Single Frequency Networks (SFN)

[Theres Always Radio]

Digital Television and the SFN

SFN

A Single Frequency Network as applied to Digital Television Broadcasts in Australia is a technique that allows multiple transmitters with overlapping coverage to operate on the same channel. Steps are taken to minimise the risk of interference between transmitters in these overlapping areas by careful control of output power and timing. It is not a perfect solution and particular care needs to be taken by antenna installers to achieve a reliable service in these overlapping areas.

Note on DVB-T

In Australia, all DVB-T broadcast channels occupy 6.656250MHz of their 7 MHz bandwidth with a Coded Orthogonal Frequency Division Multiplex (COFDM) signal comprising 6817 carriers spaced at 976.563Hz apart. Most of these 6817 carriers are quadrature amplitude modulated (QAM) and every carrier is capable of carrying different information. A few carriers are fixed to a known amplitude and phase for the purposes of synchronisation and referencing and the majority carry the sound ,vision and text etc i.e. the payload is 6048 carriers. Generating and modulating each of these 6817 carriers individually was unthinkable until recent times. Now a process called Inverse Fast Fourier Transformation (IFFT) allows all 6817 carriers to be generated at once in a manner that spaces them specifically to minimise interference between each carrier (the coded orthogonal part of the process).

The 7MHz system as used here with error correction capabilities versus desired payload allows for data rates around 19 Mbit/s (Seven and SBS) and 23 Mbit/s (ABC, Nine and Ten).

Unfortunately we cannot insert pictures on this forum so I’ll try to describe a constellation pattern and how we can use it to observe if the QAM signal is working well or not and how received noise affects this display. The version of QAM used here is 64QAM, this means for the 6048 payload carriers there is 8 discrete phase changes of a carrier and 8 discrete levels of amplitude change, from this we have 64 combinations of amplitude and phase. If you imagine an 8 by 8 matrix of squares on a page and place a small dot in the centre of each of the 64 squares this would represent an excellent received signal. Keep in mind that a small dot here means the target is being hit in the same spot time after time after time. The position of the dots away from the centre of each target gives the operator of the transmitter information on the health of the transmitter. For those at the other end receiving the signal, the dots should be always found to be in the centre of the target. Now if these small dots represent a target being hit at the same point continuously and noise is introduced, the dots being fired at the target are shifted and what was a defined small spot grows larger with increasing noise and its edges become fuzzy. Eventually the whole 8 by 8 matrix becomes a blur, but the system keeps going. All is still well so long as it still lands somewhere in its own target. Should it hit an adjacent target or worse, an error is registered and then the bit error ratio (before viterbi) starts to worsen. If things continue to deteriorate the programme is lost very quickly.

If you have noticed that you get sound blips or picture freezing occasionally when the fridge or dishwasher etc starts, then you probably have a marginal signal in your area, a poor installation, you’re in a mush zone or a bit of all. However I only seek to mention it here because it also indicates a noteworthy characteristic of DVB-T and that is as the received signal becomes weaker, reception becomes more susceptible to impulse noise. This is because the data in the payload carriers is amplitude and phase modulated (QAM) and any amplitude modulated signal can be affected. Anyone who has listened to AM radio has heard interference from car ignition systems, kitchen appliances and lightning etc. These common place interferers that cause noise on AM radio, may cause horizontal interference lines on analogue television and corrupted data in DVB-T causing intermittent freezing and audio blips or complete loss because the targets in the constellation as describe above have been missed.

Simultaneous amplitude and phase modulation has been around and in use by us all for a long time. Some recent examples are Stereo AM radio and colour television. When AM radio became AM stereo, phase modulation had been used to allow left-right information to be transmitted and in 1975 when colour television started the colour sub-carrier was added and its amplitude and phase determined the colour and its intensity.

Despite its vulnerability to impulse noise, the extremely complex error correction techniques manage in the majority of cases to cope with such impulse noise issues. It will not fix everything though, the receiver can’t recover the data if the signal is too low compared with the impulse noise.

One of the nice things about digital television is that there is no ghosting of the picture, in fact ghosting isn’t possible. The receiver is able to disregard, up to a point, reflected signals from other sources such as buildings, water towers and hills. Built into the transmitted signal is a guard interval, 1/8 for (Seven and SBS) and 1/16 (ABC, Nine and Ten). This allows for reception of the main signal and echoes successfully so long as the echo delay does not exceed 128uS for (Seven and SBS) and 64uS (ABC, Nine and Ten). This level of tolerance to reflected signals or echoes allows the digital signal to be successfully received whereas the analogue could be rendered unwatchable. Furthermore it is this tolerance to echoes that allows the possibility of a SFN.

The SFN

As any good television antenna installer knows, the choice of receive antenna is only a portion of making a reliable service for the consumer. Cable lengths, the location of the antenna and its height on the home is all-important and makes a substantial difference to the result. The reason for taking such care is the surrounding topography and buildings can play havoc with predicted received signal levels. The very same topography issues are considered when decisions are made for locations of television transmitters. Once the transmitter location is decided upon, it will be a compromise, designed to get the best coverage to most of the people, however some will miss out. In the past with analogue television, smaller translators (transmitters that receive the main transmitted signal and retransmit it on another channel) were used to help cover those who missed out. You can do the same with digital television however there is not enough spare channels available while analogue is still to air. The solution was to have them all on the same frequency. As mentioned above, digital television receivers can cope well with echoes, so if you can time a second and third transmitter’s output so it appears as an echo within the guard band of that channel at the receiver, it will be treated as an echo and stable pictures and sound are produced. To support this, all transmitters in the SFN are fed with the same programme at around the same time, usually delivered by microwave link or optical fibre. There is timing information contained within the data stream and while the main transmitter is not delayed, the others are delayed to ensure their signals fall within the guard band of the main transmitter in its area influence. All are referenced to the GPS system to ensure each transmitter is on frequency and data rate is correct.

Where does it go wrong (Mush Zones)

Whether the topography is flat or mountainous there are areas where it all goes wrong. There will be areas where more than one transmitter within the SFN is received and their timing will be outside the guard band. By design these are preferred to be placed over water or in unpopulated areas as the topography permits. There will also be areas where they do fit within the guard band but are of similar levels and the receiver struggles to cope with deciding which is the main signal and which is the echo. These can be very difficult to sort. If you are installing an antenna in these areas, suddenly, rather than just knowing the gain and front to back ratio of a receive antenna, you also need to know the shape of its pattern and if it has any considerable acceptance on side lobes. The reason here is this antenna needs to be able to discriminate between the two or three signals ahead of the receiver thereby providing the receiver a signal it can cope with. In such difficult locations, knowing from what directions signals need to be avoided becomes as important as knowing the direction of the one you need.

I hope this goes some way in helping to understand an DVB-T system operating as an SFN.

[alanh]

All,

James thanks for your previous post. It should be read in conjunction with the following links;

The how and why of COFDM

Explaining some of the magic of COFDM

Development of a digital terrestrial front end

Zarlink_MT352 specification sheet Zarlink_SL2610 Tuner Chip

If you have access to the computer program MathLab, then COFDM simulator

I agree there is no limiters in the bottom one. I am glad that this has provoked more discussion on how the RF side of DVB-T works and it is also illuminating that J Scott has been instrumental in the development of the DVB-T, DAB and DRM systems.

Thanks

AlanH

[Theres Always Radio]

Alanh

It's good to see you recently spent some time on google and brushed up on COFDM, DVB-T and DVB-T receivers.

Thanks for the suggestion that it should be read in conjunction with.........etc except I disagree. The text was intended as presented and that was to break down some of the complexity of the topic and allow those who are not involved in the industry some chance to grasp some of the concepts.

I believe this whole area of understanding MPEG, COFDM, DVB-T and SFN is a difficult one, I personally still struggle with some of it. For those who have no difficulty, absolutely, there are some good references on the net to follow up on.

I believe the nice gentleman you mentioned from the BBC would prefer to be referred to with his name spelt correctly "Mr Stott".

There was no copy and paste in my description above, its experience.

Finally, can we please leave it there and allow those who may have unanswered questions to remain interested enough to ask them?

[alanh]

James,

The reason I suggested that the posts be read in conjuction with your posts is the local application of the values used. Thanks for your experience.

I still have another point of discussion which I will have to research first, so keep looking. It is on the modulating side.

AlanH

[matthew burford]

All,

Single Frequency Networks (SFN) is where the same TV channel is used for two or more transmitters within range of each other. This system is used to conserve channel allocations and can only be used by Digital TV and Radio.

Normally the interference would be intolerable except close to one of the transmitters.

DVB-T is used for Digital TV in all areas except North America. It uses 1705 carriers (2k). The North American system called ATSC uses one carrier, this is like serial transmission.  In our system we can slow the data rate by 1/1705th for each carrier, thus is using parallel data transmission. The receiver will detect the start of the pulse on that carrier and ignore any following pulses until the original pulse is finished. The following pulse may be from a reflected signal or from another transmitter further away. This makes portable and mobile reception more achievable than the ATSC system.

When SFNs are used, the number of carriers is increased to 6817 carriers (8k mode) This increases the pulse length by 4 times, giving a greater tolerance to longer delayed signals.

The gap between pulses (Guard Interval) is selectable by the broadcaster, so can be increased with a drop of data rate. Options are 1/4, 1/8, 1/16, 1/32.

To make the system work two things need to be done

1. All the transmitters on the same channel must be at exactly the same frequency. This is done by using a satellite frequency as a common reference through GPS.

2. Delay the signals so that where the signal strength is equal the signals arrive a the same time from both transmitters. This is done by electronically delaying the main transmitter so that the time it takes for the signal to go from the main transmitter to the repeater and back to the receiver is identical.

If the signals are not arriving within the time window allowed, the only option available to the viewer is to use a very directional antenna to exclude the signal from one source. To make this easier some repeaters are on the opposite polarisation to the main station. For example Mt Tamourine is on Horizontal and Currumbin is on Vertical.

AlanH

<{POST_SNAPBACK}>

[Theres Always Radio]

Hello Matthew Burford

I am interested in why you have placed the previous post that simply quotes Alanh?

The reason I take interest was that there was so much incorrect information posted here by on this topic by Alanh that it warranted correction.

You're placing of the quote without comment suggests you have either made an error as a first time contributor on the forum or you take issue with information offered by myself.

[johntech]

Someone is populating the dba forums with misinformation which is neither warranted or welcome........I wonder if they know who they are?.

I don't pretend to be a master of Digital Modulation Technology but have a deep understanding of the practicalities of these technolgies including the DVB-S variant.

The 8k carriers are NOT differing amplitudes by design but if you are in an ordinary location this indeed may occur which is why there is 6odd thousand of them so remaining functional carriers can compensate.

Why didn't you title this thread "SFN : how it doesn't work" because if you're one of the poor saps that resides in a location where the timing of signals is outside the Guard Band then you too may have steam coming from your ears.

I have personally witnessed Broadcasters varying their SFN timing to great effect but I also know of SFN issues that seem to be unresolvable no matter what your location because of this issue.

Yeah...keep researching, then think for a while, then research again (and maybe again) BEFORE posting.

James T Kirk

You're request to "leave it there" is a noble one because whilst I have learnt something from your posts this thread is going nowhere fast with most people not being able to sort the wheat from the chaff I'm sure.....not entirely your fault.

[alanh]

JohnTech,

You stated " don't pretend to be a master of Digital Modulation Technology".

DVB-S does not use QAM, but DVB-T certainly does. If you check

DVB-T Modulation in particular the quote the phase and amplitude of the data responding to the within a bits in a byte.

The chrominance signal in analog TV uses Quadrature Amplitude Modulation on a single carrier frequency. So why can't all the carriers in a COFDM signal do similar?

AlanH

[damntv]

I have been reading this thread with great interest and I have noted a few alarming things. johntech there are 6817 carriers transmitted and hopefully the same amount recieved. alanh I do not understand how in your explaination of how a SFN works that if the main transmitter transmits its signal to a "repeater" then is delayed so it transmits its signal at the same time the repeater transmits how this works. To my understanding in an SFN the main transmitter (if a translator approach is taken) would transmit on a different frequency eg. VHF then the smaller translators would receive this signal then with the magic of GPS and using the 1PPS provided would all transmit at the appropriate time on UHF for example. This variation in timing is used as I beleive to attempt to locate a "mush zone"in an area of least population or where it is of least consequence.

Where the main transmitter is on the same frequency as all of the others within the SFN then they are all required to be feed via a source whether it be satellite, microwave link etc and then an appropriate delay is required for the setting of mush zones and stuff.

Just my thoughts

[alanh]

damntv,

A SFN is where the repeater is exactly the same frequency as the main transmitter and any other repeaters within the one coverage area. Where translators are used the output frequency of the translator is different so there is no requirement to be in time. This is because the receiver will separate the two signals with a tuned circuit tuned to your selected channel.

A typical SFN is in SE Queensland, where SBS 36 is the main transmitter in Brisbane. Channel 36 and program is used on the Gold Coast, Currumbin and the Sunshine coast.

When a SFN is used you cannot send the signal via satellite between the main and repeaters. If you do the delay is not constant which will cause problems. The signal will be sent via microwave or fiber optics. The GPS is used to keep the carrier frequencies the same.

Read DVB-T Transmission Processes You will notice all the shuffling of data. This is so that if some of the 6817 carriers are nulled out, error correction can replace the data.

Satellite distribution is used for a large number of transmitters. It is not an issue because where the same terrestrial channel is used it is out of range of any others. It this were to be the case a translator is used so that another channel is used. This is always the case for analog, but in Digitial Video Broadcasting SFNs can be used. SFNs are selected by the Australian Communications and Media Authority where there is insufficent available channels.

AlanH

[Theres Always Radio]

Sorry to those trying to follow this. Since I wrote what I consider to be a simplified explanation on Aug 20, there have been some more misinformation added by others, I'll go through them one at a time and leave it there....again, I hope.

Johntech:

The output at the transmitter sends 6817 carriers, no more, no less, I've counted them! (It is possible for less than this number to arrive at the receiver. It is generally a function of topography and the quality of the home installation)

The carriers do vary in amplitude, part of the information is contained here. Other evidence for this, for those who need to check, is the need to measure peak to average ratios of the transmitter.

AlanH:

You really should stop this, on this subject you are a menace.

The chrominance signal in analog TV does not use QAM for heaven's sake!! The phase is not shift keyed, it is continuously variable, big difference.

damntv wrote "Where the main transmitter is on the same frequency as all of the others within the SFN then they are all required to be feed via a source whether it be satellite, microwave link etc". Alanh, you replied that he was wrong, but guess what, he wasn't, all of the SBS-36 South East Qld SFN are satellite fed.

[alanh]

Sorry to those trying to follow this. Since I wrote what I consider to be a simplified explanation on Aug 20, there have been some more misinformation added by others, I'll go through them one at a time and leave it there....again, I hope.

Johntech:

The output at the transmitter sends 6817 carriers, no more, no less, I've counted them! (It is possible for less than this number to arrive at the receiver. It is generally a function of topography and the quality of the home installation)

The carriers do vary in amplitude, part of the information is contained here. Other evidence for this, for those who need to check, is the need to measure peak to average ratios of the transmitter.

AlanH:

You really should stop this, on this subject you are a menace.

The chrominance signal in analog TV does not use QAM for heaven's sake!! The phase is not shift keyed, it is continuously variable, big difference.

damntv wrote "Where the main transmitter is on the same frequency as all of the others within the SFN then they are all required to be feed via a source whether it be satellite, microwave link etc". Alanh, you replied that he was wrong, but guess what, he wasn't, all of the SBS-36 South East Qld SFN are satellite fed.

<{POST_SNAPBACK}>

[alanh]

James,

When you use a vectorscope which is demodulating a PAL or NTSC signal it displays the phase as the hue and the amplitude as a saturation. The 4.43 MHz carrier is fed into a pair of modulators with the V signal delayed by 90 degrees (Quadrature). The double sideband suppressed carriers are then added together to make a chroma signal. Why is this no Quadrature Amplitude Modulation - QAM?

How does phase shift keying work, by switching phase angles and amplitudes, to give 64 combinations of phase and amplitude. This is 64 discrete levels which is 32 levels on the inphase axis and 32 levels on the quadature axis. On operation I can't see much difference between this and colour bars in NTSC. The only difference is the steps are not equal sizes and there is only 6 dots and zero.

SBS feeds all its programs except Sydney via satellite. I was suggesting only within a coverage area. Coverage areas are specified by the ACMA and surround the highest power transmitter in the area. The whole of Australia is not a coverage area for SBS, it consists of many coverage areas.

The movement of satellites is why the delay varies for the signal. Since you have to delay the Brisbane signal and not to those further away then you are not going to use an expensive satellite circuit for an individual repeater.

I have read all the engineering articles for the tests of SFNs around Canberra when it was decided to use in in Australia. They made a point about the delays must be constant and the frequencies exactly the same.

The Magic of COFDM Shows what happens when two signals of the same frequency interfere when they are the same level.

AlanH

[Theres Always Radio]

Alanh:

Ok let's see how far we have progressed to date, you are a most frustrating person.

So far we have managed get your mental block past the following:

the correct number of carriers, no limiting, that translators are not part of SFN's, there is information in the amplitude of a cofdm signal, that DVB-T has nothing to do with FM etc etc. None of which gets openly acknowledged by you.

The latest:

I am shamed. In my zeal to compare the rigidity of the the 64 constellation targets versus the infinite possibilities offered by the colour difference signals in PAL television, I messed up, you are perfectly correct here it is QAM. I apologise to the forum for my error.

The six dots you mention represent the well defined yellow, green, cyan, blue, magenta & red of colour bars. The areas outside these target are not errors just different colours. Obviously if colour bars are used and these 6 targets are out of limits there is distortion.

Why wriggle and try to steer your statement on satellite delivery when it was clearly in response to damntv and you were just wrong. Why waffle on? You post so much misinformation and even when you finally catch on you never openly acknowledge your error. The top priority here is to pass on accurate information.

I think you've lost the plot, Alanh and James T Kirk are unimportant trivia, the only value is in the content and it only has value if correct.

[alanh]

James T Kirk quotes

"You really should stop this, on this subject you are a menace.

"The chrominance signal in analog TV does not use QAM for heaven's sake!! The phase is not shift keyed, it is continuously variable, big difference."

"Simultaneous amplitude and phase modulation has been around and in use by us all for a long time. Some recent examples are Stereo AM radio and colour television. When AM radio became AM stereo, phase modulation had been used to allow left-right information to be transmitted and in 1975 when colour television started the colour sub-carrier was added and its amplitude and phase determined the colour and its intensity."

"So far we have managed get your mental block past the following"

"I messed up, you are perfectly correct here it is QAM. I apologise to the forum for my error."

"I believe this whole area of understanding MPEG, COFDM, DVB-T and SFN is a difficult one, I personally still struggle with some of it."

"You post so much misinformation and even when you finally catch on you never openly acknowledge your error. The top priority here is to pass on accurate information."

Now I am confused!

AlanH

[johntech]

TO ALL

I FULLY REALISE THERE ARE 6817 DIGITAL CARRIERS IN 8K COFDM MODULATION TECHNOLOGY 'CAUSE I LOOK AT THEM EVERYDAY.

I WAS JUST BEING FLIPPANT BECAUSE EVERYBODY IS BEING SO DAMN PEDANTIC WHICH IF IT'S IN THE ENDEAVOUR TO BE ACCURATE IS BENEFICAL........BUT IN HERE ANYTHING MORE IS PROBABLY A WASTE OF BREATH.....err....TYPE.

FOR THE PURPOSE OF DISCUSSION 6 THOUSAND ODD IS FINE ISN'T IT?.....WHO GIVES A RAT'S ASS ABOUT THE EXACT NUMBER, THE POINT IS BEING LOST!!!!!!!.

WHY HASN'T ANYBODY ELABORATED FURTHER ON THE REAL ISSUES I RAISED????.....INSTEAD PICKING OUT IRRELEVANCE!!!!!.....HOW TRIVIAL.

THIS IS CUTTING EDGE STUFF AND EVERYBODY SEEMS TO KNOW A LITTLE BIT WHICH IS NOT CONDUCIVE TO GETTING THE FULL PICTURE (pun intended) IS IT?.

I TELL YOU WHAT....

YOU PEOPLE KEEP RANTING,RAVING AND POSTULATING (FOR ALL IT'S WORTH) AND I'LL GO OUT AND CONTINUE TO STOP DISCERNING VIEWERS DIGITAL RECEPTION FROM PIXELLATING IN THE REAL WORLD.

I DID PEEP IN HERE BECAUSE IT IS NOT ALWAYS POSSIBLE TO DO THIS BECAUSE OF REAL LIFE DIGITAL BROADCAST ISSUES.

JUST LIKE WHEN CD WAS LAUNCHED........IT WAS TOUTED AS PERFECT.

[Theres Always Radio]

Alanh:

We're getting somewhere at last, you've admitted you're confused.

Johntech:

Initially all I can suggest is take a valium or mogodon. Once settled, if you care to, raise your issue but I suspect you have little time for people like me.

I looked through what you wrote earlier and this was your question. "Why didn't you title this thread "SFN : how it doesn't work" because if you're one of the poor saps that resides in a location where the timing of signals is outside the Guard Band then you too may have steam coming from your ears." The answer is of course for the huge majority it works well, that's why. Others get to resolve guard band issues, what are you doing wrong!

I note you spend most of your time solving real world DTV issues while the rest of us do nothing, we owe you a great debt.

[johntech]

WHAT AM I DOING WRONG???......

I beg thy pardon

I know of a good number of Digital TV Installers who have REAL SFN issues in my region to the tune of none of them can get reception of particular digital channels that share with other transmitters in a very wide SFN network despite great sucess with analog TV reception at the same locations.

Additionally when I visit other installers jobs and even locations that are "converting to digital" with their existing aerials...NONE...NOT ONE OF THEM....can recieve/lock these particular channels despite adequate signal levels.

Let me repeat that 'cause I know you only half read my posts....

NO ONE, NOBODY, NOT ONE location that can recieve these particular digital channels (that are dearly loved) ANYWHERE, NOT IN ANY SUBURB, AT ALL, EVER.......clear?.

[Theres Always Radio]

Johntech:

I only have to read half your posts because you write everything twice!!

What are you doing wrong? Well I don't know!! That's why I asked!! Tell us what you do. Describe the circumstances. Give us something to work with.

It would also be clearer to all if you'd named some of the suburbs and channels you have trouble with.

By the way if the specific area is not in SE Qld, I'll have no local knowledge and I wouldn't attempt to advise with specifics, that's if you're interested, however by specifying the location, others who are close by who may have had successes may pipe up.

[bellotv]

James T Kirk

You mentioned that mush zones were arranged so as to occur in mostly unpopulated areas.

I,m curious to know if the radiation pattern from transmitter sights is stearable or are they omnidirectional ?

I am a TV tech and have know Idea what goes on at transmitter sites but would like to know a bit more.

You obviously seem qualified to comment or point in right direction

Regards Bellotv

[Theres Always Radio]

Hi Bellotv

The varaitions in the transmit antenna pattern are endless as is the power driven into them. The simplest is a single colinear that's omnidirectional or a single panel that's directional. If I'm guessing correctly you've seen the four sided panel arrays at Moombil and Nardi. These can be set up in all kinds of configurations. Generally these types of antennas have two stacks which essentially means two inpendant antennas with separate feeders back to the transmitter hall, one antenna is above the other much like you would use to pick up 3dB gain if you were in a difficult receive location. However the transmit antenna has much greater versatility.

The panels on each face can be fed with equal power giving an omnidirectional pattern (A four leaf clover with the nulls between petals filled in) or they can be fed unequally to derive endless possibilities for the pattern. This is how the horizontal pattern is derived. Now the signal is generally radiated from 250 to 1000 or so metres above sea level and we don't want the signal to pass over everyone's heads out to sea and so tilt is applied to the beam. This can vary downwards from a fraction of a degree to 3 or 4 degrees. Over the years this has been achieved by changing the phase of the signal being delivered to one stack relative to the other. (Please note that tilt can be built into a single stack antenna)

However, once the antenna is installed and adjusted correctly and commissioned that's it, a completed antenna has no "steerable" capabilities thereafter. If from time to time you notice a reduction in level compared to what you are used to seeing, it could be as you'd imagine , a reduction in transmitter power due to a fault or maintenance or possibly running the transmitters into one stack where the best you'd find from them is 3dB down.

[bellotv]

Thanks James T Kirk

So can I assume that with a 4 sided dual stack (like MT Moombil as your've correctly assumed) that there are 8 coax feeds ?

If so ,in the transmitter hall does each have its own individual PA driving it or is there one main PA and some kind of passive coupling devise that shares the power to each as required ?

Your help is appreciated.

[Theres Always Radio]

Bellotv

There would be only one feeder per stack, so two feeders connect to the transmitter hall. Each stack, say in the case of UHF, may have four faces and be up to eight panels high. If it is eight high, these 32 panels are fed from one feeder via a combination of power dividers. The lengths of the cables from the power dividers to the panels (called tails) are precision made to designed lengths for correct phasing. So essentially you have the same number of feeders as stacks.

In the transmitter hall, all individual UHF transmitters are combined together and ultimately appear combined into a single feeder, this can be 150mm diameter! This then goes to the switching frame where it is split into two separate signals each 3dB down on the original. It's these two sources which now drive the upper and lower stacks of the antenna.

In high power analogue television there is a vision transmitter and a sound transmitter per service, each transmitter can be composed of many PA's. (Years ago in valve days these transmitters were actually separate cabinets but nowadays they are all in the one cabinet.) These separate sound and vision transmitter outputs are combined and appear as a single output which feeds the channel combiner as described above.

High power DVB-T transmitters again uses multiple PA's but it is a single COFDM signal, no need for separation of sound and vision for modulation and amplification. But ultimately it's the same in that the multiple PA's are combined into one output and delivered to the combiner.

[bellotv]

One more question James T Kirk

Hi Bellotv

.......

The panels on each face can be fed with equal power giving an omnidirectional pattern (A four leaf clover with the nulls between petals filled in) or they can be fed unequally to derive endless possibilities for the pattern. This is how the horizontal pattern is derived.......

<{POST_SNAPBACK}>

So far I,m with you but I wonder how you adjust/equalize the power to each face ?

I am used to splitters and drop taps where the outputs are fixed by design.

Are the power dividers you use also fixed or are they somehow variable to achieve the desired pattern ?

Sorry to labour the point but I,ve always been curious about this

[johntech]

You don't adjust the power as efficiency is really important here.

Multiple panels allow the transmitting ERP (effective radiated power) to be increased through the property of antenna gain (just like on the recieve end).

The connecting cables to each panel are adjusted in length to change the relative phase of the array to enable beam tilt and steering into the desired area.