This is only a preview of the May 2004 issue of Silicon Chip. You can view 16 of the 96 pages in the full issue, including the advertisments. For full access, purchase the issue for $10.00 or subscribe for access to the latest issues. Items relevant to "Component Video To RGB Converter":
Items relevant to "StarPower: A Switching Supply For Luxeon Star LEDs":
Items relevant to "Wireless Parallel Port":
Items relevant to "Poor Man's Metal Locator":
Purchase a printed copy of this issue for $10.00. |
Poor
Man’s
Metal
Locator
By
Thomas
Scarborough
This unique metal locator contains just five components – a
low-cost IC, a variable capacitor, two search coils and a crystal
earpiece. And believe it or not, despite the apparently simplicity
it is capable of surprisingly good results.
T
HIS CIRCUIT represents a brand
new genre of metal locator. Although it borrows from other kinds of
locators/detectors, its basic principle
of operation is new.
And while this may seem to stretch
one’s credulity, the performance
matches that of a budget Induction
Balance (IB) detector. Build it – and
you will see!
During testing, it detected an old
English penny at 150mm (6”) in air.
However, I would only put my neck
on the block for 125mm (5”), since a
number of factors influence sensitivity.
It roughly matches the performance
of the now sadly misnamed Matchless Metal Locator (SILICON CHIP, June
80 Silicon Chip
2002), while using just one-fifth of the
components.
This locator may therefore represent
the writing on the wall for budget IB
types and even puts paid to most of the
advantages (the few remaining ones)
of Beat Frequency Operation (BFO)
detectors.
Overview
Instead of using a search and a reference oscillator (as in the BFO type), or
transmit and receive coils (as in IB),
this detector uses two transmitters
(or search oscillators) with IB-style
coil overlap.
As will be seen from Fig.1 and our
photographs, these are extremely
simple in design. Each oscillator comprises just one-quarter of a common
quad op amp IC plus the search coil!
The frequencies of the two oscillators are then mixed (in similar fashion to a BFO) to produce an audible
heterodyne.
On the surface of it, this design
would seem to represent little more
than a twinned BFO metal locator.
However, what makes it different above
all else (and significantly increases
its range) is that each coil modifies
the frequency of the adjacent oscillator through inductive coupling. This
introduces the “balance” that is present in an IB metal locator and boosts
sensitivity well beyond that of a BFO.
siliconchip.com.au
Beyond this, all that is required is
a means to control the mixer output
frequency, so that the metal locator can
be tuned. This could be accomplished
in a number of ways but the method
chosen here is a variable capacitor (a
standard 100pF AM radio tuner) wired
between the two oscillator outputs.
Since the concept borrows from
both BFO and IB, we shall give a nod to
each of these by naming the principle
“Beat Balance”.
Characteristics
The main characteristics of beat
balance (BB) are as follows:
• Depending on the way it is designed, a BB metal locator potentially
offers the same sensitivity as IB.
• It requires no receiver amplifier or
level detector, thus vastly simplifying
design and reducing cost. The present
circuit uses just two main components,
while matching the performance of a
budget IB metal locator (which would
probably have 10-20 parts).
• Both search oscillators are identical, therefore BB offers high immunity
to voltage and temperature variations.
This obviates the need for compensation circuits, including voltage reg-ulation.
• Each search coil has the opposite
response to metal, thus BB has a high
degree of immunity to ground mineralisation. At the same time, it offers
good discrimination at the point where
the two search coils overlap
This view inside the box shows the WHOLE circuit (with the exception of the
coils, of course, and the tuning capacitor hidden under the PC board).
output (pin 1) is delayed during transfer to inverting input pin 2. An approx. 8V/µs slew rate further delays
switching of IC1a, thus setting up a
rapid oscillation.
One end of the search coil is further
wired to the non-inverting input (pin
3), which stabilises operation. While
pin 3 could be left “floating”, this
would be a less satisfactory arrangement.
Since different ICs have different
slew rates, as well as different input
impedances, they are unlikely to work
in this circuit. However, the TL074CN
IC is widely available and there should
be no sourcing problems.
The search coil is a critical part of
the oscillator and needs to be suitably
designed to achieve oscillation and to
obtain the required frequency.
While this frequency needs to be
high, it should not be so high that
noise or instability are introduced.
Both the characteristics of IC1 and
the inductance of the coil influence
oscillator frequency, which lies around
260kHz without a Faraday shield. The
Faraday shield approximately doubles
the inductance of the coil, thus roughly
halving the frequency.
IC1b is wired in exactly the same
The circuit
The design is based on the simplest
of inverter oscillators. As far as I am
aware, this two-component op amp
oscillator also represents a first.
Let us focus first on IC1a.
Since an inductor resists rapid
changes in voltage (called
reactance), any change in
the logic level at IC1a’s
Fig.1: the complete circuit – it could hardly be any simpler, could it?
siliconchip.com.au
May 2004 81
Fig.2: construction of the oscillator section could hardly
be simpler. Follow the diagram and photograph above
and you can’t go wrong!
way as IC1a, except that its search coil
is connected in opposite phase.
As the search head is swept over
the ground, the presence of metal increases the inductance of L1 and then
L2, or vice versa, thus bringing about
a dip in the oscillator frequencies. A
third op amp, IC1c, is used to mix the
output of the oscillators, thus creating
an audible difference frequency, or
beat frequency.
This leads us to the one distinctive
feature of BB. Not only does the presence of metal alter the frequency of
a search oscillator but as in the case
of IB, it also influences the adjacent
coil. In fact, both coils influence each
other through mutual induction, thus
greatly enhancing the sensitivity of
the system.
Beyond this, we only need to find
a method of tuning the metal locator. This is achieved using variable
capacitor VC1, which further couples
the two inductors (the search coils),
thus offering a means of controlling
balance. Almost any variable capacitor
should work in this position, although
it should preferably have a smaller
value; eg, 47pF to 100pF. A small value
capacitor (eg 47pF) can be wired in series with VC1 to reduce a larger value.
A crystal earpiece is used for sound
output. While a piezo sounder may be
used (without integral electronics),
this is not likely to offer good volume.
If the volume in the earpiece is too
high, use a suitable series resistor to
82 Silicon Chip
reduce it. An inductive sounder or
earpiece is not recommended, because
it could overload IC1c.
Current consumption is around
15mA. Therefore an 8 x AA battery
pack should last around 70 hours.
Construction
There’s so little on the PC board
that it would be difficult to make a
mistake. OK, so you could put the IC
in back-to-front, likewise the search
coils’ starts and ends could be inadvertently swapped. Apart from that,
there’s precious little to worry about.
Fit the 12 stakes to the PC board and
solder them into place, then solder the
two jumper wires as shown. Normally
we would say use resistor lead offcuts
for this – but there aren’t any! You’ll
have to use some tinned copper wire
instead.
Now comes the challenge of popu-
lating the PC board! Since this is a
sensitive, high-frequency circuit, I
would recommend that IC1 be soldered directly to the PC board (ie,
not socketed). So long as you insert
this component the right way round,
there would appear to be little to go
wrong! The TL074CN is a fairly “tough
bird”– but be reasonably quick with
the soldering iron.
Wire up the variable capacitor VC1,
the socket for the crystal earpiece and
the battery and switch (carefully note
the polarity of the battery leads – an
error here could destroy the circuit).
It’s usual to insert the on-off switch in
the positive battery lead. Some battery
holders have solder tags, others (such
as the one we used) need a 9V battery
snap to connect them. Again, watch
the polarity!
Now mount the on-off switch and
the jack socket (for the crystal earpiece)
Two overlapping
coils are wound
using 30SWG wire
and fastened to a
non-metallic base.
Fig.8 microphone
cable connects
the coils to the
oscillator.
siliconchip.com.au
on the case. I used long bolts to clamp
VC1 underneath the PC board and
found this an easy and effective way
of fixing the variable capacitor to the
case. I used a slice of non-conductive
rubber to isolate VC1 from the back of
the PC board.
Winding the coils
Next, wind the two search coils.
They are 70 turns of 30SWG (0.315mm)
enamelled copper wire on 120mm
diameter formers. Faraday electrostatic shields are essential for circuit
stability. These are connected to 0V
and should use balanced (figure-8)
screened microphone cable. Winding
of the coils is not critical and a little
give and take is permissible. However
they should be as close as possible to
identical.
Wind the coils around the formers,
temporarily holding them together by
passing stubs of insulating tape under
them and pressing them together over
the top. Once you have wound the
coils, bind them tightly with insulating
tape around their entire circumference. Scrape the enamel off the ends
of the coils’ enamelled copper wires to
solder them to the microphone cables.
Now add Faraday shields. Prepare
some long, thin strips of aluminium or
tin-foil. Twist a 100mm length of bare
wire around each coil, over the insulating tape. This wire provides electrical
contact with the foil and is soldered to
the microphone cable screens.
Beginning at the base of the bare
wires, wind the foil around the circumference of the coils, so that no
insulating tape is still visible under the
foil – but the foil should not complete
a full 360°.
Leave a small gap (say 10mm) so that
the foil does not meet after having done
most of the round. Now again tightly
bind the coils with insulating tape
Parts List – Poor Man’s
Metal Locator
1 PC board, code 04105041,
51 x 64mm
1 plastic case, 150 x 90 x 50mm
1 TL074CN quad op amp (IC1)
1 10-100pF variable (tuning)
capacitor, with knob
1 crystal earpiece
1 3.5mm mono earphone socket
1 8 x AA battery holder and
batteries
1 battery snap (if required by
holder)
1 SPST power switch
55m (approx.) 30SWG enamelled copper wire
2 lengths twin shielded balanced
microphone coax (fig.8),
approx. 2m long
hookup wire
2 20mm lengths tinned copper
wire (for PC board links)
2 lengths of aluminium foil,
approx. 20mm wide (for shield)
1 length PVC conduit to suit
(handle)
4 20mm M3 screws and nuts
Suitable cable ties
Clear polyester resin
around their entire circumference.
Attach the coils to the circuit by
means of the specified microphone
cables, being careful to identify the
beginning and end wires correctly as
shown. If these are not correctly identified, there could be a 20% loss of sensitivity. The Faraday shields should be
connected to the cable screens and to
0V on the PC board as shown.
Hardware
The “hardware” construction is just
as simple as the rest of the design. The
two search coils are fixed to a plate of
Masonite or similar, with a single PVC
electrical conduit shaft attached. The
control box, containing all of the electronics apart from the coils, attaches to
the shaft via a pair of cable ties.
The top of the shaft is then held in
the hand just above the control box,
while its upper length rests against
the back of one’s forearm.
Use a stiff, non-metallic plate for
the search head. Masonite is both stiff
and easy to work with (I cut up an old
Masonite clipboard for the purpose).
Before the coils (or anything else)
are attached to the base plate, their
best operating position needs to be
determined.
Begin by placing the two coils on
the search head plate, directly on top
of one another (that is, “meshed”),
with their beginning and end wires
positioned as shown. Turn VC1 to its
mid-position. Switch the detector on,
then slowly move the coils apart.
When the coils have all but been
separated from one another, a tone
will be heard in the crystal earpiece.
Adjust the coil positions so that this
is a fairly low tone – then drill holes
and use cable ties (several for each
coil) to fasten them in this position
on the plate.
Once the cable ties have been tightened, carefully bend the coils until a
low tone is again heard in the crystal
earpiece. To lower the tone, create a
greater overlap of the two coils (ie, a
larger segment in the middle), and
vice versa.
To construct a shaft, saw the end off
a length of PVC piping at a 25° angle.
Drill holes through the pipe close to
its bottom end and holes through the
centre of the search plate. Then bind
the pipe to the search plate with cable
ties. The pipe (or shaft) will later be
fixed permanently to the search plate
The lid is secured to the PVC pipe by means of a couple of cable ties. When these are pulled tight, they really grip well!
Then the lid (which is effectively now the base) is screwed onto the box – and presto!
siliconchip.com.au
May 2004 83
The “works” box (left) is mounted
near the top of the PVC tube. This
view shows the front of the box, with
the earpiece socket, tuning capacitor
and on/off switch. We haven’t put a
fancy label on this project – it would
sort of ruin the effect, wouldn’t it?
Besides, the label would be on the
underside. At right is the coil end,
showing how it mounts to the PVC
tube.
with clear polyester resin (see below).
The two coils need to be set rigidly
in position on the search head, so that
they will not move even slightly when
the metal locator is in use. I would
recommend that they be potted in clear
epoxy resin, which is available from
most hardware stores, together with
the necessary hardener or catalyst.
A section of one coil should be left
exposed where the two coils intersect,
so as to enable final fine adjustment.
This section of coil may be temporarily
protected with Blu-Tac.
Be sure to plug the holes beneath
the search head before pouring the
resin, since it is very runny and sticks
faster than many glues! The detector’s
PVC shaft is bound to the search head
with the resin.
I tied the control box to the shaft
with cable ties and used a little allpurpose glue to assist. Cable ties were
further used to bind the cables to the
shaft. No hand grip was attached to
the prototype but the shaft was kept
long at the top, so as to rest against
the back of my forearm as I gripped
the shaft with my hand.
Checking it out
Once construction is complete
and everything checked, switch on
The Patent That Came Close – But No Cigar!
US Patent 4,196,391 of 1980, by Harold J Weber, was a good piece of
original thinking that makes fascinating reading. It came so close to Beat Balance – but no cigar!
The patent describes a metal locator which uses two transmitters, as does
ours, balanced by a variable capacitor. However, the patent seeks “least interaction between the inductors”, while ours is almost entirely dependent on
such interaction.
In the patent, a variable capacitor is used to balance the frequencies of
the two transmitters. Ours, on the other hand, uses variable capacitor VC1 to
obtain a beat frequency.
The patent mixes the two transmitter frequencies with a third frequency from
a Beat Oscillator. The purpose of this is to provide binaural location of metal
objects, which is the “primary object” of the patent. Ours obtains an audible
heterodyne by mixing the two transmitter frequencies directly, its primary object
being to boost sensitivity.
Harold J. Weber states in his patent that he is “surprised” to find “pronounced
improvement” by alternating the signals in the ears – for which he employs a
Gate Oscillator and a Gate Switch Driver. Why the expression of surprise? It
is hardly scientific language.
My hunch is that the pronounced improvement lay not in the ears but in
the coils. He considered that he had merely invented another “beat frequency
detector type”, thus missing a significant breakthrough by a whisker.
84 Silicon Chip
and tune VC1 until a clear difference
frequency or heterodyne is heard in
the crystal earpiece. This should ideally be a low tone of just a few tens of
Hertz. It will be found that sensitivity
is dependent to some extent on tone
and some experimentation will yield
the best frequency. If necessary, further
adjust the coils, moving them further
apart if the circuit is silent, or closer
together if the frequency is too high.
Bring a metal item close to the coils.
It will be found that one coil causes
the tone in the crystal earpiece to
rise, while the other causes it to fall,
with the overlapping segment offering
discrimination between ferrous and
non-ferrous metals.
The detector should pick up a large
coin (eg, 50c) at up to 150mm in air
(125mm is a sure bet), while large
metal objects (eg, a cake tin) will be
detected at half a metre. At close range,
it is capable of picking up a pin. When
in use, hold the search head close to
the ground, sweeping it to and fro.
Unlike IB, the two coils give opposite
responses to metal, which one soon
becomes accustomed to.
While the detector is unusually stable, readjustment by means of VC1 will
inevitably be required, particularly
immediately after switch-on.
That said, I found this to be a very
“well-behaved” metal locator. It was
easy to build, easy to set up and is a
SC
joy to use.
siliconchip.com.au
|