Contents:
0<
"
.
oQ 'Uz G')
, m
D
G)
o
m


x
Ito I D G')D 0
G) o
m
I
()
s:
I ]] <
m
OJ
s:
o o
m
r I D
o to o I 'U I
Z G')
'Z
9
�

6

K =
where
t
[ r
The coefficient of coupling
( 1)
A
r
=
diameter of loop, feet radial distance from antenna
=
taken along an axis of maximum coupling. k l
between transmitter and target will be attenuated further by sin
t/J,
where
t/J
is the target depression angle. x
Noting that
sin
t/J
=
[
B2+x2
J
1/2
and that
r3
=
[
B 2+x2
]
3/ 2
k will then be given by l
k
1
=
i (AD) 3/2
3
[
x
B2+x2
J
2
For a true 90 degree orientation of the receiving loop when centered above the target, maximum target to loop coupling exists, and the
coefficient of coupling k between target and receiver is given by 2
k 2
=
i
3
/ 1 (AD)3 2 3 x

7

The transm itter to receiver coupling via the target will be given by the product of k1 2 as given by k 16 3 Z x 1 Z
+
k1kZ
� 9"
(AD)
[
b
x
2
( 2)
J
(2)
since
The received voltage will be directly proportional to Equation
k k and the loop inductances will define the mutual inductance from l 2 transmitter to receiver via target.
For large x the receiver voltage for a given target falls off as 1/x6 just as in the induction balance and beat frequency locators. of depth follows a much less severe 1/x 2 law, For dis
tances small compared to B, the rate of signal drop off as a function allowing the detector The price paid
to retain good sensitivity to depths comparable to B. A s with all inductively coupled locators,
for this increase in detectability is a loss in resolution to small objects. the receiver signal increases
as the cube of the target diameter.
The optimum value of B is a function of the maximum desired depth of effective penetration contrasted against the minimum allowable resolution to small objects. 2 , A plot of the receiver voltage normalized and
to a given target of onefoot depth for various B and x appears as Figure plotted both in terms of decibels of received Signal, as ratios of target diam eters required to produce an equal strength signal. The loss of resolution as a function of B appears as Figure 3, plotted
as the ratio of diameters of targets required for equal detectability. The presence of a conductive target will not only increase the mutal coupling between the loops but will Simultaneously shift the phase of the received signal because of a resistive loss component introduced as the target lowers the
"Q
IT
of the coupling system. 4
Synchronous
detection can thus be employed to detect only the real component of the received signal, greatly adding to the detectability. If a
HJ.d3C
SA
NOIJ.:>3J.3C
.1.33.::1 . H J. d 3 C 9 9 00' 
08
09
Ov
0('; 
o
01.1."1:1 :::> 11:1.1. 311\1 " 1 a NOI.1.:::>3.1.30 ,,,n03
e
N 0 I n , 0 S :3 1::1
SI\
B
e �E>I::I
9
o 09
9' 099
017
oz;
o
e
'10
synchronous system is not used,
the limiting detectabil ity will be that
of a target whose mutual inductance equals the mutual inductance caused by straightthrough coupling, tortion caused by the circuitry. The mutual inductance caused by angular misalignment of the loops is given by MCos 8, where M is the mutual inductance between two parallel M is GroverTs5 ground return, and pattern dis
loops of diameter A and spacing B sharing a common axis.
obtained by making use of the extensive tabulations available. table 17 is directly applicable and is plotted as Figure 4. The induced receiver voltage will be given by
E
r
::
MwI
rms
( 3)
where I is the transm itting loop current assuming single turn rms transm itting and receiving loops. Since two mutual inductances are involved in target detection, the
induced receiver voltage due to a target increase as the square of the frequency usedo The highest possible operating frequency, consistant
wit4 terrestial absorption, sensitivity.
must be used to obtain maximum target
The model locator was designed with pipe location as a primary goal, requiring effective penetration to four feet of depth and better. value of B
=
A
4' and A
=
1 T was selected.
0'
9
9
'VIa
� '0�
'�0
J.O 0 :oj
1::::1 3 c:f
SAI::::IN3HONVN
,
j '
V
8
L,;"ti:i
, ,
12
CHOICE OF FREQUENCY
The effects of terrestial absorption and the locator detectability requirements interact to determine a suitable operating frequency.
\
by a factor of
Terrestial absorption will reduce Er further as a function of depth a.y where
a.
is the attenuation per foot in decibels and Total terrestial attenuation is
y is the round trip terrestial distance. then given by
Attenuation
=
a.
r, [ 2 +Lb 2 2J 1/ 2J
x
+
x
Decibels
(4)
A plot of the received E
is plotted as Figure 5 for B
r
that takes into accQunt various values of a.
=
4 feet.
The terrestial absorption will vary with frequency as well as the resistivity and dielectric constant of the earth at the point of penetra tion. Heiland6 has shown that the lowfrequency penetration equations and
are valid except for very high frequencies and soil resistivities,
that the results are always conservative for any frequency and con ductivity. One form of the lowfrequency equation is
d
: e..
=

60
Lf J
1/2
where d
=
p = resistivity in ohmcentimeters
1/
e:
penetration in meters
NO I.l. 'lfn N 3.l..l.'If
S/\
"
3
9 "ElI::l
9
o os
09
OV�
O 
o
14
as taken from Heiland's Figure
8.7 decibels.
1024.
At
l/E
attenuation,
the loss is
By changing dimensions,
the inverse relation
0.
=
0.045
Lp
0.
r
i 1/ 2
.J
i
( 5)
is obtained with
expressed in decibels/foot.
Values of resistivity for different media are taken from Heiland's text and plotted as a function of frequency in Figure 6. The curves indicate
an excessive amount of attenuation for most media at frequencies above
500
kHz.
favors the receiver sensitivity requirements and makes optimum use of available components. evaluation model. Neglecting the dielectric constant in the above analysis will have two opposing consequences. At the frequencies used, high values of k This is the operating frequency chosen for the
455
kHz is within a tolerable attenuation region and both
serve to decrease the terrestial absorption,
acting much as a bypass
capacitor to allow the high frequency energy to more freely travel through
a
lossy media.
At the same time,
large values of k introduce
a substantial mismatch between the loops in air and the earthair inter face, and can cause a ground reflection that could constitute a signifi Values of k encountered
cant portion of the total transmitted energy. can be extreme,
owing to water having a dielectric constant of
most soils having a moderate to substantial moisture content.
84
and
l
�z
H 11\1
'A:> N 3 n 03 1:::1.::1 J ro ,�0
OJ
0'
91
16
RECEIVER SENSITIVITY AND BANDWIDTH
The receiver sensitivity may be intrepreted in term s of that signal induced in the receiving antenna for a onedegree angular misalignment
'.
of the receiving loop from true null. Resonant loops are used for both receiver and transm itter to allow multiplication of transm itter current and receiver voltage. all metal used in the locator at neutral potential, employed. Several techniques were tried for the loop assemblies, the most practi
Q
. To place
balanced loops are
cal of which was a 10 inch by 12 inch oval loop of four conductor number 18 A WG communication cable, rewired by sequential connection The unloaded
a
to produce a four turn, center tapped configuration. loops gave
Q! s of 65, which dropped to 30 when driving
2000 ohm load.
The inductance is 18 microhenries which requires 6800 picofarads of capacity to resonate at 455 kHz. To allow for final tuning, a 700
picofarad trim mer is placed in parallel with 6500 picofarads of fixed capacity. Identical geometry is employed for both loops.
The received signal due to straightthrough coupling is given by E t  jMN N QtQ w _ . r r t r p
E
r
=
Sin 8
( 6)
where N r Nt number of receiving loop turns number of transmitting loop turns loaded receiver loop
= = = =
Qr
Qt r p
Q Q
Loaded transm itter loop and oscillator
=
equivalent parallel resistance of transm itting loop
Et
=
RMS voltage across transmitting loop.
17
For a onevolt RMS E Q r
=
t
across 1000 ohms and values of Q t
=
15 and
30,
a receiver sensitivity requirement of 73.5 microvolts results.
A full output from 100 microvolts of input may then be defined as a receiver design objective.
The receiver bandwidth is not overly critical,
and is chosen to be
narrow, while still allOWing fixed tuned RF stages using ordinary �10 percent capacitors. A 70 kHz, 15 peJ::cent bandwidth may be
selected using this criterion.
The overall selectivity and noise rejec
tion would then be defined by the receiving loop and the first RF stage impedance characteristics. bandwidth of 18.2 kHz. Using Qr
=
30 results in a 3 de,cibel system
18
SIGNAL PROCESSING TECHNIQUES
Two of the methods that may be applied to enhance the detection cap ability of an electronic locator are synchronous demodulation and non linear signal processing.
I
Because of the 90 degree phase difference
between the real part of a conductive target return and the normal inductive coupling, a synchronous system may be used to substantially reduce the straightthrough coupling as well as much of the return from the earthair interface. In such a system, the receiver output
is demodulated synchronously in phase with the transmitter reference, rejecting the unwanted quadrature signals and returning only the inphase return from a buried conductor. Certain limitations have prevented the use of synchronous techniques to date on the receiver transm itter type of locator. These limitations
take the form of pattern distortion, stability restrictions, and a reduction in utility due to the receiver and transmitter being permanently connected together. The pattern distortion problem is the most severe.
The wire interconnections between transm itter and receiver must of necessity intercept some transmitter energy and thus distort the field pattern. By keeping the interconnections directly along the null axis this form of distortion may be
of the loop as much as possible, minim ized.
A related problem is caused. by the connected wires
This may be eliminated by using balanced
themselves radiating.
receiver, transmitter, and demodulator configurations, and by thoroughly shielding the interconnections. For a minimum of 20 decibels of straightthrough rejection, the total system phase shift must be set and held to �6 electrical degrees. For
40 decibels of rejection, the total drift must be held to less than �O. 6
degrees and the system dynamic range and demodulator quadrature
19
rejection must well exceed this figure.
The ultimate limiting factor in
attainable improvement is dictated by the real term neglected in Equation (1). F or a 455 KHz frequency and a 50inch spacing, the
ultimate attainable improvement using synchronous techniques is approximately 40 decibels. Thus, a reasonable improvement may be but an extensive
expected using a relatively sim pIe synchronous system,
increase in detectability may only be obtained with an elaborate system using wideband circuitry, loops and demodulators. Two unique advantages of the receivertransmitter locator are the ability to track a buried pipe and the ability to triangulate for a depth indication. Both of these potential advantages require that the trans logarithmic amplifiers, and carefully balanced
mitter and receiver be totally separated without any disabling effects. It would then appear that a selectable dual demodulation system is of advantage, one synchronous and one asynchronous.
A large degree of clutter rejection independent of phase relationships
may be obtained by nonlinear processing of the detected signal, and
elaborate systems of this type are employed in certain military UHF mine locators, 8 Figure 7 shows a relatively simple means of nonlinear processing. a very high (3 transistor is driven from a voltage source, transconductance characteristic is produced. detected and filtered rece ver output E r If
a nonlinear the
In operation,
is summed with the base bias
voltage and the collector current is then taken as the output. If the stage is biased to point A (F igure 7), linear amplification is This operating
obtained with all signals being equally amplified.
point also corresponds to a maximum sensitivity condition, brought about by a maximum stage gain and a prebiasing effect of the base current upon the detection circuitry. Operation at point B will produce a squelch mode of operation in which no output will be obtained unless the return from a large target is received. This mode is of primary utility when tracing a single
known target.
20
(r'L
�
r;)
B
VB FIG. 7 EXPANDER CIRCUITRY

21

By biasing on the knee of the curve (point
C) low level signals will not
thereby providing expansion
be amplified as much as higher level ones,
to target signals while supressing normal background clutter.
In the demonstration model,
the signal expander stage forms a per
manent part of the circuitry, while the remainder of the circuit is designed to allow either linear or synchronous detection. The model
was built and initially evaluated on a nonsynchronous basis.
22
SYSTEM BLOCK DIAGRAM
The block diagram appears as Figure 8.
The transmitter consists of a
balanced CW oscillator driving a vertical transmitting loop, coupled to a mechanical null ad j uster to provide precise control over the null axis.
A n identical horizontal fixed loop form s the receiver front end, followed
by three stages of tuned RF amplification, required to bring the 100 f.1.volt received signals up to a level suitable for linear diode detection or synchronous demodulation. For synchronous operation, a balanced
reference is derived directly from the oscillator. Detection is followed by the expander circuit of Figure 7, which in
turn drives both a 0 1 DC m illiammeter and an integrated sonic alarm module,9 providing both visual and aural output indication.
RECEIVING
1S T
2 ND 3RD RF
AMP.
LOOP
RF AMP.
RF
AMP.
.
DETECTOR
, .
EXPANDER
"
INDICATOR
I I.
OSCILLATOR
tJ 6)
XMTG.
NUL L
LO 0 P
BLOCK
ADJUSTER
FIG.8
DIAGRAM
24
USE OF INTEGRATED CIRCUITRY
Integrated circuitry is employed for all stages except that of the signal expander, which makes use of a single, high gain silicon transistor.
\
Dual twoinput logic gates are modified to form linear differential amplifiers that serve as oscillator and the first three RF stages, while
a commercially available Ilintegrated" sonic module is employed as a replacement for the audio circuitry. The integrated circuit selected was the Fairchild !J.L914, a dual two input logic gate readily converted into a linear amplifier by using only one input on each side and driving the emitters from a negative current source. 10 The unused internal transistors are bypassed by connecting
their bases to the common emitter point.
11 The basic differential amplifier configuration is shown in Figure 9.
Normal use of this circuit requires a true current source at the eitters to ensure significant co mmon mode rejection as well as a matched pair of transistors, The fact that integrated circuitry is
used guarantees that both transistors will be at the same temperature and that both will have nearly identical characteristics. Referencing both input bases to ground via a low DC impedance elimi nates the common mode signal problem, allowing a simple resistor Transformer
and negative 3volt supply to replace the current source.
coupling of outputs forces the average collector voltage to be
V cc'
independent of any initial current unbalance, resulting in a stable yet extremely simple circuit. An input e 1 drives an emitter follower which in turn drives a grounded base stage to arrive at the right output, while the input e2 drives the Together the two produce or difference
I
same circuit as a common emitter stage. the composite right output K( e 1  e2 ) '
The opposite,
signal appears at the left and is given by K(e2  el)'
25
+'Vc
c
eout= K(e2 ell
eout= K(el e2)
(itle)+'
_
ie

1
it
 Ve
e
FIGURE 9

BASIC DIFFERENTIAL A"PLIFIER mNFIGURATION
26
The voltage gain is a function of the emitter current and the collect(\'t' load impedance. approximately 26 r be  i e where r input impedance in ohms emitter current in milliamperes. The input impedance of a common base stage is
be ie
=
=
The voltage gain of a common base stage is in turn given by RL/rbe where RL is the collector load resistance. differential stage is then given by E out E l RLi e
=
The total gain of the
z6"
(7)
with ie being onehalf the total emitter bias current it under nosignal conditions. The expected gain performance for the f,LL914 is plotted in Up to
Figure 10 as a function of emitter current and load resistance. 34
cie cibels
of stage gain may be obtained using a two milliampere The two internal resistors of 640 ohms
emitter bias current source.
each fix the maximum possible value of R at 640 or 1280 ohms, L depending upon the choice of collector connections.
The gain is easily controlled by varying the emitter current. low values of i ' e gain results. The input impedance is given by R.
For very
the RL / rbe raUo is less than unity and a negative stage
In
=
(3
r be
(8 )
27
DECIBELS
30
20
10
o
,
"
,
�
,
.2
.4
IE,
.
6
.8
1.0
MA.
FIG.10
SINGLE
STAGE
GAIN
28
where!3 is the current gain or hfe of the equivalent transistors. plotted for various f3 in Figure
R
11.
Experimental RX bridge measure
in
is
ments on several
fJ.L914
integrated circuits were made over the range of
500 kHz to 10 MHz.
The resultant room temperature values of
!3
were
in the range of 50 to 80,
producing input impedances above 2000 ohms for Since the input impedance is a function
all but the highest values of ie
of the emitter current, the interstage design must provide a relatively low source impedance and consequent power mismatch, lest the
\
changing input impedance affect the bandwidth and stability of the system. The differential amplifier configuration is also an effective limiter and operation in the limiting mode would be highly detrimental to the detection problem. Lim iting takes place when the total bias current it The total possible peaktopeak output swing is
is routed to one output.
given by itRL and the peak output is given by
E
out peak

1 i R 2 t L
(9 )
which is plotted in Figure 12 as a function of load resistance and bias current. If a large nonlimited output swing is to be obtained, the
third RF amplifier must be operated at a fixed high gain, for any attempt of gain control on this stage will be met with amplitude limiting problems.
29
3K
2K
1K
o o
.
2
.4 I E,
MA.
.6
.8
1.0
FIG. 1 1
INPUT
IMPEDANCE
11\1 n 11\11 X 'V 11\1
. 'V 11\1
z: r..
I ::I
'.1.1 . o o
go
os
31
COMPLETE SCHEMATIC
The complete schematic for a nonsynchronous integrated locator is given in Figure 13. The transmitter consists of a balanced CW oscillator,
formed by using the resonant centertapped transmitting loop as both collector loads for a single J..LL914.
A 455 kHz ceramic resonator
(transfilter) is crosscoupled from one collector to the other base to provide a feedback path. ance of 2000 ohms, The resonator selected has an input imped
an output impedance of 300 ohms, and a power Although lower in cost, the stability of
insertion loss of 2 decibels.
this circuit is not much worse than that of a crystal oscillator.
A three position OFFLOWHIGH switch chooses one of the two emitter
resistors providing transmitting loop voltages of 0.8 and 3 volts peak topeak. The LOW position is normally employed for outline purposes
after a target has been located. The receiver design philosophy is similar to that of Robertson. J..LL914 forms the first RF amplifier, 12 A
being driven in a balanced manner A singleended output
by the centertapped receiving loop assembly. is produced.
Interstage coupling is by means of 1: 1 bifialar wound
torodial transformers whose primaries are resonated at 455 kHz. The interstage Q is set to 8, allOWing fixed tuning. combining a
Indiana General Type Q1 material is used for the cores, low cost with a permeability of 125,
a high Q at the operating frequency,
and a temperature coefficient of 0.1 percent per degree Celsius. Since the internal collector resistors are smaller than the input impedance of the following stage, independent of the perstage gain. the system bandwidth is largely
1
{
x
& BEAD
"RESISIDR INTERNAL 10 Ie T3:
40 ON
T1� T2� 1llllNS ,36 BInLAR 0l0I(� RECEIVER Ql (DRE
1'
05
I
"
.
oQ 'Uz G')
, m
D
G)
o
m


x
Ito I D G')D 0
G) o
m
I
()
s:
I ]] <
m
OJ
s:
o o
m
r I D
o to o I 'U I
Z G')
'Z
9
�

6

K =
where
t
[ r
The coefficient of coupling
( 1)
A
r
=
diameter of loop, feet radial distance from antenna
=
taken along an axis of maximum coupling. k l
between transmitter and target will be attenuated further by sin
t/J,
where
t/J
is the target depression angle. x
Noting that
sin
t/J
=
[
B2+x2
J
1/2
and that
r3
=
[
B 2+x2
]
3/ 2
k will then be given by l
k
1
=
i (AD) 3/2
3
[
x
B2+x2
J
2
For a true 90 degree orientation of the receiving loop when centered above the target, maximum target to loop coupling exists, and the
coefficient of coupling k between target and receiver is given by 2
k 2
=
i
3
/ 1 (AD)3 2 3 x

7

The transm itter to receiver coupling via the target will be given by the product of k1 2 as given by k 16 3 Z x 1 Z
+
k1kZ
� 9"
(AD)
[
b
x
2
( 2)
J
(2)
since
The received voltage will be directly proportional to Equation
k k and the loop inductances will define the mutual inductance from l 2 transmitter to receiver via target.
For large x the receiver voltage for a given target falls off as 1/x6 just as in the induction balance and beat frequency locators. of depth follows a much less severe 1/x 2 law, For dis
tances small compared to B, the rate of signal drop off as a function allowing the detector The price paid
to retain good sensitivity to depths comparable to B. A s with all inductively coupled locators,
for this increase in detectability is a loss in resolution to small objects. the receiver signal increases
as the cube of the target diameter.
The optimum value of B is a function of the maximum desired depth of effective penetration contrasted against the minimum allowable resolution to small objects. 2 , A plot of the receiver voltage normalized and
to a given target of onefoot depth for various B and x appears as Figure plotted both in terms of decibels of received Signal, as ratios of target diam eters required to produce an equal strength signal. The loss of resolution as a function of B appears as Figure 3, plotted
as the ratio of diameters of targets required for equal detectability. The presence of a conductive target will not only increase the mutal coupling between the loops but will Simultaneously shift the phase of the received signal because of a resistive loss component introduced as the target lowers the
"Q
IT
of the coupling system. 4
Synchronous
detection can thus be employed to detect only the real component of the received signal, greatly adding to the detectability. If a
HJ.d3C
SA
NOIJ.:>3J.3C
.1.33.::1 . H J. d 3 C 9 9 00' 
08
09
Ov
0('; 
o
01.1."1:1 :::> 11:1.1. 311\1 " 1 a NOI.1.:::>3.1.30 ,,,n03
e
N 0 I n , 0 S :3 1::1
SI\
B
e �E>I::I
9
o 09
9' 099
017
oz;
o
e
'10
synchronous system is not used,
the limiting detectabil ity will be that
of a target whose mutual inductance equals the mutual inductance caused by straightthrough coupling, tortion caused by the circuitry. The mutual inductance caused by angular misalignment of the loops is given by MCos 8, where M is the mutual inductance between two parallel M is GroverTs5 ground return, and pattern dis
loops of diameter A and spacing B sharing a common axis.
obtained by making use of the extensive tabulations available. table 17 is directly applicable and is plotted as Figure 4. The induced receiver voltage will be given by
E
r
::
MwI
rms
( 3)
where I is the transm itting loop current assuming single turn rms transm itting and receiving loops. Since two mutual inductances are involved in target detection, the
induced receiver voltage due to a target increase as the square of the frequency usedo The highest possible operating frequency, consistant
wit4 terrestial absorption, sensitivity.
must be used to obtain maximum target
The model locator was designed with pipe location as a primary goal, requiring effective penetration to four feet of depth and better. value of B
=
A
4' and A
=
1 T was selected.
0'
9
9
'VIa
� '0�
'�0
J.O 0 :oj
1::::1 3 c:f
SAI::::IN3HONVN
,
j '
V
8
L,;"ti:i
, ,
12
CHOICE OF FREQUENCY
The effects of terrestial absorption and the locator detectability requirements interact to determine a suitable operating frequency.
\
by a factor of
Terrestial absorption will reduce Er further as a function of depth a.y where
a.
is the attenuation per foot in decibels and Total terrestial attenuation is
y is the round trip terrestial distance. then given by
Attenuation
=
a.
r, [ 2 +Lb 2 2J 1/ 2J
x
+
x
Decibels
(4)
A plot of the received E
is plotted as Figure 5 for B
r
that takes into accQunt various values of a.
=
4 feet.
The terrestial absorption will vary with frequency as well as the resistivity and dielectric constant of the earth at the point of penetra tion. Heiland6 has shown that the lowfrequency penetration equations and
are valid except for very high frequencies and soil resistivities,
that the results are always conservative for any frequency and con ductivity. One form of the lowfrequency equation is
d
: e..
=

60
Lf J
1/2
where d
=
p = resistivity in ohmcentimeters
1/
e:
penetration in meters
NO I.l. 'lfn N 3.l..l.'If
S/\
"
3
9 "ElI::l
9
o os
09
OV�
O 
o
14
as taken from Heiland's Figure
8.7 decibels.
1024.
At
l/E
attenuation,
the loss is
By changing dimensions,
the inverse relation
0.
=
0.045
Lp
0.
r
i 1/ 2
.J
i
( 5)
is obtained with
expressed in decibels/foot.
Values of resistivity for different media are taken from Heiland's text and plotted as a function of frequency in Figure 6. The curves indicate
an excessive amount of attenuation for most media at frequencies above
500
kHz.
favors the receiver sensitivity requirements and makes optimum use of available components. evaluation model. Neglecting the dielectric constant in the above analysis will have two opposing consequences. At the frequencies used, high values of k This is the operating frequency chosen for the
455
kHz is within a tolerable attenuation region and both
serve to decrease the terrestial absorption,
acting much as a bypass
capacitor to allow the high frequency energy to more freely travel through
a
lossy media.
At the same time,
large values of k introduce
a substantial mismatch between the loops in air and the earthair inter face, and can cause a ground reflection that could constitute a signifi Values of k encountered
cant portion of the total transmitted energy. can be extreme,
owing to water having a dielectric constant of
most soils having a moderate to substantial moisture content.
84
and
l
�z
H 11\1
'A:> N 3 n 03 1:::1.::1 J ro ,�0
OJ
0'
91
16
RECEIVER SENSITIVITY AND BANDWIDTH
The receiver sensitivity may be intrepreted in term s of that signal induced in the receiving antenna for a onedegree angular misalignment
'.
of the receiving loop from true null. Resonant loops are used for both receiver and transm itter to allow multiplication of transm itter current and receiver voltage. all metal used in the locator at neutral potential, employed. Several techniques were tried for the loop assemblies, the most practi
Q
. To place
balanced loops are
cal of which was a 10 inch by 12 inch oval loop of four conductor number 18 A WG communication cable, rewired by sequential connection The unloaded
a
to produce a four turn, center tapped configuration. loops gave
Q! s of 65, which dropped to 30 when driving
2000 ohm load.
The inductance is 18 microhenries which requires 6800 picofarads of capacity to resonate at 455 kHz. To allow for final tuning, a 700
picofarad trim mer is placed in parallel with 6500 picofarads of fixed capacity. Identical geometry is employed for both loops.
The received signal due to straightthrough coupling is given by E t  jMN N QtQ w _ . r r t r p
E
r
=
Sin 8
( 6)
where N r Nt number of receiving loop turns number of transmitting loop turns loaded receiver loop
= = = =
Qr
Qt r p
Q Q
Loaded transm itter loop and oscillator
=
equivalent parallel resistance of transm itting loop
Et
=
RMS voltage across transmitting loop.
17
For a onevolt RMS E Q r
=
t
across 1000 ohms and values of Q t
=
15 and
30,
a receiver sensitivity requirement of 73.5 microvolts results.
A full output from 100 microvolts of input may then be defined as a receiver design objective.
The receiver bandwidth is not overly critical,
and is chosen to be
narrow, while still allOWing fixed tuned RF stages using ordinary �10 percent capacitors. A 70 kHz, 15 peJ::cent bandwidth may be
selected using this criterion.
The overall selectivity and noise rejec
tion would then be defined by the receiving loop and the first RF stage impedance characteristics. bandwidth of 18.2 kHz. Using Qr
=
30 results in a 3 de,cibel system
18
SIGNAL PROCESSING TECHNIQUES
Two of the methods that may be applied to enhance the detection cap ability of an electronic locator are synchronous demodulation and non linear signal processing.
I
Because of the 90 degree phase difference
between the real part of a conductive target return and the normal inductive coupling, a synchronous system may be used to substantially reduce the straightthrough coupling as well as much of the return from the earthair interface. In such a system, the receiver output
is demodulated synchronously in phase with the transmitter reference, rejecting the unwanted quadrature signals and returning only the inphase return from a buried conductor. Certain limitations have prevented the use of synchronous techniques to date on the receiver transm itter type of locator. These limitations
take the form of pattern distortion, stability restrictions, and a reduction in utility due to the receiver and transmitter being permanently connected together. The pattern distortion problem is the most severe.
The wire interconnections between transm itter and receiver must of necessity intercept some transmitter energy and thus distort the field pattern. By keeping the interconnections directly along the null axis this form of distortion may be
of the loop as much as possible, minim ized.
A related problem is caused. by the connected wires
This may be eliminated by using balanced
themselves radiating.
receiver, transmitter, and demodulator configurations, and by thoroughly shielding the interconnections. For a minimum of 20 decibels of straightthrough rejection, the total system phase shift must be set and held to �6 electrical degrees. For
40 decibels of rejection, the total drift must be held to less than �O. 6
degrees and the system dynamic range and demodulator quadrature
19
rejection must well exceed this figure.
The ultimate limiting factor in
attainable improvement is dictated by the real term neglected in Equation (1). F or a 455 KHz frequency and a 50inch spacing, the
ultimate attainable improvement using synchronous techniques is approximately 40 decibels. Thus, a reasonable improvement may be but an extensive
expected using a relatively sim pIe synchronous system,
increase in detectability may only be obtained with an elaborate system using wideband circuitry, loops and demodulators. Two unique advantages of the receivertransmitter locator are the ability to track a buried pipe and the ability to triangulate for a depth indication. Both of these potential advantages require that the trans logarithmic amplifiers, and carefully balanced
mitter and receiver be totally separated without any disabling effects. It would then appear that a selectable dual demodulation system is of advantage, one synchronous and one asynchronous.
A large degree of clutter rejection independent of phase relationships
may be obtained by nonlinear processing of the detected signal, and
elaborate systems of this type are employed in certain military UHF mine locators, 8 Figure 7 shows a relatively simple means of nonlinear processing. a very high (3 transistor is driven from a voltage source, transconductance characteristic is produced. detected and filtered rece ver output E r If
a nonlinear the
In operation,
is summed with the base bias
voltage and the collector current is then taken as the output. If the stage is biased to point A (F igure 7), linear amplification is This operating
obtained with all signals being equally amplified.
point also corresponds to a maximum sensitivity condition, brought about by a maximum stage gain and a prebiasing effect of the base current upon the detection circuitry. Operation at point B will produce a squelch mode of operation in which no output will be obtained unless the return from a large target is received. This mode is of primary utility when tracing a single
known target.
20
(r'L
�
r;)
B
VB FIG. 7 EXPANDER CIRCUITRY

21

By biasing on the knee of the curve (point
C) low level signals will not
thereby providing expansion
be amplified as much as higher level ones,
to target signals while supressing normal background clutter.
In the demonstration model,
the signal expander stage forms a per
manent part of the circuitry, while the remainder of the circuit is designed to allow either linear or synchronous detection. The model
was built and initially evaluated on a nonsynchronous basis.
22
SYSTEM BLOCK DIAGRAM
The block diagram appears as Figure 8.
The transmitter consists of a
balanced CW oscillator driving a vertical transmitting loop, coupled to a mechanical null ad j uster to provide precise control over the null axis.
A n identical horizontal fixed loop form s the receiver front end, followed
by three stages of tuned RF amplification, required to bring the 100 f.1.volt received signals up to a level suitable for linear diode detection or synchronous demodulation. For synchronous operation, a balanced
reference is derived directly from the oscillator. Detection is followed by the expander circuit of Figure 7, which in
turn drives both a 0 1 DC m illiammeter and an integrated sonic alarm module,9 providing both visual and aural output indication.
RECEIVING
1S T
2 ND 3RD RF
AMP.
LOOP
RF AMP.
RF
AMP.
.
DETECTOR
, .
EXPANDER
"
INDICATOR
I I.
OSCILLATOR
tJ 6)
XMTG.
NUL L
LO 0 P
BLOCK
ADJUSTER
FIG.8
DIAGRAM
24
USE OF INTEGRATED CIRCUITRY
Integrated circuitry is employed for all stages except that of the signal expander, which makes use of a single, high gain silicon transistor.
\
Dual twoinput logic gates are modified to form linear differential amplifiers that serve as oscillator and the first three RF stages, while
a commercially available Ilintegrated" sonic module is employed as a replacement for the audio circuitry. The integrated circuit selected was the Fairchild !J.L914, a dual two input logic gate readily converted into a linear amplifier by using only one input on each side and driving the emitters from a negative current source. 10 The unused internal transistors are bypassed by connecting
their bases to the common emitter point.
11 The basic differential amplifier configuration is shown in Figure 9.
Normal use of this circuit requires a true current source at the eitters to ensure significant co mmon mode rejection as well as a matched pair of transistors, The fact that integrated circuitry is
used guarantees that both transistors will be at the same temperature and that both will have nearly identical characteristics. Referencing both input bases to ground via a low DC impedance elimi nates the common mode signal problem, allowing a simple resistor Transformer
and negative 3volt supply to replace the current source.
coupling of outputs forces the average collector voltage to be
V cc'
independent of any initial current unbalance, resulting in a stable yet extremely simple circuit. An input e 1 drives an emitter follower which in turn drives a grounded base stage to arrive at the right output, while the input e2 drives the Together the two produce or difference
I
same circuit as a common emitter stage. the composite right output K( e 1  e2 ) '
The opposite,
signal appears at the left and is given by K(e2  el)'
25
+'Vc
c
eout= K(e2 ell
eout= K(el e2)
(itle)+'
_
ie

1
it
 Ve
e
FIGURE 9

BASIC DIFFERENTIAL A"PLIFIER mNFIGURATION
26
The voltage gain is a function of the emitter current and the collect(\'t' load impedance. approximately 26 r be  i e where r input impedance in ohms emitter current in milliamperes. The input impedance of a common base stage is
be ie
=
=
The voltage gain of a common base stage is in turn given by RL/rbe where RL is the collector load resistance. differential stage is then given by E out E l RLi e
=
The total gain of the
z6"
(7)
with ie being onehalf the total emitter bias current it under nosignal conditions. The expected gain performance for the f,LL914 is plotted in Up to
Figure 10 as a function of emitter current and load resistance. 34
cie cibels
of stage gain may be obtained using a two milliampere The two internal resistors of 640 ohms
emitter bias current source.
each fix the maximum possible value of R at 640 or 1280 ohms, L depending upon the choice of collector connections.
The gain is easily controlled by varying the emitter current. low values of i ' e gain results. The input impedance is given by R.
For very
the RL / rbe raUo is less than unity and a negative stage
In
=
(3
r be
(8 )
27
DECIBELS
30
20
10
o
,
"
,
�
,
.2
.4
IE,
.
6
.8
1.0
MA.
FIG.10
SINGLE
STAGE
GAIN
28
where!3 is the current gain or hfe of the equivalent transistors. plotted for various f3 in Figure
R
11.
Experimental RX bridge measure
in
is
ments on several
fJ.L914
integrated circuits were made over the range of
500 kHz to 10 MHz.
The resultant room temperature values of
!3
were
in the range of 50 to 80,
producing input impedances above 2000 ohms for Since the input impedance is a function
all but the highest values of ie
of the emitter current, the interstage design must provide a relatively low source impedance and consequent power mismatch, lest the
\
changing input impedance affect the bandwidth and stability of the system. The differential amplifier configuration is also an effective limiter and operation in the limiting mode would be highly detrimental to the detection problem. Lim iting takes place when the total bias current it The total possible peaktopeak output swing is
is routed to one output.
given by itRL and the peak output is given by
E
out peak

1 i R 2 t L
(9 )
which is plotted in Figure 12 as a function of load resistance and bias current. If a large nonlimited output swing is to be obtained, the
third RF amplifier must be operated at a fixed high gain, for any attempt of gain control on this stage will be met with amplitude limiting problems.
29
3K
2K
1K
o o
.
2
.4 I E,
MA.
.6
.8
1.0
FIG. 1 1
INPUT
IMPEDANCE
11\1 n 11\11 X 'V 11\1
. 'V 11\1
z: r..
I ::I
'.1.1 . o o
go
os
31
COMPLETE SCHEMATIC
The complete schematic for a nonsynchronous integrated locator is given in Figure 13. The transmitter consists of a balanced CW oscillator,
formed by using the resonant centertapped transmitting loop as both collector loads for a single J..LL914.
A 455 kHz ceramic resonator
(transfilter) is crosscoupled from one collector to the other base to provide a feedback path. ance of 2000 ohms, The resonator selected has an input imped
an output impedance of 300 ohms, and a power Although lower in cost, the stability of
insertion loss of 2 decibels.
this circuit is not much worse than that of a crystal oscillator.
A three position OFFLOWHIGH switch chooses one of the two emitter
resistors providing transmitting loop voltages of 0.8 and 3 volts peak topeak. The LOW position is normally employed for outline purposes
after a target has been located. The receiver design philosophy is similar to that of Robertson. J..LL914 forms the first RF amplifier, 12 A
being driven in a balanced manner A singleended output
by the centertapped receiving loop assembly. is produced.
Interstage coupling is by means of 1: 1 bifialar wound
torodial transformers whose primaries are resonated at 455 kHz. The interstage Q is set to 8, allOWing fixed tuning. combining a
Indiana General Type Q1 material is used for the cores, low cost with a permeability of 125,
a high Q at the operating frequency,
and a temperature coefficient of 0.1 percent per degree Celsius. Since the internal collector resistors are smaller than the input impedance of the following stage, independent of the perstage gain. the system bandwidth is largely
1
{
x
& BEAD
"RESISIDR INTERNAL 10 Ie T3:
40 ON
T1� T2� 1llllNS ,36 BInLAR 0l0I(� RECEIVER Ql (DRE
1'
05
I