ve 0 - americanradiohistory.com · presents as compared to amplitude -modulated systems. 3 "A.M." AND "F.M." A concept of amplitude modulation (or "A.M.") may be obtained from the

te,..4 4 4 4.4 4 4 4 4, e4 4 4 4 4 4 44 4 4 4 4 4

dtg. 01b.,111 011111

ti

ZoRAriuz

o

mo-WnN

JAWEa 43>ve 0

...

4

The One and Only lA

PUBLIC ADDRESS

HANDBOOK

Jtl1940- 194/

AMPLIFIER

TY11,13191,1!,

25 CENTS

Most complete authentic P. A.Book Published - Written byM. Asch. C.As complete as you would expect to find anypractical handbook-this is how the radioor P. A. man finds the AMPLIFIER HAND-BOOK AND PUBLIC ADDRESS GUIDE.With essential technical data compiled froman exceptionally large number of sources,the volume covers a varied of different sub-jects coordinating every branch of PublicAddress.To actually show the scope and magnitudeof the AMPLIFIER HANDBOOK ANDPUBLIC ADDRESS GUIDE. The contents isfound at the right, showing the materialfeatured in each particular section. A tho-rough reading of the contents shows thecompleteness of this book.

Send 25c., coin or stamps and yourcopy of the AMPLIFIER HANDBOOK

will be mailed postpaid.RADCRAFT PUBLICATIONS, Inc.20 VESEY ST. NEW YORK, N.Y.

Resume of the Contents of the

Amplifier HandbookAND PUBLIC ADDRESS GUIDEFOREWORDINTRODUCTION

Definitions - decibels, frequency,input. output, impedance, etc.

SECTION I-SOURCECarbon microphones (single -but-

ton and double -button)Condenser microphonesVelocity (ribbon) microphonesDynamic microphonesCrystal microphones (sound -cell

types, crystal diaphragm types)cardiold microphonesContact microphonesPhonograph pickups (magnetic

types, crystal types)SECTION II-AMPLIFIERS

Voltage AmplificationDesign of resistance -coupled volt-

age amplifiersCommercial voltage amplifier

The Power StageClass A amplifiersClass AB amplifiersClass AB, amplifiersClass AB, amplifiersClass B amplifiersWhen to apply class A, AB, and

B amplificationPower SuppliesRalf -wave rectificationFull -wave rectificationVoltage doublers

Filter CircuitsPower supply regulation, etc.

Practical Hints on Amplifier Con-struction

MicrophonismPlacement of componentsTone compensationInverse feedbackRemote control methods

SECTION III-DISTRIBUTIONThe Loudspeaker

Dynamic speakersSpeaker performance (frequency

response, efficiency)High-fidelity speakers

Speaker Baffles and HousingsOutdoor speaker installationsPower cone speakersRadial (360. distribution) speaker

bafflesSECTION IV-COORDINATION

Input impedance matchingMiddling speakers to P.A. Instal-

lationsPhasing speakersEffect of mismatching speakers

to amplifier outputA typical P.A. installation (in a

skating rink)SECTION V - USEFUL PUBLIC

ADDRESS DATA AND INFOR-MATION

Speaker matching techniqueThe ABC of Db., VU, Mu, Gm

and SmCharts and formulas useful to

the practical P.A. sound manHandy index to important arti-

cles on public address andsound

I

Li

//ac '5, /6 7-7yi de,

ALL ABOUTFREQUENCY

MODULATION

Selected Articles Affording

A Comprehensive Insight to

The New Art of F. M. Radio

COMPILED

BY THE EDITORS OF

RADCRAFT PUBLICATIONS, Inc.PUBLISHERS

20 VESEY STREET NEW YORK, N. Y.

Copyright 1941 by Hugo Gcrnsback Przilied in U.S.A.

Contents

CHAPTER 1Page

THE ABC OF F. M.Frequency vs. Amplitude Modulation 3Basic Facts About F. M. F

CHAPTER 2

CONSTRUCTIONBuild This Practical F. M. 10

CHAPTER 3

AUDIO AMPLIFICATIONF. M. Audio Amplifier-Part 1 15F. M. Audio Amplifier-Part 20F. M. Audio Amplifier-Part 3 24

CHAPTER 4

F. M. SERVICEPart 1-Antenna Installation and 29Part 2-Receiver Alignment and 38Part 3-Test Equipment for F. M. 44

CHAPTER 5

ENGINEERINGPart 1-The How and Why of F. 47Part 2-The How and Why of F. M. 51Theory and Design Considerations of R. F. and I. F.

Coils In F. M. Receivers. 55

ALL ABOUT FREQUENCY MODULATION

CHAPTER I

THE A -B -C OF

FREQUENCY MODULATION

PART I -FREQUENCY VS. AMPLITUDE MODULATION

Frequencymodulated receiver.

strip, in scientific importance,

ONE of the pioneers inradio has againscored an outstand-

ing success. Major Arm-strong, who will be remem-bered as the Father of thesuperheterodyne type ofreceiver, has again ad-vanced the art by develop-ing a new type of trans-mission and receptionwhich bids fair to far out -

all previous developments.For years the Major worked to perfect methods and means

of eliminating or reducing static and other radio interference-probably the foremost problem confronting the radio in-dustry. After trying various methods of attacking the prob-lem, he had practically given up the idea as he and his asso-ciates could see no practical answer.

Over 10 years ago he started a new line of attack whicheventually led into a system known as the wide -band fre-quency modulation system which reduced all sorts of dis-turbances to a very small percentage of their original value.This was accomplished by transmitting a signal having suchcharacteristics that it could not be reproduced by either nat-ural or man-made static; and designing a receiver which wasnot responsive to ordinary types of modulated waves but onlyresponsive to waves having the special characteristic.

Thus this type of transmission and reception emerges fromthe laboratory and. bids fair to revolutionize the transmissionof intelligence.

It is the purpose of this article to explain what frequencymodulation is, how it operates, and the advantages which itpresents as compared to amplitude -modulated systems.

3

"A.M

." A

ND

"F

.M."

A c

once

pt o

f am

plitu

de m

odul

atio

n(o

r "A

.M."

) m

ay b

eob

tain

ed f

rom

the

diag

ram

of

a hi

gh -

freq

uenc

y al

tern

atin

gvo

ltage

whi

ch c

hang

es it

s am

plitu

de w

ithtim

e, a

s sh

own

inFi

g. 1

. Thi

s is

the

type

of

mod

ulat

ed c

arri

er tr

ansm

itted

bypr

esen

t-da

y m

etho

ds. F

igur

e 2

show

s a

carr

ier

freq

uenc

y, th

efr

eque

ncy

of w

hich

is c

hang

ing

at s

ome

defi

nite

rat

e. T

his

isth

e ty

pe o

f si

gnal

whi

ch is

tran

smitt

edby

a f

requ

ency

-m

odu-

late

d (o

r "F

.M."

) st

atio

n. N

ow le

t uk

exam

ine

tune

d ci

rcui

tsan

d se

e ho

w a

mpl

itude

and

fre

quen

cy m

odul

atio

nca

n ac

tual

-ly

be

effe

cted

."A

.M."

-Ass

ume

that

the

tune

d ci

rcui

t sho

wn

inFi

g. 3

isco

ntin

ually

sup

plie

d w

ith e

nerg

y so

that

the

alte

rnat

ing

cur-

rent

set

up

ther

ein

has

the

sam

eam

plitu

de a

t all

times

. Let

us in

sert

in th

is tu

ned

circ

uit a

mic

roph

one

whi

ch h

as th

epr

oper

ty o

f va

ryin

g its

res

ista

nce

acco

rdin

g to

the

ampl

itude

of s

ound

wav

es im

ping

ing

on th

edi

aphr

agm

. It i

s w

ell k

now

nth

at ir

the

resi

stan

ce o

f a

tune

d ci

rcui

t var

ies,

the

ampl

itude

of th

e cu

rren

t flo

win

g in

the

tune

d ci

rcui

tva

ries

acc

ordi

ngly

.C

onse

quen

tly, s

ound

wav

es s

trik

ing

the

diap

hrag

m o

f th

em

icro

phon

e w

ould

cau

se a

cur

rent

to f

low

in th

e tu

ned

cir-

cuit

such

as

that

sho

wn

in F

ig. 1

."F

.M."

-An

idea

of

the

met

hod

in w

hich

a fr

eque

ncy

-m

odul

ated

sig

nal c

ould

be

prod

uced

may

be

had

from

the

sam

e tu

ned

circ

uit a

nd a

gain

ass

umin

g th

aten

ergy

is b

eing

supp

lied

to th

is tu

ned

circ

uit s

o th

at th

e am

plitu

de o

f th

ecu

rren

t is

at a

ll tim

es c

onst

ant.

Now

let u

s co

nnec

t acr

oss

the

tune

d ci

rcui

ta

cond

ense

rty

pe o

f m

icro

phon

e w

hich

con

sist

s es

sent

ially

of

2 th

in m

etal

-lic

pla

tes,

the

posi

tions

of

whi

chva

ry in

res

pect

to e

ach

othe

rin

acc

orda

nce

with

the

soun

d w

aves

str

ikin

g th

em

icro

phon

e.A

s th

is m

icro

phon

e do

es n

ot c

hang

e its

res

ista

nce

appr

ecia

-bl

y, a

sou

nd w

ave

impi

ngin

g on

the

mic

roph

one

wou

ld s

impl

yva

ry th

e fr

eque

ncy

of th

e tu

ned

circ

uit b

ecau

se th

e to

tal c

a-pa

city

of

the

tune

d ci

rcui

t wou

ld b

e al

tere

d an

d its

fre

quen

cyis

dep

ende

nt u

pon

the

capa

city

. Thi

s ci

rcui

t is

show

n in

Fig.

4. S

ound

wav

es th

us m

ay b

e m

ade

tova

ry th

e ca

rrie

r fr

e-qu

ency

suc

h as

sho

wn

in F

ig. 2

.

.-A

MP

LIT

UD

E

..--

MO

DU

LAT

ION

-

---

.,..

- -IFR

EQ

UE

NC

YM

OD

ULA

TIO

N...

-

sill

I/11

.'11

piG

2F

161

4. 0 .... - - - -

0L

... ..... _ _ _ -

CO

ND

EN

SE

R!M

ICR

OP

HO

NE

0M

ICR

OP

HO

NE

FIG

.4F

IG 3

At

Ian

d 3

are

show

n, r

espe

ctiv

ely,

a r

epre

sent

ativ

e w

avef

orm

of a

n am

plitu

dem

odul

ated

sig

nal,

and

the

man

ner

In w

hich

it m

ay b

e pr

oduc

ed. A

n eq

uwal

enar

rang

emen

t for

freq

uenc

y m

odul

atio

n is

sho

wn

at 2

arid

4.

In3,

mic

roph

one

MI v

arie

s its

res

ista

nce

acco

rdin

g to

the

ampl

itude

of s

ound

wav

es: a

t 4, c

onde

nser

-ty

pe m

icro

phon

e M

2 va

ries

in c

apac

ity a

s th

e so

und

wan

esst

rike

the

diap

hrag

m.

In th

e fo

rego

ing

anal

ysis

onl

y th

eore

tical

arra

ngem

ents

have

bee

n co

nsid

ered

bot

h fo

r th

e am

plitu

de-m

odul

atio

n an

dfr

eque

ncy

-mod

ulat

ion

met

hods

. In

prac

tice,

of

cour

se, m

ore

com

plic

ated

arr

ange

men

ts m

ust b

e em

ploy

ed to

eff

ect e

ither

type

of

mod

ulat

ion

to g

ive

a co

mm

erci

ally

use

ful r

esul

t.S

TA

TIC

FR

EE

A s

impl

e ex

plan

atio

n of

why

the

syst

em is

so

free

fro

mst

atic

and

oth

er d

istu

rban

ces

lies

in th

e fa

ct th

atth

ese

dis-

turb

ance

s, w

hile

they

hav

e ve

ry g

reat

am

plitu

de -

mod

ulat

ion

chan

ges

have

rel

ativ

ely

smal

l fre

quen

cy c

hang

es. H

ence

byde

sign

ing

a re

ceiv

er f

or f

requ

ency

mod

ulat

ion

whi

ch d

oes

not

resp

ond

to a

mpl

itude

cha

nges

and

whi

ch o

nly

resp

onds

wel

l

ALL ABOUT FREQUENCY MODULATIONto large frequency changes practical immunity from inter-ference can be obtained.

Then by designing a transmitter which sends out a widelyswinging frequency modulated wave a full response can beobtained in the receiver. The differentiation between noiseand signal resides mostly in this special characteristic.

Estimates on the effectiveness of elimination of static dueto lightning, etc., vary from 96% to about 99%. It is certainlytrue that frequency modulation signals can be heard rightthrough a local thunderstorm while amplitude modulation

signals are entirely blotted out due tonoise.RECEIVER REQUIREMENT

Limiter.-As frequency modulationis considerably different from the stand-point of its transmission, it is only nat-ural that the receiver picking up fre-quency modulation signals is materiallydifferent in its design.

Fundamentally, the receiver has anR.F. amplifier system which should passa broad band of frequency as will be evi-dent from future considerations. It thenhas a wide -band I.F. amplifier in whichis incorporated a "limiter circuit," thefunction of which is to keep the ampli-tude of the amplified signal constant atall times. This limiter might be likenedto a very rapidly -acting automatic vol-ume control system.

Detector.-The detecting system forfrequency -modulated signals varies ma-terially from the conventional type. Thiswill be appreciated by careful analysisof the method of transmission. If we goback to Fig. 4 and consider sound wavesimpinging on the condenser microphoneit is readily apparent that the amountthe condenser plates of the microphonemove is in general proportional to theloudness of the sound waves and conse-quently the amount of variation of thecarrier frequency is proportional to the

intensity of the sound waves. The num-ber of fluctuations per second which thediaphragm executes is obviously thesame as the frequency of the audio noteimpinging on it. This means that thefrequency of the carrier will swing backand forth over its range a number oftimes per second which is the same asthe audio frequency being transmitted.Thus the detecting system must be ca-pable of producing audio voltages whosefrequencies are proportional to the rateof frequency change of the received R.F.signal; and whose magnitudes are propor-tional to the amount of frequency change.

There are several types of limiter and de-tecting circuits which may be utilized, butprobably the simplest and most effective isa limiter circuit consisting of a sharp cut-off tube and a resistor in series with thegrid of this tube developing a negative biaswhich is applied to other I.F. amplifier tubes.

For the detecting system, an I.F. trans-former somewhat similar to that used forthe discriminator of an automatic -frequen-cy -controlled receiver may be readily em-ployed, for the circuit is so arranged that ifthe intermediate frequency varies, a volt-age is developed across a resistor in thecathode circuits of a 6H6 rectifier tube.

The limiter and detecting systems areshown schematically in Fig. 5. It might bepointed out in passing that the adjustmentof the I.F. transformer feeding the 6H6 isquite critical if good quality reception is tobe obtained.

FIG 5

LIMITER65J7oa 1852

VARIES WasOTHER

I F TUBES

DET. I F.3 MC.

.05 -Ms

Detailed diagram showing the manner in which the "limiter" is connected to prirrent amplitude variationsbeing passed on to the detector circuit.

5

Sch

emat

ic c

ircui

t of f

its G

.E. H

M -

80 ta

ble

-mod

el fr

eque

ncy

-mod

ulat

ed r

ecei

ver

CO

ND

EN

SE

RS

CI -

20 m

mf.

tuni

ngC

2-0.

02.m

f., 4

00 V

.C

3-47

mm

f.C

4-1,

200

mm

f.20

0 V

.C

6-50

0 m

mf.

trim

mer

C14

-47

mm

f.C

15-4

70 m

mf.

C16

-22

mm

f.C

17-1

00 m

mf.

CIS

-10

0 m

mf.

C19

-0.0

35-m

f., 4

00 V

.

C20

-0.0

02-m

f., 6

00 V

.C

21-4

70 m

mf.

C22

-0.0

32.m

f., 6

00 V

.C

23-2

20 m

mf.

C24

-0.0

05-m

t., 4

00 V

.C

25-0

.05-

mf.,

400

V.

C26

-0.0

05.m

f.. 6

00 V

.C

27-0

.005

-mf..

400

V.

C28

-0.0

5.m

f., 4

00 V

.C

29-0

.05-

mf.,

200

V.

C30

a-20

mf.,

250

V.'

C30

b-20

mf.,

250

V.'

C30

c-40

mf.,

250

V.'

C30

d-20

mf.,

25

V.'

C31

-0 I-

mf.,

300

V.

C32

-470

mm

f.C

33-0

.1-m

f. 40

0 V

.C

34-0

.05-

mf.,

200

V.

C35

-2-1

5 m

mf.

ante

nna

trim

mer

C36

-7.2

3 m

mf.

air

trim

mer

RE

SIS

TO

RS

RI -

2,20

0 oh

ms,

R2-

6.80

3 oh

ms,

R3-

47,0

00 o

hms,

lbsW

."R

4.33

0 oh

rns,

1/2

-W"

R5-

0.47

-meg

., 1/

2-W

."R

6-47

.003

ohm

s, I/

2.W

."R

7-0.

47-m

eg.,

1/2.

W."

R8-

4.70

0 oh

ms,

1/2

-W."

R9-

220

ohm

s, 2

W."

RI0

-0.3

3.m

eg.,

1/2-

W."

R I

1-0.

1 -m

eg.,

I/2.W

."R

12-0

.1-m

eg.,

1/2-

W."

R I

3-0.

12-m

eg..

1/2.

W."

RI4

-2-m

egs.

, vol

ume

cont

rol

RI5

-82

ohm

s, 1

/2-W

."R

16-2

20 o

hms,

11/7

2.W

."

R17

-15

meg

s.,

1 2.

W."

RIB

-0.3

2-m

eg.,

1 2-

W."

R19

-0.4

7-m

eg ,

R20

-1,5

00 o

hms,

1/2

-W."

R2I

-2,2

00 o

hms,

I W."

R22

-1,6

00 o

hms.

31/

2 W

.W

. -W

.R

23-2

,200

ohm

s, 1

/2.W

."R

24-0

.1 -

meg

. ,1/

2-W

."R

25-3

3,00

0oh

ms,

1/2

W."

R26

-0.2

2.m

eg.,

1/2-

W."

R27

-2.2

meg

s.,

(Pap

er;

mie

s:'d

ryel

ectr

olyt

ic: "

care

en.)

MIS

C.

SW

. I-T

one

switc

hS

W. 2

-Pho

no s

witc

hS

W.

3-P

ower

switc

h(o

n S

W 2

)L1

-Ant

.; L2

,os

c.,

L3,

1st

1.P

.;L4

, 2nd

I.F

.;L5

-1.

F. l

imite

r: IA

-di

scrim

. tra

ns.


DYNAMIC RANGE; INTERFERENCEIf a large dynamic range is to be had in

frequency modulation transmission, thebandwidth necessary may be 100 kc. or moreeither side of the carrier. Obviously suchbandwidths can not be obtained on presentbroadcast channels and it is necessary forfrequency modulation transmission to re-sort to high frequencies where there isavailable space in the ether. Most frequencytransmissions are thus around 7 meters (be-tween 39 and 44 mc.).

When listening to frequency modulationtransmissions, the dynamic range is very ap-parent. That is, the program may go fromthe lowest (volume) to the loudest passagesof an opera with the same clearness andfidelity. Complete absence of interference isalso very apparent especially during timesof intense static when shortwave amplitude -modulated stations are completely blottedout. There seems to be very little fading.This may not be entirely due to the methodof transmission but partly because of thefrequency of transmission. Utilizing a well -designed receiver, a few microvolts signal isall that is necessary for good reception.

There can be no interference from onestation to another such as encountered inamplitude -modulated signals! If 2 frequencymodulation stations some distance apartwere to transmit on the same frequency andif a frequency modulation receiver weregradually moved between the 2 stations,either one or the other would be heard.There might be a "no man's land", so tospeak, where the program of one station andthen the program of the other station wouldbe heard, but there would be no interferenceas we know it.

It would appear that frequency modula-tion will, in the years to come, come intoits own and in a large measure supersedethe present method of shortwave transmis-sion. It certainly has in its favor programenjoyment which can not be obtained by theolder methods.

Technicians may wish to refer to one ofthe following published articles:

(1) "A Method of Reducing Disturbancesin Radio Signaling by a System ofFrequency Modulation," Edwin H.Armstrong, Proc., Lli'.E., Vol. 24, No.5, May, 1936.

(2) "New Ears for Your Radio Set andNew Fortunes in Radio." Ken, June29, 1939.

(3) "Revolution in Radio," Fortune, Octo-ber, 1939.

(4) "Science Tunes in a New Kind ofStatic -Free Radio Service," ScienceNews Letter, June 10. 1939.

To complete this reference material forServicemen the following additional infor-mation is given.

I.F. ALIGNMENTDue to the good stability of components

and the wide -band characteristics of theI.F. circuits, alignment should be unneces-sary under normal operating conditions.Should I.F. alignment become necessary, itwill require a cathode-ray oscilloscope anda 2.1 megacycle signal generator with asuperimposed ± 300 kc. sweep frequency.

This generator may be made by construct-ing an oscillator with the tank condensersemi -fixed and variable, the variable portionbeing designed to rotate with a motor andof proper capacity to give ± 300 kc. varia-tion of the 2.1 megacycle mid -frequency.

Connectthe vertical plates of the oscilloscope acrossresistor R10 and align transformers L5, L4and L3 progressively. A 2 mh. choke shouldbe connected in series with the high sideof the oscilloscope.

With the same oscillator and sweep signalas used above, connect the vertical oscillo-scope plates across resistors R11 and R12.Align transformer L6 for an x -shaped cross-over curve. Proper alignment of C13 is indi-cated when the curve crosses about midwayin the vertical plane. Proper alignment ofC12 is indicated when the sides of the curvenear crossover are nearest to a straightline.

Note: Keep signal input high enoneth sothat noise limiter is functioning. This pointis indicated when an increase in signalinput no longer changes the size of thecurve.

R.F. ALIGNMENTMake sure the dial pointer coincides with

the first division on the low -frequency endof the dial scale when the gang conden,eris completely closed.

(1) Connect a 0-50 or 0-100 microamp.meter in series with the low end of RIO.A high -resistance, 0-10 V., D.C. voltinet, ?-may be used instead of the microammeter.Connect the voltmeter across R10 with2 mhy. choke in series with the high side.

(2) Apply an unmodulated signal in theregion of 43 megacycles to the antenna.

(3) Adjust pointer so it is set to thescale mark of the signal used and peaktrimmers C36 and C35, in this order. foimaximum meter reading.


BASIC FACTS ABOUTF -M BROADCASTING

THE progress of Frequency Modulation("F.M.") as with anything that is newand not fully understood-has givenrise to a number of common fallacies,

widely spread by omnipresent pseudo -ex-perts who do not grasp the picture quite sofully as they believe they do.

Many of these fallacies deal with thecapabilities and limitations of F.M.; othersseek to anticipate public reaction. Most ofthem are sheer conversation pieces. All ofthem bear refutation, in light of the re-markable growth that has attended the newnoise -free, full -fidelity method of radiobroadcasting during recent months.

Here, for example, are a few representa-tive misconceptions about F.M. that havegained erratic circulation.

F.M. stations can't be heardmore than so miles from the trans-mitter. Therefore they can't begin toservice as great an area as the regularamplitude stations. It will take many,many more stations to cover as great aterritory as that reached by the majorstandard stations today.This is a common example of misinforma-

tion. The coverage area of an F.M. stationis based on a combination of 3 factors:

(a) The height of the antenna above thesurrounding countryside;

(b) The power used at the transmitter;and,

(c) The type of antenna employed.Service ranges of 100 to 125 miles from

the transmitter are quite possible, and manyof the applications now pending before theFederal Communications Commission will befor such service areas. The range of an F.Mstation is the same by day and night-anunvarying, unfading signal of remarkableclarity. Very few 50,000 -watt stations ofthe ordinary type reach a greater area withconsistency during daytime hours. Thenight-time coverage is greater, of course,but marred by fading, static and cross -interference beyond the primary coveragearea.

(2) P.M. networks are impossiblewith the use of telephone wires becausethese wires won't carry the high-fidelity

notes that F.M. demands for full -nat-ural quality. Therefore the use ofradio -relay --small transmitters placedat intervals across the country to carryprograms from network station to net-work station-is the only answer. Thiswould be very expensive and there is noproof that it might be satisfactory fora coast -to -coast hook-up.Wrong again. Telephone wires can carry

the 30 -to -15,000 cycle range of tone de-manded by F.M. stations. They can carryeyen much higher ranges. Such telephonelines do not exist widely at present becausethere is no great demand for them. But thephone companies stand ready to supply thissuperior service when the demand is strongenough to warrant the installation of suchnew facilities.

The development of F.M. networks on anationwide scale, co-operatively run, is ex-pected to start within another year or two.By that time the telephone companies willprobably have the new, full -range wiresready for use.

(3) The public has a "tin ear." Thepublic can't tell a high note from amedium one. Furthermore, the averagehearing doesn't register above 10,000cycles, so why bother with a lot of fancyequipment to bring in notes as high as15,000 cycles? "High fidelity" doesn'tmean anything, because the averageA.M. set today can't reproduce notesabove 5,000 cycles anytePv.This let -well -enough -alone attitude is a

poor argument. The public has a ,o -called"tin ear" only in that it has never knownwhat natural, full -fidelity radio can soundlike. Experience shows that average lis-teners, after hearing F.M. for a period ofa few days, are acutely aware of a flatnessin standard broadcast reception when theyreturn from F.M. to A.M.

The fact that the average hearing doesnot go above 10.000 cycles is no indicationthat the ear does not catch and appreciatethe many overtones created in this airyregion of the sound spectrum. It is here thatthe illusion of color, depth, extreme natural-ness is created. It is further heightened by

ALL ABOUT FREQUENCY MODULATION'

the fact that F.M. has no "carrier noise."There is no rushing sound when voice, ormusic are not present on the wave, as instandard broadcasting. F.M. is completelysilent. The faintest innuendoes of tone arenot muffled in this everpresent backgroundrush.

(4) It's proof, say the F.M. scoffers,that the public doesn't want or appre-ciate high-fidelity, since surveys showso many listeners leave their tone con-trols on the "bass" position. This cutsout the treble notes that occur uparound 8,000 cycles and above.Actually it proves nothing of the kind.

It merely shows that the average listeneris instinctively aware of the backgroundrush in amplitude or "A.M." broadcastingwhich becomes definitely prominent with thetone control at "treble." By reducing thetone control to "bass" all the highs, badlydistorted through the rushing background,are eliminated and the listener has a nearer(albeit lopsided) approximation of the real,natural thing. True "high-fidelity" does notplace any emphasis on either bass ortreble. High-fidelity reproduces preciselywhat the microphone hears, with the sameproportion of highs and lows.

(5) Why buy a new F.M. receiverwhen all the best programs arc stillon the regular stations? How can any-

listenerto hive 2 complete receivers in his liv-ing room? There are 45,000,000 receiversin this country. Why should they becomeobsolete overnight?Nobody wants them to. There are nowcompanies manufacturing the new F.M.

receivers for marketing during the nextfew months. But-in almost every case-the new F.M. sets also have a band -switchthat can turn instantly to standard broad-casts, thus giving you a choice of the oldor the new.

In addition, a number of manufacturersare making "adaptors" or "translators"that may be used in conjunction with astandard set to receive F.M. programs.Their use, however, is only recommendedwith sets that have superior tone-sincethe F.M. full -fidelity qualities may beeasily destroyed by a poor loudspeaker.

America's 45.000,000 radio sets will notbe obsolesced overnight. As the public buysnew sets, it will be urged to purchase com-bination A.M.-F.M. receivers. The process

therefore will be one of normal absorptionover a period of years.

(6) F.M. is quite beyond the rangeof the average pocketbook. P.M. sets willalways be much more expensive thanthe regular type of receiver.F.M. sets today are not produced in mass

quantities. Consequently their "per unit"cost is greater. Basically there is no im-portant difference between the componentsused in an F.M. receiver and those of astandard receiver, except that F.M. de-

mands a better loudspeaker and better -quality parts in the audio -frequency sec-tion of the set.

F.M. receivers today start at $70, runup as high as you care to pay for a fancycabinet and allied gadgets (such as phono-graph, automatic record -changer, short-wave bands, etc.). The new adaptors willsell for less than $50. As the public pur-chases larger numbers of F.M. sets, theprice will naturally tend to decrease.

(7) Even if you do purchase an F.M.receiver, there are no programs of in-terest on the air. Most of the F.M. sta-tions will just relay programb of regu-lar stations so that, from an entertain-ment angle, there's not much sense ingetting an F.M. receiver.On January 1, 1941 the new F.M.

broadcast band will be opened to fullF.M. commercial operation on a parwith standard broadcasting.

The new F.M. stations realize stronglythat they must provide a different pro-gram schedule, to a good degree, fromthat heard over the regular stations. Manyof them are already offering a daily sched-ule that duplicates only the most popularand important broadcasts. The new regu-lations issued by the Federal Communica-tions Commission require a minimum of6 hours' operation daily -3 in daylight.hours, 3 at night-with at least 1 hour ineach period devoted to special F.M. pro-gramming. Almost all of the new stations,however, will operate much longer than 6hours daily, originate far more than mere-ly 2 hours of F.M. shows a day.

Many of the new stations will have noconnection with existing broadcasters;their programs, therefore, will naturallyhave to be special originations. Purchaseof a combination F.M.-A.M. receiver istantamount to opening up a whole newworld of radio listening enjoyment

9

.4 LI, ABOUT FREQUENCY MODULATION

CHAPTER II

CONSTRUCTIONBUILD THIS PRACTICAL

F.M. ADAPTERComplete directions make it easy for anyconstructor to build and align this Fre-quency Modulation Adapter.

WITH the advent of Summer, the value of programs over thenew frequency -modulated broadcasting stations will becomemore and more apparent. Our own investigation of this newfield bears out the statements which have been made by the

engineers of several of the large companies which now have he-quency modulation receivers and frequency modulation adaptersfor use with regular broadcast receivers available.

The claim, of course, which seems most important with Summerand the thunderstorm season coming on, is that frequency -modulated broadcasts can be received through the most severethunderstorm with no interference whatever.

Here's an interesting story that illustrates this feature, whichFrank Gunther, the Chief Engineer of Radio Engineering Labo-ratories, the organization which has built most of the equipmentfor Major Armstrong, as well as for many other frequency modula-tion broadcast stations, told us that one day last Summer.

When a crowd of visitors had climbed the mountain to the loca-tion of the Yankee Network's station at Paxton, Mass., a verysevere lightning storm came up. The children of the group wereobviously frightened. The engineer in charge connected an extra,remote loudspeaker to the F.M. receiver used to pick up thetransmissions of the relay station, which was sending onward toPaxton the program from Boston, and raised the output volumeof the loudspeakers to a point where it overshadowed the thunder.While the storm was going on, the visitors enjoyed the music andpaid littl or no attention to the conditions outside, while previousto that time their attention was all on the outside.

CONSTRUCTIONFrom the accompanying photographs and drawings it will be

possible for the experienced constructor to duplicate the FrequencyModulation Adapter which we have built. By comparing the pic-tures, the circuit diagram and the List of Parts, it should be asimple matter to identify the location of all the component parts.The construction and the circuit are entirely straightforward andadjustment of the completed receiver follows normal practice. Itwill be observed that a view of the receiver has been made fromthe bottom, so as to enable the constructor to see those parts whichwould otherwise be obscured by a portion of the cabinet.

Considerably better performance is provided where an 1952(6AC7) tube is employed in place of the 6SK7 in the radio fre-quency stage. The improvement is found in much better gain andresults in considerably better limiter action. Another 1852,(6AC7) may be used in place of 6SK7V4 as indicated In the

10

The

nea

t and

com

pact

con

stru

ctio

n of

this

new

F.M

. Ada

pter

isill

ustr

ated

in p

hoto

A. E

ven

an u

nder

vIew

. pho

to B

, will

not

sca

re th

e co

nstr

ucto

r.

circ

uit d

iagr

am b

ut it

is s

omet

imes

fou

nd th

at th

is c

hang

eove

rre

sults

in o

scill

atio

n. I

n an

y ev

ent,

itis

wor

th a

tria

l.

ALI

GN

ME

NT

The

mat

eria

ls n

eede

d fo

r al

igni

ng th

e co

mpl

eted

F.M

. Ada

pter

are

a se

rvic

e os

cilla

tor

and

a 20

0-m

icro

amp.

gal

vano

met

er. F

oral

igni

ng th

e di

scri

min

ator

, T4,

a 0

.1-m

eg. r

esis

tor

mus

t be

in-

sert

ed in

ser

ies

with

the

met

er. W

ith m

eter

and

ser

ies

resi

stor

conn

ecte

d ac

ross

R23

app

ly a

sig

nal t

o co

ntro

l -gr

id o

f th

e 6S

J7.

Usi

ng 2

,100

kc.

as

the

freq

uenc

y of

the

I.F.

, and

mod

ulat

irtg

with

400

-cyc

le n

ote,

adj

ust t

he p

rim

ary

of T

4 fo

r m

axim

um r

eadi

ng o

fth

e ga

lvan

omet

er.

The

met

er w

ith it

s se

ries

res

isto

r sh

ould

now

be

conn

ecte

d to

both

cat

hode

s of

the

6H6.

App

ly a

n un

mod

ulat

ed s

igna

l to

the

cont

rol -

grid

of

the

6SJ7

and

adj

ust t

he s

econ

dary

trim

mer

of

T4

for

zero

rea

ding

on

the

galv

anom

eter

. Roc

k th

e os

cilla

tor

back

and

fort

h 10

0 kc

. eac

h si

de o

f 2,

100

kc. a

nd n

ote

that

the

galv

a-no

met

er s

houl

d sh

ow a

n eq

ual d

efle

ctio

n ei

ther

sid

e of

zer

o as

the

freq

uenc

y is

cha

nged

. The

re s

houl

d be

a c

hang

e of

vol

tage

pro

por-

tiona

te to

the

chan

ge in

fre

quen

cy e

ither

sid

e of

the

"cen

ter"

freq

uenc

y.N

ow r

emov

e th

e se

ries

res

isto

r an

d in

sert

a m

eter

in s

erie

s w

ithR

18at

the

poin

t mar

ked

X. A

pply

an

unm

odul

ated

sig

nal t

o th

eco

ntro

l -gr

id o

f V

4. S

et th

e os

cilla

tor

to 2

,150

kc.

and

adi

ust t

he


primary trimmer of T3 for maximum deflec-tion of the meter. Next set the oscillator to2,050 kc. and adjust the secondary trimmerfor maximum reading. The oscillator shouldnow be rocked 100 kc. either side of 2,100kc., and meter readings taken at variouspositions, to make sure the transformershows a symmetrical resonance curve. It isnot necessary that the transformer have aflat top (of 200 kc.) but that it should besymmetrical. It is desirable that the signalshould attenuate rapidly beyond the 2,000kc. and 2,200 kc. points.

Apply a signal to the control -grid of V3and proceed as above.

Short the oscillator coil to V2, apply asignal to the control -grid of V2, and adjust

.40,E,tizzt.r. 8::(1::1 &I' 44:

31

c.

1

_ .1 ..if

sl

yaBC.M1, I31 .. 1,,,

cl

I14+f-

1BB BO_

1....__ if 4..._ ii...1-9r---4-- 2--4-- 11

T2 as before. Next disconnect the short onthe oscillator coil and apply a modulated,43-mc. signal to V2 and turn C4 until themeter in the control -grid circuit of V5registers a reading. Now adjust trimmerC6 for maximum reading. Finally, apply a43-mc. signal to the antenna terminal andadjust C5 to maximum. The band -width ofR.F. is sufficient to pass the broad band.

When using a 6AC7/1852 in place of the6SK7 more care must be taken in theplacement of parts and in the laying ofground wires. Each circuit must be ground-ed at the socket to chassis and all points onthe chassis connected together with %-in.braid. It may be desirable to use braid toground the shield of the tube as wire has ahigher R.F. resistance.

It may be necessary to insert a 15- to25 -ohm resistor in series with the control -grid of V2, at point X, to suppress parasiticoscillation.

Coil construction: LI, L2, L3-5 turns,on 9/16 -in. form, spaced %-in. and woundwith No. 18 tinned wire. Primary -2 turnswound with No. 28 D.S.S. on lower end.Oscillator tapped 1% turns from bottom.An I.F. of 2.1 mc. is used.

CREDITS

It was our purpose in providing this de-sign to make the advantages of frequencymodulation reception available to the moreexperienced constructor. It will be recog-nized that the receiver has been designedto use items which will be found in stockin most of the leading radio stores. TheServiceman should find this design extreme-ly interesting because the construction of

12

GR

A. A

NT

65N

T6A

CT

OR

185

2 O

W

(c, O O

wai

t?.

61.4

.P

I

C2 IS

*AM

P

MA

II!

1.1.

11.1

. 3-w

OuN

Dw

1T N

5 T

AIN

S O

N A

NC

PO

RK

_S

PA

CE

D V

S:

7.11

5 T

INN

ED

CO

PP

ER

WIR

E.

SV

LI

9181

1AA

ITY

2 W

RN

S. W

OU

ND

WIT

H N

E26

055

WIR

EO

N L

ON

ER

EA

UO

SC

ILL

*TO

R T

AP

PE

D 1

41 T

UR

NS

FR

OM

bO

TT

OA

E31

4cF

US

ED

.

115

V. A

. C. .

0

o812

,000

OH

MS

,la

SW

R5

1500

0O

NU

S.

IhV

it

16

4.A

6c1C

;:118

52 (

M)

C16 01

-O

AF

R25 05

-M

EG

.

C22

100

ALA

iR

0111

6

ON

1000 AA

54E

w

LE

R (

V5)

C21 5

umP

Cl9

.411

!01

2? -

r".

TM

MP

R23

,at

-mea

.

Thi

s F

.M. A

dapt

er m

ay b

e us

ed w

ith h

eadp

hone

s an

a c

ompl

ete

rece

iver

; or

it m

ay b

e us

ed to

driv

e th

e A

.F. s

ectio

n of

any

rad

iore

ceiv

er o

r P

.A. s

yste

m.

such

a u

nit w

ill g

ive

him

an

insi

ght i

nto

the

oper

atio

n of

the

rece

iver

s of

this

nat

ure

and

will

be

very

ben

efic

ial t

o hi

m in

han

-dl

ing

serv

ice

prob

lem

s on

the

vari

ous

F.M

.se

ts n

ow o

n th

e m

arke

t. C

onne

ct a

dou

blet

ante

nna

to te

rmin

als

Ant

. and

Gnd

.T

he a

utho

r de

sire

s to

exp

ress

app

reci

a-tio

n fo

r th

e as

sist

ance

giv

en h

im b

y M

essr

s.Sh

augh

ness

y, D

ay a

ndSt

iles,

of M

ajor

Arm

stro

ng's

labo

rato

ry.

The

act

ual b

uild

ing

of th

is r

ecei

ver

was

done

for

the

wri

ter

by M

r. A

nton

Sch

mitt

,W

2RW

Y, o

f th

e H

arve

y R

adio

Com

pany

of

New

Yor

k C

ity; a

nd th

e re

ceiv

er h

as b

een

thor

ough

ly te

sted

not

onl

y in

the

labo

ra-

tory

of

the

Nat

iona

l Com

pany

at M

alde

n,M

ass.

, but

als

o in

the

labo

rato

ry o

f M

ajor

Arm

stro

ng a

t Col

umbi

a U

nive

rsity

in N

ewY

ork

City

, and

in th

e R

adio

Eng

inee

ring

Lab

orat

orie

s, L

ong

Isla

nd C

ity, N

. Y.

LIS

T O

F PA

RT

SC

ON

DE

NSE

RS

Thr

ee N

atio

nal C

o., t

ype

UM

-16

, Cl,

C2,

C3;

One

Nat

iona

l Co.

, typ

e U

M -

50, C

4;T

wo

Nat

iona

l Co.

, typ

e 3-

30, C

5, C

6;Fo

urte

enC

orne

ll-D

ubili

er,

type

DT

-6S

1,0.

01-m

f., 6

00 V

., C

7, to

C19

(in

cl.)

, C29

;T

wo

Cor

nell-

Dub

ilier

,ty

pe 5

W -

6Q5,

50

ALL ABOUT FREQUENCY MODULATIONmmf., C20, C21. (erroneously indicated indiagram as 5 =inf.);

Tnree t-orneil-punilier, type 5W -5T1, 100mmf., C22, C23, C24;

One Cornell-Dubilier type, 1W -5D1, 0.001-mf., C25;

One Cornell-Dubilier, type DT -4S1, 0.05-mf.,600 V., C26;

One Cornell-Dubilier, type BR -845, 8 mf.,450 V., C27;

One Cornell-Dubiliei, type BR -1645, 16 mf..450 V., C28;

RESISTORSTwo I.R.C., type BT%, 300 ohms, R1, R14;Four I.R.C., type BT%, 1,000 ohms, R3, R8,

R11, R16;Two I.R.C., type BT'4, 20,000 ohms. R4.

R21;Four I.R.C., type BT%, 40,000 ohms, R7,

R12, R17, R18;Two I.R.C., type BT%, 16,000 ohms, R8.

R13;One I.R.C., type BT%, 200 ohms, R9;Two I.R.C., type BT%, 0.1-meg., R22, R23;One I.R.C., type BT%, 50,000 ohms, R24;Three I.R.C., type BT1, 60,000 ohms, R2,

R10, R16;One I.R.C., type BT1, 20,000 ohms, R5;One I.R.C., type BT1, 0-1-meg., R19;

One T.R.C., type BT1, 50,000 ohms, R20;One I.R.C. potentiometer, type 13-133, 0.5-

meg., R25;TUBESOne Sylvania or RCA 6SK7, or 1852 (see

text), VI;One Sylvania or RCA 6SA7, V2;One Sylvania or RCA 1852 (see text), V8;One Sylvania or RCA 6SK7, or 1852 (see

text), V4;One Sylvania or RCA 6SJ7, V5;One Sylvania or RCA 6H6, V6;One Sylvania or RCA 84, V7;MISCELLANEOUSThree National Co. I.F. transformers, T1,

T2, T3;One National Co. discriminator transform-

er, T4;One United Transformer, A.F. transformer

type R-54, T6;One Thordarson choke, Ch.1;One Hart & Hegeman roto switch, Sw. 1;One National Co. steel cabinet, type C5W-3

(the subpanel comes with this cabinet);One National Co. dial, type 0, with No. 2

scale;'One National Co. dial drive, type ODD;Two National Co. knobs, type HRP.

14


CHAPTER III

AUDIO AMPLIFICATION

F.M. AUDIO AMPLIFIERFOREMOST among problems presented

by Frequency Modulation is the designof an amplifier which will not prove tobe the "bottle neck" of the entire sys-

tem. The new standards set by the FederalCommunications Commission for designingF.M. transmitters, that should be taken intoconsideration when designing an audio am-plifier for F.M. receivers, briefly follow:(1) The transmitter and associated studio

equipment shall be capable of trans-mitting a band of frequencies from 50to 15,000 cycles within 2 decibels of thelevel of 1,000 cycles. In addition, pro-vision shall be made for pre -emphasisof the higher frequencies in accordancewith impedance frequency characteris-tics of a series inductance - resistancenetwork, having a time constant of 100micro -seconds.

(2) The noise in the output of the trans-mitter in the band 50 to 15,000 cyclesshall be at least 60 decibels below theaudio frequency level represented by afrequency swing of 75 kilocycles (100%modulation).

(8) At any frequency between 60 and 15,-000 cycles at a swing of 15 kilocyclesthe combined audio frequency har-monics generated by the transmittingsystem 'hall not be in excess of 2%(root mean square value). This means,simply, that the transmitter should becapable of passing a band of 50 to 16,-000 cycles ;f:2 db. of the 1,000 -cyclereference; it shall have a combinedhum and noise level at least 60 db. be-low full power output; and, it shouldnot generate more than 2% totalharmonics at any frequency within itstransmitted band.

F.M. A.F. AMPLIFIER STANDARDSIn setting up standards for an F.M.-re-

ceiver audio amplifier the natural reactionwould be to use the standards set for theF.M. transmitter. Careful consideration,however, will reveal specific disadvantagesfor such an arrangement.

15

It is obvious that for ideal performance,the amplifier at the receiving end shouldhave an effectively flat frequency response,introduce Ito distortion and have no in-herent noise. With such an ideal amplifier,the full benefits of frequency modulationwill be obtained.

Any discriminating characteristics in-herent within the receiving amplifier will, ofnecessity, introduce additional detrimentalconditions, which are added to existing de-ficiencies within the transmitter to providean overall result far below a desirableideal. For example, let us assume that thetransmitter is down 2 db. at 60 cycles. Thereceiving amplifier (which was built in ac-cordance with the standards set for F.M.transmitters) is also down 2 db. at 60cycles. The overall result will be a 4 db.loss at this low frequency, which is suffi-cient to change the character of many typesof music. Similarly, an amplifier which in-troduces 2% distortion (say at an averagelevel of 1 watt) will provide an ultimateprogram having a combined distortion ofmore than 2% (which we can assume wasproduced by the transmitter). It thereforefollows that the amplifier should be definite-ly, better than the transmitter.

In addition to this, it is also feasible toassume that additional improvements willbe made in F.M. transmitters, and F.C.C.regulations may tighten their specifications.If this occurs, an amplifier which has beenbuilt to existing standards may not passon to the listener all the benefits of futureimprovements in F.M. transmission. Thepresent specification covering the width ofthe audio band is unbalanced,* and it isreasonable to assume that, in time, thelower portion of the band will ultimatelybe extended to at least 26 cycles to producea balanced spectrum.

Proof of this line of reasoning can befound in new F.M. transmitters, which arebeing constructed to exceed the F.C.C.'sF.M. requirements. For example, one of

*See "Balanced Audio Spectrums," Radio -Craft, Sept., 1940, pg. 164.

VO

LUM

E

65.:

100

NIE

C,

2 5

0-

0.45

11,

mr.

0,4,

0;5

_

.m

r

'A.

101

-V

EC

.;MFv

ME

V

L6G

OH

MS

400

10.0

0005

M5-

--ra

wie

rw-

LE4F

20

10'

170

_

525

snA

.10E

R;''

'nf E

R

.40

PT

%.1

1.-.

.13

AM

P,

it.1

52

CIG

ITr"

..Fe

lt ,

.

Sch

emat

ic c

ircui

t of t

he P

ush

-Pul

l Dire

ct -

Cou

pled

Fre

quen

cy M

odul

atio

n A

.F. A

mpl

ifier

.It

inco

rpor

ates

bal

ance

d ne

gativ

e -f

eedb

ack

and

nove

l A.G

. and

D.C

.ba

lanc

ing

circ

uits

.

the

larg

est m

anuf

actu

rers

of

tran

smitt

ers

guar

ante

esth

efo

llow

ing

audi

och

arac

-te

rist

ics:

(1)

Freq

uenc

yR

espo

nse-

Flat

-±-1

db.

from

30

to 1

5,00

0 cy

cles

.(2

) N

oise

Lev

el -

70 d

b. b

elow

ful

l mod

ula-

tion.

(3)

Dis

tort

ion-

Les

s th

an 2

%, t

otal

har

-m

onic

s.It

was

ther

efor

e de

cide

d to

ant

icip

ate

are

ason

able

am

ount

of

impr

ovem

ent a

nd d

e-si

gn th

is F

.M. a

mpl

ifie

r so

as

to p

reve

ntob

sole

scen

ce. T

he f

ollo

win

g te

ntat

ive

spec

i-fi

catio

ns w

ere

set:

(1)

Freq

uenc

y R

espo

nse-

J-1

db. f

rom

13

to 3

0.00

0 cy

cles

.

(2)

Noi

se L

evel

-At l

east

75db

.be

low

rate

d po

wer

out

put.

(3)

Dis

tort

ion

-1%

(at

ave

rage

wor

king

leve

l), t

otal

har

mon

ics.

With

an

ampl

ifie

r of

this

type

, it w

as f

elt

no u

ltim

ate

cons

umer

wou

ld e

ver

hare

tow

orry

abo

ut h

avin

g th

e "b

ottle

nec

k" o

f as

sF.

M. p

rogr

am in

his

aud

io a

mpl

ifie

r eq

uip-

men

t.Fu

rthe

rmor

e,re

ason

able

impr

ovem

ents

in F

.M. t

rans

mitt

ers

(bas

ed o

n si

mila

r im

-pr

ovem

ents

whi

ch h

ave

take

n pl

ace

in A

.M.

wor

k)w

ill p

rovi

de d

irec

t ben

efits

to th

elis

tene

r.

SEL

EC

TIN

G T

HE

FE

AT

UR

ES

The

Equ

aliz

erO

ffha

nd,

it w

ould

arm

ee,-

that

an

F.M

.A

mpl

ifie

r sh

ould

bebu

iltto

mee

t ide

alre

quir

emen

ts a

nd h

ave

unva

ryin

g ch

arac

-te

rist

ics.

Inot

her

wor

ds,

the

ampl

ifie

rsh

ould

be

devo

id o

f hi

gh -

freq

uenc

y or

low

-fr

eque

ncy

cont

rols

. Ref

erri

ng to

the

re-

quir

emen

ts s

et b

y th

e F.

C.C

., it

will

be

note

dth

at p

rovi

sion

mus

t be

mad

e in

eve

ry F

.M.

tran

smitt

erto

pre

-em

phas

ize

high

fre-

quen

cies

. Thi

s m

eans

that

hig

h fr

eque

ncie

sw

illbe

acc

entu

ated

duri

ng tr

ansm

issi

on.

The

pur

pose

of

this

pre

-em

phas

is is

to a

t-te

nuat

e re

sidu

al a

tmos

pher

ics.

ALL ABOUT FREQUENCY MODULATIONAs disturbing effects of atmospherics are

predominant in the higher audio frequen-cies, it is logically assumed that accentua-tion at the transmitter and attenuation atthe receiver will ultimately result in a flatoverall response and at the same time,materially attenuate atmospherics. This isgraphically illustrated in Fig. 1.

If we assume that a high -frequency pro-gram signal has a level of +20 VU and itis pre -emphasized to a level of +23, thissignal will be received along with an at-mospheric disturbance of say +20. Hence,without pre -emphasis, the original programsignal and the atmospheric will be of equalintensity. On the other hand, pre -emphasishas already made the program signal ap-preciably higher than the atmospheric. Byattenuation in the receiver, the programsignal is brought back to its original levelof +20 VU, and the atmospheric is reduced3 VU. The degree of attenuation of dis-turbances is a function of the pre -emphasisat the transmitter.

From a casual study of this operatingprocedure, it would appear that a high -fre-quency attenuator is the only required con-trol of the receiver. A study of existingdeficiencies in present records, however, will

(kcal:. indicate that both the high and lowfrequencies should be independently con-trolled, and the control range should piovidefor both attenuation and accentuation. An -ether very desirable characteristic in theequalizer circuit i. to have it exactly com-plement the equalizer used at the trans-mitter or in the recording studio (forrecorded programs). The equalizer shouldnot introduce harmonics, hum, or resonantpeaks in any portion of the spectrum.

The VU MeterIt was also considered desirable to have

a visual monitoring arrangement so as toindicate normai, average, and peak levels..f the program. This auxiliary feature ishighly desirame when it is required to avoidi.verload of either the amplifier or.the loud--peaker. Low -frequency speaker overload isusually judged from a distortion viewpoint,becau,c the intensity of the signal cannotbe accurately judged in view of the factthat the ear is comparatively insensitiveto low frequencies. Only critical listeners.therefore, will detect overload at low fre-quencies. The use of the meter, however,makes it possible for any average in-dividual to adjust the intensity of the prograin level so as to definitely prevent over-load at any frequency. Furthermore, it be-come, relatively simple to deter'; just whatactual effect the various settings of theequalizer controls have upon the overallprogram level.

17

Dual -Channel Input and Electronic MixerIn order to extend the usefulness of this

Direct -Coupled F.M. Amplifier, it was con-sidered desirable to incorporate an addi-tional input circuit so that phonographrecords, in addition to F.M. transmissions,may also be enjoyed.

A dual circuit input could most economi-cally be employed by the use of a change-over switch, but inasmuch as the averagevolume level of the radio program and therecorded program may be different (andtherefore necessitate a continual change),it was thought more desirable to incorpo-rate an electronic mixer. This provides 2entirely independent input channels withindependent controls so that each levelmay be set for ideal results. Furthermore,the use of the electronic mixer insurescomplete isolation of both controls, so thatthey do not affect either the volume or thefrequency response characteristics of itsassociated channel.

Details covering the design of these 3features will be described in Part II of thisarticle. A block diagram which shows therelative position of the various features isgiven in Fig. 2.

THE AMPLIFIER

As all of the several 10-, 20- and 30 -Watt Direct -Coupled Amplifiers previouslydescribed in this magazine have been de-signed around an effective drift -correctingcircuit, no immediate improvement instability seemed apparent. Subsequent in-vestigation, disclosed that unusual dif-ference in plate resistances of the inputtubes affected the performance of direct -coupled amplifiers more than resistance -coupled units. This difference in effect wasto be expected to be noticeable because ofthe increased efficiency, improved response,and lower noise level characteristic of di-reet-coupled amplifiers. Upon further in-vestigation, it was found that manufac-turers of tubes had not set close standardsfor plate resistance of preamplifier andvoltage amplifier tubes.

Although normal variations in tubes pro-duce a measurable difference in the per-formance of the resistance- and trans-former -coupled amplifiers, they have beenfound to produce another effect in direct -coupled amplifiers. For example, an un-balanced pair of input tubes would un-balance the plate current us the outputtubes sufficiently to increase residual humand require readjustment of the hum -bal-ancing adjustment. It was therefore decidedthat 2 self-correcting networks would beincorporated in this new amplifier; one toautomatically balance for difference in the


i I i:MINFINRAMMED SWAN. ATNO5rNER,C DISTURBANCE. rserigirsotg%) SiGNAL ANO DISTURBANCE AS FT AMMO AT TN!

e.

a

il:Mr. 4-Rvl.'"irfai;r1184.11.:1414,14"7"'sS Le"... C.',1t2.11-'TS

I i 4ealluti. ..c.:71.7eATIZIS.'"

A A C Tr - +.5.0....4 FIESIDuAL AITACOrNtRoC 0 STUREIMICES ARE.11Rrwelii ATTENLIATED As PRBEFAINALSIS DaRING TRANSAiss/Opv.,

MaES

fie r,,,.....c,h01gEll/

Lic,A._.,c;PIWNTVIII

/,P.4.2.4.

.....riii.MI//

SAPULL

/ CI CBALANCER

/A C

BALANCERft's"-PULL tt/ZtD

TRANSPORMSRita9 isAm:k 4- 4 4 =;) 4 Ri"

Y / POwER Surrsv ATEWS/ /,' VIIWREN / SuPPLy

FIG 2 ..BLOC.* DIAGRAM OP COMPLETE AAPLIINCRA, C INPUT

GS.IT TOPLAT* WA

63/7

PLATE(-.t..__, CURRENT

k = i a7', , CONTROL)

Gt --;=r IA

CORTROL.. itp646.

t9 fk

Dl t l P7ESS7 \ 7:1

--tcRAPENTS OF Rercie0 SaRAA.SOCO DIM T-COACO ANNINEA..~BASIC BALANCING CONTROL ,.

V2 ..

6i.E4 AIlG*RAS34000 GAPGGNAN- E 1

TUBE-L.

RAT[.,T2Te

CCYY

PM

'

TAW

a:

6sci

1 I-

I $

11II

x

I ,

V' ,-'A2

I 11 IIOV'34000V3

si 14, °H.'s.- ---NCO

SLAG

FIGS AZIG=1. .11%.M.D.00 BASIC CLECTRONICFIG 6 -.., ORIG.AL IVAN-CNC BALANCE INDICATOR.* II015-A.

TOISV

mina rc

ten:ek

.

SCTD MA

rd

rssE

Itre Tr'

AC

-EOLPTALENT CROAT Or

orte.w6-5.-._L.

.1:ChAvALCPIT 5NIKVSASIEDIN OP TSPe®MC.. GAUNT A,

.E0ONALENT NAIN-NALPLATE SALANCoNit CRcurT,.

....a=TA nut_

plate resistance of the driver tubes and theother to automatically balance for differencein gain of the driver tubes. As a furtherrequisite. it was decided that these circuitsshould provide for superior results in thedirect -coupled amplifier as compared to astandard resistance -coupled amplifier witha given set of greatly unbalanced (or evendefective) tubes.

18

THE D.C. BALANCING CIRCUITDuring the development of the 30 -Watt

Direct -Coupled Amplifier,' it was foundthat a normal variation between tubescould be compensated -for by correcting thebias on the input tubes. The basic portionof this manual balancing circuit is illus-trated in Fig. 3.


Fortunately, when an unbalance of morethan 10 milliamperes occurred in the out-put stage, the hum level came up. It there-fore became a relatively simple matter tobalance the input tubes by adjusting forminimum hum. With a change of inputtubes, it was sometimes necessary to re-adjust the initial setting. It was found,however, that some of the ultimate usersof these amplifiers would insert greatly -tin -balanced tubes, without attempting to re-adjust for balance. It was therefore be-lieved highly desirable to provide someautomatic means for balancing. The firstmethod of attack which presented itselfwas to use a tube in place of the load re-sistance of the voltage amplifier and ar-range for automatic compensation for vari-ations in plate resistance of the voltageamplifier. Another tube was to be used toaugment the bias of the output tubes, soas to compensate for variations in outputplate current.

In Fig. 4, which shows the elements ofa revised single -sided direct -coupled am-plifier, Rp is the plate resistance of thevoltage amplifier and Rk is the partialcathode resistor of the power amplifier.

Figure 5 shows the basic balancing cir-cuit originally conceived to automaticallycompensate for both variations in plate re-sistance of the input tube VI and the out-put tube V2. It will be noted that V3 isused as a plate load resistor for VI. Thebias applied to V3 through Rl depends uponthe plate current flowing through its cathoderesistor R2. The time -delay constant ofRI, Cl, prevents signal frequencies fromaffecting a change in the plate resistance ofV3, and limits automatic adjustments onlyfor "steady state" or average conditions;V4 was to be used as a shunt across Rk,so as to keep the bias across Rk constant.This circuit is likewise made responsiveonly to steady state or average unbalance,by inserting a time lag through the re-sistor -condenser network R2 -C2.

Inasmuch as the final amplifier was to bepush-pull throughout, 4 additional tubeswould be required for this balancing ac-tion. The added expense and complexityof this circuit inspired additional researchto produce a simpler and more economicalcircuit to achieve the desired results.

A side project was started to adapt theuse of the twin indicator (6AF6G) througha twin -triode amplifier (6SC7), so ar-ranged as to measure the voltage dropacross the balanced primary winding ofthe output transformer. A special trans-former was wound so that both sides of theprimary were of equal D.C. resistance (andequal A.C. impedance). The idea behindthis development was to provide a partiallyvisual check on the plate current of theoutput tubes so that should greatly un-

19

balanced tubes be used, it would become im-mediately visible, and the tube would pro-vide for read;ustment. It was found, how-ever, that the indicator with its associatedamplifier was too insensitive for the aver-age user to adjust within a 10 -ma. balance.This circuit was therefore abandoned, butit is given in Fig. 6 for the benefit of somereaders who may have other applicationsfor this particular type of indicator. Thecondenser -resistor network R1 -C1 providesa time delay to prevent A.C. potentials fromhaving any effect upon the twin -eye in-dicator. A novel portion of the circuit isthat raw A.C. is applied to the plates ofthe indicator. The flicker is not observedbecause of the persistence of vision of theeye which will tolerate interrupted imagesdown to about 16 cycles before flicker be-comes visible.

The easiest way to understand the actionof the final D.C. balancer is to substitute aresistor (rl) for the plate load and an-other (rp) for the plate resistance of thetube. If a D.C. voltage E (as indicated inFig. 7) is applied across this network, thevoltage Edc is the effective voltage appliedto the plate of the tube and is dependentupon the voltage drop across rl. Thus, ifrp is varied from zero to infinity, the volt-age will vary proportionately. The ratio ofvoltage change will depend upon the ratio

rpof . If rl is made large in comparison

rl + rpto rp, the ratio of change will be small.If an additional resistor (rc) is insertedin series with both rl and rp, as indicatedin Fig. 8, then the effective voltage E'dc

rpwould be equal to . The push-

rc + rl + rppull version of this circuit is indicated inFig. 9. If we neglect rk (which is verysmall) the voltage which appears acrossrl' + rp', is equal to Bdc which can becalculated from

Bdc =

(rl' rp') (r12 rp2)

rp' rl' rp2 r12

IC +(rl' rp') (r1= rp2

rp' rp2 rl

If Al is 100,000, rc is 500,000, and rp'varies from 800.000 to 120,000 ( which repre-sents a -± variation of approx. 20%), it willbe found that the percentage of change atBdc is 1.9r,', as compared to a changewhich would take place under conditions ofFig. 7. In other words, a 50% correction isaffected. If the same type of network isapplied to the screen -grids of the driver


tubes, as indicated in Fig. 10, still morecorrection is affected.

The practical value of this self -balancingcircuit can best be indicated by referring tolaboratory data compiled during its develop-ment. A total of 100 average 6SJ7 tubeswere checked for the maximum deviationthey produced in the output plate circuitof the 6L6G's. Two sets of the worst com-bination produced the following results:

Unbalanced OutputTube Numbers Unbalance

1 and 2 61 ma.land 3 68 ma.

When these same tubes were inserted intothe balancing circuit, the following resultswere noted:

Unbalanced OutputTube Numbers Unbalance

1 and 2 8 ma.land 3 8 ma.

As the D.C. balancer becomes an integralpart of the A.C. balancer circuit as well, itwas necessary to select optimum resistorvalues which would provide a minimum D.C.unbalance and minimum A.C. unbalance. Thedesign of the A.C. balancer circuit willbe discussed in Part II.

F. M. AUDIO AMPLIFIERPART II

IN an effort to surpass the stability ofconventional transformer- and resis-tance -coupled amplifiers, a combinedD.C.-A.C. balancing circuit was de-

veloped. The D.C. balancer provides forautomatic audio -drift correction understatic conditions. The A.C. balancer pro-vides for automatic signal balancing underdynamic conditions.

Stated in other words, the static balancer(D.C. corrector) automatically compensatesfor a very wide variation in plate resistancecharacteristics of input tubes. It preventsunbalance in the output stage with changeof emission characteristics within the inputstage. The dynamic balancer (A.C. cor-rector) automatically compensates for avery wide variation In voltage amplificationof input tubes. It prevents the applicationof unbalanced signals to the control -gridsof the output stage.

THE DYNAMIC BALANCER(A.C. CORRECTOR)The easiest way to understand the operat-

ing principles of this unusually effectivecircuit is to analyze the basic operatingprinciples of the screen -grid tube. This tubeis normally used in a conventional manner,i.e., by applying a control voltage to thecontrol -grid; a "B+" voltage, adequatelybypassed to the screen -grid; a "B " volt-age, through the load resistor to the plate;and, its suppressor -grid connected to cathode.

20

If these elements are viewed fundamental-ly as diagrammed in Fig. 1, it will be notedthat all of the grids are in the electronstream. This means that any one of themcan be used as a control -grid. Naturally, thefurther away the grid is from the emitter(cathode), the less control it has upon theelectron stream. If the grids are labelledGl, C2, and G3, in order of their distancefrom the cathode, these notations will cor-respond o control -grid, screen -grid, andsuppressor -grid, respectively.

Figure 2 shows a standard circuit, where-in the input signal (e) is applied to thecontrol -grid, a signal Vae will appear at theplate. This voltage will be out -of -phase withthe input signal. If the screen -grid bypasscondenser, C, is disconnected, and the volt-age a is applied in series with the condenser,as illustrated in Fig. 3, a voltage Va'e willappear on the plate. Va may be defined asthe control -grid to plate voltage amplifica-tion. Va' may also be defined as the screen -grid to plate voltage amplification. It istherefore obvious that the screen -grid canbe used as a control -grid. This particularapplication is important as it plays a primerole in our dynamic balancer.

The suppressor -grid may likewise beused as the control -grid in which case, thevoltage which appears at the plate would beequal to Va"e (wherein Va" may be lookedupon as the suppressor -grid to plate voltageamplification).


PRINCIPLES OF DYNAMIC BALANCINGWith the above phenomenon kept in mind,

a review of fundamental balancing circuitswill further simplify the operating prin-ciples of the dynamic balancer. If an A.C.generator E is applied to a series resistivenetwork R1, R2, as illustrated in Fig. 4, thevoltage (E) appearing across R2 is equal to

R2E. Expressed mathematically, it

R2 -I- R1 R2Ebecomes e

R2 -I- RIIf another identical generator F is con-

nected to a similar resistance network RI,R2, the voltage (f) which appears across

R2ER2 is likewise equal to . If both

R2 + R1circuits of Figs. 4 and 5 are connected to-gether, so that R2 becomes a common re-turn, Fig. 6 results. If the generators Eand F are so adjusted as to be equal in po-tential but opposite in phase, and Rithe following voltage conditions are present:

(1) The voltage across E (e) is obviouslyequal to the voltage across F (f).

(2) The voltage across Ri (el) is equalto the voltage across

(3) As the voltages are out -of -phase, itis ipso obvious that the voltagesacross R2 will cancel, and equal 0.

(4) The voltages across X and Y (e xy)wi.I also cancel and be equal to 0.

The above conditions are prevalent onlywhen the gene:ators are opposite in phaseand of equal potential. If we assume, how-ever. that one of these generators drops involtage, let us say to 505, of its originalvalue, it is apparent that the total differencewill be equal to e-f or e xy. With an un-balance in the generators it is further ap-parent that complete cancellation will notoccur across R2. In fact, some of the largervoltage will appear at this point. This volt-age (e' xy) is equal to

R2 R2E -F e' xy

R2 4- R1 R2 -f- R1An examination of this formula shows that

as R2 is increased, more of the unbalancedvoltage appears across it. If this voltageunbalance (e' xy) is applied back to F,so as to increase its voltage output, it isobvious that some balance will automaticallybe obtained.

THE DYNAMIC PLATE BALANCERHow this is actually done in the amplifier

can best be indicated by redrawing Fig. 6and replacing E and F by their respectivetube circuits, as indicated in Fig. 7. In thiscircuit, the push-pull generator EF, takesthe place of the original generators E andF. RI becomes the independent plate loads

21

PL ATEG2_,

SCREEN -G3. GRID

SuPPRESSOR-GRID

CATHODECGVITPL-

FIG I "FUNDAMENTAL SCREEN -GRID ELEMENTS",

Vae

rrewuu--NAN

rc!r, 2

..."

ri

'B+FIG.2 -STANDARD PENTODE CIRCUIT"'

VaW

eri../.

A 011k

110

rgzA.,

rt

FIG .3 "SCREEN -GRID USED ASA CONTROL -GRID",

R2EeRz+Ri %I

FIG .4 -'BASIC DIVIDER NETWORK",

'RIRzf E

F*

". 2.

RZ+FIL

FIG.5 Di VI DI 14 NETWORK OPROSafty Pi4A.;(0113 1164CIRCUIT

x Vexy

ei RI e)

_L

e e'syf f

!R2

FIG .6 '-BASIC D1 RAMC BALANCER EIRCUIT,,,


TUBE A

AIli

EE

W5I R2

. .

1,..

--3.TUBE B '.....7

FIG 7 -.THE DYNAMIC PLATE BALANCER CIRCUIT,.

20y.TUBE A /;&:\ 0.1-L--..... \ MEG.

EF

Av

111,<-6 C93

.BV -11-BV. .134-'V. 0 I-

MEC.MEG.

TUBE B-110v

FIG .B VOLTAGEWH CHPA gli;IE A REFAT IgErsFERENCE

TUBE A---14 R3

R4

r B.I.'

r...,.....\TUBE B ..."...6"

F 1 G.9 Elf MENTS OF THE EnNAMIC SCREEN -GRID BALANCER

SIGNALS DEVELOPED ACROSSPLATES OF TuBE5 A

TUBE A ARE ilaaF 1,., .iASE& B TUBE a.- /..-

---,....(I ini IC-.

1 R3

_.1.:

R4 rg

2 rC

CL

T -zix, 9 (ONE CONTROL-GR1F I G 10 -fgictCrEsr.'74S tICGaP1^-

Tv&SCREEN-GRID

IV Cg3 RIilllAce.

R3

../V.,r, --,--,,Cg Z -

R3R4 l'Ft 2

RI

BISCREEN -GRIDru m Ei '

...,-,,E AuAILIARI REGENERATIVE MAL IZER, NEE ERt NTIAL VOLTAGE ACROSS R2 DRIVES BOTHFIG I 1 1,c,RE EN -LAWS THROUGH CONDENSER Cg3"--

of both tubes, while R2 becomes the commondegenerative resistor. If both tubes A and11 have identical voltage amplification char-acteristics, the voltage which appears acrossR2 will be 0. On the other hand, if A has

twice the voltage amplification of B, thena portion of this difference will appearacross R2.

A typical example is given in Fig. 8,wherein the plate load resistors RI equal100,000 ohms each, the common degenerativeresistor R2 is equal to 400,000 ohms. If weassume that the voltage amplification of onetube (A) is 20, and the other tube (B) is10, and if a balanced push-pull signal (grid -to -grid of 2 volts) is applied to the inputof the circuit, the voltage which appears atthe plate of A is equal to say, -1-20 volts(the voltages indicated are instantaneousA.C. voltages). The voltage which appearsat the plate of B is equal to -10 volts. Ifthese signals are out -of -phase, there willbe a total voltage difference between bothplates of 30 volts (for ideal conditions,there should be a total voltage difference ofeither 20 volts [if both plates have 10 voltseach] or 40 volts [if both plates have 20volts each]).

The portion of the voltage developed byplate A which appears across the 400,000 -ohm resistor, is equal to

400,000 4

+ 20 X -20 -= + 16400.000 -r 100,000 6 volts

The portion of the voltage developed byplate B which appears across the 400,000 -ohm resistor is equal to

400.000 4-10 X - 10 > -= -8

400,000 100,000 5 voltsThe cancellation which occurs across the

400,000 -ohm resistor is equal to 16- 8 or+ 8 volts. This instantaneous value of + 8volts is fed back to the screen -grids of bothtubes to affect further automatic connection.Before considering the balancing action ofthis voltage, let us briefly look into thescreen -grid circuit.

THE DYNAMIC SCREEN -GRID BALANCERThe fundamental principles involved in

the dynamic screen -grid circuit are virtuallyidentical with those for the plate dynamicbalancer. There are. however, 2 importantexceptions.

In our conventional circuit of Fig. 2, itwill be noted that the screen -grid was by-passed to ground through C. If this con-denser is entirely removed, a voltage willappear at the screen -grid, which is equalto Va"'e (Va " being the control -grid toscreen -grid voltage amplification). If therest of the circuit of Fig. 2 remains un-changed, it will be found that the voltagesVae and Va-e will be in -phase. The voltageVae however will be decreased. This iscaused by the degenerative action of thevoltage which appears at the screen -grid.Its degenerative action can best be analyzedby referring again to Fig. 1. If a positiveinstantaneous voltage is applied to GI, theelectron stream is increased. The increased

22

ALL ABOUT FREQUENCY MODULATIONcurrent through G2, produces a drop acrossits supply resistor. This, in turn, decreasesthe applied potential of C2 to retard theflow of electron streams to the plate. Asthe control -grid to screen -grid voltage am-plification increases, the control -grid toplate voltage amplification decreases. Verylarge signals can easily he handled by thescreen -grid under this condition.

Figure 11 shows the elements of the dy-namic screen -grid balancer circuit, ar-ranged to simulate the plate dynamicbalancer of Fig. 8. It will be noted, however,that an essential difference is the inclusionof the condenser Cg2. If the control -gridto screen -grid characteristics are identicalin both tubes, complete cancellation of thevoltages which appear at both screen -gridswill take place, as discussed for the condi-tions illustrated in Fig. G. Let us assumefor a moment however, that the control -grid to screen -grid characteristics of tube B,are lower than that of A. This naturallymeans that complete cancellation will nottake place across both screen -grids and aresidual potential will appear at the screen -grid of tube A. This voltage will then drivethe screen -grid of B in a very conven-tional resistance -coupled circuit, which caneasily be perceived by redrawing Fig. 9,as indicated in Fig. 10.

Here it will be noted that the screen re-sistor R3 of tube A acts as an equivalent"plate load". Condenser Cg2 assumes therule of the common coupling condenser.The screen -grid of B acts as the control -grid. The voltage which appears at theplate of B will be out -of -phase with thatwhich appears at the plate of A becauseof the following reasons: when the con-trol -grid of tube A is being used as a driver,and it becomes instantaneously positive,loth the plate and its screen -grid becomeinstantaneously negative. The negativescreen -grid of tube A is coupled to drivethe screen -grid of tube B negatively. Thisin turn produces an instantaneous positivepotential at the plate B.

With correct selection of values, this cir-cuit may be made to operate as a perfectinverter, and shows how complete balanc-ing may be attained even though the con-trol -grid of tube B is entirely inoperative.In actual practice, however, such a conditionis rarely encountered. What usually hap-pens is the control -grid to screen -grid volt-age amplification of both input tubes arenot always equal. This coupling circuitequalizes the difference within the firststage so that practically equal- but op-positely -phased voltages appear at thepush-pull output plates of A and B.

In addition to the dynamic screen -gridbalancer and the dynamic plate balancer,there is an auxiliary regenerative balancerwhich comes into play when the common

23

k I

-,,HE A Tuhf,(EP %tR,,,N TIE DIFFERENTIAL

I,RiS l (..HCo,1- of f ir

RI 'C93SCRELN-

GR1C,Tuht A

SCREEN -

TUBE8

FIG.12

C91

LAL.BRATED%ARABLE \'-"" A rrt NuATOR"AL.NEWAI,CNL LiENERATOR

/' TI

TRANsFOERBALANCERM D

CAAnylp 4R, AA TT ot wr,

BALANCE:,

J ME

TOAMPLIFIER

UNDERTEST

1113111Mr6

2:

FIG.I3

%Au...no %IKE

VI) MC I,lab ILLTc C A o:Z:47,uuER 57

L L-R4.0Eh OIAC,RANI OF LARORATOR.' SETuPTJ(MR,. TIN,n1 C RALANCiNG HCUir-

couplingng resistor of the plate supply. R2is coupled to the common coupling resistorof the screen supply, R4. through con-denser Cgfl, us indicated in Fig. 11. If weredraw this schematic again so as to takethe form of a more familiar coupling cir-cuit, we have Fig. 12. Here, it will be noted.the full potential difference which appearsacross R2 (400.000 -ohm resistor of Fig. 81is applied through Cg3 and through bothR3 resistors directly to the screen -grids ofboth tubes. If we assume that the control -grid to screen -grid voltage amplificationof both tubes is equivalent (for simplicityof explanation 1. then the residual in-stantaneous - 8 volts of Fig. 8 is applieddirectly to both screen -grids without anyadditional cancellation. This instantaneouspositive voltage also acts as a driving volt-age to the screen -grid of tube B to furtherincrease the negative swing of its plate.In actual practice, circuit values can beadjusted to automatically correct for anydesired range of variation between tubes.Laboratory tests, however, simplify thedetermination of optimum values for maxi-mum D.C. static correction, maximum A.C.dynamic correction and minimum loss ofoverall gain.

LABORATORY TEST SET-UPFor checking the degree of balance ob-

tainable, the laboratory equipment indicatedin Fig. 13 was used. The coupling trans-former T1 was used to obtain a push-pullsignal. Two vacuum -tube VU meters were

ALL ABOUT FREQUENCY MODULATIONused across each half of the push-pull inpusignal to enable exact adjustments of inputvoltages. Individual calibrated attenuatorswere used to vary the amount of inputsignal fed into each half of the push-pullstage.

It was found that when full signal wasfed into one grid and no signal into theother, a 50% balance occurred. In otherwords, one output grid developer: a voltage50% of the other and exactly 180' out -of -phase. With a 50% variation in input signal,80% balancing occurred. In other words,when half as much signal was fed into ore

input grid, as compared to the other, itsassociated output grid had 4/5 of the volt-age which appeared on the opposite push-pull output grid. This signal was alsoexactly 180° out -of -phase. Both of theseconditions represent extreme abnormalities.Over 100 combinations of input tubes werechecked for variations in voltage amplifica-tion. It was found that the greatest varia-tion of tubes produced a difference of lessthan 5% between both output grids.

How this output grid voltage is furtherbalanced by the action of the feedbackcircuit will be explained in the nextpart.

F. M. AUDIO AMPLIFIERPART III

ANUMBER of questions have been re-peatedly asked of the writer sincethe initial article describing this F.M.amplifier appeared. Among these were:

"Why is it necessary to extend therange of the amplifier from 13 to 30,000cycles?""Why is a 24 -watt amplifier requiredfor reproduction of phono or F.M. pro-grams in an average home?"Both of these questions are answered in

this article, after the technical descriptionhas been completed

THE BALANCED FEEDBACK CIRCUITThe voltage which appears in the bal-

anced 500 -ohm winding of the output trans-former is fed back to the cathode circuit ofthe 6SJ7 drivers through a bridged circuit.This particular feedback circuit can bestbe studied by redrawing the original cir-cuit, which appeared in part one, asshown in Fig. 1 herewith.

An analysis of this bridge circuit willshow that, under normal conditions, novoltage will appear from cathode to cathodeof the input tubes. The A.C. voltage acrossrk2 will he equal to 0. If the feedback resis-tors rb, rbl, or the cathode resistors rk, rkl,or the feedback windings, are unbalanced, avoltage will be present across rk2. As theinput tubes are operating in push-pull, thevoltage which appears across rk2 must bein -phase with one of the cathodes, and out -of -phase with the other. It therefore degen-erates with the cathode circuit with which

24

it is in -phase, and regenerates with thecircuit with which it is out -of -phase. Thisaction, in turn, tends to further balance thevoltage across the plates of the driver tubes.Its overall effect greatly increases the over-all dynamic stability of the amplifier.

The advantages of running the feedbackloop from the secondary of the output trans-former back to the cathodes of the inputtubes are as follows:(1) By embracing the output transformer,

the feedback loop corrects for frequencydiscrimination.

( 2) Most effective circuit stability is at-tained by coupling the balanced outputfeedback circuit directly to the push-pull drivers.

If the feedback voltages were taken di-rectly from the plates of the output tubes,it is apparent that compensation for dis-crimination within the output transformerwould not be effected. During the develOp-ment of this unit, a tertiary feedback wind-ing was checked, and it was found that adistinct phase shift occurred between theprimary of the transformer and the terti-ary winding. This latter winding was notalways in -phase with either the secondaryor the primary. Such a condition naturallyresults in feedback regeneration at somefrequencies. This confirmed a long-standingtheory that tertiary windings are not idealfor feedback purposes.

By alight adjustments of feedback re-sistors, they may be coupled directly to anyone of the output taps, so that any varia-


SOO.OHNIweNOIN5

rb

CATHODESOF ORIvEa

TUBES

OUTPUT- TRANS

I rb,

rK2

"BASIC BALANCED -FEEDBACK

1620DEPENDUPONS

APPL1CAT

rs

2.200 iL10 PA F.

5.000

2 1.4F

C'TOEOUALIZEO

SELF -BALANCINGINVERTER

ti0.1.-MEG

-50.000.n.

10.000.n.

TOCATHODESOF 6L6G".!

PRELIMiNAllv INPU TUBE CiaCulT,TWO OF THESE CIRCUITS ARE EMPLOTECI,

ONE FOB EACH INPUT

tions in the coefficient of coupling betweenthe used output terminals and the 500 -ohmline (if these terminals are not used), willnot have any effect upon the desirable ac-tion of the feedback loop.

THE EQUALIZED SELF -BALANCINGINVERTEROne of the major problems in developing

an ideal inverter is to be able to obtainequal voltages (out -of -phase) from eachside of the inverter output. Reasonable vari-ations in tubes should not produce objec-tionable unbalance. A basic circuit for apopular self -balancing inverter is given inFig. 6. In this circuit, balancing action isobtained by including a common grid -re-turn resistor (rg2) in the push-pull stagefollowing the inverter. Balancing action forvariation in the amplification factors ofVI and V2 is obtained by applying thedifferential voltage which appears across-g2 back to the grid of V2. While this actionis very effective, it does not provide a per-fect balance when VI and V2 are reasonablymatched. In fact, normal operation of thiscircuit provides an unbalanced signal at

the grids of the output tubes. While thisunbalance may not be serious, it neverthe-less, introduces distortion. To correct thiscondition, the equalized self -balancing cir-cuit, illustrated in Fig. 7, was developed. Inthis circuit, a common plate load resistoris inserted in series with both plateloads of VI and V2. Any unbalance in out-put voltage appears across r2 and is cou-pled through condenser cg to the grid ofV2. If this unbalance is opposite in phaseto the signal being impressed upon the gridof V2 through the dividing network rg andrgl, then degeneration takes place so as todecrease the output of V2, which in turn,equalizes the signals appearing on the gridsof the output stage. On the other hand, ifthe residual voltage appearing across r2,is in -phase with the voltage being impressedto the grid of V2 through the dividing net-work rg and rgl, then, regenerative couplingtakes place, to increase the output of V2.The great advantage of this equalized self -balancing inverter over the standard self -balancing inverter is that 100% balance isnormally attained.

In order to ascertain the relative effec-tiveness of both circuits, the amplification

FUSEPT

.2A)

[-----.\ ITC -0 a. OR 2 WHOM auLes.770 110V., A C. BUSS

.-- AFTER MASTER SWITCHIngl= -THE D.C. NEATER &ORLY (HUM RIPPLE LESS THAN I %)-

83

TO

HEATEROF PRE -ANAL WIER,

&INVERTERTUBES

2000

25v.

25

BALANCED LOADINGTERMINALS OF OUTPUT

WEB TRANS .RMANIEN5004frak

ALL ABOUT FREQUENCY MODULATIONfactor of V2 was altered. In one case theoutput voltage of V2 was normally adjustedto produce twice that of V1, and in theother case, it was adjusted to produce one-half of VI. Both of these adjustments weremade without either balancer in the circuit.Then both balancers were incorporated, andthe following data tabulated: (In order toevaluate the effectiveness of the balancingactions, the percentage of unbalance is givenunder various conditions).

Percentage ofUnbalance at

P.P. gridsCondition Self- Equalized

balancing Self -Inverter balancing

Inverter10% Perfect

balanceV2 Unbalanced +50% 20% 22.5%V2 Unbalanced -50% 9% 11%

It should be noted that while the standardself -balancing inverter provides a slightlybetter (by 2.5%) balancing action underwidely unbalanced tube conditions, the equal-ized self -balancing inverter provides for bet-ter balancing under normal operating condi-tions. It should be remembered, that thedata indicated, was obtained by unbalancingV2 50% in either direction. This representsa far greater change than ever encounteredin actual experience.

Balanced Tubes

SELECTION OF THE INPUT TUBESAs the input tubes and their associated

circuits determine the residual noise levelwithin the amplifier, it was decided to care-fully check all prevailing tubes and stand-ard circuits in an effort to attain a condi-tion which would provide the highest gain-to -noise ratio.

In conducting these tests, a wide varietyof tubes were set up in standardized cir-cuits. Both the gain and their noise weremeasured under a wide range of operatingconditions, so as to obtain the ,optimumgain -to -noise ratio. One of the preliminaryacceptable circuits is given in Fig. 2. Itwill be noted that the type 1620 triple -gridamplifier tubes- are used. It is not to be con-strued, however, that these are the onlydesirable tubes. As a matter of fact, anumber of other types may be used, depend-ing upon the ultimate application of theamplifier.

In checking residual noise, it was foundthat in many tubes, hum constituted a sub-stantial portion of noise. By substituting astorage battery supply, many "tummy"tubes had their overall noise reduced from8% to 10 db. It was therefore decided thata D.C. supply would be incorporated withinthe amplifier to heat the preamplifier andinverter tubes.

THE D.C. HEATER SUPPLYBy connecting the 1620s in parallel, and

then in series with the 6N7s, a 12 -volt 600ma. supply was required. Work done in ourlaboratory about 2 years ago, utilizing atype 83 to deliver 1A. at 24 V. proved theadvisability of building an "A" supplyaround this tube. Our lab records had shown1,000 hours' test without any measurablelecrease in emission. The "A" supply recti-fier system was therefore built around thistype of tube circuit, which is illustrated inFig. 3.

The writer feels that many readers willthink that the 83 would be considerablyover -worked in this circuit, in view of thefact that its published D.C. output currentrating is 225 ma., maximum. Its peak platecurrent, however, it will be noted, is 675milliamperes, maximum. This rating, though,is applicable to a 450 -volt condition. It ap-pears from empirical data, that when low -9r voltages are applied to the plates of the83, such as 30 or 35 volts, a much highercurrent can be drawn, and still obtain rea-sonable life from the tube. A number ofphotocell exciter lamp supplies, incorporat-ing four 83s and delivering 10 volts at 4amperes, have proven the dependability ofthis type of circuit. For sceptical readershowever, the amplifier can easily be rede-signed to accommodate standard tungarbulbs in place of the 83 rectifier.

It will be noted that a pair of R.F. chokesand bypass condensers are employed in thefilter circuit to avoid any disturbances frominterfering with A.M. tuners, should they beused with this amplifier.

26

PLATE AND BIAS SUPPLIESA study of the original power supply cir

cult will indicate that 2 rectifiers are em-ployed in a tandem power supply circuit. A5U4G supplies plate voltage to the drivers,preamplifiers, and inverter tubes, while a6V4G supplies plate voltage to the poweroutput stage. As a 5V4G is a slow -heatingrectifier, plate voltage cannot be applied be-fore the full bias appears at the control -grid of the output stage.

Thus by carefully designing the powertransformer and its associated filter thiscircuit affords increased life of power out-put tubes as compared to circuits enemy-ing 2 rapid -heating or 2 slow-headnit bisand plate supply rectifiers.

THE BALANCED OUTPUT CIRCUITAlthough the original circuit showed a

balanced output transformer equipped with4/8/500 -ohm taps, this transformer can besupplied with any variation of impedances.It has been standardized, however, with4/8/16/500-obm windings. The balanced na-ture of the transformer provides a wide va-


OUTPUTTUBES

B+.

FIG S

OUTPUT

CTR ANSA

1111121111.

MI IMIMEMEM.41 NMI=MBE

--UNRAL5NCE0 IMPEDANCE COMBINATIONS,.17V335/1/2/4/6/to/u/sito02050/66 op.s.

-STANDARD SELF -BALANCING INVERTER.NORMALLYufrali FURNISHES A 10%1,451404CM S.GNAI. TO THE PUSH -PULL CAWS-

-riety of impedances, which are obtainedeither by balanced or unbalanced loaning.A balanced loading circuit is illustrated inFig. 4. Figure 6 shows the unbalanced out-put terminals available, ranging from 9.175 -ohm to 166 ohms. It will be noted that a totalof 16 impedance combinations are available,ranging from 0.175 -ohm to 500 ohms.

Although the transformer may oe loadedin an unbalanced fashion, true balancedfeedback and push-pull action throughoutthe driver and power output stage still takesplace. An analysis of the unbalanced load-ing circuit diagram, will clarify this point.Regardless of where the load is applied, thevoltage from either terminal of the 500 -ohmline to ground, would be identical. If anyvariation does exist, it would be caused bya difference in the coefficient of couplingfrom the loading portion of the secondaryto the 500 -ohm terminal on the same side.It is a relatively simple matter, however, toover -design the output transformer so asto provide a unity coefficient factor underany conditions of normal unbalanced load-ing.

POWER OUTPUT RATING OF THEAMPLIFIERIn rating the power output of an ampli-

fier used for F.M. applications, the readershould dis-associate himself from conven-tional P.A. amplifier ratings, as unfair eval-uation will take place, if this factor is nottaken into consideration.

Ordinarily, a P.A. amplifier can safely berated up to 5% or 7%. In most P.A. appli-cations, this amount of distortion wouldnot be readily detected. In F.M. work, how-ever, it is imperative that the amplifier beoperated at not more than a total of 1%distortion. This precaution must be taken,as originally outlined, to prevent the am-plifier from becoming the bottle -neck ofdistortion in the entire F.M. transmission -reception chain.

Although the amplifier delivers a maxi-mum output of 30 watts, it has been ratedat 24 watts for 1% total harmonics. It laintended, however, to be normally operated

27

at an output level of 12 watts which pro-vides less than 1/2 of 1% total distortion.These unusually low ratings are advocatedso as to virtually eliminate distortion con-siderations from the amplifier. It should beborne in mind, however, that if the unitis operated at an average level of 3 watts toproduce unmeasurable harmonics, transient.ncreases of level of 9 db. will bring thepower output up to 24 watts with its in-tended 1% of total harmonics. Many so-called de -luxe F.M. radio receivers employrelativily low power output stages to ef-fect appreciable economies, particularlywhen large quantities of receivers are in-volved.

WIDE -RANGE RESPONSEThe development of an amplifier having a

response of from 13 to 30,000 cycles ± 1 db.,obviously increases its overall cost, andsometimes raises the question, "Why shouldI buy an amplifier with such a wide -rangeresponse, when F.M. broadcasts only runfrom 50 to 15,000 cycles? Furthermore, theaverage human being can not hear 30,000cycles."

To answer -this question intelligently, wemust first acknowledge the fact that thelatest findings amongst young listenerswith acute hearing clearly indicate that 30,-000 cycles CAN be perceived! Furthermore,fundamentals and sub -fundamentals, shouldbe reproduced, in order to avoid destructionof original tone qualities. This can easilybe proven by a difference in quality of re-sponse of bass drums or organ programswhen fundamentals are cut off. In view ofthe fact that the response range of am-plifiers has been continually increasingthe writer believes that it is only a ques-tion of time before the ultimate amplifierwill extend out to the outermost limits ofalma* hearing, and, if this can be accom-plished now, why shouldn't the amplifier beremoved as a restricting link in the chain ofreproduction?

We note that in the past, loudspeakermanufacturers have consoled themselves for


VP

EQUALIZED SELFBAL ANCING INVERTER NORMALLY FURNISHUrfa ES A PfiFECTLY BALANCED SIGNAL TO THE PUSH- PULL GRIDS -

restricted response by contending that no"program or amplifier can reproduce morethan 5,000 cycles." Record manufacturerscomplained that no phono pickup could re-produce more than 0,000 cycles, and pick-up manufacturers contended that no ampli-fier passed more than 8,000 cycles. This vi-cious circle naturally hindered projected improvements in any one branch.

These illogical assumptions really haveno place in modern communication equip-ment. If F.M. stations are forced by theF.C.C. to providc 50 to 15,000 cycles, thewriter believes it is only a question of timebefore this spectrum will be balanced andsome of the better stations will eventuallyextend this range to the very outer limitsof human hearing.

In providing this extremely wide rangewithin the present amplifier, any possibili-ty of early obsolescence is completely elimi-nated, for further extension of the rangeis obviously unnecessary, unless the humanrace during the process of evolution willacquire an extended hearing range.

Be. "Balanced Audio Spectrum." Mlle-Crld4.September. 1940. Page 184.

CONCLUSIONThe writer wishes to caution readers not

to compare this amplifier with conventionalpublic address units, by checking power out-put, distortion, hum, or tube components,as a number of essential features have beenincluded in its design, which are highlydesirable, in order to attain the full bene-fits from F.M. broadcasting.

Because of its technical excellence, it canof course, be used in any other applicationrequiring the ultimate in design and per-formance.

Additional information on the currentsand voltages at the outputs of thehigh -voltage and bias supplies of theF. M. Audio Amplifier are describedhere.The normal D. C. out of the 5V4G'sfilter system (high -voltage supply) is160 ma. at 590 V. (to ground.) Nor-mal D. C. out of the 5U4G's filter sys-tem (bias supply) is 22 ma. at 185 V.(to ground).

28


CHAPTER IV

F. M. SERVICE

PART 1

Antenna Installation and Service

THE advent and increasing popularityof Frequency Modulation receivershave presented new problems to theradio technician. Frequency Modulated

receivers require installation and service:adjustments somewhat different from stand-ard procedures. Knowledge of these dif-ferences is essential to develop and applythese new techniques.

From a standpoint of design and circuit:.rrangement, F.M. receivers are similar tothose with which we are already familiar.It is the fact that F.M. receivers operateon ultra-highfrequencies, that new con-siderations are introduced, considerationsinvolving antenna installation, receiveralignment, oscillator stability, and insula-tion leakage. To facilitate discussion of thesubject, the material herein is presented in2 installments, Part I-Antenna Installa-tion and Service, and, Part II-ReceiverService.

RECEPTION CONDITIONS ON F.M. BANDTo understand and appreciate the im-

portance of a suitable antenna and itsproper installation, a brief explanation oftransmission and reception conditions onthe 42-50 megacycle band is of consequence.

The strength or signal level and distanceover which signals are received at theantenna depends upon several factors:power of transmitter, location and type ofantenna, time of day and season of year.The theory of radio wave propagation mostgenerally accepted is that advanced byl'rofessors Kennelley and Heaviside, inwhich 2 radio waves are radiated from atransmitting antenna, (a) the ground orsurface wave, and (b) the sky wave.

The surface wave follows the curvatureof the earth and is absorbed by the earth,metallic deposits, steel buildings, hills, trees,and bodies of water. However, the surfacewave is steady and reliable in that it

29

travels an equal distance both day andnight. When waves of a high frequency aretransmitted, less of the radiated energy istransformed into surface waves (possiblybecause it is absorbed faster by the earth)and more of the energy is changed into thesky wave. This latter wave does not 'followthe surface of the earth, but travels instraight lines and behaves much like lightand radiant -heat waves.

The sky waves travel out from the earthin all directions and are thought to en-counter layers of ionized gas in the earth'satmosphere or ionosphere. These layers ofgas reflect and bend a portion of the skywave back to earth, so that signals maybe received at considerable distances far be-yond the range of the surface wave. Partof the sky wave is also absorbed or passesdirectly through into the ionosphere. Theamount of reflection and absorption of thesky wave is dependent upon the densityand ionization of the gas layers, caused byultra -violet radiation from the sun, whichionizes the gas molecules and produces freeelectrons. Since the degree of radiationfrom the sun and its influence is a variablefactor changing with the time of day andseason, the density and height of the"Heaviside Layer" above the earth is al-tered. For this reason, the amount ofenergy reflected or refracted to earth andthe angle to which the waves are bent islikewise a variable factor.

The degree of reflection of the sky wavealso depends upon the frequency. Thehigher the frequency of a propagated radiowave, the farther it penetrates the ion-osphere and the less it tends to be bentand reflected to earth. At frequencies ashigh as 42-50 megacycles, the F.M. trans-mission band in which we are interested,radio waves are bent so slightly that theyseldom return to earth.

Thus, it can be seen that reliable, con-sistent, and dependable reception of signals

FIG 2

ALL ABOUT FREQUENCY MODULATIONin the F.M. band during both day andnight, because of their ultra-highfrequencynature, is limited to the surface wave, ortine -of -sight propagation. For all practicalpurposes, the surface waves may be likenedto beams of light traveling straight outfrom the transmitting antenna to the hori-zon and beyond into space. To receive thesewaves, the receiving antenna must be within"seeing" distance of the transmitter. Thegreater the height of the transmitting andreceiving antennas, the greater the horizonrange or area over which dependable sig-nals may be received. Although receptionof ultra-highfrequencies beyond the hori-zon range of a transmitter is not impos-sible and is frequently reported, strengthof signals is not constant, and thus unre-liable.

CURRENT\ VOLTAGE

ANTENNA ...i, w- -'"RESONANTE ..TO Y ..peRvE- ..--

leNGTN

lAt

CCCC N6TNVOLTA GE AND CURRENT DISTRIBUTION

ON A HALF -WAVE ANTENNA -

ANTENNA Vs- ANTENNA / ANTENNA /VEINWESENGTIV 1.0A6 WOVELENGEN /DVS PorfieNOTA C /ONO

EVRECTiviry CNARACTER/S TICS OF AORIZONTALANTENNA, FROM ABOVE --

( I/4 btutvei !NanoEACH) ROOS

AORCELANNSCRIP, eve

ye-wAVEZ ENGTI. EACH)NE. /2 SOLIOCDPPER WIRE

SUPPORTINSULATORS

TRANSMISSIONLINE

STANO-OPPffiv.SULATDR3

NALF WAVE 0/POLE ANTENNAUSING TUBING OR w/RE

DIPOLE ANTENNA ARMS

PORCELAIN RIBBEDcINSULATOR

.416,

PORCELAINSCREW -EWE

MOuNTINO0,RoLe T- PRACRET0 S4.ppORTINO ARMS !

D/POLIf STANDARDFIBER /NsimArwo BLOCKS

Ft 4

STEEL ANGLEIRON

DIPOLESTANDARD

/DIPOLE ANTENNA ARMSScllows

SuPpoRT APFOR DIPOLERs

WOOD SO .PORTINGEVAN (EGO.

SNARED)

ENROL! CONS UCTION, ILLUSTRANNGMETHODS OF SUPPORTYNG 0/POLE ARMS

SCREWSHOLD/NOSTANOAND

--METHODS OF RA/.!/NO DIPOLE ANTENNAWISNI EV EXTENONIO DIPOLE STANDARD --

30

F.M. ANTENNA REQUIREMENTSProbably the 2 greatest advantages pos-

sessed by F.M. receivers, when comparedwith those of conventional standard designor amplitude modulated receivers, are (1)substantial freedom from natural static and"man-made" interference, and (2) extendedfidelity and dynamic range of reproduction.Most listeners are more concerned withthese aspects of F.M. receiver superioritythan with other claims. To derive thesebenefits, only one requirement must be sat-isfied, that of adequate signal pick-up, tooperate the limiter stage of the receiver.Without sufficient signal to saturate thelimiter tube, elimination or clipping ofsignal peaks will be incomplete, and recep-tion will be generally unsatisfactory, as aresult of noise amplification and distortion.

To assure adequate signal pick-up, atten-tion must be directed to the antenna sys-tem. Of all factors contributing to strongsignal pick-up of F.M. signals, antennaheight is the most important. Hundreds oftests have proven conclusively that thestrength of the signal picked -up by the re-ceiving antenna is almost directly propor-tional to the height of an antenna. Forexample, by doubling the height of anantenna, an increase of 100% in signal levelis secured. For every foot the antenna israised, a proportionate gain results. Ofcourse, there must be a substantial increasein antenna height before a real improve-ment is noted.

Although the F.M. receiver provides noisereduction by reason of its limiter stage, therange over which the limiter operates isnecessarily restricted. For this reason, everyprecaution and device must be employed toreduce, as much as possible, all noise pick-up by the antenna system. The principalsources of interference are diathermy andX-ray automotive ignition, signfikshers, neon signs, oil burners, high -volt-age power transmission lines, electric light-ing and power plants, and all electrical

ALL ABOUT FREQUENCY MODULATIONapparatus of a high -frequency nature. Thisinterference may be picked up ty bothdirect and indirect radiation.

Locating the aerial in a noise -free areaand utilizing a balanced transmission lineto conduct signal voltage to the receiver,is the general approach toward solution ofthis problem. Increasing the height of theantenna not only serves to increase thesignal strength, but aids in lowering thenoise level.

LOSSESAt ultra-highfrequencies, the matter of

insulation is of considerable importance.Materials such as unglazed porcelain andmolded bakelite, which are satisfactory atpowerline and broadcast frequencies, arerelatively unfit at U.-H.F. because of lossesresulting from high power factor and ab-sorption. This is particularly true of thecheap rubber, impregnated -cotton insulatedtwisted -pair wire so often mistakenly usedfor transmission lines. Only insulated wire,insulators, ternfinals, and terminal blocksof the highest quality should be coun-tenanced in the installation of an F.M. an-tenna to avoid all possible losses and conse-quent reduction in signal strength.

STRAPS MELDBy SCREWSOR LAO

BOLTS

COPING

RAGGING ORCHIA/NEI,'

BLOCKS HELD 70FLAGGING WITH LAGBOLTS AND EXPANSIONPLUGS IN BRICK OR CEMENT

--MOLINE/NG DIPOLE 70Una FLAGGING OR CHIMNEY

Fortunately, the current and voltage dis-tribution on a half -wave doublet or dipoleantenna, which is the most effective andpopular type employed for F.M. reception,is such that the voltage at the center of thedipole is theoretically zero. This is shown atFig. 1. For this reason, and since the 2sections of the antenna are generally sup-ported at the center, it is not absolutelyessential that extremely -low -loss insulatorsor supports be used for this purpose.

Losses in a twisted -pair transmission linemay be high due to high carbon content ofthe rubber insulation, and absorption of

31

DIPOLESTANDARD

BRICKCHIMNEY

STRAPSBOLTS ANO

WOOD NUTS.5 TRIPS

ROOP

- METHOD OF moriHT,HGmien DIPOLE 70 CHIMNEY

moisture by outer cotton insulation. Onlylines that are sufficiently weather -proofedand storm -proofed, and use a high gradeof rubber, are acceptable. Ordinary lampcord is definitely "out" because of lack ofweather -proofing. Some manufacturers sup-ply transmission line cable especially suit-able for F.M., with a high percentage purelatex rubber insulation on each wire and anouter covering of the same material. Wheretransmission lines must be run distances ex-ceeding 200 feet. losses become prohibitivelyhigh. In such cases, concentric or coaxial linecables are desirable and essential. Concen-tric cables consist of a solid or strandedconductor, rubber insulated, enclosed in acopper braided shield, with over-all weath-er -proofed cotton insulation. Coaxial cableis similar with the exception that insulatedbeads or spacers are used to maintain afixed clearance between the solid conductorand conducting shield. Of the two, coaxialcable possesses lower losses per foot, but isdecidedly more expensive.

DIPOLESTANOARO HELDON BLOCKS BySTRAPS AND SCREWS

WOODENBLOCKS

DIPOLESTANOARO

WOODSTRIA.

3uPPORTS

MOUNTING DIPOLE ON GAacro4POS,uram HOMES OR GARAGES-


O'POLESTANDARD

S.RING

SCREWSTEEL

GUY WIRESBRACKET EYES

DIPOLE STANDARD

11111f1SOCKET FOR

BLOCK OFWOOD BOLTED.

TO ROOF --- LAOBOLTS

FORTNER METN003 OF MOUNTING0,1201E-USE OF STEELBRACKET1it OR mOunerwG BLOCA--

'1.1'eotTS MOLDINGDIPOLE STANDARDTO BRACKET

/MSS NOLOING-BRACA-e T TO

ROOF

a/RoLESTANDARD

niAAP"

IMPEDANCE MATCHINGWhen a manufacturer's antenna kit is

used, the question of impedance matchingusually presents no problem, since these kitsare available complete with dipole, trans-mission line, insulators, and matching trans-former. The importance of impedancematching between dipole, transmission lineand receiver input cannot be over -empha-sized. The advantages gained by locating theantenna high in a clear area may be mini-mized through losses resulting from mis-match.

The impedance at the center of a half -wave antenna in free space is 73 ohms. How-ever, because of the presence of insulatorsand dielectric material in the vicinity of theantenna, and the height of the antennaabove a conducting ground as well as theefficiency of the ground, this value is al-tered considerably. As a result of these fac-tors, the impedance of the average half -wave antenna may be taken at approximate-ly 100 ohms.

The surge impedance per unit of lengthof a transmission line consisting of 2 paral-lel conductors, depends upon the size andspacing of the conductors, and the dielectricconstant of the insulation between the con-ductors. The larger the conductor and thecloser the spacing, the lower is the surgeimpedance. For example, the impedance ofa twisted -pair line consisting of No. 14 rub-ber -covered, cotton -braid insulated wire,similar to that used in house wiring, is ap-proximately 100 ohms. This type of wire,when suitably weather -proofed and connect-ed to obtain an impedance match in a man-ner to be described later, may be used as ashort transmission line. In any event, toobtain the greatest transfer of energy fromthe dipole to receiver, a line of the correctimpedance must be employed. Transmis-lion line of the twisted -pair, concentricind coaxial type, of various stated surge

32

impedances, are commercially available fromlarge supply sources.

Of equal importance in the matter ofimpedance matching is that of terminatingthe transmission line in the correct charac-teristic impedance. By this is meant that theimpedance of the receiver input must matchthat of the transmission line. Most receiversare designed-witit antenna inputs whose im-pedance varies from 100 to 300 ohms, neces-sitating the use of an impedance matchingdevice in some cases to obtain maximumperformance. A description of a simple auto -transformer to serve this purpose is latergiven.

Because of the short physical length of anF.M. half -wave antenna, and character ofthe transmission line, swaying in the windmay produce signal variations and noiselevel changes of large magnitude, sufficientlygreat to place these variations Beyond thecontrol of the leveling action of the limiterstage in the receiver. For this reason, theantenna and transmission line must berigidly mounted, by whichever means areavailable and necessary, to prevent possiblesignal voltage changes and distortion as aresult of reception of out -of -phase signals.

POLARIZATIONMost F.M. transmitters employ antennas

which are in a horizontal plane and radi-ate energy which is horizontally polarized.To receive this signal, the receiving antennamust also be horizontally polarized to havea maximum voltage induced in it. Some fewstations employ vertical polarization, thusmaking a vertical receiving antenna neces-sary for maximum performance.

The horizontally polarized half -wave an-tenna is more widely used because of anadvantage insofar as noise pick-up is con-cerned. Since noise originates from nearbysources, and a good part of it, especiallyautomotive ignition interference, is verti-cally polarized, less of this noise voltageis induced in a horizontally polarized anten-na than in one that is vertically polarized.

On the other hand, the horizontal half -wave antenna is directional, the greatestsignal voltage being induced when thelength of the antenna is placed in a posi-tion which is broadside or at right -angles tothe signal source.

Antenna polarization is important at dis-tances relati;iely close to the transmitter.At farther distances, the plane of polariza-tion of U.-H.F. waves has been known tochange as much as 90°, so that correctpolarization becomes a matter of experi-mentation to provide best signal pick-up.Tilting the dipole at various angles fromthe horizontal to the vertical position, andchecking receiver response to these changes,will determine which position is best. Whenan antenna for an P.M. receiver is installed

ALL ABOUT FREQUENCY MODULATIONin areas close to several transmitters, oneof which may radiate vertically polarizedwaves, it may be necessary to utilize botha vertical and horizontal dipole to receiveall signals with sufficient signal strength.Or a compromise may be effected wherebyone dipole is used at some angle between ahorizontal and vertical plane.

A resonant antenna, such as the half -wave dipole, possesses marked frequencydiscrimination by which signals at otherthan the resonant frequency of the antennaare sharply attenuated. Since the F.M.band of transmission covers a wide range offrequencies, from 42-50 megacycles, this fre-quency discrimination would prove unde-sirable were it not for the fact that theloading of the transmission line producesa broad frequency response. It is customarytherefore to cut or design the half-wa' 2 an-tenna to resonate at the center of the F.M.band, at approximately 46 megacycles. Itmay be advisable, inasmuch as line lossesincrease with an increase in frequency, toresonate the antenna at 47, or possibly 48megacycles to compensate for these losses.

DIRECTIVITYBecause of the fact that the horizontal

half -wave antenna is bi-directional, this dis-crimination may prove undesirable whenthe signals of widely separated transmittersmust be received at one location. The direc-tional characteristic of a horizontal half -wave antenna is illustrated in A of Fig. 2.

As a general rule, this situation provestroublesome in urban areas or close to anumber of transmitters which lie in variousdirections, one or more of which may beoutside the directional pattern of the an-tenna. No such problem is presented whenthe receiver is far removed from this area.As a matter of fact, in the latter instance,attempts are usually made to increase thedirectivity of the horizontal antenna. Onesolution to this bi-directional effect isthrough the use of a vertical half -wave an-tenna, but this results in a loss of signalstrength and an increase in noise pick-up.In some cases, an antenna designed to res-onate at a frequency equal to twice or 3times a half -wavelength, is employed toovercome the bi-directional effect of a hori-zontal half -wave antenna. The changes inthe directional pattern from that of a half -wave are shown at B and C of Fig. 2.

It can be seen that the angle of the nullpoints is decreased thus producing lessdirectional effects, but the maximum signalvoltage possible to be induced into the an-tenna are also less. In addition, the resist-ance of such an antenna increases consider-ably, from 25 to 35%, necessitating a trans-mission line of higher impedance andfurther impedance matching at the receiver.Another method is that of orientating thehorizontal half -wave antenna in some cons -

33

-GIN WIRE STRAIN INSIRATOR'EGG. roar

SOLDEREDSPLICES

GOOD*ROMP TAPETroNrLy.OvErreaCm SPLICE

OWE -ALLPuRaERrape AND

W/C.1101VTAO(

-Aarro00 OF STAGGERINGSAL 'CEO CONNECTIONS WHENEY rENO/NO riPANSmrssiore

LIME --

promise position, whereby adequate signalpick-up is obtained in all directions.

As mentioned previously, the half -wavedipole antenna is usually designed to reso-nate at a frequency in the center of theF.M. band. When calculating the length ofa half -wave antenna, it must be rememberedthat the physical length usually averages5% less than the electrical length. This isdue to end effects occasioned by the pres-ence of insulators, and the fact that theantenna has resistance. A simple conversionformula, which considers these end and re-sistive effects, for computing the physicallength of a half -wave antenna, is the fac-tor 467.4 or 468 (which is accurate enough),divided by the desired resonant frequency ofthe antenna. 'f he figure obtained is thelength of the entire antenna in feet. Eachhalf of the dipole or doublet should there-fore be cut to half this amount. An antennawith each half of the dipole cut to 5 feet, 1inch, has been found satisfactory for thereception of signals in the F.M. band.

F.M. ANTENNA TYPES AND INSTALLATIONAt locations close to F.M. transmitters

and where the problem of adequate signalpick-up and noise pick-up does not exist,any short length of wire will serve as theP.M. antenna. When operating conditionsare favorable, an ordinary inverted -L typeantenna of 20-100 feet will provide satis-factory reception of F.M. signals. Upon sev-eral occasions, an antenna comprising a20 -foot length of wire extending downwardfrom a window was found sufficient to ob-tain good reception on an F.M. receiverlocated 25 miles from a number of trans-mitters. Operating conditions were ideal,however, the noise level being particularlylow with the receiver in a home built on ahill with a good line -of -sight. The distancefrom a transmitter that such an antenna,or one of the inverted -L variety, will pro-vide a signal of sufficient intensity and highsignal-to-noise ratio, is a matter of con-jecture and must be left to trial. In mostinstances, an antenna efficient at U.-H.F.must be installed.

The most satisfactory antenna for F.M.reception is the horizontal half -wave dipole.Essentially, this Antenna consists of 2metal -tubular rods or wires placed in linewith each other, as shown in Fig. 8. Themetal rods may be either copper or alumi-num. Solid rods are not recommended. To

ALL ABOUT FREQUENCY MODULATIONDIPOLE

REFLECTORSt GHTLY LONOER'war. 92-wwwILENOTH

TRANSN.S.SIONLINE

TRANSMITTERS

DRECTroN

RimOReATKA,

00 TO TA )WAVELENGTH

STANDARD

-ADDED REFLECTOR 7D0/POLE TO INCREASE'SIDA/AL LEVEL ANO oiacer/y/rY

obtain the requisite rigidity, only wire ofa heavy gauge, either No. 10 or No. 12,should be used, supported by low -loss in-sulators. The transmission line must be runat right -angles to the dipole for at least a14 -wavelength, at least 6 feet, before anybends in the line are made.

ROOF MOUNTINGSVarious methods have been devised to sup-

port the dipole arms. These are merely me-chanical arrangements and usually have nobearing upon electrical efficiency. The onlyrequirement is substantial construction.Several of these arrangements are shown inFig. 4.

At A, a 2 -piece cast aluminum bracketassembled by means of a number of boltsand nuts, is used to couple the 2 dipole sup-porting arms and standard. A center sup-porting insulator of ribbed construction isbolted to the bracket. Glazed porcelainscrew -eye insulators serve to support thedipole rods at the far end of the support-ing arms. The rods are threaded onto screwsemerging from each end of the center sup-porting insulator. The transmission line isconnected to these screws internally. Lami-nated fibre or hardrubber blocks are usedto anchor the dipole rods in position in thearrangement illustrated at B; and, an egg -shaped wood form is employed for the samepurpose at C.

In manufactured kits, the standards aresupplied in 5 to 6 foot lengths of straight-grained knot -free material, and more thanone is used to elevate the dipole as high aspossible. These sections are joined by meansof a 2 -piece sheet -iron or cast -aluminumbracket held by bolts and nuts, screws, andelectricians' strap, as shown in Fig. 5.

Methods of anchoring and supporting theantenna vary in each individual case. Gen-erous use is made of expansion plugs, lagbolts, and electricians' strap.

On apartment buildings and privatehomes, the dipole may be erected as in Fig. 6.The wood blocks are held in position withlag bolts whica are fastened in expansionplugs snugly fitted into holes or openings in

34

the brick, stone, or cement, made by a"star" drill and hammer. The standard forthe dipole is mounted by straps and lagbolts. When brick or atone chimneys serveas the "foundation" for the antenna, it isessential that the construction be solid andsubstantial. Too often, loose bricks and ce-ment prevent firm insertion of the expan-sion plugs. Also, use of the hammer and"star" drill may dislodge bricks and cementand so weaken the chimney that solid an-choring of the antenna is impossible. Thedangers of a weak chimney further bur-dened by the weight and stress of an an-tenna in the wind are apparent. When suchchimneys offer the only available means ofsupport for the antenna, the constructionshown in Fig. 7 has proved safe and sub-stantial, provided suitable guys are em-ployed. It may be seen that 4 lengths of2 x 4 in. or 2 x 3 in. lumber are held inposition by long bolts and nuts. These boltsare procurable from large hardware supplystores or may be ordered from the localblacksmith or iron -works. The dipole stan-dard is held erect by straps and lag bolts.

The problem of installing the dipole onhomes with peaked sloping roofs often pre-sents itself. When access to the chimneymay be gained without danger of slippingor falling, the construction already shownin Fig. 7 is recommended. Alternative meth-

the dipole are illustratedin Fig. 8. When the eaves of the buildingdo not jut out beyond the vertical sidewalls, the standard or the dipole may besecured directly to the wall. Otherwise, it isnecessary to build up or pile up a sufficientnumber of wood blocks, securely fastened tothe wall and to one another, so that thestandard may clear the eaves and be erect-ed vertically.

Another method of installing the dipoleemploys a steel bracket or wood block whichserves as the base for the dipole. This isshown at Fig. 9. The bracket is bolted to theroof and the dipole standard held to steelcross -bars by means of threaded U -bolts.The wood block which is a 12-18 ft. lengthof 2 x 4 in. lumber, is bolted to the roof.The dipole is mounted in a hole or socketmade in the wood block. In the latter meth-od, guy wires are absolutely essential, sincethis is the only means of holding the dipoleerect and steady. The methods and mannerof erecting the dipole antenna herein de-scribed are by no means the only possiblearrangements. Others will suggest them-selves.

TOWERSDipoles which "tower" or rise more than

b ft. into the air usually must be supportedrigidly to prevent swaying. This is done Plotonly to assure good reception but to preventweakening of supports and excessive strain-


ing of mooring bolts and plugs. For use asguy and supporting wires, No. 10 solid orstranded galvanized -wire "rope" should al-ways be employed to effect a "permanent"installation.

Anchoring of the guy wires may takeany number of means. Probably the simplestof all methods is the use of large eye -screwsfastened securely to any solid wood mooringor to expansion plugs inserted into brickwalls. Turnbuckles should be employed totake up the slack in long cables when thewires cannot be tightened otherwise byhand. Of particular importance at this pointis the necessity of breaking up the guywires by insertion of strain insulators ofthe "egg" variety to avoid possible res-onance effects at the receiving frequency.For our purpose, guy -wire lengths should al-ways be less than 10 feet. This is illustratedat Fig. 10. Although any system which willsupport the dipole is suitable, the 3 -guy -wire arrangement has proved most satis-factory. In any event, the guy wires mustbe kept out of the field of the dipole.

The generous use of tar compounds andsynthetic resins is recommended for weath-er -proofing lag bolts, straps, supports anddipole standards. When a steel or wood baseassembly is bqlted to the roof, applicationof a tar compound is essential to avoid leaksand seepage in inclement weather. This ma-terial is available from many supply sources.

TRANSMISSION LINEThe transmission line must be run at

right -angles to the dipole for at least 5 ft.Avoid right-angle or all unnecessary bends,and doubling back of the transmission line.Tape up the line where the insulation isslit to make connection to the antenna toprevent entrance of moisture. Although itis best to keep the transmission line in onepiece from antenna to receiver, when it isnecessary to extend the line, solder and tapethe lengths of line, but stagger the connec-tions as shown in Fig. 11. Only high-graderubber tape may be used at any point in anF.M. installation. Ordinary friction tape in-troduces high losses at ultra-highfrequen-cies. Every means must be employed to re-duce this possibility of leakage.

Anchoring of the transmission line shouldbe accomplished without danger of snapping.Protect the line from abrasion at all pointsof contact. Use standoff insulators whereverpossible to keep the line clear. Insulatorsof the "nail -it" knob type should be avoided,since with these there is danger of bruisingthe insulation of the line. The transmissionline may be brought into the building invarious manners. The porcelain "feed -through" insulator is practical. When a holeis drilled in the window casement or framefor the line or feed insulator, it should bedrilled downward from inside of building to

35

prevent entrance of moisture. The use of thecommon lead-in strip should also be avoidedas the insulation is usually inadequate andnecessitates a break in the transmission line.

Lightning protection and installation asprescribed by National and local Boards ofFire Underwriters, and in accordance withlocal fire and building department codes, areessential. Technicians should familiarizethemselves with these regulations, such as,height of antenna above building, type oflightning arrester and type of ground.

The ground connection for the receiver isimportant. Although in A.C.-operated re-ceivers, the "ground" is obtained throughthe line bypass condenser or capacity of thepower transformer windings, reliance uponthis grounding effect is not advised. Addi-tion of a good ground connection oftenspells the difference between good satisfac-tory reception and noisy operation.

ANTENNA SERVICE POINTERSAfter the F.M. installation is complete

and reception is found satisfactory, no fur-ther thought of F.M. antenna requirementsis necessary. When operation of the receiveris generally poor and it is known that thereceiver is not at fault at the time of instal-lation, we must look to the antenna instal-lation.

The cause for unsatisfactory F.M. re-ception may be attributed to many factors:inadequate signal pick-up, excessive noisepick-up, incorrect polarization and directivi-ty of dipole, and losses due to impedancemismatch and poor insulation, any or allof which tend to produce weak, noisy anddistorted reception.

In areas close to transmitters, the ques-tion of adequate signal pick-up does not ex-ist. At remote locations, this considerationis of importance. It is assumed that thedipole was erected as high as possible inthe first place. The method generally em-ployed to increase signal pick-up in thiscase is by the use of a reflector.

This is a metal rod, similar to that of thedipole itself, but slightly longer, placed par-allel with the dipole from 2 to 5 feet be-hind it, as shown in Fig. 12. This proce-dure not only greatly increases signal pick-up from one direction, but increases the di-rectivity of the antenna, since signals ap-proaching from the rear are greatly attenu-ated. This latter result is especially advan-tageous when it is desired to reduce noisepick-up from a nearby source. The reflectoris used just as often to obtain this effect asto increase signal level. No electrical con-nection exists between dipole and reflector.In instances where it is necessary to obtaina further increase in signal strength, adirector is used.

The director is a metal rod slightly small-er in length than the dipole, and placedabout 2 to 5 ft: in front of the dipole. This

VONAL..sooRCE

Ll

folrECTOR

ALL ABOUT FREQUENCY MODULATIONcommaion will provide maximum signalvoltage transfer.NALFve

DIPOLE

REFLECTOR

StraGE5rFO4WDSrAniCE3 OF

F M. ANTENNAI/ /0FrL2 SFr. for

!OF'. 41,113.01. 2ro5 Fr illecorww022,05Arl offptrr.

FM Anoro-AiNA WITH OiRfOrOV ANO AVFLECPOrtrr,,CFEAsZ 5OrIA, [Frei AND 0recrtwrr-

is shown in Fig. 13. Approximate lengthsand distances are given.

Before installing a reflector and then adirector, it is common practice to rotate thedipole to change its directivity, although itmay have been positioned properly (atright -angles) with respect to the transmit-ter, in order to increase signal pick-up.INTERPHONE

When 2 men are engaged in the instal-lation of the antenna, a telephone or othermeans of communication may be rigged up,so that the effect of rotating the antennato various positions upon reception at thereceiver may be ascertained. On a 1 -manjob, this is more difficult, since checkingreceiver response to the change in directionmust be made after each change. The direc-tivity of a horizontal half -wave antenna iscritical, depending upon the distance fromthe transmitter. Directional changes of onlya few degrees are often sufficient to affectsignal strength considerably. These factsapply also to polarization of the antenna.The tilted position at which a stronger sig-nal may be received must be checked aftereach change in polarization, but thesechanges are not critical.

When a transmission line is used to cou-ple the dipole antenna to the receiver, lossin signal strength is possible due to impe-dance mismatch, when the surge impedanceof the line is not nearly that of the anten-na. The loss due to mismatch often hasbeen found to lie with the use of a trans-mission line whose impedance was higherthan that of the antenna. To correct thiscondition, the ends of the transmission line,where they connect to the dipole, may befanned out, as shown in Fig. 14. This is pos-sible since the impedance of a half -waveantenna, which is at a minimum at thecenter, increases toward the outside ends.The distance of the fanned portion of theline from the inside ends of the dipole de-pends upon the line impedance. It is neces-sary. therefore, to fan the leads out a littleat a time, noting the effects upon reception,until a point is found on the antenna whichmatches the line impedance. This point of

36

MATCHING TRANSFORMER

Another cause for loss in signal level isdue to mismatch between the transmissionline and the receiver input circuit. Onemethod used to effect correct matching ofthe 2 impedances at this point is throughthe use of a matching transformer or auto-transformer. The construction of such deviceis relatively simple.

A total of 28 turns of No. 18 enamel- orcotton -insulated wire is wound on a 1 -in.form, and tapped every 2nd turn. The cen-tei-tap or 14th turn is grounded to the re-ceiver. Referring to all taps with respect tothe center -tap, there are 7 pairs of tapswhich connect to an equal number of turnson the coil. In other words, taps No. 1 areeach 2 turns from center -tap, taps No. 2 areboth 4 turns from center -tap, etc. By suit-ably connecting the transmission line andthe balanced receiver input to the couplingcoil, a correct impedance step-up or step-down match may be obtained.

When the transmission line is of lowerimpedance than the receiver input, it maybe connected across a fewer number of turnsthan the number across which the receiverinput is connected, to provide the step-upratio. When the line impedance is higherthan that of the receiver input, the receiverinvit is connected across fewer turns thanthe line. Once the correct impedance ratio isfound, the coil may be enclosed in a suitableshield and all leads soldered to the correcttaps. The nature of the autotransformer re-quires a balanced receiver input, whereinneither end of the primary of the antennacoil is grounded. Lack of this balanced con-dition will unbalance the transmission lineand introduce undesirable effects. Fortun-ately, almost all F.M. receivers have thisbalanced input.

Poor P.M. reception may be the result ofa high noise level, despite adequate signalpick-up. In such cases, it is assumed thatthe dipole has been erected high in a noise -free zone and a balanced twisted -pair trans-mission line has been used to connect thedipole to the receiver. Unless the receiverhas a balanced input, the advantage or pur-pose in using a balanced line is lost, andnoise voltages induced into the transmissionline may be induced into the antenna coilof the receiver. Because the conductors ina twisted -pair line are close, any voltageinduced in them, whether it be signal ornoise voltage, will be equal and in -phasewith each other so that their polarities arealways similar. Consequently, the polarityof the ends of the line which connect tothe primary of the receiver input trans-former are both either positive or negative.

ALL ABOUT FREQUENCY MODULATIONTherefore, no current will flow in the coilsince there is no difference of potential, andno noise voltage will be induced because ofthis inductive relationship. However, somenoise voltage is always transferred to thereceiver because of capacitative coupling be-tween primary and secondary of the anten-na coil.

REDUCING CAPACITY COUPLINGTwo methods may be employed to reduce

this capacity coupling. One involves the useof an auxiliary transformer in which agrounded Faraday electrostatic shield is in-terposed between primary and secondarywindings of the transformer, This trans-former is connected between the transmis-sion line and the receiver, and it is unim-portant whether one side of the receiverinput is grounded or not. The second meth-od, and one which has been employed withsuccess is the use of a center -tapped coil,such as the impedance matching autotrans-former described, connected between thetransmission line and receiver. The center -tap of the coil is grounded. The pair of tapsto which the receiver input and lines areconnected is best determined by trial. In anyevent the receiver input must be balancedor ungrounded.

Probably the most frequent cause for lowsignal pick-up is the unwarranted use oftransmission line with low-grade rubber in-sulation. The losses and leakage at U.-H.F.of this type of wire are great enough topractically short-circuit the line and there-fore the signal voltage. The use of line withgood rubber insulation must be stressed.Generally, transmission line rubber insula-tion which possesses good elasticity andstanding "plenty of stretch," may be con-sidered "live" enough for use as a line

1-0,45,2E EYOSD/POLL ARMS

1:)/00,e

FANNED )wANSML33.ONL,Ne Ze403 ,.001,nnoeCA,CeMA `-erAo

-A-4N/VINO M4NSWII53,0N LLNE LeA0.3TO Oar., ,mPerAWCE rwArC.,OrrWeEn1

0/00:5 44,0 LLNE ov,".?bvi0304,Ce--

in an ultra highfrequency antenna systemWire with rubber insulation which can b.torn or peeled easily without stretching 1.1

"dead" rubber and usually presents highlosses.

Although there is no doubt that the com-parative bulk of good transmission line isunsightly from an artistic point of viewwhen used in the interior of a building.twisted or parallel lamp cord, which is cer-tainly more appealing, should never be usedto continue the transmission line from thewindow to the receiver. Ordinary twin lamp -cord conductor offers high losses at F.M.frequencies.

There is much to be written of F.M. in-stallation and service, more than space lim-itations will permit, but an effort has beenmade to include most essential requirementsand considerations. To sum up briefly, theF.M. antenna system should possess the fol-lowing characteristics:

(1) Dipole as high as possible.(2) Good low -loss insulation.(3) High-grade transmission line.(4) Proper directivity and polarization.(6) Correct impedance matching.(6) Balanced line and receiver input.

37


PART II

RECEIVER ALIGNMENTAND DIAGNOSIS

N servicing Frequency Modulation re-ceivers, knowledge of the fundamentaloperating differences between AmplitudeModulation and Frequency Modulationreceivers is essential. The subject of F.M.antenna and installation was covered ascompletely as space limitations would per-mit, in Part I of this discussion on F.M.Installation and Service. It is the purposeof this article (Part II) to discuss F.M.receiver service requirements and proce-dures. We will start off with a brief résuméof First Principles and then launch intoour discussion of practical F.M.-receiverscry icing.

Frequency -modulated signals are obtainedby varying the frequency of the carriersignal at an audio or sound frequency rate.The amount the carrier frequency is variedon both sides of the mean or carrier fre-quency determines the intensity or volumelevel of the signal. Basically, the F.M. re-ceiver is similar to the A.M. superheterodynereceiver except for 3 major differences.This may be seen in the block diagram ofthe 2 receiver types pictured in Fig. 1. Bothtypes may have an R.F. stage, the primaryintention of which is to provide adequateselectivity and voltage gain. A converterstage, consisting of a single tube function-ing as mixer and oscillator, or 2 separatetubes performing these functions is com-mon to both.

THE I.F. AMPLIFIERAlthough an I.F. amplifier of one or morestages is also common to each receiver, theI.F. amplifier in an F.M. receiver differsfrom that of an A.M. receiver by reason ofits wide -band characteristics. In an A.M.receiver, the I.F. amplifier is designed toreject a signal more than 10 kc. to 15 kc.from that to which the amplifier is tuned.On the other nand, the I.F. amplifier in anF.M. receiver is designed to pass a signal,without appreciable attenuation, as muchas 100 kc, either side of the frequency towhich the I.F. transformers are aligned.

Various means are utilized to secure thisband -width. In some instances, the primaryand secondary windings are over-coupledto broaden -out the response curve. The

General Electric HM 136 receiver and themore recent Pilot FM -12 receivers areexamples of this practice.The majority of F.M. receivers, however,employ shunt resistors, to load up eitheror both the primary and

secondary windingsto obtain the required 150 kc. to 200 kc.band -width. The use of such high -gain tubesas the 1852, 1853, 7G7, and 7V7, more thanoffsets the loss in gain as a result of resis-tive loading. In the early Pilot FM -1Rmodel, as well as several P.M. adapters,both: secondary of the I.F.transformers are shunted by resistors, asshown in Fig. 2. Only the secondary wind-ing of the I.F. transformers in almost allStromberg-Carlson F.M. receivers, is shunt-ed resistively. The value of these shuntresistors varies with each receiver model,and depends upon transformer design anddegree of loading required in each case tosecure the band -spread. Resistor valuesfrom 10,000 ohms to 50,000 ohms are mostcommonly used for this purpose.

Early -model F.M. receivers employed anI.F. amplifier tuned to 2.1 megacycles, witha few using 3 mc. Modern F.M. receivershave I.F. amplifiers aligned to 4.3 mc., thestandard set down by the Radio Manufac-turers Association (R.M.A.). One modelmanufactured by Zenith has an 8.6 mc.I.F. amplifier.

THE LIMITERIn the block diagram for an F.M. receiver,a limiter stage may be seen. The limiterstage, essentially an I.F. stage, consists of1 or 2 amplifier tubes so arranged as todeliver constant output despite wide varia-tions in signal input. The tubes employedas limiters are usually pentodes havingsharp cut-off characteristics, and operatedat low plate and screen -voltages, so thatplate current cut-off occurs with relativelysmall grid bias or signal input.

Normal signal input will swing the gridvoltage considerably above and below thelinear portion of the tube's characteristiccurve. Positive peaks beyond the range ofthe limiter tube will be clipped by grid -biaslimiting, whereas negative signal peaks willhe clipped due to plate current cut-off. In38

ALL ABOUT FREQUENCY MODULATIONthis way, variations in signal voltage de-livered to the limiter which are greaterthan the operating limits of the tube arecl.pped and have no effect upon plate cur-rent.

Since static and noise disturbances, pri-marily, produce amplitude changes in thesignal, as do tube noises, the clipping ofthe amplitude changes removes the disturb-ing effects but leaves the frequency -modu-lated signal unaltered. This action is illus-trated at Fig. 3. For complete noise-ejimi-nation, it is essential that the signal volt-age appearing at the limiter grid be suffi-ciently great to swing the grid bias toplate current cut-off and saturation points.

Limiter tubes are generally operated atzero bias or with a small bias voltage. Thelimiter circuit utilized by Stromberg-Carl-son, shown at Fig. 4, is representative ofmany F.M. receivers. The circuit of theGeneral Electric H,M. 136 receiver, at Fig.5, typifies another method wherein theload resistance is connected in the second-ary -return. In this case, the tube is sup-plied with a small initial negative bias. Inthe Pilot FM -12 receiver, a low -value re-sistor is connected in series with the limiterload resistor shown at Fig. 6, so that anindicating meter may be conveniently con-nected to the receiver for alignment pur-poses. An example of 2 tubes arranged incascade or series to operate as more effectivelimiters is seen in the GeneralJFM 90 receiver, whose limiter circuit isshown at Fig. 7.

THE DISCRIMINATORThe 3rd major point of difference be-

tween an A.M. and F.M. receiver lies withthe type of 2nd -detector or demodulator. Inthe F.M. receiver, a discriminator detector,as shown in Fig. 8, is used. The discrimina-tor consists of a push-pull diode detector inwhich opposing voltages developed acrossload resistors are equal and opposite solong as the carrier frequency rests at theintermediate frequency. The resultant volt-age across the 2 load resistors, from pointA to ground is zero, and no audio voltageis developed.

When the signal impressed upon the dis-criminator transformer is frequency modu-lated, due to phase changes as a result ofboth magnetic and capacity coupling, thevoltage drops across the load resistors willbe unequal as the frequency varies aboveand below the intermediate frequency withmodulation. The resultant voltage measuredacross both diode load resistors will thenbe equal to the difference between the volt-ages developed across each; and will varyin polarity from point A to ground as themodulation swings the frequency higher andlower than the resting or resonant fre-quency. The degree of modulation, or fre-quency swing, determines the magnitut;, of

39

AMPLITUDE MODULATION

R A0010F = LF 2tiP 01,7

RECEIVERJJ

IOSCILLATORS

FREQUENCYFRECILIENCY MOOLILATTON

R F IF A00/0

RECE/VER

IOSCILLATOR1

- SIMPLIFIED &OCR 01A6RANti SNORINGNMI SIOUEACE &TAM ARO FM RECEIVER MARES. -

I F FM IF. 50000 IfTRANSi op/M3 I

.-,. FM

,-..5(4000ONA5 A.M.IF

TRANS

A.M.

GOALSONAL5., -

-'1,P.

,

I2

2a000OHM!

- USE OT RESISTIVE uNioNve 70 IBROADENRE5P0v5E CHARACTERISTICS. PILOT FM I2 -.

Ip I

Ey - 1.,,e CHARACTERISTIC

AA IR AT ..M1.ff,47

Ey mg rImeNI -IP-

MM. ...,.... INPUT GINO-- VOL TADSnALE

-RON

-ACT/ON OTLAA,PER III amtmwe mOOLAAroav MANSOproLL0 LLNEAR PoorLom or cHARAcremsTIC -

5O MIFF LIMITER65.17

- --LIMITERCIRCUIT.

1/5E0 IN THESTROAraew-CARL so,515 AAPRECEIVER,

22.000OHMS

.0I- ..---A.-Me

TYPICAL OF THATTumme USED INmoLcArow

Amp MANYRees, VITITS --

%00CCONLASWMLIMITER155J7

-SB:,---L--.

( ,....-.I-,I

LIMITER CIRCUIT,.....7/ U5E0 IN THE

GENERAL ELECTRICNM /S6 RECEIVER,

T-KmA-/E5 ANOTHERa33- METHOD OF OPERATINGNEG....4' 22 L iMI TER -.

AMAX

5O MME LIMITER%i

05./7

1142,0000~31.600 OHMS

11=61

- LIMITER °actorUSEDm TTLERILOT PA"2RECEIVER,

METER I IMIVO IOWLJUSTT7co,,,4,c7,0,,FROM INDCATANO

...LOCO INSTRUMENT-.ao2

ALL ABOUT FREQUENCY MODULATIONthe voltage; and, the numbsr of times peisecond, or late at which the I.F. signalswings shove and below the resonant fre-quency, produces the audio

NEED FOR ALIGNMENTAlthough mot t common troubles encoun-

tered with A.M. receivers, such as faultycondenser. and resistors, are not foreign toP.M. receivers, and are located in likemoiler. probably the major difficulty withF.M. receivers is that of alignment. In anA.M. receiver, the result of I.F. stages onlyslightly misaligned is a loss in sensitivity., id slectivity.

TI ;. result of misalignment in an F.M..ece ver is much more apparent and serious,

nless the I.F. stages and discriminatorare accurately aligned. not only will lossin sensitivity he experienced, but also thesymptoms of distortion and noisy receptionwill be encountered. The need fur precisealignment of the I.F. and disci iminatorstages cannot be overemphasized. Despite.stutements to the contrary, new receiversoften require realignment, due to changesoccurring in shipment and delivery, whereadjustment trimmers or circuit leads shiftsufficiently to influence alignment. This isso, notwithstanding precautions Lateen bydesigners to avoid, by careful packing,bonding, and engineering, the possibility ofchanges. For this reason, the alignment ofF.M. receivers should be thoroughly under-stood. It must he remembered, however, thatmost new receivers reach their final destina-tion in perfect condition, and the impressionthat all reccirers must be realigned is notintended. Before alignment is attempted, thesymptoms displayed during receiver opera-tion must warrant the need for the pro-cedure. Noisy operation and distortion is

rot always the result of misalignment. Moreoften, it is due to insufficient signal pick-up,the cause of which may be a faulty antennasystem- faulty in the sense that the dipolemay rut resonate at the F.I. band or beproperly oriented-transmission line mis-match to the dipole or receiver, or the useof a t:Tnsini..sion line with high R.F.dos c

The procedure followed in P.M. receiveralignment is similar in many respects tothat employed with A.M. receivers, but withseveral important differences. In A.M. re-ceiver alignment, a low input signal is usedto adjust tuned circuits to prevent exces-sive action of the automatic volume control(A.V.C.) and thereby circumvent the useof an output indicator. A strong signal isnecessary to align F.M. receivers, a signalstrong enough to saturate the limiter tube.since this condition is normal during regularoperation. The output indicator is no longera copper -oxide rectifier type of A.C. motorconnected to output plate or voice coilcircuits.

40

There are 2 recognized methods of align-ing the I.F. and discriminator stages of anF.M. receiver. One is through the use of awide -band, frequency -modulated signal gen-erator, and oscilloscope, which is calledvisual or variable -frequency alignment. aprocedure employed by many manufacturerswho claim that this is the only proper wayto obtain perfect alignment. A secondmethod, referred -to as fixed -frequency align-ment, makes use of a signal generator anda sensitive milliammeter, such as is used in20.000 and 10,000 ohms/volt voltmeter in-struments.

VISUAL ALIGNMENTThe vertical plates of the oscilloscope are

connected to the "high" side of the limiterload resistor, as shown in Fig. 9. and theground terminal is connected to ground orchassis of the receiver. The signal generatori, connected to the control -grid of thelst-detector or mixer, and ground, the samegunund connection used for the oscilloscope.

In tly. General Electric F.M. adapter orrecsl.er using 2 1st -detectors, the signal isfs.. into the 2nd mixer).

The wide -band frequency sweep cscillatormay he incorporated within the signal gen-erator or contained in the oscilloscope. Inany case, the sweep oscillator is hetero-dyned with the signal generator unmodu-loted output to produce the specified inter-mediate frequency signal for the receiverunder alignment. Use the widest sweepfrequency that is possible with the equip-ment employed. since correct adjustment issimplified. In some signal generators andoscilloscopes. a sweep frequency as high as750 kc. is available. In others. a :weep fre-quency of only 200 kc. to 300 kc. is possibleWhen the frequency modulator is containedwithin the oscilloscope, synchronizing. is noproblem. Otherwise, a synchronizing voltagefrom the "wobbulator" must be injected intothe oscilloscope to obtain a steady pattern.Starting with the limiter input transformer,and proceeding back to the 1st I.F. trans-former, adjust trimmers of each I.F. trans-former to obtain a symmetrical responsecurve closely similar to that illustrated atFig. 10.

In some receivers, a stage -by -stage orprogressive alignment, rather than an over-all alignment will produce better results.The frequency -modulated signal generatoroutput is connected to the control -grid ofthe I.F. amplifier preceding the limiter andthe limiter transformer is adjusted to obtaina curve with steep sides and wide peak.Each transformer is then adjusted after thesignal generator connection is made to thecontrol -grid of the tube preceding thattransformer. These I.F. transformers shouldbe adjusted to give maximum width consist-ent with maximum vertical deflection. No

ALL ABOUT FREQUENCY MODULATIONover-all adjustments are made after thisstage -by -stage alignment is completed.

A dummy antenna, consisting of a

0.05-inf. or 0.1-mf. condenser, must be usedin series with the "high" side of the signalgenerator and tube control -grid. When thegenerator is coupled to the grid of themixer, insufficient output may result due tothe low impedance of the R.F coil in themixer grid circuit. In this case, it may benecessary to temporarily disconnect theR.F. grid lead from the mixer control -grid,and couple the generator through the dum-my antenna directly to the mixer. A re-sistor. from 10.000 to 25,000 ohms in value,must be connected from tube grid toground to complete the grid circuit.

To align the discriminator, the oscillo-scope is connected across the diode loadresistors. The vertical plates are connectedas shown in Fig. 11, and the ground termi-nal to the ground or chassis of the receiver.Without altering or disturbing the fre-quency setting of the frequency -modulatedAgnal generator, which is coupled to thecontrol -grid of the mixer, adjust theprimary and secondary trimmers of thediscriminator transformer, referred to asCp and Cs, to obtain the "S" or "X" traceon the screen of the oscilloscope as shownin Fig. 12.

The type of pattern depends uponwhether single- or double -trace alignmentis employed. It can be seen that the centerpart of the single trace is a straight lineand is centered and symmetrical with re-spect to all axes. The double -trace dis-criminator characteristic must also be centtered and symmetrical, as shown, with thecrossover at the horizontal and vertical axes.When trimmer Cs (Fig. 11) is adjusted tomore than correct capacity, the traces il-lustrated at Fig. 13 will be obtained. Thetraces shown at Fig. 14 picture the condi-tion resulting when Cs is adjusted to lessthan correct capacity. The adjustment oftrimiuer Cp (Fig. 111 determines theline:, rily of the center portion of the trace.which must be straight. Incorrect align-ment of Cp may provide a trace shown atFig. 15. Since the adjustment of ('p andCs inter -lock to some extent, it is usuallynecessary to readjust Cs again after Cp isnliened. When the "S" or "X" trace seenat Fig. 12 is obtained, the discriminatoralignment is complete.

FIXED -FREQUENCY ALIGNMENTHighly satisfactory and accurate align-

ment of the I.F. and discriminator stagesof an F.M. receiver may be accomplishedwithout a frequency modulator and oscil-loscope. For this purpose, a calibratedsignal generator and a sensitive D.C. in-dicating instrument are essential. The in-dicating instrument may be a D.C. vacuum -

41

tube voltmeter or a 10,000 or 20,000ohms/volt volt -ohmmeter combination withseveral low -current ranges. Stage -by -stagealignment is employed, starting with thelimiter stage input transformer.

The signal generator is adjusted to pro-vide an unniodulated output at the correctintermediate frequency for the receiverunder alignment, and is coupled to the con-trol -grid of the I.F. amplifier tube pre-ceding the limiter input transformer. Adummy antenna consisting of a 0.1-mf. con-denser, connected in series with the highside of the generator to the tube grid, is

rtant. The ground lead or shield of thegenerator output cable must be properlygrounded to the receiver.

Since the current flowing in the gridcircuit of the limiter is proportional to thesignal, connection of the output indicator ismade to the limiter stage. The mode ofconnection depends upon the type of indica-tor and limiter. When an electronic or D.C.vacuum -tube voltmeter is employed, it isonly necessary to connect it from the con-trol -grid of the limiter tube and ground,or across the load resistor as shown inFig. 17. Otherwise, a milliammeter is con-nected in series with the load resistor,shown in Fig. 18.

In some receivers, a low -value resistor ofabout 1,000 ohms, part of the limiter load,is provided so that the milliammeter may be

across the resistor. Inasmuch asthe meter resistance is much less than thatof the 1,000 -ohm resistor, the greater por-tion of the grid current will flow throughthe meter. This circuit is shown at Fig. 19.It is advisable to use twisted -pair leads toconnect the meter to the circuit.

Each 1.F. stage is adjusted to providemaximum deflection or reading on theD.C. electronic voltmeter or milliammeter,coupling the signal generator successivelyto the control -grid of each 1.F. amplifierto the mixer grid. The adjustment will befound to be broad due to the wide -bandcharacteristics of the I.F. transformers.When this procedure is completed, and withthe signal generator coupled to the mixergrid, the frequency setting of the genera-tor should be changed 75 kc. or 100 kc.above and below the correct intermediatefrequency, noting the reading on the in-,'icating instrument in each case. When theI.F. amplifier is correctly and accuratelyaligned, the reading on the meter shoulddecrease an equal amount above and belowthe intermediate frequency. The alignmentshould he repeated until this condition re-sults. Keep the signal generator output lowwhen adjusting I.F. stages, just below thepoint where further increase in signal

output produces no change in the meterreading.

To align the discriminator, the signal

qQ 47

65J7

12"

F, 6

5J7

Oppr

4c

/0.0

00O

HM

S

0.44

-A

yEC

I.

47 *RA

F

6gi

47.0

00-*

-)O

NLy

i

"-D

OU

BLE

LIM

ITE

R C

IRC

UIT

OF

GE

NE

RA

LE

LEC

TIN

G ..

IFM

SO

RE

CE

IVE

R 7

0 °B

MW

WO

WO

E E

FF

EC

TIV

E A

CT

/ON

-D

I SC

RiA

NN

AT

OR

FIG

8

DIS

CR

IMIN

AT

OR

LOA

DR

ES

/ST

OP

S

- T

YP

ICA

L D

ISC

RIN

I/NA

T0R

CIR

CU

IT.

TH

E O

FA

IOD

UL

AM

R O

F P

IEF

1,4

RE

CE

/veR

-M

OM

LIM

ITE

R

ri-O

calo

Sco

PE

L__i

_.

I: VE

RT

ICA

L P

LAT

ES

--C

ON

NE

CT

ING

OC

ILLO

SC

OP

E T

O E

IME

R T

YP

EO

F L

iNN

TE

R F

OR

wS

uAL

ALI

GN

ME

NT

OF

I F

.57:

44E

5 -

L /A

N T

ER

-770

/CA

L T

RA

CE

OF

T F

RE

SP

ON

SE

CU

RV

E O

F F

AA

RE

CE

/wE

Rv/

SuA

L A

L/O

NA

TE

NT

-

OS

CIL

LOS

CO

PE

r

L

f. b.

-MN

/a-A

L"A

rES

-CO

NN

EC

T/N

8 7N

E O

SC

/LT

OS

CO

Pe

FO

RD

ISC

RIM

WA

TO

R A

L O

N M

EN

gene

rato

r is

cou

pled

to th

e lim

iter

cont

rol -

grid

or

the

mix

er g

rid in

acc

orda

nce

with

indi

vidu

al m

anuf

actu

rers

' ins

truc

tions

.In

this

inst

ance

, a s

tron

g in

put s

igna

lis

re-

quire

d. T

he o

utpu

t ind

icat

or is

con

nect

ed to

the

disc

rimin

ator

load

res

isto

rs a

s sh

own

at F

ig. 2

0A. W

hen

a V

.-T

.Vm

. is

used

, the

prob

e or

hig

h si

deis

con

nect

ed to

poi

ntA

or

B, a

nd th

e gr

ound

term

inal

to th

ech

assi

s of

the

rece

iver

. The

prim

ary

trim

-

SIN

GLE

TR

AC

E

DO

UB

LE7R

AC

E

-DIS

CR

WIIN

AT

OR

ahu

mc-

ree,

srIc

RE

SP

ON

SE

EO

RY

ES

06T

A/N

ED

WV

ISU

AL

ALI

ON

AA

EN

T-

Err

algl

i

-TR

AC

ES

SN

OW

ING

EF

FE

CT

°FA

D-I

L/S

T/N

G C

s T

O M

OR

ET

HA

N C

OR

RE

CT

CA

PA

CIT

Y.

CR

OS

3OvE

R n

orno

mA

S x

-

mer

Cp

is a

djus

ted

for

max

imum

rea

ding

on

the

met

erw

ithth

epr

obe

atB

.T

hese

cond

ary

trim

mer

Cs

is a

ligne

d fo

r ze

rore

adin

g w

ith th

e pr

obe

at A

. Whe

n a

D.C

.m

illia

mm

eter

is e

mpl

oyed

as

the

outp

ut in

-di

cato

r(a

zero

-ce

nter

type

isad

van-

tage

ous

but n

ot e

ntire

ly e

ssen

tial),

it is

conn

ecte

d in

serie

s w

ith th

e lo

wer

Mod

elo

ad r

esis

tor

and

grou

nd a

s sh

own

in F

ig.

20B

. Trim

mer

Cp

is a

djus

ted

for

max

imum

-EF

FE

CT

OF

AD

JUS

TIN

GC

s 70

LE

SS

TN

AN

CO

RR

EC

TC

AP

AC

ITY

. CR

OS

SO

VE

R

Mai

nN

OT

AT

AX

IS -

ES

ITel

fall

--T

RA

CE

S S

NO

WIN

G E

FF

EC

TO

F IN

CO

RR

EC

T A

DJU

ST

-M

EN

T O

F C

p-N

ON

-LI

NE

AR

/TY

-

read

ing

and

trim

mer

Cs

alig

ned

for

zero

read

ing.

Car

e sh

ould

be

exer

cise

d w

hen

usin

g a

mic

roam

met

er w

ithou

t zer

o ce

nter

sinc

e th

e m

eter

will

rea

d be

low

sca

le w

hen

the

seco

ndar

y tr

imm

er is

adj

uste

d in

the

othe

r di

rect

ion.

A h

igh

-res

ista

nce

D.C

. vol

tmet

er o

f 20,

000

or 1

0,00

0 oh

ms/

volt

sens

itivi

ty m

ay b

e us

edas

the

indi

catir

g in

stru

men

t and

isco

n-ne

cted

from

poi

nt A

to g

roun

d, b

ut th

e


FIG 17

0

LMI/TER

...CI:h4FC77N6 A V71,41 TO E/77-lER rYPE L/4.1/

e me,7E47

TER C/RCCUT TO 4Z/GN / ',STAGES'.

F IC; 18

La,

/4-fr 7'E:0

alit,004Aifteerew?ORAfax/44we7e)?

A-COV+ECr/NO A APCRO44i4107271? OR M/LL hibseAlE7Z7? 7V L/41, TER Cheafir 70.4L/6N ZRSTACE:5",

precaution mentioned in connection withmeters without zero center must be ob-served. It is possible to change the zeroadjustment of the meter to read up -scaleto avoid slamming of the needle off -scalewhen the secondary trimmer is adjusted.

The following procedure is suggested onlyas a means of accomplishing 1.F. and dis-criminator alignment when other essentialequipment is not available.

Since an audio signal is developed acrossthe limiter load resistor in the circuit shownat Fig. 5, a modulated signal and conven-tional copper -oxide rectifier type of outputmeter may be employed to align the I.F.stages is an F.M. receiver. The meter isconnected in the usual manner to the outputtube plate circuit, and the input of theaudio amplifier is connected to the limiterload resistor. The I.F. transformers arethen peaked for maximum reading on themeter. The discriminator is aligned by con-necting the audio -amplifier input to pointB. the junction of the diode load resistors,shown at Fig. 8, and the discriminatortransformer trimmers adjusted for maxi-mum indication on the output meter.

R.F. AND OSCILLATOR ALIGNMENTThe alignment of the radio frequency and

oscillator stages in an F.M. receiver presentsno problem and is carried out in conven-tional manner, similar in all respects toA.M. receiver alignment. Calibration of thehigh -frequency end of the F.M. band is pos-sible with the oscillator shunt trimmer, butbecause of the narrow spread or limits of

FIG 20",414,7N00 CO-CONAW-7"/Ne /1/0/C47,1017, 0/-9C0/41014,7 AIL/64,44e/vrn-,

is reve:24.FAVOCIENC/A5

Alf/ 770N Of H/6,4-SWEVUE7VCI,rrENZA4 77416 rz.ev,v4KSCAVA.V.V.477246-

4(.1010 C/A7CL//7" Rei7t/C 770N /A, A> 04.7CvALC.424 Pv/ZZ hvcRe.e.ser rweor-Agewesseror.socie

43

ALL ABOUT FREQUENCY MODULATIONthe band, no oscillator padding condenseris provided. When padding of the oscillatorcircuit to calibrate the low -frequency endof the band is required, an end turn of theoscillator coil may be shifted slightly toeffect the alignment.

In connection with oscillator alignment,it must be remembered that the oscillatorin an F.M. receiver of conventional designis operated below the signal frequency, notabove, as in A.M. receivers. This is done toavoid possible image frequency interferencefrom the television channels that lie abovethe F.M. band, and to secure greater oscil-lator stability. The dummy antenna forR.F. and oscillator alignment consists ofa resistor of 100 ohms value connectedacross the antenna input to the receiver.

One of the advantages claimed for F.M.receivers is the extended high -frequencyrange. In some cases, brilliance in repro-duction of the high frequencies may belacking or inadequate, despite correct align-ment and operation. It is only necessaryto change the constants of the filter circuit,shown in Fig. 21, connected between thediscriminator and audio amplifier. Thisfilter, consisting of resistance and capacity,is used to attenuate the high -frequencyboost introduced in transmission, thereby

tending to flatten out the overall audio fre-quency response. Removal of the filter en-tirely is not recommended, but a change inthe time constant, usually 100 microseconds,may be warranted. By reducing the valueof the resistance or capacity, an increasein high -frequency reproduction is secured.

On the whole, the radio technician whorecognizes and understands the major dif-ferences between F.M. and A.M. receiversshould have little difficulty in the installa-tion and service of F.M. receivers.

Another common trouble with F.M. re-ceivers is distortion. This condition iscaused by any one or more of a number offailures, but principally, incorrect I.F. dis-criminator alignment. When the discrimina-tor characteristic is not linear, frequencychanges in the signal will not be convertedinto a sine audio voltage. Misalignment ofthe I.F. amplifier may result in unequalamplification of the wide frequency bandessential for correct operation, find conse-quently a non-linear discriminator response.Both the I.F. amplifier and discriminatormust he aligned to the same frequency andthe response of the I.F. amplifier must beequal over the required spread on bothsides of the intermediate frequency.

PART III

TEST EQUIPMENT FOR

F. M. SERVICINGMost Servicemen already realize that their

present test equipment will not adequatelytake care of the requirements of F.M., butthere is much confusion and difference ofopinion regarding the equipment which willbe required. Most of this confusion is dueto the fact that few people have had anyactual experience with F.M. receivers inorder to really know the requirements.

The majority of the present worries havebeen based upon the problems arising fromthe wide band of transmissions. Experiencewill soon show that the band -width is theleast of the worries, although it does enterinto the picture somewhat. The biggestproblem is to get test equipment which will

44

function at all at 40 to 50 megacycles!In order to simplify an understanding of

the major factors in servicing F.M. radioreceivers and adapters the following de-scription is divided into 5 sections identifiedas follows: (A) I.F. alignment; (B) dis-criminator adjustments; (C) R.F., detectorand oscillator adjustments; (D) locatingreceiver troubles; and, (E) testing tubes.

TEST UNIT NO. I:I.F. SERVICE OSCILLATORA.-In aligning F.M. receivers there Is

no way of controlling the band -width. Thathas been left entirely to the manufacturer.


Illustrated here are rep-resentative, Weston pro-cisiol. instruments suit-able for efficient testingand alignment of fre-quency modulation re-ceivers. They are identi-fied as follows: A, model776 oscillator; B, model787 high -frequency oscil-lator; C, model 669 V. -T.

voltmeter.

The Serviceman cannot possibly change thecoupling between the coils on intermediatefrequency transformers and expect any de-gree of success. In actual practice all I.F.adjustments are made for peak performanceat the intermediate frequency which usuallylies between 1 and 5 megacycles, dependingupon the individual manufacturer. Thetransformers cannot be stagger -tuned in anattempt to broaden the band-pass character-istics. Any attempt in that direction willresult in all kinds of distortion. The mostimportant point in connection with the I.F.alignment is that all adjustments be madeat exactly the same frequency. This requires,therefore, a service oscillator which is freefrom drift at 1 to 5 megacycles. There isone such oscillator available, designed witha negative feed -back circuit which is re-ferred -to as automatic amplitude control.This results in remarkable stability at therequired frequencies.

TEST UNIT NO. 2:U.-H.F. SERVICE OSCILLATORB.-The only other circuit in F.M. receiv-

ers which is particularly annoying to theServicemen at present is the discriminatorcircuit. However, if we analyze this portionof the receiver very carefully it will soon beseen that this is exactly the same as similar

45

circuits which were used for automatic fre-quency control in A.M. receivers. See Fig. 1.The adjustment of these circuits is exactlythe same as shown.

C.-The last adjustment required in thenew type of equipment is that of the R.F.,detector and oscillator circuit. These adjust-ments must be carried out using a serviceoscillate[ which is capable of accurate tun-ing at between 40 and 50 megacycles. Suchequipment is extremely rare at the present.Any attempt to use harmonics of lower fre-quencies will be very annoying and confus-ing. Furthermore, it will be impossible tocheck the band -width characteristics of thereceiver as is often desired. An oscillatorusing inductive tuning capable of operationfrom 30 to 150 megacycles is particularlysuitable as its tuning characteristics at F.M.frequencies is better than anything elsenow available. On this unit one division ofthe dial at 40 megacycles represents only 40kilocycles, making it convenient to tunethrough the desired band of 100 to 150 kilo-cycles.

At the present stage of affairs there hasbeen an expression on the part of somepeople of the desire for a service oscillator,capable of being 100-kc. F.M. wobbled, to b3used in conjunction with an oscilloscope. Ifsuch equipment were available it would serve


Chassis of the Weston model 787 oscillator tuner,assembled.

FROM

.5+'

FIG .1

6H6DUODIODE

R.F.C.

ti TYPICAL A.F.C.

TatlifASLTUBE

FROMI.E

6H6 TO AUDIODUODIODE AMPLIFIER

ti F.M. DISCRIMINATOR",

to very little advantage as all adjustmentswould still have to be made for peak re-sponse. The oscilloscope would function asnothing more than a visual type of volt-meter. The various I.F. transformers couldnot be adjusted any more accurately orquickly with such a setup.

TEST UNIT NO. 3:V. -T. VOLTMETERD.-In locating the troubles in F.M.

receivers, it will once again be necessaryto resort to fundamental test equipmentsuch as 20,000 ohms/volt analyzers, vacuum -tube voltmeters, etc. Due to the frequency ofoperation the many types of so-called signaltracers will be of doubtful assistance as theyhave all been designed for operation at thepresent A.M. broadcast frequencies. Themost logical instrument for localizing re-ceiver troubles in F.M. equipment will be avacuum -tube voltmeter capable of measuringthe A.C. signal voltages at from 1 to 50megacycles. There are a few such voltmeterswhich have been available for some time.

E.-The testing of tubes in F.M. equip-ment will present quite a problem, as manytubes capable of satisfactory operation atthe lower frequencies will fail completely atthe P.M. frequencies. Tubes such as the1852 and 1853 must be used in order to ob-tain sufficient gain at high frequencies.Those tubes have very small elements closelyspaced, and are much more susceptible toleakage and shorts than the other, morecommon types. Experience has also proventhat no commercial tube tester will satis-factorily indicate the quality of a tube whichis to be used as an oscillator at 50 mega-cycles.

Because of the fact that very little workhas previously been done at these higherfrequencies, it is perfectly safe to prophesymany new and radically different types oftubes during the coming few years. The tubechecker picture, therefore, is the only un-certainty in the consideration of suitabletest equipment for F.M.

It is hoped that in the excitement andrush resulting from the widening of activityin F.M. the Serviceman will keep his headin the selection of equipment. Do not be tooanxious to jump at the purchase of equip-ment which is represented to be a cure-all.

Above all do not let any one give you theidea that there is anything tricky or com-plea about F.M. receivers. When it comesright down to reality the circuits and com-ponent parts which are used are no differentthan those already familiar to everyone. Thereal issue is the operating frequency of 40megacycles, and the only problem is to getequipment which is designed carefully andaccurately enough to reliably operate at thatfrequency of 50 megacycles.

46


CHAPTER V

ENGINEERINGThe How and Why of Frequency Modulation

THERE is widespread belief that manypresent-day developments are funda-mentally new within the last few years.On the contrary, the bases of many so-

called new developments date back a greatnumber of years. It is true, however, thatonly recently has it been possible to utilizeto fuller advantage the possibilities of manyof the ideas passed down to us by earlyinvestigators of the Radio Art. New instru-mentalities have made it possible to ex-plore these fundamental ideas to muchgreater extent. Transmission and receptionat ultra -short wavelengths, or ultra -highfrequencies, is one such outstanding ex-ample.

U.-H,F.The use of the ultra -high frequencies for

sound broadcasting offers technical advan-tages, not only to the broadcaster but tothe public, which is much more important.The technical advantages consist of (a)escaping the 10-kc..channel limitation, (b)getting away from static, and (c) elim-inating all except spasmodic long-distanceinterference.

We've known this for years, have ex-perimentally operated low -power U.-H.F.stations since Way Back, and have enjoyedthe experience of receiving Clean Stuff fromour little ultra -high frequency transmitterswhen QRN, with devastating wallops,washed out our temporarily musclebound50 kw. steamrollers. Five years ago theF.C.C. had applications for, or had licensed,over 100 ultra -high frequency transmittingplants and it seemed that a trend was de-veloping toward ultra -high frequency broad-casting, but this trend was not sustained.Interest has been revived in recent monthsthrough the promotion of F.M. on the ultra-high frequencies.

47

WHAT DOES F.M. OFFER?Frequency Modulation is a weapon

against noise, a sword if you please, withadvantages which can be calculated ac-curately and simply, as we shall see. Butunreasonable powers should not be attribut-ed to it. The pen should not be mightierthan the sword.

Your scribe bows low and humbly at-tempts, with these hesitant strokes, to bringto you gentlemen of the A.T.E. Journalwhat the Lower Classes vulgarly call theLowdown. A snack of inside dope.

Let's get to the point. What advantagesdoes F.M. really give over A.M.? Using thefrequency deviation approved for the in-dustry by the F.C.C., F.M. UNDER THEOPTIMUM CONDITIONS gives (a) anadvantage of 20 to 1 in background noisesuppression, (b) an advantage of at least 30to 1 in rejection of shared -channel interfer-ence, depending on the beat frequency, and(c) some advantage to the broadcaster incapital expenditures and operating costs.There you have it.

F.M. IN 19021One frequently meets laymen who have

the mistaken idea that F.M. is a revolution-ary new invention. The justry proud fatherof your profoundly humble scribe boughthim his first lace velvet pants in 1902. Mostof you were still unborn during that antedi-luvian era.

It was in that year that a gentlemannamed Ehret applied for a patent whichwas issued in 1905 covering the basic meth-od of F.M. for voice and code transmissionand reception!

Mr. Ehret proposed to shift the carrierfrequency by means of a voice -actuated con-denser. He proposed an off -tuned circuit inthe receiver for converting the frequency -modulated waves into waves of varying am-plitude.


With certain improvements these are themethods now used. For code signalling heproposed to key the transmitter inductanceor capacity to change the carrier frequency.Before the No. 1 war this method was verywidely used for many years on longwavetransmitters. Remember how discombobu-lated one could become by trying to read thebackwave when fatigued?

"WIDE SWING" F.M.Frequency Modulation research has been

carried on for over 30 years and, exceptfor 1918, 1920 and 1924, patents have beenissued on F.M. methods and devices eachyear for the last 25 years. They were grant-ed mostly to a number of inventors in theemploy of organizations which spend largesums on research, such as G.E., Westing-house, A.T.&T. and RCA, and to a few in-dividuals, particularly Major Edwin H.Armstrong who has promoted use of thefeature of "wide swing" in F.M.

Other features are important in F.M.such as limiting. Gentlemen named Wrightand Smith filed a patent application cover-ing it 15 years ago. Fourteen years ago,and subsequently, patent applications werefiled and granted to Westinghouse, A.T.&T.and RCA on balanced, or "back-to-back"F.M. demodulators. The most commonly useddiscriminator today was patented by S.Seeley of RCA. Frequency multiplication ofun F.M. wave to increase the frequencyshiftis covered in patents issued to West-

"W2XWG's field intensitypattern in micro -volts/meter (irrespective ofthe type of modulation).These measurements rep-resent a power of 1,000W., a transmitting an-tenna height of 1,300 ft.,a receiving antennaheight of 30 ff. and anoperating frequency of

42.6 megacycles."

inghouse, and G.E., for whichwere filed in 1926 and subsequent years.High frequency pre -emphasis and de -empha-sis circuits were patented by S. Seeley andothers of RCA. Its introduction to the in-dustry was due in considerable part to theefforts of N.B.C.

At the close of 1939 more than 250 patentshad been granted on either Frequency orPhase Modulation, of which more than 160covered F.M. About 10 years ago R.C.A.C.was trying F.M. on channels between ourEast and West coasts. About 12 yearsago your scribe co-operated with Westing-house in F.M. tests between New York andPittsburgh. So you can see F.M. isn't new.

HI -Fl A.M.There is a popular impression that by

use of F.M. and "wide swing" the publicmay only now enjoy high fidelity. The factsare that with ultra -high frequencies thefidelity can be made as good as anyonewants it to be with either frequency oramplitude modulation. Any improved fidelityis made possible by getting away from the10-kc. channel allocations of the StandardBroadcasting Band and not by using F.M.

Furthermore, to get "high fidelity" inA.M. or F.M. receivers the listener mustpay exactly the same high price for high -power, low -distortion audio amplifiers, loud-speakers and acoustical systems. However,the time may come when High Fidelity willreceive the widespread recognition it merits.

48

ALL ABOUT FREQUENCY MODULATIONThere is much more interest now in low re-ceiver prices which preclude high fidelity.This is unfortunate but incontestably trueregardless of any wishful or idealisticthinking to the contrary.

There is no lack of satisfactory fidelity inpresent-day transmitters because, if for noother reason, the F.C.C. requires it. Theloss of fidelity rests in the home receivers.Medium-priced receivers satisfy the publicdemand and high fidelity cannot be ob-tained in those models. The price paid forso-called high-fidelity amplifiers and loud-speakers is in itself more than the cost ofmost receivers. Possibly 1 person in 6 has areceiver of good fidelity. Many of these lis-teners.normally operate with the tone con-trol adjusted for the lowest degree of fideli-ty possible with such receivers. It appearsthat the public is not suffering any lack offidelity because of the present broadcastingsystem.

We in N.B.C., and others, have been pro-viding transmission of excellent fidelity forat least 15 years (network lines excepted)and will continue to do so. We believe in itand endorse it. But we have no illusionsabout the public reaction toward it.

NOISE THRESHOLDFrequency Modulation would under favor-

able conditions, but not all conditions, re-duce static about 20 to 1. But what staticare we talking about? Static practicallydoesn't exist on ultra -high frequency. There-fore, isn't its absence mainly due to theshift to the ultra -high frequency band? Itis.

Don't think that your humble servant isbearish on F.M. because that would be in-correct. It is cold professional realism, notbearishness. An F.M. station will providenoise -free service to a much greater dis-tance than an A.M. station of equal powerbecause F.M. can suppress receiver hissnoise, auto ignition noise and other ultra-high frequency disturbances about 20 to 1,if the carrier is stronger than the noise andif the receivers have enough gain to makethe limiters limit at low field intensities.Some F.M. receivers begin to slack off atabout 100 microvolts. To obtain the fullbenefit of F.M. out to the "noise threshold"limit they should hold up down to 10 micro-volts. This noise threshold is strictly anF.M. phenomena; more on this later.

We are all confident that Television hasa most brilliant future. We are not entirelyclear on the position that Ultra -High Fre-quency Sound Broadcasting will have withrespect to it. Those of us who have livedwith television for many years feel thatsound is supplemental to sight but defi-nitely second in importance. When tele-vision hits its stride, sound broadcasting

1) fleetht first item under "SOUND" In the June,'11. Issue of Radie-Cratt, pg. 715.-Edits

49

may assume the status of silent pictures.Who knows? Nobody does. In any event,sound broadcasting will be with us formany more years and we should give fullopportunity to improved methods and de-vices. F.M. is one of them. N.B.C. has oneF.M. station and will build more. F.M. isbeing given its chance to prove itself.

The N.B.C. has for many years viewedrealistically the advantages of the ultra-high frequencies and has been confident thatthe industry would, in time, do likewise.Five years ago Mr. Hanson and your pro-foundly humble scribe wrote a long reportan the subject forecasting the growth ofultra -high frequency Sound Broadcasting by6 -month intervals and hitting very close.Frequency Modulation had such promisingtheoretical advantages that we undertook afull-scale field test to determine the extentto which they could be realized in practice.

$30,000 WORTH OF TESTSAs a result we completed, last year, at a

cost of over r40,000, the most thoroughfield test of F.M. ever undertaken and wehave the information we sought.

It was obtained, not by laboratory work,which had been done before by others. in-cluding R.C.A.C., nor merely by operatingan F.M. station, but by building specialtransmitters, receivers, measuring instru-ments, etc., and then painstakingly makingthousands of measurements at distantpoints over many months and under a vari-ety of conditions.

A special 1.000 -watt transmitter was or-dered from the R.C.A.M. Company. It hadfacilities for both A.M. and any degree ofF.M. deviation or "swing" desired, with re-mote control facilities for instantaneouslyswitching to either system. Since the F.M.deviation varies directly with the audio in-put level, remote controlled pads could beand were used to select the deviation de-sired.

W2XWG was installed in the EmpireState Building. Special authority was ob-tained from the F.C.C. to use amplitudemodulation as well as F.M. on 42.6 me. forthe term of the project. The television videoantenna, having a pass band extending from30 to 60 megacycles, was used for most ofthe W2XWG transmissions although a spe-cial folded dipole was used when the videoantenna was transmitting "pictures."

W2XWG was equipped with means forcontinuous variation of power between 1/10 -watt and 1,000 watts, and a vacuum -tubevoltmeter for accurately measuring thepower.

The modulation conditions selected wereA.M./F.M. 15 (deviation of 15 kc., or totalswing of 30 ke.), and F.M. 75 (deviation of75 kc. or total swing of 150 kc.). Tonemodulation was used for most measure-


h- AMPLITUDE MODULATION -4;)

F/G. 2

O /3 30 45 60 70K/LOCYCLES

0 6 96 /2 /4DB.04/61

FIG. 3

ments. For measuring distortion, or noiselevels with modulation present, the tone out-put of the receivers was cleaned up by pass-ing it through filters and then impressedupon RCA noise and distortion meters.

Four special receivers were built by theR.C.A.M. Company for this project. Eachwas equipped for instantaneous selection ofA.M./ P.M. 15 or F.M. 75. Two complete I.F.systems were built-in, one 150 kc. wide andone 30 kc. wide, each having 5 stages, withboth A.M. and F.M. detectors. All receiverscontained meters, controls, de -emphasis cir-cuits with keys, 8-kc. cutoff filters with keys,separate high -quality amplifiers and speak-ers, cathode-ray oscillographs, etc. Each re-ceiver had sufficient R.F. gain to give fulloutput with limiting at input levels muchlower than required, theoretically doing sowith only 1/10 -microvolt input. These re-ceivers were made as good as receivers canbe built in order that our conclusions onF.M. would not be clouded by apparatusshortcomings. Sacrificing good receiver de-sign to price will not permit the full gainof F.M., as reported herein, to be realized.

FIELD INTENSITYAs a part of the project, a field intensity

survey was made of the W2XWG transmis-sions. The map is included herein for 1,000watts, 1,300 feet antenna height and .7 an-tenna gain. It is Fig. 1.

Measurements and electrical transcrip-tions were made under a variety of con-ditions at the following locations:

MilesCollingswood, N. J. 85Hollis, L. I. 12Floral Park, L. I. 15Port Jefferson, L. I. 50Commack, L. 1. 36Riverhead, L. I. 70Hampton Bays, L. I. 78Bridgehampton, L. I. 89Eastport, L. I. 65

50

N.B.C. Laboratory 1

Bellmore, L. I. 23

All above stations are temporary, with theexception of the last two, which are perma-nent.

Most of the measurements were made atthe Bellmore station. For the temporarystations, 2 automobiles were equipped andused, one a Radio Facilities Group measur-ing car, the .other a borrowed R.C.A.C. truckfull of recording gear. The receiving sta-tions represented a cross-section of ruraland suburban Americana.

Let's next see what theoretical advantageF.M. has in noise suppression and how it isobtained. Later we will see what we meas-ured.

In F.M. the deviation of the carrierfrequency can be made as great as desired.If it is 15 kc. and the audio band -width is15 kc. the deviation ratio is 1, correspondingto the deviation divided by the audio band-width. If the deviation is 30 kc. the devia-tion ratio is 2, etc.

The advantages of F.M. over A.M. in noisesuppression are contributed by 3 factors:

(1) The triangular noise spectrum ofF.M.

(2) Wide swings, or large deviationratios.

(3) The greater effect of de -emphasis inF.M. compared to A.M.

Let us consider them in order.

TRIANGULAR NOISE -SPECTRUMAn F.M. system with a deviation ratio of

1 has an advantage in signal-to-noise ratioof 1.73 or 4.75 db. for hiss or other types offluctuating noise.

Since the figure 1.73 applies to such noisesas tube hiss, which is comparatively steadyin amplitude, we will consider this typeof noise. It differs from impulse noise suchas is produced by automobile ignition sys-tems.


Tube hiss consists of a great many close-ly overlapping impulses or peaks There areso many of them at all audio frequencies,we are concerned with, that the noise hasa steady characteristic. When combined witha steady carrier of fixed frequency, the noisepeaks beat with the carrier. The noisepeaks also beat with each other. When thecarrier is considerably stronger than thenoise peaks, beats between the noise peaksbecome negligible in amplitude and thepredominating noise is due to the combina-tion of carrier and noise peaks.

Since a combination of 2 carriers differ-ing in frequency produces a similar phe-nomenon, we will treat both cases at thesame time. The effect is most easily shownand understood by means of a simple vec-tor diagram.

The strongest carrier vector continuouslyrotates through 360° and is indicated onFig. 2. The weaker carrier, or the "noisevoltage," rotates around the carrier vectorat a frequency which is equal to the differ-ence between the desired carrier and un-desired frequency.

It will be seen that amplitude modulationis produced. If the undesired frequency is50% as strong as the desired frequency,50% amplitude modulation results. As theundesired vector rotates around the desiredvector, phase modulation also is produced

the limits A and B. The faster theundesired vector rotates, or the faster therate of phase change becomes, the greaterbecomes the momentary change in frequencyand, therefore, the greater the frequency

modulation becomes, because FrequencyModulation is a function of the first differ-ential of phase modulation. Therefore, theamplitude of the frequency modulation noiseor beat note varies directly with beat fre-quency. With both frequencies exactly thesame there is no amplitude modulation nor

is there any frequency modulation.Such being the case, the noise frequencies

close to the carrier produce little frequencymodulation noise but as the noise com-ponents further from the carrier combinewith it they produce more frequency modu-lation, Therefore, the higher the noise beatfrequency the higher its amplitude. Thisresults in a frequency modulation noisespectrum in which the noise amplitude risesdirectly with its frequency. In other words,it is a triangular spectrum.

In amplitude modulation there is no sucheffect as this. All noise components combinewith the carrier equally. Therefore in am-plitude modulation there is a rectangularnoise spectrum. The ratio of noise voltagesin F.M. and A.M. is therefore the ratio be-tween the square root of the squared ordi-nates of a triangle and a rectangle. Thisratio is 1.73 or 4.75 db.

DEVIATION RATIOFor an F.M. System the suppression of

fluctuation noise is directly proportional tothe deviation ratio.

On Fig. 3 the A.M. noise spectrum cor-responds to the total hatched area below 15kc. because the I.F. system would cut offthere. The F.M. 75 receiver I.F. systemactually accepts noise out to 75 kc. and ithas the usual F.M. triangular characteris-tic. However, the receiver output and theear responds only to noise frequencies with-in the range of audibility, around 15 kc., andrejects everything else. Therefore, the F.M.75 noise we actually hear corresponds onlyto the small cross -hatched triangle and allthe rest is rejected.

The maximum height of this F.M. tri-angle, corresponding to voltage, is only1/5th of the height of the A.M. rectangle.Such being the case the F.M. 75 advantageis 5 to 1, or 14 db.. Simple?

PART II

HOW AND WHY OF

FREQUENCY MODULATIONTRANSMISSION and reception on ultra -short wavelengths is not a new idea-not even sound programs utilizingthe technique of Frequency Modulation.

The truth of this statement was discussedin Part I of this article

51

in which it was shown that thebasic method of F.M. for voice transmissionand reception was the subject of patents is-sued in 1905. From this general introduc-tion, the writer proceeded to discuss subse-quent technical developments culminating


i 1=11811111111M =IMO MIIIIIMIIMMINIMIIIIIIIMFREQUENCY P400(1111770N !1o 1 AMPLITUDE

MODULAT/ON II.

1 II MOHWla0 7 i II NMI Ii 6i 5

IIt..;::i

NO GE.E.wo

1 11 WA 11

lad 111II, ,,,,,,!.t- 3 MI II PI II< t§f43. II Ii Ell,w 2 ":....%a ,,i..-.4.1.7muMill

a E)4:veimi NIF NE R G ' ".5% 4

F.4.. 3inglad0114 G 8

FREQUENCY10 12

KC

'4

IINu

150 2 4 6 8 to 12 14 15 0 2

FREQUENCY KCIriOril

2. 60 =ME excel.ENT7g 55 IMELLIII

! 50 MIa AMPLITUDE MEM ectii`"

ii 45gu EMERMill:,_ 40

ss INIIIIWIII D3C 0, 611

IP4.2 2_ MIMEO0 7718E 1115.5 AND MISCELL-1111f, 20 AA/E01.45 CE.NETr BORN000

15

NE/OH-NO,GE ism l'orir

litanawf/Trifq4 vv2xv3.NY

- CrYNMi 1U71

ai, ' '1

RECe,vERBELLMORE,

aocar,o. v..1

Ct 0

c6 5 0 200 In0 50 20 10 5 2 1

MICROVOLTS AT RECEIVER INPUT TERMINALS286 114 572 28.6 11.4 5.7 2.86 1.14 .57

FIELD INTENSITY 941(110YOLTS 011174ETER

4 94 hi 55.5 65 78 88 98 112 -122015-794CE 05.740 NORTON AVERAGE

CURVE 1000'-30', 1 KW, 50 me®AVERAGE SZR,NCE RANGES

RNEERA EEihmoRe NC/s.-EASED t/PoN AUX.mERsuRff mffedrs

IM 'KW =ADO FOR Ezz3 ADD FORSRN SO KW

111.1111amoloruorpao,

fooFM 75

AmPutu0E 1111111111.111.111FN ISEN 30

I7.5

ANALITUOIE 1.11.111.10.111ffillFN IS

INCELLEN7"6008 EMS

I F C:t LE:T

G000

;11:35 ..trow 40 08 EMS

20'4 '0 80 10'0 120 130

MILES SERVICE RANGE USING NORTON AVERAGE'CURVE FOR SO MC 1.00.0- 30.NTEN.A

mEIGWTSFIG 6 .T.1.1.4.YE.0 BY matswoLo)

52

4 Fig. 4. High -fre-quency pre -emphasis

it used in the transmitterand de -emphasis is usedin the receiver to im-prove the signal/noiseratio. The importance ofthis characteristic inultra - H.F. broadcaitingand its greater effective-ness In F.M. are illus-trated at left. Compari-son is on the same scale

of relative amplitude.

I/

in the wide -band system of F.M. espousedby Major Armstrong. The pros .and cons ofsuch characteristics as fidelity, noise thres-hold, triangular noise spectrum, deviationratio, field intensity, etc., upon which testswere made at a cost of $30,000 by N.B.C.,were described.

We now continue with the further detailsof these tests, including discussion of pre -emphasis, de -emphasis, F.M. noise threshold(the effects of ignition interference, etc.),and the simultaneous operation of 2 F.M.stations on the same channel.

DE -EMPHASISWhen the high frequencies are attenuat-

ed in a receiver, the high -frequency noise is,of course, attenuated by the same amount.This may make a noisy signal more pleas-ant to the ear, but it degrades the fidelity.However, if the high frequencies are in-creased in amplitude in the transmitter, theoverall fidelity will be restored. Neverthe-less the noise which comes in at the receiverremains attenuated and therefore a reduc-tion of noise results from this practice.

The use of a 100 -microsecond filter to ac-complish this purpose has been adopted asstandard practice in Television and ultra-H.F. sound broadcasting by the Radio Man-ufacturers Association and recently by theF.C.C. It has actually been in use for sev-eral years. (Italics ours.-Editor) A 100 -microsecond titter is a combination of resist-ance and capacity which will charge to 63%of maximum, or discharge to 37% of maxi-mum in 100 microseconds.

It was shown that in F.M., the noise am-plitude decreases as its frequency decreaseswhereas in A.M. it doesn't. Therefore, de -emphasis is more effective in F.M.

Consider Fig. 4. The full rectangle at theleft is the A.M. noise spectrum. The fulltriangle at the right is the F.M. spectrum.The application of de -emphasis reduces

ALL ABOUT FREQUENCY MODULATIONthese areas to those combining the hatchedand black sections. Squaring those ordi-nates gives the black areas, correspondingto power, or energy. Extracting the squareroot of the ratios of these black areas givesthe r.m.s. voltage advantage of F.M. overA.M. It is 4, corresponding to 12 db. Bearin mind that this 12 db. includes the gainscontributed by both the triangular noisespectrum and de -emphasis. The spectrumadvantage was 4.75 db. Hence the de -empha-sis 'advantage is 12 db. minus 4.76 db. or7.25 db.

All commercial F.M. receivers include de -emphasis and all F.M. transmitters includepre -emphasis. It's an F.C.C. requirement.(Italics ours.-Editor)

Now let's sum up. We.saw (Part I) thatthe F.M. noise spectrum advantage was 4.75db., the de -emphasis advantage was 7.25 db.and the deviation ratio of "F.M. 75" was 14db. Combining these gives us 26 db.

Let's now see what advantage we actual-ly measured as part of the field test project.Your attention is directed to Fig. 5 whichhas on it a great deal of information.

It actually condenses to one illustrationmuch of the data we sought and obtained.Many pages could be devoted to it. Thecurves may be extended to the upper -leftin parallel lines as far as desired. The actu-al field intensity of the noise can be deter-mined from the A.M. curve. For instance,for 10 microvolts at the receiver terminalsthe A.M. signal-to-noise ratio is about 25db. or 18 to 1. Hence the noise is 1/18 of10 microvolts, or 0.6 -microvolt r.m.s.

The ordinates are identified in receiverinput microvolts, microvolts -per -meter andmiles distance. Use the one you are mostinterested in. If you want condensed dis-tance tables refer to the bar chart, Fig. 8.

Compare the measured gains with thecalculations we went through. They look tobe the same. They are. That means we foundthat the theoretical gain of F.M. can beand was obtained in practice.

Note the dotted sections of the F.M.curves. They are dotted to indicate thatoperation is not only below the "noisethreshold" but is far enough below it thata noticeable increase of noise results as soonas modulation occurs. The dotted sectionsrepresent noise in the unmodulated condi-tion. During modulation they break evensharper than indicated. Since there is nosuch thing as a noise threshold in A.M.there is no such break. Wherever usableA.M. entertainment service is provided"F.M. 15" is 12 db. quieter and "F.M. 75"is 26 db. quieter.

F.M. NOISE THRESHOLDAn interesting series of events takes

place in a Frequency -Modulated system

53

55

SO

4540

35

25

FREQUENCY MODULR770N FLUCTUR-TION NOISE THRESHOLD

7-1,4N3A4/77ele INOOULATE0 17.000TONE. TANS ANGRY 0,5104770N 1,100.,Crs

RECEIVER Wing /4,000,LON-4,45.5 r/L TER, LERYING ONLY ,V0. -SE

nomimili35

30 miliiii,..."!11.II_ 15

2520

1.111M.11111111111.1.1111=1.11=111111111.11111111.11.1111160 1.11111.1111iiMI'

MINN

FM

L114aneer-. 1111Wilii1W

FM75

o " 20 40 60 80 100PER CENT OF MAXIMUM FREQUENCY

RESPONDINGFIG. 7

DEVIATIONPER CENTCOR m0OuLATION

TO

when the noise peaks equal or exceed thepeaks of the carrier. The result is a rapidincrease of the noise level or decrease ofthe signal-to-noise ratio with modulation.

In Frequency Modulation wherein themaximum swing is 150 kc. the point wherethis begins to occur is reached when theunmodulated signal/noise ratio is about 60db. When the unmodulated signal/noise ra-tio is less than about 60 db., or 1,000 to 1,the noise level rises with modulation, andas the noise peaks exceed the carrier peaksby a considerable amount, this noise levelmay go up 20 db., or 10 times. When operat-ing above the threshold limit the noisechanges very little as the station is modu-lated. Below the threshold limit the effectis not unlike harmonic distortion in anoverloaded amplitude transmitter.

In Frequency Modulation of a lesserawing, such as 30 kc., a similar effect oc-curs. In this case, however, the thresholdlimit occurs at about 35 db. signal/noiseratio. Figure 7 shows the results of someof the measurements we made. In order thatthe noise would not be confused with thesmall amount of inherent distortion in apractical F.M. system, the measurementswere made in such a manner that the effectsof distortion were eliminated. This was doneby modulating the transmitter with a 17,000 -cycle tone and eliminating at the outputof the receiver with a 14,000 -cycle low-passfilter, not only the fundamental modulating

F/ G. 8

ALL ABOUT FREQUENCY MODULATION"F.M. 40" system having a total bandwidth of 100 kc. occurs at about 43 db. Sincethis provides a very good signal/noise ratioand the required band width is only 100 kc.,F.M. 40 is believed by many to have moreoverall merit than F.M. 75 when the com-parative gains and limited space in theallocation spectrum are considered.

So far as is known, the data on theF.M. threshold effect presented here, anddata published by Murray Crosby ofR.C.A.C. constitute the only measured dataever published.

Figure 8 shows ignition noise measure-ments with peak noise input microvoltsplotted against peak signal to noise ratio,based upon the signal resulting from maxi-mum 400 -cycle modulation. The "F.M. 15"threshold is shown. The F.M. 75 thresholdis not shown because at the time the meas-urements were made A.C. hum within thesystem made the accuracy of S./N. measure-ments in the 60-db. region uncertain.

It should not be assumed that peak S./N.ratios of 20 or 30 db. are unusable when thenoise arises from ignition systems becauseit isn't true. The relative infrequency of ig-nition peaks produces an audible resultwhich is very deceiving. Ratios as low as10 db., while distracting, do not entirelyruin service as is the case with fluctuationnoise.

It will be noted that the curves of igni-tion noise threshold flatten off at the bot-tom. This is to be expected from the char-acter of ignition noise. The impulses arevery short in duration, very high in ampli-tude and (relatively) widely separated. Theyliterally blank -out on4y small portions ofthe signal waves, without impairing the re-mainder: The short, blanked -out intervalsof the signal change little over a wide rangein noise peak amplitude. Once an ignitionpeak has risen to the value required to con-trol the receiver and blank -out the signala further rise in the noise level will notoccur until the peak increases in breadth,or duration, or until there is a sufficient risein certain low -amplitude components of ig-nition noise having fluctuation noise char-acteristics.

The peculiar shapes of such curves belowthe threshold values are due to the waveshapes and crest factors of ignition noise,but they are also influenced by the methodof measurements.

45° 5c ling1.0

40

0 35I

E. 30

AMPLITUDEtwool.u.orlow

RECENERiSrc.fMNO

wiDTH

/ N/T/ON NOISEFRE.04/ENC11 AND

AMPLITUDEMODULATIONRECEIVERS

8A-C.Amo geeo warI IIIIS KC. F.M.

RECE VEQ

25kr

11/1 I,COVERFM

1111118.4*/OwiDTAI

25 1611111:110111111111111co 20

1111inniggr775r? :111111111.4.1116x ,,,,,

..x,....0C,

1111111.161112ii 60%'g007 .00

co 5

0 1111111.1111111111k.0 100 500 1000

PEAR NOISE INPUT MICROVOLTS

am.0Et' za

ONaoIn=

ozo

3.t4'

FIG. .9

40

30

20

10

0

OPERATION OF 2 F.M.57,477ONS ON SAME

CHANNEL

CAL CLAATEDAFAR.7m. 30

10 20 30 40 50 60MBE LOW FULL MODUI ATION-NOISE ON 0E -SIRED STATION CAUSED BV UNDESIRED STATION

WORST CONDITION -BOTH S-ATIONSU./MODULATED

INTOLERABLE

ACCEPTABLE

UNDETECTABLE

ACIPICENT-CNANNELOPERATION-EMr ZSw2rWO.42.6,4C-.osimeoA004.4 sea WIYN300CYCLE roHEirTormN 4 2.004C

0113, -IT..4 0014-41V0 WiTIVpao.seA.4

DEGREE OF 1111.111111CROSSOVER

4 8 12 16. 20 24AMOUNT UNOES RED CARRIERETCEEDED DESIREO IN DB

R

FIG. /0

tone but all distortion products, leaving onlythe noise.

This effect has ..o doubt been obseryea bymany without being understood. It is in-herent in a frequency modulation system.

The noise threshold in the case of an54

OPERATION OF 2 F.M. STATIONS ON THESAME CHANNELBy referring to the section covering noise

interference it can be seen that the worstcondition of shared -channel operation oc-curs when both stations are unmodulatedand a fixed beat -note, therefore, results. Itwill also be seen that the higher this beat


note the greater will be its amplitude. Fig-ure 9 was made on the basis of the worstconditions, which occur when the differencein carrier frequency reaches approximately5,000 cycles. Were it not for the effect of de -emphasis in the receiver the beat -note am-plitude would rise with frequency. How-ever, de -emphasis of the high frequenciesprevents that from happening and the ef-fect may be further understood by referringto the section on pre -emphasis and de -em-phasis. It will be noted that the noise onthe desired station caused by the undesiredstation varies inversely with the deviationratio; F.M. 75 has a deviation ratio of 5compared with 1 for F.M. 15.

When either of the stations producing thebeat -note becomes modulated, the beat -notedisappears because one carrier sweeps acrossthe other one. When the desired station isapproximately 20 db. stronger than the un-desired station, interference and cross -talkeffects become unnoticeable. At 12 db. dif-ference they are noticeable but it is theopinion of some engineers that the 12-db.ratio would be tolerable. Frequency Modula-tion offers a great advantage over Ampli-tude Modulation in the allocation of sta-tions on the same frequency. In A.M. the

carrier amplitude of the desired atatioumust be 100 times, or 40 db. greater thanthe undesired carrier amplitude for a 40-db.signal to beat -note ratio. For F.M. 75 itneed be only 10 db., or 3 times greater. ForF.M. 30 it need be only 17.5 db. or 8 timesgreater. For F.M. 15, it need be only 24 db.,or 10.5 times greater.

The result is that F.M. stations can belocated much closer geographically, andtherefore many more station assignmentscan be made per channel. All interferencedue to sky -wave transmission from distantstations is automatically rejected in F.M.because the interfering signals never reachthe high amplitude required. This is not scin A.M. transmission.

Figure 10 shows the results of adjacent -channel measurements using one of the RCAField Test receivers and 2 commercialmodels of other manufacture. It should benoted that the undesired station was modu-lated with fixed tone of uniformly highmodulating level. As a result the interfer-ence was probably somewhat ;more severethan would be the case for program trans-mission in which the average modulatinglevel is rather low.

lh¢oty and azilyn eon.lidetationi

R. F. AND I. F. COILS INF. M. RECEIVERS

DURING the last 12 months radio iter-ature has contained a number of arti-cles dealing with the various phasesof Frequency Modulation from re-

ceive/. design to antennas and service. (e)Today the average reader of Radio -Craft ison speaking terms with F.M. even thoughsituated in localities not yet having thebenefits of F.M. service.

It is our plan not to'describe any particu-lar F.M. receiver or the operation of thevarious circuits in an F.M. receiver sincethis has been so well covered in the past.The art is new and rapidly progressing; theinformation which has been printed re-quires no summary and we shall thereforedevote this article to the design and theoryof the varioug coils and I.F. Transformersused in a representative F.M. receiver.

The ability of an F.M. receiver to receiveand demodulate commercial wide -band F.M.transmissions depends on the band -width

aee listing at end of article.

55

which the amplifier is able to accept, andsince this band -width is determined chieflyby the I.F. characteristics, we shall discussthe I.F. amplifier before turning to theR.F. and Oscillator stages.

THE I.F. STAGE -4.3 MC.The Radio Manufacturers Association,

after considerable deliberation, selected afrequency of 4.3 megacycles as a recom-mended intermediate frequency for use inF.M. receivers since a frequency above 4 mc.precludes the possibility of image frequencyinterference within the band of 42-50 mc.An I.F. of 4.3 mc. provides a guard bandbetween the amateur 3.5-4 me. band and the75 kc. transmitter frequency excursion.Thus the first design factor, the frequency,is already established.

The ideal overall response of a perfectI.F. amplifier for use with present trans-mission standards is shown in Fig. 1, at a.

Pra

ctic

alco

il ar

rang

emen

ts. A

-Wire

d an

d te

sted

rem

oved

to s

how

cer

amic

win

ding

Fort

unat

ely,

this

cur

ve d

oes

not h

ave

tobe

dup

licat

ed s

ince

the

actio

n of

the

limite

rte

nds

to r

emov

e al

l inc

rem

ents

in a

mpl

itude

beyo

nd a

cer

tain

leve

l as

dete

rmin

ed in

the

desi

gn o

f th

e re

ceiv

er. T

he d

otte

d lin

e re

p-re

sent

s a

suita

ble

leve

l of

abou

t 4 v

olts

at

the

limite

r gr

id.

Ban

d-pa

ss.-

If th

en, w

e de

sign

an

am-

plif

ier

havi

ng a

cur

ve s

uch

as b

in th

e sa

me

figu

re a

nd d

epen

d on

the

limite

r tc

ope

rate

as b

efor

e, w

e m

ay r

etai

n th

eat

lily

toal

ign

the

tune

d ci

rcui

ts w

ith a

n or

dina

ryun

mod

ulat

ed s

igna

l at 4

.3 in

c. a

nd s

till h

ave

esse

ntia

lly th

e sa

me

band

-w

idth

cha

ract

er-

istic

s as

the

idea

l ban

d -w

idth

cur

ve s

how

nat

a. B

y us

ing

a cu

rve

of th

is s

hape

, max

i-m

um b

and

-wid

th is

not

obt

aine

d at

ful

l re-

ceiv

er s

ensi

tivity

, but

this

sen

sitiv

ity w

ould

B.F

. and

osc

illat

or tu

ning

ass

embl

y; B

-F.M

. dis

crim

inat

orfo

rms;

C-F

.M.-

1.F

. tra

nsfo

rmer

for

4.3

mc.

(ca

n is

2V

2 in

s.

not b

e us

able

in a

ny c

ase,

sin

ce th

e ab

ility

of th

e lim

iter

to r

emov

e al

l tra

ces

of a

mpl

i-tu

de m

odul

atio

n, s

tatic

, etc

., re

quir

es a

sig

-na

l of

abou

t 4V

. as

alre

ady

stat

ed. I

n ad

-di

tion,

the

abili

ty o

f th

e lim

iter

to d

iscr

imi-

nate

aga

inst

am

plitu

de c

hang

es in

crea

ses

in p

ropo

rtio

n to

the

I.F.

sig

nal a

vaila

ble

atits

grid

.B

and-

pass

cha

ract

eris

tics

such

as

we

re-

quir

e m

ay b

e ob

tain

ed b

y th

e us

e of

tune

dci

rcui

ts, c

oupl

ed b

y a

mut

ual c

oupl

ing,

the

valu

e of

whi

ch d

eter

min

es th

e pa

ss b

and.

Con

sulti

ng F

ig. 2

and

dis

rega

rdin

g fo

r th

em

omen

t the

fac

t tha

t sev

eral

of

the

curv

esha

ve p

rono

unce

d "d

oubl

e hu

mps

," w

e se

eth

at 2

cir

cuits

of

cons

tant

"Q

" an

d ha

ving

vari

ous

degr

ees

of c

oupl

ing

desi

gnat

ed a

sk,

exh

ibit

vari

ous

band

-w

idth

s fr

om w

hich

tran

sfor

mer

, with

com

bine

d "a

ir an

d m

ica

tuni

ng".

long

). C

arro

n co

ils a

re il

lust

rate

d.

we

may

mak

e ou

r se

lect

ion.

The

val

ue o

f k

may

be

pred

icte

d w

ith s

uffi

cien

t acc

urac

yfr

om th

e fo

rmul

a:W

idth

of

Pass

Ban

d(1

)k

Mea

n In

term

edia

te F

requ

ency

afte

r w

hich

the

mut

ual i

nduc

tanc

e, M

, may

be s

et to

the

prop

er v

alue

fro

m:

(2)

M =

k V

Li,

L.

In p

ract

ice

this

val

ue is

set

app

roxi

mat

ely.

Fina

l adj

ustm

ent i

s m

ade

in th

e ac

tual

am

-pl

ifie

r, s

ince

unk

now

ns in

the

form

of

un-

pred

icta

ble

capa

cita

tive

and

indu

ctiv

e co

u-pl

ing

occu

r. T

hese

may

incr

ease

or

decr

ease

the

effe

ctiv

e co

uplin

g.A

lthou

gh a

sat

isfa

ctor

y cu

rve

coul

d be


4.1 2 4.3 4.4 4.5FREQ. /N MEGACYCLES

selected from Fig. 2 curve k .035, we findthat is has 2 distinct peaks with an areaof reduced response between them, a condi-tion entirely unsuitable for F.M. reception.Consult now, curve Q = 55 in Fig. 3 whichrepresents 2 tuned circuits having a valueof k = .035 suitable for our purpose buthaving the same "double hump" response.If we vary the Q of the coupled circuits,maintaining the coefficient of coupling con-stant, we may alter the "double humped"response to any intermediate value. Bear-ing in mind the action of the limiter, andthe band -width selected, we find that byemploying tuned circuits having a Q of 40a satisfactory response is secured. Thisvalue of Q may be calculated from thefc rmula :

1.4

(3) vQ. Q. = - or Q. = Q. -.=. 40k

If we desire the curve to be essentially"flat topped" the formula would read:

1.6

(4) VQ. Q. - or = Q. 42.9k

Q.-There are several methods of ob-taining the required value of Q. The coilsmay be wound on high -loss cores, woundwith high -resistance fine wire, or shuntedwith a value of resistance to lower the Qfrom a Mgher value. In actual practice,either the 1st or 3rd method is preferable,since fine wire usually results in finer in-sulation and increased distributed capacity,resulting in a lower L/C ratio with conse-quent reduction of resonant impedance. Theauthor prefers the 3rd method in whichthe coils may be designed for the most sat-isfactory winding from a mechanical stand-point. This results in coils of consistent

57

4.1 4.2 4.3 4.4 4.5FREQ. /N MEGACYCLES

characteristics, which are symmetricallyloaded with shunt resistors across the pri-mary and secondary. These resistors havethe two -fold purpose of providing (a) theproper Q and (b) a dissipative circuitwhich prevents the rapidly changing fre-quency from setting up transients. The lat-ter may be heard as a "fuzz," particularlyon loud, high -frequency passages.

L/C Ratio.-Although treated last in thisdiscussion, the L/C ratio of the tuned cir-cuit inductance to its shunt capacity mustbe borne in mind all during the design ofthe I.F. transformers. To provide a reason-able amount of gain at th, relatively highfrequency of 4.3 mc., we must employ asmuch inductance as possible compatiblewith stability. We know that the sum ofwiring, tube input and output, and coil dis-

4.2. 43 4.4 4.5FREQ. //V MEGACYCL ES

I- -

-I

21.4

71/1

0 43

S1-

4741

1-7'

.74,

7t64

7/..7

7.01

.0 /.

2.4r

44P

le40

.4e

# 47

1W

s.5

C04

41 4

5245

.01.

44 to

vs25

51.

051

50(C

,A,e

()

Jrz

6,4C

7 34.0

0004

,45

(EA

CH

)/.0

00.

0777

4S

L

A

u.kr

65,4

7

/000

0744

.5 410

W

RO

,20

0 O

uow

s

5000

007

7.5

172.

4

41/0

00 0

7445

//2.K

,

724.

4,00

0t

(.4

r,f,

6146 62

4

I II

Cl2

410

500

142

2R

73, I

lucO

.Y3

wm

e0 6

/16

42/ 5

0 W

O O

ww

3P3

figfa

.40

0V

One

.1/

3w65

,37

SO

S

43. )

000

Oor

ms

- 47 50

000

/gee

off6

0

/fa51

4/00

7

MA

IIAN

830

874,

g,5

;.043

.00.

..,42

5, Im

g-G

472

///79

. 4/2

. R20

ovc

ios/

uo

V3.

".ga

TC

/ 70

C/3

Avc

ios/

v2

'LW

0 /-

ME

G /.

/.2.4

/.0

5..7

414,

400

17.

R.F

.and

I.F

. ope

ratio

n of

a ty

pica

l F.M

. rec

eive

r. N

ote

unit

-bui

lt R

.F. s

ectio

n en

clos

ed w

ithin

dot

ted

trib

uted

capa

city

will

be

appr

oxim

atel

y 15

mm

f. W

e ha

ve a

vaila

ble

a m

idge

ttr

imm

erha

ving

a m

id r

ange

or

mos

t sta

ble

capa

city

of a

ppro

xim

atel

y 50

mm

f. T

his

resu

lts in

ato

tal s

hunt

cap

acity

of 6

5 m

mf.

Cal

cula

ting,

we

find

from

sub

stitu

tion

in th

e fo

rmul

a:1

(5)

2sV

LC

that

the

requ

ired

indu

ctan

ce is

21

µH. W

eth

en m

ay c

alcu

late

the

appr

oxim

ate

mut

ual

indu

ctan

ce b

etw

een

prim

ary

and

seco

ndar

yas

alre

ady

men

tione

d (f

orm

ula

2). A

mic

atr

imm

er m

ount

ed o

na

cera

mic

bas

eis

chos

en,

sinc

eno

elec

tric

alim

prov

emen

tco

uld

be o

btai

ned

from

the

use

of e

ither

perm

eabi

lity

or a

ir tu

ning

in c

onju

nctio

nw

ith th

e re

lativ

ely

low

Q n

eces

sary

. The

irus

e w

ould

res

ult i

n un

nece

ssar

ily in

crea

sed

cost

. So

far,

we

have

dis

cuss

ed th

e I.F

. Tra

ns-

form

ers

as a

gro

up. A

ctua

lly w

e ha

ve 3

type

s:(1

)In

put/I

nter

stag

ew

hich

are

iden

tical

, (2)

Lim

iter

inpu

t, an

d (3

) D

is-

crim

inat

or.

The

Lim

iter

tran

sfor

mer

diff

ers

from

the

Inpu

t and

Inte

r -s

tage

onl

y in

that

no

load

ing

resi

stor

is n

eces

sary

for

the

seco

ndar

y, b

e-ca

use

itis

load

ed to

a s

omew

hat g

reat

erde

gree

by

the

grid

cur

rent

dra

wn

by th

elim

iter

tube

. It i

s pe

rmis

sibl

e to

incr

ease

the

coef

ficie

nt o

f cou

plin

g in

the

limite

rtr

ansf

orm

er to

mai

ntai

n th

e sy

mm

etry

of

resp

onse

. The

diff

eren

ce in

cou

plin

g is

so

lines

.(S

ee N

OT

E a

t end

of a

rtic

le.)

smal

l tha

t it i

s no

t pra

ctic

al in

pro

duct

ion

to m

aint

ain

and

is s

eldo

m ta

ken

into

ac-

coun

t.D

iscr

imin

ator

-T

rans

form

er D

esig

n.-U

n-lik

e th

e Li

mite

r, th

e D

iscr

imin

ator

diff

ers

cons

ider

ably

from

the

I.F.s

bot

hin

me-

chan

ical

con

stru

ctio

n an

d in

ele

ctric

al c

har-

acte

ristic

s. T

he v

olta

ge in

put t

o th

e pr

i-m

ary

of th

e D

iscr

imin

ator

tran

sfor

mer

ispr

actic

ally

con

stan

t at a

bout

20

volts

und

erus

ual o

pera

tion.

The

vol

tage

del

iver

ed to

the

diod

e re

ctifi

er is

not

con

stan

t, ho

wev

er,

and

varie

s w

ith th

e fr

eque

ncy

of th

e ap

-pl

ied

sign

al a

t any

inst

ant.

The

prim

ary

and

seco

ndar

y w

indi

ngs

of th

e di

scrim

ina-

tor

tran

sfor

mer

are

iden

tical

to th

ose

ofth

e I.F

. tra

nsfo

rmer

s. T

he c

oupl

ing

betw

een

the

2 co

ils is

mad

e by

mea

ns o

f 2 in

depe

n-


dent methods. The coefficient of coupling isselected to give a peak separation of 250 kc.for reasons which will be discussed later.In addition, a capacity is placed betweenthe plate end of the primary and a center -tap on the secondary.

Let us consider the relationship betweenthe voltages in this transformer and theireffect on the design. It is necessary forproper discriminator operation that thevoltages developed in the 2 halves of thesecondary be EXACTLY equal and 180°out -of -phase with each other. This requiresan exact electrical center -tap. Even capaci-ties to ground must be equal and a smallcompensating condenser (a few mmf.) isoften placed externally from the unground-ed diode to ground to compensate for thesmaller diode capacity of this circuit.

In the operation of the discriminator thesecondary is tuned to exact resonance withthe mean I.F. Any mis-tuning of this cir-cuit results in phase shift and consequentnon-linear operation. Therefore, best de-sign practice dictates that we employ a coilwound on a ceramic form and tuned bymeans of an air trimmer. The primary wind-ing need not be so exacting since phaseshift occurring in this resonant circuit willnot affect the operation. For this reasonwe may use the more economical mica trim-mer. The primary must pass the full band-width of frequencies from 4,225 to 4,375 kc.with but little frequency discrimination. Wewill show how this is accomplished. Forproper demodulator operation the recoveredvoltage versus frequency curve should belinear within the frequency deviation en-countered.

The response curve of the discriminator,is a combination of the voltages developedacross the secondary due to the inductivecoupling with the primary, and the primaryvoltage introduced into the center -tap. Inform, it resembles a "double hump" I.F. re-sponse curve with one hump reversed tocomplement the other. This resembles a let-ter "S" laid on its side. Our object is tomake the center of the curve as straightas possible over the range of the frequen-cies included in the maximum deviation ofthe transmitter.

Inspection of curve k = .058, Fig. 2 willshow that the sides of the curve betweenthe mean frequency and 75 kc. are fairlystraight. The peak separation in this curveis 250 kc. Using this as a design factor wemay construct our discriminator transform-er. Although the sides of this curve are notabsolutely straight their differences aresuch that they cancel each other and alinear response vs. frequency results. Oneother factor, the primary response, affectsthe shape of the curve. In operation theresistance reflected by the closely -coupledsecondary tends to broaden the primary ac -

59

ceptance band to the full band -width whichsatisfies the condition mentioned in the be-ginning of the discussion of the discrimina-tor. No resistance loading is necessary onthe secondary as this is provided by theresistance of the diodes and their associatedcircuits.

R.F. AND OSCILLATOR STAGESFrequency Modulation transmissions oc-

cupy the band between 42 and 50 mega-cycles. Any experimenter who has builtamateur 5 -meter receivers is familiar withthe small size of the coils and condensersnecessary to tune to these frequencies. Sev-eral facts, however, are usually overlooked.The length of leads from the coil, size ofthe variable condenser (length of electricalpath), bypass condenser leads, etc., all con-tribute to the inductance in the circuit. Forthis reason trouble is usually encounteredwhen gang condenser operation is at-tempted.

Coil Design in Theory and Practice.-Atthese frequencies it is ABSOLUTELY IM-POSSIBLE to design a set of coils on paperand expect them to work perfectly. The onlymethod by which perfect results may beobtained is to follow the preliminary designwith actual application.

Here the exact circuit and proposed lay-out are constructed and final inductance ad-justment is made. Coils so designed will givethe maximum performance in THIS LAY-OUT ONLY and any attempt to installthem in a similar circuit with a differentphysical arrangement will give unsatisfac-tory results. For this reason manufacturerswho supply F.M. components in kit formgenerally furnish the complete R.F. and Os-cillator sub -assembly completely wired andtested. Commercial receiver manufacturersare specific in their service notes to theextent of stating: "If it becomes necessaryto replace any part be sure to put the re-placement part in the exact position occu-pied by the defective part and use exactlythe same lead length originally employed."Bearing these remarks in mind we willproceed with the actual coil design.

Condenser Tuning-Ratio.-As stated, wewish to cover a range of 42-50 mc. If weuse a condenser to tune to these frequencieswe may employ the formula for inductance,capacity and frequency to determine thetuning ratio of the variable condenser(ratio of total minimum capacity to maxi-mum capacity). To permit a slight spaceat each end of the dial we assume a rangegreater than actually required and in thecase of F.M. our total range will be 39-51mc. If the inductance L is held constant wewill require a condenser ratio which isequal to the square of the frequency ratio.We will therefore have:


CFI( /8R.4 rE vat r;,.E rTe 1

°5C/L[470R -4,47;7"-I

(---

C,R

le. /..1nMEM 1

TEST I I C' ''=' I

I

L_ _ _ _ _ I

L- - - - _Jflar4M1

jp"N himmiCANW

-- 1-'c' 2

lE"'\b ___---...11---_______--

4111111--...M.3 2

i III

,; MAPIJII_____11110

./111111Ate

4.

FIG.5B -4' FIG SCFic. 5-Testing and applicational circuits. A-Block diagrams of coil -testing setup with which finaladjustments are made. Band -width and gain -per -stage may be measured by means of the instrumentrepresented hero. B-Diagram showing the introduction of undesired inductance at (I) due to bypassingthrough a long lead to ground; (2) long lead to tuning condenser; (3) long lead to screen -grid, withregeneration introduced due to the common inductance of the 9round path between the gang andground (4) and bypass ground (4a). Long lead (5) contributes Inductances to the tuned circuit butintroduces a loss of gain. C-Detail showing how overall inductance may be increased by moving turnsas shown by I, or decreased, 2; or the effective tap position raised, 3, or lowered, 4. Larger changes intap (tracking) position may be made by moving the position of the turns above and below the tap.

(6)

(51 mc.- )_ Condenser ratio = 1.7139 mc.

This value L71 is very small and is notordinarily obtaihed with any availablevariable condenser. There are, however,several solutions to this problem:

1) We may place a large shunt capacityacross the coils and use a small variablecapacity for actual tuning. This methodhas the disadvantage that unless the coilsare very small, an extremely small variablecondenser is required for tuning. If the in-ductance is reduced the L/C ratio becomesunfavorable, the parallel resonant im-pedance is lowered and the available gainis small.

(2) A small series condenser (semi -fixed)may be used between the coil and the tun-ing condenser. From the loss of gain stand-point this method is superior to Method 1.It has several disadvantages such as criti-cal adjustment (stray capacities), and dif-ficulty in obtaining proper tracking be-tween the oscillator and R.F. stages.

(3) A combination of both these systemsmay be used which will result in retainingthe better features of both. This consistsof placing a small semi -fixed trimmer di-rectly across the coils to provide alignmentat the high -frequency end of the band. Astandard small variable tuning condenser

60

may be tapped down on the coils. This taphas the same effect in limiting the tuningrange as the series condenser in Method 2.It has the additional advantage that ad-justment of the inductance above and belowthe tap allows us to vary the tuning ratioand accomplish tracking over the narrowband required.

Its disadvantages lie in a somewhat high-er distributed capacity and a tendency toresonate at 2 distinct frequencies. This lat-ter fault is not important in a superhetero-dyne, since the effect is such that it is im-possible for the oscillator to beat with theundesired frequency to produce the I.F. Nopadding condenser is necessary with thiscircuit (general high -frequency practiceusually eliminates the padding condenser)as this adjustment may be made by as sim-ple a procedure as bending the tap leadwhere it leaves the coil. By this means itsmutual inductance may be added or sub-tracted from the effective inductance tuned.

Design Example.-For the sake of designlet us use Method 3. Connect a small "high -lift low -capacity trimmer" across each coil,assuming a wiring and input capacity, plustrimmer capacity, of 22 mmf. for R.F. andOscillator circuits. For the time, ignoringthe maximum tuning capacity, we can cal-culate the inductance necessary to resonateat the highest frequency (51 mc). Employ -1", formula (5).


We find the required value to be approxi-mately .44 µH. This value may be obtainedby winding 5 turns of No. 18 bare wire,spaced twice its diameter, on a 1/2 -in. O.D.bakelite tube. Using this value of inductancewe find by again substituting in formula (5)that approximately 37.7 mmf. will be re-quired to tune to 39 mc. Rather than use asmall condenser (37.7 mmf. - 22 mmf.15.7 mmf.) as a tuning condenser we maychoose a condenser several times as largeand tap down on the inductance to give thesame result. If we select a condenser hav-ing a range of about 90 to 100 mmf. we maytap down to about 1/6th of the total in-ductance. This method of design neglectsthe minimum capacity of the tuning con-denser but due to the impossibility of de-signing these coils without actual applica-tion, this discrepancy may be corrected 'inthe final adjustment. This method permitsus to utilize approximately 80% of the con-denser rotation to cover the required band,resulting in an easily read dial.

R.F. and Antenna Primaries.-The designof the R.F. and Antenna primaries also rep-resent a compromise, since extreme caremust be taken to keep the capacity betweenthe primary and secondary low. One ofthe most satisfactory methods is to space -wind the primary between the spaced turnsof the secondary. Care must also be takento prevent the R.F. primary from resonat-ing with the tube output and wiring capaci-ty, Within the band employed. Since atbest, the impedance of the primary aspresented to the plate of the preceding tubeis far too low to permit a reasonableamount of gain, it is necessary to use asmuch inductance in the primary as possible,without conflicting with the conditions men-tioned. A ratio of 2 primary turns to 3 sec-ondary turns represents a workable value.

The antenna primary is a less criticalmatter. The impedance reflected into atransmission line from a doublet antennashould be approximately 70 to 100 ohms.The impedance of the primary may be con-sidered as the sum of the impedances ofthe primary alone plus that coupled into itby the secondary. This coupled impedanceL equal to:

(7)(2n Mr

Z.where M is the mutual inductance betweenprimary and secondary and Z. is the paral-lel resonant impedance of the secondary.

The condition for maximum transfer ofenergy is such that the coupled impedance

is equal to the primary impedance and thecoupled reactance is equal but opposite insign to the primary reactance. Formulashave been developed for the calculation ofthe ideal conditions, but due to their com-plexity they are seldom used and the samevariables as we encountered with the designof the I.F. transformers affect the couplingand alter it from the calculated value.

For our purpose we may calculate theprimary impedance based on about 1/2 thedesired impedance and from formula (7)and a knowledge of the effective circuit Qdetermine the approximate required mutualinductance. We may assume that the cou-pled reactance will approach the requiredvalue and from this starting point thepriniary may be constructed. On the R.Fcoils mentioned above this results in 2turns interwound with the primary starting1/2 -turn outside of the end secondary turn.

Oscillator Coils.-The exact constructionof the oscillator coil will depend on the typeof circuit employed. Electron -coupled oscil-lators may be used with the 6SA7 or witha separate electron -coupled oscillator. Thesecircuits require a tapped coil.

The position of the tap should be deter-mined experimentally to give the most suit-able amount of feedback. Coupling shouldbe close in any case (interwound if a 6A8converter is used). The oscillator induc-tance should be adjusted to a slightly lowervalue than that used in the R.F. stagesand a ceramic -insulated air trimmer shouldbe employed to assure permanence of dialcalibration and freedom from drift.

By this time the reader has probablycome to the conclusion that the design oftuned circuits for frequencies such as em-ployed in F.M. is a rather haphazard busi-ness. This is true to a certain extent. Weshould remember that the principles in-volved, are a starting point in the design,and their use at lower frequencies may bemade with greater accuracy. At these lowerfrequencies stray capacity and lead lengthsdo not represent so large a proportion ofthe total circuit constants.

Note.-In the bank of 8 condensers which arebypasses for the input 6A7 tube, the unit whichbypasses the 1,000 -ohm filter resistor to groundshould be instead a 0.001-mf. mica condenser.

It may be of interest to note that condenserC which shunts the heaters serves only to reducethe impedance of the heater circuit close to thesockets of the assembly. This is necessary be-cause of the relatively long heater wires whichhave considerable inductance and which thereforemay, under certain conditions, allow a certainamount of feedback between these tubes.

61


F.M. PHONO PICKUP

Pictorial illustration of the new F.M. Phono Pickupshown in more detail in Fig. I.

HARD on the heels of Frequency Modu-lation broadcasting comes this in-genious development of a Bridgeport,Conn., radio man. Here for the first

time in any radio magazine are the com-plete details for the home construction ofthis "wireless" P.M. Phono Pickup, a unitwhich bids fair to replace, in time, all othertypes of pickups. Patent applications havebeen made by its inventor.

FIDELITY

The most amazing thing about this Fre-quency Modulation Pickup, developed byLeslie A. Gould, is its extreme simplicity.If this were its only achievement the newinstrument would be outstanding; but itgoes much further! Its fidelity range-theband of audio frequencies which it is ableto transmit-is said to go considerably be-yond that of the ordinary crystal and mag-netic types. Being a Frequency -Modulateddevice, its inherent range of frequencies is

62

limited primarily by themechanical serrations in therecord groove.

What is the frequencyrange? It must be remem-bered that this is an experi-mental pickup. More highlyengineered commercial mod-els would undoubtedly exhibitbetter performance. There-fore it is especially inter-esting to note that as nearlyas the inventor can judge thefrequency response of the model here illustrated is approximately 16 to 8,000 cycles.It is expected that with an improved stylusholder it should be quite possible to repro-duce up to 15,000 cycles. The latter fre-quency "top" of course presupposes thatthe recording extends out to this high fre-quency.

However the bottle -neck in present-dayphono record reproduction is not in the re-cording but in playback. How does the out-put voltage compare with crystal and mag-netic units? With present models, approxi-mately the same. Whereas formerly, thevarious types of pickups were used as ameans for modulating a carrier frequencygenerated by any other instrument such asan oscillating tube this unit performs thefunctions of both. The R.F.-carrier fre-quency generated by a built-in vacuum tube

is shifted back and forth infrequency, a process whichin itself is a form of modu-lation.

ReadRADIO -

CRAFT

LIBRARY

BOON"

SERIES

HaveYOU

The

Complete

List?

CIRCUITA mach better picture of

the extreme simplicity ofthis device may be had byreferring to the circuit shownin Fig. 3. Here a type 6C5 isused as an oscillator tube in

the simplest type of oscillatory circuit im-aginable. To get down to the FrequencyModulation band an oscillator coil consistingof 9 turns of No. 20 enameled wire wound ona %-fn. Lucite form is used. This coil ismounted at the forward end of the pickup.

Like any other oscillatory circuit, anymetal placed in the vicinity of this oseillat-

US

E 1

/46'

TN

/ER

Aiu

mm

umO

R O

TH

ER

ME

TA

L/2

-LO

NG

x'7

2' T

YT

OE

SID

ES

MA

KE

CO

VE

R O

KS

AM

E M

AT

ER

/AL

mE

N-

RE

mO

VE

ST

YLU

SM

OLD

ER

DR

ILL

9450

/4 H

OLE

IN

®.w

ear

5.oe M

M -

DIS

CA

RD

r/w

5P

AR

T

TY

PE

6C

5 O

SC

0A0/

0 T

UB

E BY

PA

SS

CO

ND

EN

SE

R.0

06 -

MR

EN

D O

F S

WIF

T, A

ND

INS

ER

T W

ON

T-F

ar/N

0B

RA

SS

RO

O O

R W

IRE

I ER

YE

RU

BB

ER

ON

SH

AF

TS

MO

DIF

IED

HO

LDE

R

1111

012

TO

NE

AR

M

5010

ER

A ,/

2R

/NO

TO

TH

E I"

LO

NG

"5.E

XT

EN

DE

O S

HA

FT

RE

PLA

CE

TH

E S

TY

LUS

HO

US

ING

US

ING

TW

O15

02 S

CR

EW

S A

kEl N

oss

INS

TE

AD

OF

RIV

ET

S

DR

ILL

2 H

OLE

ST

HR

OU

GH

Bor

row

. OF

TO

N1E

AR

M F

OR

ST

YLU

SA

,0D

ER

SC

RE

WS

(US

E L

OC

K w

AS

NE

RS

....IT

N N

UT

S)

NO

TE

T

ES

T T

O S

EE

TA

N, R

w5

AN

DS

TY

LUS

5w

w0

'RE

EL,

QF

ILA

ME

NT

VO

LTS

FIG

. 1

45.9

0 V

OL

r5

LIG

N7

ME

TA

L R

ING

//2-0

/A F

AS

TE

NE

D 7

0S

HA

FT

OF

ST

YLU

S H

OLD

ER

RIN

G /5

PLA

CE

D R

flOuT

%B

-F

RO

ME

NO

OF

05C

CO

IL (

SP

AC

ING

NO

T C

R,T

,CA

L)

GR

ID R

ES

IST

OR

50.0

00O

HM

S

6R/0

CO

ND

EN

SE

R25

0 M

MF

GR

IDW

IRE

CA

TH

OD

E

AD

JUS

TA

BLE

CO

ND

EN

SE

RF

OR

TU

NIN

GA

CK

UP

OS

CIL

LAT

OR

WIR

E?O

DE

S/R

ED

PLA

CE

BA

ND

EN

O W

IRE

OF

Cow

SO

. CE

DE

DT

O F

RN

NE

OS

C C

OIL

N92

071

1N,N

GG

RID

RE

SIS

TO

RE

NX

M W

IRE

, CLO

SE

- C

ON

DE

NS

ER

500

00 O

HM

SW

OU

ND

25 lO

omm

FG

R,0

250

c0m

m,H

oeN

sER

6C5

11

TA

P 2

RO

uNC

TT

UR

NSF

RO

MC

-WA

ICS

CO

NN

EC

rioN

V2'

°IA

ME

TA

LR

ING

FA

ST

EN

ED

TO

ST

YLU

S H

OLD

ER

SH

AF

T

F I

G3

FIL

45-9

0V

FO

R 4

2 T

O 5

0 M

EG

AC

YC

LE B

AN

D,

CO

IL W

OU

ND

mw

rm 9

TU

RN

S N

O 2

0E

NA

ME

L W

IRE

,Ce0

5E-W

OU

ND

.TA

PP

EO

AT

SE

CO

ND

TU

RN

FO

R C

ON

NE

CT

ION

TO

CA

TH

OD

E O

F 6

C5

'12.

BA

KE

L Ir

e oR

lucT

TE

-Col

LF

OR

A..

3/4

LON

G T

HR

EA

DE

D IN

CE

NT

ER

FO

R 6

-32

SC

RE

W. F

OR

FA

ST

EN

/ND

CO

IL F

OR

M T

O 3

10E

OF

TO

NE

AR

M.

DR

ILL

V4'

0. M

OLE

FO

RN

EE

DLE

TO

Eyr

ewo

TH

RO

UG

H

ST

YLU

S N

OLO

ER

(SE

E F

IG 2

)

ST

Y! U

S((

OLD

ER

NE

ED

LE

FIG

.4

OS

CIL

LAT

OR

CO

ILO

R/D

RE

SIS

TO

R

ME

TA

LR

ING

AD

JUS

TA

BLE

CO

ND

EN

SE

R

BY

PA

SS

GR

IDC

ON

DE

NS

ER

,C

ON

DE

NS

ER

C

11O

SC

ILLA

TO

RT

YP

E 6

C5

Con

stru

ctio

n de

tails

of t

he n

ew F

requ

ency

Mod

ulat

ion

Pho

no P

icku

p sy

stem

. The

circ

uit o

f the

acc

ompa

nyin

g F

.M. o

scill

ator

isgi

ven

inF

ig. 3

.

ing

coil

will

cha

nge

the

freq

uenc

y of

the

circ

uit.

Cap

italiz

ing

on th

is p

heno

men

on,

Mr.

Gou

ldm

ount

sa

smal

lm

etal

ring

adja

cent

to th

is o

scill

atin

g co

il, a

nd s

ince

the

ring

is m

echa

nica

lly f

ixed

to th

e ne

edle

-ho

ldin

gst

ylus

orar

mat

ure,

the

vibr

a-tio

n of

the

need

le is

tran

slat

ed in

to th

e

mec

hani

cal m

otio

n of

the

met

al r

ing

whi

chin

turn

shi

fts

the

freq

uenc

y of

the

circ

uit

back

and

for

th o

ver

the

defi

nite

ran

ge d

eter

-m

ined

by

the

reco

rdin

g.T

hat i

s th

e en

tire

unit.

Few

res

onan

cepo

ints

, no

expe

nsiv

e pa

rts,

no

com

plic

ated

elec

trom

echa

nica

l sys

tem

s re

quir

ing

delic

ate

arm

atur

es, a

nd c

ryst

als

or p

erm

anen

t -m

ag-

nets

.

CO

NS

TR

UC

TIO

ND

etai

ls f

or th

e co

nstr

uctio

n of

this

pic

kup

may

be

obta

ined

fro

m th

e va

riou

sill

ustr

a-tio

ns w

hich

acc

ompa

ny th

is a

rtic

le. T

hese

ALL ABOUT FREQUENCY MODULATIONare explanatory and complete in themselves.

Tlie body of the pickup was made froman old, east -oft crystal pickup. The oscilla-tor tube for shortest possible leads is mount-ed directly on the pickup arm. The readermay wish to improve upon the method ofmounting the stylus and can usually do so.The method shown in the drawings is sim-ple and very effective. However, there is noquestion but what better methods can befound and employed. It is merely necessaryto say that the minimum amount of fric-tion or damping should be used in the me-chanical attachment of this stylus to thearm, since the frequency range of the unitis limited mainly by its mechanical system.

The 1/2 -in. ring used in this pickup can bean ordinary brass curtain ring found inmost 5c -and -10c stores. These rings, beinghollow, and very light and stiff, are idealfor the purpose. When soldering the ringto the stylus shaft use as little solder aspossible thereby keeping the stylus as light

in weight as possible which will give thebest results on the high -frequency portion ofaudio reproduction.

`WIRELESS" PHONO-OSCILLATORNo antenna is necessary. The pickup

transmits a Frequency -Modulated wave di-rect to your F.M. radio receiver up to 50ft., or above, under usual conditions. If anantenna is desired, a short piece of insulat-ed wire about 6 ins. long can be connectedto the cathode termitlal of the radio tubesocket and allowed to extend through therear end of the tone arm.

The F.M. signal from this pickup can alsobe received on a superregenerativd type ofreceiver. It may be necessary in some loca-tions that are noisy to disconnect the regu-lar antenna of your F.M. Receiver and con-nect in its place a short, 1 -wire antennaabout 3 to 6 ft. long when tuning -in on thispickup.

List of Articles in Past Issues ofRadio -Craft on Frequency

ModulationNew Circuits in Modern Radio Receivers (Dept.):

"New Circuit Using 'Electric Eye' as F.M.Tuning Indicator" (Pilot) ; "Direct -CoupledLimiter Tubes Used in F.M. Receiver" (Zenith),April 1941. "Frequency Modulation Receiveruses 2 Limiters in Cascade" (Scott), Dec. '40."Supplementary Shadows Indicate F.M. Reso-nance" (Meissner), Nov. '40. "Pushbutton Am-plitude-Frequency Modulation Changeover"(Stromberg-Carlson), Aug. '40. "Same TuningIndicator Used Both for Amplitude and Fre-quency Modulation Receiver" (Stromberg-Carl-son), "Tuning Indicator for Frequency -Modulation Receiver" (Stromberg-Carlson).March '40.

Recent Improvements in F.M.-Receiver Design,March 1941.

A New A.F.-Drift Correcting, Signal -Balancing,Direct -Coupled F.M. 24 -Watt Audio Amplifier,Part I, Dec. '40; Part II, Jan. '41; Part III,Feb. '41.

Circuit Features of the Latest Ultra -$.F. "DX"F.M.-A.M. Receiver, Feb. '41.

F.M. Servicing Pointers, Jan. '41.The ABC of Frequency Modulation, Dec. '40.Station WOR Gets F.M. Voice. Dec. '40.Present Status of F.M. Broadcasting, Dec. '40.F.M. Servicing Procedure. Nov. '40.Servicing P.M. Receivers. Oct. '40.Servicing F.M. Receivers. Sept. '40.Latest 28 -Tube DeLuxe High -Fidelity F.M. and

A.M. Broadcast Receiver, Sept. '40.Build This Practical F.M. Adapter, Aug. '40.Choosing an F.M. Antenna. July '40.Data Sheets (Dept.): "Stromberg-Carlson. No.

425 Frequency Modulation Receiver and 'Con-verter'," Feb. '40.

Frequency -Modulated Programs on Your PresentReceiver!-with This Easily -Built F.M.-A.M.Ultra -Shortwave Adapter. Part I, Dec. '39;Part II, Jan. '49.

64

Radio -Craft Library SeriesPRESENTED on this page are the other hooks in the RADIO -CRAFT LIBRARY SERIES

-the most c piste set of volumes treating individually important divisions of radio,refrigeration and air conditioulug. Each book has been written for the purpose of givingyou the opportunity to specialize In one or more of the popular branches of the subjectsmentioned. The material in these hooks is very helpful . . you'll find them valuable Inyour work.

The authors of these books are well known-they are authorities on the subjects onwhich they have written. This is the first time that you are enabled to build a library ofuniform technical books by such popular writers.

Here Are the Titles-Book No. 13

ABC of Mr ConditioningTechnical Review of the Fun-damentals of this Latest

Branch of Engineering,Including Service Data.

Ity PANE D. HARRIGAN

Book No. 11ABC of Refrigeration

A Compendium of the Princi-ples of Refrigeration with aDetailed Listing of Defectsand How to Overcome Them,with Hints on Setting Up a

Service ShopBy TRAFTON MASON

Book No. 113Practical Radio Circuits

A Comprehensive. Guide toRadio Circuits for the Service

Man. Constructor andExperimenter

By DAVID IIELLARE

Book No. ISPoint -To -Point Resistance

AnalysisTheory-And Applications ,ofThis Modern Test Procedure

In Service Problems.By BERTRAM M. FREED

Book No. 19Practical Radio Kinks -

and short CutsA Compendium of Practical,'time -saving Methods for theConstruction. Operation andRepair of Radio Receivers.By It. BAKER BRYANT

Book No. 20The Cathode -Ray Oscilloscope

Theory and PracticalApplications

By CHARLES SlcURANZA

Book No. IIBreaking Into Radio ServicingSimple instructions and Pro-cedure for Starting a Profit-

able Radio Businessof Your Own.

By ROBERT EICHBERG

Book No. 22New Radio Questions

and AnswersAnswers to question.. mostfrequently asked by both no-

vices and experts.By ROBERT EICHBERG

Book No. 23Practical Public .tddress

Modern Methods of Servicingand Installing Public Address

EquipmentBY B. ItAKEit BRYANT

Book No. 24Automobile Radio

Principles & PracticeA Complete Treatise on the

Subject Covering AllPhases from Installing to

Servicing and Maintenance.By B. BAKER BRYANT

Book No. 25Home -Made Radio Test

InstrumentsContains Circuits and Ideasin Test Equipment WhichMake It Possible to Service

Radio Receivers MoreQuickly and Profitably.

Compiled by the Editors ofRADIO -CRAFT

Book No. 20Modern Battery Radio Set

Fully Describes BatteryPortables, Direction

Finders, Home Radio Sets.Boat Radio Receivers,

Experimental 1 -Tube Sets,"Personal" Receivers, etc.

Compiled by the Editors ofRADIO -CRAFT

Book No. 27Modern Radio Servicing

TechniqueVarious Met boils of

Dynamic ServicingStage -by -Stage Analysis,Emergency Servicing and

Repair of Unidentified -MakeSets are Described in this

Book.Compiled by the Editors of

RADIO -CRAFT

Book No. 2RAll About Frequency

ModulationBasic Principles. and theDesign, Construction andRepair of F.M. Receivers

are Described for theRadio Serviceman and

Experimenter.Compiled by the Editors of

RADIO -CRAFT

ALL BOOKS UNIFORMThe books In the RADIO -CRAFT LIBRARY SE-RIES are all uniform . . .

size 6 x 9 inches. Each vol-ume contains an averageof 50 to 150 Illustrations.The books are printed onan excellent grade of hookpaper.

RADCRAFT PUBLICATIONS, INC. * 20 VESEY ST. NEW YORK, N. Y

CLIP COUPON-AND MAIL TODAY!

RADCRAFT PUBLICATIONS, IN('., 20 Vesey St., New York, N. Y.I have circled below the numbers of hooks in the RADIO -CRAFT LIBRARY

SERIES, which y-ou are to send me POSTAGE PREPAID. My remittance ofIs enclosed nt the cost of fifty cents for each hookCircle numbers wanted:

Name Address

City StateSend remittance by check. or money order: register letter If you

send cash or stamps.

13 15 16 IS 19 20 21 22 23 24 25 26 27 2/4

Ma:M.3M' f 0331

14 New "Type Oa SetviceRADIO CIRCUIT MANUAL -1941

The Only EDITED ManualEver Published !

DIRECTORY OF RECEIVERSMANUFACTURED IN 1940AND UP TO JUNE, 1941

THE ACTUAL'AGE SIZE IS

10 IT 11Vis Inches

7718 oln TIPS: MASI ALS-6.11ty. Ikea bare to handle.

A PERMANENTINSTITUTION

1 11- I NMI-, A (IONS, IN('_. J. .rsey St.New ...irk 4 't

l( ADP t1 51 Ntst Al 1141

MORE INFORMATION IN HALFTHE NUMBER OF PAGES

The valise of a service manual u measured not by tlot rut,. rof page. but by Ilse mosint of useful inform..., I'm, in.sly 736 pages this Rad, Circuit Manual cur." oser ..0,e.. models MORE than does any sits, rnmoeur.manual in twice the number of pugHOW DID WE DO 11'7

. . By i mg tbe sire of mu page: by discarding nressential data and editing the balance; by listing only thoosereceivers .high the Egon,. Engineer wil definite:, have iorepair (no communications or export receivers no shoetrees..'et, or amplifiers. no electronic devices, etc 1; by mansmontb nf lord work based on a definite plan of pros -m.11.0and a clear understanding of the acts,' requirements of lb.Semler. En M "deadgineer. ire is no "d weight' infounatnoto add bulk to thisManual &my word rooms. Every mir iiof reading time ...ell spent.

OUTSTANDING FEATURES Contain* data on more than

.800 race .vm noodelel-nrorthan any other redo servicemanual.

Only 736 ' lest thanhalf the bulk of any othermanual and more tkian 1.3lighter.

All information EDITED! --all non -essential data deleted andthe halanee checked modtoted with the schematics andLetches

4.1'1. I*.,,, page permits lutingof all inhumation nn one page.(A few idable caeca surcepted. 1

I.F. peal. for all supehet me-culla are boldly displayedblack bones,- none mown., allmount,

No sptee wasted on conununicalion. and import receivers milplifiers, electronic nu...al instruments. etc 100'; Se.,ice Engineer's Manual

A "CUSTOM-TAILORED" MANUALFOR SERVICE ENGINEERS

Here. at last, is a Service Mmissal deliberately PLANNED for the Service Engineer In.Need of mere hodgepodge collection of rervice data. as manusils l irebeen in the nestthiii RADIO CIRCUIT MANUAL it orderly compilation of esonntial rad, diagramsand service talormatinn. carefully edited and uniformly presented for the frustu,' conye nience of the busy Service Engineer. All lime...morning. non.emential dat hams

beenweeded out. and the remaining information, vitrilly important to the unfurl and efficientservicing of modern radio receivers. has been laid oui in logical. emy.redorg stylewhich cuts time from the Jay'. worts. Bream. of this and other Natures schich are sell.evident upon first observation. it has been possible to list all information pertaining to agiven model an single ate..le 736 page. thia Manual 'semen. essential oervice data on over 1800 re. eiv tr nint1els,more than soy other ....I... service manual an tbe market'

ONLY ONE MANUAL PER YEAR!The new techmoue used compiling Ili. RADIO CIRCUIT MANUAL 7041pouible to include in a merle book all the net. receiver model* which the radio indususcan product in tingle year. This factor alone represents an important saving to oltService

ORDER YOUR MANUALIMMEDIATELY -NOW:

ONLY

1000RADCRAFT PUBLICATIONS, INC.

20 VESEY ST. NEW YORK, N. Y.

Documents

ve 0 - americanradiohistory.com · presents as compared to amplitude -modulated systems. 3 "A.M." AND "F.M." A concept of amplitude modulation (or "A.M.") may be obtained from the