A Study Of The Feasibility Of Designing A Superregenerative Reciever To Meet Certain Critical Requirements..pdf

1

A STUOY OF THE FEASIBILITY OF DESIGNING A SUPER-REGENERATIVE RECEIVER TO MEET CERTAIN CRITICAL

RE~Urn.EMENTS

Thaddeus Francis Kycia, B.Sc.

A thesis submitted to the Faculty of Graduate Studies and Research, McGill University, in partial fulfilment of the requirements ror the degree of Master of Science.

ABSTRACT

A super-regenerative reeeiver has been designed for linear mode operation in a band from 455 Mc ./see. to 510 Me ./see.

For

1inear mode operation it was neeessary to use automatie gain stabi1ization, whieh a1so kept the oscillator output pulses at a steady amplitude. The reeeiver was constructed, having a band-width of 630 Ke./see and a noise figure of 20 db.

The experimenta1 resu1ts were in

good agreement with the theoretiea1 estimates, and numerical values for eomparison are provided wherever possible. The reeeiver ean deteet a minimum signal of

1.5~volts,

and

it therefore IOOre than meets the sensitivity recpirements for use as a deteetor in a l.I.h.f. "bridge".

ACKNOWLEOOEMENTS The writer wishes to express his appreciation to Dr. J.R. Whitehead who originated the project and under whose direction i t was carried out;

to Dr. H. G.I. ivatson for allowing the use of his laboratory

equipment and workshop, and to Mr. W. Avarlaid, Mr. B. Meunier, Mr. M. Kingsmill, and the staif of the Physics building for their

kind cooperation. Special thank s are due to Dr. T.

w.

h. East for hi s frequent

assistance and advice during both the project and the writing of this thesis. The writer is also indebted to the Defence Research Board, whose financial assistance made this work possible.

FIGURES 2.1

P~Gé

Block diagram of the super-regenerative receiver

l'

3.1 The \.I.h.f. oscillator 3.2 Analagous lumped circuit of the

13

.h.f. oscillator

~o

3.3 Plate current vs. grid bias voltage of the 5876

Jf

3.4 Transconductance vs. grid bias voltage of the 5876

2>

3.5

Determination of the gm cycle

3.6 Determination of the conductance cycle 3.7 Relative frequency response of the receiver

'1

3.8 Noise band-width measurement

'J ~

3.9 a) Output noise for no input signal

"

3.9 b) Detected pulses which form the noise

."

3.10 a) Output for large input signal

37

3.10 b) Superimposed output pulses of equal amplitude

37

3.11 a) Output for an amplitude modu1ated input signal

li

3.11 b) Formation of the sine wave

"38

3.11 c) Relative separation of the pulses

'3<1

3.11 d) Superimposed pulses for the waveform of fig. 3.11 a) 31 4.1 Frequency bridge of the quench oscillator 4.2 The quench oscillator 4.3 \.I.h.f. detector and automatic gain stabilizer 4.4 The meter circuit 5.1 Receiver, front view 5.2 Receiver, back view 5.3 Receiver, bottom view (without shi~ld) 5.4 IIBridge" measurement setup

~I

CONTENTS

1.

2.

3.

4.

Introduction 1.1.

Fistory of super-regenerative receivers.

1

1.2.

Purpose of the receivers.

g

Princip1es of Super-Regeneration 2.1.

Super-regenerative action.

2.2.

The conductance cycle.

If)

2.3.

Linear mode and a.g.s.

Il

2.4.

General description of the receiver.

II>

12.

The Super-Regenerative Osci11ator 3.1.

Construction of the oscillator.

Ir

3.2.

The 5876 penci1 triode.

/1

3.3.

The equivalent circuit of the plate càvity.

1'1

3.4.

The graphical determination of a-.

'-:1-

3.5.

Theoretical formula for the output.

!J.:J

3.6.

Super-regenerative band-width.

'30

3.7.

Noise figure.

'3~

3.8.

Oscillograms of oscillator output pulses.

'31'

Other Sections which constitute the Receiver 4.1.

Quehch osci1lator.

4.2.

The u.h.f. detector and a.g.s. system.

4.3.

The mater circuit.

5.

Construction and Use of the Receiver 5.1

The 1ayout and operation of the receiver.

5.2

The super-regenerative receiver as a "bridge" detector. tI?

Summary and Conclusions

Appendix

References

1.

'7

1.

INTRODUCTICN

. 1.1 History of Super-regenerat1va receivars The principle of the super-regenerati va osci1lator Was first introduced by E.H. Armstrong in 1922.

Its high gain 1fas reallzed, but its other remark-

able features were ovar100ked and i t was on1y thirteen yeaxs later that turther investigation was pursued into the subject.

H. Ataka, 1935; M.G.

Scroggie, 1936 and F.W. Frink, 1938 published papers exp1aining the phenomenon of super-regeneration.

Lt about the

SaIlle

time severa1 unique circuits,

making use of super-regeneration, appeared in print. W.B. Lewis and C.J.!. M.ilner (1936) designed a duplex operation transceivar (simu1taneous two wsy cOImnUnication) by mald.ng use of the reradiating properties inherent in the super-regenerative osci1lator.

It acted alterna-

t! valy as a recei ver and transmi tter during each quench periode

Each trans-

ceiver would be sensitive to in-coming pulses at that specifie time when the signal pulse just arrivad from the other transceiver. S. Becker and

L.M. Leeds (1936) described a two

w~

police radio system.

This radio installation consisted of duplex operation trom headquarters to eaeh car and simplex operation between the cars.

The reeeiver had a com-

bination of both the super-regenerative and superheterodyne princip1es and the resultant sensitivity was more than sufficient to fUlfillthe rigorous requirements. The super-regenerative circuit then found its

w~

into numerous radio

amateur receivers, as weIl as the "walkie talld.e" developed by the U.S. A.r'tlJy Signal Corps and later modified for use during the Seoond World War.

It

bad the desired features of baing sensitive, light, simple and economioal. Other contributors who promoted the understanding of super-regenerati "f9 theory were W.E. Bradley (1948); A. Hazelti"s; D. Richman and B.D. Lo\l..ghlin

(1948); Macfarlane and Whitehead (1946, 1948); H.A. G1uckSlD;Bll (1949) and L. Riebmann (1949). The super-regenerative receiver became we11 known for its part in the wartime I.F.F. (identification friend or f'oe) Mark 111 responder, deve10ped at T.R.E. in Great Brita1n.

The purpose of' the I.F.F. responder was to

indicate to an observer at a radar station whether soma distant aircrait Was t'riandly.

The respondar on the a1lied aircrait wou1d amplif'y the re-

ceived radar pulses and reradiate them.

Theae were picked up by the radar

station and observed as long traces on the P.P.I. screen, compared to the small dots which represented the refleoted pulses t'rom an enemy aircraf't. The use of' a.g.s. (automatic gain stabilization) was made in these responders and this gave them the added f'eature of good stabi1ity over a wide t'requency range (157-lS7Mc./ sec.).

More than 200,000 of' these t'ully auto-

matic responders were produced to be fi tted into al1 the allied ships and planes as a joint effort by the United states and Great Britain. "Super-Regenerative Receivers" by J.R. Whitehead (1950) describes the circui t of' the Mark 111 responder, as wall as ma.ny other circuits making use of' the super-regenerative principle.

The book a1so contains a comprehen-

sive treatment of' super-regenerative theory as we11 as physical interpretation of' the resu1ts. Another major piece of work published after that time on super-regenerative theory was by HeA. Whee1er in the form of two monographs on the analysis of super-regenerative selectiv1ty.and design f'ormulas. 1.2 Purpose of the receiver The

o~ect

of' the work undertaken was to stuQy the feasibi1ity of a

super-regenerative receiver operating in the Unear mode at 500Mc./sec., with specific properties which could make it use able as a "bridge" detector.

" It would 'he used as a null indicator for a General Radio admittance me ter • The specifications are that it would be able to detect IO/f'\t of signal or better and that it would have a sufficient amount of stability.

The par-

ticular receiver discussed here cao detect 1.5;'f'V: of signal and achieves its very reliable operation through the use of the automatic gain stabilization.

The purpose of the a.g.s. is to keep the receiver in the linear

mode of operation, thus adding to the complexi ty of the receiver.

On the

whole, however, it is much simpler than a superheterodyne operating at the srune trequency.

/0

2.

PRINCIPLE

OF SUPER-REGENERATION

2.1 Super-regenerativa aotion A parallel tuned circuit can maintain oscillations i f enough energr is supplied te it te overcome the resistiva losses.

If, at a.ny time, more

energy is supplied than dissipated, the oscillations will grow. A. positive feedbaek amplifier, which may consist of one tube, can be

made to supply the nacessary energy.

The condition that oscillations be

maintained is that the transconductance of the tube equa.l a certain value which depends on the losses in the circuit.

The transconductance is a

function of both the plata voltage and grid bia.s voltage.

By periodical1y

varying ei ther ·o f the se , one can get tœ tube into a s tata where the oscillations grow, or wbere they decay.

The general practice is to modulate

the grid bias voltage with the output of a quench oscillator whose waveform, for simplici ty, is sinusoidal.

Over part of the quench cycle the grid bias

is very negative and no oscillations existe

The grid then becomes Iess œga-

tive until it reaches the grid bias voltage, which is just necessary for oscillations to existe

The osoillations begin to grow as the bias becomes

still less negati va, until they reach a peak in amplitude which comes at a time wben the grid bias returns to its critieal. value.

Beyond that instant

the oscillations decay and, under proper opersting condi tiens, are well below noise at the time ..han the next cycle of build-up begins.

The output

of the super-regenerative oscillator is in the form of pulses at intervals equal te the period of the quench frequency. 2.2

The conductance Cycle There is an al ternativa method of stating the cri ta ri on for

tions in a tu.ned circuit.

o~illa­

In effect, the tuned circuit can be represented

1/

by an

inducti~~capacitance and conductance, all in parall.el. The conductance -. ~/

part of the circuit · damps any oscillations that may be induced in it.

The

positive feedback amplifier, to which the tuned circuit is connected, acts like a negative conductance and, when the net resultant of the conductance across the tuned circu! t ls zero, the tube is in a state of steady oscillations. That i8 equi valent to saying that enough energy ia being supplied to overcOlll8 the resistive losses.

Basides acting as a negative conductance, the 08cll-

lator tube also has some small reacti ve effect, which causes a variation in the frequency of oscillation during the build-up cycle. When the oscillations are growing, i t indicates that the net conductance across the resonant circuit is negative.

The net conductance varies directl1'

as the transconductance of the tube which, in turn, varies as the grid bias voltage.

The conductance cycle, as a function of time, can be obtained gra-

phically by knowing the variations of the grid bias volt&ge with time. A complete theory of the super-regenerativa princip1e has been formulated in which use i8 made

ot the conductance cycle curve. A numerical example,

showing ho. the experimenta1results compare wi th the theoretical formula, is illustrated in a later section. 2.3 !anser mode and a.g.s. It was stated earlier that the super-regenerati va (uhf) oscillator was to operate in the linear mode.

Under that type of operation, amplitude modu-

lation on the input signal is linearly reproduced by the output pulses.

The

requirements on the uhf oscillator are that the groTdng oscillations œver reach saturation, if i t is to operate in the linear mode. The oscillations always start tram noise and the signal induced in the oscillator at the input.

If the oscillator had i ts mean D.C. bias so adjusted

as to give adequate gain for low signal amplitUdes, it wou Id not require much

or. an inerease in signal amplitude to have the oscillations reaching saturation. Automatio gain stabilization is, therefore, necessary. The detected output pulses are _amplified and rectified by the a.g.s. system to give a negative D.C. voltage which oontrols the mean bias of the u.h.f. oscillator tube.

Now, if a large signal is applied to the u.h.f. oscillator,

the a.g.s. system makes the grid bias more negative, thus keeping the output pulses of the oscillator approximatelY constant. 2.4.

General description of the receiver

Jt. block diagram of the receiver is shown in figure 2.1.

The output of

the quench oscillator is at a frequency near 50 Kc./see. and bas a peak amplitude of about five volts under normal operating conditions. The a.g.s. system consists of the amplifier, phase inverter and rectifier.

The a.g.s. amplifier is tuned to quench frequency Md so i t responds on1y to the .general frequency component of the detected pulses. fore, a quench frequency sine Wave.

Its output is, there",

The phase inverter, followed by a double

diode deteotor, gives out the D.C. control voltage.

This voltage provides

the bias for the u.h.f. oscillator, thus completing the feedback loop. The other rectifier, connected to the output of the phase inverter, gi ws out a positive output which is compared to an adjustable D.C. voltage by the microammeter.

For this case, the meter ia at the "Signal Strength Indicator"

position, since changes in the output of the a.g.s. are caused by changes in the input signal amplitudes.

~

QUE:.NCH 05C.1l-l..ATOR

RCC.T\F\E.~

~

1

METE.R

1

1 QÙ~NGH

VOLTAGE

INPUï

SI~NAl..

SOPE'R'R(GENEAATNf R . y:: ~1'-l..ATOR

r---

OCTEC.ïOR

Q .•.

À.G.5. AMPl-IFïE:R

Q .F.

f---

Q.F. PHASoe:. IN'ILRTER

Q.I='.

D.C..GON""O\.. '101-,.,o..ca~

F"IG. 2.\ ~'-oc.K OlAG'R AM OF' THE: Sl,)~~R·Rt:CaENERAi\VE. RE.CE:IVER

REC.T\F'~R 1---

"-

V-\

Ir

!ha oa'dtv oœb1nation bad t.en pre'tioualy llM la aD ad ... oonwrted 1Dt.o

UN t~r aD

o..lllator br 1ll8Olatlng ths grid OODDl.Uoa

tJoca tt. aromd ad t1tUng in a teedbuk loop.

n.cun

'.1,

u.h.t. apl1ts..r

!ba oeell1ator, u

.bon la

oozud.lIt. ot • 'tame! iDplt aaYitq at tM aat.bode and. a t1uw4 oat.-

pu.t o.Y1:ti7 at tJ. plate of tbl S8'16 paDaU triode. b

gr14 i . grounded to li .h.r. 0e01l1at1o~ tbroqh C, ... 06.

c,

0 ...

111... of oapult.aDn betwHn the tn 11198 ot the tlup holder &Bd poo11D4.

C,

~m be!JII tbI dieleotn.. IIiaId tfta tœ paraUel plate

bu a ftlue et 100 pt. &Bd

t_'a te haw a 'ftlue of 300 pt.

oapaoltaœe. aM up to gs.w 4DO pt., wb10h bu

aD.

0, ...

deter-

!bit two

!Jape4aH of 0.'5 . . a"

SCOllo./.... In ao'bal. practio., the Wo aadUe. of the 5876 t.lng the ~ part

or

&1"8

le •• than 1 ca.

a~

tœ tube _"ea t.bI NdUe ••

o8oth. and plate of the tube tlt lnto 1IOa:D.t11J1 blooka, wbioh are

ooupàd to the quarter waw lim..

.ra

Or ad 02,

l.eDgth of

Ioth thI oaPUlt.aJ»~­

., iDlreu1.ag tM oapMi'" iD t.be

the .tteoUw elHtrioal

gl'i4 ct the triode ' 1 . gifta ..

tàI f'l.aIIp

2~5

1enath

of

ta l1aI 18 lDOftU84,

...u.. 'thu

pt. !hie doe. Ilot afteot the electorioal

tt.. liDi very .uah, e1Doe 1'.m oathoda 1. ooupled to 1t at tba low

1apeàalMJe end.

The plate, on 't.J» ott..r haII4, 18 ooa.pled to the bigh iape4aHe-

_. of the lJ.llII and

.0

the plate to

lrid oapuita.Doe ot 1.4 pt. oontribtte.

_h to the .hartleaing of the Uœ iD t.œ plate aad'tu, !lai oeaU1&tioo freqano;r oan be ftJ"1ed

&8

b7 aeau ct \

1. lhowD ln til. J.1. and °2

trca 455 W

'10 ~.I..o. '1'be lnpllt aIId output loopa are 81tuaW at the low iapedanoe, ar hip

U. H. F. OSC'LLATOR

INPUT C.ONNt.C. 'TOR

INPUT

1..00P

6.l \1..

lAI'"

i

c..

C:A'T"'OCE C...."(\TY

~

T

J:'i:['DeJlCK

-=?"

OUTPUT OF'

QUE:NC.H 05C. ~

~.P.

c.. ,

R.F.' c.. 2

(

bblJG

\0 K,n.

RI

C2

C.OUP\..lNG

OUTPUT

1

or A..cas.

1 [ --,'-----------------'mTIT?--R.".C.. & 1

1

Pl.-A.'TE.

stI'?6

C"''II'T'(

OUTPU'T

F\G.3.\

CONNE.CïOFt ïO oe:"Tt:.C,"TOA

2.:'K•.n. ~ &

& .. 300'1.

...... vt

16

UHF OSCILLATOR Li st

0

f Compon En t s

Cathode cavity Plate cavi ty V1

5876

~.f.

High-Mu Triode

tl.h.f. trimmers Input 100p Output loop Feedback coupling C3

100 pf. ceramic condenser

C4

coupling capacitance of 130 pf. to cathode cavity

65

coupling capacitance of 130 pf. to plate cavity

C 7

0.05/,",f.

R1

l<X..n.

R2

2.7ILn...

R.F.C. l

choke to quench oscillator

R.F.C. 2

choke to a.g.s •

.tl.F.C. 3, H.• F.C. 4

chokes to heaters

R.F.C. 5

choke to cathode

R.F.C. 5

choke to plate

C6

consists of capacitance, amounting to 300 pf. between the grid and the cavity through 0.25 cm of polystyrene insulation.

/7

current ends of their respective cavities.

The loops are coupled to the strong

magne tic fields which exist at those ends.

The feedback loop, on the other

band, is coupled to the magnetic field in the plate cavity and electric field in the cathode cavity.

The electric field coupling is in the form of a capa-

citive probe. Preliminary test showed that, at its best, ·the receiver could just detect rive microvolts of input .signal.

This indicated that the input termination

was weale and so the inpat ' loop was increased in size. was inserted at the output.

The same type of loop

Atter these modifications, - it was evident frem

an observation of the low D.C. control voltage supplied by the a.g.s., that the DeW

loops increased the lQading on the cavities.

To overcome this extra load-

ing, the feedback loop was improved by lerghening the capaci tive probe. lowing this change, the D.C. control voltage went to its normal value.

FolIt

was then found that the minimum detectable signal decreased to one and a halt microvolts. The super-regenerative gain innepers is given as

à/4J

where à represents

the integral wi th respect to tim.e of the negative part of a conductà,nce cycle and C 1s the equi valent capac1 tance of the plate cavi ty.

One would then ex-

pact to have the gain of the rece1ver to 1ncrease with frequency , s1nee C decreases and the corresponding D.C. control voltage from the a.g.s. todecrease the gain by an equal amount, by simply becoming more negative and tlms deereasing ~

This was the case from 455Mc./sec. to about 490t6.c./sec.,

after which the D.C. control voltage started becaming less negat1ve as the frequency was increased to 510 ~ c./sec.

S1nce the 5876 can osc1l1ate at a fre-

quency as high as i700 MC./seC., it is quite justifiable to assume that the transi t time of the 5f!J76 would not cause such a large effect.

The conclusion

ia that the efficiency of the feedback loop drops with 1ncreasé in frequeney

in the receiver band. 1 . 3.2 The 5876 pencil triode The 5876 is a new tube developed only' recently for u .h.f. frequencies and can be used as an oscillator up to l700M c./sec.

It has the features of

minimum. transit time, low cathode to plate capaci tance, 10.. lead inductance and good thermal stabili ty.

As much as 6 watts of energy can be dissipated

at the plate if the plate cylinder should have a large surface of contact wi th i ts support.

The tube, when oscillating at 490M c./sec., dissipates at the plate 2.7 watts for no signal input, 1.5 watts for a large continuous wa~rt~~, f.~I;'''' · 0.5 volts and 6.0 watts if the grid should accidentally became ~o~ed~

If

the oseillator should be detuned so that no oscillations exist, the plate dissipation would be 3.7 watts.

The tube should not be allowed to dissipatë.

more than 3.5 watts at the plate for aD1 long length of time, sinee the polystyrene insulation for the plate support would weaken with the bigh rise in tempe rature •

3 • .3 Equivalent circuit of the plate cavity In order to give a numerical example of the oscillation build-up and deca:y, it is first necessary to find an equivalent lumped resonant circuit for the plate cavi ty. in fig. 3.2.

An analagous lumped circuit of the oscillator ls illustrated

No attempt is made tofind the equivalent resonant circuit far

the cathode cavity, sinee it does not enter the theory directly, but bas soma second order effects. The line of the plate cavity can he represented by an inductance, a capacitance and a

conductance, all in parallel, while the trimmer can be assumed

to he an extra variable capaoi tance across the tuned circuit.

Of the three un-

knowns, the capaci tance and inductanoe arma will be first determined. The trimmer aots llke a paralle1 plate oondenser with adjustable plate separation and air as a dielectric. 2 to be 2.2 cm and their separation to 455 MC./seo. (fI).

The effective area of the plates was round

w~s 0.10 cm. when the osci1lator was tuned

For an oscillator frequency of

average separation was 0.18 cm.

480

MC./sec. (f2 ), the

The capacitance of the paral1el plate conden-

ser is given by the expression 1.11 KA/4rrd pf., where K is the dielectric constant and equals uni ty for air A is the area of each plate in cm2 d ia the separation between the plates in cm.

The formul8. gives a value

of 1.95 pf. for 455 Mc./sec. and 1.08 pf. for 480 MC./sec., the change in capaci tance

C thus being

O.~

from the fundamental relation

where Substituting in the values

pf.

The value of the inductance L is found

" .

'.:,.

~ ,

....

" ~~"

.-

...

..

;.

:J.D

5 0-

B ~~AWI'II~

It-

u.. ~~

l1 2

~"4:

.

~

-

ta..

:J/

::

0, (;) l-a

y, ""',

For f - 488 Mc./sec. the equivalent capacitance f • 500 Me./sec., it is 7.4 pf.

1/

(2nf)2 is 7.7- pf. and for

The next problam was to determine the equiva-

lent conductance. It is knoWll that the bandwidth of a quiescent circuit is gi ven by b

G

-2 ::'c

at - 3 db., where Go i6 the equi valent paraIIel conductance of the qui-

escent circuit. been calculatad.

To find Go i t is only necessary to measure b since C bas already The procedure was as follows:-

The power in the receiver was turned off while the output of the signal generatorconnected to the output loop of the receiver was increased to above

0.5 volts at 1$0 Me./sec. tuned to

480

Me./sec.

The cathode cavity was detuned and the plate cavity

For a fixed output of the signal generator, the detected

portion; which was fed into a D.C. amplifier on an oscilloscope, gave an indioation of the amount of off-resonance of the plate cavity from the signal generator frequency. At resonance the cavity absorbed little power and thus a larger output was observed on the C.R.T.

When not in resonance, the impedance of the

cavity, observad at the loop, decreased and so did the output voltage.

The de-

tector had been previously calibrated so that the frequency could be set 't othe values_ giving an output of 3 db.balow resonance.

The advantage of this method

was that there were always large detected signaIs to be observed, even at -1 neper from resonance. "as

The quiescent band-width obtained for the plate cavity

b. 2.6 Mc./sec.

(at -3 db.)

The equation for the conductance in parallel 1d th the resonant circuit is

(3.4.1.)

then

• 12y mhos. The equi valent circuit for the plata oan ty is now approXimataly known.

3-.4 The graphical determination of aA graphie method for f1nding the value of a- under a partieular set of condit10ns is descr1bed fully in Chapter 5 of "Super-Regenerative Receivers" by

J.R. Whitehead.

It i8 the most accurate method sinee it makes a1.most no use of

approximations for the osc11lator tube characteristic curves. As had been menrJ-U"-' a. ~l1 d..Ltioned previously, a- ia the integral with respect to time ?f the negative part of the resonant circuit conductance G.

where

The expression for G is

qUi.s scent conductance of the tuned circuit k:

constant for gi ven circuit conditions

gm •

transconductanee of the oscillator tube

From this equation i t is apparent that when oscillations Just start, G ia zero and gm has its cr1tical value &no.

k:

The expressioll f9l'kis" .thereforé.,

Go/&no

The charaeteristic values of the quieseent circuit were

2.6 Mc./sec.

Band-width

b •

Capacitanee

C • 7.9 pf.

480

Resonant frequency

fo •

Me./sec.

Shunt conductance

Go • l29~ mhos.

Other data consisted of QueÏlch frequency

fq: • 54 Kc./sec.

Quench amplitude

Jq • 5.0 V

5876 plate voltage

Va • 250 V

Output pulse peak

Vl • 0.25 V

Input signal

V •

l2.5~V

Bias for oscillations to begin Vgo • - 3.4 V Mean grid bias voltage

(at-3db)

. . 'r . . ~In

grid bias voltage ourve for the 5F!f76 triode is not !. availab1e in the published tube data and so it was neoessary to obtain it exThe transoonduotanoe

perimentally.

First, the P+ate ourrent la was reoorded for different grid bias

voltage values Vg , the plate being kapt at 250 V. and shown in fig.

3.3. The slope of the ourve

The resu1ts were p10tted

6LJAEa whioh

is defined as &Il

was then found for differenot Vg values and similar1y p10tted in fig. 3.4. required part of the transoonductance (gm) cycle oan be seen in fig.

The

3.5. It

is simp1y obtained b,y projeoting the grid bias voltage as a funotion of tima on the tube transconductallce as a function of time.

a.. grid bias

In the third quadrant i8 a 5 volt quenoh oyo1e super-

imposed on a -7.4 volt mean grid bias. osoi1latory reg10n being above gmo • of - 3. 4 volts.

voltage, and getting the transoonductanoe

In the first is the resu1t, with the

3.5 m. mhos. corresponding

to a grid bias

The value k is noll' found from e quation (3. 4. 2.) to be

0.037 x 1~·. Knowing this, the conductanoe oyole in fig. 3.6 is oonstructed from the transoonduotanoe oycle by using equation (3.4.3.).

On the same figure i8 a

grid bias voltage oyole showing fram the conduotanoe ourve how the osoi1lator behaves during various parts of the quenoh periode

The· regenerative period,

during whioh the input osoil1ations are amplifie d, is followed by the superregenerative period, during which the oscillations grow rapid1y. damping period and the existing osoillations deoay.

Then comes the

The detected output pulse

is also sketohed from the traoe on the osoillo.oopa, with i ts peak at the end of the super-regenerative periode

The area a- Was measured and found to be

l05.~

mhos.~seo.

3.5 Theoretioal formula for the output Provided that several oondi tions are met, the expression for the osoillation amplitude aorOS8 the plate oavity is given at any time t b,y

\8

IGt

S8r6 U.H.F.

HICaH MU

T~IOCE:

v.,, zSOV.

14

IOJ IYnI.OJ.

10

,

-JO

-8 F'ICa.l~ PLATt.

-~

-6 ·

tr

:

VOL..T~.

C.URRENT V~ CaRlO al~ VOLTACaE

o

58"7E, U.H.r. H1GH MU TRIODE Vit>..:' ZSO V.

'7

6

3

/ z

-8 ~

-.

o

VOL..TSa

F'lCa. 3.4 TRANSaC.ONOUCi'ANC.ë: vs. GRIO

!aIA~ VOL...TAC.e:.

...' .

-.: ~

w ~ ~

~~

.

\J

~

....t) ~

2

'" J

0

)0-

0

( ~

w r

~

0 1:

t-

l:~

~

ls.

,,~l:Ul

0

z

0

Q

~

z

~

,

N

OC

Id 1IJ

il

\II

.

.ri

~

•

"

li.

..•

•

W II)

~

0-

~

III

•

~

'"

.~

... 0

Id J

>

2,

"

~~

10

JI ~

~

v

lai

il ~ tr ~

-

\l

H

0

?

•

' ''f

o GR\D &l~$

-2

vo\..TAC.t. c."'(CL.E.

-4

-8 0

10

1 1 1

'I~

- -

1

CAM~ Re..c..t.~I:.AATIVE.tR(~yP~\'Itt-PERIO, . ç:>e..RIOO P6Gào

Df'\MPINCa PERloe

1

100 C.ONO\JCT~Ct.

G

CYC~&

1$

/-< MHOS.

O+--------+~~--_+----~

___r_------~------------~-

~O

-u

fi'- SEC..

-s 0·1

v,

Ol

VOL.TS 0.1 o+-----------~-----------~--~~-----~------------~10

s

F"IG.3.tO DtïERMlNAT'ON OF' Tl-\~ CONCUC.,.ANCe: CfCI-E.

~

y,~)

:: Vs K

G.r

..l-

(!

~~J z ( :.) ~ [ ;t

:b JG{X) ~

~[

-e

(;;;)z]

~ [WoT -f(W-W~Z-,]

.;(',

wbere

vI.. t) = oscillation

= absolute

Vs

amplitude at time t

.

r

signal amplitude

Wo

= 21ffo

: radial frequency of the osc111ator

LV

- 211" f

: radial frequency of the signal

= capac1 tance of the plate cavi ty = derivative of the conductance with respect

C G'(t,)

~

;...

~ 1..

,. .

)

to time at t

=t,

The ratio of V, over Vs cail be obtained from above and then greatly aimplif1ed when the input signal i8 tuned to the rece1ver.

v~ = G. [ s

:,pc;I7 ~[ ;~ f G('()~

whare the éinusoidal For t

=t 2

;t-

1

c

It ia given by

('3. 5 .:1)

~

variation with time is also laft out.

the oscillation amplitude V, ia simply the peak of the output

:r~

pulse and -

Jz,

G(x) dx becomes a- as Was defiœd previously.

The last equation

then reduces to

V~=G.

(3.5.3. )

.J'

wbich can he rewri tten 80 as to get an expression for a-, namely

a--

= -'c h

[[C~,)]:f ~~:J

(3.5.4.)

= II~~~.~~ which compares favourably wi th

lO~

mhos.ft sec. from the gr.aph1cal method.

Judging from the reading accuracy in conducting the valillus measurements, i t is estimated tbat the value of C is lalown to witbin : known to w1 thill t 10%.

20%

and that of b ia

The possible error for a- is about 20%, since most of

its contribution comes tram the C outside the logarithm in equation (3.5.4.).

T-he discrepancy in the two values of a- is well within this error. An expression for the total receiver gain Nt CaD. he obtained from equation

(3.5.3.) by taking the natural logarithma of both sides.

~*

=

This gives

[ v'::-~ J]

h

= .~ [ Go [FK,(;t-,)] { ] ~ ~ = so that

#0::

~

)10 +}1s

(3.5.5.)

~ [G.[ C;'C.i"J1] ~

(3.5.6.)

and

(3.5.7.) No is called the slope gain.

It ia the contribution to the total gain that

oceurs during the regenerative period, when

the

net conductance across the reso-

nant circuit is positive but is dropping to zero.

Ns has already been mentioned

bafore and is called the super-regenerative gain, since it acts during the superregenerative periode

No worka out te he 2.62 nepera or 22.S db and Ne, as ob-

tained graphically, is 6.65 nepers or 57.0 db.

Under normal operating conditiops

the mean grid bias voltage may be much smaller, in which case the super-regenerative period would be longer and the regenerative pariod shorter.

That would

be accompanied by a 1arger Ne and smaller No_

There were several conditions which had te he met in order that equation (3.5.1.) be valide

Firstly, it is a formula for the slope-contro1ed state and

the requirement.'for slope control ia that

-

12 C

wbere to is defined in fig. 3.6.

-*, -..;r

0

-:.:

(3.5.8.) It is found that

"3. S"./"'"'~.

-

/2 ~

G.

(3.5.9.) (3.5.1~

which satisry the condition stated in equation (3.5.S.). The oscillation amplitude ia actually obtained as the sum of two parts, of

which VI is the more important term if the condition for slope contr01is satiefied and that (3.5.11) The first condition has been met and the second is also satisfied from knowing that (3.5.12)

t2 - to • 3.~ see. and fram equation (3.5.9). 3.6 Super-regenerative band-width Equation O.5.1) shows how the super-regenerative oscillatbr

~eaets

to fre-

The frequency response S(f) oan be

queneies other than that of the reeeiver. described by srJ_) !t'Y'

=:.

a

~ ~j

e

[

Wo --~,

::L

~r

.

l

(c.v- W.,)

G'U,)

-m,II!'

~

]

(-1-;-0)']

G ,~,)

which behaves Iilœ a Gaussian error-curve for f

(3.6.1)

not[!~:~different

from to;

The band-width b s • 2(f-fo) is obta1ned approximately when

~ [ -~1T'~:lf.)'J" ~[-/J ks

=

~[

S:,J( ~

(3.6.3)

Fig. 3.7 gives the relative response of the receiver in db. width

91 is measured at -8.7 db.

to be 0.63 Mc./see.

from equation (3.6.3.) ls

t

/J."

lite.

!.AJ..e..

0.6.2)

The band-

The expeoted band-width

'3/

- 0 ~'9.0

~6,.O

F"~. ~.'"l RE~Tl\IE F'RE.QUE:NC.Y RESPONS[. 01=' ïHE ~ECElve:R

which compares very well with the experimental value. The ensrgy band-wi.th for noise calculations ia derined as +OQ

J 1sr-hl" 4-

--

:
where the experimental results of IS(f)/2 are given in fig. 3.8. S(t o ) • l, the area under the

band.-width.

C'l.n'Ve,

Since

which is just the integral, gives the

This is illustrated by means ot a rectangle having a béight

equal to uni ty aDd a width equal to the band-width, which turned out to be 0.38 Mc./sec. 3.7 Noise figBt! The performance ot the receiver must be compared with an 1dea1 receiver with respect to noise.

For normal purposes it ls only possible to receive

signaIs in a band whose width equa1s the quench frequency f" so that the 1dea1 comparison receiver must have a band.-width

t,,-

54 Kc./sec. a1so.

The noise generated in a 54 EC./sec. band-w1dth 1s

7 where K • Bottzman's constant 1.37 x 10- 23 jou1esl'°K T • Temperature

o~

the resistanoe in degrees Kelvin

R • InternaI res1stance of the source in ohms. f • Rece1ver noise band-width in cycles/sec.

It was necessary to determine the amount ot noise present in the receiver • .1. 100% amplitude modulated signal was fed to the receiver and the signal am-

plitude was decreased ontil the peak noise level in the valleys ot the

'.' -

~

'.

.

",

419.0

480.0 S\GNÂ\.. F"RtQUt.NC't IN Me:/ ~EC..

flCa,3 ,B

NO\~~

aANO-W 10iH Mt.A~U~E.M~Nï

modulation rose to the mean carrier level.

The r.m.s. carrier voltage Was

3.0"pv. which indicated that the "1% leveln had a value of 4.24,r v.

If we

assume what happens to be the peak value of noise as the level which is exceeded for 1% of the time 1 1

./ :

"Jo ,

1.82~v.

l.

this level is 2.3.3 x r.m.s. value, thus giving

for the r.m.s. noise.

.. .

Now

! '.

.fl. /-1- - s

.

' ..

.!lr

= 10

=

and the approximate noise figure

~ 0 x~. 10

::L:l 0

d!.J,-

It has been shown by H.A. Wheeler that the noise figure of a super-regenerative receiver is larger than that of a straight amplifier, having the sarna band-w.i.dth, by a factor which equals

where b s • the super-regenerative band-width and

ff. the quench frequency The reason for this is that the super-regenerative receiver is sensitive

to noise in a band equal to b s , but i t heats wi th the component frequencies of the general spectrum and appears in a band who se width equale the quench frequency ft" The ratio can he worked eut and is

I:J .0 Subtracting from the noise figure the value of this ratio, yeu obtain

9 db,

which is a reasonable noise figure for the 5876 triode alone in a first-class conventional circuit.

3.8 Oscillograms of oscillator pulses The 500 MC./sec. oscillations leave the oscillator in the form of pulses at quench frequency.

These pulses are detected and observed on an oscilloscope.

Fig. 3.9a shows the randam noise output wben no signal is

presen~while

in

fig. 3.9b one can see the superimposed output pulses for only noise at the raceiver input.

The pulses have a Gaussian wavef6rm and a wide distribution

in amplitude corresponding to the wide amplitude variation of the noise at the input.

Fig. 3.10a shows the detected output for a large, continuous wave sig-

nal at the input, with fig. 3.10b giving the constituént superimposed pulses. These are line.

ail of the same amplitude and he.nce the trace appears as a solid

One can observe how the pulse resembles a Gaussian waveform, as had been

pradicted by the ory • Fig. 3.il iilustrates how the pulses f»l'Im an amplitude modulated sine wave.

In fig. 3.lla is shown a linearly amplified output of a signal whose

amplitude was modulated with l Kc./sec.

In the next pictura, one can see the

individual pulses, of which it takes 54 to reproduce one wavelength.

The third

picture shows the relative separation of the pulses and fig. 3.111 gives the pulses aIl superimposed.

There are sharp limits at the top and bottam which

correspond to the pulses describing the top and bottom peaks of the amplitude modulated signal.

FIG-

3.CJ (~)

OUT'PlJT NOISE

Fo~

No

INPuT

SIG-NAL

1 1

1

1

1

,

. Fl a.

:S .l O Cc:).)

,! OUTPVT

FOR

L..AR.<;.~

" INPUT

SlcPNfH_

FI~ 3. Il

(0)

FïG- 3.11 (b)

OUTPUT

FoR

FORMATiON

AN

AM'?L.I(UDE

OF THe. S/WE WAVE

M0.DUL.A'EJ) INPVT SI6-NIH..

'3f

FIG- 3.(1

Cc)

FIG s.1I (cL)

REl.AT'''E

Sç'PARATIDN

SIJPEllIM'PoSE:D

PUt..SES

O~ T"HE

puLSeS-

FoR THG w~VEFORM OF FIG- 3.II{o.J

4.

OTHER SECTIONS WHICH CONSTITUTE THE RECEIVER

4.1 The guench oscillator As had been mentioned bafore, the quench oscillator voltage modulates the The frequency can be varied from 44 Kc./sec.

u.h.r. oBcillator bias voltage. to

58 Kc./sec. and normally operates near 50 Kc./sec. The oscillator is of the

bridge type, being a modification of the Wien type. network is shown in fig. 4.1. voltage to the amplifies.

ei. (ef - e r ) and eo are the input and output

The ratio of er to e o is given by 1

-!:i-=

..L

The frequency determining

(4.1.1)

0

--.-L :3

=

1

-= ::.

A (-!::::i

:e." D

11{"i

- ~ ) -t.- 0

-k)

It is, the re fore , necessary to make K greater than 3 and to have no phase shift (no imaginary part in A) at resonant frequency. in fig. 4.2.

ft

A diagram of the circuitVshown

The 3 watt tungsten 1amp R is a non-linear arm of the bridge and

accounts for the stable oscillations.

Its temperature i6 high enough to pre vs nt

instability.

Rp

acts as a grid leak and RL as a cathode bias resistor.

couple the quench oscillator trom the restor the circuit.

C4 and R4 de-

It was also impor-

tant to shield the oscillator trom the a.g.s., which is tuned to the detected quench frequency pulses from the super-regenerative oscil1ator.

The band-width

of the tuned amplifier stage in the a.g.s. is 2 Kc./sec., a maximum drift of

..... . . r

" - :.

~-!... ~.. _

: •• -.

'1/

.

.

.

~

.~. ; . ~ ...: ~' "

'.

,

...

'

t ,

"

",

'

, ..

• • Ô:"

"

.,

'

, '.

0

w

: ... .

. . ...

... .. ,~

.

OC

0

,

tJ ~

~

0

J J

OC

\1)

-dl v-

0 ~ ÙJ: z W 1:. :l

..,

' : ';

"

lX

J

-v. . "-

...

.~

l.L

~

8tt Wcr IL.

Z 1~

0 ,. ?"

'

co

v

;,r~l': ;~l~~~~{

·/,'tt,~f;·f.;) ':· l' .

~.:

...

..

.-:'

.

;1 . t

\~i;~'.'E>

""

.....

.,.:

',,',',' *~~ ~1,

... " .')

.,,:

.If

• ••

QUE:NCH OSCILLATOR RIS r----------------------------------------------------------,-------------------r-------.v·~..~.~~~~

22K.n

.. ~()()\I.

.3O~.n.

RJ

lS~4~..

(

..

III

_CaRID. OF

U.H.r. 05C.

o.001p"

Cil

III...C.~I~O_--,

n --- 1

Q.I)AF'

'JA'

S.S Rg

0.&.,. ...

=r <:'14 ~

TEST . - -_ _ _1'-';.1

JAC.t<.

k

JI

1I0PF"

30t<4

ca

Ru L.A.MP RIO

"Il

2

~~

~6SNl

Y26SN7

if":i Ga .4. 2. '

,}

... . ~~ 1;

~

'.,.~~_'::d. ~.~~~, ., . ·.~~:. :i\:~(:

' -" f "

" '

..

QUENCH OSCILLATOR List of Components V2 ' , v 2 r,

6SN7

R3

30K~

R4

25K.n.

2 watt polentiometer

R5

22K..n

1% tolerance

R6

22K..s'\.

1.% tolerance

R7

10K$I.

RS

10K..n. .

potentiometer ) ) ganged together potenticmeter )

R
5.5K

RIO

3 watt lamp (non-linear resistance)

Rll

30K .n..

R12

~.n.

R13

22K

Cs

110 pf.

1%

tolerance

C9

110 pf.

1%

tolerance

CIO

O.l/"f.

CIl

1.1"'" f •

C12

O.5..;ttf.

C13

0.OO7~f.

014

0.5 ~f.

JI

test jack

double triode

.n.

...1'\..

l Ke./sec. or 2% cao be tolerated in the quench fraquency, which ia given by l/2ffRs C9 •

Rs consists of a fixad 22 K.n. resistor in series with a fraction of

an Ohmi te 10 K.n. potentiometer.

The maj or variation in Rs will be due to tem-

perature changes, whereas C9 has a zero temperature coefficient. change by more than -0.02% per Oc rise.

Rs will not

For a 200 C change from the normal opera-

ting tempe rature only 0.4% change would be observed in the frequency, which is well within the a.g.s. band-width. A maximum amplitude of 25 volts peak cao be fad into a 7K.o. load with very little distortion in the waveform.

Almost no distortion exists for an output

of less than .5 volts. 4.2

The u.h.f. detector and a.gaS. sIstem A schematic diagram of the circuit is shown in fig. 4.3. The output pulses from the u.h.f. oscillator are detected by a IN 54 A ger""'\

manium diode and fed into a tuned amplifier.

The tuned amplifier has a t6~id

and a capacitor C16 for the tuned circuit and R4 as a damping resistance to obtain a band-width of 2 Kc./sec.,

R4 (47K~) reduces the Q of the plate load

from 45 to Z7, thus decreasing the equi valent damping resistaoce t'rom 30.5 K.Jl. to 18.5 K..n...

The amplification at resonance is given by

w" '-, :;

G~

w~ L,

Q

Cf

The measured amplification was found to be 28 which is in close agreement with the ory.

It was necessary to decouple the tuned amplifier from the other

stages in order to reduce the feedback through the B + supply, which had originally reduced the gain of the tuned amplifier to 23 and sometimas aven maintained oscillations.

"

, ~". '

.

U.H.F: DE:TECTOR AND AUTOMAï\CGA1N STAS\L\Ze:.R

8+ 300V.

.~ C. 16 l 't
0. !J,lf'

c.n

~----~----~

Er.1'7 A

OUTPUT OF'

U. '"'. F'. OSOC..

U. H.'-. DE.TEC.'ïOR

E.F 37,...

EF'3'7A

,.UNEO AMPLlF'\ER

F'IRST

6SN'7

SECOND VOLïA<6L

VOL.T~GE

~MPL.'F\ER.

ÂMPLlFI!:.A

F'G.4.3 .

PHASE INvE:RTER

GRIO U.\oI.F.QSt

0.1»"

Cza

E»ALS

REC.",~\ER

~

GRIO OF'

urtF. ose.

The . tuœd amplifier ..as rolloœd by two stagee

ot volte.g8 ampl1ticat1oa,

each haviDg a gain of about 50 at a frequency ot 54 le./seo.

tir.t

'V01.~

plate load.

The gain ot the

amplifier was made var1able by having a po"tentiolD8ter tor \hl

The outplt ot the seoond voltage amplifier red 1nto the pLu. 1n-

verter .hoae output ilIlpeda.nce, at the plate, equalled R33 or 5.6 I..n , and at the oathode equalled the re01prooal ot the trazwoonductanoe or 2.. 30,.JL val• •

or

the oomponents

tor these aeotioœwe:re oho88n

obaraoterietios t~ the Er 'J1J.. and the (mr1.

54

rco./..o.

rran

•

'fbI

obae:nr1ng the ttt'be

A double diode "7 reotifia4 the

voltage at the plate and oathode ot "6 togiw out a negat1ve D.O.

oon'b'ol voltage tor truf grid of the u.h.t. oaol11ator. In ordsr to 8AU8ry the NYCJll.t orlter1on tor 8tabillty1n the te"baok loop, it 1fU naoellsary te in'troduoe a t.1.m lag wh10h would be greawl" than

br the

tw10e tl» 100p gain, mult1pl1ed

whole loop.

81Dl or the phase !age and delay.

This _. the reuon for the large time oonstant (0.05 . . . ) of

R38 and 029 , and the slow re.pon.. ot the lIIystem. in 8ignal amplitude at the input

large Pllse for a fraotion ot

ot

the a.g.8. 1fas POl"

quirement

ot til»

li.

or

B.r

maJd ng a rapid l M " "

the reoelwr, 1t Was possible

to .. e a

second at the orystal IN 541 bafore the oontrol

r~etored.

the u.h.t. ol!lcillator

to al..eys operate in

tœ

on the a.g.s. ls simply to ha," .nough gain

input a1gnal the osoillationa do not reaoh sa.turation. the lnput signal 18 decreaeed;

tœ

11n.eel' mode, the re-

.e. that for the large.t It ie obviou. that, U

outPlt pulsee oan Dot lnorease and 210 tœ

oscl11ator can never f'ind itselt in the loger1thmic mode.

4.3 Tm aster oircuit 1 clrcuit ehowing the met3r connections 1111 civen in tig.4.4. double rectifier; being simi1a.r 1n operation ta V ,

7

voltage 1.natead.

rut

Vs

18 a

giv1.ng out a poaitiw

For po si tion 3 of the two-pole six-poai tion swi tch,

ODi

: ~~rf': ',.··.· ,,~~~~~;t.. . "" t'},

~ :

' ' ;'<:,:~~' ,: .~ -

,:~.

~

/.,;.,

1.' •

'1.

C,:,' .~ .

,;

'* :

~

THE: ?\..ATt

Sl~t-IAL.. 5iRtNGiH. INOiCATOR 3

•

4

Me:T~R

'-,

.

CIRCUlT

or &8'76

0·$00\1.

R.'.c:..S. CURREN'i IN

o-so

M.A.

587'-

H.T. 'IIOL..TAC.E.

.~

s·o-SOO\!-

3~

I.$.A a~

ON

O·~S".

!ta..,.

en •

.1

TO

C.~8-

=:;1

or s&'T6

.... .....

~

6eME.T&R OUT OF

-...l

C.I~C.UI T .3!~A.

TO

~TOGRIt:>

,. . . .

Zll<.n.

~

~"

6ALS

S:-\G.4.4

.. --~---

,' .

I{f

terminal of the mater 18 connected to the output of

Vs

and the other 1. b&ebd

off a variable D.C. voltage with Bi" as It8 souree, as ehawn in tige 4.4.

50 K.Q. potenticaeter ie ueed as the

~ro

centerlng ot the meter.

fi»

An exterDlll

meter cm be connected to the jack below the meter, i f the panel .. ter le DOt sutf'ioie ntly &Ccur ate.

All the other positions of the Ind toh are qui ta simple. For poei UODa l, 5 and 6 tbrJ mater bas a large resistan.ae in Mrie. wi th i t and aota lik. a YOl't_ter.

For poeitionil,it ha.e a 1.S..n.. Bmmt, thu,. acting lib a milllaDIMter.

The eT-ternal meter oan ha ueed wi th poe! tion l, 4 or 6 if it ha., a 5'!t ap.

._tull

moye_nt and poetion 2 if it bas a resleteDoe ot 1.S I..n.., and still haw'tJ»:.:o scale dètlecUon ail the panel meter.

5.

CONSTRUCTION OF THE RECElVER

5.1 The layout and operation of the receiver Three views of the receiver are shown in fig. 5.1, 5.2 and 5.3.

AlI the

controls are on the front panel wi th only two test jacks at the raar of tha chassis, as S3en in fig. 5.2. served on the oscilloscope b,y Pulses".

The wavaform of the datected pulses can be obco~necting

to the jack labeled ttOetected Output

The output from the other jack gives the quench voltage, which is fad

to the grid of the 5876.

It enables one to observe the waveform of the quench

frequency, measure i t9 amplitude and frequency, as weIl as to use i t as an external trigger for an oscilloscope if one should try to observe the oscillator output pulses. The two controls on the front panel,for the frequency and amplitude of the quench voltage, are shielded te minimize frequency drift of the quench oscillatore

The a.g.s. gain control is not provided with a

kno~

since,after it is set

and locked, it does not need resetting very frequently. The cathode and plate cavities are tuned by means of the two knobs on the front panel and can aIso be locked if desired, since the tuning shafts are provided with locking nuts.

The other controls have already been mentioned in

para. 4.4.

5.2.

The super-regenerative receiver as a "Bridge" detector A view of the equipment necessary for making measurements with the admit-

tance meter (u.h.f. "Bridge") is shown in fig. 5.4.

The signal generator on

the right provides the u.h.f. power for the "Bridge", while the receiver at the centr~with

its

po\~r

supply on the laft, acts as a datector.

The advantage of

this type of detector i9 the fact that,at such high frequencies, the super-regenerative principle is more efficient tham the superheterodyne principla.

3

-uJ

">

-

w u

llJ

cl

-

~ L-L J,

5"/

)

-> U)

v"-

ct ~

cL LU

:::> lU

u W

P'

y

.

01.

g

UJ ù

U)

~

-

lL.

0:>

'"'

llJ

II)

~

~ lU

~ lU

~'-) ct

lU

••

I"

l

~

-

A

...

(J/.

~ ~ ~

~

~

'"' ~

~

"

:t ~ ~

~

The suggestad method of detection is to m1x the unbalanced signal from the admittance meter with the output of a local osci1lator and detact the intermediata frequency with an amplifier.

The difficu1tJ lies in finding the desired

frequency !rom the beats of the harmonies. of approximate1y

T

This arrangement has a sensitivity

5,..u.. volts.

It doos not compare very we1l wi th a sensi ti vi ty of 1.5"P vol ta, which ia possible with the super-regenerative receiver.

The reradiating oscillations

of the super-regenerative receiver is of no disadvantage, sinee their presence is equally felt in all the arms of the admittance meter and doos not affect the balance.

SUMMARY AND

CONCLUSIONS

A super-regenerative receiver was designed, and found to be feasible, for operation in a frequency band from 455 MC./sec. to 510 Mc./sec. a- was obta1ned, graphically, to be

105~

The value of

mhos. /" sec., while from the formula

it turned out to he 114/"mhos./sec. The super-regenerative band-width was measured to be 630 Kc./sec. under normal operating conditions at a general frequency of 54 Kc./sec. and a quench amplitude of 5 volts peak.

This result of the band-width was in excellent agree-

ment with the theoretical value of 660 Kc./sec. The noise figure was determined to be 20 db., of which Il db. was calculated to be due to the super-regenerative principle, thus leaving 9 db. for the 5876 triode, which is what one would expect from a first class conventional circuit. The receiver was able to detect 1.5fo volts of input signal and hence easily fUlfilling the sensitivity requirements for its use as a "bridge" detector. The receiver is also sufficiently sensitive field strength meter.

and stable to be used as a

At the "Signal Strength Indicator" position of the switàh

on the panel, the meter has an approximately logarithmic response. be

It would

possible to get signal strength directly in db. when the meter ls calibrated.

APPENDIX l

The Power Supply for the Heceiver The pwwer supply should have an internal. resistance of less than 50 J\.. at D.C and at 54 Kc ./sec.

In the power supply that was us ed,

there were two filament transformers.

One of these with its centre

top grounded was used for all the tubes whose cathodes were near ground The other was used for the fila.ment of the 6SN7 phase

potential.

inverter whose cathode was at a mean potential of fi) volts and a possible peak potential of 120 volts. The power supply is connected to the receiver through a six The places on the power supply

conductor cable with an octal socket.

to which the wire are connected can be known by observing the colours of the wires.

The connections to the octal socket are numbered below

together with the colour of the wire going to each connector.

1.

ground

blue

2.

general filaments

black

3.

B+ 300V

white

4.

not used

5.

V6 filament

brown

6.

V6 filament

green

7.

general filaments

red

8.

not used

57

REFERENCES

Armstrong, E. H., Some recent developments in regenerative circuits. Proc. Inst. h.a.dio Engrs., N.Y. 10, 244 (Aug. 1922). Ataka, H., On super-regeneration of an ultra-short wave receiver. Proc. Inst. Radio Engrs., N.Y. 23, 481 (Aug. 1935). Schroggie, M. G., The super-regenerative receiver. 13, 581 (Nov. 1936).

.iirele as Engr.,

Frink, F.W., The basic principles of super-regenerative reception. Proc. Inst.Radio Engrs., N.Y. 26, 76 (Jan. 1936). Lewis, 'tI. B. and }ülner, C. J., A portable duplex radio telephone. wireless Engr., 13, 475 (Sep. 19)6). Becker, S. and Leeds, L. M., A modem two way radio system. Radio Engrs., N.Y. 24, 1183 (Sept.1936). Bradley, W. E., Super-regenerative detection theory. 96 (Sept. 1948).

Proc. Inst.

l!..1.ectronics 2l,

Hazeltine, A., Richman, D. and Loughlin, B. D., Supe r-regenera ti ve design. Electronics p. 99 (Sept. 1948). ' . " Macfarlane, G. G. and~itehead, J. R., The super-regenerative receiver in the linear mode. F .': Instn. Elect. Engrs., 93 Part III A, p. 284 (March-May 1946). Macfarlane, G. G. and 'W hitehead, J. R., The theory of the superregenerative receiver operated in the linear mode. F. Instn. Elect. Engrs., 95 Part III, p. 143 (May 1948). Glu ck sman, H. A., Super-regeneration; an analysis of the linear mode. Proc. Inst. Radio Engrs., N.Y. 37, 500 (May 1949). Whitehead, J. R. ~ Super-Hegenerative ft.eceivers, Cambridge University Press (1950). i'lhcùer, H. A., Wheeler l{onographs, Volume l, Numbers 3 and 7, Wheeler Laboratories Great Neck, New York (1953).