20 PHILIPS 'fECHNICAL REVIEW Vol. 7, No. 1

THE MEASUREMENT OF PEAK VOLTAGES IN A STUDIO INSTALLATION

by F. de FREMERY and J. W. G. WENKE. 621,.317.726:621.396.712.3

A measuring apparatus is described which is used in a broadcasting studiofor a continuous controlof the alternating voltageswhichare sent to the trans-mitter. In connection with the purity of the broadcast these voltages mustnot exceedcertain limits. Since there is a danger of their doing so, particularlyat the peaks of the recorded sound, specialprovision has been made for con-venientreading offofthe peak voltagesbymeans of a retardation arrangement.

The amplitudes of the alternating voltages grees of amplification for each part of a piecewhich are generated in the electrical transmis- of music is the task of the technician who.seatedsion of sound vibrations must lie between two at a control table, supervises and control" thelimits: the voltages must on the one hand project whole broadcast:far enough above the level of interferences In order to carry out this regulation the tech-(noise), and on the other hand the amplitude nician must know in advance the moments atmust not become so great that a disturbing which peaks may be expected as well as theirnon-linear 'deformation occurs in one of the intensity. In this he is aided by the score whichlinks of the transmission. lies before him and the experience gained at the

An especially important part is played by rehearsals which precede the actual broadcast.these two conditions in the running of a broad- In addition, however, the technician must havecasting studio. In this case the aim is to secure some method of checking whether the regulationthe greatest possible freedom from interterenee he performs i.s having the desired effect. Theand purity of reproduction, while the amplitude desi.red check is obtained not only by meansof the sound vibrations, at least when it is a of a loud speaker which is connected with thequestion of music, exhibits great variations. It line to the transmitter and with which it isis in just these variations that the dynamic possible to determine subjectively whether orquality of the music lies. As an illustration a' not distortion occurs, but also by means of aregistrogram of the sound intensity in a concert 'measuring instrument which indicatés continu-hall during the performance of a piece of music ously the amplitude af the alternating voltageis reproduced in fig. 1. In order to provide that on the line to the transmitter.the alternating currents which are sent from The principle of the measuring instrumentthe studio to the transmitter are sufficiently which was used for this purpose in a studiofree of distortion even at the pea k s of sound, installation previously described in this perio-it will often be necessary to reduce the peaks dical ') is shown in fig. 2. It contains two ampli-artificially by decreasing in a suitable way and fier stages in push-pull connection, followed byat the correct moment the amplification which a rectifier and an output amplifier stage. Theis applied to the microphone voltages. On the anode current of the final stage, which is aother hand in the sof t passages of the music n;teasure of the amplitude of the input A.C.it will be desirable to increase the amplification voltage of the apparatus, is. measured with a,

somewhat in order that the reception of those moving coil instrument with a light pointerpassages may suffer as little as possible from 1) F. de Fr em er y and J. W, G. Wen k e,interferencès. The regulation of the correct de- Philips techno Rev. 6, 139, 1941.

AI

----11 min:__ Fo OUVERTURE LOHEN6RIN

22367

Fig.!. Variation of the sound intensity in a concert hall during the perfor-manceof a pieceofmusic:Overtureto Lohengrin. (From: R. Ver m e u Ie n,The relationship between fortissimo and pianissimo, Philips technoRev, 2,266, 1937).

JANUARY 1942 PEAK VOLTAGES IN A STUDIO INSTALLATION 21

which is built into the control table, see fig. 3.In series with this instrument a second meteris connected, which, together with the actualamplifier connection, is assembled on a panelof the amplifier racks of the studio installation,so that the broadcast can be controlled there

, also. The scale of the instrument is calibratedin decibels and runs from -45 to +45 dB. Thevalue 0 dB thereby corresponds to an input

reason a device has been employed which verymuch facilitates the readi"!'..goff of voltagepea k s in par ticuiar, with which we arechiefly concerned. The rectified signal voltageis not fed to the output amplifier valve directly,but via the time circuit indicated by T in fig.2. This consists of a connection in parallel of acondenser 0 and a potentiometer, the latterbeing composed of a resistance R and a high

-c+ !

+ 40'62

Fig. 2. Connections of the peak voltage meter. T time circuit consisting ofa condenser 0 and two resistances Rand S. For reasons explained belowthe leakage resistance of a blocking-layer rectifier is used for S. A ammetercalibrated directly in dB of the input signal. Dl and D2 differential ammeters.

A.C. voltage of 1.55 VR.M.S., which is consideredas the normal, permissible zero level in broad-casting tr ansmissions over lines with an impe-dance of 200 ohms (transmission power 12 mW).The part of the scale above 0 dB is colouredred as an indication that the A.C. voltáges hereexceed the prescribed limit and that there istherefore danger of distortion.

Upon rapid alternation of the amplitude ofthe input voltage it would be very fatiguingfor the person at the control table to followthe light spot on the scale of the indicator as itcontinually jumps back and forth. For this

Fig. 3. Indicating instrument of the peak voltage meterbui It into the cöntrol table of the broadcasting studio.

resistance S, which we shall discuss later. If anA.C. voltage of constant amplitude is now fedto the input of the connections, the condenserC is charged to a certain voltage and a verysmall current flows through Rand S. The gridof the output amplifier valve thereby assumesa certain negative voltage. If the amplitude ofthe input A.C. voltage increases, the condensero is .very quickly charged to a higher voltageand simultaneour-ly the grid of the output valvebecomes more negative, whereupon the indica-tion of the anode current meter immediatelyadjusts itself to the new value. If, however, theamplitude of the input A.C. voltage now beco-mes smaller again, the condenser 0 can onlyadjust itself slo w 1y, by discharge over SandR, to the new, lower voltage value, so that thelight pointer of the anode current meter movesback only slowly. In this way provision has beenmade that upon the occurrence of a voltage peakthe indicator reaches practically its final value(maximum deviation 1 dB) within 0.005 sec,while it falls back at a rate of only about 20dB per second.The relation between the anode current of

the output stage and the input A.C. voltage,i.e. the calibration of the scale of the indicator,depends in the first instance on the characteris-tics of the amplifier valves used, in particularon that of the output valve. Since with increa-

22 PHILIPS TECHNICAL REVIEW Vol. 7, No. I

sing input A.C. voltage to the whole circuit thegrid of the output valve becomes more stronglynegative, with increasing input A.C. voltage theanode current will decrease. Due to the curva-ture of the ia-vg-characteristic the variation inanode current thereby becomes smaller andsmaller just in the region of high input A.C.voltages in which we are especially interested,so that there as it were the scale is compressed.In order to avoid this undesired effect, a normalresistance is not used for the resistance S ofthe potentiometer in fig. 2, but one which isdep end ent 0 n vol tag e, namely the

Fig. 4. Relation between leakage current and blockingvoltage (blocking characteristic) of a selenium valve,at two different temperatures. (From: D. M. D u i n-k e r. The use of selenium valves in rectifiers, Philipstechno Rev. 5, 199, 1940).

leakage resistance of a blocking-layer rectifier(selenium valve). In fig. 4, where the blockingcharacteristic of such a valve is given it maybe seen that the very high resistance of thevalve in the blocking direction (leakage resis-tance) decreases with increasing voltage andchanges correspondingly the voltage divisionthrough the potentiometer R-S in such a waythat with increasing input A.C. voltage thenegative grid voltage of the output valve in-creases less than proportionally. By giving SandR suitable dimensions a fairly linear dB scalecould be obtained for the indicating instrument.Upon changing valves, etc. the calibration of

the scale is checked in two steps: an A.C. voltageof 1.55 V R.M.S. is applied to the input terminals,and by adjusting the regulatory resistance R2'which affects the grid bias voltage of the outputvalve, the pointer of the dB-meter on the am-plifier rack is set exactly at 0 dB. In order tofix a second point on the scale, by reven ingthe switch H, the grid A.C. voltage of the firstset of amplifier valves is then reduced by afactor which corresponds to an attenuation ofthe input signal by 30 dB. By adjustment ofthe resistance RI the indication of the anodecurrent meter is then set at -30 dB. Since

Fig. 5. The peak voltage meter opened. At the topcentre is a panel for all the amplifier valves with theirgrid resistances, decoupling condensers. etc. Below itthe potentiometer of the time circuit. The blocking-layer rectifier in this part is composed .of three sele-nium valves connected in parallel. The potentiometeris covered with a double-walled metal cap.

only a small anode current flows at a high inputA.C. voltage (0 dB), so that the adjustment ofRI has little effect, the first calibration pointremains practically unaffected in the fixing ofthe second.

Fig. 4 shows that the blocking characteristicof the selenium valve still depends very muchupon the temperature. This dependence can be

Fig. 6. Front view of the peak voltage meter mountedin an amplifier rack. The dB-meter is in the middleand below it the two differential meters with t.hocorresponding keys. The amplifier valves are accessiblethrough the door above.

,JANUARY 1942 J;'EAK VOLT~GES, IN A STUDIO INSTALLATION. '.23

expressed by a temperature coefficient of theleakage reeistance, which coefficient is found tbbe approximajêly independent of the voltage inthe voltage region in question. It i s thus possible

. .to make the calibration of the .instrument hide-pendent 'of the '.temperaturé' by making theresistance R of a material whose specific resis-tance has the same temperature coefficient as'the leakage resistance of the valve. When thatis done the voltage division through the poten-biometer remains unchanged upon variations inthe temperature. Care must now of course betaken that the resistance R and the valve arealways at the same temperature, but this is

. easily ensured by housing them both in a double-walled heavy metal box, which because of itslarge heat capacity makes it possibleIor tempe-rature fluctuations of the sUrFoUl\dings to -in-

,"

flusnee only the temperature inside the boxand not the temperature di. s tri b ti t ion.In .fig. 5, in which the peak voltage meter i!3shown open, -this box has been removed in orderto show the párts of the ·potentiometer ...

The influence of mains voltage fluctuationson the calibration is also eliminated by stabili-zing with.neon lamps the plate voltage which isprovided by the supply apparatus. Moreover,the plate· voltage of the first two amplifierstages can be checked with two meters. Theseare differential meters (see fig. 2) which in nor-mal use indicate the difference in current be-tween the two valves in push-pull connection ofeach stage. By reversing keys which are situatedunder the meters, see the photograph fig. 6 ofthe complete apparatus, the plate current ofeach valve can: however, 'be mea~ured separately ..

'.•

"

, ,