BBC RD 1996/5

Research andDevelopment

Report

EXPERIMENTAL SATELLITE BROADCASTOF EUREKA 147 DAB FROM SOLIDARIDAD 2

J.T. Zubrzycki , B.Sc.(Eng.), Ph.D., C.Eng., M.I.E.E.,R.H. Evans , B.Sc.(Eng.),

P. Shelswell , M.A., C.Eng., M.I.E.E.,M.A. Gama (Telecomunicaciones de Mexico) , and

M. Gutierrez (Instituto Mexicano de Comunicaciones)

Research and Development DepartmentPolicy and Planning Directorate

THE BRITISH BROADCASTING CORPORATION

EXPERIMENTAL SATELLITE BROADCAST OFEUREKA 147 DAB FROM SOLIDARIDAD 2

J.T. Zubrzycki , B.Sc.(Eng.), Ph.D., C.Eng., M.I.E.E., R.H. Evans , B.Sc.(Eng.),P. Shelswell , M.A., C.Eng., M.I.E.E., M.A. Gama (Telecomunicaciones de Mexico), and

M. Gutierrez (Instituto Mexicano de Comunicaciones)

Summary

BBC Research and Development Department, in collaboration with Telecomunicaciones de Mexicoand the Instituto Mexicano de Comunicaciones, has carried out highly successful test transmissionsof the Eureka 147 Digital Audio Broadcasting (DAB) system in Mexico from the Solidaridad 2satellite from 17 to 21 July 1995. This included tests of mobile reception conducted in the suburbs ofMexico City. Even though the DAB signal is very different from the usual signals handled by thesatellite, it was transmitted without distortion, verifying the feasibility of satellite DAB fornationwide reception of CD quality radio by car listeners. The Solidaridad 2 was designed fortelecommunications, based on Inmarsat C technology, operating at L-band to and from mobiles andat Ku-band for the fixed earth-to-space/space-to-earth links. The L-band frequencies are just abovethose designated for DAB by ITU-R and so the availability of this satellite gave an excellentopportunity to test the Eureka 147 DAB system.

Research & Development DepartmentPolicy & Planning DirectorateBRITISH BROADCASTING CORPORATION

Issued under the Authority of

General ManagerResearch & Development Department

BBC RD 1996/5

(R029) 1996

British Broadcasting Corporation

No part of this publication may be reproduced, stored in aretrieval system, or transmitted in any form or by anymeans, electronic, mechanical, photocopying, recording,or otherwise, without prior permission.

(R029)

EXPERIMENTAL SATELLITE BROADCAST OFEUREKA 147 DAB FROM SOLIDARIDAD 2

J.T. Zubrzycki , B.Sc.(Eng.), Ph.D., C.Eng., M.I.E.E., R.H. Evans , B.Sc.(Eng.),P. Shelswell , M.A., C.Eng., M.I.E.E., M.A. Gama (Telecomunicaciones de Mexico), and

M. Gutierrez (Instituto Mexicano de Comunicaciones)

1. INTRODUCTION ................................................................................................................. 1

2. BACKGROUND DISCUSSION ......................................................................................... 1

2.1 Listener’s interests ........................................................................................................ 1

2.2 Options ..................................................................................................................... ...... 1

2.3 The Eureka 147 system ................................................................................................. 2

2.4 Other systems ............................................................................................................... 2

3. SOLIDARIDAD SATELLITES ........................................................................................... 2

3.1 Introduction to the Solidaridad satellites .................................................................... 2

3.2 General description ....................................................................................................... 2

3.3 Forward transponder .................................................................................................... 3

4. TEST SYSTEM .................................................................................................................... 4

5. EXPERIMENTAL RESULTS ............................................................................................. 6

5.1 Back-off tests ............................................................................................................... .. 6

5.2 Power budget ................................................................................................................. 7

5.3 Mobile reception ............................................................................................................ 8

6. DISCUSSION ....................................................................................................................... 9

7.7. CONCLUSIONS .................................................................................................................. 9

8. ACKNOWLEDGEMENTS .................................................................................................. 9

9. REFERENCES ..................................................................................................................... 9

(R029)

EXPERIMENTAL SATELLITE BROADCAST OFEUREKA 147 DAB FROM SOLIDARIDAD 2

J.T. Zubrzycki , B.Sc.(Eng.), Ph.D., C.Eng., M.I.E.E.,R.H. Evans , B.Sc.(Eng.),

P. Shelswell , M.A., C.Eng., M.I.E.E.,M.A. Gama (Telecomunicaciones de Mexico) , and

M. Gutierrez (Instituto Mexicano de Comunicaciones)

1. INTRODUCTION

Satellite digital audio broadcasting (DAB) is anattractive option for broadcasters in many parts of theworld. Over the past year a start has been made intransmitting DAB terrestrial services, with the Londonarea of the United Kingdom now having five SingleFrequency Network (SFN)1 transmitters to serve it.The UK will have a 60% coverage in less than 2 years(1998). Countries that are very large, or those that aresparsely populated, may find it uneconomic to providea complete terrestrial service. In these instances, thereis a strong argument for using satellite services to ob-tain a large-area coverage. Also, there is a particularlygood case for providing international services via sat-ellites when access to terrestrial transmitters is notpossible or feasable. Satellite broadcasting would thenimplement the European Union policy of ‘broadcast-ing without frontiers’. It would also make multilingualbroadcasting a particularly effective option, as DABmultiplexes are readily reconfigurable.

Methods for delivering radio programmes from satelliteshave been under discussion for many years. In theearly days, there were studies of AM or FM transmissionsand of using short wave frequencies,1,2 but these werefound to be impractical. There is now a general accep-tance that digital techniques are best suited to thepurpose, and that an L-band frequency is technicallyoptimum.2,3

Part of the difficulty for introducing satellite soundbroadcasting economically is that the link budget formobile and portable reception is rather tight. Tomake satellite sound broadcasting a success, it isnecessary to ensure that full use is made of all theimprovements in technology; without them, the sys-tem would be marginal and, consequently, lesslikely to succeed.

The question of mobile use does not arise whenconsidering satellite television broadcasting. Usually,dish antennas between 60 and 90 cm in diameter areused, fixed to the exterior of the house. But for soundbroadcasting, mobile use becomes a major considera-tion; occupants of cars, coaches and lorries areperhaps bigger users of radio at some times of the day

than are householders! Small antennas (not needingcritical alignment) are the main requirement for themobile radio listener. It is the difference between theuse of antennas for the non-mobile television vieweron the one hand, and the regularly-mobile radio lis-tener on the other, that makes the difference in the linkbudget for each system.

Even though satellite sound broadcasting offers only asmall margin in the link budget, there are possibilitieswhich may be exploited for offering practical systemsthat would provide a useful service.

2. BACKGROUND DISCUSSION

2.1 Listener’s interests

Not all listeners necessarily demand a high quality ofsound reproduction; ruggedness, access and reliabilityare often more important to them. The listener wantsprogrammesto listen to, the technology provides ameansto that end!

There are already several digital sound broadcasting sys-tems available on satellites, including the ITU-Rrecommendation BO. 789.4 These have not been a re-sounding success. Though there is some interest, themajority of the public have not taken up the new serv-ices with any enthusiasm. The new digital soundservices are based on the 12 GHz television systemsand so require a fairly large dish antenna. Most radiousers require to move their radios around, for at leastsome of their day; restricted positioning of the set isnot acceptable. Also, satellite radio receivers are moreexpensive; the majority of people do not wish to spendas much on radio receivers as they are prepared to payfor television sets – radios must, therefore, be reasonablycheap; as a corollary, there is a much larger potentialmarket than for television, on a global scale.

2.2 Options

The existing analogue radio service has the advantageof having developed a common system, where virtu-ally any radio receiver capable of using AM and FM

(R029)

can be used anywhere in the world. This means thatreceivers can be made in quantity, cheaply. Currently,the use of digital signals for terrestrial radio broadcast-ing is undertaken using a single worldwide standard,the Eureka 147 system. However, there are other pro-posals which are being considered, especially in theUSA.

A system that can make optimum use of a satellitechannel may not be the one that would be chosen assuitable for terrestrial use. Clearly, a common systemis desireable if cheap receivers are to be made, even ifa less efficient method of satellite transmission werefinally to be used to achieve this end. Only when thenumber of listeners is large is it going to be possible tomake effective use of the medium, so the early avail-ability of cheap receivers is crucial to any businessplan. The Eureka 147 DAB system has been devel-oped to respect this need and is now acceptedinternationally as a standard DAB system, and theonly one with universal delivery application.

2.3 The Eureka 147 system

This system is a high capacity multiplex which isbroadcast on a large number of closely-spaced digitalcarriers based on COFDM (Coded OrthogonalFrequency Division Multiplex).3,5 It is designed tooperate ruggedly in the presence of low signalstrength, multipath interference and Doppler shiftcaused by mobile reception.

Very early on in the Eureka project, the advantagesof being able to use a single system for all types ofchannel was considered to be a key requirement.Consequently, Eureka 147 DAB is designed to beused on satellites, terrestrially or in a hybrid combi-nation, where a satellite signal is rebroadcast on thesame frequency to fill in the signal in places wherereception is poor. The system, principally designedas a rugged terrestrial option, is not the optimumthat could have been developed for purely satelliteuse; in its current terrestrially-orientated implemen-tation, it does require more power than some of theoptions that have been designed specifically forline-of-sight satellite channels.

2.4 Other systems

Other systems being proposed optimise the use of thesatellite transponder. This gives improved perform-ance of the satellite, at least for fixed reception, and solowers the cost to the broadcaster. The unansweredquestion, at the moment, is whether this reduces thecost sufficiently for listeners to be prepared to acquiretwo separate receivers – for respective satellite andterrestrial use.

3. SOLIDARIDAD SATELLITES

3.1 Introduction to the Solidaridad satellites

The BBC, in collaboration with Telecomunicacionesde Mexico (Telecomm) and the Instituto Mexicano deCommunicationes, carried out DAB experiments inMexico City using the Solidaridad 2 satellite. It wasconsidered that satellite digital audio broadcastingusing the Eureka 147 system would likely be an attrac-tive proposition for many parts of the world where aterrestrial-only service would be impractical. Solidari-dad 2 was designed for telecommunications –operating at L-band to and from the mobiles, and Ku-band to and from the earth station. But note that onlythe Ku-band up-link and L-band downlink (for trans-mitting the DAB signal earthwards) were used for theDAB experiments and tests (see Fig. 1). The Solidari-dad 2 L-band frequencies are just above thosedesignated for DAB (1.452 to 1.472 GHz). The experi-ments were mainly based on fixed reception using ahigh gain antenna, but some simple mobile tests werealso carried out.

3.2 General description

The two Solidaridad satellites, flights one and two,were launched in November 1993 and October 1994.They have been sited in the geostationary orbit locatedat 109.2° W and 113° W, giving them an elevationangle of approximately 63° in the Mexico City area.They are principally designed for communicationspurposes to furnish a diversity of domestic andregional services; so the Effective Isotropic RadiatedPower (EIRP) of each satellite is not as high as wouldbe desirable for broadcasting DAB signals directly tothe public.

Solidaridad embodies payloads in ‘C’ ‘Ku’ and ‘L’bands. In particular, the L-band payload has beendesigned for multicarrier use to provide voice, fax,data and position-reporting to aeronautical, maritimeand land mobile commercial services. Transportable orsemi-fixed terminals in remote areas of Mexico alsouse these services.

The satellite performs the conversion between the L-band signals sent and received by the mobile terminalsand the Ku-band used by the earth station. Fig. 1 showsthat both the forward and return transponders on eachsatellite are divided into a number of subbands; theseare interleaved between the satellites. The subbandsare independently controllable with bandwidthsranging from 2.5 MHz to 5.5 MHz. Parts of the bandare used by other operators, such as Inmarsat. How-ever, as stated in Section 3.1, for the purpose of theDAB satellite tests,only the Forward Link transponderwas used, with the L-band signal acting as the DAB

(R029) - 2 -

broadcast transmitter; so the bi-directional communi-cation aspect of the satellite (using the Return Linktransponder also) wasnot used for the DAB tests – seeFig. 1 for the frequency plans for both satellites.

The preparations for the experiments had includedcareful examination of the suitability of the Mexicansatellites for use with the Eureka 147 DAB signal.Special emphasis was put on parameters such as phasenoise, signal-to-noise ratio, transponder gain and therequired bandwidth. This information was used tocalculate the link budgets and to select the subbandconsidered suitable for the tests; subband one onSolidaridad-2 was chosen.

The DAB experiments were carried out using1.52915 GHz in the L-band. This frequency waschosen with due consideration to internationalfrequency coordination, bearing in mind that the DABspectrum has a bandwidth of 1.5 MHz.

3.3 Forward transponder

The Forward Link transponder receives Ku-bandsignals from the fixed ground station and convertsthem to the L-band for onward transmission to mobileor semi-fixed terminals. The Ku-band uplink signals,received by the satellite’s vertically polarised receiveantenna, are fed to the low noise amplifier (LNA) ofthe Ku-band transponder. The output of the LNA isthen equally divided between the Ku-band transponderand the L-band transponder by a hybrid circuit. ThreeLNA outputs are configured in this way, to provide 3-for-1 redundancy in the receive section of the L-bandtransponder. The signals for the Forward Link trans-ponder are coupled from the output of the LNA andbefore feeding the Ku-band receiver, in order to mini-mise the number of frequency conversions in the Ku/Ltransponder; thereby maintaining the requirement forfrequency precision.

After amplification, the received signal is applied to

1,626.5

1,629.5

1,635.0

1,638.51,639.0

1,644.01,645.5

1,649.01,650.0

1,652.5

1,657.0

1,660.5

9

1

2

3

5

6

8

9

1

2

3

5

6

8

248.5

251.5

253.5

257.0258.0

260.5261.0

265.0266.0

268.5

IF

RETURN LINK

11,948.5

11,951.5

11,953.5

11,957.011,958.0

11,960.5

11,965.011,965.0

11,968.5

Not used for DAB tests

1,525.0

1,528.0

1,533.5

1,537.01,537.5

1,542.51,544.0

1,547.51,548.5

1,551.0

1,555.5

1,559.0

9

1

2

3

5

6

8

9

1

2

3

5

6

8

245.0

248.0

250.0

253.5254.5

257.0257.5

261.5262.5

265.0

IF

FORWARD LINK

14,245.0

14,248.0

14,250.0

14,253.514,254.5

14,257.0

14,261.514,262.5

14,265.0

Used for DAB tests

14,257.5

Ku-band

L-band

receivingantenna

Solidaridad 2satellitetransponders

satellite newsgatheringvehicle dish antenna

Fig. 1 - Basic frequency plan for the Solidaridad 2 satellite DAB tests.

(R029) - 3 -

the input of a 17 MHz bandpass filter, which removesany frequencies destined for the Ku-band transponder.The contiguous 17 MHz band, used for mobile com-munications, is passed through this filter and thentranslated to an intermediate frequency (IF) between245 to 265 MHz by the Ku-band-to-IF downconverter.

The IF-to-L-band upconverters provide channel filter-ing, channel gain control, and eventual upconversionto the L-band transmitting frequency. Each upcon-verter contains independent subband filters.Channelisation of the IF signal into the subbands is ac-complished with surface acoustic wave (SAW) filters.

Once channelised, each subband signal level may beadjusted independently via the channel control attenu-ators resident in each IF-to-L-band converter. Thecontrol attenuation is achieved by a series of single-step attenuators, which allow a downward adjustmentof channel gain over a range of 15 dB from the nomi-nal channel gain in steps of one dB. The subbands ineach converter are then recombined and upconvertedto the appropriate L-band frequency plan (34 MHzbandwidth). This combined signal is first divided intofour separate inputs and fed to a 6-for-4 redundantswitch ring of SSPAs (solid state power amplifiers), tobe amplified for transmission by four SSPAs operatingin phase.

A master reference oscillator, which provides coherentlocal oscillators for all upconverters and down-converters, is employed in the L-band transponder toensure compliance with the requirements forfrequency translation accuracy during its operationallife, and also between each of the subband channels. A

10 MHz crystal reference oscillator has been chosenbecause of its demonstrated short- and long-termstability characteristics.

4. TEST SYSTEM

The object of the tests was to verify that the Eureka147 system is suitable for satellite transmission. Inparticular, it was hoped to measure the link marginsfor both static and mobile reception. The experimentalsystem used is outlined in Fig. 2. Two separate sourcesof stereo audio from two CD players were each Musi-cam-coded; usually one at 256 kbits/s and the other at224 kbits/s (one of the feeds was buffered and takenoff to be fed directly to the receiving equipment for A-B comparisons to be made). The two stereo feedswere then multiplexed into the first two slots in theEureka 3rd generation fixed multiplex. The signal wasthen COFDM-modulated and converted to an interme-diate frequency of 70.15 MHz, for transfer to asatellite newsgathering (SNG) vehicle (shown inFig. 3) provided by Telecomm, via an attenuator to setthe uplink level. The uplink was in the Ku-band, such,that it was translated by the satellite transponder to1.52915 GHz (L-band) for broadcasting. This waschosen because a bandwidth of at least 1.7 MHz wasneeded, which meant that it had to be outside the bandused by Inmarsat. The uplink site was the TelecommEarth Station in Mexico City.

For convenience, the receiving equipment, outlined inFig. 4, was set up adjacent to the uplink equipment. Ayagi antenna of about 1 metre in length (see Fig. 5(page 6)), modified prior to the tests to receive circular

audiocoder

audiocoder

audiomatchingamplifier

pre-amp

CD

multiplexer

sourcemonitor

256kbits/s

224kbits/s

laptopPC

2Mbits/s

COFDMcoder

COFDMmodulator

digitalI & Q

analogueIF

SNGvehicle

50 mcoax

70.15 MHz(1.5 MHz BW)

inputIF

Mexicanuplink

Ku-band

Solidaridad2

CD

audio audio

Fig. 2 - Outline of the experimental transmission equipment.

(R029) - 4 -

Fig. 3 - The satellite newsgathering vehicle at the Telecom ground station, Mexico City.

1.52915 GHz

~1.3029 GHz

amplifier

226.25 MHz

226.25MHz

audioplug-inoption

L-bandRHCPpatch

Solidaridad2

L-bandRHCPyagi

plug-inoptions

50 mtwisted

pair

15 V

50 mL-bandcoax

headend frequencyconverter

spectrumanalyser

DABreceiver DAT

audioamp

fromsourcemonitor

head-phones

speakers

Fig. 4 - Outline of the experimental receiving equipment.

(R029) - 5 -

polarisation, was used for the fixed tests; while for themobile tests, a patch antenna of about 18 cm diameterwas used (see Fig. 6). The gains of the fixed and mo-bile antennas were 14.5 dBi and 8.3 dBi respectively,each to an accuracy of ±1 dB as measured at the Na-tional Physical Laboratory, in the UK. The antennaswere located outside so as to have an unobstructedview of the satellite. The yagi antenna enabled suffi-cient signal to be collected to determine the influenceof the satellite systems on the DAB signal; the patch an-tenna is more appropriate for normal use.

The signal from the antennas was immediately amplifiedby a headend unit (as shown in Fig. 5) which also con-tained an L-band filter (with 2 dB insertion loss). Thesignal was then sent through 50 m of low-loss cable tothe rest of the receiving equipment located indoors andwas initially converted down to Band III. Then it wassplit, with one half feeding a Philips 452 test receiverand the other half being further amplified for displayon a spectrum analyser. The audio signal from the

receiver was played through loudspeakers (as couldthe direct feed of one of the audio sources mentionedabove) and also recorded on Digital Audio Tape (DAT).

For the mobile tests, most of the equipment could bepowered directly from 12 V. The exception was theMarconi synthesiser which had to be powered via aninverter. In the mobile tests, the signal was monitoredon headphones as well as being recorded.

5. EXPERIMENTAL RESULTS

The experiments were carried out at each of the threeDAB transmission modes; Mode I (1536 carriers,1 kHz spacing), Mode II (384 carriers, 4 kHz spacing)and Mode III (192 carriers, 8 kHz spacing). OnlyMode III has been specifically designed for L-bandsatellite operation, so it was interesting to see how wellthe other modes performed. The performance indicatorused was the bit error rate (BER) before Viterbi decod-ing (no error rate measurements were made after errorcorrection). The BER reading was readily availablefrom the Philips 452 test receiver that was used, as ithad been found to be a sensitive measure of audioquality. Measurements of the satellite Effective Iso-tropic Radiated Power (EIRP) were relayed from thesatellite control centre located in a building near to thetest site.

5.1 Back-off tests

The back-off measurements were carried out by vary-ing the signal level uplinked to the satellite and thenmeasuring the BER while noting the uplink power andthe satellite EIRP. The maximum permissible EIRPwas 43.5 dBW, this is short of the maximum possible(47 dBW) as the operators were being cautious regard-ing the effect that the DAB signal might have had onthe satellite transponder. Fig. 7 shows the resultsobtained, where the ‘threshold’ indicates the point atwhich bit errors produce audible impairment of thereceived audio signal.*

The differences in performance between the threemodes of operation can be seen to have been verysmall, but there is a noticeable increase in power pen-alty of about 1 dB between modes I and II. The mostlikely reason for this could have been the possiblepresence of phase noise in the upconverter oscillatorused in the SNG vehicle that did the uplinking – thesatellite on-board oscillator was known to have excel-lent performance. The slight increase in the scatterwhich shows in the results of modes I and II can becorrelated with the presence of dense storm cloudsduring that part of the tests. Although the L-banddownlink should not be affected by rain attenuation,

Fig. 5 - The encapsulated Yagi receiving antenna(necessary to receive the low-power satellite signal)

and associated headend.

Fig. 6 - Patch antenna being fitted to the roofof the pick-up truck. * The impairment was audible under non-critical listening conditions.

(R029) - 6 -

this is not the case for the Ku-band uplink (which wascosited); also, there is the associated change in skynoise temperature (i.e. background cosmic noise).

The results of L-band back-to-back tests, carried outby BBC R&D at its UK Kingswood Warren headquar-ters’ site beforehand, are also plotted in Fig. 7. Theyindicate that there is a power penalty when using thesatellite system. However it is the difference in slopethat is of main interest, showing that even in mode IIIthere is still a residual effect from the uplink and satel-lite which increases slightly with power. It is not clearwhether this effect is due to oscillator phase noise, am-plifier non-linearities or other causes.

Fig. 8 shows the variation in the uplink High PowerAmplifier (HPA) output against satellite EIRP, fromwhich it can be seen that the satellite transponder islinear. This is as expected because the transponderincorporates linearizing circuits for multi-useroperation.

5.2 Power budget

The power budget can be checked from Fig. 7 bytaking the satellite EIRP which crosses the thresholdof impairment; this value is 37.5 dBW for Mode III.Table I (overleaf) demonstrates the link budget forthis system. At 37.5 dBW transmitted power, a C/Nof about 7.8 dB could be expected. Previous tests2

indicate that at this C/N a BER of about 3 × 10-4

should be observed after error correction (4 × 10-2 isobtained before error correction). Additional differ-ences may be caused by the phase noise of theoscillators, but this should account for significantlyless than 1 dB of the difference. There may also beinaccurate assumptions about the figures assumedfor sky noise. Under these conditions, the results areclose to those predicted.

Another way of assessing the power budget was byoperating the satellite at high power (43.5 dBW) andinserting an attenuator between the high gain receivingantenna and the headend amplifier. The insertion lossof the attenuator with its extra lead was about 1.2 dB.Fig. 9 shows a graph of the measurements; from whichit can be seen that the curve crosses this threshold atan attenuation of about 2.0 dB. However, the actualmargin is greater than the 3.2 dB implied due tothermal noise introduced by the attenuator.

Table II (overleaf), following, shows the power budgetobtained with the attenuator set for operation at thethreshold point. This power budget also includes anextra 0.8 dB on the C/N to compensate for the distor-tion of the DAB signal at high power; as shown by theincreased separation of the Mode III and back-to-backcurves in Fig. 7. Note that the G/T has been modifiedto take account of noise introduced by the attenuator.

100

10-1

10-2

10-3

BER

33 35 37 39 41 43 45

Mode IMode IIMode IIIBack to backThreshold

satellite EIRP, dBW

Fig. 7 - Bit error rate against satellite power.

HPALinear

12

10

8

6

4

2

00 5000 10,000 15,000 20,000 25,000

satellite EIRP, W

uplin

kH

PA

outp

ut, W

Fig. 8 - Satellite transponder input/output characteristic.

1.0

0.1

0.01

BER

attenuation, dB5 4 3 2 1 0

Mode IIIThreshold

Fig. 9 - BER against received signal attenuation.

(R029) - 7 -

By comparing Tables I and II, the true power margincan be seen to be 6 dB using the high gain (yagi) an-tenna and with the satellite transmitting at 43.5 dBW.

Table I:Solidaridad power budget for DAB reception.

Satellite EIRPSpreading lossReceived signalEffective area of isotropic antennaReceiver G/TBoltzmann’s constantBandwith

Received C/N

37.5 dBW–163.0 dB–125.5 dB(W/m2)

–25.0 dB(m2)–8.5 dB/K

+228.6 dBJ/K–61.8 dBHz

7.8 dB

Table II:Power budget for DAB reception with an attenuator.

Satellite EIRPSpreading lossReceived signalEffective area of isotropic antennaReceiver G/TBoltzmann’s constantBandwithAttenuationReceived C/N

43.5 dBW–163.0 dB–125.5 dB(W/m2)

–25.0 dB(m2)–10.7 dB/K

+228.6 dBJ/K–61.8 dBHz

3.2 dB8.6 dB

Table III:Solidaridad power budget for mobile DAB reception.

Satellite EIRPSpreading lossReceived signalEffective area of isotropic antennaReceiver G/TBoltzmann’s constantBandwidthMarginReceived C/N

43.5 dBW–163.0 dB–125.5 dB(W/m2)

–25.0 dB(m2)13.2 dB/K

+228.6 dBJ/K–61.8 dBHz

0.5 dB8.6 dB

During these tests, it was not possible to see any in-crease in intermodulation products (IPs) within thelimits of the received signal-to-noise ratio. The IPlevel could be expected to be low because the SNGuplink was operating a 300 W TWT high poweramplifier (HPA) at well below its maximum outputpower (at about 10 W), and also the HPA on Solidaridadsatellite is equipped with linearizers; though there isthe unexplained 0.8 dB power penalty at high power.

5.3 Mobile reception

A pick-up truck was equipped to receive the satelliteDAB signal from a patch antenna mounted at 63° onthe roof using a magnetic base (Fig. 6). This antennahad a measured gain of 8.3 dBi and a 3 dB beamwidthof about ±35°. The receiver was found to be operatingjust within threshold at a BER of between 2 to 4 × 10–2

when the vehicle was stationary and the satellite wasoperating at an EIRP of 43.5 dBW. Table III shows themobile power budget, confirming that the operatingpoint was indeed close to threshold:-

It was clear that the mobile tests would have to be per-formed with the vehicle travelling in a straight line, inorder to keep the antenna orientated correctly toreceive the signal. A series of measurements of BERagainst velocity were recorded as the vehicle drovesouth along a roadway within the Telecomunicacionesde Mexico (Telecomm.) compound. Tests were carriedout at velocities of 15, 30, 45 and 60 km/hr in each ofthe three DAB modes; the maximum speed was limitedby the length of the roadway and by necessary safetyconsiderations. The measured BER was within the range3 to 5 × 10–2, and did not depend on vehicle speed.However, a slight increase in BER was sometimesrecorded during the acceleration period, but a firm cor-relation could not be proven within the limitedacceleration capabilities of the vehicle used for thetests. One test was carried out with the vehicle driving inthe opposite direction, but no differences in resultswere noted.

The component of the vehicle’s velocity in the direction ofthe satellite was about half its road speed due to thehigh elevation of the satellite (63°). Thus, at the highestspeed of 60 km/hr, the Doppler shift was 38.6 Hz whichcan be accommodated even in Mode I – which has theclosest carrier spacing. It is important to note that theseexperiments have not tested the system performance un-der conditions of differential Doppler shift caused bythe presence of signal reflections (multipath); a higherpower satellite would be needed for such tests, whichshould also include measurements at motorway speeds.

The vehicle was driven through the suburban streets ofMexico City for about 40 minutes at speeds of around40 km/hr in order to qualitatively assess the reception.Because the antenna needed to be kept manuallyaligned with the satellite, it was not always possible toensure that optimum alignment was maintained. How-ever, some interesting observations were made:

• It was found that when the antenna wascorrectly aligned on the satellite, and thatthere was a clear unobstructed view, thengood reception down to a BER of about3-5 × 10–2 could be obtained.

• Any shadowing, e.g. by buildings, bridges andtrees, caused the signal to be lost for thatmoment. It is, therefore, clear that there wasinsufficient power margin to test the capabil-ity of the DAB receiver for making use ofreflections.

(R029) - 8 -

• A simple test of indoor reception of satelliteDAB for portable receivers was carried outthrough the window at the satellite earthstation. But it was found that there wasinsufficient power available to receive ausable signal on the lower gain patch antenna.

6. DISCUSSION

These tests have shown that satellite broadcasting ofDAB is feasible, even in Mode I, if the satellite eleva-tion is high enough to reduce the Doppler shift of anyreflections. The Solidaridad transponder has linearizercircuits which keep the IPs to low levels, althoughabout 0.8 dB of power penalty did occur under highpower operation. The slight differences in the recep-tion quality that did occur between the modes is mostlikely to have been due to residual phase noise in theequipment of the SNG uplink vehicle used; if so, therewould be scope to have a better performance from aspecially designed uplink.

It is possible, though, to make a first-cut estimate ofthe requirements for a practical satellite DAB broad-casting system to mobiles from a geostationarysatellite, assuming:

• that mobile reception requires a non-steerableantenna with an omnidirectional gain patternin the horizontal plane, this (along with massproduction considerations) means that the an-tenna gain will probably be about 4 dBi;

• that the fade margin required to cope withshadowing is 6 dB (and this may cause addi-tional noise from the environment);

• that 0.2 dB margin is required for rain attenu-ation (contributing some noise);

• that 1 dB margin is required for system penal-ties (partially contributing noise).

Table IV shows that the satellite EIRP, as seen fromthe edge of the service area, needs to be 55.2 dBW.The satellite power could be reduced by minimisingthe system penalties and the distortion.

A satellite with a high power transponder will beneeded for the next stage of mobile reception tests inorder to fully explore the ability of the Eureka 147system to make use of reflections from buildings. Sucha high power transponder would also allow indoorreception experiments with portable equipment to bemade.

7.7. CONCLUSIONS

Tests of satellite DAB broadcasting in the L-bandfrom the Solidaridad 2 geostationary satellite, coveringMexico, confirmed that the Eureka 147 DAB system issuitable for use in satellite broadcasting.

Neither the satellite nor the uplink significantly af-fected the DAB signal.

It was possible to use Mode I, which is not intendedfor use above 375 MHz. There was just sufficientpower received to allow tests of mobile reception to becarried out. This showed that Doppler shift was not aproblem at road speeds of up to 60 km/hr with a geo-stationary satellite (which appeared at high elevationin Mexico).

It has been possible to estimate from these tests theaggregated power required for a satellite broadcastservice to mobiles. These early results obtained fromthe use of higher transmission powers clearly indicatethat further tests are required to discover what valuesignal reflections may have for mobiles passingthrough cluttered streets. Such tests would also beneeded to examine the possible use of reflections forfixed installations which were located in areas wheresatellite signals were shadowed by high-rise buildingsor steep terrain.

8. ACKNOWLEDGEMENTS

The authors thank colleagues at Telecomunicacionesde Mexico and the Instituto Mexicano de Comunica-ciones for the use of the Solidaridad 2 satellite, theearth station facilities, test equipment and their enthu-siastic help given during these tests.

9. REFERENCES

1. LEE, M.B.R., 1993. Planning methods for anational single frequency network for DAB.Proceedings of International Conference on

Table IV:Estimated power budget for a satellite DAB service.

EIRP at edge of service 55.2 dBW

Subtract:Attenuation at L-bandReceiver C/N requiredRF bandwidth occupiedRainSystem penaltyNoise temperatureDistortion penalty

187.4 dB8.0 dB

61.8 dBHz0.2 dB1.0 dB

22.6 dBK0.8 dB

Add:Boltzmann ConstantAntenna gain

228.6 dBJ/K4.0 dBi

Fade margin 6.0 dB

(R029) - 9 -

Antennas & Propagation, IEE Conference Publi-cation No. 370, pp. 970-947.

2. ITU-R Special Publication, 1995. Terrestrial andSatellite digital sound broadcasing to vehicularportable and fixed receivers in the VHF/UHFbands, Geneva.

3. ALLARD, M. andLASSALLE, R., 1987. Princi-ples of modulation and channel coding for digitalbroadcasting for mobile receivers. EBU TechnicalReview No. 224, August, pp. 168-190.

4. ITU-R Recommendation BO. 789, 1994. Digitalsound broadcasting to vehicular, portable andfixed receivers for BSS (sound) bands in the fre-quency range 500 – 3,000 MHz. March.

5. SHELSWELL, P., 1995. The COFDM modulationsystem: The heart of digital audio broadcasting.Electronics & Communication Engineering Journal,7(3), June, pp. 127-136.

(R029) - 10 -

Published by BBC RESEARCH & DEVELOPMENT DEPARTMENT, Kingswood Warren, Tadworth, Surrey, KT20 6NP