MASSACHUSETTS INSTITUTE OF TECHNOLOGY
Depart,rnent of Ele<:tricalErlgirleerir~grxrld Cnr~~pllt,er
S<:ier~(:e
Spring Term 2007
Tllerrrlal Design Prol~lern
Introduction
This 1,111 gives ~ ~ 0 t1 l1~ eopport,llnit,y to evaluat,e the (1ynarnic:s of a ternperat~lrecorltrol systerrl rxrld desigr~a
<:ornpensatorfor it. The systerrl illat yo^^ will investigi~t,eis patternecl on orle that mas (ievelo~~e(i
a t Lirlcolr~
Lat)oratory to cor~trolthe ternperat~lreof a laser rliorle illat is part of a space-t)ase(i o~~ti<:al-cornrm~r~ici~t,i(>r~
experiment.
Tlle Lirlcolr~La11 systerrl hrxs der~~nr~strrxte(i
lnr~g-terr~~
ter~lperat~lre
stal~ility\%-it,hir~
100/r0C. Your syst,ern
has a sorr~e~vl~at
lligller rloise floor, primarily t)ecallsr t,lle <:ontrol circuit,ry has 11ee11sirr~plifie(ias cornpared
~vit,htlle origirlal system.
This type of syst,ern car1 l ~ ~lserl
e
i~na r ~ yi~ppli(:r%t,i~~~
tllat requires precise t,e~r~perr%t~~re
c011trn1. 011e ~ ~ n s s i l~ilityirlvolves stal~ilizingt,lle operating ternperat,~lreof ir~ert~ial
s11c11
rrleasllrerrlerlt corr~poner~ts
as gyroscopes.
For this lab, you will neecl to check o ~ a.~proto-t)oarrl
t
desk (38-501). You frorr~t l ~ efiftl~floor eql~iprr~ent
will need your part,s 11ox frorrn 6.002. If you have lost your parts t ~ o x(or never llarl one,) rnrxlce a list of t l ~ e
parts that yo^^ rleerl after you rlesign yollr corr~prnsat,or.and ask at the desk for therr~.
There is a ~IOII-trivial
arrlo~lr~t
of work t,o t)e done for t,l~islat). Please st,art earl!:
Experimental Configuration
Irr~port.ants t r ~ ~ c t ~featnres
~ r a l of t,lle tllerrrlal syst,ern are sllo~vr~
in Figure 1.
t,lle t,e~r~perr%t~~re
c011tr01 syster~lis S~IOTVIIi11 Figure 2.
I
1
f i ~ ~ ~ < : tl~loclc
i ~ ~ ~cliagrarn
r%l
of
Ramator
Plate
Insulation
f
Thermistor '
Reslshve-film
heater Tlle ternperat~lreof the cor~trollerls~lrfaceof t,lle tllerrrlal st,n~<:t~lre
is sensed with a device \%-hoseresistarlce
is a filrlct,ior~of ternperat~lre.This device is caller1 a t,herrnist,or. This tllerrr~istorforrr~sorle leg of a 4 elerrler~t
l i e . The l~riclger11111 terq)erat,~lrehas t ~ e e nset t,o approximately 60°C. and t,l~is~rrx111eis the syst,e~r~
referer~(:et,e~r~perrxt~~re.
'4 differerltial arr~plifierc:onr~ecterlt o t,lle l~riclgeprovides a lirlearizerl o u t , p ~ ~
oft
1P7/"C of t,err~perat~lre
error. This sigrlal is availat)le rxt,ernally to enal~leyou to rrleasure the rlynarni<:s of
Cite as: James Roberge, course materials for 6.302 Feedback Systems, Spring 2007.
MIT OpenCourseWare (http://ocw.mit.edu/), Massachusetts Institute of Technology. Downloaded on [DD Month YYYY].
External
Power
Monitor Compensator
Compensator Input
External Drive
Compensator
-
"1
-
Power
Monitor
Compensator
-
0"
N
Drive
Thermistor
Bridge and
Amplifier
s
(inverting)
3
w
c
-
?
Ternperam
Error lVPC
Thermal
Dynamics
+
Output Temperature
Thermistor
Bridge and
Amplifier
--
-2 watrslvolt
-
Temperature Error 1V/oC
41
Dynamics
0
w
9
-,
Heater Power
Amplifier
Thermal
Power
+ Loss (radiated)
Thermal
-2 watts/volt
Compensator
--
External
(inverting)
P.
c'
Thermal Power
Loss (radiated)
o
Reference Temp. (+60 C)
4
Functional Block Diagram
Temp. (+6O0C)
Figure by MIT OpenCourseWare.
Cite as: James Roberge, course materials for 6.302 Feedback Systems, Spring 2007.
MIT OpenCourseWare (http://ocw.mit.edu/), Massachusetts Institute of Technology. Downloaded on [DD Month YYYY].
I
Output
Temperature
t,lle t,l~errr~al
system. Provisior~t,o ark1 an i n p l ~ tt,o t,lle error sigrlal is also ax-ailal~le.This test i n p l ~ tis also
11se(1to evalllate tllerrrlal rlynarnics.
Tlle S ~ I I ~ofI t,lle error sigrlal and the test signal provide t,lle i n p l ~ tto the syst,ern <:ornpensat,ionamplifier.
The input arlrl olltpl~tpoints of t,lle cornpensator are arxi1at)le alor~gr r i t l ~a s ~ v i t c lillat
~ rliscor~r~ects
tl~e
ir~terrlalcornpensation arnplifier. This cornl~inationallor%-sy o l ~to use the <:ornpensat,orthat y o l ~rlesignecl to
1:10se the 10op.
Tlle out,pl~tof the <:ornpensat,oris applied to a porrer amplifier t,llat drives a heater locat,ecl on the opposit,e
slirface of the plate on r r l ~ i c tl ~
l ~ etllerrr~istoris rnoi~ntecl.Tlle power amplifier is rlesignerl so that the heater
~ ~ o r r is
e r linearly proport,ional to the voltage applierl to its input,. This type nf arrlplifier is require(i si11c:e
t,ernperatllre and pox%-erare linearly relat,e(l in tllerrrlal systerns.
Tlle volt,age a t t,lle porrer arnplifier i n p l ~ tis also arxi1at)le for use in e r x l ~ ~ a t i nloop
g (lynarni<:s. The scale
fact,or illat relat,es t,lle voltage rrlor~itorerlt,o lleater porrer is -2 matt,/volt. Note that this rrlor~itorvoltage is
valirl only rvller~t,lle ir~terrlalcornpensation is t~eingilsed. Horrever. the scale factor of -2 rrat,t/volt rerrlair~s
t,rile rnller~llsirlg exterrlal cornpensation.
Tlle tllerrrlal systern loses power corlstar~tlytllrollgh t l ~ e~ ~ r ~ i r ~ s l ~rarliator
l a t e r l plate. Since t l ~ earr11)ient
t,ernperatllre and t,lle ternperatllre of t l ~ et,l~errr~al
st,n~<:tllre
rerr~air~
essentially const,ant, this heat loss is
<:onst,ant. The rlet porrer applie(i to t l ~ et,l~errr~al
is t,lle (iifference t)et,meen the porrer delivered t)y
st,n~<:tllre
t,he lleat,er arlrl t l ~ epower lost throllgh t l ~ erarliat,or. The power lost via the radiator allorrs t l ~ esyst,ern to
eitller raise or lnmer t,ernperatllre using orlly a cor~t,rolle(i
heater.
Cite as: James Roberge, course materials for 6.302 Feedback Systems, Spring 2007.
MIT OpenCourseWare (http://ocw.mit.edu/), Massachusetts Institute of Technology. Downloaded on [DD Month YYYY].
Prelab
1. Explain why tlle cornpensator in t,lle t~lockrliagrarn on t,lle next page has t,o t)e invert,ing
2. Corlsirler t,lle circuit in Figure 3,
Ignoring tlle (iiocle. (iet,errnine tlle trarlsfer f i ~ r ~ c t i o#.
r~
Use the forrrl
(r.3
+ 1) for poles ancl zeros.
IJ-hat p ~ ~ r p o sdoes
e tlle (iiocle serve?
3. Go irlto la11 and (lo tlle .'Initial \Ieas~~rernents"
section, rrleas~lrirlgthe open-loop l~ehaviorof t,l~e
t,llerrrlal
systern. Turn in a M.iTL.iB Bncle plot of your rlat,a.
4. (Optional) Explain 11orr t,o design a lleater power amplifier sll<:hillat
to input voltage. If you llave had 6.301, rlrarv tlle circuit schernati<:.
O I I ~ ~ porrer
I I ~
is lirlearly relat,ecl
Cite as: James Roberge, course materials for 6.302 Feedback Systems, Spring 2007.
MIT OpenCourseWare (http://ocw.mit.edu/), Massachusetts Institute of Technology. Downloaded on [DD Month YYYY].
Laboratory Setup
!Ye will attempt to llave a t least t l ~ r e e(3) <:orrq)lete,i(ie~lti(:r%l
tllerrrlal systerrls arxilat)le in t l ~ e1al)oratorj~
a t all t,irnes. At least orle of t,l~esemill l)e connecte(i t,o a sigrlal generat,or arlrl storage os<:illoscope so that
you can rrleasllre t,lle relat,ively slow trarlsier~tresponse of the syst,e~r~.
-1separat,e syst,errI is perrr~e~~r%r~t,ly
<:onnect,e(lto a. dynamic analyzer t,llat allows yo^^ t,o rrleasrue the t,rar~sferfilr~ct,inr~
nf the therrr~alsysterr~.
N O T E : You sho~ll(lnot change any of t,lle <:onnect,ionsin the dynamic arlalyzer config~lration..i-\rl(lit,ionally, you s h o ~ ~ lrlot
( l cllar~geany of t,lle conne<:tionsin t,lle step response set,up urltil you are ready t,o try out
yollr cornpensator.
Tlle t,l~errr~al
systerrl t,akes a l ~ o 30
~ ~rrlirl~~tes
t
following t,urr~on to stal~ilize,and t , l l ~ mill
~ s l)e left r~lnning
(hiring rlorrrlal 1al)oratorj~hours. Do ~ ~ , ti~:i"r~,
o t r ~ f l tlrf! system! If you are cor~cerr~erl
that a syst,ern is riot
st,al~ilize(l
or riot operat,ing properly. check t l ~ evoltage a t t,lle Porrer klonitor DNC connect,or. If this voltage
is riot in t l ~ erarlge of -0.5 t,o 5V. sorr~etl~ir~g
is wrorlg.
Caution
T l ~ eradiator plate rrlust face ~lp\%-arrl
(luring operation, and norrnally is a t a l ~ o +GOo.
~ ~ t !Yhilr t,l~isternperat,ilre is rlot rlangerol~slyllot in sllort rloses, you r r o ~ ~ lrlot
( i \%-antto t,oll<:ht l ~ eplat,e for very long. ill~norrnal
operation (sll<:has p ~ ~ t t , i a
n grlegative voltage of rrlore tllar~a. few volts on the exterrlal cornpensation connect,or, or <:ir<:~~it
11igl1plate terrlperat~lres. Tlle plate will 1)e locate(1 t,omar(i
faihlre) can reslllt i r ~dr%r~gern~lsly
t,lle rear of t l ~ ela11 l)er~<:h,
and illere is no reasor1 for you t,o t,oll<:hit. Do not 1111" anyt,hing on top of it,!
Initial Measurements
It is rliffic~lltto ex~~erirnentally
evaluate t,lle open-loop performance of illis systern. This is a <:ornplication
share~ll1y rrlarly feerll~acksysterns, sirlee keeping tllerrl i r ~the linear regior~mit11011t feerll~r%ck
is ge11era11y
corr~pensationin our ele<:troni<:sso t,llat you car1 rr~akeyollr rr~eas~~rerr~ents
(lifficlllt. !Ye 11ave incl~~rle(i
\%-it,h
t,lle loop closerl. n l l i l e ollr corr~pensationrloes st,al)ilize t l ~ esystern, t,lle performance \%-hichit acllieves is
extremely sluggish when <:ornpare(l\%-it,hthe ~llt,irr~ate
systerrl capal~ility. First, rrleasllre t l ~ eclose(1-loo11st,ep response l~et\%-een
t l ~
eExterrlal Drive and t,lle Ternperat,ilre Error
~ ~ o i n tNote
s . frorr~t,lle l)lo<:k(liagrarr~(Figure 2) that, t l ~ efee(1l)ackloop for t,l~isrrleasllrerrlerlt has imity gair~
in t l ~ efee(it)a<:kpath. l i 1 1 1 sho~ll(ladjust the arr~plit,ll(leof t l ~ elo\%-freqllrncy sqllare warre \%-hichprovi(1es
t,lle st,eps so that the systern stays in t,lle linear region. l i 1 1 1 car1 check t l ~ eperforrr~anceof t,lle systern t)y
ol~servingt l ~ ePower \Ionit,or point. Urlrler rlorrrlal con(lit,ions. t,lle voltage a t illis point will l ~ -1.5
e
to -3Y,
<:orrespon(iingto 3 to 6 matt,s of heater porrer. .i-\rlj~lst
t,lle square wave arr~plit~lrle
t,o give 11eak (ieviations
frorr~this operating point of at)out 1117(121'1;). The porrer rrlor~itorvoltage rrlllst rerr~air~
negat,ive a t all tirrles
~
as tn a110\%-a<:cllrat,ernean~rernents
t,o irlsure lirlerxr ~ ~ ~ e r r % tli oi r)~ot)jective
.
is to get a. large erlo~lgllsig~lr%l
~vhilerr~air~t,air~ir~g
linearity.
t,llat t l ~ esyst,err~response car1 t)e a<:cl~rately
.issl~rr~e
represente(i as illat of a cornplex pole pair. Extirr~at,e
pararnrt,ers w, arlrl for the approxirr~r%tir~g
pole pair. .ilsn esti~r~at,e
the (:10se(i10011 t)an(irvi(ith (w,,,) for this
<:onfig~lration
frorr~yollr step response rr~eas~~rerr~ent.
Next use t,lle H P 3562d Dynamic: .inalyzer to rrleasure t,lle trarlsfer fim<:tion frorr~heater po\%-erto tern~ l e r a t ~ l rerror.
e
Kote illat this trarlsfer fim<:tion inchl(1es all of your elerr~er~ts
in t l ~ eloop ex(:ept for the
(iesigrl ~ ~ 0 1 1~1: o
r r q ) e ~ ~ s a t ( ~ r .
<:ornpensator,arlrl tl111s 1:ontains all of t l ~ einforrn,ltion you nee(1 to s~~c(:essfi~lly
.is rnentionr(i earlier. t,lle conne<:tionst)et,meen t l ~ t,l~errr~al
e
systern arlrl t l ~ edynamic analyzer are in place
arlrl rrlust riot 1)e <:hangerI. Cllarlrlel 1 011 the arlalyzer is connect,e(l to t l ~ epower amplifier input P ( t ) . arlrl
<:hannel2 is conne<:tr(ito t l ~ ternperat~lre
e
error o l ~ t , pT~(~t t) .The analyzer drives the systerrl fr01r1 its ir~terrlal
source and <:hara<:terizesthe trr%rlsferfilrlct,inr~t)et,~%-eer~
<:l~r%r~r~el
1
<:l~r%r~r~el
2. (:nrlsi(ierir~g(:ha1111e11 tn t)e
t,he inpllt to tlle syst,err~llrlrlrr test.
Tlle analyzer is a. very capat)le and <:orrespon(iingly <:ornplex inst,nlrr~ent,In order to sirnplify it,s use
in lal). Ire l ~ a v epre-prograrnrned t,lle rlecessarji test steps. There are trro opt,ions r r l ~ i c you
l ~ may clloose in
<
Cite as: James Roberge, course materials for 6.302 Feedback Systems, Spring 2007.
MIT OpenCourseWare (http://ocw.mit.edu/), Massachusetts Institute of Technology. Downloaded on [DD Month YYYY].
t,esting t,lle tllerrrlal t,rar~sferfilnction. One uses rarlrlorr~rloise for t l ~ edrive signal. the ot,ller uses a sr%-ept
sine Tvrme.
These trro prograrns are accessecl t)y p~lshingt l ~ e.iUTO SEQ key m11i~:his locat,ecl on t,lle ilpper right,
orti ti on of t,lle front panel. After yo^^ plls11 this key, lal~elswill ap11ear next t,o t l ~ erow of soft-keys along t l ~ e
right hancl edge of t,lle analyzer CRT. Tlle t,mo upper keys are 1at)ele~l6.302 RdNDOhl arlrl 6.302 SITEEP.
<:orrespon(iingto random rloise testir~gand sr%-eptsine t,est,ing. respectively. Y ~ I may
I
rrleasllre t l ~ etrarlsfer
filr~ctior~
using eit,ller one, or l~ot,hif you like. Tlle test is st,arterl l1y 1111s11ingthe appropriate soft-key. arlrl
each reql~iresat)out t,mo rr~irl~~tes
t,n (:n111111ete.
111 t,lle random rrlorle, t,mo trarlsfer filrlct,ior~sarnples are t%verage(iin order t,o rer111ce the effects nf rloise.
No data apl~earsurlt,il t,lle first r r ~ e a s ~ ~ r e r ris~ <:ornplet,e,
er~t
at r%-hichpoint a. field in t,lle ilpper part of t,l~eCRT
will ir~clicate1 AITG. IYkler~t,l~ese<:nr~rI
rrleasurerrlerlt, is c:orr~plet,e,the field will ir~rIic>~t,e
2 .iYG. 111the sr%-ept
sine I I I O ~ Pt,lle
. sr%-ee11
st,arts at low freqllencies, arlrl data points are rlisplaj~e(ias they are rneas~lrerl.111 l~ot,h
rr~ocles.t l ~ et,est sigrlals rrlaji t)e ol)serve(i on cllarlr~els1 and 2 of t,lle oscillos<:ope.
Tlle ~ r ~ e r % s ~ ~ r ecnver
~ r ~ ethe
r ~ t ,fre(l~lerl(:y
s
rr%rlge0.1 HZ to 10 Hz, as t,l~isis the regior~of interest,. Uecallse r%-e
are rneasllring 10x3- freq~lencyl~eharior.t,lle rrleasllrerrlerlts require a lor~gtirr~et,o corr~plete.In otller mor(is.
t,lle several rrlirl~~tes
involver1 are not rlue t,n prncessir~gtirr~ei r ~the analyzer; they are rlue to the rather 10119
t,irr~ecor~st,ar~ts
of t l ~ etllerrrlal systern. For yollr interest, r%-el~avein<:l~~(ie<l
a plot in Figure 1 which shor%-s
~ I O I I ~tn
S (:nrr1111et,e.
t,lle t,rar~sferfilrlct,inr~irler%s~lred
frnrr~0.01 Hz tn 0.1 Hz. This test r e q ~ ~ i r eseverr~l
d
tV11e11t l ~ etrarlsfer filrlct,ior~rr~easl~rerr~er~t
is complete. you may use t l ~ ecursor
to take rlat,a on rrlagnit~l(ie
arlrl phase as a. filrlct,ior~of freq~lency. The cursor is the srrlall circle r r l ~ i c lappears
~
on t,lle t,mo plots. Its
~~osit,ion
is <:ontrolle(it)y the large knoll to t l ~ right
e
of t l ~ eCRT. The freq~lency.rr~ag~~it,l~(le.
and phase ap11ear
in fields on the CRT. 011illis display. rrlagr~it~lrle
is incli<:ate(iin rlU (rlefined as 20 loglo # ) .
Tlle phase is
inciicaterl in degrees. Note t,llat zero degrees of arlciitional phase sllift corresponrls to +180° on t,l~eanalyzer
plots (111e to the irlversior~frorr~po~verr%rr~plifier
i r q ~ t,n
~ ~ternperat,llre
t
error.
a rr~irlllt,eto corr~plete.
Tlle arlalyzer oc<:asionally er~t,ersan al~t,ocalil~rate
cycle r%-hichr e q ~ ~ i r eat)out
s
Durir~gillis tirr~e,keyl)oar(i input is clisal~le(i.and t l ~ ea~~toc:al
st,atns is inrlicat,ecl on t,lle CRT.
If you are ir~teresterlin ler%rrlir~g
r1101.e r%1)011tt,lle r%rlalyzer.a copy nf it,s i ~ ~ t r n d ~ ~ c 11ser's
t , n r y irlr%rl~lr%l
is
ar~ailal~le
in the lal]. The rletailerl operat,ing rr~arlllalis in the instr~~ctor's
office, arlrl rnay l ~ read
e
there l~ut,
riot rerr~overl.
1
Discussion of Measured Frequency Response
Certairl asl~ectsof illis trarlsfer filrlct,ior~car1 l)e est,irr~aterlfrorr~the properties of t,l~errr~al
str~~ct,l~res.
Cor~sirler
a serni-infinit,e <:onstant cross sectior~bar of heat <:or~cl~~ct,ing
rr~aterialillat is t,herrnally ins~~lat,e<l
except at
%
irnpeciance. or the ratio
for tl~errr~al
varial~les,at t,l~eer~rl
one end. It car1 1)e S~IOTVIIthr%tthe tl~errr~al
of t,lle l ~ a ris proport,ional t,o
Ever1 t,ll011g11t,here are very sigrlifi(:r%ntrliffererlces t)et,meen t,llis idealizer1
system. t,lle trarlsfer filr~ct,ior~
sllo~vsa fairly r%-iderange rrl~erethe angle is <:loseto
rr~ocleland our t,l~errr~al
-15" arlrl the rnagnit,l~(leis proport,ional t,o
.it lligller freqllencies, the trarlsfer filr~ctior~
exhit~itsnon-rr~inirm~rn
phase lorr-pass characteristi<:s. Konrr~ir~irrl~~rr~
phase rrlearls that t,l~ereare ot,ller t,rar~sferfilr~ctior~s
with i(ienti<:alrrlagr~it~lrle
<:hara<:teristicsm11i~:h
phase systern.
ex11il)it less rlegative phase shift,. A -\111retirr~erlel:%y ( f ' - " " ) is an example of a non-rninirr~~nr~
S11c:h <:hara<:terist,icscar1 l ~ esllo~vr~
t,o place irnport,ant lirr~it,~
on t l ~ e~llt,irrlatesyst,ern systerrl ~~erfnrrr~r%r~ce
t,llat car1 l ~ achievecl.
e
This port,ion of the t,rar~sferfilrlct,ior~reflects t,l~errr~al
(liffi~sio~~
frorr~t,lle heater to t l ~ e
t,l~errr~istor throl~ghthe t,herrr~ist,oritself.
5.
4.
Compensator Design and Evaluation
Sirlee power is lost tllro~lgl~
t,lle rarliator plat,e <:ontimlol~sly,
it is necessary t,o ap111y heater porrer <:ontim~ol~sly
t,l~isrequired pnr%-erwit11 ilnrr~i~lr%lly
zern t,e~r~perr%t~~re
error. ~ ~ 0 1 1 1
t,o rrlake 1111 for t,l~isloss. In order to s11p111~7
<:ornpensator rlesign sho~llcli r ~ c l ~ ~ar l pole
e
at t,lle origin. d s~~ggest,ecl
topology for t l ~ ecornpensator mas
analyzer1 in t,lle prelal].
Cite as: James Roberge, course materials for 6.302 Feedback Systems, Spring 2007.
MIT OpenCourseWare (http://ocw.mit.edu/), Massachusetts Institute of Technology. Downloaded on [DD Month YYYY].
Low Frequency Transfer Function Data
-5
Mag (dB)
-7
-9
-11
-13
-15
0.01
Freq (Hz)
0.1
Low Frequency Transfer Function Data
180
Phase (degrees)
135
90
45
0
0.01
Freq (Hz)
0.1
Figures by MIT OpenCourseWare.
Cite as: James Roberge, course materials for 6.302 Feedback Systems, Spring 2007.
MIT OpenCourseWare (http://ocw.mit.edu/), Massachusetts Institute of Technology. Downloaded on [DD Month YYYY].
Design a corn~~ensator
t,llat achieves a. systerrl phase rr~argir~
of 50". ancl try to rnuirnize the c:rossover
freclllencjr as is cor~sister~t
with illis phase rnargin. li111rcornpensator rrlaji or may riot use all of tile cornpofeel free t o use a (iifferent cornpensator topology if you \%-ish.1)ut it
rlerlts sllorvr~in the prelall. l i 1 1 1 sl~ol~lrl
sl~olllrlnot llr rr111c11rrlore cornplicat,ecl tllar~illis topology. If yo^^ (lo use RI ancl CI, rr~akeR1 >
You
sl~olllrlfir111that start,ir~gwit11 Cp = O.llrF rrill reslllt ~ I rI e a s ~ ~ ~ r %
-\:or~q)or~e~~t
l)le
vallles. or1 ~ ~ 0 1 1prnt,ot)oard,
1
and conne<:t it to one of t l ~ et,l~errr~al
systerrls wit11 BNC
Duild $0111 -\:or~q)e~~sr%tnr
c:at)les. The i n p ~ to
~ t your <:ornpensat,orcorrles frorr~t,lle DNC 1at)elecl Cornpensat,ion In, and its o ~ ~ t ,goes
p~~t
t,o Exterrlal Cornpensation. To rrleasure the st,ep r e s ~ ~ o n sofe yollr firla1 c:lose(i loop syst,ern, yo^^ may fir111
t,llat llorro~%-ing
two '.E-Z llook DNC learls" frorr~t,l~e<:orn~~onrnt
desk in hel~1f111
in llookir~g111)yonr <:ir<:~~it
t,o t,lle tllerrrlal systerrl in lat).
Note that since yollr cornpensator has a. pole at t l ~ eorigin. it,s operation cannot l)e checkecl on an open
loop llasis.
klonitor the o ~ l t p l ~volt,age
t,
ancl s ~ v i t c t,lle
l ~ systerrl from 111terrlalt,n Ext,errlal COIIIof your corn~~ensat,or
1
cloes rlot set,tle to t l ~ erarlge -0.S to -S volts, rvitllir~a. few
~lensat,ion.If the vnltr%geO I I ~of ~ ~ 0 1 1-\:or~q)e~~sr%tnr
secon(is. s ~ v i t c lllack
~ to Ir~terrlalCornpensatio~~
and figure O I I ~~vhr%t
yo11 rlirl wrong. (Long ~lerioclsof large
negat,ive volt,ages rrill 1:allse t,lle rarliator plat,e t,o get extremely hot!)
On<:et,lle syst,ern is working properly wit11 yollr cornpensation, rrleasure t,lle st,ep response of t,lle lirlear
region. Cornpare t,lle performance wit11 that of t,lle original syst,ern to sllo~vt l ~ eirnprovernent achievecl t)y
yollr cornpensator.
tV11e11 you are ready to (iisconnect your <:ornpensator frorr~the systerrl. first srrit<:h llack t,o ir~terrlal
cornpensation. If yollr rernove or t,urr~off yollr cornpensat,or wit11 your syst,ern still operat,ing in the Exterrlal
Cornpensat,ion rnocle. the systerr~t , e ~ r ~ p e r r %will
t ~ ~rlestal~ilize
re
ancl the next, user will t)e very 11r111appy\%-it,h
vn11!
G.
Write Up Yn-~rlat) \%-rit,e1111 sho111(iinchlcle t,lle follorrir~ginformation: 1. The st,ep r e s ~ ~ o n of
s e the syst,ern wit11 our <:ornpensation,as rrell as an est,irrlate of
on a clorninant pair a~)~lroxirr~at,ion.
d,,,
<. ancl d,,,l~asecl
2. The t,rar~sferfim<:tionyou ellose for your cornpensat,nr. S ~ ~ p p oyour
r t elloice \%-it,harlalysis ancl Do&
plots to sllorv illat you olltain 50° of phase rnargin.
3. The circuit you ~lserlt,n irrlple~rler~t,
~ ~ 0 1 1(:nrrq)e~~sr%t-\~r.
1
1. The step response ancl t l ~ est,irnate(i
e
clllantit,iesreq~~est,ecl
in pr%rt1f ~ the
r syst,e~r~
\%-it,h
$0111 (:nrrq)arisor~.
I
r
~ s 1 1t 1 I e
l I
1
1 i t t l i 1 1 1 rrlaji hanrl-dram t~lockcliagrarns and arly ot,ller
sketcl~espr~viclerlthey are easy to read. Tr%l)lllat,erlat,a in tal~les.IJ-hen presenting experimental rlat,a. you
neatly sketching (rvit,haccurate
sl~ol~lrl
clearly explain exact,ly \%-hatexperiment is 1)eing clone. This ir~cl~~rles
lallels) t,lle systern input,. o ~ t , p ~arlrl
~ t any
,
rrleasllrerl clllantit,ies.
Check Off
Make sure t,llat you (in riot (iisassernl~leyollr corn~~ensator
ur~tilafter checkoff, as the T.4s mill \%-antto see
t,lle <:ir<:~~it
\%-llichyo^^ t ) ~ ~ i l t .
Cite as: James Roberge, course materials for 6.302 Feedback Systems, Spring 2007.
MIT OpenCourseWare (http://ocw.mit.edu/), Massachusetts Institute of Technology. Downloaded on [DD Month YYYY].