Presented at The Society for Psychical Research Centenary Conference/Parapsychological
Association Jubilee Convention, held at Trinity College, Cambridge,
16 - 21 August 1982.
PARANORMAL METALBENDING, RESEARCHES WITH PIEZOELECTRIC SENSORS
J.B. Hasted, D. Robertson and P. Arathoon
Department of Physics Birkbeck College (University of London)
Malet Street London WC1E 7HX
Introduction
It has been our experience(1-6) that by far the most suitable
experimental methods of investigating paranormal metalbending
are those in which the detection of psychokinetic effects at
a metal target is achieved without the target being touched by
the subject to whom it is exposed. Clearly such an exposure requires
careful observation of the subject during the entire experimental
session. No-touch protocol is also required in psychokinetic
experiments with random event generators, but in paranormal metalbending
a simple physical target, as opposed to a 'black box' piece of
electronic circuitry, is exposed to the perception of both subject
and experimenter.
The simplest physical quantity whose paranormal variation can
be observed is strain, or physical deformation of the otherwise
stable metal target. This can be either temporary (dynamic, or
elastic) or permanent (beyond a yield point). Visual observation
of strain is insensitive, even more in the case of elongation
than in that of bending. However, various transducers are available
by means of which even micro-strains, proportional changes of
dimension delta l / l ~= 10^-6, can be detected as electrical
signals, or voltages. The simplest of these is the resistive
strain gauge, with which much paranormal metalbending information
has been obtained.
Another transducer suitable for such observations is piezoelectric
material, particularly piezoelectric ceramics such as lead zirconate
titanate (PZT), which are used as gramophone pick-ups and in
other applications. These have very much greater sensitivity
than the resistive strain gauge, strains as small as delta l
/ l ~= 10^-10 being readily detectable. The background noise
from a resistive strain gauge, with its attendant bridge, amplifier
and chart recorder, is electrical in origin, whilst the acoustic
noise is much smaller. But with a piezoelectric sensor, amplifier
and chart recorder, the acoustic sensitivity is so high that
the background noise can be dominated by acoustic rather than
electrical contributions. On these grounds the piezoelectric sensor
is superior to the resistive strain gauge for the detection of
paranormal metalbending signals.
This paper describes the experiments we have carried out during
the past year on the detection of paranormal metalbending signals
in the presence of metalbending subjects Stephen North, Heloise
Gr. and Willie G.
Properties of Piezoelectric Sensors
In a piece of piezoelectric material the inherent electric charges
are separated when mechanical stress is applied to produce a
strain. The material is a good insulator, but has metal electrodes
covering two opposing faces, and when the material is first prepared
commercially, a high voltage is temporarily applied between these
electrodes, so that the material becomes 'poled'. On the subsequent
application of stress parallel to these electrodes, a difference
of electric charge is produced between the electrodes, this difference
being proportional to the stress; the constant of proportionality
is the appropriate 'piezoelectric constant' of the material.
In the simplest arrangement the polarization, P, which is essentially
the difference of charge, is given by
P = dX + YE
where Y is the mechanical stress and d the piezoelectric constant;
X is the dielectric susceptibility and E the electric field.
He need not at this stage consider the tensor calculus which
describes the array of such constants more rigorously.
The resistance of the piezoelectric material is almost infinitely
high, so that the charge does not leak away, and can be taken
to be approximately proportional to the stress. It might be measured
on an electrometer amplifier, whose resistance is also almost
infinitely high. Unlike the resistive strain gauge, the piezoelectric
strain gauge is a very high impedance device.
Unfortunately this makes it almost useless for detection of paranormal
metalbending signals without modification. Even with some electrical
screening an electrometer amplifier input terminal is enormously
sensitive to stray electric charge, arising from triboelectric
effects, atmospheric ions, and even paranormal (otherwise inexplicable)
electrical effects(5,6). Our experience both in thirty years
of orthodox physics and seven years of research into psychokinesis
suggests that exposed targets connected to high impedance electrometer
amplifiers would not with certitude remain electrically stable.
We therefore virtually short-circuit the piezoelectric, whose
capacitance is C, by connecting it across a low resistance (R
= 3.5 kOhms). In this way the charge leaks away very rapidly,
so that an amplifier and slow response chart recorder only records
a very small fraction of it; most of the sensitivity is lost;
however, there is ample to spare. The time constant for the leakage
is RC, 3.5 kOhms x 500 pF = 1.75 microseconds. Owing to the slow
response (O.1 s) chart recorder, we therefore lose a factor of
about a hundred thousand in sensitivity. Moreover, since the
voltage across R is recorded by the Amplifier, it is the current
through R which determines the signal; the current is proportional
to the first differential of the stress with respect to time,
that is, to the rate of flow of charge. It follows that a rapid
elastic dynamic strain, such as a sharp tap on the specimen with
a pencil, is recorded not as a single 'peak', which would be
the case for a resistive strain gauge, but as a peak in one direction
immediately followed by a peak in the other direction. The contrast
between up-and-down piezosensor signals and unidirectional strain
gauge signals obtained paranormally on a single metal specimen
is shown in the tracing of chart-records in Figure 1.
The hanging of a weight on the specimen for a period of time is
recorded as an up-and-down signal occurring at the moment of
hanging; there is no deflection while the weight is in place,
and an up-and-down signal in the other direction at the moment
when the weight is removed. This contrasts with the record obtained
from a resistive strain gauge, which appears ideally as a square
pulse whose duration is the period during which the weight is
hung.
However, a single peak from a piezo-sensor on the chart recorder
trace does not usually correspond to a permanent deformation,
but rather to a sullen elastic deformation immediately damped,
so that the specimen returns to its original form sufficiently
slowly for the reverse signal to be too small for recording.
_
With this arrangement of piezoelectric the sensitivity is not
to be defined in terms of signal per unit stress, but rather
in terms of signal per unit rate of variation of stress. Shocks
produced by falling weights show sufficiently fast variation
for them to be recorded rather more sensitively than they are
with a resistive strain gauge. Falling weight calibration of our
own system is as follows: 5 g falling through 10 mm onto the
end of the piezoelectric produces an 8 mV chart recorded signal.
The signal across the 3.5 kOhm resistor is amplified by TD 1034A
operational amplifier of gain 100, before chart recording. With
this arrangement, the manufacturer's static calibrations cannot
be used, although it can be shown that our experiments are consistent
with them. For their piezoelectric strain gauge BSG/1-1 (length
9.5 mm, width 1.6 mm, thickness 0.3 mm) Messrs. Vernitron give
a figure of 2.2 x 10^5 V / strain and capacitance 500 pF. For
the ceramic PZT5A the appropriate piezoelectric constant is given
as 2 x 10^-2 V/m for unit stress 1 N/m^2.
The piezoelectric is mounted on or in the metal specimen, together
with its electrical connections, using epoxy-resin. It is then
further insulated with epoxy-resin, and electrically screened
with layers of conducting paint, which connects to the metal
itself and to the screening sheath of the electrical leads; variation
of the capacitance of the latter produces signals, so that they
must be carefully hung and not touched, and kept as short as
possible. For comparison of piezoelectric with a resistive strain
gauge, the latter is mounted on the back face of the metal, opposite
the piezoelectric.
Another experimental method has been to expose to the subject
a larger block of piezoelectric ceramic (50 mm x 7 mm x 7 mm)
without attaching it to a metal specimen. Electrical screening
is still important, and this is conveniently done with solid
plastic coat covered with conducting paint.
Environment and Protocol of Sessions
Environment and protocol have been determined by the following
aims of the experiments:
1. To test, and learn about the behaviour of, piezoelectric sensors
exposed to metalbending subjects.
2. To learn about the psychological conditions under which metalbending
action is most easily produced.
3. To make comparisons between piezoelectric sensors and resistive
strain gauge sensors.
4. To test the accuracy with which a subject can forecast the
production of a metalbending signal.
The environment of the experimental sessions with Stephen North
and with Willie G. has been an isolated basement physics laboratory
in Birkbeck College. Both subjects have previous experience of
visiting the laboratory
for experimental sessions. The sessions with Heloise Gr. were
carried out in the living room of the family house in a West
Country town, since these were the first experiments in which
the subject had participated. All sessions were held in the afternoon.
Within the basement physics laboratory an electrically screened
room of dimensions 2.25 m x 1.6 m x 2.2 m high was constructed
of 0.7 mm aluminium sheet. Care was taken to make good electrical
contact between the individual metal sheets and especially to
the metal sheet which formed the floor. A door 0.9 m, 2 m high
was constructed of expanded aluminium with diamond-shaped apertures
of axes 30 mm and 15 mm. Tests were made of the attenuation of
commercial radio signals with door open and door closed. Since
the screened room was in a basement laboratory and the door faced
solid ground, the attenuation was only slightly affected by closing
the door. This enabled us to conduct the exposure of the specimens
to the subject while he was sealed inside the room and observed
through the expanded aluminium door, either open or closed. A
plan view of the arrangement of subject, specimens, apparatus
and experimenters is shown in Figure 2.
The protocol of the experiment is as follows:
(i) The equipment is allowed to run in the absence of the subject
for a period of about 30 minutes. An experimenter must be present
for part of this time and must enter the Screened room and sit
in the subject's chair. No signals must be recorded during any
of this period. Any drifting exhibited in the chart record is
adjusted by the chart recorder zeroing control.
(ii) The equipment must contain a 'dummy charnel'(3). This is
an unscreened input resistance exposed within the screened room
and connected to a separate amplifier channel ant chart recorder
channel. All amplifiers are battery-operated, but it is possible,
at least outside the screened room, for electrical mains pulses
to reach the circuitry despite screening. Such a pulse would
show up on both dummy and sensor channels, so that such an appearance
cannot be mistaken for a paranormal pulse. Inside the screened
room, however, we have never observed pulses on the dummy channel.
(iii) With the equipment in a stable condition, and no pulses
previously observed, the subject is earthed and introduced into
8 metal chair in the screened room. The instructions previously
given to him are to see if he can interact with the target or
targets he sees in front of him, thus producing signals on the
chart recorders; he is told that he must not touch the target
or electrical leads, and that he should report on inadvertent
touch; no other instructions about what he does with his body
are given to him. No violent movements have taken place, but
a few inadvertent touches have been reported.
It is the responsibility of the senior experimenter to terminate
the session when he thinks that further work would not be profitable.
Clearly the factors which influence such a decision are tiredness,
success rate in producing signals, boredom, condition of the
equipment, including possible permanent bends on the specimen,
electrical failure, or termination of chart paper roll.
It is of course important that the experimenters observe the subject,
and especially his hands and feet, during the time he is within
reach of the specimens and leads. The mounting of the specimens
is by elastic suspension from a stand which is sufficiently rigid
and heavy for no signal to be produced if the stand is touched
or jogged.
Obviously we cannot expect a completely watertight observation
of a subject by two experimenters for a period of, say, thirty
minutes. The eye can wander momentarily unless great concentration
is exerted, and if there is such continual concentration, the
intense atmosphere is not conducive to relaxed experimentation.
However, we do not believe that more than a very small proportion
of the signals could be due to touch, since we are confident that
our visual observation would detect a proportion of such events.
Since the walls of the screened room are of smooth polished aluminium,
it is possible to watch carefully without appearing to do so.
We have at no time seen unequivocal cheating by touch in this
series of experiments, but we have become sensitive to the conditions
when this might occur. Willie G. and Heloise Gr. have not had
sufficient experience with these experiments to reaise exactly
how easy it might be to produce signals by touch, and they have
shown no desire to do so. Stephen North admits what we have known
for some time, which is that when bored or frustrated he feels
tempted to touch, in order to stimulate further signals; but
because he can hear the chart recorder pen, he knows that signals
occur when he is not touching. It is our continual responsibility
to avoid the onset of feelings of temptation, and to distract
him or terminate the session temporarily should we suspect such
an onset. Sessions with Stephen North are often divided into
comparatively short periods (5 - 10 minutes), during which we
can be confident of the absence of touch.
(iv) The presence of other observers is permitted, and records
of these have been kept. Observers are not introduced before
the subject is seated; they must sit in chairs placed behind
the experimenters, although the view that they have of the subject
is still a good one. They are permitted to take part in the conversation,
which is not recorded. Prior to the session, the observers are
informed that touch of the specimens and electrical leads by
the subject is not permitted, and are asked to report subsequently,
making criticisms or suggestions about the session.
Psychological Factors in Sessions
Since the subject can hear the movement of the chart recorder
pens, and sometimes can even see them move, the precise moments
at which signal pulses occur can become known to him; he can
with experience use these as biofeedback to assist him to obtain
further signals if he is motivated to do so. The experimenters
are interested in establishing the ability to obtain signals
reliably from all subjects who have demonstrated any ability at
all. All subjects from whom we have obtained signals have claimed
to have had private experiences of a psychokinetic character,
usually the bending of metal in the home, and occasionally the
unusual behaviour of wristwatches. It is possible to some extent
to obtain the benefits of notoriety by demonstrating the bending
of cutlery by stroking, but such benefits do not to any great
extent follow from the ability to produce signals on instruments.
The attitudes of Willie G. and Heloise Gr. are not appreciably
influenced by the search for notoriety, but Stephen North has
already been exposed to publicity, 60 that his attitude must be
carefully assessed. The experimenters have had regular sessions
with Stephen North over a period of years and know him well.
His ambitions over the past year have been to establish himself
in a career, or at least in work, after leaving school, and to
maintain a rewarding relationship with his girlfriend; for most
of the gear he has been unsuccessful in finding work, but in
December 1981 he started work in a recording studio, and in this
work he finds much fulfillment. During most of the sessions,
therefore, he was out of work and badly needed encouragement.
For some sessions he arrived preoccupied, and success in obtaining
signals eluded him. The experimenters believed that stimulation
was necessary, and arranged visits from suitable stimulating.
people; this produced results. Regularity of sessions, held on
a specific afternoon of the week, was found to be advantageous.
Previous experience has suggested that signal R occur at sudden
changes of psychological state ('gaps between thoughts')(4).
The first of such changes to be recognized was the sudden lapse
in concentration, brought about, for example, by a suggestion
to break for tea. This technique is still successful, and we
have recorded recent examples of it; but it must be remembered
that a break of concentration implies that there was real concentration
to start with. Thus the notion that metalbending signals are
entirely spontaneous is incorrect.
It is important that there should be short periods of concentration,
and these may indeed produce signals. But this success cannot
usually last for long. After a period during which a succession
of signals was obtained, a quiet period would follow. We have
attempted analysis of the time intervals between signals in order
to test the validity of this claim. If we make a somewhat arbitrary
criterion that a gap of longer than 3 minutes will be considered
to be a 'quiet period'; the numbers of 'active periods' which
alternate with quiet periods are summarized in Table 1 for the
sessions recorded.
A more general example of sudden change of psychological state
is that induced by a sudden surprise. However, the subject's
emotions generated by a surprise may not always he known to the
experimenters, or even to the subject himself. We have not made
an analysis of the success rate of the surprises we have deliberately
introduced into sessions, but some are known to have been successful
Another technique for producing signals is to find a stimulus
to try something new. "See if you can do it when the specimen
is under water"; "See if you can count one, two, three,
go, and produce a signal"; "See if you can do it while
holding your breath" and so on. The subject must not realise
that these challenges are usually as much for the sake of the
stimulus as for the sake of the experiment.
Comparison of Effectiveness of Piezoelectric and Resistive Strain
Gauge Sensors
Experiments have been carried out with piezoelectric
sensor and a resistive strain gauge mounted 'back to back' -
i.e. on opposite sides of a thin metal specimen (0.8 mm x 12
mm x 150 mm). The specimen was then exposed radially from the
subject, hanging horizontally. In previous experiments of this
type, with two resistive strain gauges of equal sensitivity(7)
no right-left asymmetry was found with Stephen North. Nevertheless,
alternate sessions were held with piezoelectric on the left and
on the right, in case some asymmetry of Stephen's action had
developed subsequently.
The purpose of these experiments was to compare the effectiveness
of our two devices as detectors of metalbending action. As has
been explained above, the physical sensitivities are not directly
comparable, in that the piezoelectric sensitivity depends upon
the time variation of the metalbending 'action'; the limits of
the time variations are not known accurately. The sensitivity
of the piezoelectric sensor could be greatly increased by raising
the value of the shunt resistance; but the equipment might then
be less secure against electrostatic artefacts. Therefore the
best procedure is to compare the effectiveness of the two devices,
as we have developed them, in terms of the relative numbers and
magnitudes of signals detected on each.
In the first session (June 4 with Stephen North), similar numbers
of signals were obtained in either channel: 7 on strain gauge
only, 13 on piezo- sensor only and 4 synchronous on both channels.
But in subsequent sessions, with the piezo-sensor channel operating
with the higher sensitivity available when the electrical noise
is reduced by improvement of the electronics, essentially all
signals were recorded in the piezo-sensor channel, as will be
seen from Table 2.
The superior sensitivity of the piezo-sensor to rapid shocks was
demonstrated by mechanical tests. When the metal specimen was
manually touched, signals were recorded in both channels. But
when the specimen was tapped sufficiently lightly with 8 pencil,
signals of several millivolts could be obtained in the piezo-sensor
channel without any signal being recorded in the strain gauge
channel. This demonstrated the response of the piezo-sensor to
time variation of stress, in contrast to the response of the
strain gauge to stress or strain. The paranormal signals were
more rapidly time-varying than those produced by human touch;
and it is possible that since paranormal metalbending action
appears to occur in rapid bursts with millisecond rates of rise,
much data may have been missed in the early sessions described
in previous publications, using strain gauges only.
Some sessions were held with three separate sensors exposed equidistant
from the subject, in horizontal line with 20 cm spacing: on his
left, metal with strain gauge; centrally, a block of piezoelectric
ceramic; and on his right, metal with a piezoelectric sensor.
It is difficult to draw conclusions from the findings presented
in Table 3 about the psychological attractiveness to Stephen
North of the three targets.
Timing Control of the Production of Signals
The purpose of timing control experiments is to determine with
what accuracy the subject can predict the arrival time of a signal.
Stephen North is the only subject who has taken part in these
experiments. They have only been undertaken when Stephen was,
in the opinion of the experimenters, already producing sufficient
signals to encourage the belief that he would succeed with a
timing control experiment. Stephen was then asked to say, "One,
two, three, go", and attempt to produce a signal on the
word "go"; he could do this however frequently or infrequently
he liked, and could choose the times when he did it.
An experimenter marks the chart paper, either with an electrical
marker signal, or in the first experiments with a pencil, when
he 'hears the word "go"'. Since Stephen counts slower
than one count per second, and counts in even rhythm, it is possible
to mark the paper within 0.3 s accuracy; this has been proved
in a subsidiary experiment, using acoustic recording on a parallel
channel. A standard deviation of 0.271 s was obtained over 25
attempts. Multiple channel chart recording has been used throughout.
It is surprising how successful Stephen has been in producing
a signal under such protocol. Typical success rates are demonstrated
in Figures 3, 4, 5. Signals more than 60 s late are regarded
as 'spontaneous', and not as late arrivals intended for a target
time. A number of other spontaneous signals, unaccompanied by
"One, two, three, go", are also found.
Signals occurring more than about four seconds prior to "go"
are of course open to the objection that Stephen might hear the
pen move, and immediately start to say, "One, two, three,
go", so as to claim a certain measure of success. There
have, however, been no such signals, other than signals designated
'spontaneous', and after these Stephen did not immediately attempt
to count.
The graphical method of displaying the success of an experimental
session enables the subject, as well as other students of psychokinesis,
to assess the important features of each session in terms of
the following questions: does the success improve as the session
continues? How long are the periods over which success can be
maintained? Are the signals predominantly early or late? How
frequent are the very good and the very bad-attempts, and the
failures?
The extraordinary success in producing signals early, but within
one second (i.e. on the count of "three" rather than
on "go") is the most surprising feature of these experiments.
References
1. J.B. Hasted, J. Soc. Psych. Res., 48 (1976), 365-383.
2. J.B. Hasted, J. Soc. Psych. Res., 49 (1977), 583-607.
3. J.B. Hasted and D. Robertson, J. Soc. Psych. Res., 50 (1979), 9-20.
4. J.B. Hasted ant D. Robertson, J. Soc. Psych. Res., 50 (1980), 379-398.
5. J.B. Hasted and D. Robertson, J. Soc. Psych. Res., 51 (1981), 75-86.
6. J.B. Hasted and D. Robertson, Psychoenergetic Systems, 4 (1981), 169- 18
7. J.B. Hasted, Zeit. Parapsych. und Grenz. Psych., 20 (1979), 173.
Table 1
| Date (1981) | Subject | Length of active period (min) |
Number of signals in active period
|
| 4 June | SN | 10.0 | 22
|
| 18 June | SN | 3.9 | 7
|
| | 2.3 | 4
|
| | 4.5 | 10
|
| | 5.0 | 32
|
| 25 June | SN & WG | 6.8 | 5
|
| | 4.5 | 4
|
| 5 Aug | HGr | 7.8 | 10
|
| | 8.4 | 11
|
| 5 Aug | HGr | 3.9 | 8
|
| | 10.7 | 8
|
| 12 Aug | SN | 4.6 | 11
|
| | 8.2 | 8
|
| | 1.2 | 5
|
Table 2
| Date (1981) | Subject | Numbers of signals
|
| | Piezo-sensor | Strain Gauge
|
| 18 June | SN | 45 | 0
|
| 25 June | SN & WG | 12 | 1
|
| 5 Aug | SN | 3 | 0
|
| 7 Aug | HGr (1) | 21 | 0
|
| HGr (2) | 16 | 0
|
Table 3
| Channel | Number of signals
|
| August 12 | September 2 | September 11
|
| Strain gauge | 0 | 8 | 2
|
| Piezo block | 27 | 0 | 21
|
| Piezo sensor | 0 | 0 | 1
|
| Strain gauge and block | 0 | 0 | 10
|
| Strain gauge and sensor | 0 | 14 | 14
|
| Block and sensor | 0 | 0 | 3
|
| Strain gauge, block and sensor | 0 | 0 | 4
|
Figure Captions
1. Parts of Chart Record of no-touch session held on 11 Sept.,
1981, with Stephen North. Subscript 1 represents strain gauge
records and subscript 2 represents piezoelectric sensor records.
Picture 1
Picture 2
2. Plan view of electrically screened room and apparatus. T. t,
tables. S. subject. Sp, specimen. L, electrical leads. E, experimenters.
O. observers. A, amplifiers. C, chart-recorders.
3, 4 and 5. Time differences delta t between signals and predicted
signals (i.e. "Go" in "One, two, three, go")
during 30 October, 13 November and 19 November sessions with
Stephen North. Vertical broken lines represent failures, broken
line circles represent spontaneous signals.
Back to English research in psychokinesis
Back to main index
Back to Uri Geller's home page.
