Signals and Systems Project Assignment (200 Points) : Name
Signals and Systems Project Assignment (200 Points) : Name
Signals and Systems Project Assignment (200 Points) : Name
SignalsandSystemsProjectAssignment(200Points)
Due10:59PM,Wednesday,November29th,2017
Rulesforassignments:alltheworkshouldbewordprocessed.Turninyourmfileyouusedand
Multisimfilealongwithresultsandfigures.Yourcodesandresultsshouldbeneatly(legibly)presented
forfullcredit.IfIcannotfollowthroughwhatyouwrote,youwillnotgetanypointsatallforthose
portionsIcannotfollow.
1. YouwillbedesigninganLTIsystem,especially,adeviceutilizingthefrequencyselectivityofan
LTIsystem,calledfilter.Thedeviceyouaredesigningisavitalpartoftheoverallsystemwhichit
willgetintegratedintoasasubsystem(antialiasingfilterforanalogtodigitalconverter).
AnalogButterworthlowpassfilterhasbeenchosen;thefollowingisaspecificationofthefilter
forthegivensystem(8bitADCwith50kHzsamplingrate)towhichitwillbeconnected:
vIN(t) H(s) vout(t)
fc(3dBcutofffrequency)=7.8kHz
fs(stopbandfrequency)=25kHz
As(stopbandattenuation,indB)=50dB
1) FindoutandlisttheadvantagesanddisadvantagesofaButterworthfilter.
2) DeterminetheorderoftheButterworthfilter,N,usingthefollowingformula:
10( AS /10 ) - 1
log10
0.99526
N = round 0.5 +
2 log10 ( fs /fc )
whereNistheorderofButterworthfilterandroundisafunctionroundinganumbertothe
nearestintegernumber.
3) UsingMATLABfunctionbuttapalongwiththeorderofthefilterNyoucalculatedin2),
youcanfindthetransferfunctionofprototypeButterworthfilterHprototype(s)andthe
transferfunctionH(s)whichsatisfiesthespecifications.
a. Afteryoudeterminethelocationsofpolesandzerosalongwithgainusing
buttapfunction,UseMATLAB[Num,Den]=zp2tffunctionandfindthe
coefficientsofthenumeratoranddenominatorofHprototype(s).UsingMATLAB
SYS=tf(Num,Den)function,nowyoucandetermineHprototype(s).Includethis
prototypetransferfunctionyoufoundinyourreport.Also,includeaplotshowing
thepolesandzerosofHprototype(s)(calledpolezeroplot)inyourreportusing
MATLABfunctionzplane.
b. TheprototypeButterworthfilteristheNthorderfilterwith3dBcutofffrequency
1rad/sec.Therelationshipbetweentheprototypefiltertransferfunction
Hprototype(s)andtheH(s)whichsatisfiesthespecificationsintheboxforButterworth
Lowpassfilteristhefollowing:letssaythecutofffrequencyofH(s)is (3dB
cutoffforButterworth),anditsorderis3,then
b b
H prototype ( s ) = H (s ) =
3 2 3
s + a1s + a2s + a 3 s s 2 s
+ a1 + a2 + a 3
w wc wc
c
IncludeH(s)youfoundinyourreportalongwithitspolezeroplot.
c. PlotthefrequencyresponsesofbothHprototype(s)andH(s)usingMATLABfunction
bode.VerifythefrequencyresponseofH(s)satisfiesthegivenspecification.
Includetheplots(frequencyresponseplots)inyourreport.
4) Youaredesigningananaloglowpassfilter,especially,anactivefilter.Higherorder(3rdorder
orhigher)activeanalogfiltersaretypicallyimplementedbycascadingsecondorderfilters(if
theorderisevennumber)orsecondorderfiltersandfirstorderfilter(iftheorderisodd)
duetoreducedsensitivityandmodularity.Whenyouconnectthesecondorderfilters
(calledsecondorderstages/sections),youneedtoarrangethesectionswiththelowestQ
(Q:QualityFactor)neartheinputandthehighestQneartheoutputtohelpavoid
saturation.
Figure1:BuildingEvenOrderFiltersbyCascadingSecondOrderStages
Figure2:BuildingOddOrderFiltersbyCascadingSecondOrderStages
5) Now,letsderivethesecondorderstagesforthedesignedtransferfunctionH(s).The
secondorderstagescanbederivedbypairingupcomplexconjugatepolepairsfromyour
designedtransferfunction.StartingwithMATLABfunction[Z,P,K]=buttap(N)you
usedin3),youcanpairupcomplexconjugatepolepairsfromtheprototypetransfer
functionHprototype(s).LetsdenotetheNthstagesecondorderfilterofHprototype(s)byHPN(s).
Then,youcanconvertHPN(s)totheNthsecondorderstageofyourH(s),thetransfer
functionsatisfyingthegivenspecificationsinthebox,andletsdenotethatbyHN(s).The
relationshipbetweenHPN(s)andHN(s)forlowpassButterworthfilterisasfollows;notethat
theprototypefilterhasgain1:
1 1 wn2
H PN ( s ) = HN (s ) = =
s 2 + 2zns + 1 s 2
s 2 + 2zn wns + wn2
+ 2z s + 1
w n
n wn
1
where zn is the damping factor, Qn = , and wn = FSFn wc
2zn
Intheaboveequation,Qnisthequalityfactorofthenthstage,FSFniscalledFrequencyScaling
Factorofthenthstage,whichis1forButterworthfilter(notethatitisNOT1forothertypesof
filters),and istheoverallNthorderfilterscutofffrequency(forButterworthfilter,cutoff
frequencyis3dBcutofffrequency;forothertypesoffilters,cutofffrequenciesaredefined
differently).Notethat FSFn = an2 + bn2 where an and bn aretherealandimaginarypart,
respectively,ofthepoleoftheNthstageHPN(s)oftheprototypefilterHprotoype(s)(since 2zn isreal,
thepolesofHPN(s)arecomplexconjugatepair: and ).Include
HPN(s)andHN(s)youfoundforallthestagesinyourreport.Also,usingbodefunctionin
MATLAB,plotthebodeplotsofallHPN(s)andHN(s),andincludetheminyourreport.
Figure3:PolezeroPlotforHprototype(s)
2. SofaryoufoundthetransferfunctionH(s)andcorrespondingtransferfunctionofsecondorder
filtersforeachstageHN(s);hence,youcometoastagewhereyouneedtoimplementthe
/
transferfunctionH(s)bycascadingHN(s)( forevenorderand
/
_ foroddorder).Foractivefilters,althoughdifferenttopologies
areavailable,unitgainSallenKeyarchitecturehasbeenchosen(youcanaddamplifyingopamp
attheendofthefilterifyouneedgain).ThefollowingistheSallenKeylowpassarchitecture.
DeriveitstransferfunctionHSK(s)intermsofs,C1,C2,R1,andR2andincludeitinyourreport.
Figure4:SallenKeyLowPassFilterArchitecture
3. NowmapeachsecondordersectionHN(s)toHSK(s)todeterminetheresistorandcapacitor
values.BycomparingHSK(s)andHN(s),youcanderiveequationsfordeterminingtheresistorand
capacitorvalues.Thefollowingarethoseequations;derivetheequationsEQ(1)andEQ(2),
andincludethederivationinyourreport:
2
2znC 2 ( 2zn ) C 22 - 4C 1C 2
R1,2 = EQ(1)
2wnC 1C 2
4C 1
C2 EQ(2)
2
( 2zn )
where wn =FSFn wc with wc being the overall system cut-off frequency
1 1
HN (s ) = H PN ( s ) =
2 2
s s s + 2zns + 1
+ 2zn + 1
w w
n n
4. IftheorderofthetransferfunctionH(s)isodd,youneedafirstorderactivelowpassfilter
whichappearsinthefigure2inpage2.Deriveitstransferfunctionalongwithits3dBcutoff
frequency(thecutofffrequencyisexpressedintermsofRandC);letsdenotethefirstorder
transferfunctionbyHreal_pole(s),andincludeitinyourreport(dothissubproblemiftheorder
oftheoveralldesignedfilterisodd;iftheorderoftheoveralldesignedfilteriseven,youcan
skipthispart).
5. Finally,usingtheequation(1)and(2)above,youcandeterminetheresistorandcapacitor
valuesofthesecondorderfilterHN(s)ofeachstage(section):first,pickthevalueforC1and
determineC2accordinglysatisfyingEQ(2).PickthevaluesforC1andC2whichareavailablein
thestandardvaluesofcapacitorswith5%tolerance.Afteryoudeterminethevaluesof
capacitors,determinethevaluesofresistorsusingEQ(1).Also,picktheresistorvaluesfrom
availablestandardvalueswith1%tolerance.Now,however,thechoiceofC1andC2satisfying
EQ.(1)isinfinite,andmanydifferentwaysofdeterminingC1andC2areavailable.So,letstry
onealgorithmwithwhichwecanchoosevaluesiteratively.First,youpickaresistorvalue
accordingtothefollowingguidelines(pickstandardvalueswith1%tolerance):
IfwesetR1=R2inEQ.(1),usingEQ.(1)andEQ.(2),wecanderivethefollowingsimplified
algorithmforcalculatingR1,R2,C1,andC2:
Figure5:AprocedurecalculatingR1,R2,C1,andC2values
Whenyoupickresistorsandcapacitorsforfilterdesigninpractice,alsokeepinmindthe
followinggeneralguidelines:
Capacitors Resistors
Avoidvalueslessthan10pF Valuesintherangeofafew
UseNPOorC0Gdielectrics hundredohmstoafew
Use1or5%tolerance thousandohmsarebest
components Usemetalfilmwithlow
Surfacemountispreferred temperaturecoefficients
Use1%tolerancecomponents
Surfacemountispreferred
DetermineR1,R2,C1,andC2values(use1%tolerancestandardvalues(E96)forresistorsand5%
tolerancestandardvalues(E24)forcapacitors)foreachstageHN(s)ofthefilterH(s)usingthe
procedureinFigure5.Letsgoforprecisionapplication;then,theexponentforresistorsR1and
R2canbe3.
6. BuildtheoverallfilterH(s)inMultisimsimulationenvironmentusingthecapacitorandresistor
valuesyoufoundin5usingtheprocedureinFigure5.Use1%toleranceforresistorsand5%for
capacitorswhenyoubuildandrunsimulation.Makesureyouarrangethesecondorderfilters
withthelowestQneartheinputandthehighestQneartheoutputtohelpavoidsaturation.
DofrequencysweepanalysisusingMultisim(simulatefromDCto10MHz)andverifythe
frequencyresponseofthefilter.Checkwhetherornotthefilteryoudesignedsatisfythe
specification.Ifyourdesigndoesnotmeetthespecification,increasethefilterorderand
redesign.IncludetheMultisimcircuitschematicsanditsACanalysis(FrequencySweep
Analysis)results(plot)inyourreportalongwithyourconclusionwhetherornotthefilter
youdesignedmeetsthespecification.
7. Now,letsuseatoolwhichcangeneratethevaluesfortheresistorsandcapacitorsofthefilter
youaredesigning.ItiscalledWEBENCHandyoucanaccessitatwww.ti.com/WEBENCH.You
canregisterforfreetousethetool.SelectButterworthwithSallenKeyarchitecturewithE96
standardvaluesforresistorsandE24forcapacitors.Includethegeneratedresultbythetool
(schematicsforeachstagealongwithBOM(BillofMaterial))inyourreport.Comparethetool
generatedresultwithyourowndesignintermsfrequencyresponsethroughMultisim
simulation(ofcourse,youwillgetsomewhatdifferentresultforthevaluesofresistorsand
capacitorsbythetoolfromyourowndesignsincethetoolmightuseadifferentalgorithmto
determinethevaluesofRandC).
8. Answerthefollowingquestionsconcerningthedesigngeneratedbythetool:
1) YouwillseeatermGBWPshowingupinthegeneratedresult.FindoutwhatGain
BandwidthProduct(GBWP)ofanOPAMPmeans
2) RecordtherequiredGBWPforeachstageandcheckthesuggestedOPAMPmeetsthe
minimumGBWPrequirement
3) Findoutandrecordwhattolerancesareusedforthegenerateddesignforresistorsand
capacitors
4) Recordwhattypeofmaterial(technologyformakingthecomponent)usedforthe
suggestedresistorsandcapacitors