presence in the first place), tune in one of your receivers. Even if you get no audio, knowing the signal strength indicated by the receivers’ antennas, and the type and orientation of its antennas will give you a lot of useful information. Olympics ceremonies use multiple receiver ‘nodes’ positioned about the stadium for redundancy. These all have their own directional antennas, in different positions around the stadium, all pointing in different directions. A Lot of position information will be revealed by these different antenna/receiver positions. The one thing to remember when put in charge of spectrum — either from a spectrum planning, or monitoring and policing perspective — is that most people you find who are using incorrect frequencies won’t even know. The last thing you should do is to track them down, and turn up with a handful of analysers and licence listings while reading them the riot act. Most of the time, they are completely unaware of the frequency their kit is using, and even more unaware of any spectrum planning in place. 9 out of 10 people I have found in this situation are happy to be educated, and will do as you ask; a good portion will ask for your help. There is no better option than to ask them if they would like to be given a suitable frequency to use. After all, if you include them in the plan, it’s better for them and you. Get their phone number, get them to call you if they have any issues, and most of them will become another set of eyes for you. After all, any other rogue transmitters have just as much chance of hitting them as they do you. When I do a spectrum solution for an event — even one that does not necessarily attract the services of the local Spectrum authorities — I always include a handful of the most popular bands used for Broadcaster and ENG equipment in the frequency solution. This way, I already have a pool of compatible frequencies to hand out. This simple additional has saved much time and effort on so many jobs. Most of the broadcasters that are recurring on these events often drop me an email long before the event to obtain the situation on spectrum availability, and often ask to be included on my solutions.