special issue on new spread spectrum, lan and pcn productssss-mag.com/pdf/sssv1n7.pdf · nthly news...

nthly News Magazette

0 199-7 by RF%SingIr Copy Plice $3.00

Special Issue on New Spread Spectrum, LAN and PCN Products

ynJj&$J& j&-y&(&

Article mRumors & Ramblings 2Deciphcrings 2Editorial 2The Aerial 3Technical Trendsin Education 4New Products 9,14,15,21DSP for SS 5International Scene 5Washington Sccnc 5Cartoon 5Secret SS Signals 6A 16 Khs GPSS Radio 7Introduction to The Navy’s

IPANSAT - Part II 10DSP Tutorial 17More New Products 22Novemher SSS Preview 32

I

The FulLANs and PCNs

Some of the latest wirelessproducts for LANs and PCNs arepictured here, Do these productsindicate the filtllre d irect ion ofSpread Spectrum applications?

Are we about to enter theera of the PDA (Personal DigitalAssistant, or Appliance)? Or are Bthese sleek new products justmore misguided marketing ideas

4

that represent sidetracks to the 51

direction of progress in thisbusiness? This month’s editorial 1’d iscusses these and otherimportant issues for Spread

Are We Ready For This?

Spectrum’s future.More New Products

see EDI’roI<IAL ,“C’ 2 Inside Spread SpectrumScene

Spread Spcctruni Scent rIOctober, lYY2

SPREAD SPECTRUM SCENE isdedicated to the Spread Spectrumprofessional and is committed tobeing the primary source for the lat-est news and information about thegrowth, regulation, and opportunitiesin this emerging science.

SSS provides a forum for publicationof technical information, advertising,editorials, opinions, and newsrelating to the emerging fields of ourcoverage and emphasis. SSS isread by over 5000 technical decisionmakers each month. SSS canpresent your advertising message tothe key designers, programmers,system integrators and end users inthis new industry. Call our 800number Hotline to request a MediaKit.

Editor & Publisher:Randy Roberts

Editorial Consultants:Koert Koelman,M. N. RobertsContributors:

Tom Diskin, Dan Doberstein,John Greene, Matthew

Johnson, Peter OnnigianCartoonist:

Rich TennantR&D Staff:

Walt Jensen,Chris Kilgus,

Dr. Andy Korsak

Published bv: RF/SSP. 0. Box 2199

El Granada, CA 94018-2199

TeleDhone Numbers:Voice: 415-726-6849FAX: 415-726-0118

InterneVUUCP Email:[email protected]. RROBERTS4

AOL: randy&s

Advertisins & SubscriDtionHotline:

Voice: 800-524-9285

SSS is published monthly and isavailable by subscription. You canreceive a complimentary copy bysending us a self addressed 9 x 12inch stamped (S.75 US postage)envelope.

Subscr!p_llo~ Rates:12 Months - US First Class Mail - $29.9512 Months - Foreign, AIR MAIL - $45.00

US Funds

0 Telxon Corporation announcedthat its joint project with Wal-Mart Stores, Inc. to implementthe wireless spread spectrumradio communication network hasbeen successfully installed inrecord time in all of Wal-Mart’s1,804 stores nationwide.

0 Telxon Corporation alsoannounced a unique plan withARDIS, a joint venture of IBMand Motorola, to allow fieldservice organizations to test thep o w e r o f h a n d h e l dmicroprocessors and wide areawireless data networks.

0 Bell Atlantic and Cellular DataInc. are installing a new wirelessdata system based initially inBaltimore. Bell Atlantic is alsoworking with I B M a n dWestinghouse Electric to provideother facets of the proposed newservice.

l Heard a good rumor -- want to“leak” some info to yourcompetition -- call our 800number and we may print it!

lDecipheringsl

MAKE VOYAGES.ATTEMPT THEM.

THERE-S NOTHING ELSE.

- Tennessee Williams-

Don’t miss an issue ofSpread Spectrum Scene.

Subscribe now!

I mamBmWhither Wireless LANs

and PCNs?The front cover of our

rather late, special October issueshows some of the latesthardware for LANs and PCNs.Are these products representativeof the future direction of appliedspectrum technology? Weattempt to answer this interestingquestion in this issue and in thiseditorial.

First let me say that thisspecial October issue is devotedto what industry is currentlydoing with spread spectrumtechnology. You will see quite afew new products and someinteresting advertisements in thisissue. Some of these productswill undoubtedly find a marketniche. Other products, that areless well marketed, conceived orfunded will probably go the wayof the dinosaur. Honestly, evenmy crystal ball is a little foggywhen it comes to picking thesurvivors from the dogs in thisbusiness!

Given the tough highlycompetitive marketplace out thereand the fact that no majorsegment of the consumer markethas demanded large volumes ofthese products yet, thismarketplace is like a horse racerun on an old fairgrounds’ dirttrack during the off season -- alot of nags competing with a fewdark horses. Are we ready yetto see a shake out ofmanufacturers and products?Well, not really -- some 31vendors or interested parties sentthe FCC comments on theproposed revisions to Part 15spread spectrum rules, some 2l/2 years ago. Now only about15 of those companies are still in

Spread Spectrum Scene October, lYY2

business. Thus, shake outs seemto be a definite on-going process,not always related to themarketplace or product mix.

As we all have noticed, oureconomy is a mess! It may betough for any new, currently non-essential technical gadget to gainmarket acceptance. Everybodysays things will be better after theelection. But, I wonder! Justmaybe, the economic malaise weare suffering is due more to ourgadget orientation, our fascinationwith high technology and B-schoolMBA company Presidents/CEO’s,who have forgotten (or neverknew) the art of critical thinking.

This country needs to getback to basics: we need to createjobs! This means manufacturingjobs -- not service sector jobs. Weneed to stop exporting rawmaterials and resume exportingand dominating the industries ofthe world through high technolocv.Forget the frivolous -- come upwith products that improveproductivity, reduce costs andtill a real world need in anenvironmentally sound way!

Maybe when engineers andscientists run our companies, likethey do in Europe (not B-schoolMBAs) -- just maybe, we’ll have achance again. Until then, guessingthe direction of SS or any othertechnology is just a waste of time!

There were two very meaningfulconferences in the last 6 weeks orso: (1) The Wireless DataConference and Exposition, Oct.13-15, in Santa Clara and (2) TheIEEE sponsored First AnnualWireless LAN ImplementationConference, Sept. 17-18, inDayton. If any of you areinterested, next month we’llhighlight the most interestingpapers from one or both shows.Call, write or FAX your wishes.

)y Peter Onnigian, P.E., W6QEU

Gain Over What?Recently I gave a talk on

Yagi antennas at a nationalamateur radio convention. WithIearly 200 in attendance, i tsecame apparent in the Q and Asession, many were antennaengineers and technicians, somefrom large defense firms. It wassad to hear a few did not knowthe di f ference between anisotropic (dBi) and a half wavedipole (dBd) reference. Theyalso had difficulty understandingwhy the gain of the same antennamay be expressed as a highernumber in dBi than in dBd!

Antenna Gain ReferencesTo understand an isotropic

antenna, imagine the radiatortotally enclosed in a hollowsphere. The radiation from itscenter is distributed uniformlvover the interior surface of thesphere. This uniform radiation issaid to be isotropic by definition.

Assume further, it takesone watt of power to cover theentire surface of the sphere with100 milliwatts of intensity. If wewere to illuminate only a smallportion of the sphere with thesame surface intensity, it followsthat the radiation source powerrequired would be much less thanthe one watt required toilluminate the entire sphere. Infact, a dipole would illuminate awide band only on the spherewith the same 100 milliwattsintensity.

In fact, it would requireonly 0.61 watts of power for the

Octohcr, lYY2 Spread Spectrum Sreuc

same intensity as required to coatthe entire sphere. This reductionto 61 percent by a dipole is equalto -2.146 dB, rounded off to-2.15 dB. Thus the isotropicsphere has 2.15 dB more gainsince it requires more power toilluminate the entire sphere!Gain is the ratio of the maximumradiation in a given direction tothe maximum radiation producedin the same direction from areference antenna both with thesame input power.

Another definition: Thedirectivity is the antenna’s abilityto concentrate radiation in aparticular direction. Usefulantennas exhibit some directivityunl ike an isotropic, whichradiates equally in all direction.As stated in last month’s column,an isotropic antenna exists onlyas a mathematical model, and isnot realizable in practice.

The gain of an antenna isa basic property and is frequentlyused as the figure of merit. Gainis the directivity of an antenna,less the various losses inherent init. These include IR, dielectric,VSWR mis-match, undesired sidelobes, front to back ratios andother losses.

Gain NumbersOur interest is the relative

gain of commercially availableantennas. The common practiceexpresses gain in decibels relativeto that of a half wave dipole.This gain is expressed as dBd,that is decibels over a dipole.However this does not hold truefor all those available in themarket place.

For example antennabrand X is rated as 8.5 dBd.This is equal to brand H antennawhich is rated as 10.65 dBi [8.5+ 2 .15 = IO.651 Misleading‘!

3

Acrlal from page 3 - To some yes, Verymisleading!

Suppose you have tochoose an antenna for the highergain. They are both pricedequally. Would you pick the onerated at 10.0 dBi or the one rated9.15 dBd? To find the answer,level the reference. Eithers u b t r a c t 2 . 1 5 dB from theisotropic to equal dBd, or add2.15 dB to the one rated in dBd.Don’t compare dBi with dBd.That’s comparing apples withoranges. Most engineerscompare with t h e d i p o l ereference, so subtract 2.15 fromthe dBi value, to equal 7.85 dBd.The antenna rated 9.15 dBd has1.3 dB higher gain than the onerated as 10.0 dBi.

If the manufacturer insistson playing the numbers game andrates his antennas in dBi, thenyou should normalize it to dBd toget the true value.

Truth in LabelingThere is nothing that beats

measuring antennas on a certifiedand accredited antenna range.While expensive, the numbers areaccurate. This procedure ofrange measurement eliminates thesales hype and gives thebuyer/user confi rmat ion ofantenna gain numbers.

Unfor tunate ly in thespread spectrum and burglara l a r m fields there are veryunscrupulous manufacturerswhose antenna gains are highlyover rated. Then there are thosethat don’t indicate if the gainstated is simply guess work,measured or developed by thesales department! The worstoffense is stating gain in dBwithout reference to isotropic ordipole. Maybe it’s gain over awet noodle!

by Tom Diskin

Technical Trends InEducation

Some years ago when Iwas a budding electronicshobbyist, I remember reading anarticle in a leading electronicsmagazine enti t led “MeTechnician, You Engineer.”This article carefully delineatedthe separate tasks of bothtechnicians and engineers, aswell as how they must worktogether to achieve theircommon goals. I believe theauthor’s goal in writing thispiece was to emphasize thatthese two job titles are bothmutually exclusive of eachother, while at the same timehighly dependent in terms of theend product they would jointlyproduce.

Around the same time, apaper was delivered toelectronics instructors inCalifornia at a meeting of theElectronics Instructors Councilon November 2 1, 1959 entitled“ B u i l d i n g S t a t u s f o rTechnicians.” In hispresentation, Mr. Charles R.Mulkey of Monterey PeninsulaCollege criticized those whodescribe technic ians as“assistants to engineers” or onewho ” -works for an engineer or

I under the close supervision ofan engineer.” He maintainedthat technicians, like engineers,create, design and build, and insome cases do the job wellenough that they are often moreh i g h l y p r a i s e d b y t h e i rcompanies than the bestengineers. In some cases, seniortechnicians often act asconsultants to engineers, forthey are the members of theteam who understand thematerials, processes andlimitations of techniques neededto successfully complete aproject. Mr. Mulkey furtherstated that the technician “...mayin reality be the functioning partof the engineer. He may be theengineer’s only real connectionwith a real world. He is veryoften the ‘practical’ brain of theengineer. He is, above all, aman who knows the newmaterials and the newtechniques, and he knows andunderstands their limitations.”In other words, he is the onewho really applies thetechnology of our modernWestern civilization, withunderstanding of its scientificbasis and a practical knowledgeof its limitations.

While we may take issuewith some of Mr. Mulkey’sstatements, we must temper thiswith the understanding thatthese statements were madenearly 33 years ago! Howmuch has the status of thetechnician changed in theensuing years‘? While we haveadded many related job titles,s u c h a s “ E n g i n e e r i n gTechnician”, “Field ServiceTechnician”, and the like, hasthe relationship See Education page 29

4 Spread Spectrum Scene October, 1992

&~&~ti~g)&J,)$& &+@m@

G e r m a n y r e c e n t l yderegulated and re-structured thePostal Administration. TheDeutsche Bundespost (DBP) nowhas the Ministry of Posts andTelecommunications (MPT) withthree separate departments incharge of all telecom.

The British “Duopoly” oftelecom includes BT plc andMercury Communications Ltd.T h e U K O f f i c e o fTelecommunications (Oftel)s tar ted a “duopoly review”process that initiates a newcompetitive era in the Britishtelecom industry.

The French have also re-vamped their century old telecomlaws. It’s still a governmentmonopoly, but value-addedservices, cellular telephones andterminal equipment can be soldcompetitively.

so close, Congress went home forthe year. They left a lot of workundone! Both S. 73, the AmateurRadio Spectrum Conservation

Technologies Initiative died frominaction. Maybe next year theywill really get to work --especially if we see some newfaces in Congress because of theelection!

The FCC has been prettybusy lately, however. They gavePioneer’s P re fe rences to :American Personal Commun-ications, Cox Enterprises ofAt l an ta and O m n i p o i n tCommunications of ColoradoSprings (R. C. “Bob” Dixon’scompany). All of thesecompanies were rewarded with anentitlement to a license in themarket of their choice for PCNservices.

The FCC also issued a 97page NPRM on the PCN/PCSservices. So many companies aremore than eager to enter this $35to $40 Billion a year market(according to a recent Arthur D.Little report) that the FCC justcan’t wait for congressionaldirection and enabling legislation.Look for some of the currentFCC commissioners to take jobsin the PCN industry, just after theelection!

DSPBill and S. 218, the Emerging I

In my last column, Ipromised to write on HDTV.Unfortunately, I made it soundlike 1 would cover much more ina single column than makessense, so this will be the first ofseveral columns on HDTV

So what is HDTV andwhat is it doing in a DSPcolumn? HDTV is a genericname for any one of a number ofschemes for t ransmit t ing atelevision signal with about twicethe horizontal resolution andtwice the vertical resolution ofthe NTSC system we use forcolor television in the US today.Since N T S C r e q u i r e s 6MHz/channel (and that is with avery narrow guardband) doublingboth the horizontal & verticalresolution demands 2x2x6 = 24

see DSP page 27

By Rich Tennan t

Octohcr, 1002 Spread Spectrum Scene 5

---_ --===--r_ __ = -= q =--- ===--__ ___ -- =-=== =---_ __------=----e e----_ __ __ ___ __ __ __ P :: ::----__-_--=-_=_====_ __-GE ZZLC=-__ _-__-

------__- -==-_=_--_-=-_= == =_ __ __- --= q L’= q ==--r= = =-_= =====,= 3 _ __ __-_G-4 = _ _-___-_ -- ___-___

----- - =-_-___ _-_ _-_ = =_ _-_a _ _ - = ,I,

p=-------= F__ ___ __- __ ___--_--_=-_-----_=-_-____--

In our last column wediscussed some of the availableliterature on this fascinatingsubject. We found an absolutedearth of useful informationabout “Secret Signals.” Noreader has bothered to share anyspecial information with us either.

Is there really anythingthat can be done to explore thesesignals we have seen and knowthey are on the air? The answeris Yes and No! Yes there issome general tutorial informationthat can be shared with ourreaders. But the answer is alsono, because to give you detailed,how-to information on crackingthese signals would be a Federalcrime.

So, we come to the firstreal meat in this column: Howd o y o u (theoreticully) t/-y t oohtuin uxftil i~fi,rniution .fkni“secret” spread spectruni signuls,when you do not know the exuctcode being ured ? The followinginformation should be useful inanswering this question -- anycoded system (or cipher for thatmatter) can be attacked by astraightforward crypto-analyticattack, whether you know thecode, key, or are authorized ornot! Whether the attack issuccessful depends on manyvariables -- but, let us say thatg i v e n p a t i e n c e , a d e q u a t ecomputing power (and somecustomized software) and just abit of luck you may find that

J

today’s commercial secret signalsand even N B S s t a n d a r dencryption are merely child’splay!

The basic principle toapply is called a “known cleartext” crypto-analytic attack. Theidea is very simple: somehow,you obtain a sample of “cleartext” -- which could be a picture,a valid user ID code, anauthentication key, a valid billingnumber / account number, etc.You may actual ly have tosubscribe to the service for ashort period to obtain the valid“clear text.” However, once youhave obtained a sufficiently longsample of “clear text” that youhave reason to believe will besent again sometime in the future- you are ready to attack thesystem.

Most commercial systems,re-key (or change keys/codes)infrequently, usually sometime inthe late evening / early morningor even only once a week. Sothe same keys (read: spreadingcodes) are used for many hoursor days. This presents theopportunity to record largeamounts of encoded stuff for lateranalysis.

The analysis is very,very simple: you are looking fora repeat of your known “cleartext.” Of course it would help ifyou knew when the known “cleartext” was going to be transmittedagain -- but this is not absolutelynecessary, is it?

You simply correlatelinear samples of incomingdemodulated data with your“clear text,” if something thatlooks like a repeating, fixed codepops out, voila - you ji)und it!Next month we will give a fewmore theoretical details.

FEDERALCOMMUNICATIONS

COMMISSION RULES -PART 15

Paragraph 15.249OPERATION WITHIN THE

BANDS 902-928 MHz,2400-2483.5 MHz, 5725-5875MHz, AND 24.0-24.25 GHz.

(a) The fleld strength of emissions fromintentional radiators operated withinthese frequency bands shall comply wtththe following:

Fundamentalfrequency

Field strengthof fundamental(mUUvolts/meter)

902-928 MHz 502400-24X3.5 MHz 505725-5875 MIIz SO24.0-24.25 GIIz 250

Fundamentalfrequency

Field strengthof harmonics(mUUvolts/meter)

902-928 MHz so02400-24X3.5 MHz so05725-5875 MHz so024.0-24.25 t:IIz 2500

(b) Fleld strength limlts are specified ata distance of 3 meters.

(c) Emissions radlated outside of thespeclfled frequency bands, except forharmonics, shall be attenuated by atleast SO dB below the level of thefundamenlal or to the general radiatedemlsslon llmlts In 15.209 whichever Isthe lesser attenuation.

(d) As shown In 15.35(b), for frequenciesa b o v e 1 0 0 0 Mllz, the above fleldstrength UIII~~S are based on averagelimits. Ilowever, the peak fleld strrngthof any emission shall IIU~ exceed themaxlmum permitled avrrage l i m i t sspeclfled above by more than 20 dl%under any condillon of modulallon.

(e) Parllrs consldrrlng the manuracturc,Importation, marketing or operation ofequipment under tlils scc1l011 should alsonutr ll~e rcqulremeut In 15.37(d).

Thanks to:

Rules Scrvicc CompanyRockville, MD - (301) 424-9402


A 16Kbs Full DuplexSpread Spectrum RF

Data Link

by Dan Doberstein, PresidentDKD Instruments

This is the first in a seriesof columns describing a fullduplex Direct Sequence RF datalink using the cordless telephonechip set. Although this frequencyband is not allocated for spreadspectrum use for very low poweruses it is legal and with a coupleof mixers and associated LO’s thefrequencies used can be any thatare currently allocated for SpreadSpectrum.

The system block diagramis shown in figure 1. TheMC145168 and the MC3362form the heart of the system. TheMC145 168 is a Dual I’LL thatlocks both the transmitter VCOand the first LO of the receiverportion. The frequencies used fortransmit and receive are burntinto the internal ROM. Thismeans you cannot change thefrequencies used for xmitt andreceive to any arbitrary onesdesired but will have to live withthe ones provided for. A table of16 transmit and receive pairs areavailable in the 45 to 49 Mhzband. The channel select inputsare used to select which channelpair you are on.

The MC3362 is a singlechip dual conversion receiver thatwas designed for the cellular andcordless telephone applications.In this design it used to convertthe received signal to the first IFof 10.7 Mhz and then to the 2ndIF of 455 KHz. A single tunedbandpass is used for the 1 st IF of10.7 Mhz as a wide bandwidth isneeded here due to the fact that

1 signal is not despread yet. The-chip rate of about 1 Mhz dictatesthe approximate 1 Mhz BW ofthis filter. The 10.7 is then mixedwith the 10.24 Mhz LO which isBi-phased modulated with areplica of the transmitted phasecode. Assuming correlation andcode lock the signal is nownarrow band and is passedthrough the 455 Khz ceramicfilter. This filter has about a 12Khz BW. This is just wideenough to pass the 16Kbs digitaldata stream. the data stream canbe user data or using the CVSDchip speech data. The data is bi-phase modulated onto the carrierafter being exclusive OR’ed withthe spreading code. Aftercorrelation the data can berecovered at the receiver endu s i n g t h e costas l o o p d a t ademodulator.

The MC341 8 chip is aContinuously Variable SlopeDemodulator/Modulator(CVSD).It takes in speech or other analogsignals and converts them to aserial bit stream. You can choosethe bit rate by changing thesupplied clock. A bit rate of16KBS is a little marginal butintelligible speech can be sent.The use of this chip reduces thebit rate for speech considerablyover straight ahead application ofsampling at say 8 bits every 125usec. One of the features of thischip is its ability to modulateAND demodulate the serial bitstream.

One of the design goalsfor this system was to minimizethe number of crystal sourcesused. This has cost as well aspractical benefits. The PLL’s, 2ndLO, Code Clocks and DataClocks all are locked to orderived directly from one crystal

oscillator at 10.24 Mhz. Thismultiple use of one frequencyreference simplifies the hardwareconsiderably, rather than havingseveral xtal oscillators to worryabout one good one is used.Another spin off of the entired i r e c t s e q u e n c e / c o h e r e n tarchitecture is that the data clockis AUTOMATICALLY recoveredonce code lock is achieved. Thissimplifies the data recoveryproblem considerably.

There are two codegenerators for each station, onefor the transmitted data and onefor correlating with the incomingreceived signal. The 10.24 Mhz isdivided by 10 for the transmittercode generator. The receiver codegenerator is clocked by the CodeSync and Clock Generatorcircuity. T h e c o d e u s e d f o rtransmit WILL NOT be the sameas for the received code,i.e. onecode pair will be used for eachstation pair. The user will be ableto specify which code pair is tobe used by changing jumpers onthe code generators. The use ofdifferent codes for xmit andreceive helps with the rejectionproblems of a shared antenna.

The Code Sync and ClockGenerator circuitry provides amodulated version of the IO.24Mhz reference divided by 10 togive approximately a 1.024 Mhzcode clock for the receiver codegenerator. The 1.024 Mhz clockis modulated in such a way as tokeep the receivers code alignedwith the incoming code fromtransmitter at other station. At theheart of the code sync method isthe tried and true tau dithertechnique. We will discuss thedetails of the implementation ofthis technique next month.

October, 1992 Spread Spectrum Scow-

7

Low power for mobile communications equipmentThe UMA1014T single-chipfrequency synthesizer is a low-

power device which provides acost-effective solution to frequency

synthesis in battery-powered

mobile communications equipment

chip all the functions required In an

RF frequency synthesizer. Functions

microcontroller.

The charge pump output of thecomparator, which charges/discharges

an external capacitor to provide thecontrol voltage for a voltage-control-

led RF oscillator (VCO). settles rapidly

after frequency changes. Only minimal

passive filtering IS required betweenthe comparator output and the VCO.

_ such as cellphone handsets,PMR (Private Mobile Radio) and

pagers. It is an integral part ofthe Philips chipset for mobile

communications equipment and

complements the companyS exist-ing range of mobile radio ICs from

Europe and the USA.

The UMAI 014T provides in a single

integrated into the device include anosci l lator c i rcui t , prescalers and

programmable dividers. These are

necessary to ‘frequency divide’ thereference and RF frequency inputs -

thereby completely eliminating the

interfacing problems assocrated with

external prescaling. The divided down

reference and RF frequencies arethen compared in an on-chip phase/

frequency comparator. In addition, theUMA1014T has an 1%bus interface so

that it can be controlled by a system

Providing a cost-effective solution to frequency synthesis m battery-powered mob//ecommunica tlons equipment - the l/MA 10 14T chip.

Finally, the excellent stability of the

charge pump current-both to changes

in temperature and output voltage -allows the UMA1014T to be used in

accurate and stable phase-locked-loopcircuits containing very few external

components.The UMA1014T is available ‘in a

16-pin small outline plastrc packageand has a current consumption of 13 mA

when active and only 2.5 mA when In

power-down mode, which contributes

significantly to extended battery life Inportable equipment.

An extensive applrcations report

d e s c r i b e s t h e o p e r a t i o n o f t h e

UMA1014T. I t also gives d e t a i l e dtheory for calculating the values of the

loop elements, describes applrcationsand gives sample pnnted circuit board

layouts.

Block diagram of the UMA 1014T angle-chip frequency synthesizer

October, 1992 Spread Spectrum Scene 9

Acknowledgements for the NPS PANSAT Article Series

SSS wishes to gratefully acknowledge the help, encouragement and co-operation of

Mr. John Sanders, Deputy Director of Public Affairs at NPS. Professors Rudy

Panholzer, Ed Euler and Tri Ha were also extremely helpful in the preparation and

review of these articles. Finally, we would like to thank Mr. Dan Sakoda, PANSAT

Systems Engineer for showing us around the ultra-modern shop, lab and classroom

facilities at NPS.

Prof. Rudy Panholzer (center). Prof. Tri Ha, NPS grad student (1. to rt.)

John Sanders, Dan Sakoda (1. to rt.) Lt. Tom Fritz, USN

Spread Spectrum Scene October, 1992

Introduction toThe Navy’s

PANSATProject - Part 2

by Randy Roberts,KC6YJY & RF/S!3 Director

The author visited the NavalPostgraduate School (NPS) in Monteny,Caltfornia on July 28, 1992. Randy wasgraciously hosted by Professor RudyPanholzer, PANSAT P r i n c i p a lInvestigator and his staff Most of ourquestions about PANSATand how it willoperate were answered at a weeklydevelopment/project status reviewmeeting chaired by PANSAT ProjectLead, Pn?fessor Ed Euler The,following concluding article highlightsthe mission, current status and plans.forthe 1995 launch of PANSAT

PANSAT Project RecapPANSAT is the acronym

for “Petite Amateur NavySatellite”. T h e N a v y ’ se x p e r i m e n t i s l i s t e d a sExperiment Number NPS-901.Last month we presented thebasic details of the PANSATproject. This month we concludethe description of the Project. Thebasic concept of PANSAT is tostore and forward a packetizeduplink message until the destina-tion station is in view of the

satellite and then re-transmit themessage on the downlink. ThusPANSAT will be functionallyanother Pacsat -- the big differ-ence is the signals needed toaccess the satellite will be SpreadSpectrum (SS), with direct se-quence modulation. The primarymilitary o b j e c t i v e o f theexperiment is to significantlyenhance the education of militaryofficers in the NPS’s SpaceSystems Curricula by the design,fabrication, testing and operationof a low-cost, low earth orbit(LEO), digital communicationssatellite.

The ac tual PANSATsignal design is still being studiedand the ground station modemdesign is the subject of thef o l l o w i n g N P S graduatestudents’ Masters Theses. Thepreliminary thesis titles are:

“BPSK SS Modem Design,” Lt.T. Fritz, USN

“QPSK SS Modem Design,” Lt.J. Mckinstry, USN

“4-FSK SS Modem Design,” Lt.T. Murray, USN

“DPSK SS Modem Design,” Lt.L. Rasnick, USN

The project is currently inthe phases of design, analysis,parameter optimization andcircuit breadboarding. A designreview is being held this Octoberto finalize many of the designdetails of all the satellite’s sub-systems.

PANSAT DESIGNDETAILS

The major spacecraft sub-systems include: (1) Structureand Configuration; (2) PowerControl Unit; (3) Spacecraft

C o n t r o l Un i t ; a n d ( 4 )Communication Control Unit.

The CommunicationControl Unit (CCU) will utilizean 8088 micro-processor withwatchdog timer and up-loadedcommands. The communicationspayload will comply with currentFCC part 97 rules for amateurradio SS operations and willprovide 10e5 bit error rates andwill utilize redundant (mutuallyexclusive) SS modems. Thet e l eme t ry and command(TLMITCM) functions will use asimple non-spread BPSK modem(NPS use &).

The PANSAT solar cellpower generation subsystem willoutput 24.5 Watts, minimumaverage electrical power at End

Octpher, 1992 Spread Spectrum Scene-

11

of Life, even with a tumblingspacecraft. The onboard powerbus will be an unregulated 12Volts.

Mm& FliiN

k.&#& : i... .--....‘--., !

~~,~‘~--

‘krq’wi I . . . . . . . . . . . . . . . . ._ L-i ., ,___,,_,,bw

The physical satellite will be a26-sided polyhedron with 17solar sub-panels. The PANSATis sized for a Space Shuttle GASlaunch, but is amenable tol a u n c h e s o n v e h i c l e s o fopportunity including: Pegasus,Delta II, Titan II and Scout.

A 5 Watt (+7 dBW)satellite transmitter with a 0 dBisatellite antenna is envisioned.The ground based user stationcan use a 1 Watt (0 dBW)transmitter with a l/4 waveground plane (+2.15 dBi) or a l/2

crossed turnstilearray would be even better.

The satellite antenna willbe a four element TangentialT u r n s t i l e w i t h c i r c u l a rpolarization. The patternsimulations so far suggest nullsless than 3 dB deep andpolarization losses less than 3 dBare possible. The 26 sidedPANSAT will continually tumble,and any more complicatedantenna than this 4 elementcircular one is just not feasible ina low cost satellite.

PANSAT GroundSegment

The figure above showsthe preliminary concept for NPS?ground control faci l i ty foiPANSAT. It is not mucldifferent than a well equippecamateur radio OSCAR 13 stationNPS will however, be the onl\ground control station foiPANSAT.

During this author’s visito NPS, Lt. Russell GottfiiedUSN, presented the preliminar:resul ts o f h i s n e t w o r lcommunication model simulationand described the results of aiinformal amateur radio poll take]by the local radio club. Severakey parameters of PANSAT’d e s i g n ( i n c l u d i n g onboar

memory size, average messagelength and processor loading) area function of the- user trafficloading. Further, the very natureof an SS demod-remod satellitecommunication payload makescertain limits as to how manyusers can access the bird duringany given pass. In fact, the birdmay appear to be blocked fortraffic if too many ground userstry to access PANSAT at onetime. Since only a single user“connects” to PANSAT at a time,queuing delays of up to severalseconds are possible with evenmoderate traffic loadings. Lt.Gottfi-ied did an excellent job ofmodelling and simulating all ofthese complex interactions. Hismain findings are that messagesshould be a maximum of 2 to 4K i l o b y t e s l o n g a n d t h a tsimultaneous use by a world wideham community of 200 to 400users should provide optimumPANSAT performance.

Where will 200 to 400space equipped, spread spectrummodem equipped hams comefrom? Inherent in the researchinto modem techniques beingdone at NPS is the desire todesign a simple, low cost grounduser modem/RF package. Thisauthor will f o l l ow the sedevelopments and report on themin the future. NPS wants torelease to the public domain, allinformation necessary to buildground user station modem/RFe q u i p m e n t f o r PANSAT.Perhaps SSS or another magazinewill publish these details andeven offer a kit for potentialPANSAT enthusiasts. In themean time, start thinking aboutspread spectrum, learn somethingabout it and get ready to enterHam Radio’s 21st century.


FrrePort oprrutrs Contact: Windata, Inc.. IO.searnlrssl~ Bearfoot Rd., Northborough.u?th existing MA 01532, (508) 393-3330:wired LA Ns. fax (508) 393-3694.

F irLAN, a IO-Mbps wire-less Ethernet LAN, con-

sists of two tranceivers and ahub. Able to run transparent11across all Ethernet platformsand media, FirLAN allowsseamless integration into ex-isting wired LANs.

The ET300 transceiver hasa 90-foot range and supportsup to 29 Ethernet connec-tions. An extended-rangeET330 transceiver has arange of 300 feet. The EH360hub covers a 90-degree areaand supports as many as 250transceivers.

Communications, Inc.,102-21 Antares Dr., Ottawa,Ontario. Canada K2E 7T8.(613) 723-l 103; fax (613)723-6895.

pP vs. DSP - What’s the Difference?The main distinction between a general purposemicroprocessor (Intel 808x, 80x86, Motorola 68Oxx,etc.) and a DSP engine Q’I TMS32Oxx, MotorolaDSPSGOxx, AT&T DSP-xx, etc.) is replacing themicroprocessor’s built-in software routines(microcode) for numerical calculations withdedicated hardware circuits. This makesmath computations much faster.

For example, a microcoded multiply requires 12-41clock cycles in an Intel 80386. A TMS32025performs the same multiply in l-cycle, thanks to abuilt-in modified Booth algorithmmultiplier circuit.

DSP chips add extra buses, registers, and othergoodies to further speed up the math. Theperformance improvements are so dramatic thatmicroprocessor manufacturers are making theirnew pP chips more and more “DSP-like”.

DSP chip makers, pressured by more sophisticatedapplications and the demand for high-levellanguage support, are making their devices moreand more “VP-like” by including flexible memoryaddressing and management, once thedomain of the pPs.

The development of faster pipelined RISCmicroprocessors on the one hand, andsophisticated floating-point DSP chips on the otherhand, means that the distinction between PP andDSP will grow smaller as time goes on.

(If all of this confuses you, read the tutorial on DSPEngine Architectures, scheduledfor our next issue.)

The TMS32OE15 at a Glance.

The Texas Instruments TMS32OE15 is an EPROM(electrically-programmable read-only memory)version of the f&-generation ‘INS32010 DSPengine. The part was introduced in 1987.

Arithmetic: Fixed-point.

Precision: 13/ 18bit multiply (MPYK/ MPY),for .012%/ .0015%32-bit accumulator.

accuracy.

Speed: 160-nS instruction cycle (6.25 MIPS)with 25MHz clock (TMS320E1525).

Memory: 4096 16-bits on-chipprogram &ROM, 256 x 16-bits on-chip data RAM, and 4096 x 16-bitsexternal program or data memory.

Price: $120/ $86.10 (US) (qty l/ 1000)for 25-MHz version.

(Thanks to HamPute, May/June 1992)

October. lYY2 Spread Spectrum Scent IS

DSP TutorialModulation, Fourier Series, and Sampling Theory

Some Ground Rules.

One of the horrors of using a newtachnology is that you have to learnsome new theory to know what

u’re doing.go ” form

DSP theoby in its

szlicated. ‘I%e dzvation of ?%involves the kind of heavy-dutymathematics that the “Einsteins” ofthis world really enjoy.

Good for them, bad for us.

As it turns out, much of the theorybehind what we do in amateurcad:involves some heavyFortunately, most of it can be boileddown to some fairly simple “rules ofthumb” that make the technologyuseful to us. We didn’t have tostudy electron ballistics in order touse vacuum tubes. Understandiigthe physics of hole and electronmigration, a n d t h e Ebers-Mollequations, wasn’t really necessary tobias a transistor. This theory hasbeen “canned” into forms that wecan use. So it is with DSP.

Of course, by streamlining thetheory you lose the ability torigorously prove things as you

resent them. Many items have toL accepted on “faith”. While thisdrives engineers andmathematicians crazy, most hamshaveLleyoeA&iive ;zsh i t . The

R done awonderful job of streamlining theoryfor amateur use. We hope to followin the same spirit.

DSP lvpl&xs traditionalcomponents by mathematicalequati.ons. As a result, the mathunderpinning of DSP can’t becompletely eliminated

We will assume that you canremember a little of your high-

school algebra and trigonometry.We will jog your memory as we go.Beyond that, we will spare you therigorous mathematical proofs.Hams learn by doing. We want youto be “doing” DSP as quickly aspossible.

Let’s begin . . .

The Rod to Sampling Theory.

Sampling theory is what links theanalog and digital worlds. We couldjust present it in “here it is” fashion,but we won’t. The concepts behindit are too important to gloss over.Instead we will build up to it, step-by-step, so that you will have a“feel” for it. So that you can useDSP, not just parrot the buzzwords.

If You Can’t Handle It, Map It!

Mathematics is a tool that we use torepresent the world around us. It isa language. A way to look at things.If a given mathematical“environment”, or domain becomesawkward when dealing with aphysical reality, we are perfectlyfree to re-defme the domain andlook at the reality through differenteyes. We transform the problemfrom one domain to the other, ormap it. Mathematicians have beendoing this for centuries.

As amateurs, we use an oscilloscopeto look at signals in the timedomain. The ‘scope plots signalamplitude versus time.Unfortunately, things like harmonicdistortion and sideband splatter arehard to see on an oscilloscope.Instead, we use a spectrum analyzerto look at the signals in thefrequency domain. Were lookingat the same signal, only throughdifferent eyes (gee Figure I).

We all learned mapping, in @ghscnooi. rsememoerthe quadraticequation? It isshown as equationpair (1) and (2).Everything was

until bz-4acGy-e less thanzero! How do youtake the squareroot of a neeative

(al Ttnr Donaln tbJ Frequency Donaln number? You map

Figure 1. Two-tone SSB with 5% Croeeoverthe problem to a

new domain, of course!

‘

a.x +b.x+c =O (1)

x = (2)2.a

Imagine if there were a square rootof -1. We know it doesn’t reallyexist, but just imagine. Let’s defmethis imaginary number i as thes uare root of -1, whatever that is.4e could then map all possiblesolutions to the quadratic equationin this new domain of real andimaginary numbers. Put the realnumbers along a horizontal axis,and imaginary along a vertical axis,

lnaq T

r*cos 8 real

Figure 2 Coordinate Mapping

and you have rectangular orCartesian awrdinatee, ~LJ shownin Figure 2. Every number can berepresented as a real part and animaginary part:

c = (real part) + i~(imagim3r-y part)

These numbers are also calledcomplex numbem. “i” is called theimaginary operator. Since we useI, -1,t to represent current inelectronics, we can use ‘7’ as theimaginary operator to avoidconfusion.

We found out in high school thatdealing with rectan a rcoordinates got haii when X”loingthings like multiplying and dividing.Solution: map it again!


We know that multiplying anddividing numbers with exponents iseasy (you just add or subtract theexponents). 18th century Swissmathematician L,eonhard E u l e rproved that :

j@a = cos(6) l j.sin(B) (3)

This equation is known as Euler’8identity. If we draw a line from theorigin (0,O) to our number, we canrepresent it as a ray “r”, or distancefrom the origin, and an angle withthe horizontal x-axis “0”. We havepolar coordinatee. Thetrigonome

9.involved is shown in

Figure 2. e 1 1s the polar operator.

Polar representation is widely usedin signal processing. “r” is theamplitude of the signal, while “0” isthe phase at any instant in time. Incase you forgot? “e” is the base of thenatural loganthmic system, andequals 2.718281828... Rememberthat the different domains representthe same thing (the signal). This isall just mathematical smoke andmirrors!

Negative Frequencies?

In Euler’s identity, 8 represents afxed phase angle. Each completerotation, or cycle, represents 360-degrees, or 2.7~ radians. Torepresent a continuous signal, weexpress it in the form:

jotA.0 ,whexe (0

A = the amplitude of the signal0 = 2.x.f= the frequency in radians/second

t=time

Okay, so far so good. But look whathappens when we solve Euler’sequation for cos(wt), which is asimple sine wave:

jot -jote + e

coe(ot) = - (5)2

,whexe

-jot= cos(ot)

Eluation ( 6 )- j .sin(ot) (6)

is the c o m p l e xconjugate of ejat. It represents anegative frequency! What on earthis a negative frequency?really knows.

NobodyIt’s just like an

imaginary number. Mathematical

smoke and mirrors. But it makesthe math work.

Let’s illustrate by looking atsomething more familiar:

Mixing and Simple Modulation.

How do you build a mixer in DSP?Just multiply! No tomids, no hotcarrier diodes, no FETs, simplymultiply the signal by a carrier, andyou’re done.

Let’s follow this one case through toillustrate how negative frequencieswork in the math. Remember fromalgebra that you derived things bysubstituting, expanding all of thet=ms, collecting terms, andsubstituting back:

cospc -t].cos[os .t] =

j.0 .t -3.0 .tC C

e’ + e.

2

j.io .t -j.o -ta s

e + ex

j[oc +os ]l -j[oc *asIt

e + e

j[oc -tie 1: -‘[WC -us ,:

e +e=

co*[pc + @s !I,t OJSB)

coqu; - OS ,t;

(7)

(LSB)

2Equation (7) is the modulationequation. It states,mathematically, that a perfect mixerproduces the sum and differencefrequencies of the incoming signals.The negative frequency componentsgenerate the difference frequencies.

Our DSP mixer is a double-sideband@SB) modulator. The signal

Figure 3. AM Generation

frequency w, is the baseband audiocomponent.

A DSP multiply is equivalent to aperfect “double-balanced” mixer. Noinput components appear at theoutput!

How do you make AM? Add DC. Asshown in Figure 3 and equation (8),the result is ideal lOO?h modulation.This is what Kenwood does in theTS950SD.

[’ t C08b8 .t]].cos[Oc ‘“1 =

COB 0 .t +c 1 (CABBIER)C

cos[pc + OS 1’3 t (USB)

.os[pc 2- OS It]

(LSB) (8)2

We’ll cover other modulationtechniques in a future tutorial. Fornow, let’s get back on the road tosampling theory.

Linear Sy8tems: Superpoeition

So far, we’ve demonstrated how toprocess a carrier and a sinusoidalsignal component. But actualsignals contain many components.Doesn’t this complicate things? Notif the system is linear.

A eyetern is any combination ofsignal processing functions. It caninclude mixers, filters, detectors,what have you. Linear eystems, bydefinition, adhere to the principle ofsuperposition.

Superpos i t ion means that ifsignals are added together at theinput, the output will be the sum ofthe system’s response to each of theinput signals. We can superimposethe individual responses. There isno interaction among the signals.

IX Spread Spectrum Sccnc Octchcr, lYY2

A ham receiver, for example, issupposed to be a linear system. Itsinput consists of signals throughoutthe radio spectrum. It’s outputresponse to a particular signalshould not be distorted by othersignals in the spectrum. There is nointeraction (intermod) among theincoming signals. When a receiverstage overloads, intermod occursand the system is, by definition,non-linear.

Mixers, filters, phase-shiftnetworks, and amplifiers are alllinear, so long as there is no clippingor compression. Detectors may ormay not be linear. A diode, forexample, is not linear. You may beable to restore linearity, however, ifyou can Nter out the intermodcomponents.

Luckily, most DSP involves linearcircuits. Superposition applies. Wecan treat each component of a signalindependently of all the othercomponents, and then add theindividual results to g;;v;;; totaloutput response. andconquer. Now if we only knew howto process waveforms other thansinusoids . . .

The Fourier Series.

Jean Baptiste Joseph Fourier to therescue! This 18th century Frenchmathematician devoted much of hisca.reer to mathematically linking thetime and frequency domains. Heproved that any periodic, orrepeating, waveform is equivalent toa sum of sines and cosines:

ar(t) - " +a>: aP

P

The z symbol in equation (9) is justa shorthand notation for adding alarge number of terms together.We’ll elaborate on this notation infuture tutorials. The I symbolrepresentscalculus).

integration (from OYou don’t have to learn

calculus to use the Fourier series,since the series for a wide variety ofwaveforms have been tabulated in anumber of references. In theequations, P represents the periodof the waveform, and t is time.

waveform. You may also getterm, depending on symmetry.

Figure 4 illustrates how a periodic

2nd HARM

I I I

I I I 14th HARM

t

- L-._-_l

t = w1 I L 1

Figure 4. Constructing aPeriodic Pulse Train from

Sinu8oid8

waveform can be constructed byadding sinusoidal components. Thisshould give you an intuitive “feel”for what’s going on.

Th.e fbdamental freqzdency (I%rst”harmonic) of the sequence is alwaysthe inverse of the waveform period(Z/P). Which harmonics you getdepends on the symmetry of the

Figure 5. Unit Pul8ee(Power = 1.0)

a DC

Let’s look more closely at one specialcase: a “unit” pulse train (Figure S),with period P, width w, andarea/period (power) = 1. TheFourier series for this waveform is:

l[nwsin -

L + a.c

(-1?I IP axnt

‘COB -ILnW [ IP

n-

P

P,if amplitude = -

x(10)

Remember that (-1)” equals +l if nis even, and -1 if n is odd. Thespectrum “envelope” follows theform sin(X)& where X is [xnw/Pin this case. This envelope isimportant whenever “wide” pulsesare used in DSP. The frequencyrolloff must be corrected in this case.Sin(xXy(xX) is plotted in Figure 6.

As the pulse width gets narrower,the sin(Xfl envelope broadens. Theextreme case is when the pulsewidth is “zero”. Periodic zero-widthpulses are called an impulee train.With the help of some liiit theory(tpyVeyrry about it) equation (10)

e

i + a.c

zlmt(&as -

[ I(11)

P

but equatyon (5) tells us that thecos(Zxn4/P) equals a positive andnegative frequency component, eachof amplitude l/2. Result:

-, -4 -, -* -0 * , , J 4 5

Figure 6. Sin(lrX)/(lrX)


A periodic impulse tmin contains allharmonic components, including aLX term, and has a “flat” spectmmenvelope: all components have thesame ampliZ&.

Figure 7 shows this graphically.

. ..?...-., -,I -*I -r DC , *r ,r 4r

Figure 7. Unit Impul8e !lkainSpectrum

Recap.

L,&s~~all where we have traveled. .

aWe reviewed the concept ofmapping and domains, andtransforming problems from onedomain to another using operators.

*We looked at rectangular andpolar coordinates as ways to mapsignals.

l Euler’s identity introduces theconcept of negative frequencies.

*We derived the modulationequation as a practical applicationof Euler’s approach. This is thefirst trigonometric identity in our“bag of tricks”.

l We learned that we can work withcomplex signals one frequencycomponent at a time, thanks tosuperposition.

l The Fourier equations allow us torepresent any periodic signal as asum of sinusoids.

l One very special case is theperiodic impulse train.

l All of the above applies equallywell to analog or digital signals.

We told you all of that so that wecan tell you this . . .

Sampling Theory

An analog signal is continuous:there e an infinite number ofpoints on the waveform, each havinganue~t.e number of possible

Digital cumputzers can onlyhandle discrete signal8 (finite

number of points, or eCample8), Withquuntized level.8 (fmite number ofpossible values). How do we bridgethis gap?

Sampling theory was developed inthe 1930s and 1940s to determinewhat happens when you repmsent acontinuous signal by a finite numberof samples. We’ll deal with theeffecta of quantizing the signallevels in a future tutorial.

Etiyrn:Et p r o v e d tt&;f pam@ Is

sampled at a periodic mte that is atleast tw ice the highest fmpencycontainedin the signal, the analogs&n& can be perfectly n3xmstructedby an ideal loWpasS filter. This i8

known a8 the sampling theorem,and the minimum sample rate iscalled the Nyquist rate.

The sampling theorem says that asequence of numbers can contain allof the information we need toprocess and reconstruct the signal.The infinite number of pointsbetween the sampling points containredundant information!

Sampling means that we extract thesignal’s value at discrete points intime (Figure 8). Thio i8 exactly thesame a8 multiplying the eignalby a periodic impulse train. Weknow the spectrum of a periodicimpulse train, we know themodulation equation, and we knowabout superposition. Therefore:

fkquency components withinthe eignal to all harmonic8 ofthe sampling rate (includinga DC term).

Figure 9 illustrates the samplingtheorem. The image components,created by mixing the basebandsignal and the impulse train, arecalled alia8ing by DSP engineers.Signal components are aliused to

Figure 8. Sampling

. . .. . .- f, dC f,

Figure 9. Nypuiat Sampling

other frequencies (just like acriminal aliases hi name when heskips town). Aliasing is the same aaimages in your receiver, or the slow-motion effect obtained by viewinghigh-speed action with a strobelight.

The Nyquist theorem requires thesampling rate to be high enough toavoid overlap of the aliased spectra.Then you can reconstruct the signalwith a lowpass filter.

For ham applications, the Nyquistrate is sometimes overkill. Figure 10shows what happens if you samplenarmwband s ignals at a lowsampling rate. The spectra areinterleaved, but they might not

. . . .

- f, L DC-_I

f ,BASEBAND

Figure 10. NarrvwbandSampliw

overlap. You could then reconstructthe baseband signal with suitablebandpass filtere. This is whatKenwood does in the TS-950SD.

One word of caution: signal“bandwidth” here is considered to bethe width between the stopband(typically -40 dE5 to -80 d_B) pointa,J& the passband (-6 d.B) width! Asyou work with DSP, you are free tochange the sampling rate as long a8you do not allow the stopband pointsto overlap. There are advantages indoing this . . .

-NW-%.

RCfCEllCCl

A. Oppenhei A. Wiikk’y, Sigmb andsystems.Prentice-Hall, 1983.

A. Oppenheim, R. Schafer, Dighal SigmlProcessing, Prentice-Hall. 1975.

B. Gold, C. Rader. Digilal Processing o/Sigmk,McGraw-Hill, 1969.

B.P. hthi, Communicahon Sysums,Wiley, 1968.

E. Craig, LaPlace and Fourier Tmnrforms forElcclrical Engineers, Holt. Rinehart andWinston, 1966.

1990ARRL Handbook, Chapter 18.Arnedatn Radio Relay League.

Sfandard Ma~hcwdcal Tables,chemical Rubber Company.


IWP from pge s - MHz if we were tostick with the NTSC scheme.This is clearly unacceptable, somuch effort has been expendedin trying to reduce the bandwidthrequirement, with some success.

These efforts have yieldedtwo basic classes of schemes:analog, such as Japan’s MUSE,and digtal, such as the four USproposals among the fiveschemes currently under FCCconsideration for the US HDTVstandard. The fifth is a narrowbandwidth version of MUSE(6MHz). It would beembarassing for the USelectronics industry if MUSEwins the competition! Thedigital schemes are Zenith &AT&T’s DSC-HDTV, GeneralInstruments DigiCipher, Sarnoff sADTV, ATVA and an MIT

proposal.This column will

discuss only the MUSE system,since it allows me to develop theterminology needed to describethe all-digital schemes, and thereis already a family of DSPprocessor chips tai loredspecifically for MUSE on themarket. I should then be in agood position to explain why theFCC has taken a lot of flak fromsome very professional engineersfor hastening the entire approvalprocess with significant bias forall digital schemes.

Next month I will presenta block diagram of the processesdiscussed here. I will also givemore details about today’s digitalHDTV schemes and how they fitinto continuing political/FCCapproval battleground.


Education from page 4 - b e t w e e ntechnicians and engineers reallychanged all that much in thepast three decades?

I recently had theopportunity to review a newbook on this subject which isdue to be released in November.E n t i t l e d B e c o m i n g A nElectronics Technician:SecurinP Your High TechFuture by Ronald Reis, this newbook describes the electronicsindustry as it exists today. Itexplores the background andhistory of the electronics field,current avenues or research anddevelopment and the trends thatwill shape the future of theindustry. Of particular interestis a chapter which describes thetechnician in the workplace ofthe nineties with respect to thee l e c t r o n i c s e n g i n e e r ,technologist and assembler. Itshows students how the careerthey’ve chosen fits into theoverall workplace picture andwhat type of relationships theycan expect to find once they’reon the job. This book gives aclear, concise look at theelectronics industry, how toprepare for that first job as at e c h n i c i a n , a n d mos timportantly, how to continuetechnical growth to stay abreasto f n e w t r e n d s a n ddevelopments. The book isbeing published by MerrillPublishing, of MacmillanPublishing, New York.

Next month I willexplore current trends ineducation across the countrywhich address the incorporationof technical preparation coursesfor all students. Stay tuned!

October, 1992

For information or to order 7Call (404) 939-0256F A X (404) 939-0157

IYISA

Or write:Eaaleware Coworat ion ’ ”Li!ilml

‘̂

1750 Mounta& GlenStone Mountain, GA 30087USA

Programs rut7 on stnndard IBM rmrf:ompatible PCs, ir7clude a 30 day satisfnctior7

;zzarnn~ee. fire szcpport find 171~ ~rrrnzral fees.mmedznte slzipment available.

‘. . 1hne prices internntionnlly. Direct fnctoryiales ntzd zzser szrpport by phorre or FAX.

Spread Spectrum Scene 29

New! =MATCH=Suite of matching network solutions

&Narrow or broadband. comulex loads

=OSCILLATOR=Unified oscillator design process:

YILC, distributed, SAW, & crystal designs~Estimates SSl3, residual M & PM noise,*$m- only, W%, with .$+@a~=

=FILTER=Complete L-C filter synthesis package:

=TLINE=Relate physical & electrical parameters:

~Microstrip, stripline, coplanar & coaxti Accurate algonthms with referencesthesis & anal sis

,i$%- oaly !§%Jwith =SuperStar=

=SuperStar=RF/Microwave circuit simulation, tuning,optimization and statistical analysis:

Who is RF/Spread Spectrum?

SSS is published by RF/Spread Spectrum(RF/SS for short). RF&S is an independentconsulting and product development firmwith a staff of several highly experiencedengineers, programmers, systems designers,strategic marketeers and project managers.RF/SS is poised to help the evolving PCS andSpread Spectrum industries define, developand market profitable new products. CallRandy Roberts at (415) 726-6235 or send us aFAX at (415) 726-0118 to find out more aboutour capabilities, experience, staff specialtiesand availability. We’re here to help you!

RF/Spread SpectrumP.0 Box 2199


Conclusiun: Conclusion ofDan Dobcrstcin’s Idea Articleon a General Purpose SSR a d i o (GPSSR) f o rcommercial, industrial or hamU S C, designed around the 49MHz cordless telephone chips.

Intcrcstin~: More on DSP,“Secret” Signals, The Aerialand Technical Education inour columns.

Tutorial: More hands-ontutorial materaii in TechnicalTricks and the feature articleon low power 915 MHz links.

News: Latest news on SpreadSpectrum regulatory, newproduct and marketingdevelopments.

gj jge&gy/

L&)&i~&qg~ &&yJJ$:,

An RF/SS PublicationP. 0. Box 2199


FORWARDING Sr ADDRESSCORRECTION REQUESTED

BULK RATEU.S. POSTAGE

PERMIT NO. 17SANTA CRUZ, CA

special issue on new spread spectrum, lan and pcn productssss-mag.com/pdf/sssv1n7.pdf · nthly news...

Documents