APRS in the Sydney Olympics
APRS in the Sydney 2000 Olympics
I cannot remember exactly when it was, but I remember the day wellÖ I was at work early making sure that everything was operating properly, and my cellphone rings. Caller ID says phone number is not available, which is unusual, but not too so. Answering the phone the guy on the other end said he was calling from the USA, and asked if I wanted to work on the Olympics, doing APRS?
The guy on the phone was a Ham from central Pensylvania, and
worked for Winemiller Communications who I found out later was the principal RF
contractor for the Sydney Olympic Broadcasting Organisation. I aggreed to get
whatever maps I had available together and copied to CD-ROM for Jeff Winemiller
when he visited Sydney a few weeks later. This article is all about what
Sydney has a very poor RF environment ñ there are no great
tall buildings that you can see from everywhere to point a microwave receiver
at. From most parts of Sydney you cannot see the CBD nor the mountains to the
west from ground level. This causes lots of problems if you need to get video
signals from street level back to a control room. Although the AMP tower (Centerpoint)
looks tall from inner Sydney, when you go move towards Olympic Park, it
For an event such as the marathon, you could always use
cables, but can you imagine laying the video cables for the entire 42 Km of the
Marathon? This left one option, wireless video cameras with some sort of a
repeater system to get the microwave signals out to the rest of the world.
It would have been possible to put cranes up throughout the
city with microwave receivers but this would not have looked very good. It also
assumed that there would be line of site to the crane which would not have been
possible all the time.
So the only way to get line of site for the microwave signal
was to have a repeater that somehow followed camera vehicles as they went along
the course. The only solution that worked technically was to have helicopters
follow the vehicles as they followed the course.
Three motorcycles and a lead vehicle providing live video
and audio from the course.† With microwave antennas pointed straight up from
ground level, a helicopter would need to be well under one KM from the vehicle
it was receiving in order to get a decent signal.
But there is no way that you could keep the four vehicles so
close together during a race so having a single helicopter was not really an
option. There was also the problem of signals potentially interfereing with
each other in the receiver, and also what happens when the helicopter needs to
More than one helicopter was needed, but how many more? It
turned out that the best solution was to have a helicopter assigned to each
vehicle. That way when a helicopter needed to refuel two vehicles would need to
stay close together, sharing the one helicopter until the other one returned.
The Auburn control room was one of the hidden sites during
the Olympics. Security was almost totally non-existent and parking was usually
plentiful. There were certainly no security checkpoints or parking
restrictions. Located on the top of the old nurses home at Auburn Hospital,
no-one knew we were there and we liked it that way.
During the games the building had been housing health care
workers for the games, and many of these had the night of the closing off work.
That evening many of them stumbled across our control room on their way to the
roof to see the fireworks. To say that they were stunned was an understatement
when they saw what had been going under their noses during the games.
Not that we had not caused problems of our own. So that we
could safely have our people working on the roof, we got scaffolding put up,
including new stairs. This scaffolding could be seen from the Olympic Stadium
it was so visible. But the problem with the scaffolding is that it tends not to
create problems when you build it around a TV antenna, which is of course what
had been done.
On the day of the opening ceremony, someone worked out that
all except for one TV channel worked ñ the one channel that did not work was
the Olympic broadcaster, Channel 7. So with just hours to spare we move the
antenna for them, and even replace the co-ax which looked as if it had been put
in with when the building was built in the early 1960ís. [As an aside, there is
evidence to suggest that the bricks for this building came from the State Brick
Works that were located in the middle of the Sydney Olympic Park site]
Our control room was actually located in a plant room above
the 6th floor of the building. The plant room contained two large
copper water tanks that were used to maintain mains pressure water in the
building. These water tanks used up about a half of the room, and we had the
other half. The entire room was about 5m x 5m. Not very large when you consider
that the usable space was about 2.5m x 5m, and in that space we placed about
ten 19î rack cabinets, and had another one positioned between the two water
tanks. This room also contained a PC for my work, and up to about seven people
working, with spectators occasionally.
Each rack contained at least one video monitor ñ I think we
had about 31 video monitors in† this small room which required some squeezing
inÖ Many of the racks contained monitors four high. This created a video wall
which would have been impressive if you could see it all. But as I mentioned
before size was limited. Most of the monitors were on the 4m wall, with about
0.5m between the backs of the monitors and the wall. When you realize that most
monitors are about 0.5m deep themselves, the 2.5m usable space becomes 1.5m.
This does not leave much room for people wanting to watch the screens. It was
often so cramped that anyone moving required 3 or 4 other people to move to
allow you get out of the room.
There were a couple of telephones in the control room, along
with three links direct to the repeater system. The telephones were used
extensively, but there were two main phone numbers next to the main phone.
The first number was an emergency number in case of Hijack
or Air Piracy of an aircraft ñ which was particularly absurd considering that
the aircraft would only have a pilot and a technian in them, and could not
really seat a hijacker. Besides which, the technician worked for us, and it was
unlikely that the pilot would hijack is own plane. Still, we kept the number
there just in caseÖ
The second number was far more useful ñ it was the number of
the local Pizza Hut. They got to know use well. The best order was for 9 large
and extra large pizzas, and about 16 liters of softdrink for lunch on the day
of the Menís cycling. The order was so large they let me use the back entrance
to the Pizza Hut normally reserved for their delivery people.
As I mentioned before, nestled between the two water tanks
was a 19î rack. This rack had 23 fibre optic cables coming in from various
sites thanks to our friends in Telstra, as well as one underworked UPS in case
we ever lost power. Each fibre contained a 230 Mbit/Sec data stream of
uncompressed PAL video. When all links were operating this equated to about 5 Gbits/second
through the Auburn control room alone. When you consider that Packet Radio
operates at 1200 bits/sec and Ethernet is about 10 Mbit/Sec you realize that
this is a lot of data.
In the rack, CODECís convereted from PAL Video to Fibre or
Fibre to PAL depending on the direction of the link. The CODECís also
transmitted full stereo audio, 4 switch inputs and an RS-232 link allowing
almost any possible signal to be moved around.
The fibres came from each of the receive sites (Uni of NSW,
and Equestrian), and sent data a feed back to each site, and also sent video
back to the International broadcast centre. In addition the IBC (International
Broadcast Centre) also sent back video to us from the Manly Ferry, and program
video. There were also backup links direct to the IBC from the Equestrian
Centre and UNSW in case there was a major breakdown at Auburn.
Having 2 or 3 different receive sites means that decisions
need to be made on which signal to use. Whilst it might be possible to automate
the switching, we just threw people at the job during the Olympics. This
allowed us to anticipate what would provide the best signal, providing a better
signal most of the time.
To do the switching, each switcher was given a Jaycar
project box with 4 or 6 buttons in two columns. Each column selected a
different camera, and each row selected a receive site. The switches were set
out so that they could be held by both hands, and use their thumbs to press the
switches ñ much like playing some video games ñ but in this case with only
about 1 million of our closest friends watching in Australia, and probably over
a billion throughout the world.
[Insert Kolm cartoon]
During some races such as race walking or Road Cycling where
the course was fairly small video switching became more of an exercise of
trying to keep interested since the pictures were always so good from the
receive sites. During the Marathon and Triathlon more effort was required
making the job more enjoyable [albeit a bit more stressful].
Quality Control and Video Coloring
One of the hidden jobs of the Auburn control room was to
maintain all the cameras so that they looked the same. This was not as easy as
it sounds. All cameras are different and react differently to different light
levels. To fix this problem, remote control of the iris and coloring of the
camera is possible thanks to a cute packet radio transmitter from Total RF.
This unit used a 500 MHz band radio signal running at two watts broadcast to
the motorcycles where the radio signal was decoded and fed into the camera.
We had one transmitter on Waverly Tower east of the Sydney
CBD connected by a modem, and another transmitter on the roof of Auburn. The
antenna at auburn was a simple 6 element Yagi cable tied to a paper so that we
could change direction easily. We would have liked to get an Omni, but at that
frequency the Yagi was easier to buy, and provided some gain if we needed it.
At times the video coloring dropped out, but this usually
did not last very long. It meant that all the camera operators needed to worry
about was zoom and focus ñ which was good since they were from Spain and did
not know English. We were not so fortunate with the Helicopter video cameras,
as the operators needed to manually adjusted the coloring ñ and once again they
did not know English leading to a new kind of Chinese whispers during Olympic
Whenever the colouring control did drop out it was my job to
jump a fence on the roof of the hospital and more the direction the antenna was
110V & Electrical Safety
The electrical environment was quite bizzare. With an american
company came all the 110V equipment, whilst the Australian gear all operated on
240V. In the Auburn control room were about six 2000 watt auto-transformers to
change the voltage. They were large, and came with metal handles for carrying.
However we did discover the problems with importing
equipment that was not Australian Made only after almost blowing up some leased
line modems. We discovered that the handles would actually become a shorted
turn when they touched and spikes were getting into the equipment. This
equipment should never have been allowed into Australia.
Working in dual 110/240V created some problems even for
people like me who should have known better ñ I got one computer out of itís
box, plugged it in, and BANG. Power supply set for 110V. Fine, only the PSU
destroyed, and we had a spare one in another computer where the shipping
company destroyed the motherboard.
My boss actually blew up a couple of laptops thanks to the
use of UPSí and differing grounds. Donít ask me how, but somehow they managed
to get the laptop to float above ground, and blew the laptop when they plugged
in to the serial port.
One of the more bizzare problems was when we were powering
up the control room one morning. One of our team went behind the racks to the
power points to plug in the voltage conversion transformers that we had
disconnected the previous night, in case of surge or an electrical storm.
The guy plugged the transformer in, and the equipment did
not work. He looked at the transformer, and decided that it had failed and was
about to replace it when I had a look. The first thing I did was plug the
transformer in elsewhere, and it worked. Then I put it back in the power point
and turned it on, and everything was fine. After that I proceeded to tell him
that these things work better when they are turned on. He did not even think to
turn the power point on, since they do not have switches in the USA.
That same morning the roof of the control room with all the
microwave dishes and personel experienced an unnerving freak of nature. As we
were waiting for a rehearsal in severe fog there was an almighty bang and flash
on the roof. Static electricity had built up and had suddenly discharged. This
caused the people on the roof quite some shock ñ and they quickly left the roof
for close to an hour when some of the fog had lifted. No damage was done to
which we were all grateful.
UNSW and Auburn Rooftops
With microwave signals, line of sight is a must. The problem
comes that a helicopter flying at 500 ft is obscured by some Sydney buildings
when they are flying near the City and the receiver site is at Auburn. But as
the ëchopper gets close to the UNSW the coverage at Auburn improves. However as
it gets very close, the beam width of the receiving antenna is so narrow that
it is effectively impossible to get a good signal. For these reasons we had two
main receive sites for microwave signals.
Each site had six microwave dishes on tripods, tracked by
hand. Three dishes were for motorcycles, one for the lead vehicle, and the
remaining two dishes were for two helicopter video cameras. Each dish had
tunable microwave receiver in the 1.9 GHz to 2.5 GHz band and an LCD monitor so
that they could see the picture they were receiving. When the helicopters were
on the other side of the city, they normally had an easy job, but had to
maintain concentration when the helicopter was close.
Each dish needed to be put away each night, and assembled
again each morning. But once the equipment was set up, the trackers would
normally have up to an hour before a rehearsal, and after the rehearsal often
an hour or two before the race started, meaning that they got to spend a lot of
time on the roof sunbathing. Some of our team got quite respectable suntans ñ
something that would have been impossible back home in the USA.
Once they obtained an Esky, and borrowed an old car radio
the conditions on the rooftops became quite civilized.
City Wide Repeater system.
For any large event, good communications is essentialÖ And
you could say that the Olympics is the Ultimate large event. With about 40
people working for the company, and more than that out in the field needing to
hear instructions from the director required a good repeater system was used.
Five channels were used in the 500 MHz band connected to a
voting repeater system. A Voting repeater system has a number of different
receiver location, and repeats the one with the best signal. In our case we set
up receiver sites in North Sydney on the Hyundi Building, in eastern Sydney on
the UNSW, and at Auburn near Sydney Olympic Park. Our transmitter was located
on Waverly Tower also in eastern Sydney. All the sites were connected by
telecom ë4 wireí circuits providing excellent repeater coverage from the 5 Watt
Icom HTís that we used.
This is not the full story though ñ During races we had the
repeater running full time. In addition the director and our control room had
direct access through the ë4 wireí circuit so that we didnít actually need a HT
to transmit, and we could operate full duplex, even if those with HTís couldnít.
The worst job during the whole Olympics must do to Casey
[#insert Call] who looked after the repeater at the Waverly telephone exchange.
Not only was Casey effectively alone much of the time, he did not even have a
TV set to watch the Olympics, nor was there anyone to relieve him for lunch.
Since he was baby-sitting the repeater he did not actually have much to do
either. About half way through the games things improved when he got his VK ham
license and was able to access the Waverly Amateur Radio Society repeater,
direct from the console.
The Manly Ferry
One of the technical successes of the whole Olympics must be
the live video link from the Manly to Circular Quay ferry. During the 8 Km
journey, the ferry travels an ëLí shaped path through some of the harshest
microwave territory in the country.
The things that make Sydney so beautiful are what causes all
the problems, The Sydney Harbor bridge causes reflections, the Sydney Opera
House stops a single receive site access to the entire path of the ferry with
the entire terrain just adding to the problems.
This all called for a unique engineering solution. French
company Sagem has recently released a COFDM transmitter which was put to good
use. COFDM, or Coded Orthogonal Frequency Division Multiplexing is a digital
transmission standard soon to be used in most of the civilised world. The COFDM
system digitizes the video into an MPEG data stream and then creates an analog
microwave signal using 2000 carriers and modems. Error Correction and Coding is
used to reduce the number of errors on the received signal.
The COFDM system was installed with an Omni-directional
microwave antenna on the top of the mast on the Manly Ferry. The antenna was
connected to a transmitter in the body of the ferry. Since the power in the
ferry was variable, and often disconnected we installed a UPS to protect the
sensitive computing equipment. At the front of the ferry a remote controllable
camera was mounted in a waterproof housing. The camera was controlled by a UHF
radio link to the microwave receive site, and then connected to the IBC
The receive site at Neilson Park was more interesting
because of the itís unconventional antennas. Being at the middle of the L meant
that the antennas needed be open well over 90 degrees, whilst maintaining reasonble
gain. The solution was to use three cresent dishes. The signal from each dish
was fed into an LNA, and then into a combiner. Finally the signal was fed into
the decoder, where the signal appears out about 1 Ω seconds later as a perfect
PAL video signal.
Over the path of the ferry, there are a few places where the
picture dropped out, but in most cases the picture was what could only be
described as perfect. The picture was even fantastic where the ferry was moored
at Circular Quay and the signal was bounced off the Harbor Bridge.
Watching the coverage of the Torch Relay just before the
opening where the Flame was on the Collaroy ferry showed just how good a job we
had done. Channel 7 had a camera on the ferry with a microwave uplink, although
they did not have a good signal path to their helicopter, so the signal kept
dropping out. Whenever this happened, the producer switched to our video camera
providing the perfect pictures.
[It is amusing to write that the ferry that beached itself
as this article was being written was the same one as used for the torch during
International Broadcast Centre
The International Broadcast Centre was where most of the
worlds electronic media congregated to bring all the signals in from all the
venues, process them and distribute them world wide. As you can imagine with so
many events the IBC was a huge undertaking. The main building of the IBC was
about 200m by 330m, and contained about 25 different roads ñ each named in some
uniqely australian way [Such as The Dogs Leg, Lamington Drive, Wedgie Way, The
Main Drag etc]
At the centre of the IBC was SoboTech, where all† the
signals came in for distribution throughout the IBC. SoboTech had about 400
video and audio signals coming in through fibre optic cables. Each signal was
then displayed on itís own monitor on a 400 TV video wall. Then the signals
were distributed to each of the broadcasters either as an Analog or a Digital
signal [About 50% of users used digital].
All the video signals inside the IBC were PAL, even for the
countries that used NTSC or SECAM. The only exception was Japanese TV who also
had some HDTV cameras.
The control room was basically a switching station
Security at the Olympics was a pain at times. They relied on
your honesty. One afternoon after work I went around Sydney Olympic Park with
one of our team members, into the area outside the security zone. He was
wearing a pocket knife on his belt which he needed for work. On returning
though security he had placed the pocket knife in his bag, and left the holder
on his belt.
After going through security he and having his bag searched
they asked if he had a pocket knife on is belt. He said no, and they let him thorugh.
It was in his bag.
Another time I was taking some equipment into the Stadium,
and was walking thorugh a checkpount. I had 2 backpacks, a 1.5m antenna, 2U
rack mount case and some other equipment. I was loaded. I took all this off,
and failed the metal detector test. They then put the wand over me and I
passed. They declined to search my bags at that point, and asked if I was
carrying any ëExplosives, firearms or knivesí. Since I had my butane soldering
iron with canister of butane I said yes, to which they asked ëdo you have any
chemical or nuclear weaponsí. The obvious answer was yes, and I went through.
Car searches were equally as thourough. They had glove box
Wednesdays where they would search every glovebox in cars coming into the Olympic
Park. Thursdays it was boots, Tuesdays it was engine compartments etc. The only
way to get out of a search was if the compartment was sealed with a sticker at
departure ñ but I saw vans whoís back door was sealed, but where there was easy
access from the front door.
Also cars were only searched if there was someone available
to search the car ñ no search if the searches were searching someone else.
I did not know this, but in NSW there is a registration
label that you can get if you are an overseas visitor and you want to drive
your own car around. The registration label states that it is registered as an
unregistered vehicle, sort of a contradiction. And to get it the vehicle must
be insured. Why an I telling you this. Well, we imported [by sea] a few
vehicles for the games, all from the west coast of the USA. The tallyÖ
††††††††††† 3 Outside Broadcast vans ñ About the side of
Ford Econovans. Each with a retractable 50 foot microwave antenna
††††††††††† 4 motorcycles.
††††††††††† 1 bike trailer for the four bikes.
I was going to say that they were all driving around Sydney
with Pensylvania number plates, but that is not quite right. Seems that someone
stole the plates of two of them during shipping so they went round without
plates at all. At one stage I tried to get a copy of the NSW Police Report
concerning stolen USA plates so that the Pensylvania DMV would replace them and
ship them to Sydney. Lets just say that we placed this in the Too Hard basket.
The three vans were all left hand drive of course. This made
it fun driving around Sydney for our people. One of our team, obviously an
armchair lawyer, decided that it was not legal to be driving around with no
steering wheel on the right hand side ñ so he bought one. Technically it was a
steering wheel cover, but it was the right shape at least ñ and attached it
with Gaffer Tape. Now they were legal.
APRS – Air Traffic Control
As mentioned before the reason I was actually there was so
that I could monitor the locations of motorcycles, lead vehicle and helicopters.
Actually I was also monitoring the position of the Auburn control room, but
luckily that was not moving.
I guess the most important part of the whole exercise was
the work I was doing with APRS. APRS, or the Automatic Position Reporting
System. APRS is a technology designed to collect and distribute information on
the position of objects. Normally these objects are cars or houses, but could
be anything from border crossings in Bosnia, to sheep on a New Zealand farm, or
in my case video cameras.
GPS (or Global Positioning System) receivers are used to
work out where the object is, although this is not always the case. The GPS
receivers are connected to Packet Radio TNCís and then connected to radios
creating a network of stations.
The TNCís were programmed to transmit position information
every 20 seconds. This was not ideal, but it was good enough. The problem was
that we had a lot of stations, and if we had chosen more than 20 second
transmissions, they would not have all fitted in.
One of the problems in this type of situation is that many
of the stations will want to transmit their data at the same time. This would
have been a real problem since our transmitters were all low power, and they
could not always hear each other. The solution was to schedule every
transmission on the network. This is actually built into the TNCís which helped
me considerably. To make sure that all the TNCís used the same timeslot, the
internal software synchronizes the transmission to the highly accurate time
from the GPS receiver.
Placing the transmissions into timeslots took some planning,
but it was worth it.
As I mentioned before, all the transmitters were operating
on low power ñ only 5 watts on the 500 MHz band. There was no way a 5 watt
transmitter on the wrong side of the Sydney CBD was ever going to get data back
to the Auburn receive site. The two options were to use a second receive site,
like with the repeater system, or use a repeater.
We chose to use a digipeater since it could operate in
half-duplex mode. Normally digipeaters are fixed objects on a building
somewhere. In our case setting the equipment up on a high building would have
required too much work, and would not guarantee the performance of the whole
So we turned one of the helicopters into a digipeater. This
had the effect of putting the digipeater on a 1000-1400 foot mast, easily
higher than all the buildings in Sydney. The five ground units were then
programmed to use the helicopter when it was available [In-Air refueling of
Helicopters was not considered viable].
For obvious reasons 5 other helicopters and the light plane
did not need to use the repeater as they could be heard direct, just as the
repeater could be used direct. What made this solution even better was that we
had APRS equipment in all the helicopters so all I needed to do was to
reprogram the TNC to turn it into a digipeater. [The reason we did not use the
light plane for a digipeater was that it was there for emergency use during
important races, meaning that there were times it was not available. It was
also the only aircraft that had laptops meaning that we could reprogram it to†
become a digipeater in case of emergency anyway]
There were times that the relay helicopter was not available
because of refueling. At these times we generally lost the position of the
motorcycles, although we would occasionally get the position updates through
I am told that the light plane used APRS in an rather
interesting way. As soon as they got to altitude of about 14,000 feet the pilot
would cover all the windows with cardboard and fly by instruments. He would use
a laptop in the plane to tell him where to fly, and where he was. Of course he
used his other instruments too, but APRS told him where he was needed.
The pilot had a hard job, because he needed to stay within
about 2Km of all the motorcycles, and at the same time circle doing flat turns,
which I am told is a rather nasty way to fly requiring a lot of concentration.
But the light plane was essential to the success of the
games because for the start of the womenís marathon it was providing the only
pictures from the ground because the helicopters could not take off because of