Effects of the Cathode Capacitor on P-P Output Stage - Tube CAD ...

Fig. 4. Same measurements as Fig. 3 withdifferent pair <strong>of</strong> tubes in <strong>the</strong> amplifier.These curves are plotted a different scale.equal to <strong>the</strong> sum and difference <strong>of</strong> <strong>the</strong>two input frequencies, or <strong>of</strong> one inputand harmonics <strong>of</strong> <strong>the</strong> o<strong>the</strong>r, etc. Thisaction is <strong>the</strong> basis <strong>of</strong> our common system<strong>of</strong> amplitude modulation, where <strong>the</strong> production<strong>of</strong> such sidebands is a necessary feature <strong>of</strong><strong>the</strong> process. In <strong>the</strong> commonly used (andmost familiar) method <strong>of</strong> intermodulationtests, a high-frequency tone (<strong>the</strong> carrier) ismodulated by a low-frequency tone, and<strong>the</strong> resulting side-bands detected andmeasured by a suitable circuit.Although this method is an excellentindication <strong>of</strong> nonlinearity which generallycorrelates closely with listening tests, itwas found that ano<strong>the</strong>r method gaveresults which sometimes seemed to correlateeven more closely. This secondmethod <strong>of</strong> intermodulation distortionmeasurement also utilizes <strong>the</strong> fact thatdifference frequencies are produced. In thismethod, generally called <strong>the</strong> CCIFmethod, two high-frequency tones <strong>of</strong>equal amplitude and separated in frequencyby a fixed amount. are applied to<strong>the</strong> device under test. The output ismeasured by a wave analyzer tuned toone <strong>of</strong> <strong>the</strong> distortion frequencies such as f2– f1, 2f2– f1 or 2f1 – f2, and <strong>the</strong> distortionvoltage is expressed as a percentage<strong>of</strong> <strong>the</strong> sum <strong>of</strong> <strong>the</strong> two input voltages.Figure 6 shows <strong>the</strong> two types <strong>of</strong> intermodulationin graphic form. One noticeablefeature <strong>of</strong> <strong>the</strong> CCIF type is <strong>the</strong>location <strong>of</strong> <strong>the</strong> difference-frequency componentf1 – f2. Since this component is s<strong>of</strong>ar removed from <strong>the</strong> two input frequencies,<strong>the</strong>re is less tendency for <strong>the</strong> signalto mask <strong>the</strong> distortion as <strong>the</strong>re might bein <strong>the</strong> SMPTE method, or in <strong>the</strong> CCIFmethod for <strong>the</strong> 2f1 – f2. and 2f2– f1 terms.This difference tone is particularly objectionablewhen it falls in <strong>the</strong> region <strong>of</strong>maximum sensitivity <strong>of</strong> <strong>the</strong> ear (400 to5000 cps), and is noticeable in amounts assmall as a fraction <strong>of</strong> one per cent.It must be pointed out, however, thatdue to <strong>the</strong> difference in method <strong>of</strong> expressingpercentages in <strong>the</strong> two methods, equivalentpercentages are not indicative <strong>of</strong>equivalent degrees <strong>of</strong> nonlinearity. TheSMPTE values appear relatively highwhile <strong>the</strong> CCIF values appear relativelylow for <strong>the</strong> same nonlinearity. 2In using <strong>the</strong> CCIF method <strong>of</strong> measurementtwo procedures are commonly used.One is to select two fixed inputfrequencies and measure <strong>the</strong> distortionas <strong>the</strong> power output is varied. The o<strong>the</strong>ris to select a fixed power output level andto vary <strong>the</strong> two input frequenciessimultaneously, maintaining a constantdifference frequency. The first isra<strong>the</strong>r readily accomplished, whereas <strong>the</strong>second requires ei<strong>the</strong>r a special oscillatorsuch as <strong>the</strong> General Radio 1303-A, ortwo oscillators calibrated with sufficient ac-Fig. 5. Comparative distortions withdifferent types <strong>of</strong> tubes in <strong>the</strong> output stageshow different results.curacy to enable <strong>the</strong> difference to bereadily distinguished from <strong>the</strong> dial settings.The stock oscillator in <strong>the</strong> UTClaboratories is a decade-type oscillatorwith an accuracy <strong>of</strong> four places, andconsequently is ideally suited to any suchapplication.Two such oscillators are used with <strong>the</strong>simple mixer-potentiometer circuit shown inFig. 7 to provide control <strong>of</strong> frequency andoutput <strong>of</strong> ei<strong>the</strong>r oscillator and <strong>the</strong> overallvoltage output. The isolating resistornetworks prevent any interaction,2A discussion <strong>of</strong> <strong>the</strong> two methods <strong>of</strong>intermodulation distortion measurementwill be found in <strong>the</strong> General Radio CompanyTechnical Publication B-3, The Measurement<strong>of</strong> Non-linear Distortion, by Arnold Peterson.Fig. 7. Method <strong>of</strong> connecting two audiooscillators to provide CCIF type <strong>of</strong> IM signal—without introducing intermodulationin <strong>the</strong> mixing circuit.and consequent intermodulation between<strong>the</strong> two oscillators.The CCIF intermodulation tests wereperformed using both methods mentionedabove. When distortion was measured as afunction <strong>of</strong> power level, <strong>the</strong> followingrelations were taken into consideration :Since two frequencies are involved, acomplex (non-sinusoidal) wave is produced,and <strong>the</strong> indication <strong>of</strong> an ordinaryvacuum tube voltmeter is, <strong>the</strong>refore, notvalid in determining <strong>the</strong> power level by<strong>the</strong> customary formula P=E 2 /R. Fur<strong>the</strong>rmore,since one wave rides <strong>the</strong> o<strong>the</strong>r,<strong>the</strong> peak value <strong>of</strong> <strong>the</strong> two waves mayreach a value equal to <strong>the</strong> algebraic sum<strong>of</strong> <strong>the</strong> individual waves, with <strong>the</strong> resultthat overload can occur for two frequencieswhen each is only one-half <strong>the</strong> amplituderequired for overload by a singlefrequency. (See Fig. 8) Since each wave isonly one-half <strong>the</strong> maximum amplitude, itcan produce only one-quarter <strong>the</strong>power <strong>of</strong> a single maximum-amplitudewave. The power available from two suchwaves without any possibility <strong>of</strong> overloadis, <strong>the</strong>refore, only one-half <strong>the</strong>power available from a single-frequencywave. The curves for <strong>the</strong> CCIF intermodulationtests are calibrated in terms <strong>of</strong>volts as read on an audio-frequencyFig. 6. Comparison <strong>of</strong> frequenciespresent in two different types <strong>of</strong>intermodulation measurements.