Are you sure?
This action might not be possible to undo. Are you sure you want to continue?
LOUIS H. HIPPE~
A simple method for obtaining information about the characteristics of unknown transformers.
I~ ~[AX,' AUDIO IN.sTA!..L~TIO~ the technician is sometimes faced with the problem or determining quickly and witti reasonable accuracy the unknown impedances 01 irunsforrner windings, At other times. the technician may rin himself with a transformer that could be put to good u e, but uniorrunately he is una hie to obtain sufficient data on the impedance capabilitie ilt the transformer to make ir usable in a practical applicarionc A: other times he may find rh ~ he has avail. ble several of the 400cps power rransforrcers 01 the type used ill surplu military equipment. TI e e transformers can often be used in audio installations where the power requirements are not too greac=depending, 01 course, upon the internal insulation of the power transformer but" pee" sheets on winding impe nee for at; lio ervice are not av ilable for this type oi transformer.
Once the impedances DE primary and secondary windings of allY trensiormer arc known the transformer then becomes valuable and usable as a component in construction of new equipment or replacement in equipmer already in use. However, unless the technician experiments by cuenedtry, he is nor apt to know, even in a general sense, just what rubes or other components the; trails, former wi I allow him to march. Since curantitry requires a ot 01 time, and since there is no 10giCJlI place to start, the transforrncr is like y to be relegated to the junk box where it will kick around until it eventually finds its way 0 the ash can. Good equipment can be aved irom such a fare with a Ii tie effort and a minimum oj equipment.
C.lcul1lion Methods
There are everal possible transformer impedance calculation methods and techniques a ailable which wi!. give results or Tea enable accuracy. Although not of the caliber of laboratory measurements, the :olerallces arc accurate enough for average service.
lI.Io·1 au io technicians OWI1 or can harrow II volrohrnrnilllammeter (pre
106J6 Fieser» m«.
¥Ctrlil Hl!I!)~ oOd. Calif,
Ieralily of the vacuum tube rype) : a 1,OOOcps audio oscillator can be built easily. With these two pieces of equip. mcnr plus a few odds and ends, the unknown impedances of any transiormer winding or choke can be quickly computed.
Figflrl! 1 is the schematic of a 1.000 cps audio oscillator which will Drove a valuable asset to the workshop ~nrl laboratory ;4 addition to the specific use about to be described. It i inexpensive to build since parts are held to a minimum and may be selected Ir'orn spare part components.
In making impedance calculations, it is well La remember that One of its constituents i reactance. Reactance of a given coil or transformer winding changes with the frequency applied, Because reactance changes with frequency, it follows that impedance also changes.
We are therefore inrcres ed in ;:. .. oscillator as a source of voltage at 400 to 1,000 cps because :r allows us to obtain a greater degree oI accuracy in making
imped snce checks on a transformer that will eventually be u ed in the voiceirequency ran e. A 6Ocp= test voltage source is somewha 1e$5 accurate especial y ii he transformer has poor response at. 60 cps.
With the meter and test oscillator we can conduct our impedance determining experiment on the assumption hat uoltage ratio is p oportional to the turns ratio and that the impedance of G usindi. g varies as the square of the turns. This is expressed by the formula:
(1)
Where Zt i the known primary impedance.
Z; is the unknowi imped nee, V, i tile known applied voltage. and
VI is the voltage measured across the unknown winding.
Note rae sra ement, where Z, is the imown primary impedance." Tills value [Colltim,ed 011 page 43]
(5 C7
01 01
C, C2 (+
C3 c. O~IDUI
001 001 .DO ,001
C1<
J. 1, ~ll
~ 0
., _ :E c
.~ '!l ~ '" 0
Cl .. :
0 0 0 "
ci N .... ". .~ e
c: 0: '" 0:
~
'"
   Fig. 1. Simple lOOO·cps oscillator which can be constructed readily and which is useful in making measurements of the type described in this (tide, CH is small a.c.vd.c. filter choke; SR, and SR:: are 50ma selenium rectifiers:
T, and T, are 53volt, 1.5 amp filament transformers "back to back."

UNKNOWN IMPEDANCES IN TRANSFORMER
[from page 21]
must, of course, e found first in order to make the iormul. a operative.
J mpedance, being made up 0: resisranee, can be determined by Ohm's Law applied to ! ose circuits I avi Ig Impedauce. The formula ior Ohm's Law in a.c. circuits is:
z
E I
Where E = e.rn.L, in \'01t5
! = current, in amperes .~ = impedance, in ohms.
Procedure
Ad' usc the output of the audio oscillator to about 25 vol on the meter. Apply this voltage to one winding of the transformer under lest and at rhe same time measure the curren drawn. (See Fig. 2.) Since reflected impedance .from
Fig. 2. Schematic of arrangement u.sed for measuring voltage ratio between transformer windings preliminary to
calculating impedance ratios.
secondary to primary under Joad will have an important bearing on the final resul • the secondary of the tran iormer under tesr should carry a oad, This C:'!}) be a resistor, a speaker or a pair oi he dphonesas all example. If a r iSLOf is used it should. e 01 the noninductive carbon type to avoid reflected reactance, Once tWO known alues are found, the Impedance may be calculated from the above formula.
A more accurate method or rnakinc the measurements is to rrear the r imarv of the transformer as it choke, lea v log t ie secondary unloaded fur the mcment, and determine ri e inductance of the winding and find irs reactance, This can be done by subs irution as shown in Fig. 4.
\.djust R so .11" the voltage drop across Jl bala ces and equals the voltage <1 rap across e primnry. The vol !:a.'e drop aero __ the rransforrner pr'inJary is, of course, 1101 due to inductance alone, bur is caused by irs impedance. ~:eas\lre the d.c. resistance of that section of R in vhich [he voltage drop occurs, :hen so ve for the in uctance Of I' e primary
with the Iormuh : L =,!i,. Once the in
~hJ
(2)
DQUARTERS FOR
Po cb:!~ ris for sucerb Drof~ .• sJonal sound Rp.rodu~tion. EaSy c~ating. (!xce.;irianal high teet
i . the iln<Sf &n Magnetic recording. 10 _U, eutpur,"
Complete Systetns :om $548
ELECTROVOICE MICROPHONE :;:650 Response ± 25 db; 40 15,000 cps. Extremelv high ~urpur level  46 <!p. HnPEdaqce setecter.
~:~hr:~;,~~dI'~~~~IJ~07~~~~ "::~~~~
$9t net .EItutroVojce..~ficrop.h.oJUJ uu.d. ud:ui:Jltljt (If Robiu Hood D,U '(JICeIrIS.
MCINTOSH AMPLIF"fERS
Highest quality. effidcm:y",, FuJI riy ami: range. Frequoncy ran~elO to 20,GOO plUS or minus. 0.1 db: 10 tc 200.0.00 ph..s or mlnus .2. db. !...:owcsl phase shift OtS'"clion; lowest noise. level,
50W~2 lor less ttl." % distortlcn
• Conrtpucus Sifi.gle. trco, rat O.g
• 50 watts RNISPeak 100 $249.50 net
20W·2 for tess lnan ,% dIstortl,,"
• CpnHf1uous sfn~le freq. r+3i:lf".g
• 20 watt. RMSPeak 40
 $149.50 n.~
ELECTROVOICE SPEAKE.!< ;tSP 15
t={e.~!....Ierq response 70·13000 .c.p..s., ~ 3dh. M~~lm 1T' 4ns.lyfaneCI~ cower tricot; 30 wafts. Field (!.OI(C'i1:l!Itlon: 51,t4 lb. Alnico V Opng:e Suuk. \1.agnel. .aC3X ir05~O"C( ireo~ne:y: 3OtJO c.c.s. 8'a!!e' cone ie:it?Oatlce: 3~ c.p,s, HI~ frequency cone resonance: no c, .s.
$5;.00 net
n v tAM TUNERS
j:{~~~~ ,0; ~:;t~~r b~;'dvi~; ~i
os ~ 53loa Me. Ft~ f;J:t1a ·lnC'ciDor.!t. ~g A.vfO"J"i't.a ec frequcnCy ccnrrol fo, sfrrphfie...:t F VI ttJ.nl,..~.
From $87,75
PERMANENT DISPLAY E<;uipm'.f Show'; e+ Audio F3ir
I'
• GIVES HIGH QUALITY REVERBERATION EFFECTS WITHOUT EXPENSIVE EQUIPMENT
• SOUND IS NOT DelAYED AND PLAYED BACK, BUT DIES
AWAY NATURALLY!
• THE RATE OF DELAY IS ElECTRONICAllY CONTROLLED
• ENTIRE UNIT OCCUPIES ONLY 14" OF PANEl SPACE
• PRODUCES AN INFINITE VARIETY OF EFFECTS
The basic assembly requires only 14" of panel space [contrcs+ed with the thousands of cubic feel in on echo chomber). lts inexpensive, for costly rewind and tokeup motors with switching ecces
.$Ories and expensivefloor; and studio space are not required. Yet, in terms of program use il will provide exactly the effects clients specify with excellent reverberant qualilie.s in on infinite variety of uses.
* DEVELOPED BY AUDIO FACILITIES CORP., N. Y.
Manufactured and sold exclusively by Tech labs.
Patents applied. for.
HCHNICAl SPECIFICATIONS
Input impedance .. __ Bridging for 600·ohm line
Inpu & output level _ 0 vu
Output impedance 600 ohms/250 II.
Controls _._ .. _._._ __ Single control operation
Reverbero ion lime 0 to 10 sec., continuously vcrleble
Flni.sh ... __ ... _ .. Grey wrinkle. Others to order
Power requirements 117 volts 60 cps single phose 120 wotts
For Litera/ure & Demonstration Records Wrile To:
ductance is known. rhe inductive reacranee (in ohms) of the primary may be found by the formula: XL = 2",jL. Then the impedance. may be Iound by measuring the d.c. resistance of the winding with an ohm meier an solving lor the impedance with the formula:
Z='\jW+X,," 3}
Where Z = impedance
R = d.c, resisrance X; = Net reactance.
Since we are subsrirutine XL (incluctive reactance) lor X n in the above formula, it is assumed that the net reactance is equal to :Ite inductive reactance. This
Fig. 3. IA) Method of measuring voltage ratio of simple transformer. (S) Tapped transformer measurements give sufficient information for complete calculations. (el Example of tapped
transformer.
i nor completely accurate, hut the result is sufficient I)' close to make the transformer usable in many audio applicarions.
With the proper load on the econda ry the impedance reflected back to the primary would be such as to lower the primary impedance consideralJly. For general application this can be assumed
Fig. 4. Method of making measurement by substitution. Signal Generator voltage is kept constant while adjustments are made. The technique is described and formula given in the text.
tv I wer the primary impedance 10 one fourth. This gives the primary impedance at Iii. J!I'1(;"st [rcouency responses at the transjormer.
v ith tI e Impedance of the primary winding known and with the output of the audio oscillator sri I applied to the
primary, but with
removed. voltage
ratio measurements can be made on the .secondary winding, as at A in Fig. 3, and irs impedance d termined by the formula (l). Succeeding measuremen and calculations can be made on any ,number 01 multip e secondary windings.
Tapped Winding C3Ieulat;ons
Now' suppose we have a transformer
. with arapped winding in which the impedance of two sections i known, but the third unknown, as at B in Fig. 3 The im dance of he unknown section can be ca culated from tile! formula:
'Where Z,,=unknown impedance;
Zu = impedance Ot known section and
Z, = total impedance of sections Z~ an Z~.
In order to make our impedance calculations complete 10: a gi 'en transformer, it is not only necessary to know the impedance oi individual windings
nd tapped sections, but also those tappedwindings 011 tile .rom,> core ',1 combih'<llioll. The commercial y made variable impedance transformers have his inforrnaricn in chart form ior easy refer·enee.Such ready information make the transformer more versatile for any given .application and saves the builder much time and many a headache. We can index the impedance for our transiormer in thesame manner. The third am! last formula makes this camp eteness P05 sible.
Suppose we have a transformer in which we ha V" a twosection tapped 'winding haying impedances oi 500 ohms .and 10 ohms respectively, as shown at 'C in Fig. 3.
Substituting in the formula :
( ROo ~ o
2,:=10 lii1) = 3~ O/:ULS.
The impedance oj any other rapped winding combinations C2.0 be calculated 'in the same manner.
Note also rha in this formula it is not necessary to know the primary impedance. The calcu arions deal with only be knewns and unknowns oi the secondary windings. The resultant values "PI> y to rapped windings on the same core; tills fact sl auld be kep .in mind w en making calcularions employing the last formula.
3 H EA D 5: Erase ... Record •.. Playback Designed To Monitor Directly From The Tape
While You Are Recording!
Developed from the famous PT6, the new MAGNECORDER PT63 offers 3 separate heads  Erose, Record, Playback  for MONITORING FROM THE TAPE while recordinq. The addition of this feature aS5UJes finer results by eliminating most recording errors. All heads, contained in a single housing, are individually alignoble and replaceable, and each one is triple shielded 10 ellmincte crosstalk and hum. 71h inch and ISinch tope speeds, easily interchangeable. (3 speeds ... 3 3/.. ", 7 %" and 15" ... are also available at slight addilionol cost). Attractive bloc grain leotherelte over wood construction. $350.00 Net, "lith Case.
PT63J AMPLIFIER
A new single channel portable amplifier which contains a separate record and playback amplifier so that you can monitor from the tope while recording. In addition, 10 watts of audio is provided to drive bolh the 5inch monitor speaker or on external loudspeaker. Response Rot from 50·15kc at 15" /sf!c. The 19" x 7" control pcnel provides a switch to chcnqe equalizalion
for either 7 % II or 15" tope speeds. Switching is also provided for record, playback or bios readings on the
3" VU meter. May be directly rack mounted when removed from case. $387.00 Net, with Case
VISIT THE AUDIOTORtUM
HOT!; 'tI: ... ;. • e the ropi4fy ctu:Gti S m::,ktl coe!ition..," ell p.fiu:r sko·«n '=If'lob.lul I. du~ng .. wi,k,1! tlO';U and ett Ntl, F.O."., ILY.C.
(om. in Qnd.' .. '~il .ur new IOlIn.l d.po,,· ISIIII ••• 0.11 thu. ih:m1 and mony",of. on wot~i"S dhplut' at 011 'itJ!u.
Harmonic Distortion in
IronCore Transformers
T. WILLlAMS* and
R. H. EASTOP;;'"
A discussion of a simple method of measuring total harmonic distortion with accuracy adequate for routine check purposes.
O NE. ~IC~ION upon th~ design of an audio ~!fansformel' IS that rhe inherent distortion be confined to insignificant proporrions. It is therefore necessary to be a Ie to predict the distortion Introduced by a given transformer from data obtained from measurement" upon samples or the core material. This informacion is particularly necessary in conr.ecuon with large output transformers in which flux density may be quite high a the lowest frequency or me pas. band.
. 'ormnUya transformer is driven from a c i rcu i t conraining a th er mionie valve and in con idering the distortion inrroduced it is necessary to distinguish between (a) distorrion introduced by the valve OWUlg to its Own . onlinear characrerisrics and to the fact that it is working into a complex nonlinear load; and (b) distortion introduced by tile transformer owing to the nonlinear cliaracterisric of its core materi Only the latter type oi distortion is here considered.
This particu ar roblem has been inves igared by _ . 'P<i:rridge1 and it may be useful to restate some of his conclusions:
( I) Wnh respect to non inear distoJ'~ tion in uic transformer a circuit consisting of a source (i!l!l';~al resistance r), transformer and load, Fig. I may be replaced
~' the equivalen circuit oi Fig. 2. In tbe case where all circuit elements are linear, :101.;; reduces to Thevenin's Theorem.
(2) If no can~la'1t polarizing current i present; cnly oddnumber harmonics are produced. Third harmonic predominates and if an accuracy oi i per call is accepiable, higher harmonics rna}' be neglected,
(3) The existence of a l:&)ns1aJlt component .of magnetizing ior~e results in an asyrnmetrical hysteresis loop and even as well as odd ilarmonic% appear. If the .peak 6= densiry i 1= t:T<l!l 10,000 gauss {ior
• Chie! E1't;ir.ur, Pori idge Transiormers Ltd., London, £"OioJJ'o.
~. Euqineer, Pa,.,r:d§~ Transjormcrs Lt d., London, EU9/(llId.
1 N. Partridge, "Harmonic dis rtion in .audio ireq ency transfcrmer ," ji! ireless Euoinur, Sept. and Nov., !94f.
18
siliconiron laminations") the harmonics above the third can be n~!ected with less than 5 per cent error.
(4) The percentage harmonic distortion appearing across ;I. transformer in a circuit sum 2S Pi!). 2 is:
VI> = K S~;(l ~:)
Vr T Zt
Where K is a numerical constant. depending upon the core dimensions
S Is a factor depending. upon the core material and the peak flux density
R is the value or Z:r and Z. in parallel
Z f is the primary opencircuit impedance at the Iundnmenral .irequcncy [.
Methods 01 Measurin.g H .. monic Distortion
Methods available f{IT rueasur ing distortion may be classified into three groups:
(1) Wave analyzer methods in which the amplitude (and perhaps be phase, roo) of each frequency component it the distorted waveform is measured directly by a selective circuit. These methods give the fullest inforrnarion but are slow and require expensive and bull")' appara us.
(2) Fundamental suppression methods in which thedistorted velragewaveform is fed througha passive highpass rilter which rejects the fundamental but does not attenuate the harmonics. Such filters are not inexpensive aud one filter is suitable for measurements ar only one frequency.
(3) Fundamental suppression meth.00s in which thedistorted voltage waveiorm is balanced against a pure waveform at the fundamental frequency and rhe difference is measured. 'The~e are several methods of deriving .he pure reference waveform by using special
. rrnnsformers or oilier passive networks, or by using vacuum tube circuits.
Simple Method for Mc>surin~ Total Distortion
A slmple method ior the rapid routine measurement of. distortion has been de
'eloped by the authors and does not eem 0 be widely known, By the usc of inexpensive apparatus total core distortion can be measured quickly under condieions strictly comparable with specified operating conditions.
'The method fulls into category (3) as listed above. Fig"";: 3 shows a modifica ion of (c) in Fig. 2 in which a paralle ne york composed or C" R, and r is connected to the same source of e.m.i. .and a resistance R. is shown across 'he rransiormer primary inductance to represent the core losses.
\ i~b an indicator connected across.
A am! E this will be recognized as Maxwell's Bridge. By atljustmen or C» r, and R the iun amenral voltage across A C can be made equal to that across Be This setting will held whatever the frequency of E so long a. Land R arc CCOns.~R!.
The balance conditions are then; L=R R, C,
and:
R _RR,
1:
r
In praccice a pertec balance cannot be obtained since the nonsinusoidal magnetizing current through: R, causes a distorted voltage waveform across R, (and L) and this caunoc be completely bala need by the sinusoida I voltage across
Fig. 1. Circuit of transformer working between impedances of Z. and Z\.
R. The residual voltage across AD when C" r, and R are a ijusted ior "balance" is t:le total harmonic dlstor ion appearf1'!g across the transformer winding.
AUDIO EN INEERING • APRIL, 1951
The measuring procedure then is as io lows:
(1) Sec R, to we value Zs.2L/(Zs+ Z
.L
(2) SCi E to make the voltage across AC correspond 10 u e desired power in theloadZL;i.e.., (V~)!=ZLWD
(3) Adjust R and r to make the volrage across AB a minimum. This minimum reading is the harmonic voltage figure required, and if ruultiplied by 100/V.tii gi,,'es tile percentage distortion iactor. There is _ 0 need to rea the values oi Rand r :;0 that uncalibrared ,a.ri.lll<le resistors may be used lor hese circuit elements.
Error' of th~ Method
[0 the foregoing brief description it has been implicitly assumed that (1) the voltage source E is free of harmonics, (2) the voltage source has an internal impedance of zero, and (3) the i'balance" indicator has an infinite input impedance and s ray.capaciea ces are negligible. Fortunately these conditions can be closely approached ana in practice the errors introduced are oi .small order. Each source of error is examined separately below.
The conditions of balance ior the Iaxwell bridge contain no frequency
(C)
Fig. 2. Equivalent circuit of Fig. 1. [Al represents transposition of transformer impedance ratio ; IB) represents transposition using Thevenin's theorem; and (C) represents simplification of (8l.
term and ';0 [he bridge once balanced ,f r one frequency remains so t all frequencies for which the numerical :nagriitude 0_ the impedance elements. remain. unchanged, Uniorrunntely the inductance of an ironcored transiormer is considerably Jower at the third harmonic than it is at fundamenra frequency. Thus clearly can it be seen how 'harmonic irequency components of the source e.rn.I, grve a false value to the reading.
To examine the order of magnit de of tiiis error le it be supposed that the
bridge is balanced at the fundamental frequency and so the fundamental voltzg\! across AB is zero. Now suppose that the inductance has a magnitude of L at fundamental frequencyand l;U at third
harmonic frequency. Let l~rJ x E be the. magnitude of tills third harmonic fl'equency present in the sou rce (funda
ental magnitude, E). The in iicatcr across A13 would give 110 reading of harmonic from the source jf SL were zero. If will be seen therefore that the
false reading due to l~O E win be the difference between the fractions
iw(LSL)
o
Fig. 3. Modification of (e) in Fig. 2 to permit 'of measurements of transformer characteristics.
that is
[ j.,(L8L) j",L] n
R, + jw(L8L  R, + j.,L_ 100 E Expressed as a percentage of the 'fundamental across the inductance, this is
jloL t:
R,+ jwL ~
x 100 pe.r caul which simplifies to 1
By means of this expression the false reading is related to the percentage harmonic, ,~, In the 'SOUTce and from a knowledge of the factors the value of 11 can be deduced for a given maximum permissible error reading. 8L/L is the fractional change of' inductance as between fundamental and harmonic when the cwo.are applied simultaneously. uuder this condition it bas been found experimentally that 8L/L is a small fraction of the order of a tenth to a fifth, measurements being taken with a fundamental frequency of 50 cps and third harmonic.
The order of magnitude of 1!w' (LSL)t
can be obtained from the.
knowledge that at any fundamental irequency in the pass band .,L will be numerically greater than 2RJ or at the third harmonic frequency greater than 6Rp which makes w(L~) greater than
. I w' (f.aL)!
5R,;sothar\il+ R/ becomes
'26 \ hich is ne r enough to 5,
The percentage 1/ in the source SlVe5 rise .0 a false reading of magnitude less ilia l 1j25tb of 11. _0 ii a false reading or less than 0.1 DcT cent can be colcrared, the source U1l1S~ no! ccntain more dian 2~{ per cent harmonic,
Er,o, Due to Finite Sou.rce lmpedanee
Let it be supposed that the balanced bridge setup driven from a pure tone source, of impedance Rs. is indicating an h rmonic voltage across ."B_ The readiog acr AB ca . be considered as due !0 a fictitious generator or haTmonies in series with L; as he source! E docs not aii= the volraee across A]3 under balanced condition it is shortcircuited, ~\"ing the equivalen circuit of A in Fig, 4.
T3e true reading of distortion appears across A.B when Rs=O. How is this reading affected when Rs is not zero ?
T:." complete analysis 0': this circuit leads inte some rather tedieusnlgebra but ii the shunting effect of C and R aCT~ Rs can be. ignored and if Lcl > 2RJ when the equivalent circuit can be greatly irnplified as follows:
(1) with respect to a voltage introduced. as .lie e .. the voltage across AB is almost the same as U1C voltage across AD. This is easily seen by drawing the vector diagram of the various voltages.
(2) The arm CBD can therefore ue completely omitted. A 0£ Fig. 4 then reduce; to c, applied to LJ R" and RsiIl series. The indicated harmonic voltage is the voltage across R,. III general, volt
age across R" is R,.
(Rs+R,) + )wL x e,
R}
'Nhen R3 = 0 this becomes
R., ;.iwL xc", rhe true reading Ot harmonic vo tage, The ratio of these two v(j'l:~g~ is R; + j.,L
D _ r> , ' T' To test the magnitude of J' EJ • J\. r /"'L
error introduced by Rs assume as beion, ",L=2R, and also R5=0;)R,.
TI  _,' R1 + J",L
Ie 1TdCtlOll . r then reduces
Rs+RJ . !",L
I + j2) ,. .
to (11. '1 . The modulus or this ratio ;.... or J
is '15 ....
 :: =O.9S, i.e., H R. 1.$ 10 per cent
21 
of R, the error introduced into the reading is err e order of 2 per cent or
For many measurements the 6Ocps line voltage provides a sui ble source
10
with an internal impedance oi a. few ohms.
Error D.." 10 Siray c.,p~cit1lnc,es and fndic.tor Impedance
There is no difficulry in providing as an indica 0: v cuum ~ube voltmeter of substantially infinite input impedance, and as iistortion measurements are of irnnortance onlv at loy v tl'eQuc.nci even large "strays" inrroduce ~egligibie error. For exa rple, 0.0 ",t at 60 cps is aJI impedance or 0.27 megohms. However, the choice 3.[ grounding point on the bridge does need some considerati n. The vacuumrube voltmeter will probably have one 'input terminal which IIIU': be grounded d the best point f r a ground 011 the bridge :is A. The indicator can then 'be connected acr 55 ,Ae, 'while Lite volt ge I .Ie is adjusted to gT"e the appropriate flux riensirv :11 the core, and rhen s "itched to AB while Rand r are adjusted to £:he a minimum read in t, The brill t: must Ll.' driven irom 3 floatinu winding: 011 a transformer and the capaci nee of this \ inding to ground will shunt one or more 0' tbe bridge elements.
A vacuumtube voltmeter with threeterminal input being developed by til" authors will, give greater ireedom in
o.a
Fig. 4. Bridge of Fig. 3 driven by generator of finite source impedance'. Al actual circuit, and (Bl equivalent
circuit.
measurements of thi::. kind. Suppose :be three input terminals are T"" Tu. Te. The terminal T is permanently grou ided and he input im edance between any two terminals is "cry high. The Indica or reads the voltage between T" and Tn irrespective or ":1)' voltage T,,1'o or T»Te.
Cho'.ce of Ccmpanenrs
S. mmiug up, the method offers a means oi rapid measurement oi ((lta\ distortion using 3 minimum nurn er 0,[ inexpensive component and of sufficient accuracy for /I osr purposes,
20
The resistance R, may 'be built up (rom elected compo ition resi tors as the impedance it represents is not k .. nown to a high degree of accuracy. A fixed tubular paper capacitor o( suitable value i used for C J' The resistance R ron)" conveniently he ::L chain of fixed. v ue composition resistor in series with a composition potentiometer giving a coarse adjustment and ,a fine adj'1 tment over a wi le range. Another composition porenriemerer is suirab e: ior r. For distort jon measurements the cl~ments C" r, and R need no! be calibrnted ~ [hey are merely . dju ued _ Iha. he indicator reading is at a minimum.
Strictly speaking. the indicator itseli should g:ive an indication of tile r.rn.s. value of rhe c mplex waveform oi dL,· tortion voltage appearing across AB. Usually the third harmonic predominates and in this case the mean value. as indicated on a recti fiertypc m vingcoil instrument, will 1)(, close to the r.rn.s. value.
If laboratory type decade boxes can be used ill me setup, and a suitable source of variable frequency is available, it may be noted ihat all the other important pa rarneters of transformer performance can be measured with only a small change of setup.
(1) Incremental. f>riJl"I1"}' inductance a ui care loss.
Use' as Maxwell bridge .. Frequency II'S Jbr di torrion measurcrnerrt bur voltage reduc 'J.
(2 P"i7JIllrT lrakagl! inductance and ,oppt'r loss.
Short circuit secomla.ry of component on test and increase frequency lO midd c of passband. U$C as Maxwell bridge. For this measurement it is more COIlvenient to have C, and ,. ill parallel ~1\"1I1g the leakage inductance and equivalent series resistance directly as series alues.
(3) Equi alent sluiut copacitoncc.
Increase frequency well above passband, r nrerchangc C and R. Use as simple capacitor bridge.
Measurements can be made with a polarizing direct current through the windings if a source of. d , c. is connected
to A and D, Of course, the d.c. path througb E should then be blocked with a capacitor and the d,c, supply must be fed through a choke .0 present 'a high impedance to E.
Results Obtained
F j!Jure 5 StOW 1.\\'0 curves taken in tlte manner described, Curve 1 is the measured distortion characteristic of a. typical P. rtridgc High Fidelity output transformer (rype PPO;2) employing 'a h 1"'h ","de siliconiron core. Curve 2 i the same characteristic taken for the
same rransformer hut with a . diomeral core suhstitnted for the sificon iron. It will he noticed that really low percentazes . re ob i ied by 'he usc of the more expensive nickeliron COre marerial wl;ich also greatly increases the hand wi th obral na le w j th 3JIY gi ven decree of primarysecondarv intersectioning. It will be noted. however, that the power handling cap city of the transformer is nor i ncrcascd y the use of the nickel iron core. In oeh cas the percentage disrorrion incrc es rapidly for power levels above watts, This should he exnected as bath m: ierials saturate at abou the same AlIx density.
H.,monic Di~tortion and Intermodulalion
Tr js ow widely recognized rhat intermcdulation di tortion is more impor  tanr han harmonic dist rtion as a C' •. LU5C of pMI' fidelity in au lio reproducing ysterns, In conclusici therefore it may
be ss well to justify :11,; auention paid (0 he measurement of harmonic distor
io .
It is wellknown rhar harmonic disrorrion arises from the nonli neariry or the m zneric prop rtics or cere materials in common use. '[he production oi harmonics is, i 1\ itself, not very ohjectioaa lc since almost all transmitted sounds arc rich in harmonics to various degree i ior example, the note of a violin is chn racterized by the presence of mill har: Ionic irequcnci in proportions which \'01.1')" r different intrurnents and ior rhe same iusrrument
played different y. Hence, the proper
[C ontinucd Q11 page 331
Fig. 5. Typical distortion curves taken on two transformers with identical windings and different core materials. I l ) siliconiron core, and (2) Radiometal
core.
AUDIO ENGINEERING • APRIL, 1951
HARMONIC DISTORTION
(from pa!p 20J
tion Ot harmonic can be v .. ried within wide lirnits without desrroying the rearism 0: the reproduced sound.
However. the effect rpon a compound 'tone or harmonic distortion i.:l·the reproducing system is mor objectionable since ir is :ncvita I)' accompanied by t •• 1! production oi other tones which are not musically related ro the ilUlclamenta1.
These :O;ICS are known variously as "interrnodula ion tones.' "combination tones" and "sum .. anddifference tones." Possibly the vor result of their presence in addi ion to their unmusical character) is the effect on . e clarity 01 reuroduciion. 'Indivicual tones 10 e ditil;ction an become merged, one with another, due to the wide spectrum of sumanddifference tones.
Ir would seem t1i;r. he best way 01 specifying the distortion introduced by a circui element would e to measure rhe interrn dulation distoruon. How.ever, it is 11:0. desirable to express the di torrion by a sin le number which will offer <1 direct co ipar ison between e uiprnerus ; this cannot be done for intermodulation disrortion without making some very arbitrary assumptions. To measure in ermodulation distortion the procedure is:
(1) Feed in two rest tones simultaneously, each of known amplitude and frequency (four numbers .
(2) 11 easure the T.m.S. value of combination tones and harmonics appearing at the. output. This i  usually expressed as. a percen age oi t' e r.m.s, input voltage (one umber).
F ive numbers are therefore required for the complete pecification at intermo u ation distortion.
Total harmonic di ortion, on the other hand. i completely expressed by three numbers; frequency and amp itude of tC3C tone amI r.m.s, value of harmonics appearing in output. The obvious choice of frequency is that of the lower limit of the passband, and the amplitude can be hat corresponding: to maxirnu rated o· !pUI; these ligures mils: be stared anyway, and only one more figure is needed to in icare he harmonic distortion.
:!1CC e rwo types. of distortion are 50 cl ely rel led, re magnitude of the harmo ic distortion i some indication oi the exrent of the more objectionable inrermodularion. and is more easily m ured. K either figure, or course, is irnporta _c in itseli; the measure of electrical distortion must be correlated to the degree oi aural "nuisance value" :0 which it corresponds.
UA Craftsman Is Only As
 ........ ,~, Good As His Tools!"
Benjamin Franklin
You will find the best in recording apparari come from the: .1eeves Soundcrafc Laboratories. Magnetic tape with ten distinct features chat contribute to its higher efficiency and fidelity; an assortment of recording discs to answer every requirementall are backed by the greater integrity a nd. experience of (he Reeves name foremost manufacturer of recording and electronics accessories.
Sounderafr rape is made J in all types and lengths to " accommodare :aJl tape re .. ~ edees,
Soundcrafc recording diSC5 available in • ,,;u:iery cf sizes, ~ single and. double face.
UoOIti(ms i UIPMi'JI to P, 10 m S7.d snm, JHIII Oif 11. H. L
How Good is an Audio Transformer?
N. H. CROWHURST*
A thorough understanding of transformers is essential to the selection of the proper type for any particular application. The fundamentals are here presented by the author to introduce the series on one of the most important components in highquality audio equipment.
T BE>:;: \\'_\S.\ 1'1 £ when audio transiormers were specified simply by turns rano. All inrerstage transiormer would be designated, according to ratio e.g., 3:] or 5: 1. But one component would give much better periormance. although having me 'arne turn  ratio as another. .he rna lern iran former manuir cuircr knows that there is iar more to de igni g a ~ ood t ransiorsner [han jt t putting windings 011 a care 0 as !O have the correct rums rario. Each design must in i. t be suited for the partie lar Job in hand. To help tile prospective user. rhe component is generail}' specified by be circuit ror which ii is intended, e.g. "lO,OOOohm plate to single grid." together with some staremen: at frequency range. Thi~ is much more iniorma rive, but a still further and more detailed understanding (if audio transformers will Enable them to be used to best advt mage in every application.
Auy audio rran Iorrner is essentially :1. matching device, b It no transformer is a perjecr matching device because it introduces i own losses and defects. It is the designer's job to se thar these losses and deicers are kept to suitable proportions. If he has done a good job, then the USer call get the best performance from the component by following the manui durer's recommenda ions as 10 circuit value . But it would not be economic to desi n a di fier t transIorrner to suit Ever)' po sible application or circuit, so a standardized design often has to core r a range oi U ·C$. A perfectly good rran former. connected into a circuit II ited ior i , will give poor results, but correct understanding oi the problem can help rectiiy these deficiencies.
How Big Murl it Be?
A popul r fallacy has been hat audio transformers always follow the. principle "rhe bigger till; better." Sometimes size is essential to quality, bu {or other applications he smaller component is the better one.
All modern tran O~II:er core ma erials have a well defined sarurariou density. Jagnetizatien above it will considerablv distort the current waveform, aid 'he' distortion wil reflect into the
* 82, Canterburv Grove, London, S.B. 27, r.unl'Jur.
circuit. The _ aturation density will correspond io different voltage levels t differeni frequencies. so the handling capacity of a trausformcr depends on rhe frequency considered. A winding that will accept 5 volts ill' ()Q cps before distorring will accept 10 volts at LO cps.
Applying rhis tact to audio power I mnsforrners, a COl1lpOIII!.nt rated to deliver 10 watts over a irequency ra ige down to 60 cps will deliver considerably
ore power if the lowIrequency CUIoff is rai ed to, say, 240 cps. Conversely. a sun ller tr nsformer may be u ed for the same power rating ii a higher cutoff is employed. In practice. the fact that
nly signals of 10\'1 Ievel are needed a the lowfrequency end oi the audio snectru n enables smaller transformers to be employed than would be. poss ible if full output were requir d down 0, say, 60 cp .
Another factor that affects size is the presence or otherwi e of d.c, polarizing'. Where plate current passes through Oll~ winding (not in pushpull), the core becomes polarized. To minimize this polarization. a gap is Lett in the core by the manufacturer, according to the intended current. Polarization reduces the allowable a.c, magnetization, but the gap ill the core reduces the primary inductance unless rhe turns aft increased. 0 to achieve suitableprimary inductance and sati factory a.c, magnetization limits, without excessive insertion 10', rhe pre ence of polarizing current requires more turn on a larger core size,
For input transformersline to grid, or microphone to gridsize is seldom au advantage, the best transformers designed being of "mall size. For interstage transformers, where appreciable voltage swing is required, a slightly larger rransformer rna)' e necessary.
r
c.
Fig. I. Equivalent drcuit of ~udio tr~n.fotllt""
c, = priiaC1~dtance~ ~ : r:!,;:~ i~~~:~:
C!= sC'CCndarv ra =.primary re.sI<S!i1Oce
 capacitance r" e secbndarv
u = interNifld.inp, < resistance
cepacltooce RD e ccee, 10.$$ referred
to primary
For driver rrnnsiormers. where power must be delivered :0 the rid circuit, considerably larger ius are required.
Its Electrical Specific.ation
The ITa fermer's elecrrical properties have. a direct bearill!,: on it; performance, bot me user will not iecessarily be orhere with hem directly. His concern is wi:h the performance oi rhe lini shed article, The principal properties aTC OO\\'n ior reiercncc a Fig. 1, which will : sis in unders anding the behavior n: thl! tran·f rrner under different condirion • and what P' rr each contributes. 1'111' is one way o~ representing the equi alent :~'cuit 0: any audio transrorruer. For convenience, the fact thai :lte transformer provides it stepup or stepdown is not shown in the diagram. This action or the rransiormer may be regarlC4t O!S periecz. it imperfections being presen d in he circuit values of Fig. 1. The egcnd under the diagram explains what each symbol represents.
There are four elemen  hown as
hU:1uug rlie tran mitred signalprimary and ceo dar)' winding' capacitance, primary inductance, and core loss referred to he primary.
There arc' jour elements. shown as in series wirb signal tr nsmis ionprimary and secondary winding resistance.
,.,
co,
Fig. 1. Equivalent circuit lor low· frequency cutoff, in a directcoupled a, f. transformer. IAI Complete .mneemenl: R, = primary source impedance; R,= secondary load impedance. (6) S,mplified equivalent of IAI: R = parallel
combination of R:. R,. and R e
leakage inductance, and interwinding capacitance,
All these circuit elements must be referred to one winding when considering heir effect. If the primary is the chosen reference winding. then element'S actually due to the secondary windinz or its associated circuit must be "referred" in value bv a factor of the turn fa io squared, Suppose the ratio is 3: 1 S[~PuP, then secondary winding resistance is divided by 9 and is capacirancc
multiplied by 9 in referring to the primary. Usually the "referred" winding resisrancesoi 00 h windings are or the ;Qm1! order. but [ e referred windii is capacitance of the high windm i much larger than thnt 01 the low one about !I times for .3: J ratio . Hence the effect of the lowwi ding capacitance can often be ignored, only tha; ol the high one being taken int account.
Insertion Loss
A a middle frequency, he hum rc.actance will be high e: ough to exercise negligible c, ecr, while the series reneranee of 1.. wi I be low enough to ignore, so the transformer is virtually a "T" resister net vor '. the attenuation of which can be calculi ted wit. reference to I!~:ternal circuit vnlucs=also usuallv r  sisrances ouly=at midfr \rem:)'.' This auenuanon, e~_C1 . tI!lS a measure, of pOWEr 10:;"', is known as the insertion loss of the rra .S orrncr.
For input and imerstaze transformers, where voltage tran icr is he irnporta: t feature. lnser aon II)~ i not generally given serio < consideration. But Ior driver and au OUt transformers, where insertion 10' n.eans valuable watts are lost, i must be considered . Insertion 105:' may be expressed in d , when the fractional power ratio corresponding 0 the db figure ~\'cs the proportion of input power reachin the OUtput, Of ;1 may he expressed directly a a percentage. For example an insertion loss oi 1 £Iv represents an efficiency oi a. proximarely SO pe cent,
For a given transformer size, corredly designed, the ratio 01 r, and r~ to Rp is fixed by the geometry of the COIllponent. Var iation or rum and wire
Fig. 4 (left), Equivalt>nt circuit for lewIrequency cutof!'n a """Uel·fed a.I, transformer:
C~~coupling c.opae.tur Fig, 5 (right l , Rearrangemenr of Fig, ~. TO show it as J resonant
circuit.
gauge to suir, var ies windin resistance and referred corel . shunt in the arne proper ion. TIIII_ a winding with tOO icw tum' resul  in low corel shunt, while too many rum' produces excessive win ling re istnnce. Conversely. [rom he user's vie" Din, connecting a transiormer to circui c or 'lower impedance than desizn v rue introduc hi h series 10 s dueto windin resistance, while connection ro righer impeda ICes results in seri u hunt loss  due to core l11agnetizarion.
In generally. n variation oj impe lance resul in, deficiency in frequency respouse :!S well, and he choice of the number (If tUIT.. in design must be a compromise to achieve goodrespcn e to both low and ligh ends oj the spectrum,
.. '\t lo\\' freque:1cies. the ·enect oi L '.
C" C •. and C. i5 11egligi.hle. but ;h. t of
•
~. ~I I II
I I
0 I ~r 1I1I
I I I II
Y I I
I r
~II 1t ! It I
I V II! I II
/1 I I Iii I(
I 1.1 I I UJ III
0 II . I II iI'll
0 .• 0.' 03o.¢ ~o.e. 2 !.<{i~E7om Fig. 3. lowfrequency cur off attenuation characleristic, 6 db
per octave.
•
inductance Lp will become appreciable. For the purpose of comparative rcspouse, the values of. winding resistances I't and r, can generally be ignored a small compared to the circuit resistances R, and R., 1"0 which the transformer is connected. Thus the circuit can c redrawn as at Fig. 2 irom he viewpoint of relative I.f. response. . r
A) the relevant values arc shown, R, being rhe source impednnceei.e., the pial, resi lance of the tube if he rrausformer i connected in \1 plate circuit:
Fig.. 6 (Ief~). Equivalent circuit for hish frequency cutoff in a stepup rransfcrrner , and fig,. 7 (right!. in J stepdown transformer.
R, is he econdary load resi ranee, if any. referred hy turns ratio squared th" primary ·ide. For example, a shun resist r of 1 mcgolun on the secondary of as: I tepup will give it referred R, ui 1 megohm divided by 5' or 25, which i, 40.000 ohms.
The hunting effect of L» must be considered relative to the parallel resultant of R" R. and R". Tin: B) in Fig, 2 hows n simplified theor e ical iorm oi the circuit, where R is the resismncc of RJ, R., and R; in parallel, This circuit produces 1\ simple 6 dl,! oc ave .f. cutoff, of U1C type shown at Fig. 3. in which the 3·db point is the ire,quency where the reactauce of Lo is equal to R. 1 us the cutoff frequency can he mollified by changing either L. or R. As L» and R. are fixed by the rransformer, the remaining po sibilities ior adjusting I.i. cutoff are R, ami R,. Reduction of the eIT ective parallel reo sistance of these two will lower the l.f. cutoff, extending the frequency range.
\\lith a directcoupled transformer, where the core is polarized, the core is gapped so that the highest possible inductance is achieved when the curren ior which it i, designl:d paSSel; through the winding. It is wa~te£1I1 to work the
trnn former r1 n current . E:)"erin r apprt:clablv rrom thi. va lue.
A' direcccounled tr.ln~:ormer operatin in a 1111 li·pull plate circuit does not have i core polartzcd. 0 a mallcr component can be us d anrl the circuit oi Fig_ 2 ca be applied. taking care hal all impedances arc referre :0 either the \'. hole. or half. of be primary windiuz, correctlv. However, the mduc ance "aries ~\'idely will both frequency and level. 50 t:1C respon c . hown t Fig. J will not be ( pplicable. hut the same principle for adjll5tillg l.f. C1:IOh pplics,
Distortion is closelv as iociated with 1.1. rei' nse ecause  it anoenrs 1101'( strongly at this end. rc, s ilun improve l.t, rcspo . cabo reduce distortion lueto harmonics in the m: gnetizing current.
omct.rn ' transformers are ]'Iar. llel Ie ~ to avoid Pi': inc rhc plate current ~thrQuglJ the primary wi.ulin . Since a couplin capaci r is then necc __ arv. its reactance lso ecomes effective at low i.equ('ncies, so the cireui ior!.f. reponse tnkes tJH, i rm shown til Fig_ 4. Here Rt is he eff . ive resistance zivcn
y the p.ate resistance oi the tube and j coup.;n!! rC'i tor in parnllel ; R,. as ieiore, i the second« ry .10:1 I r esisrancc, if any, ~cferrcd to tli" primary. Thi circuit is in the form of a resonant ,ircuit, 'eel: more dearly as redrawn at Fig. S. where R, i in sec!':" wit 1 heresonant components C,· • nd L!', and the re is nces Rp and R, arc h: shunt. For this reason, Itt gel' values oi R, or
mailer value i R. increase the dampillg (or reduce he Q) Di the resonant circul .
Lessthancritical damping results in ~ I.i. peak Jn rill! vicinity oi cutoff. CritiC111, or mnre ancr irical. dam ing~ eliminates any tendency to ;;eak, Choice of suirnble valu s for R,. «" and C r can re tilt ill I '1 po.; ible operarion oi the iransiorruer ill a circuit. :1 will he shown in detail in following artie \!S.
Response .. , High Frequencies
Ate hizh frequencies the effect of Lp, an Co if used, will be n ligiVle, and where any appreciable rran formation ratio is employed. 01 e or rhc wincli:1g capacitances may usually lw neelecied. 'Thus Fig. (j shows the h.i. circuit equi v  ;U"1l for;! stepuri transformer. while
(C"",irtll.'iI ~II />11'1,' .'.il
TABLE"
Pteferred change in Icv,,1 db
From speech to speech From music to speech From speech to loud
starting music From speech 11.\ quid 3tar:illg music
F'rurn speech to interval ignal (Bow Bells)
o ~ to5
+2 to> 3
I
V'o!t~,g., Regulator Tubes
An article 0:1 e use vf nCOIl voltage
regulator ubes by C. 'Tuppin appears in the September' 95{) Tout c III Radio. Techniques [or the stabilieation of both d.e. and a.c. voltages are discussed, and trigger tubes are mentioned briefly. Of impor ance is the table oi r egulator tubes available in France. Some of these IUDeS are capable of stahi tzing voltages as high as 3io vo rs while others have current range> up to 200 rna. Most oi these rubes are not marketed ill this country, although the>' would appear to be quite useful in many applicalions where designers are now iorced to use serie or paral el connection of available rubes.
HOW GOOD IS AN AUDIO TRANSFORMER?
(from page I
Fig. 7 shows that ior a s epdown type. As in the previous sec ion, i is assumed rhar winding resistances are negl.igible in comparison with the circuit impedances. ~
complicating factor is the interwinding capacitance C" so it is usual to take steps 10 eliminare its effect. Some transformers incorporate shields between windings, connected to ground, so that in erwin ing capacitance is replaced h. win ingroshield, i.e, to ground, capacitance, effectively increasing existing winding capacitance slightly.
Where shields are DOL employed, it is
~~~I
"m"'Sl!
I
Fig. 8 ([eftl. Elimrnalion O'f C. iTom Fig. G. and
Fig. 9(right) from Fig. 7.
generally possible to connecr . he transformer so as to' practically eliminate 111t! effec of inrerwiading capacitance. Interwinding capacitance is the capacitance between the layer of turns thar are nearest together in the IWO windings.
hese layers will e turns electr ically ose to one end of their respective windmgs. By connecting the windings into the external circuit so hat the signal potential to ground from one of these ends is zero, the imerwinding capacitance will become virtually the capacitance. from the other windinz to ground. If the ends nearest together in both
~. RUGGED
• C.ONSTRUCTION
• EXCEPTIONAL CAlIBRATIO'N STABILITY
• UNAFFECT£O
BY VARIATIONS IN HUMIDITY
HIGH·FIDELITY i'S 0 quorterly mooo'%;ft e , dovot.d to tho ;,'\iO'O:)II of peopl ..... hO' _"loy Ih. lalt in homo nunicol en eJlot,.mon,. It h G lo(g esi::,e mo;o:clno.. buoutHuilv iIIu"rclcd, ond hondloOrtlolr ~rinr.d on fina pap.cr.
?lIbt;,hcd by MilTON II. 5lEEP£R
Buill io give ears of service wirhour change in ce llbradon, Even witho ut desstcarors, KelloggMidgec Condenser Microphones show 00 detectable drifc in lnborarory ser ice. Proven in use (or
optimum performance ror more chan 15 years.
PRINCIPAL CHARACTERIStiCS
MODE\. A
MODEL II
RESPONSE <Refc:rred 10
1 yolt/dynr/cm:)
± 3 db ',om 20 10 11,000 ± 4 do hom '20 '0 15.000 cp~
CPl., fet dcaed ccvl y wi,hGfw;tMtt1!ltilleofoqn,pen'
wi,".oui 9rU1~... cficv!:r Indd.:nc~ In f,e. field,
55 db
59 db
fLAT TO WITHIN
SEND fOlt FULL !.?EClFlCATlON' TODAY I W1tITE DEPt. ~7·C
SUPPLY COMPANY
Before you spend money for RECORDS, consult
Hiyk;Jidelihj
the Big Maguine ._ f Information about Records and HiFi Equipment for Your Home
BEFORE YOU BUY NEW RECORDS: Moke it 0 hobit 10 ~h.c~ the te'liew.s in H1GH·fJOElITY by such oUU'ooding (.fili" 0"' John C,only, Alfr e d fran~.,n.'ein, C. G. Burlce. and John Indto~. If 0 new retecse h worth buying. you·1t Rnd if li1iled. Als.o, you'lI fjnd it compa,ed wilh other ,ecord~n9' or ~he some, compo~ilion. Thus HIGH·Fl0UITY CO" help you 10 ovoid m.i'tak.es, ond to eheeue the beU 1c:("or~in9 of Gny (ompoiiJion you wont. Playing time Ol1d price of e..ach recerd is )h'Own,
SPECIAL FEATURE: The Spring Issue conloin. C. G. Burke. monumenJal re secrch wotk which r.evie ..... $ eery lP ever mode of Beerhove_n's mualc, In book form, it wou1d occupy 120 pog·es.,
BEFORE YOU BUY HIFI EQUIPMENT: Moke it a hobil la con.ult 1he elaboralely·illu.lraled erttctes in HIGHfiDELITY on reccrdplcvees. slylit tope reeerderr, FM end TV tuners, amplifiers, speoJce;s, and modern decorative ideal lor home inuollol~ons, 0_11 described ;n ncnteebnlcet language.
Mail Coupon Today! Big Spring i.Hue now ready fot moiling.
~ ..  ... .. ..  .. . .. ..
 Mr. Charles fowler, Editor. Hrgn Fidelity Magazine a
• 26 Audio 8uiJdlng, Greco' Barrington, Man. •
Pfecue enter my lubf'!1iplion 10 Hrgh rideIity Moga.zine: :
OS6.oo lor thr .. ,'0:0 ISAve $63 0$,3.00 for ou tOOl ISAVE SI) •
Addre .. _
Add $01 p'r T.or lor Canedo ... .$ r.oo 10,. ~t:. ccvnl,ifl
...... 
winding. have zero igna potential to ground, the effect oi inrerwinding capacitance is completely eliminated, not even adding to efiecti\'c winding Citpacitance, because there is no signal potentia across it. (Any d.c, potential between windings ill nor produce c  pacirance currents.)
When the effect of imerwinding capacitance has been eliminated. the circuits oi Figs. 6 an 7 become those
hown at Figs. 8 and 9 respectively. 1 'ow, as jn the 1.£ case, each circuit ecomes OJ. simple resonant circuit, in which the impedance connected 0 the
low side (:If the transformer is series damping, while that connected 1,0 the hjgh side is shunt damping. Figlt'r" 10 shows the b.f. response due to (a) essthancritical damping, (b) critical damping, (c) morethancritical damping.
Common cases where inadequate damping gives rise to a h.i. peak are iuterstage transiormers operated in :he plate circuit of triode , and output trans[ rrners for tetrode or penrode rubes. Careful attention to the circuit can eliminnre an undesirable peak or loss and produce the best possible. rcspcn e.
FIFTH PRINTING
Just off the press
AND STILL GOING STRONG!
audio
anthology
Paper cover
Distributors write JOI quantity disaoun:
The loog awaited compilation of reprints from early issues of AUDIO E GINEERL G, most of which are unobtainable.
• Amplifiers
• Speakers
Contains 37 articles on the following subjects:
• Phonograph Equipment
 Tone and Loudness Controls
 Noise Suppressors  Dividing Networks
These articles have been oi great interest to readers of AUDIO ENGINEERIPG over the past three years, Assembled in on~ volume, Ihe)' compris. tho most authoritatlo« rejerence work [or th» audio ~ob.b)'isl.
ICUT OUTMAIL TODAY ,
I Book Division, Dept. A9 I
I Radio Magaxines, Inc. I
I 342 Madison Ave., New York 17, N. Y. I
I Sin: Enc1'~tI ii my 0 check 0 money order lor copin of tile I
I Audio Antholoe· I
J N~m. (print CiretuUyl    .......•.............
L!~r~.:::~:::::::::::::::::::::::~i~~~·::::::::·St"tt'::~~~::::~:~_l
The inter al properties oi an audi tr msiormer that t a limir 0 its frequency range :\1 the upper end aTC it leakage inductance ami the winding capacitance of 'the high side (includingad IlLio!.1;!1 eanacitancc connected aero' the Willdill~r  externally, such as gri,1 input ami strays}, The ultimate curoff frequency is inver ely proportional to the squ: re root or the product of this indue alice and capacitance. Both these <1U,1II 'ties increas e with the sice of UlC transformer so it is: de: r ih smaller transformers have an inhcrcntlv hi;!h"r
frequency r:mge. 
• I II 11
I /' 1'
;; I r I \
0  I \
I I j I
.. j I \
f
:J I I! \ 1
~~. II I \
:g I I II
,. I I \
1 II I \
I I II \
0.' 0.' 1 2 ,
".
I')
0 I I III
I I It., 1
I I )... I
I I 1,,\
~ I I I I '\
I I I \
~ t.o ' I \ I
. I I I
~ I I \
0
I I I I \
''l,.'i II I ~
O~ , • >
II,
I!l
0 I J l l
N I 1
I .I I
I "'!.L
~ If'..
c I II <,
'2 to I I ,
~ I I '\
1: 1 1"\
I I '\
ee I I _\
0.' G. 1 t •
IIr,
U:.! Fig. 10. Attenuation responses for highhequency CJ.ltoff; IAt lessthancrltlca! damping; (8) critical damping; and CI morethan
critical damping.
_ uccessfui operation oi any transformer depend af1 reciably upon the circuit in which it is worked, and the ruanujac ure r's snecificaticn docs not wholly determine 'how good any pa rticul r rransforrner is for [he job in hand. illl lc methods oi measuring up the qualities oi it transformer ami of determining circuit modifications {or optimum performance will be discussed in subeque 1l ar icl
This action might not be possible to undo. Are you sure you want to continue?
Use one of your book credits to continue reading from where you left off, or restart the preview.