This is only a preview of the July 1989 issue of Silicon Chip. You can view 44 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 "Experimental Mains Hum Sniffers":
Articles in this series:
Articles in this series:
Articles in this series:
Articles in this series:
|
BUILD THESE EXPERIM
MAINS HUM SNIFFERS
In this article, we present two simple
circuits for sniffing out mains hum
signals. Both can be built using parts
from your junkbox and can be quickly
assembled.
By STEVE PAYOR
These two circuits will let you
check out the extent of electric
fields permeating the average
household, mostly originating from
the mains wiring. Some of the
things worth checking out are
fluorescent lights, TV screens, appliance cords, power points and
electric blankets.
They are not intended to be practical devices for tracing mains wiring, however. Although appliance
cords will register strong indications on these devices, wiring that
is buried in walls is easily masked
by the surface material. We'll have
more to say about these performance limitations later.
That said, both circuits are
worth building, if only for their
curiosity value. The simpler of the
two let's you listen in to mains hum
fields on a pair of headphones
while the other circuit uses a meter
to indicate the source of the hum.
when the headphones are unplugged. The current drain is around
3-5mA.
Without the optional earth lead,
the probe will respond to both live
and grounded objects, since your
body will be picking up a certain
amount of hum.
If the circuit is grounded the probe will become silent when moved
close to a grounded object. Thus it
can easily distinguish between correctly earthed 3-wire appliances
(eg, a toaster) and appliances
which only have two wires (eg,
some desk lamps).
PROBE
G
The el-cheapo circuit
Fig.1 shows the simpler of our
two circuits. It consists of nothing
more than a JFET in series with a
battery and a pair of headphones.
The circuit is best built on a small
scrap of perforated board, with the
gate lead of the JFET close to one
end. No on/off switch is needed
since the battery is disconnected
+
1.5V
Gffio
VIEWED FROM
BELOW
OPTIONAL EARTH
LEAD
.,-
Fig.I: this simple hum sniffer circuit
uses a JFET in series with a battery
and a pair of headphones.
The circuit shown in Fig.I can be built up on a scrap piece of perforated board. Note that the JFET (at end of
pen) is mounted close to one end of the board, with its gate lead nearest to the edge.
42
SILICON CHIP
ENTAL
Running the probe along the outside of a power cord will quickly
tell you if the cord is plugged into a
live socket. Not only that, but you
can actually pick the active lead
itself.
Wiring buried inside walls is a
little trickier. Often the conductivity of the wall material masks the
exact location. It is not unusual for
the hum to be spread over a general
area of about half a metre in width.
Let's now look a little more closely at the circuit operation.
With the gate of the JFET left
floating, the gate-source voltage
tends towards 0V, due to leakage of
the gate-source junction and surface leakage on the board between
the gate and source wiring (the
source wiring acts as a guard or
shield between the gate wiring and
the rest of the voltages in the
circuit).
With a gate-source voltage of
zero, the JFET current is at a maximum. The impedance of the gate
circuit is very high, so the electric
field from nearby 240V 50Hz wiring
induces a signal of several volts into the gate. Negative voltages turn
the JFET off (somewhere between
- 2V and - 8V for the 2N5459),
positive voltages can't turn it on
any more that it already is, and
voltages above + 0.6V are clipped
by the diode action of the gatesource junction.
Thus a 10V p-p signal will cause
the gate voltage to swing between
0V (approx.) and - 10V, giving a
net negative bias to the gate and a
reduction in current through the
JFET.
Try this simple experiment: plug
a multimeter into the headphone
socket and note the zero-signal current. Now bring a 50Hz source near
to the JFET (your finger will do) and
The differential hum detector uses a
meter to indicate the direction of the
hum source. You will probably
already have most of the parts in
your junkbox.
watch the average DC current
decrease as the signal increases.
Now back to our buried wiring
problem. You could try mapping out
the field strength on the wall, using
the multimeter instead of the headphones to get quantitative measurements. With enough readings, a
contour map could be plotted and
the contours would indicate the
most likely location of the buried
wiring.
PARTS LIST
1 PC board, code
SC04105891, 168 x 88mm
1 centre-zero meter movement,
± 1 00mA full scale sensitivity
(see text)
2 2N5459 JFETs (see ,text)
1 2k!J horizontal or vertical
mount trimpot
1 1 .5V AA battery holder
1 1 .5V AA battery
1 miniature toggle switch or
slide switch
1 perspex handle to suit
There is an easier way - two
detectors placed symmetrically
either side of the wiring will both
pick up the same amount hum, but if
one is a little closer than the other,
the balance will shift. A pair of
detectors with a meter to measure
the difference between them
should, in theory, be able to zero in
on the buried wiring.
Differential detector
This leads us to the differential
hum detector as shown in Fig.2. The
2k!J trimpot sets the initial balance
between the zero-signal JFET
resistances. If both probes pick up
the same amount of hum, then both
JFETS will turn off by approximately equal amounts and the meter
needle will remain centred. If one
probe picks up more hum, the current through its associated JFET
will be reduced and the meter is
wired so that the needle swings in
the direction of the probe receiving
the most hum.
Thus the needle actually points to
the source of the hum.
]UL Y 1989
43
Choosing a JFET
s:\
1.5V;r..
GUARD
GUARD
250•0·25QuA
G<at>D
VIEWED FROM
BELOW
DIFFERENTIAL HUM DETECTOR
Fig.2: the differential hum detector circuit. If one probe picks up
more hum than the other, the current through its associated JFET is
reduced and the meter needle swings in the direction of the hum.
Layout is important in a circuit
like this. Symmetry of capacitance
must be maintained and DC leakage
currents must be controlled with
guards.
If you are using our PCB to build
this device, it is best to obtain it undrilled since a number of hole locations have been provided to accommodate a variety of components. So
dig deep into your junkbox and see
what you can press into service.
Meter movement
The board will accommodate a
centre-zero tuning meter (as sold by
DSE, Jaycar, Altronics, etc) in the
lowest two holes. "Edge" meters
can be fitted to the middle two holes
while the upper holes, drilled to
3mm, will accommodate an MU45style unit which can be bolted
directly to the board.
Suitable meter sensitivities range
from ± 50µA to ± 250µA full scale
(optimum is about ± l00µA) .
Don't worry about which way
around the + and - connections
are arranged - the PCB has four
pads marked A, A and B, B which
effectively form a double-pole,
double-throw (DPDT) switch to
reverse the meter connections if it
happens to swing the wrong way.
Start by soldering pads B and B.
If the meter subsequently swings
backwards, desolder B and B and
solder A and A instead.
Above: the two probes are nothing more than fancy
copper patterns on the back of the PCB. At right is
another version of the detector using 2N5459 N-channel
JFETs and a small centre-zero tuning meter.
44
SILICON CHIP
2N5459 N-channel JFETs are
recommended but lower current
2N5458s or 2N5457s can be used
with a 5kQ trimpot and a lower current meter (eg, ± 50µA).
P-channel JFETs such as the
2N5460 can also be used. In this
case, reverse the battery holder
and solder pads A and A for the
meter. Note that the pin connections for P-channel JFETs are different and the leads will have to be
bent so that they fit the board
correctly.
On/off switch
Either a toggle or a slide switch
can be used here. In addition, optional solder pads are provided so
that " up" can be either ON or OFF
as you please.
Trimpot & battery holder
Any trimpot from 2kQ to 5kQ,
horizontal or vertical, can be used,
preferably something with a plastic
knob or thumbwheel. A 5kQ or 10kQ
pot could also be used provided it is
shunted with a pair of 2.2kQ or lkQ
resistors.
The single cell battery holder is
available from Tandy Electronics
(Cat. 270-401). Alternatively, this
item is also available from
Radiospares (stock number 489908).
Perspex handle
An insulating Perspex handle
How to Bend Perspex
ZONE BEING SOFTENED
(OD BOTH SIDES)
PERSPEX
To bend Perspex, you have to heat a localised strip
to a temperature of around 120°C. You can do this
with a hot air gun such as a heatshrink gun or electric
paint stripper, or possibly even a hair dryer'. You can
also use an electric bar radiator or a hot stove element as shown in Fig.3 . You will need two sheets of
insulating material such as Fibro (or asbestos cement, if you can) or, in a pinch, heavy gauge
aluminium to mask off all but the desired bending
area.
Perspex absorbs long wavelength infrared quite
well. It won't actually melt but at the correct
temperature goes "rubbery" and is easily formed.
Overheating will blister and discolour the surface so
don't overdo it.
\
I
ALUMINIUM------'
SHEETS
:-,-::=----,---
''.\!// \\\/// \\\i/f \\\JI/ ~
c:::::::i c:::::::J
RED HOT STOVE
ELEMENT
c::::]
6mm OR LARGER GAP
(ABOUT 2x SHEETTHICKNESS)
c::::]
f
Fig.3: Perspex can be easily bent by heating a narrow
strip of the material to about 120°C.
flt\
tr~s
2N5460
VIEWED FROM BELOW
Fig.5: this diagram shows the
pinouts for the 2N5460 Pchannel JFET. You will have to
bend the leads so that it fits
the board correctly.
Fig.4: this version
uses N-channel JFETs
and an MU45-style
meter. Don't forget to
bridge the
appropriate pads on
the back of the board
associated with the
meter and the on/off
switch (see text).
can be fitted using the same screws
and nuts used to secure the battery
holder. If you have never bent
Perspex before, ref er to the accompanying panel.
Whether you hold the unit by an
insulated handle or by the meter
case depends on what sort of field
you are standing in. Sometimes you
may need to earth yourself or even
hold the unit by the battery to earth
it as well. In short, you will have to
experiment to determine the best
way to hold the unit for a given
situation.
Start by bridging the B pads on the
back of the board. If the meter
swings the wrong way, desolder them
and bridge the A pads instead. You
must also bridge one set of pads
associated with the on/off switch.
Performance & limitations
The unit may give misleading
readings when overloaded. Too
strong a signal will cut off both
JFETs so keep the signal strength
within reasonable limits by varying
the way you hold it. In most cases,
holding the meter case between
outstretched finger tips, with no
earthing, will give the best results.
Appliance cords will register at a
distance of 100mm or more while
fluorescent lights will register at
200-300mm. Wiring in walls may be
masked by metal conduits, metal
plates in the wall or door frames, or
uneven conductivity in the wall
material or surface coating;
For these reasons, the unit
JULY 1989
45
~
I
•
,.. jl
-
lo~~--
L
_J
Fig.5: the PC pattern is designed to accept three different meter types and virtually any size trimpot.
The meter here is pointing in the direction of the mains switch wiring. Note
that misleading results will occur if the JFETs are overloaded by strong hum
fields. In most cases, holding the meter case between outstretched finger tips
will give the best results.
46
SILICON CHIP
should only be used to confirm the
presence of mains wiring, not its
absence.
Static charges on surfaces may
also present a problem in dry
areas. Try this: wave a plastic bag
or polystyrene "biro" near the electrodes. Even if these objects have
just been sitting around, picking
them up will probably charge them
to several thousand volts.
The RC time constant of the JFET
gate circuit is determined by the
board and junction leakage and is
of the order of several seconds.
This means that you should move
the detector very slowly when in
the presence of large accumulations of static. Some improvement
can be obtained by fitting a resistor
between the gate and source of
each JFET but, unfortunately, to
maintain full sensitivity at 50Hz,
the resistor value needs to be
around 2-10 gigohms. This is why
we elected to use the board leakage
instead.
By the way, it will be necessary
to minimise this leakage by removing the solder flux around the gate
connections of the JFETs. This can
be done by cleaning the board with
a cotton bud moistened with
methylated spirits.
[§;I
|