Geotech Page 1

ET

I: H

ow T

o B

uild

Gol

d &

Tre

asur

e D

etec

tors

, 198

1 —

Cop

yrig

ht

Mur

ray/

Mod

ern

Mag

azin

es, r

epro

duce

for

per

sona

l use

onl

y

“GOLD FEVER,” shrieked the newsheadlines following the finding of the27 kg Hand of Faith nugget at Wedder-burn in Victoria recently. It wasunearthed by a couple of amateur fos-sickers using a metal detector, justabout the most sophisticated tool everbrought to bear in the hunt for gold.

Designs for metal detectors genu-inely able to discriminate between‘trash’ and ‘treasure’ have generallybeen well kept trade secrets. Even thegeneral principles of operation havebeen veiled in mystery. However, weare indebted to Lee Allen of AltekInstruments of the UK for providing uswith the circuit design of this metaldetector project via our British edition.The design incorporates all the featuresand refinements of modern commer-cially-made instruments and featuresperformance equivalent to units costingtwo to three times as much.

Principles of operationThis detector employs the basically

well-known induction balance tech-nique to detect the presence of ametallic ‘target’ in the ground, but

includes a number of refinements whichrespond to certain characteristics of thetarget. The ‘search head’ contains twocoils: an outer coil which is connectedto a low frequency oscillator operatingsomewhere in the range 15 - 20 kHz,and an inner coil which is placed so thatit is only very loosely coupled to the

the coupling between the transmit andreceive coils will be increased. Thisgenerally produces an increase in thesignal from the receive coil. In simpleinduction balance detectors, such as theETI-549 (May 1977), this signalincrease is detected and used to gate anaudio oscillator on so that a tone is

A �discriminating� metal detectorThis metal detector operates just like the bought onesbut costs only one-third to one-half as much to build ityourself. It features three discriminate ranges plus VLFoperation and includes an auto-tune button.

design: Lee Allen, Altek Instruments, UK

article: Phil Wait

outer coil. The latter is connected to the‘receiver’ input of the instrument. Beingonly very loosely coupled, the signalinduced in the receiver (inner) coil fromthe transmit (outer) coil is very smallwhen a target is not in the vicinity of thesearch head.

When the search head approaches ametallic target, the target will have anumber of influences on the two coils.Firstly, the magnetic field pattern of thetransmit coil will be disturbed, and thus

Geotech Page 2

ET

I: H

ow T

o B

uild

Gol

d &

Tre

asur

e D

etec

tors

, 198

1 —

Cop

yrig

ht

Mur

ray/

Mod

ern

Mag

azin

es, r

epro

duce

for

per

sona

l use

onl

y

passed to a speaker or headphones.That’s all quite straightforward, but

there are other influences to be takeninto account. The ground in which a tar-get is buried can have quite a profoundeffect on the coils in the search head.Firstly, if the ground is basically non-conducting, then it will have a perme-ability considerably different to that ofair. This will affect the couplingbetween the two coils in the searchhead, increasing the coupling if thetransmit and receive coils are initiallyset up away from the influence of theground. You can compensate for thiseffect by physically varying the posi-tion of one coil in relation to the otherwhen the search head is near theground. However, different soils willhave different compositions and thushave different values of permeability —even within quite a small area. The bestway to compensate is by electronicmeans and well go into that shortly.

If the soil contains an appreciableamount of iron minerals (maghemite,hematite etc... often referred to as “ironstone soils”), or mineral salts of onetype or another, then it will be partlyconducting.

Such soils will have a permeabilityoften greater than basically nonconduct-ing soils, affecting the couplingbetween the coils in the search head in asimilar way to that just explained.Again, as the composition of the soilsvaries, so will the coupling. Another

effect is that of ‘eddy currents’ inducedin the conductive soil. The ac magneticfield of the transmit coil will induce acurrent in the ground beneath the searchhead and the eddy current has an effectopposing the permeability effect of thesoil — and the whole effect varies in acomplex and unpredictable way as yousweep the search head over the ground.

The only way to compensate for thesevarying, and generally unpredictableeffects, is to devise circuitry that ‘recog-nises’ the effect.

Permeability effects will vary thephase as well as the amplitude of thesignal coupled into the receive coil fromthe transmit coil while eddy currenteffects vary the amplitude. Knowingthat, one can devise appropriate cir-cuitry to take the effects into account.

However, we need to know how ametallic target affects the phase andamplitude of the receive signal. If thetarget is ferrous, it will have a muchgreater effect on the magnetic field ofthe transmit coil than will the surround-ing soil as its permeability is greater andit will ‘bend’ or concentrate the fieldlines to a much greater degree. If thetarget is non-ferrous it will have a per-meability effect opposite to that offerrous targets, deflecting the field lines,but eddy currents also have someinfluence.

The eddy current effect in a targetdepends on the electrical and physicalcharacteristics of the target. Metals

which are good conductors will havegreater induced eddy currents than met-als which have a higher resistivity. it isa fortunate accident of nature that goldand silver are good conductors (lowresistivity) while iron (especially if itsoxidised or rusty) is not so good aconductor.

If the target is ring-shaped then theeddy current effect is enhanced,whereas if its a broken ring or just apeculiarly-shaped mass, the eddy cur-rent effect is less pronounced. The‘attitude’ or orientation of the targetwill also affect the eddy current effects.If the main plane of the target object isaligned such that the field lines from thetransmit coil cut it at right angles, thenthe eddy currents induced will be at amaximum. If the main plane of the tar-get is aligned parallel to the transmitfield then the eddy currents induced willbe at a minimum. Obviously, the atti-tude of the target with respect to thetransmit coils field will vary as the headpasses over it and the eddy currenteffect will vary accordingly — it maynot be maximum beneath the centre ofthe search head.

The permeability and eddy currenteffects combine in the receive coil andthe signal varies in phase and ampli-tude in characteristic ways.

The instrumentThe best way to understand how this

instrument operates is to look at it in

Geotech Page 3

ET

I: H

ow T

o B

uild

Gol

d &

Tre

asur

e D

etec

tors

, 198

1 —

Cop

yrig

ht

Mur

ray/

Mod

ern

Mag

azin

es, r

epro

duce

for

per

sona

l use

onl

yblock diagram form. The accompany-ing diagram shows the basic circuitblocks employed. The transmit oscilla-tor drives the transmit coil in the searchhead and supplies a signal to two phasecontrol circuit blocks. The signal fromthe receive head is first amplified andthen ac-coupled to the input of a gate(gate 1). This gate is controlled by theoutput from one or other of the phasecontrol circuits via a block which‘squares up’ and inverts the signal. Theoutput gate consists of an ac signalsuperimposed on a dc level. This passesto an integrator which obtains the aver-age dc level of the composite signal.This is then passed to both a dc ampli-fier which drives a centre-zero meter,and to a ‘sample and hold’ circuit. Theoutput of this block provides a dc levelto the input of gate 1 which is a mea-sure of the average dc level of thecomposite signal. The initial dc levelapplied to the input of gate 1 is actuallyestablished by the tune control. Thus, adc negative feedback path is provided.

In addition to meter indication, anaudio indication is provided. The out-put of the dc amplifier driving the metercontrols a gate which switches on or offthe output of an audio oscillator. This isapplied to an audio amplifier and an on-board loudspeaker or headphones.

Power for the audio amplifier is pro-vided by two 9 V batteries in parallel.The rest of the circuitry requires twosupply rails at +10 V and +5 V withrespect to the common rail (0 V). Thisis supplied by a regulator from an 18 Vsource consisting of two 9 V batteriesconnected in series.

Initially, the instrument is set up inthe ‘VLF’ mode. The search head isheld in the air and the tune controladjusted to bring the meter to centrezero. This is done with the tune mem-ory button depressed. This activates thesample and hold circuit, storing the dclevel set by the feedback loop in thecapacitor of the sample and hold block.Thus, a particular dc level at the outputof the integrator corresponding to metercentre zero is set up.

The search head is then lowered tothe ground. Naturally, this will upset thecoupling between the transmit andreceive coils and the output at gate 1will change. This will change the dc

Page 4

ET

I: H

ow T

o B

uild

Gol

d &

Tre

asur

e D

etec

tors

, 198

1 —

Cop

yrig

ht

Mur

ray/

Mod

ern

Mag

azin

es, r

epro

duce

for

per

sona

l use

onl

y

Geotech Page 5

ET

I: H

ow T

o B

uild

Gol

d &

Tre

asur

e D

etec

tors

, 198

1 —

Cop

yrig

ht

Mur

ray/

Mod

ern

Mag

azin

es, r

epro

duce

for

per

sona

l use

onl

y

As

the

gene

ral

prin

cipl

es o

f op

erat

ion

have

been

dis

cuss

ed w

ith r

egar

d to

the

blo

ck d

ia-

gram

in t

he t

ext,

this

des

crip

tion

is c

onfin

ed t

oth

e C

ircui

t alo

ne.

Com

men

cing

w

ith

the

tran

smitt

er,

Q3

isco

nfig

ured

as

a C

olpi

tts O

scill

ator

, th

e tr

ans-

mit

coil

in

the

sear

ch

head

fo

rmin

g th

ein

duct

ance

whi

ch r

eson

ates

with

the

com

bina

-tio

n C

9 an

d C

10.

Bia

s is

app

lied

to Q

3 vi

aR

18,

D1

and

R19

. E

mitt

er b

ias

is p

rovi

ded

byR

17. T

he tr

ansm

it si

gnal

to th

e ph

ase-

lead

and

phas

e-la

g ci

rcui

try

(gro

und

and

disc

rimin

ate

cont

rols

) is

tap

ped

off

the

colle

ctor

of

Q3

via

C8

to th

e ju

nctio

n of

R11

and

R14

. The

gro

und

cont

rol c

ircui

try

conn

ects

via

R11

whi

le th

e di

s-cr

imin

ate

circ

uitr

y co

nnec

ts v

ia R

14.

The

sig

nal

from

the

rec

eive

coi

l is

ampl

ified

by Q

1 an

d ap

plie

d to

gat

e 1

(see

blo

ck d

ia-

gram

), o

ne C

MO

S g

ate

in I

C1

(IC

1b),

via

an

emitt

er-f

ollo

win

g bu

ffer

stag

e,

Q2.

Not

e th

atQ

1 is

ope

rate

d as

a g

roun

ded-

base

am

plifi

er.

The

pha

se o

f th

e re

ceiv

ed s

igna

l th

roug

h Q

1an

d Q

2 is

not

alte

red.

Out

put

from

the

em

itter

of Q

2 is

app

lied

to th

e dr

ain

of IC

1b.

The

bas

e of

IC

1b

is

driv

en b

y a

squa

rew

ave

deriv

ed

from

th

e tr

ansm

it si

gnal

, th

eph

ase

of w

hich

can

be

varie

d by

eith

er t

hegr

ound

or

disc

rimin

ate

cont

rols

.In

the

VLF

mod

e, t

he p

hase

of

the

tran

smit

sign

al t

appe

d of

f fr

om Q

3 ca

n be

var

ied

usin

gR

V1.

Thi

s pr

ovid

es a

pha

se-a

dvan

ced

sign

alth

at c

an b

e va

ried

over

the

ran

ge f

rom

abo

ut+1

0° t

o +

40°.

A l

eadi

ng p

hase

RC

net

wor

k is

form

ed b

y R

10 a

nd R

V1

in c

onju

nctio

n w

ithC

7. T

his

sign

al is

app

lied

to t

he in

vert

ing

inpu

tof

an

op-

amp,

lC

4b.

As

this

is

oper

ated

at

max

imum

gai

n w

ith a

hig

h si

gnal

lev

el a

t th

ein

put,

it w

ilI ‘s

quar

e up

’ the

sig

nal a

t its

out

put

(pin

7),

whi

ch d

rives

the

gate

of I

C1b

.In

the

dis

crim

inat

e m

ode,

sw

itch

SW

1 co

n-ne

cts

the

tran

smit

sign

al

to

circ

uitr

y w

hich

prov

ides

a l

aggi

ng p

hase

sig

nal

whi

ch c

an b

eva

ried

over

a r

ange

set

by

RV

2 (t

he d

iscr

imi-

nate

con

trol

) an

d a

set

of

‘rang

e’ r

esis

tors

:R

12,

R13

, R

15 a

nd R

16.

The

se f

orm

a l

ag-

ging

ph

ase

RC

ne

twor

k in

co

njun

ctio

n w

ithC

11.

The

sig

nal

is t

hen

buffe

red

by a

non

-

inve

rtin

g op

-am

p,

lC3b

, an

d ap

plie

d to

th

ein

vert

ing

inpu

t of l

C4b

via

SW

1c.

The

sou

rce

of I

C1b

is c

onne

cted

to

an in

te-

grat

or s

tage

for

med

aro

und

lC3d

. T

he o

utpu

tof

thi

s st

age

is c

onne

cted

dire

ctly

to

the

sens

i-tiv

ity

cont

rol,

RV

3.

The

w

iper

of

th

ispo

tent

iom

eter

goe

s di

rect

ly to

the

inpu

t of

a d

cam

plifi

er,

lC3a

, to

w

hich

w

e sh

all

retu

rnsh

ortly

. T

he w

iper

of

RV

3 is

als

o co

nnec

ted

toth

e sa

mpl

e an

d ho

ld c

ircui

t, vi

a R

25,

whi

chin

volv

es I

C5,

IC

1a,

the

tune

con

trol

RV

4 an

dth

e tu

ne m

emor

y pu

shbu

tton,

PB

1.T

he s

ampl

e an

d ho

ld c

ircui

t wor

ks in

the

fol-

low

ing

way

. T

he ju

nctio

n of

res

isto

rs R

25,

R26

and

R32

will

be

at a

dc

leve

l det

erm

ined

by

the

dc le

vel a

t th

e w

iper

of R

V3

and

the

dc le

vel a

tth

e w

iper

of

RV

4, t

he t

une

cont

rol p

oten

tiom

e-te

r.

The

dc

le

vel

at

the

wip

er

of

RV

3 w

illde

pend

on

the

sign

al le

vel a

nd p

hase

sw

itche

dth

roug

h to

the

int

egra

tor

by I

C1b

. W

hen

the

tune

m

emor

y pu

shbu

tton

PB

1,

is

pres

sed,

IC1a

(al

so a

CM

OS

sw

itch)

will

app

ly a

dc

leve

l to

the

inp

ut o

f th

e sa

mpl

e an

d ho

ld c

ir-cu

it pr

opor

tiona

l to

the

dc le

vel a

t th

e ju

nctio

nsof

R25

, R

26 a

nd R

32.

Thi

s w

ill c

harg

e C

13an

d th

e ou

tput

of

IC5

will

set

tle a

t th

is v

alue

.T

his

dc l

evel

is

then

app

lied

to t

he d

rain

of

IC1b

, via

R30

.T

hus,

the

rec

eive

d si

gnal

and

thi

s dc

lev

elar

e ‘m

ixed

’ at

the

inp

ut t

o ga

te 1

(i.e

.: IC

1b),

the

com

posi

te

sign

al

bein

g ap

plie

d to

th

ein

tegr

ator

.T

he m

eter

, M

1, i

s dr

iven

by

a dc

am

plifi

er,

IC3a

. T

he in

put

to t

his

op-a

mp

com

es fr

om t

hese

nsiti

vity

con

trol

and

is

appl

ied

to t

he n

on-

inve

rtin

g in

put

(pin

3).

Thi

s st

age

has

a ga

in o

fab

out 3

0 an

d a

little

‘sm

ooth

ing’

(in

tegr

atio

n) o

fth

e si

gnal

is

appl

ied

arou

nd t

he f

eedb

ack

byha

ving

a c

apac

itor

(C14

) co

nnec

ted

in p

aral

lel

with

the

feed

back

res

isto

r, R

36.

Apa

rt f

rom

driv

ing

the

met

er,

the

outp

ut o

fIC

3a i

s fe

d to

the

sou

rce

of l

C2b

whi

ch g

ates

the

audi

o os

cilla

tor

thro

ugh

to t

he a

udio

out

-pu

t st

age

(i.e.

: ga

te 2

). T

he d

c le

vel f

rom

pin

1of

lC3a

goe

s vi

a D

2 an

d R

39 to

pin

11

of IC

2b.

A p

ositi

ve b

ias

is a

pplie

d to

the

cat

hode

of

D2

from

the

+5

V r

ail

via

R38

. O

nly

whe

n th

e dc

leve

l at

the

out

put

of l

C3a

goe

s hi

gher

tha

n0.

6V a

bove

the

bia

s ap

plie

d to

the

cat

hode

of

D2,

will

lC2b

be

bias

ed o

n.T

he a

udio

osc

illat

or i

nvol

ves

lC4a

, co

nfig

-ur

ed a

s an

ast

able

mul

tivib

rato

r op

erat

ing

at a

few

hu

ndre

d H

ertz

. T

he

outp

ut,

pin

1,

isap

plie

d to

the

gat

e of

lC

2b.

Whe

n lC

2b t

urns

on,

the

sign

al i

s ap

plie

d to

the

inp

ut o

f th

eau

dio

outp

ut s

tage

.O

ne g

ate

from

IC

2 is

bia

sed

into

its

lin

ear

regi

on a

nd a

cts

as a

sou

rce-

follo

wer

buf

fer

atth

e in

put

of t

he a

udio

out

put

stag

e. T

he v

ol-

ume

cont

rol,

RV

5, is

the

sour

ce r

esis

tor

for

this

stag

e an

d th

e ou

tput

is t

aken

from

the

wip

er o

fR

V5

to t

he b

ase

of Q

4, c

apac

itive

ly c

oupl

edvi

a C

18.

The

out

put

stag

e is

a s

impl

e co

mpl

emen

tary

clas

s-B

st

age

empl

oyin

g a

low

po

wer

NP

N/

PN

P t

rans

isto

r pa

ir. T

he c

olle

ctor

of

Q4

driv

esth

e ou

tput

sta

ge,

its c

olle

ctor

loa

d al

so p

rovi

d-in

g bi

as t

o th

e ou

tput

pai

r (R

48).

Bot

h dc

and

ac f

eedb

ack

is a

pplie

d to

the

bas

e of

Q4

byR

49 f

rom

the

out

put.

Aud

io o

utpu

t ca

n be

fro

man

8 o

hm s

peak

er o

r he

adph

ones

, via

a d

c is

o-la

ting

capa

cito

r,

C20

. H

eadp

hone

vo

lum

e is

redu

ced

by a

470

ohm

res

isto

r, R

50,

in s

erie

sw

ith o

ne le

ad to

the

head

phon

e so

cket

, SK

2.P

ower

sup

ply

for

the

circ

uitr

y is

spl

it in

to tw

opa

rts.

The

aud

io o

utpu

t st

age

is s

uppl

ied

bytw

o 9

V b

atte

ries

conn

ecte

d in

par

alle

l (B

1 an

dB

2).

The

se a

re c

onne

cted

via

a r

ever

se-p

olar

-ity

pro

tect

ion

diod

e, D

3, a

nd o

ne p

ole

of S

W2

whi

ch is

a s

witc

h on

RV

5.T

he r

est o

f the

circ

uitr

y re

quire

s a

+10V

and

a +

5V r

ail,

with

res

pect

to

the

com

mon

rai

l(0

V).

Thi

s is

der

ived

fro

m t

wo

9V b

atte

ries,

B3

and

B4,

con

nect

ed i

n se

ries

and

appl

ied

to a

regu

lato

r ci

rcui

t vi

a a

reve

rse-

pola

rity

prot

ec-

tion

diod

e, D

4, a

nd th

e ot

her

pole

of S

W2.

The

regu

lato

r is

ba

sica

lly

a co

nven

tiona

l se

ries-

pass

circ

uit,

Q9

bein

g th

e re

gula

tor

tran

sist

or.

The

zen

er d

iode

ZD

1 pr

ovid

es a

sta

ble

ref-

eren

ce v

olta

ge f

or a

diff

eren

tial

pair,

Q7

and

Q8,

the

lat

ter

cont

rolli

ng t

he b

ase

curr

ent

toQ

9. R

esis

tor

R54

allo

ws

a sm

all

amou

nt o

fcu

rren

t to

pa

ss

to

Q7/

Q8

at

switc

h-on

to

ensu

re

the

regu

lato

r ‘s

tart

s’

corr

ectly

. T

he

base

of

Q7

is b

iase

d at

hal

f th

e up

per

supp

lyra

il vo

ltage

by

R51

and

R52

. T

his

volta

ge i

sbu

ffere

d by

an

op-a

mp,

IC3c

, co

nfig

ured

as

avo

ltage

fol

low

er,

and

used

to

driv

e th

e +

5Vlin

e. D

ecou

plin

g is

pro

vide

d by

C19

, C21

and

C22

. N

ote

that

‘ba

ttery

tes

t’ fa

cilit

ies

are

pro-

vide

d by

R44

/PB

2 fo

r th

e 18

V s

uppl

y an

d by

R45

/PB

3 fo

r th

e 9

V s

uppl

y.T

UN

ING

Dur

ing

the

tuni

ng o

pera

tion,

whe

n th

e in

stru

-m

ent

is b

eing

initi

ally

set

up,

the

circ

uit

wor

ksin

the

fol

low

ing

way

: W

ith t

he t

une

mem

ory

push

butto

n op

erat

ed,

IC1a

is g

ated

on

and

adc

neg

ativ

e fe

edba

ck lo

op is

est

ablis

hed

from

the

outp

ut o

f th

e se

nsiti

vity

con

trol

, ba

ck t

oth

e in

put

of g

ate

1, t

he d

rain

of

IC1b

, vi

a th

esa

mpl

e an

d ho

ld c

ircui

try.

A p

ortio

n of

the

volta

ge fr

om th

e w

iper

of R

V3

is a

dded

to

the

volta

ge

dete

rmin

ed

by

the

volta

ge

divi

der

R33

, R

V4

and

R34

. T

his

is a

pplie

d to

the

sour

ce o

f IC

1a. A

s th

e tu

ne m

emor

y bu

tton

ispr

esse

d,

IC1a

is

co

nduc

ting

and

capa

cito

rC

13 w

ill c

harg

e to

the

val

ue o

f th

e co

mpo

site

volta

ge a

pplie

d to

the

sou

rce

of I

C1a

. T

heop

-am

p IC

5 is

a lo

w in

put

curr

ent

devi

ce a

ndth

e ou

tput

, pin

6, w

ill s

ettle

at a

val

ue e

qual

toth

e co

mpo

site

vo

ltage

ap

plie

d to

its

no

n-in

vert

ing

inpu

t (p

in 2

). T

his

dc le

vel i

s ap

plie

dto

the

sou

rce

of I

C1b

and

the

sig

nal

outp

utfr

om t

he r

ecei

ve c

oil a

mpl

ifier

is m

ixed

with

it.

Thi

s w

ill b

ring

abou

t a

redu

ctio

n in

the

dc

leve

l of

the

sig

nal

appl

ied

to t

he i

nteg

rato

rin

put,

and

thus

a r

educ

tion

in t

he d

c le

vel

atth

e in

put o

f IC

1a a

nd, w

ithin

a s

econ

d or

tw

o,a

new

dc

cond

ition

is

esta

blis

hed.

Whe

n th

edc

leve

l aro

und

the

loop

set

tles,

the

met

er w

illre

ad

zero

(c

entr

e)

and

the

tune

m

emor

ysw

itch

is r

elea

sed.

The

dc

leve

l at

the

outp

utof

IC

5 (a

nd th

us a

t the

dra

in o

f IC

1b)

is m

ain-

tain

ed b

y th

e ch

arge

on

capa

cito

r C

13.

Inpr

actic

e, it

will

drif

t ver

y sl

owly

, as

C13

will

be

grad

ually

dis

char

ged

by t

he i

nput

cur

rent

of

IC5

and

the

capa

cito

r’s o

wn

leak

age.

For

thi

sre

ason

, th

e tu

ne m

emor

y bu

tton

is lo

cate

d on

the

croo

k of

the

han

dle

whe

re it

can

be

oper

-at

ed b

y yo

ur th

umb

ever

y no

w a

nd th

en to

re-

cent

re th

e m

eter

.

FE

AT

UR

ES

•

VL

F a

nd T

/R o

pera

tion

•T

hree

ran

ges

of ‘d

iscr

imin

ate

’ (T

/R)

oper

atio

n •

Can

tun

e ou

t alu

min

ium

rin

g-p

ull t

abs

•G

roun

d b

alan

ce c

ircu

itry

incl

ude

d•

Tun

e m

emor

y (‘a

uto

-tun

e’)

butto

n

•H

igh

sen

sitiv

ity (

will

det

ect

20c

pie

ce a

t dep

ths

over

250

mm

)•

Pre

-wou

nd a

nd a

ligne

d w

ater

proo

f sea

rch

head

•S

trai

ghtfo

rwar

d co

nstr

uctio

n, n

o a

lignm

ent

nece

ssar

y•

Low

bat

tery

dra

in•

Use

s co

mm

on N

o. 2

16 tr

ansi

stor

rad

io b

atte

ries

•C

osts

aro

und

$200

in k

it fo

rm

HO

W IT

WO

RK

S —

ET

I 150

0

Geotech Page 6

ET

I: H

ow T

o B

uild

Gol

d &

Tre

asur

e D

etec

tors

, 198

1 —

Cop

yrig

ht

Mur

ray/

Mod

ern

Mag

azin

es, r

epro

duce

for

per

sona

l use

onl

y

level at the output of the integrator. Theground balance control is then adjustedto bring the meter back to centre zero.What the ground balance circuit does isto provide a signal which leads thephase of the transmit signal and thusleads the phase of the signal induced inthe receive coil without the presence ofground. The ground balance controlvaries the phase of this signal over arange of about four to one. Thus, whenyou vary the ground balance control,this varies the phase of the signal con-trolling gate 1, thus varying the averagelevel of the signal passed to theintegrator.

The process is then repeated until nochange occurs when the search head islowered to the ground. This establishesa ‘normal’ condition for the output ofthe integrator and the sample and holdcircuit maintains the appropriate dclevel at the input to gate 1 such that themeter remains at centre zero.

If the search head then approaches ametallic object, the amplitude of the sig-nal in the receive coil will vary as thecoupling between the coils and thephase of the signal will be altered by the

Geotech Page 7

ET

I: H

ow T

o B

uild

Gol

d &

Tre

asur

e D

etec

tors

, 198

1 —

Cop

yrig

ht

Mur

ray/

Mod

ern

Mag

azin

es, r

epro

duce

for

per

sona

l use

onl

y

target. This will change the averagelevel of the composite signal out of gate1 and thus the dc level at the output ofthe integrator will change. This will beamplified and the meter will show anindication. Also, gate 2 will be oper-ated and a tone will be heard in thespeaker.

However, this method of operationwill not indicate the difference betweenthe characteristics of different targets.

In the discriminate or TR mode, theground balance control is not used. Theinstrument is initially set up using thetune control to bring the meter to cen-tre-zero. The different TR ranges permitvarying degrees of control with the dis-criminate potentiometer. The phase-lagcircuit block generates a signal whichlags the phase of the transmit signal andthe discriminate control provides aphase-variable signal to drive gate 1.

When a ferrous target is approached,the combined permeability and eddycurrent effects tend to reduce the ampli-tude of the signal picked up by thereceive coil. This will cause a reductionin the dc level of the signal out of gate 1and a reduction in the dc level out of the

integrator. Thus, the meter will move tothe negative (left hand) side of thescale. This side of the scale is marked“bad”, obviously.

When a non-ferrous target isapproached, the combined eddy currentand permeability effect tends to increasethe amplitude of the signal picked up bythe receive coil. This will cause anincrease in the dc level of the signal outof gate 1 and an increase in the dc levelout of the integrator. The meter willthus move toward the positive (righthand — “good”) end of the scale.

The effects we are considering areactually quite small, hence the circuithas a considerable amount of do gain.

Gate 2 only operates when the outputfrom the do amp increases (goes posi-tive) and thus the audio output is onlyheard in the discriminate mode whenthe meter shows “good”.

It is unfortunate that ring-pull tabsfrom drink cans are aluminium and thusindicate along with other nonferrousmetals. But, the discrimination ability ofthe instrument can be adjusted toexclude the small effect these targetsgenerate — along with small trinkets,

the smaller gold nuggets, etc — butwho wants the tiddlers anyway!

If the dc level applied to the input ofgate 1 drifts — and it may do for a widevariety of reasons, operating the tunememory button will restore the balanceof the circuit and re-centre the meter.Quite a cunning arrangement.

Search headThe most important properties of the

search head are its size, the relationshipbetween the transmit and receive coils,and the shielding against capacitiveeffects between the coils and theground. Surprisingly, the actual induc-tance of the coils is not of primaryimportance.

The greater the coil diameter thegreater the penetration depth but theless sensitive the detector will be tosmall objects. Penetration using simple,circular coils is about equal to thesearch coil diameter for small objectssuch as coins, while sensitivity isroughly proportional to the cube of theobject diameter (expressed as a func-tion of the search coil diameter).Sensitivity is also inversely propor-

Geotech Page 8

ET

I: H

ow T

o B

uild

Gol

d &

Tre

asur

e D

etec

tors

, 198

1 —

Cop

yrig

ht

Mur

ray/

Mod

ern

Mag

azin

es, r

epro

duce

for

per

sona

l use

onl

ytional to the sixth power of the distancebetween the coil and the object.

All this means that if the object size ishalved the sensitivity is reduced to one-eighth. If the depth is doubled the sensi-tivity is reduced to one sixty-fourth. Seewhy metal detectors designed to pick upsmall objects use small coils and reallyonly skim the surface? If the search coilis doubled in diameter for greater pene-tration the sensitivity to small objectsfalls to one eighth, apart from the coilassembly becoming mechanically lessrigid. The law of diminishing returnsagain or ‘you don’t get something fornothing’.

Our new detector improves penetra-tion while retaining sensitivity by usinga co-planar arrangement of coils in thesearch head which gives a slightly mag-nified field pattern downwards, into theground.

We mentioned earlier that the twocoils are only loosely coupled. Thepositioning of the receiver coil in rela-tion to the transmitter coil is verycritical and is the major factor affectingthe performance of the instrument. Infact. misplacement by a millimetre or sowill markedly affect the performance.

As the search head is moved around,

the changing capacitance between thecoils and the ground could completelymask the minute changes in the field weare looking for. To avoid this affect thecoils are enclosed in a Faraday shield.

By now it should be obvious that con-struction of the search head is not a taskto be tackled on the kitchen table on arainy Sunday afternoon. In fact, con-struction and alignment of the searchhead would be beyond most readers’resources (anyone who has attemptedour earlier induction balance metaldetector knows what it’s like). With thisin mind we chose to use the commer-cially built, pre-aligned search headmade by Altek Instruments. This will beavailable in Australia through All Elec-tronic Components in Melbourne whohave agreed to make the unit availablewholesale to other suppliers, as well asretailing parts themselves, along withhardware — the plastic extendable han-dle and the case for housing theelectronics.

ConstructionThe mechanical components for this

project — the search head, handle andcase, are available through Altek’s Aus-tralian agent, All ElectronicComponents, as mentioned earlier. Werecommend you obtain these as yourfinished instrument will then be a pro-fessional looking piece of equipment,with the features and operation of a‘bought one’ two to three times theprice. However, you can suit yourselfand make your own handle if you sodesire and we have designed the pcboard such that it will also fit in a largejiffy box. You will have to use thesearch head recommended though, forthe reasons we have explainedpreviously.

All the electronics mounts on a sin-gle, double-sided pc board. The Altekcase has two clamps on the rearenabling it to be clipped on to the han-dle. The Altek handle has two sections,the lower section sliding inside theupper section enabling the operator to

Geotech Page 9

ET

I: H

ow T

o B

uild

Gol

d &

Tre

asur

e D

etec

tors

, 198

1 —

Cop

yrig

ht

Mur

ray/

Mod

ern

Mag

azin

es, r

epro

duce

for

per

sona

l use

onl

yadjust the length of the handle to suithis height. Connection between thesearch head and the electronics is via alength of shielded cable (supplied withthe head) and a five-pin DIN plug/socket arrangement. The tune memorypushbutton is mounted in the end of the‘crook’ of the handle (see photographs)where it can be easily operated by thethumb. It connects to the electronics viaa length of shielded cable passedthrough the handle.

Construction should commence withthe pc board. As it is a double-sidedboard (i.e: copper tracks on each side),first identify the ‘front’ and ‘rear’ side.These are marked, respectively, ETI1500f and ETI 1500r. The rear side hasthe more complicated pattern of tracks.The components are mounted on thefront of the board, where there is theless complicated set of tracks. Note thatsome of the resistors, IC pins and pcboard pins (used for connecting exter-nal wiring to the board) must besoldered to copper tracks on each sideof the board.

Commence with the resistors. Takecare with those that cross tracks thatyou don’t create a short circuit whereit’s not wanted. Next mount the capaci-tors. Take care with the orientation ofthe electrolytics. Note that capacitorsC2, C7 and C11 are styroseal types,used for their good temperature stabil-ity. Be careful when soldering them inplace that you don’t overheat the leadsas this can cause melting of the capaci-tor’s case, possibly damaging it. Thesample and hold capacitor, C13, mustbe a low leakage type, preferably poly-carbonate or mylar. We used a greencapsuccessfully, but whatever you manageto obtain, make sure its a good qualitytype from a well-known supplier.

Now mount the semiconductors. Takecare with the orientation of these as youcan destroy devices if they are incor-rectly inserted when power is applied.Finally, solder the pc pins in place andthe four battery clips. The latter all goalong one edge of the board.

Overall assembly of the pc board isclear from the overlay picture on page17.

Once you have everything in place onthe pc board and you’re satisfied that allis OK, you can turn your attention to the

hardware. Start with the case thathouses the electronics. If not pre-drilled,you’ll need to mark out and drill all theholes in the case lid. The panel artworkcan be used as a template. Note that wedressed up the case with a Scotchcalpanel. These should be availablethrough the usual suppliers. Centrepunch holes before drilling. The cutoutfor the meter can be made with a holesaw or by drilling a series of 4 mmdiameter holes just inside the markededge of the hole. When you completethe circle, the centre piece can besnapped out and the edge of the holecleaned up with a half-round file untilthe meter drops in neatly.

The speaker is mounted on the lefthand side of the case lid (as you wouldnormally view the unit in use). It is heldin place by large washers placed underthe nuts of three bolts spaced around theouter rim of the speaker. Alternatively,you can glue it in place. Be careful notto get any glue on the speaker cone oryou might end up with a rather ‘stran-gled’ sound!

The pc board mounts in the bottompiece of the case, along with the DINsocket (for the search head connector)and two battery test push buttons. Asmall hole in the bottom passes thecable to the tune memory button. Thecase bottom has four integral mouldedstandoffs to provide support for the pcboard which is held in place withscrews.

Once all the mechanical work on thecase is satisfactory, the Scotchcal panel

may be stuck on. Take care when posi-tioning it as it’s almost impossible tomove if you misalign it. Carefullysmooth out all the bubbles toward theedge of the transfer.

The controls, meter etc. may bemounted next. Then you can wire all theexternal components to the pc boardpins. We used lengths of ribbon cablewhere possible to simplify the wiring.The easiest way to accomplish this partof the assembly is to place the bottomof the box, with the pc board mountedin it, on your left and the lid, with themeter and controls etc. mounted, facedown on your right. Follow the wiringdiagram on page 16 and complete allthe interconnections. You should nowappreciate pc board pins!

The tune memory button mounts in ahole in the end of the ‘crook’ of thehandle, as we explained earlier, and theshielded cable connecting to it passesthrough the handle, emerging through asmall hole drilled in the handle nearwhere the cable can enter the hole pro-vided for it in the bottom of the box.This cable is best inserted before youmount the pushbutton. Remove thehandgrip. Push the cable through thehole in the handle near the case, until itappears through the end of the handle.Solder the end of the lead to the push-button and mount the pushbutton in thehole in the end of the hand grip (easiersaid than done!). Put the handgrip backand you can pass the business end of thecable into the case and terminate it. Ifyou’re lucky, kit suppliers may sell theunits with this part already assembled.

Holding the batteries in place is gen-erally left to your ingenuity. We useddouble-sided sticky tape (ah, that’s use-ful stuff...). The battery life is quitegood as the circuit has been designedfor low current drain. Reverse polarityprotection is provided on the pc boardto avoid problems should you inadvert-ently attempt to connect a battery backto front.

When all wiring is complete, push thecase onto the handle and drill a smallhole through one of the clamps and thestem of the handle. Insert a nail or a boltand this will prevent the case fromrotating on the handle. Mount the searchhead and adjust the length of the stem tosuit yourself. Wrap the cable from the

Geotech Page 10

ET

I: H

ow T

o B

uild

Gol

d &

Tre

asur

e D

etec

tors

, 198

1 —

Cop

yrig

ht

Mur

ray/

Mod

ern

Mag

azin

es, r

epro

duce

for

per

sona

l use

onl

ysearch head around the stem so that it isheld quite rigidly and plug it into theDIN socket on the case.

You’re ready to roll!... once you’vetested it.

If you wish to make the search headcompletely waterproof, seal the holethrough which the cable passes withSilastic rubber or some similar caulk-ing compound.

OperationWhen construction is complete and

you’re satisfied all is well, turn thedetector on and advance the vo1umecontrol. Set all other controls tomidrange and switch the mode selectorto VLF. Hold the search head up in theair and well away from metal objects,press the tune memory button and rotatethe tune control. The meter shouldswing either to side of the centre posi-tion. Set the pointer to centre scale andrelease the tune button. The metershould remain at this position but maydrift slightly, which it will tend to doimmediately after switch on. Pressingthe tune memory button at any timeshould return the meter to centre posi-tion, set by the tune control.

The next step is to determine that thepolarity of the receive coil is correct.After tuning the detector as described,bring a piece of iron near the searchhead. If the meter swings to the rightyour circuit is correct, if it swings to theleft you will have to reverse the twowires on the DIN socket that connect tothe receiver section on the pc board.The meter should now swing to theright.

Ground balanceWith the detector tuned, lower the

search head to the ground. The metermay swing off scale. If it swings to theright turn the ground control to the left,if it swings to the left turn the groundcontrol to the right. Raise the searchhead from the ground, press the tunebutton and the meter will return to cen-tre scale. Lower the search head againand repeat the procedure until there islittle difference in the meter readingwhen the search head is lowered. Set-ting the ground control is quite criticaland may take some time to achieve thefirst time around. The detector can now

be used in the VLF mode.

Sensitivity controlThe sensitivity control sets the gain of

the dc amplifiers in the detector and willgenerally give best results at midrange.If the control is set fully clockwise thetuning will tend to drift, requiring morefrequent operation of the tune memorybutton.

Discriminate controlsThe mode switch selects one of three

discriminate ranges: TR1, TR2 or TR3,while a vernier action is provided by thediscriminate control. The discrimina-tion ability of this circuit is extremelyeffective and it is possible to discrimi-nate between an aluminium ring pull taband a gold ring. Remember that dis-crimination depends on the resistivity ofthe target object.

When set to TR1, discriminate con-trol at mid-range, the meter shouldshow ‘bad’ for ferrous objects and‘good’ for non-ferrous objects alongwith a tone from the speaker. As thediscrimination controls are advanced,some nonferrous objects such as brasswill start to give a ‘bad’ reading, whilegold and silver will give a ‘good’ read-ing. As the controls are advancedfurther aluminium will start to give a‘bad’ reading, and so on. As you use thedetector you will become familiar withits operation.

The best way of setting the discrimi-nation controls is to carry around a fewsample objects of the type you want todiscriminate against just for this pur-pose. One thing to remember is that acorroded object will require a differentsetting of the controls to a non-cor-roded one so carry samples typical ofwhat you are likely to dig up.

By careful setting of the controls,unwanted objects can be tuned out, giv-ing no meter movement at all so thedetector can be used to reject particularobjects and at the same time discrimi-nate between others.

Well, its now up to you. Remember,the secret of success in metal detectingis more knowing where to look than thetype of detector you have. There aremany books available on the subjectwhich could help put you on the righttrack.

Notes for constructors(courtesy of G.N. Vayro, Broadmead-ows, Victoria).

• Take special care with the orientationof IC5 (CA3130) if an 8-pin TO-5(circular metal case) type is supplied.Refer to the pinout diagram below.

• Take care with the wiring of theheadphone socket as not all types havethe same, or similar, connections.Check this by examination or with amulti- meter before wiring.

• Take care when wiring the DINsocket that connects the search head.The search head wiring is colour-coded,as shown on the circuit diagram. Thered and black wires come from thereceive coil. This coil has a dcresistance of around 50 ohms. Thetransmit coil is connected via the cableshield and the white wire. It has a dcresistance of around 12 ohms. Theremay be a yellow wire in the cable.Ignore it as it is not connected. TheFaraday coil shields are internallyconnected to the cable shield.

• The wiring to the two pushbuttonsPB2 and PB3 should first be sorted outwith an ohmmeter before soldering it inplace.

• The pushbutton in the handle needs tohave good ‘feel’ and positive contact.One of the small C & K or Swann typesshould fill the bill.

• If you have or are using a metal frontpanel, it should be earthed to reducespurious capacitive effects. The body ofthe discriminate control should connectto 0 V (Pin i) and a star washer shouldbe inserted under its nut to provide agood contact to the panel. Otherwise, aplastic Scotchcal panel is recommended(one was used on the prototype).

• It is strongly recommended that aflux-removing solvent be used to cleanthe pc board following assembly.Whilst flux does not cause problemswhen ‘new’, many atmosphericallybourne chemicals can and do react with

Geotech Page 11

ET

I: H

ow T

o B

uild

Gol

d &

Tre

asur

e D

etec

tors

, 198

1 —

Cop

yrig

ht

Mur

ray/

Mod

ern

Mag

azin

es, r

epro

duce

for

per

sona

l use

onl

y

Geotech Page 12

ET

I: H

ow T

o B

uild

Gol

d &

Tre

asur

e D

etec

tors

, 198

1 —

Cop

yrig

ht

Mur

ray/

Mod

ern

Mag

azin

es, r

epro

duce

for

per

sona

l use

onl

ythe flux in time. This causes a leakagepath to be established between thetracks and is especially troublesome inhigh impedance circuits, such as aroundIC5. A de-fluxed pc board will obviatelater (or early) problems with the auto-tune circuit; it also looks moreprofessional and aids identification ofdefective solder joints. The effort isworth it.

• If you have trouble with handcapacitance effects, plastic knobs or

collet knobs may be used to advantageon the controls, particularly the variablediscriminate control.

• The wiring to the pushbutton in thehandle should be done with shieldedcable, passed through a hole drilled inthe rear of the case to avoid fouling thetelescopic shaft in the retracted position.

• A battery clamp, fashioned from asmall strip of aluminium, is recommen-ded.

• The case should be mounted as closeto the curve in the handle as possible foroptimum weight distribution.

• A screw or bolt should be placedthrough the rear case mounting clip tostop the case rotating on the shaft. Therear clip is recommended to allow theshaft to be telescoped to minimumlength.