sweep.signal.generator.1955

7/28/2019 Sweep.signal.generator.1955

1/8

A p p l i c a t i o n s o f a Sweep Signal GeneratorF R A N K G . M A R B L E , Vi ce P resident- Sales

The design techniques used in the develop-ment of a new Sweep Frequency Signal Gen-erator were discussed in the Spri ng 1955Number 5 Issue of T H E NOT EB OOK .*That discussion covered the methods used toobtain the performance required of a pre-cision A M modulated Signal Generator,sweep and marker system in a single instru-ment. This article continues the discussionby considering some of the various methodsby which such an instrument can be used.One advantage of a sweep frequency signalgenerator lies in its abil ity to save time andthus economize engineering manpower, free-ing it forother constructive work. One might,for instance, use an adjustable frequency, ad-justable level cw signal generator to obtainoutput-vs- input data for an if ampli fi er ateveral discrete frequency points. This datacan then be plotted on a graph showing re-sponse-vs-frequency to obtain the pass-bandof the circuit. For each circuit readjustmentthis procedure must be repeated. For a nar-row pass-band circuit this process is at besttedious, but for a broad-band circuit its timerequirements are virtually prohibitive.The simultaneous display of the response-vs-f requency curve of a circui t on the screenof a cathode ray oscilloscope by a sweep fre-quency signal generator system, and the in-stantaneous indication of changes caused byadjustments expedites the engineer's workenormously. Another advantage of the sweepmethod is the practical fact that some of thetime so saved will be used to obtain refine-ments which would have been overlookedusing the slower single frequency methods.*" Sweep Frequency Signal Generator DesignTechniques* J ohn H . Metmie and Chi LungK m g - The Notebook Spring 1955 Number 5.

YO U WILL FINDApplications of a SweepSig nal Generator-- Page IUs e of the RF VoltageStandard Type 245A Page 4Calib rati on of th e Internal ResonatingCapacitor of the Q Meter- Page 7Editors Note Page 8

2 4 0 - A RECEIVER

F igu re 1. The author ad'usts the pass - band of the I F amplif ier below the 4.5 mc hwerl imi t of the Sweep Signal Generator 240-A by using the Univerter Type203-B. Thi s com-bination permits cw and sweep measurements from 0.1 t o 120 megacycles per second.Besides the savings in time and the greater

practical refinement obtained, some informa-tion is immediately observed by sweep meth-ods which can be easily overlooked in thepoint by point method. Regeneration effectsand "suck-outs" may cause changes in theresponse curve which persist over only a verynarrow range of frequencies. Since cwmeasurements are made at discrete frequencypoints only, it is possible to obtain a smoothresponse curve excluding these effects if theyhappen to lie between the selected measure-ment points. A Sweep Signal Generator pre-sents data which is continuous with fre-quency. This removes the possibil ity of miss-ing important information.Th e Basic Measuring SystemFig 2 shows the 240-A , a broad banddetector, and an oscill oscope interconnected.The resultant information which appears onthe screen of the oscill oscope is shown inenlarged form in the photograph in F igure 3.The display is a graph wi th abscissa pro-portional to frequency and ordinate propor-tional to the amplitude response of the detec-tor circuit.The interconnections in Figure 2 requiredto obtain the display include the connectionof the rf output of the signal generator tothe detector whose output connects to a

marker adder ci rcuit "T est In" in the SignalGenerator and hence to the vertical deflectionampli fi er of the oscil loscope. The sawtooth of

CRYSTALMARKS&

Figure L. I ypicar interconnections o t weepSignal Generatw Type 240 -A , Test Circuit(Broad Band Detector) and Oscilloscope.PULSE MARKS -

AMPLITUDE-F R E r ) " E N C Y -.-

Figure 3. Enlarged phot ograph of thedisplay appearing on the screen of theoscil l oscope in f igure 2.


2/8

BOONTON RADIO CORPORATION

TH E BRC NOTEBOOK i s publishedfour h i es a year by the Boonton RadioCorporation. I t is mailed free of chargeto scientists, engineers and other inter-ested persons in the ronznzunirationsand elertroiii cs fields. The contents maybe reprinted only wlth wi tten permis-s i on from tbe editor. Your conmentsand suggestions dre wel come, and.rhould b e addressed to: Editor, T H EERC NO TE BO OK , Boonton RadioCorporation, Boonton, N. .

voltage which frequency modulates the con-stant-amplitude RF output of the Signal Gen-erator is connected to the horizontal deflec-tion amplifier of the oscilloscope.The sawtooth voltage, while increasing inamplitude, modulates the constant ampli tudeRF output from a minimum to a maximumfrequency. Simultaneously it deflects theoscilloscope from left to right. At the maxi-mum frequency point the sawtooth starts de-creasing in amplitude, the rf output of thesignal generator is reduced to zero, the oscil-loscope deflected from right to left and thetuning mechanism of the signal generatorreturned to the minimum frequency point.The constant ampli tude rf output reappearsand the cycle is repeated.The lower l ine in the display in Figure 3represents the reference line of the graph orline of zero input and the upper line thedetector response curve.The frequencies along the Abscissa mustbe identified if the response curve is to havemeaning. I n the Sweep Signal GeneratorType 240.~4, frequency identification is ac-compli shed by two types of marks. T he marksat (a) in Figure 3 appear at the harmonicsof a crystal oscil lating at 2.5, 0.5, or 0.1megacycles. The marks indicated by thearrows marked (A ) have separation of 2.5megacycles. The center frequency can beidentif ied from the tuning dial of the SignalGenerator. W i th the center frequency iden-tified the frequency of each of the othermarks can be deduced since the frequencyspacing is known.The marks at (B ) on Figure 3 are sharpnarrow pulses, The position of these pulsescan be adiusted along the frequency axis by

Figure 4. Oscil loscope display result ingfrom insertion of a seiective circuit andassoc iated detector in Sweep Signal Generatortes t c i rcu i t .

front panel controls, The crystal marks canbe switched off after the pulses have beenpositioned to coincide with any two of thesemarks. This leaves these two frequenciesmarked in a manner which causes minimuminterference with the reference curve. Thepulses (B ) can also be positioned betweentwo crystal marks (A ) since the frequencychanges linearly with distance. The crystalmarks can then be switched off. I n this way,any two frequencies along the frequency axiscan be marked.The marks shown at (A ) and (B ) onFigure 3 are added to the display in themarker-adder circuit through which the sig-nal from the detector (shown in Figure 2)passes before it is connected to the verticaldeflection amplifier of the oscilloscope.

Determination of Selectivity andSensitivityThe elementsof a Sweep Signal Generatorsystem for measuring selectivity and sensitiv-ity of a test circuit are the same as shown inFigure2with the test circuit inserted betweenthe RF output and the detector. If the test

circuit contains a detector, the detector inFigure 2 can obviously be omitted. The re-sultant display appears in Figure 4. The con-stant amplitude signal source is frequencymodulated or swept from a low to a higherfrequency at a slow rate compared to thesignal frequency. When the maximum fre-quency of the sweep is reached the signalsource output is turned off and the generatorreturned to the low frequency point for asubsequent sweep from low to high fre-quency.

Figure 5. lnterconnections for observa-tio n of pass band of a single stage with-in an IF amplif ier.The test circuit detector provides responsecurve of attenuation vs. frequency and fre-quency identif ication marks are added to thevarying signal from the test circuit.The horizontal deflection connections ofthe oscilloscope are connected to the samevoltage that sweeps the signal source. The dis-play on the oscilloscope, Figure 4, includesthe response vs. frequency response of thetest circuit, the frequency identification marksand a base or zero reference line indicatingthe level out of the test circuit with no input.The selectivity of the test circuit is apparentfrom a comparison of the change in responsevs. the number of megacycles or kilocyclesper inch along the horizontal axis of the dis-play. Thi s frequency calibration of the hori-zontal axis is deduced from the markersshown.Selectivity usually varies with signal levelas a result of A GC , limiters, non-linear

amplifier, etc. Therefore it is important totest it at various operating signal levels. TheSweep Signal Generator T ype 240-A, men-tioned in the article cited i n the first para-graph of this article, provides calibrated out-put level from 1.0 to 300,000 microvoltswhile sweeping. Its leakage is sufficiently lowto permit use of an external 20 db attenuatorto obtain outputs down to 0.1 microvolt.

m o n n e c t i o n s or stua, fcable and cable termination ch aracteri s-t i cs .Selectivity of Single StagesThe system of connection in Figure 2 issuitable for receivers, filters or amplifiers.

The terminated rf cable (a 50 ohm system)is connected into the input of the test circuit.T he detector of the test circuit is connectedto the marker adder circuit i n the sweepsignal generator (input impedance 1 meg-ohm). The use of a sweep signal source isnot limited to complete receivers or ampli-fiers, however. So long as arrangements aremade to avoid any effect on the sensitivity orselectivity of the circuit under test by theimpedance of the rf output of the SweepSignal Generator or of the detector the selec-tivity-sensitivity characteristics of any circuitmay be observed within the sensitivity limitsof the oscilloscope being used.A convenient method of observing the passband of a single stage appears in Figure 5.The output of the Sweep Signal Generatorconnects to the grid of the tube which con-tains the test stage in its plate circuit. Abroad band detector is connected to the plateof the following stage through a couplingcondenser. T he low input impedance of thedetector lowers the Q of the circuit in theplate of this tube so materially as to make itseffect on the final result insignificant. T hetuned circuit of tube V1 is operating underits normal condition and its sensitivity-selec-tivity characteristic can be observed on theoscilloscope.

Study of Pass Band CharacteristicsThe Q of the pass band of a test circuitcan be approximately deduced by use of asweep signal generator. As discussed abovethe response curve of a circuit can be dis-played on an amplitude vs. frequency graphon the face of a cathode ray tube. The mark-ing system of the Sweep Frequency SignalGenerator makes it possible to identify anyfrequency along the horizontal axis. Sincethe response in the vertical direction on theoscilloscope is linear, a point 0.707 times thedistance from the zero or reference line to thepeak of the response curve can be locatedon each side of the peak. From the frequencymarking system the frequency difference

2


3/8

T H E NOTEBOOK( n f ) between these two points and the fre-quency of the peak can be obtained. Q canthen be obtained from the followingformula: Q =f,AfAdjustment or Stagger Tuned CircuitsBroad pass bands are often obtained byadjusting the resonant frequencies of suc-cessive single tuned circuits to slightly differ-ent frequencies within the desired pass band.The overall result i s a relatively flat pass bandbroader in frequency than any one of the in-dividual tuned circuits.T o adjust this type of amplifier, it isnormally quicker to first resonate each indi-vidual circuit to the proper frequency with acw signal generator. A fter completion ofthis procedure, the overall pass band con-figuration can be investigated and touch upadjustment made with a sweep signal gen-erator. The Sweep Signal Generator Type240-A is excellently suited to this proceduresince it operates as a cw (with or withoutA M ) or sweep signal generator, withoutthe necessity of disturbing the input or out-put connections to the test circuit. A VacuumTube Voltmeter can be bridged across theinput to the vertical deflection amplifier in-put of the oscilloscope for the single fre-quency work. T he oscilloscope of course isused for the overall investigation and touchup. Since the output monitoring and at-tenuation system is equally applicable to cwand sweep work the sensitivity can easily bechecked under either condition.

Study of Cable CharacteristicsThe characteristics of high frequency cablesmay be investigated by use of a sweep fre-quency signal generator. In Figure6a sweepsignal generator is shown connected to theinput of a length of high frequency cable.A lso connected to the input of the cable is awide band detector. The low frequencysweep voltage from the sweep signal gen-erator is connected to the horizontal deflec-tion input of the oscilloscope. The RF signal,swept or frequency modulated at a low rateof 60 times per second, is fed into the cable.Reflected signal from imperfections in thecable or the termination arrives back at the

Figure 7. Oscilloscope display indica-ting amplitude of reflect ed energy from atermination coaxial cable.input a finite time later. Since during thisfinite time the input signal has changed toa new frequency, an audio difference fre-quency (input frequency minus reflected fre-quency) appears across the output of thedetector. The amplitude of the input signal

is great and constant and the reflected fre-quency amplitude for a near match is smalland variable. The amount of energy reflectedfrom the end of the line depends on the cor-rectness of the termination and varies fromzero for a perfect match to afinite value pro-portional to the mismatch for mismatchedlines. Since the termination impedance will ,in general, vary with frequency, the amountof energy reflected wi ll also vary. T he audiofrequency from the detector appears on theoscilloscope, The envelope amplitude of thedisplay is proportional to the instantaneousreflected signal and the abscissa is propor-tional to frequency of RF input as shown in

Figure 8. Diagram of equipment and con-nections for measurement of l in earit y ofF M discriminator.W ith a perfect termination over the fre-quency range in question, various cables canbe observed for imperfections in construction.A periodic variation in dielectric constant ofthe cable insulation will exhibit itself on theoscilloscope display.Adjustable resistance load will permitquick determination of the Z, for long cablelengths.

The linearity of F M DiscriminatorsThe Sweep Signal Generator T ype 240-Aprovides a powerful method of determiningthe linearity of an FM discriminator. Themethod is indicated in Figure 8. A low fre-quency (60 cps) sweep, adjusted to sweepthe ful l frequency range of the discriminator,and a higher f requency sweep (400 cps) isfed into the EX T sweep input of the SweepSignal Generator. T he high frequency volt-age is adjusted to sweep only a small fractionof the frequency range of the discriminator.I n effect the high frequency sweep exploresthe slope of each section of the discriminatorwhile it is slowly moved from section to sec-tion by the low frequency sweep. T he outputis detected and passed through a high passfilter which passes only the resulting 400 cpsnote. T he display of the amplitude of thi snote vs. the low frequency. sweep affords avisual display in which the slope of theamplitude of the envelope of the 400 cpsnote is proportional to FM discriminator line-arity. A constant amplitude indicates a lineardiscriminator whereas a varying amplitudeindicates a variation in linearity.

The Study of Crystal ModesThe rapid location of the several frequencymodes at which a crystal oscillates is impor-tant but tedious by discrete frequency meth-ods. The Sweep Signal Generator Type 240-A provides a frequency sweep onwhich indi-vidual frequencies can be identified by the

marker system included in the generator. hcrystal, however, has such a high Q thatsweep rates must be very low to prevent ri ng-ing and spurious responses. By using anoscilloscope with low frequency sawtoothsweep available, the 240-A can be swept atfrequencies of 1 or 2 cps by connecting theoscilloscope sweep output to the External

SIGNALGENERATOR

. re 9. Equipment arrangement fo rmeasurement of quartz crystal character-i s t i cs .Sweep of the 240-A . The system is thenconnected as shown in Figure 9. By varyingthe center frequency of the 240-A and itssweep width the crystal can be explored forresponses over a considerable frequencyrange.

Extension of the Frequency RangeThe lowest center frequency of the SweepFrequency Signal Generator T ype 240-A is4.5 megacycles. A t this frequency the sweepfrequency capabilities of the instrument are2 1 % to 130% of center frequency or2 4 5 K C to : 1.35M C. For applications intelevision video amplifiers both for color andblack and white, and aircraft navigation re-ceiver intermediate frequency amplifiers,lower center frequencies and/or broadersweeps are required. Both these requirementscan be met by use of the Univerter T ype 203-B with the Sweep Frequency Signal Gen-erator Type 240-A . The 203-B consists of abroad band mixer with local oscillator at 70M C. fol lowed by a broad band amplifier wi tha 50 ohm output. T he gain of the 203-B isset at unity. Figure 1shows the 240-A, 203-B in a measuring set-up. I n use the 240-Ais tuned to a frequency equal to 70 M C plusthe desired output center frequency from thesystem. Sweeps from 2 0 . 7 M C to 1 1 5 M Care available. Single frequency outputs un-modulated or with AM modulation can beobtained. Thus single frequency or sweepoutputs are made available over the bandwidth of the 203-B which is 0.1 to 25 MC .SummaryThe Sweep Signal Generator is a powerfultool of considerable flexibility. I t not onlysaves considerable time but makes refine-ments possible in circuit adjustment and de-velopment which would not normallv beposble T H E A U T H O RFrankG.Marb1e.s career covers a broad fi elaof engineering experi ence: design Gdevelop-ment work for Philco: coordinator on variousgovernment projects; two years with WesternElectric s electrical research division G en-gineering administrative posts with Prutt GWbitney Aircraft G Kay Electrical Co. Mr.Marble bas been with Boonton Radio sinceI951 G a Vi ce-president-sales since 1954.M I. Marble has aBS in E E (M ississippi StateColleQe 1934) G an MS in EE (M.I.T. 1935).

3


4/8

BOONTON RADIO CORPORATION

Use of the RF Voltage Standard Type 245-AWhen discrepancies exist among measure-ments made with different signal generatorson the same radio receiver, it is often very

difficult to determine just which of the in-struments is performing correctly. The intro-duction of a reference standard usually willresolve the dilemma so that effort can profit-ably be applied to the offending units. How-ever, care must be exercised in the use ofsuch a standard and understanding appliedto the interpretation of the results. T hisarticle discusses the use of a source of stand-ardized voltage at a known impedance.The R.F. V oltage Standard Type 245-A ,shown in Figure 1, is designed to deliver,across the BNC output jack at the end of itsOutput Cable, open-circuit radio frequencyvoltages of s,1, and 2 microvolts througha source impedance of 50 ohms over the fre-quency range of 0.1 mc to 500 mc. I t canbe used in conjunction with a signal gener-ator as a source of known voltage and im-pedance for determining receiver sensitivityperformance. Using this source of vol tageas a point of reference, it is also possible toperform relative comparisons, with othersources of radio frequency voltage whose fre-quencies li e within the specified range. I naddition, the input system is calibrated f oruse as a 50 ohm rf voltmeter at 0.05 voltsover a wide frequency range.

Principle of OperationA description of the RF V oltage Standardis given in the Spring, 1955 issue of the BRCNotebookl. T he system block diagram of theRF V oltage Standard is shown in Figure 2.A n external source is used to supply rf volt-age to the Input Cable. T he voltage at theoutput end of this cable is indicated by anRF V oltmeter at the point where the cable isterminated by the input to the Coaxial A t-tenuator2. The low voltage output of theCoaxial A ttenuator appears in series with animpedance matching resistor.

W. C. M O O R E , Engineer ing ManagerCALIBRATION SET FULLSCALE

INPUT CABLE

SET Z E R O /FINEFigure 7 . RF Voltage Standard Type 245-A.

A n input level of 0.05 volts is establishedacross the input to the coaxial attenuator byadjusting the voltage output of the externalrf voltage source until the indicating meteron the RF V oltage Standard reads at the1 microvolt calibration point on the meterscale wi th the range switch set to 1micro-volt. T he 25,OOO:l coaxial attenuator di-vides the 0.05 volts down to 2 microvoltswhich appear across the 0.0024 ohm resistorin series with the 50 ohm impedance match-ing resistor in the rf assembly.Since the 50 ohm characteristic impedanceof the Output Cable is matched by the 50ohm resistor, its length i s electrically inde-terminant. I n fact, its length may even beconsidered zero, and the 50 ohm terminatingresistor is effectively connected to grounddirectly f rom the 50 ohm impedance match-ing resistor in thc RF Assembly. Thisdixidesdown the 2 microvolts delivered by the1UTPU T TERMINA TING IMPEDANCECABLE RESISTOR MATCHING OUTFRESISTOR - RESISTORVOLTAGE

R F , , OUTPUT C A B L E IA S S E M B L Y A S S E M B L YIFigure 2. R F Voltage Standard System Bloc k Diagram.

The low voltage output from the RF AS-sembly, which presents a source impedanceof 50 ohms, drives the 50 ohm Output Cablewhich is terminated by a 50 ohm coaxialteiminating resistor. T he terminating resistoris followed by a 25 ohm impedance matchingresistor to raise the equivalent source imped-ance at the end of the output cable to 50ohms.

Output Voltage CalibrationFigure 3 shows the distribution of voltagesthroughout the instrument when the meter isset at the 1 microvolt level and there is noexternal load connected to the output cable.

Coxial Attenuator to 1microvolt across the50 ohm terminating resistor.The meter on the RF V oltage Standard iscalibrated in terms of the open circuit voltageappearing across the BNC output jack on theoutput cable, with no load connected to thecable.

Output ImpedanceT he output system of the RF V oltageStandard is based on a 50 ohm characteristicimpedance The optimum conditions forpower transfer and control of voltage stand-ing waves on the cable as the load impedanceis varied are described in the Fall 1954 issue

ZERO ADJUSTR F OUTPUT- C A B L E

of the BRC Notebooks.The output impedance of the RF Assem-bly is 50 ohms, determined by the 50 ohmimpedance matching resistor, which is thetermination of a specially designed section ofcoaxial transmission line 1, Looking backalong the coaxial output cable from the 50ohm terminating resistor toward the RFAssembly, one sees the 50 ohm characteristicimpedance of the cable in shunt across the50 ohm terminating resistor. T he net resultof this parallel combination is 25 ohms ofresistance which is then built up to the de-sired 50 ohms by the 25 ohm series Imped-ance matching resistor located between theterminating resistor and the l3NC output1ack.The open circui t output impedance at theoutput jack on the Output Cable is 50 ohms.

Measuring Receiver SensitivityThe sensitivity of a radio receiver has beendefined by the Institute of Radio Engineers4as the number of microvolts required to pro-duce standard output when applied to thedummy antenna in series with the input im-pedance of the receiver For asystem consist-ing of a 50 ohm transmission l ine system anda50ohm receiver, this means that a 1micro-volt receiver will produce standard outputwhen 1 microvolt is applied across the seriescombination of the 50 ohms antenna imped-ance and the 50 ohm input impedance of thereceiver This yields y2microvolt across thereceiver input terminals.Figure 4 shows how this condition is metby the voltage calibration and output imped-ance characteristics of the RF Voltage Stand-ard Type 245-A The actual circuit can bereduced to a schematic circuit because thecharacteristic impedance of the cable ismatched at the voltage source as describedabove The diagrams show the distributionof voltages and impedances along the circuitsfor the loaded and open circuit conditions

4


5/8

T H E NOTEBOOKwhen the meter indicates 1 microvolt. Atthis level setting, the circuit is being drivenby 2 microvolts out of the Coaxial A tten-uator.The equivalent circuit diagrams show thatthe same loaded and open circuit character-istics of vol tage and impedance will be pre-sented to the load i f we assume a simpleseries circuit consisting of a 1microvolt gen-erator in series with 50 ohms. This resultcould have been obtained directly by an'ap-plication of Thevenin's Theorum to the orig-inal circuit. A dditional diagrams and ex-planatory information can be found in theInstruction Manuals for BRC Signal Gen-erator Types 202-B and 211-A , and Uni -verter TvRe 207-A.

F igu re 3. Voltag e Distribut ion for OpenCircu i t Vo l taqe of I UV at End of OutputCable.The sensitivity of a receiver designed towork with a 50 ohm antenna line impedancecan therefore be read directly from the meteraty2,1,and 2 microvolts because the equiv-alent source impedance of the RF VoltageStandard provides the 50 ohms to which yz,1, or 2 microvolts are applied.If higher values of antenna resistance areinvolved, direct readings of receiver sensi-tivity can be obtained by merely adding inseries with the output cable a suitably-mount-ed, non-reactive resistor whose resistance isequal to the desired antenna resistance minusthe 50 ohms already presented by the R FVoltage Standard. For example: to readdirectly the sensitivity of a receiver designedto work from a 75 ohm line, such as RG-

11/U, a 25 ohm resistor must be added inseries with the inner conductor at the BNCoutput jack on the output cable to obtainthe correct impedance match. If values ofantenna resistance less than 50 ohms are in-voived, it is necessary to use an impedancematching pad and allow for its insertion loss.Checking Signal Generator OutputThe use of the RF V oltage Standard tocheck the output from a signal generator isbased on using a receiver as an un-calibrated

50n

EQUIV+L_E NT-qIECUlT EQU IV +EN T CIRCUI

Figu re4. Derivarron ur Lyurva lent Cir cu i tof R F Voltage Standard Output SystemAssuming a Matched L oad & 7pV Setting.

transfer indicator to compare the outputsfrom the two sources at a fixed signal level.Figure 5 shows the steps for the case of asignal generator having 50 ohms output im-pedance at the output jack.The method shown in Figure 5, in whichthe same Output Cable is switched from theRF V oltage Standard to the signal generatoroutput jack is valid only for signal generatorshaving a 50 ohm source impedance at thepanel output jack.Some signal generators, however, present50 ohms only at the output end of their ownspecial 50 ohm terminated output cable. I nthis case, the receiver input must be trans-ferred between the terminals of the outputcable on the signal generator and the outputcable on the RF V oltage Standard. Only inthis way will the comparison show up stand-ing wave errors in the signal generator outputsystem.

F VOLTAGE AT HIGH LEV EL IS OBTAINED FROM THE SIGNALENERATOR AND ADJUSTED TO GIVE THE DESIR ED OUTPUT.EFERENCE READING NOTED.HE A T T E NUA T E D O UT P UT IS PICKED UP AND R E C E I V E DAPID A

CW S f f iN A LGENERATOR @z=son2 4 5 f AOUTPUT CABLE

HE LOW LE VEL OUT PUT OF THE SIGNAL GENERATOR I S ADJUS-ED TO PRODUCE THE S4ME REFERENCE LEVEL READING ONHE RECEIVER AS WAS PRODUCED BY THE K NOWN LOW LEV EL

Figure 5. Comparison of Voltage Outputfrom a 50 Ohm Signal Generator wi th th eR F Volt age Standard, Usi ng a Receiveras an Oncalibrated Transfer Indicator.I n case the signal generator has a sourceimpedance of 50 ohms, i t is not necessarythat the receiver input impedance be matchedto the signal generator output impedance toobtain avalid comparative reading. Since thetwo sources of voltage present the same im-

I50 OHMSOUTPUT IMPEDANCESIGNAL GENERATOR25 OHMSOUTPUT IMPEDANCE

Figure 6. Comparison of Voltage Outputfrom On-equal Source Impedances by Addi-tion of an External Impedance MatchingResistor.

SY ADDING AN'EXTERNALSERIES IMPEDANCEMATCHING RESISTOR TOTHE SIGNAL GENERATOROUTPUT T HE TWOSOURCES CAN BE.COMPARED DIRECTL Y

pedance, it is necessary only that the receiverinput impedance remains constant, at, what-ever value it may have, throughout the com-parison process. For this reason, only thesignal generator frequency can be changed topeak the receiver response, since smallchanges in receiver tuning may result inappreciable changes in input impedance.A n ampli tude modulated signal can beused with an A M receiver and an audio volt-meter, provided the amplitude modulationis kept below 3070.

Unequal Sburce ImpedanceThe problems of i nterpreting signal gen-erator output readings increase when check-ing the calibration accuracy of a signal gen-erator whose output impedance cannot bemade the same as the reference standard bysuitable resistive pads, as shown in Figure 6,or whose output cable system sets up standingwaves at critical frequencies. These sameproblems arise in the use of such a signal gen-.erator for receiver sensitivity measurements.The necessary information to make these cor-rections i s given i n some detail in cataloguesand instruction manuals by the major manu-facturers of signal generators." L C L l Y L " I m r U I LR">".L

STANDARDINPUT TO RF VOLTAGE STANDARDADJUSTED TO PRODUCE IxVMETER INDICATION NOTERECEIVER REAOINGOUTPUTDJUSTEDROMO IGNALRODUCEENERATORAMERECEIVER INDIC ATI ON AS WASPRODUCED BY RF VOLTKESTANDARD

F igu re 7. Signal Generator Calibrationwhen al l Three mpedances are Different.Figure 7 shows a case in which the imped-ances of the RF Voltage Standard, the signalgenerator and the receiver are 50 ohms, 300ohms and 150 ohms respectively. T he gen-eral equation shown in the figure gives thenumber of microvolts actually delivered bythe signal generator for any combination ofimpedances in terms of the indicated outputlevel of the RF V oltage Standard.T he presence of standing waves in the out-put system of a signal generator which is notmatched internally will produce errors incalibration which must be corrected by usingdata supplied in the signal generator instruc-tion manual. These errors are a function offrequency and must be taken into account ateach frequency setting.I n summary:1. Determine the output impedancecharacteristics of the signal generatorbeing calibrated.2. A ttempt to modify it to 50 ohms bythe use of pads or dummy antenna sys-tems, taking into account their effect onthe calibration dues to attenuation char-acteristics.

5


6/8

BOONTON RADIO CORPORATION3. If the output impedance cannot bemade 50 ohms, determine the compleximpedance of both the receiver and thesignal generator and calculate the result-ing voltage divider. A lso calculate thevoltage divider consisting of the re-ceiver and the 50 ohm impedance ofthe RF V oltage Standard.4. Since the outputs of the two voltagedividers are equal when the signal gen-erator output is adjusted to give thesame receiver reading as the RF V oltageStandard, we can equate the two expres-

(E).(Esg) =ions as follows*Zr = receiver input impedanceZag=signal generator output impedance50 =RF Voltage Standard output impedanceEsg =signal generator open circuit voltageE=RF Voltage Standard open circui t vol t age

Then the signal generator setting, Esg,which will produce the same receiver re-suonse as the outDut of the RF V oltage

where Zr + s g Z,+ 50

Gandard, E, can 6e determined from tceequation: E Zr +Zsgsg Zr+ 5 0 (E)Use A s A 50 OHM RF VoltmeterT he input system of the RF V oltage Stand-ard is shown in Figure8. I t contains a lengthof coaxial cable which conneets the sourceof power to the coaxial head, which con-sists of a diode voltmeter in parallel with theinput to the precision coaxial attenuator. T hediode voltmeter reads the input voltage di-rectly at the input to the attenuator, and thecalibration of the RF V oltage Standard is notaffected by standing waves on the cable aheadof this point.The 60ohm attenuator input impedance isshunted by approximately 300 ohms diodeimpedance, which together form approxi-mately a 50 ohm termination for the 50ohminput cable. T he voltage seen by the diodevoltmeter at the input to the attenuator wi llbe nearly the same as that applied at the in-put BNC connector, subject to voltage stand-ing waves on the cable. Variations in thecharacteristic impedance of the cable and thediode impedance introduce a moderate stand-ing wave of voltage on the cable which in-creases with frequency.The ratio f or each coaxial attenuator isindividually determined, and the correct in-put voltage for the 1 microvolt level metersetting is given on the voltmeter calibrationdata plate on top of the instrument. This in-formation can be used for checking the in-strument at low frequencies (below 500 kc)and for measuring rf input voltages. W iththe range switch in the 1 microvolt position,the input voltage is increased unti l the meterindicates 1 microvolt. T he input voltage isthen equal to the value stamped on the dataplate. Input voltages of y2 and 2 times thisvalue can be determined by adjusting theinput for meter indications of 0.5 and 2withthe range switch in the corresponding posi-tion. AccuracyThe method of setting up the calibrationof the RF Voltage Standard at the factory issuch that the initial accuracy is determined

by the care with which the 60 ohm and the0.0024 ohm resistors are measured, the accu-racy of the voltage source used to set up therf voltmeter circuit, and the Voltage StandingWave Ratios (V SWR) of the input to thecoaxial attenuator and the impedance matchof the output cable termination. Of these,the VSW R is the least accurate measurementand also the greatest contributor to the over-all tolerance.

F igu re 8. RF Attenuator and Voltmeter.The GO ohm film resistor is the center con-ductor of a terminated transmission line2and together with the 0.0024 ohm disc re-sistor i t provides an accurate attenuator use-ful over a very wide range of frequencies5.The actual ratio is taken into account in set-ting up the voltage into the attenuator andadjusting the meter to read the desired out-put voltage. T he uniformity with frequencyof the attenuation ratio is determined by com-paring each unit against acarefully measuredstandard unit at several points over a widefrequency range.

gure 7 . m r v oirage aranaara-oasrcC i rcu i t .The long term accuracy, which is of con-siderably more importance and upon whichthe specifications are based, includes thestability of several components not involvedin the initial calibration. A circuit has beenchosen in which these variations are minim-ized by the procedure used to place the instru-ment in operation.T he simplified circuit of F igure 9 showsthe basic dc metering system associated withthe rf voltmeter. T he rectification efficiency,or ratio of rectified dc current to applied acvoltage, of a semi-conductor diode at a con-trolled value of bias current is a very stablecharacteristic.The transistor is used in conjunction withthe diode to raise the impedance level pre-sented to the meter for proper damping. Thediode current passes through the junctiontransistor with a constant efficiency of about

98% regardless of resistance changes. Thiscurrent transfer factor, known as alpha, isvery stable and therefore does not contributeany significant variation i n accuracy. Theaction is somewhat analagous to the unityvoltage gain characteristic of a cathode fol-lower circuit which also presents a large im-pedance ratio between input and output cir-cuits.As seen in F igure 9, the transistor imped-ance is located in one arm of a bridge. Hencethe bridge can be brought to balance by vary-ing the transistor impedance by means of itsbase voltage. This is done during the initialadjustment procedure with the SET ZEROcontrol. This does not affect the 98% effici-ency of current flow through the transistor.PrecautionsSeveral points of technique in handlinglow-level radio frequencies become of par-ticular importance when checking the cali-bration of a signal generator. RF voltageleakage out of the signal generator, some-times along the power cord, will cause troubleif the receiver is not well shielded. L ikewise,interfering signals from adjacent equipmentor broadcast transmitters wi ll affect poorlyshielded receivers and prevent accurate meas-urements.The conditions of impedance match andcorrections for standing waves on outputcables must be accounted for before the per-formance of a signal generator can be evalu-ated. The connections between the outputcable from the RF V oltage Standard and thesignal generator to the receiver input shouldbe as short as possible. T he insertion lossofany matching pads must be included in thecomparison.Sharp receiver response will cause criticaltuning and stability problems, and will passonly the low frequency components of thenoise which make the meter bounce. A widerpass-band will produce a higher, but muchsteadier, noise level to which the desiredsignal is added.Tune only the signal generator whensearching for maximum receiver response toavoid changes in the receiver input con-ditions.A lways check the signal generator tuningwhen going from the condition of high levelinto the RF Voltage Standard to the lowlevel into the receiver. I t is sometimes ad-visable to re-tune the signal generator fre-quency each time the low level output isre-adjusted i n order to get significant results.When fi rst placing the RF Voltage Stand-ard in operation, it is advisable to re-checkthe SET FU L L SCAL E and SET ZERO posi-tions. Initial drift can be caused by changesin battery voltage when the instrument isfirst turned on and by changes in the resist-anceof the transistor due to a sudden changein temperature, such as bringing the instru-ment from storage into a warm laboratory.There is no significant heat developed insideof .the instrument. Re-adjusting the SETFULL SCALE and the SET ZERO controlsrestores the calibration accuracy of the instru-ment even though the transistor and diodedc resistances may have changed.

6


7/8

T H E NOTEBOOKSummaryBy judicious use of the RF Voltage Stand-ard Type 245-A it is possible to check thelow and high level calibration of signal gen-erators over a wide range of frequencies, andto establish signals for testing receivers atthe microvolt level with a confidence notformerly possible.

Bibliography1. A n RF V oltage Standard Supplies aStandard Signal at a Level of One M icro-volt, C . G. Gorss, BRC Notebook No.5, Spring 1955.

2.

3.

4.5.

Radio Frequency Resistors as UniformTransmission L ines, D. R. Crosby andC. H . Pennypacker; Proc. I .R.E., Feb.1946, p. 62.Signal Generator and Receiver Imped-ance-To Match or Not to Match, W .C. M oore, BRC Notbook N o. 3, Fall1954.Standards on Radio Receivers, Insti-tute of Radio Engineers.Accurate Radio Frequency M icrovolt-ages, M . C. Selby, Transactions ofA IEE, May, 1953.

C a l i b r a t i o n of t he I n te rna l Resona t ingCapac i t o r of t h e Q M e t e rSAMUEL W A L T E R S , Editor, T he Notebook

Q METER TO BE CALIBRATEDSHIELDED COIL

F igu re 7 . Interconnections of equipment that can be used in the calibration o f the in-ternal resonatina caDacitor of a Q Meter. Here shown are Q Meters Type 260-A and aGR prec is ion cGaci tor Type 722-0.Recently we have received a number ofinquiries on this subject. They are numerousenough to indicate a wide-spread interest inthe technique of calibrating the InternalResonating Capacitor of the Q Meter. Thisinterest is understandable since the versatilityof theQ Meter in performing a host of func-tions besides measuring QF depends, i n somespecial cases, on the additional accuracy ob-tainable from an error curve for the InternalResonating Capacitor.The Q Meter contains (1) an RF oscil-lator, (2) a measuring circuit includingthe main and vernier tuning capacitors (I n-ternal Resonating Capacitor), (3) a vacuumtube voltmeter and ( 4) a system for inject-ing aknown amount of the oscillator voltagein series in the measuring circuit.The Internal Resonating Capacitor is usedto adjust the value of capacitance so that thecircuit under test can be resonated at themeasurement frequency. Calibration of thiscapacitor should be done at a relatively lowfrequency wi th respect to the instrumentsoperating range in order to prevent strayinductance effects.The calibration method described here isbased on substitution of a known amount ofcapacitance from a precision capacitor for anindicated amount of capacitance in the QMeter, using a resonant circuit on a secondQ Meter for the comparison.

Equipment Required1Q Meter to be calibrated-BRC 160-Aor 260-A (referred to as No. 1).1 Q Meter (B RC 160-A or 260-A ) usedas an Indicating unit (referred to as No.1 Precision Capacitor wi th a range cover-ing at least 600ppf (G .R . 722 or equiva-lent).1 Shielded Coil that will resonate between

2) .

200-500 K C.Preliminary CheckBefore beginning calibration it is advisableto inspect the Internal Resonating Capacitorto be calibrated. A quick check of the fol low-ing points may save a needless repetition ofcalibration and avoid waste of time since theinstrument can not be calibrated properly ifany of these mechanical conditions prevail:(1) Examine capacitor for foreign mat-ter, specks of dirt, etc. Such matter tends tolower the Q of the capacitor by introducinga spuri ous resistance across it.(2) Check main bearing of rotor sectionsof both main and vernier capacitors. Shaftsshould be firm to prevent mechanical back-lash or electrical instability.(3) Check spring gear take-up of bothcapacitors. Improper loading of gears wi llalso cause backlash.(4) M ake certain rotors are centered wi thstators. Check plate spacing visually. Runout rotor plates to notice any wobble.

Procedure o f CalibrationA Calibration of main Q Capacitor(1) Set the QMeter, N o 2, which is tobe used as the resonance indicator, to450ppf and turn on the power M ounton the instrument a suitably shieldedcoil that will resonate between 200 kcandgookc such as the 103-A32(2) Connect the precision capacitor tothe H i and Gnd terminals of the indicating Q Meter through a short pieceof coaxial cable N ow set the capacitorin the indicating Q Meter, N o 2, to theminimum value of 30ppf(3) Connect the grounded terminal ofthe precision capacitor to the G nd term-inal on the Q Meter being calibrated(N o 1) with a N o 18stranded copperwire A rrange another lead from theinsulated terminal on the precision capacitor to apoint in air Y8to Y2 abovethe Hi capacitor terminal post of the QMeter being calibrated (N o 1) usinga no 20 A W G bare tinned signal conductor copper wire The tip of this self -suspended lead must be straight, withouthooks or loops, and must point downto the Q Meter terminal Isolate thislead from surrounding objects(4) Set the precision capacitor to 600-ppf or more Now adjust the oscillatorfrequency control of the indicating QM eter, N o 2, for amaximum indicationof Q Resonance will occur at a lowerfrequency than in step 1 Note the set-ting of the precision capacitor, call ingthe reading C,( 5 ) Set the main capacitor dial of theQ M eter being calibrated (N o 1) to30 and the vernier dial to zero Do notenergize this Q Meter(6) Touch the suspended lead, movingit as little as possible, to the Hi terminalpost on the Q Meter being calibratedand re resonate with the precisioncapacitorThe dif ference between the two recorded readings, C,-C,, plus 0 l S pp f is thetrue capacitance corresponding to a dialreading of 30t(7 ) Using the reading noted in step 4asC1, other values of capacitance on the

Note this reading as C,

r 4

t 3

> + az::+

2 0

z

L

G- Ii e - 2( 0 - 3. A> 4

- 4

0 100 2W 300 400 500DIAL RE A DI NG LUC

Figure 2. Correction Chart,

7


8/8

B OONTON RA DIO CORPORATION THE NOTEBOOKQ Meter main capacitor can be checkedby successive settings of the unknowncapacitor and the precision capacitor asabove to obtain new values of C,.B. Calibration of Vernier Tuning CapacitorThe same procedure is followed as aboveexcept that the range of the precision capaci-tor must be expanded to obtain greater accu-racy of calibration. T he main tuning capacitoris left at 30ppf, and the vernier capacitor ismoved successively from 0 to f l , +2, +3,and -1, -2, -3. The amount of changeon the precision capacitor dial necessitated ineach case for resonance on the indicating QMeter represents the corresponding value onthe vernier capacitor.By subtracting the calibrated values fromthe dial readings and plotting the errorsagainst the dial readings as sbown in Figure2, a calibration chart for the main capacitorcan be drawn up. I t is possible through thismethod of calibration to obtain an error curvewhich permits use at an accuracy somewhatbetter than our specified tolerancel, depend-ing of course on the skill of the operator and

the accuracy of calibration of the precisioncapacitor used. A similar but expanded chart(since the actual error wil l be in tenths) canbe drawn for the vernier capacitor..* See lead arti cle in Winter, 1955 issue ofNotebook on A Versatile Instrument-TheThe Q Meter by L . 0. Cook.f The VTVMadds about 0.15Wf when themeter i s energized for normal Q Meteroperat on.$ Specif ied accuracy i s plus or minus lwffrom 30 to 100 Wf and plus or minus 1 %abwe I O 0 pp .

A NOTE FROM THE EDITORW e have noticed, as the publication datefor each issue of TH E NOTE BOOK drawsnear, that members of our engineering de-partment pause when passing the editorialsanctum on the way to the water cooler andgape over our shoulder at the three-inch layerof chaos sp rad over the desk. This we charit-ably attribute to the engineers curiosity con-cerning the mysterious journey of The Note-book to the printed page, (rather thanwonderment as to why anyone would getpaid for doing that sort of thi ng), and wefeel that our readers, members of the samegenus, might also be curious, if not in themechanical process of preparation of T H ENOT EBOOK , then certainly in the mysteryof why another group of engineers shouldbeso interested.Dispensing with a description of theblood, sweat and tears generated by the au-thors in the course of their creative labors(many of our readers are painfully familiarwith the picture), we will begin the journeyat the point where the copy is ready for type-setting. T he type for T H E N OT EB OOK isset by a monstrous machine called a L ino-type, which spews castings, or slugs, each ofwhich corresponds to a line of type. T hismachine also has the ability to make an evenright hand margin by regulaitng the spacingbetween individual letters and words.When the copy has been linotyped andedited as carefully as time and the humanfactor permit, it is cut up and pasted in pageform on large sheets of paper. T he largertype used for headings is set by hand, usingcommercially available pads of paper letters.

Glossy photostats of the line drawings arealso pasted in position.When our eight repro pages are ready,we take them to the offset printer, who pro-ceeds to photograph them with a camerawhich is roughly half the size of a masterbedroom. I n this photographic process hereduces the size of our repro pages, whichhave been arbitrarily set up 10% larger thanthe final page size. The developed negativesare then placed over a light box and allextraneous lines, marks and paste-up detailspicked up by the camera are removed bymeans of opaquing fluid. T he negatives arethen carefully laid out in two rows of foureach on a large sheet of paper, each pagehaving a special position with respect to theothers. Thi s operation is called stripping.A large plate of thin, sensitized metal is thenexposed to light through this bank of nega-tives. When the plate is developed the ex-posed areas retain a.greasy substance whichattracts and holds printing ink. T he plate isnow mounted on the cylinder of the offsetpress and rotated, fi rst against ink rol lerswhich deposit ink only on the greasy areas,then against a large rubber roller which, inturn, transfers the ink to the paper passingthrough the press.Each sheet of paper, when printed on bothsides in what the printer terms a work andturn sequence, contains two complete copiesof T H E NOT EBOOK . These sheets are thenfed into a folding machine which simultane-ously folds and creates the glued binding.There remain only to cut, trim and punchthe loose-leaf holes and our NOTEBOOK isready for shipment.

n

AL BA N Y. N e w Yo rkE. A. OSSMANN 8 ASSOC., INC.28c Weis RoodTe lephone : 8 -5833AL BU Q U ER Q U E, N e w Mexico

W EEL Y EN TER PR I SES107 Washington Street, S.E.Tslephone: 5-5586ATL AN TA, G e o r g io

B l V l N S 8 C A L D W E L L26 7 East Pacer Rood , N. E.Telephone: Ch erokee 7522

BOONTON, New Je rseyB O O N T O N R A D I O C O R P O R A T I O NIn te rva le RoadTelephone: Dserfio l d 4-3200E. A. OSSMANN 8 ASSOC., INC.43 Greenfie ld Avenue,Hamburg, No w YorkTe lephone: Fron t ie r 9563

BU FFAL O , N e w Yo r k

C H I C AG O 40 , I l l i n o i sAL FRE D CROSSLEY ASSOS. , INC.4501 N Rovenswood AvenueTelephone: Uptown 8-1 141

D A L L A S 9, Te x a sEA R L L I PSC O M B ASSO C I ATESP.O. Box 7084Te lephone: Elmhurs t 5345

D A Y T O N 9, OhioAL FR ED CROSSLEY ASSOS., INC.53 Pork AvenueTelephone: Oxmoor 3594H I G H PO I N T , N o r t h C o r o l i n a

B I V I N S 8 C A L D W E L LSecur i ty Bonk Bu i ld ingTe lephone : High Po in t 3672HOUSTON, Texor

EAR L L I PSC O M B ASSO C I ATESP.O. B ox 6573Te lephone : Jackson 4 -9303

NEW H AVEN , C o n n e c t i c u tROB ERT A. WATERS, INC.1150 Whalley AvenueTe lephone : F u l t o n 7-6760

PH O EN I X , A r i r o n oN E E L Y E N T E R P R I S E S64 1 East Missouri AvenueTe lephone : Cres twood 4 -5431P I T T S B U R G H 19, P e n n r y l v o n i aH.E. R AN SFO R D C O M PAN YGrant Bunldmng

Te lephone : Gran t 1-1880R O C H ESTER 10 , Ne w Y o r kE.A. OSSMANN 8 ASSOC., INC.3 Juniper StreetTe lephone : Cu lve r 7640

LO S A N G E L E S 46 , C a l i f o r n i o SAC R AM EN TO 14 , C a l i f o r n i oN E E L Y E N T E R PR I S ES , N E E L Y E N T E R PR I S E S3939 Lonkersh im Bou leva rd ,Telephone: Stanley 7-0721

1317 15th StreetI Nor th Ho l lywood , Ca l i fo rn ia Te lephone : Gi lbe r t 2 -8901

B O O f i O N R A D I O

SAN D I EG O 6 . C o l i f o r n i oN E E L Y E N T E R P R I S E S1029 Rosecmns StreetTelephone: Academy 3- 8106

SAN FR AN C I SC O 18 , C o l i f o r n i aN E E L Y E N T E R PR I S E S2830 Geory Bou leva rdTelephone: Walnut 1-2361S T . P A U L 14, Minnero toAL FRE D CROSSLEY ASSOS. , INC.2388 U n i v e rs i t y Avenue

Telephone: Midway 5-4955SYRACUSE, New York

E.A. OSSMANN &ASSOC., INC.308 Merr i t t AvenueTelephone: Syracuse 9-3825B A Y L Y E N G I N E ER I N G , L T O .F i r s t Street,A j o x , Ontar io , ConadaTs lsphone : Emp ire 8 -6866 (Toronto)

TORONTO, Ontario, Canoda

A i o x 118W AL TH AM 5 4 , Massachuset tsROB ERT A. WATERS, INC.4 Gordon Street

Telephone: Waltham 5-6900

8 Printed in U.S.A.

sweep.signal.generator.1955

Documents