Note: Descriptions are shown in the official language in which they were submitted.
~:f~~7~i''~9'a
Description
In-Line Metallic Debris
Particle Detection System
Technical Field
The present invention relates to particle
detection in general, and more particularly to
arrangements for detecting metallic particles carried
by a fluid, especially lubricating oil or fuel for gas
turbine engine applications.
Background Art
There are already known various constructions of
detecting arrangements capable of detecting the
presence of metallic particles in a flow of a fluid.
The traditional approach to detecting the presence of
metallic particles in flowing fluid systems has been
the use of magnetic plugs that are exposed to the
fluid flow and attract and capture ferromagnetic fluid
borne particles. In the simplest embodiment of this
approach, the plug must be periodically removed from
the system and inspected for the presence of
ferromagnetic debris thereon. Presence of material of
this type in the flowing fluid has been found to be
correlated with the overall health of the system and,
therefore, the number and sizes of the ferromagnetic
particles that have been scavenged from the fluid and
magnetically captured by the plug over a given period
of time provide an indication of such health. Tn
F2-3150As-
- 1 -
~.(~~lEi ~9Ea
particular, this approach provides information on the
degradation of bearing and/or gear components prior to
the onset of occurrence of catastrophic failure of
such components.
While the magnetic plug approach has been found
to provide the capability of detecting debris
particles in flowing fluid systems, it provides only a
partial and less than satisfactory solution to the
system health monitoring problem. Specifically, the
magnetic plug must be removed from the system and
visually inspected in order to detect ferromagnetic
debris buildup. This has the significant disadvantage
that potential component distress problems occurring
between two successive magnetic plug inspections may
go completely undetected until it may be too late,
with substantial or even catastrophic failure of the
affected component taking place prior to the time
scheduled for the next inspection. Secondly, in order
to perform the inspection of the magnetic plug, it is
necessary to disturb the integrity of the system being
monitared. This is known to have had catastrophic
consequences in several cases. Additionally, the
magnetic capture efficiency of the plug is less than
100 and it varies with debris particle size and is
affected by flow velocity. Moreover, since a magnetic
capture approach is being used, only ferromagnetic
debris can be detected.
To avoid at least some of the aforementioned
drawbacks, further developments have been pursued in
this area to improve the debris particle detection in
fluid flows, resulting in advanced magnetic plug
designs. One of such developments has been to provide
the magnetic plugs with open electrical gaps that are
- 2 -
M()~lf~'~9E;~
electrically bridged and shorted following the buildup
of a significant amount of the debris material on such
plugs. This was intended to reduce or even eliminate
the need for visual inspection and to provide
continuous monitoring of the health of the fluid
system. While constituting an advancement over the
original magnetic plug detectors, such electrical gap
systems still suffer from certain major shortcomings.
Notably, since such systems still employ magnetic
capture, they axe capable of detecting only
ferromagnetic debris. A further, and an even greater,
drawback of such systems is that the magnetic plugs
preferentially scavenge small submicron debris
material from the flowing fluid, due to the lower
momentum or inertia associated with small particles.
Yet, the submicron debris is not symptomatic of an
unhealthy situation or of component distress; rather,
it is a naturally occurring phenomenon associated with
normal wear encountered in any mechanical system in
which components rotate or are otherwise displaced
relative to one another in physical contact with each
other. As a consequence, the magnetic plug systems
employing the electric gap shorting concept for
detection are prone to extremely high false alarm
rates. So serious has this latter problem been found
to be that it is currently a common practice not to
electrically connect the electric gap arrangement,
thus forfeiting its advantages, and to revert once
more to reliance merely on the magnetic capture
feature coupled with visual inspection.
Yet, the value of monitoring the health of
equipment employing flowing fluid lubrication and/or
fuel supply systems on the basis of the characteristic
- 3 -
Mt.)~D~)'~9
properties of the debris carried by such fluids has
been clearly recognized industry-wide. This
recognition has spurred interest in the development of
inductive debris detecting systems for this purpose.
The advantages of the inductive detection approach
include a continuous in-line monitoring capability and
the ability, in principle, of detecting not only
ferromagnetic, but also non-ferromagnetic, metallic
particles. While inductive debris detecting systems
using multiturn coils. surrounding a passage through
which a fluid to be monitored flows and supplied with
alternating current represent an advance over the
aforementioned approach employing magnetic plug
detectors, they still possess a fundamental
shortcoming, namely that, because of the nature of the
coil design employed therein, the coil and its
associated,resonant bridge circuitry are sensitive not
only to the eddy current flow produced in the debris
particle, which is the fundamental mechanism utilized
in inductive debris particle detection, but also to
changes in the dielectric constant of the fluid
flowing through the detection region of the coil.
This sensitivity to the dielectric constant of the
fluid and to changes therein results in a high rate of
false indications. A reason.for this is that what is
commonly found in oil lubrication systems is not only
oil and debris particles carried thereby but also foam
and small and even large entrapped air bubbles.
Typical extent of entrapment of large bubbles (i.e,
bubbles having diameters equal to the flow tube
diameter) in the flow can amaunt to upwards of 50% of
the total flow. Furthermore, finely suspended dirt
clouds and other substances, such as water, are also
- 4 -
~EDE)E~'~;~f
frequently found in lubrication system fluid flows.
The presence of any or all of these types of
inclusions in the flowing. fluid leads to large changes
in the dielectric constant.of the fluid which, in
turn, are perceived by the inductive detecting systems
sensitive thereto and thereby adversely affect the
accuracy of the indications provided by such systems.
An attempt to~ avoid this problem caused by
entrapped air bubbles and other dielectric inclusions
is presented in the British patent No. 1,348,881,
where a radio frequency bridge circuit is employed in
conjunction with a pair of multi-turn induction coils
for detecting the presence of fluid borne debris
particles. In the system of the above patent, the
flow of the fluid is split into two parallel paths
each passing through one of the induction coils. The
theory behind this approach was that large air
bubbles, for instance, would be evenly split between
the two parallel.flow paths, thus leading to a signal
of a similar magnitude in each of the coils; in
contrast, a metallic fluid-borne debris particle would
only be carried in one of the paths, thus resulting in
a signal only from the coil surrounding this
particular path. This then was supposed to provide a
mechanism for differentiating between these two
effects, that is that of the entrapped air bubbles, on
the one hand, and that of the metallic debris
particles, on the other hand. However, experience has
shown that, the nature of things being what it is, the
air bubbles and similar dielectric inclusions are not
evenly split between the two paths in the real world,
which results in a signal level in one leg of the
bridge circuit exceeding that in the other leg even if
_ g -
~~~J~~'"~9E~
only dielectric inclusions and no metallic particles
pass through either one of the parallel paths, thus
giving a false indication indistinguishable from that
attending the passage of a metallic particle through
this detection system.
Other attempts have been made as well to
eliminate the air bubble sensitivity problem by
employing pairs of split parallel path coils spaced in
the flow direction. Notwithstanding the complexity of
the arrangements of this type, even this approach has
not resulted in an unqualified success, as evidenced
by the lack of acceptance of arrangements of this type
in the marketplade.
A subsequent development in this field is
reflected in the British patent application No.
2,101,330 A, published on January 12, 1983, which
discloses a system for detecting particles in flowing
fluids utilizing two inductive coils that are spaced
from one another along a section of the path of flow
of the fluid and each of which surrounds a portion of
this detecting path section. As ferromagnetic
particles and other inclusions entrained in the fluid
pass through the detecting path section, they cause
changes in the electrical impedance of such coils and
these changes are then evaluated> The coils and the
evaluating circuitry together constitute a detector
arrangement which is supposed, in principle, to
eliminate the false signal indications generated in
response to the passage of dielectric non-uniformities
through the detecting path section, by limiting the
detection capability of the system to a selected phase
angle thereby to screen out the influence of gas
bubble discontinuities. Unfortunately, this also led
- 6 -
CA 02006796 1999-03-30
to the loss of capability on the part of this system
to detect non-ferromagnetic particles, thus rendering
this system, despite its high complexity, not
significantly better than those employing the magnetic
plug approach. Moreover, in practice, the inductive
debris detecting system of this type still suffers
from a high false-alarm rate because the gas bubbles
do not generate exactly the same response in each of
the two coils. This is at least partially
~o attributable to the pulsating nature of the flows
typically encountered in lubricating systems.
A current adaptation of this latter approach
is the Sensys/Ferroscan (a trademark of Atomic Energy
of Canada Limited) technology; however, it also
suffers of the disadvantages discussed just above.
Other developments have resulted in new approaches to
magnetic capture systems. The most noteworthy of
these approaches is that employed in the quantitative
debris monitor system manufactured by the Tedeco
2o Corporation; however, while this system employs a
variation of the magnetic capture technique, it still
has the drawback of reacting to gas bubbles because
its detection means is also sensitive to dielectric
constant changes in the flowing fluid background
medium. To deal with this problem a new air bubble
separator system (marketed by the Tedeco Corporation
under the trade name or trademark Lubriclone) has been
developed, as described in an article by F. DiPasquale
entitled "Field Experience with Quantitative Debris
3o Monitoring" (SAE Paper No. 871736, October 5-8, 1987),
to remove air bubbles from the fluid flowing to the
quantitative debris monitor of the above type. Yet,
the complexity, size penalty, cost penalty, and low
~:~~lf~'79~
effectiveness of such an approach more than offset its
benefits.
Accordingly, it is a general object of the
present invention to provide an inductive type debris
detection and/or monitoring system which avoids the
disadvantages of the previously proposed systems of
both the magnetic, and the inductive type.
More particularly, it is an object of the present
invention to develop ari inductive particle detection
system which is insensitive to temperature changes, to
changes in the dielectric properties of the flowing
medium, and to overall system vibrations.
Still another object of the present invention is
to design the inductive particle.detection system of
the type here under consideration in such a manner as
to be able to determine, on an individual particle
basis, both the size and the magnetic or non-magnetic
character of fluid-borne metallic debris particles.
It is yet another object of the present invention
to devise an inductive particle detection system of
the above type which is non-intrusive to the flow
system, thus causing no additional flow resistance.
Yet another object of the.present invention is to
provide an inductive debris particle detection system
that is insensitive to naturally occurring submicron
wear debris particles.
A concomitant object of the present invention is
to develop an inductive debris monitoring and/or
detection system arrangement of the above type which
is relatively simple in construction, inexpensive to
manufacture, easy to use, and yet reliable in
operation.
_ g
~:~~~fi'79fi
Disclosure of the Invention
Brief Description of the Drawing
The present invention will be described in more
detail below with reference to the accompanying
drawing in which:
Figure 1 is a perspective view of an arrangement
of the present invention for detecting the presence
and character of metallic particles in a flowing
fluid;
Figure 2 is a graph~depicting the response of the
arrangement of Figure 1 to.various metallic and
non-metallic inclusions entrained in the fluid flowing
therethrough; arid
Figure 3 is a~simplified diagrammatic view of a
circuit of the present invention capable~of
determining the character of~any entrained metallic
particle from the response of the arrangement of
Figure 1. '
Best Mode for Carrying Out the Invention
Referring now to the drawing in detail, and first
to Figure 1 thereof, it may be seen that the reference
numeral 10 has been used therein to identify a tubular
probe housing or pipe section. The pipe section 10,
which is of a non=conducting unity magnetic
permeability material, bounds a passage 11 for the
flow therethrough of a fluid that is to be examined
for the presence therein of various inclusions, such
as magnetic and non-magnetic metallic particles.
A probe member 12 of a highly electrically
conductive material, such as copper, is arranged in
such a manner as to be stationary relative to the pipe
- 9 -
~a.D~f~'79f~
and to circumferentially surround the passage li.
For instance, the probe.member L2 may be embedded or
potted in, or arranged around, the pipe section 10.
The probe member 12 is constituted by a single turn
coil and it has respective marginal portions 13 and 14
which bound a gap 15 with one another. The probe
member 12 should have an axial length to diameter
ratio greater than one. So, for instance, the axial
length of the probe 12 may be 1-1/8" and its diameter
about .71", and it may be made of 3 mil thick copper
sheet. It will be appreciated, though, that the above
dimensions, while they have been carefully chosen for
a particular construction of the detecting arrangement
of the present invention, may be altered without
departing from the present invention, so long as the
altered dimensions satisfy the~operating criteria that
will be discussed below.
In the probe member construction illustrated in
Figure l, the marginal portions 13 and 14 overlap one
another, and a capacitor arrangement 16 is interposed
in the gap 15 that is situated between the overlapping
regions of the marginal portions 13 and 14. The
capacitor arrangement 16 may include merely a layer or
slab of dielectric material; in which case the
overlapping regions of the marginal portions 13 and 14
constitute respective capacitor plates. However, more
often than not, the surface areas of the overlapping
regions of the marginal portions~l3 and 14 are
insufficient to provide the required capacitance. In
such a case, in accordance with the present invention,
the capacitor arrangement 16 may be constituted by a
single multilayer.capacitor device, or preferably by a
number of such multilayer devices which are
-10 -
:ED~Ei'7~~fi
distributed at predetermined, such as substantially
identical, intervals along.the gap 15 between the
overlapping regions of the marginal portions 13 and 14
of the probe member 12. In any event, the capacitor
arrangement 15 is situated at the gap 15 and is
isolated from the passage 11.
As further shown in Figure 1 of the drawing, the
marginal portions 13 and 14 have respective electric
leads 17 and 18 connected to them. The electric leads
17 and 18 serve to supply alternating electric current
to the marginal portions l3 and 14. When this occurs,
the probe member 12 forms a parallel tank circuit with
the capacitor arrangement 16.
In accordance with the present invention, the
capacitance of the capacitor arrangement 16 has a
relatively nigh value. The value of this capacitance
is chosen in such a manner with respect to the
inductance of the probe member 12 as to achieve an
inductance to capacitance ratio on the order of one to
4 or even less. The probe member 12 and the capacitor
arrangement 15 form a resonator that is. operated at
resonance, as will be discussed in more detail later.
The resonance characteristics of this resonator are
influenced by the electromagnetic properties of
inclusions present in the passage 11.
Because of the elongated single turn coil
configuration of the probe member 12 and the location
of the capacitor arrangement ~16 at the gap 15, that
is, as close as physically possible to the marginal
portions 13 and 14, this tank circuit has a high Q
factor. It will be appreciated that important
criteria to be considered when altering the dimensions
of the probe member 12 (and/or the capacitance of the
-11 -
~:(D~)~f~'~;~fa
capacitor arrangement 16) include the preservation of
this high Q factor; the preservation of uniformity of
the magnetic field in the central probe passage
region, and the maintenance of a low inductance to
capacitance ratio.
Two major advantages are obtained when the
resonator or tank circuit constituted by the probe
member 12 and the capacitor arrangement 16 are
constructed in the above-discussed manner for. the
detection of metallic debris, in fluid flow systems.
First, the thus obtained resonator has a very high
quality factor, which means that the effective
electrical impedance of such a resonator is greatly
affected by even relatively small perturbations in the
characteristic properties of the fluid present in the
passage 11, as caused by entrained inclusions. This
high sensitivity renders possible the detection of
minute metallic particles then present in the passage
11.
However, the thus obtained high sensitivity to
the electromagnetic characteristics of inclusions is
not sufficient when it is desired to construct a fully
operational and reliable system for detecting
fluid-borne debris.. This is so because, as will~now
be explained in connection with a lubrication system
with.a liquid lubricant, such as oil, that is being
recirculated, without being limited to this particular
fluid, the contents of any pas$age in a typical
lubrication system varies with time between the
extremes of substantially none of the lubricant to
substantially all lubricant while the system is in
operation. Lubricants typically have a dielectric
constant on the order of three relative to that of
-12 -
~:()~,)f~'~96
air. Thus, the filling and emptying of the passage in
which the debris detection is to take place causes
changes in the passage contents electromagnetic
characteristics, namely the passage contents
dielectric constant and hence the passage electric
displacement field, as affecting the behavior of any
surrounding coil or resonator. In~the inductive
debris detection arrangements of the prior art that
employ, as explained before, conventional multiturn
coil configurations, this kind of fluid level
variation produces coil or resonator performance
changes which inherently result in false indications
of debris presence. hn contradistinction thereto, the
employment of a large fixed and isolated capacitance
in the arrangement constructed in accordance with the
present invention results in the second of the
aforementioned two~advantages,.namely that'the
characteristic behavior of the resonator is only
insignificantly affected by these changes in the
electric displacement field existing in the passage
11. The reduction of this effect in the arrangement
of the present invention is of such a magnitude that a
.0~7" diameter ferromagnetic sphere passing through
the passage 11 having a .75'° inner diameter produces a
signal of a magnitude three times that of a noise
signal produced by completely emptying and~filling the
passage 11 with a typical lubricant.
Thus, it may be seen that the achievement of the
high Q factor means that, when the electric current
supplied to the marginal~portions 13 and 14 through
the electric leads 17 and 18 alternates at such a
frequency that the tank circuit operates at or close
to resonance in the,absence~of any inclusions from the
-13 ~°
~~~'~~ i'~:~~
fluid present in the passage 11, any change in the
characteristic response of the contents of the passage
ll.caused, for instance,. by the presence of metallic
particles in the fluid flowing through the passage 11,
introduces an imbalance into the operation of this
tank circuit in a manner dependent on the
electromagnetic properties.of such inclusions. More
particularly, metallic particles influence the
electromagnetic field generated by the probe member 12
and thus the electric current flowing in the tank
circuit differently, and to a much greater extent,
than dielectric. particles. or other dielectric
inclusions, and non-ferromagnetic metallic particles
influence the electromagnetic field differently than
ferromagnetic metallic particles, resulting in a
different phase shift in each instance, while the
magnitude of the change depends, by and large, on the
size of the respective particle or inclusion. At the
same time, however, the elongated single turn coil
configuration of the probe member 12, coupled with the
low L/C ratio employed in the tank circuit, results in
a situation where the electric displacement field
within the probe member l2~is as low as possible in
relation to the electric field existing in the
isolated, fixed capacitance region, so that air
bubbles which are frequently~encountered in lubricants
will affect the operation of the aforementioned tank
circuit only to an insignificant~extent, if at all.
The phase shift response of the tank circuit that
is constructed in accordance with the present
invention to changes in the electromagnetic properties
of the contents of. the passage 11 is diagrammatically
depicted in Figure 2 of the drawing in which the point
-14 -
~~D~IE~ ~9f
of origin O represents the conditions encountered when
the passage 11 is filled with lubricating oil devoid
of any inclusions. If the dielectric constant of the
fluid present in or flowing through the passage 11
changes, which may occur, for instance, due to
replacement of the original dielectric fluid by
another.dielectri:c fluid, both the relative
resistivity ( R/R) and'the relative impedance ( L/L)
of the overall tank circuit (which includes the fluid
present in the passage 11 in addition to the
aforementioned tank circuit proper that is constituted
by the probe member 12 and the capacitor arrangement
16) change generally to the same relatively small
degree. This is indicated in Figure 2 by the point A
located on a straight line D; the distance OA being
representative of the worst case~scenario involving
complete replacement of lubricating oil by air. It
may be seen that the above distance is rather small.
On the other hand, this.distance would be much greater
in an inductive debris detection system employing a
conventional, multiturn coil:
~,On the other hand, when a ferromagnetic particle
enters the internal passage 11 that is surrounded by
the probe member 12, then both the relative impedance
and the relative resistivity change in dependence on
the size of the particle so as to be located on a
curved.line F which is applicable when the
ferromagnetic particle is substantially spherical. As
an example, point B of the curve F may be reached when
the spherical ferromagnetic particle is about 7 mils
in diameter, and the distance on the curve F from tha
point O will be lesser for smaller and greater for
larger spherical ferromagnetic particles. For other
. -15 _
~~~~~E~ ~9f
shapes of the ferromagnetic particles, other curves
akin to curve F and forming a family therewith apply,
but all such curves are always located in the first
quadrant of the graph depicted in Figure 2. Thus, it
may be seen that the~values located in the first
quadrant are indicative of the ferromagnetic character
of the respective particle, and that the extent of
deviation from the point O~is indicative of the size
of the respective ferromagnetic particle.
In contradistinction thereto, when the particle
entering the internal passage 11 of the probe member
12 is metallic but non-ferromagnetic, the relative
resistivity still changes,in the positive sense, but
the relative impedance changes in the negative sense,
in accordance with the representative curve N of~a
curve family akin to that mentioned above, with all
curves of this family this time being always located
in the fourth quadrant of the Figure 2 graph, and the
distance along the respective curve, such as N, being
again indicative of the. size of the respective
metallic non-ferromagnetic particle. Thus, when it is
determined that the value lies in the fourth quadrant,
then the particle must be metallic and
non-ferromagnetic, while the distance from the point
of origin 0 gives the size of such particle.
A circuit constructed~in accordance with the
present invention to gather and decipher the above
information is presented in Figure 3 of the drawing
where. the same reference numerals as before have been
used to identify corresponding parts (i.e, their
electrical equivalents). The lead 18 from the tank
circuit 12 and 16 is shown to be grounded, while the
lead 17 is connected to one end of one transformer
-16 --
~'(~~i)~~ ~9f
winding 19 of a driving and pickup transformer 20.
The transformer 20.further includes another
transformer winding 21 whose one end is grounded while
the other end thereof, is supplied with an alternating
electric current from a voltage controlled oscillator
(VCO) 22. The alternating electromagnetic field
generated by the other transformer winding 21 induces
a correspondingly alternating electric current in the
one transformer winding 19; and this latter electric
current drives the tank circuit 12 and 16. The
frequency of the alternating electric current issued
by the oscillator VCO is such that the tank circuit 12
and 16 operates at or close to resonance.
The alternating electric current is also supplied
directly to one input of a first mixer 23, and through
a 90° phase shifter 24 to one input of a second mixer
25. A line 26 supplies an alternating electric
currant derived from the one coil 19 and thus
representative of the alternating electric current
flowing through the one coil 19 and thus into and out
of the tank circuit l2.and l6~to a pre-amplifier 27
from where the amplified electric current is supplied
to another input of the first mixer 23, as'well as to
another input of the second mixer 25, where the
respective incoming alternating electric currents are
mixed with one another, with the result that
respective in-phase and quadrature error signals
indicative of the difference between the output
frequency of the VC0 22 and the resonant frequency of
the tank circuit 12 and 16 appear at respective
outputs 28 and 29 of the mixers 23 and 25. These
error signals are then filtered by respective low-pass
-17 -
N~~Y~1
filters 30 and 31 to obtain respective resistive
(in-phase) and reactive (quadrature) error signals.
The reactive error signal is supplied to a
reactive error amplifier 32 which amplifies this
reactive error signal,. and this amplified reactive
error signal is then supplied to an input of the VCO
22 which changes its operating (output} frequency in
dependence on the magnitude of the amplified reactive
error signal. Similarly, the resistive error signal
is fed to an input of a resistive error amplifier 33
which amplifies this resistive error signal, and this
amplified resistive error signal is then supplied to
an input of a voltage controlled resistor (VCR) 34
which is interposed between the other end of the one
transformer winding l9 and. the grcund and whose
resistance varies in dependence on the magnitude of
the amplified resistive: error signal. The resistive
and reactive error amplifiers 33 and 32 are'
constructed to operate with relatively large time
constants, so that the resistance of the VCR 34 and
the frequency.of the'VC0.22.change gradually in
response to relatively long-term changes, especially
those due to temperature variations, of the resonance
characteristics of the tank circuit and/or of the
characteristic properties of the contents of the
passage 11. On the other hand, short-lived changes in
such characteristic properties, such as those caused
by the passage of individual metallic particles
through the interior of the probe member 12, will
leave the performance of the VCO 22 and of the VCR 34
virtually unaffected.
The output signals of the low-pass filters 30 and
31 are also supplied to an evaluating circuit 35 which
-18 -
'~~,'Qf~f:'~9f~
is constructed to evaluate the reactive and resistive
error signals to determine therefrom the character and
size of any metallic particle then present in the
passage 11. A quite simple~exemplary implementation
of the evaluating circuit 35 is shown in Figure 3 of
the drawing, but it is to be.uriderstood that the
evaluating.circuit 35 may have other configurations,
depending on needs, or requirements for accuracy. The
illustrated implementation of the evaluating circuit
35 incorporates a voltage divider 36 and a plurality
of comparators 37a to 37n (n being any arbitrarily
chosen integer) each of which has two inputs one of
which is connected to an associated section of the
voltage divider 36~ while the other input is supplied
with the filtered resistive error'signal appearing at
the output of the low-pass.filter 30. Thus, the
comparators 37a to 37n compare the filtered. resistive
error signal voltage with various reference voltage
levels derived from the voltage divider '36, and that
or those of the comparators 37a to 37n at which the
filtered resistive error voltage exceeds the
respective reference voltage issues an output signal
or.issue respective output.signals which is or are
then supplied to a drive circuit 38 of any known
construction which drives a display 39. Furthermore,
the filtered reactive error signal 'appearing at the
output of the low-pass filter 32 is also supplied to
the drive circuit 38 and is used to drive the display
39 accordingly.
It will be appreciated that, in the construction
of the evaluating circuit 35 depicted in Figure 3, the
drive circuit 38 and the display 39 may be constructed
in any well-known manner,to present a numerical
-19 -
~:fl)~(i'~;~~
indication of the value of the resistive error signal
which, as a reference to Figure 2 will reveal, is
indicative of the size of the respective metallic
particle; whether such particle is ferromagnetic or
non-ferromagnetic, and to present a simple, for
instance on/off, indication of the sign of the
reactive error signal to distinguish ferromagnetic
metallic particles from non-ferromagnetic ones.
However, it ought to be realized that it is also
contemplated by the~present invention to provide other
constructions of the evaluating~circuitry 35 and/or of
the display 36, which present more sophisticated
and/or mare accurate results. So, for instance, the
reactive and resistive error. signals from the outputs
of the filters 30 and 32 have~been supplied to an
oscillograph for recording thereat, and the thus
recorded traces of the reactive and resistive error
signals have been compared and evaluated in view of
one another.to determine both-the size and 'the
magnetic properties of respective particles. Of
course, it is also contemplated to automate this
cross-referencing procedure to determine the exact
location of the response to the respective particle on
the graph of Figure 2 with attendant more precise
determination of the characteristics (size, magnetic
properties) of the respective particle.
Thus, it may be seen~that the electric driving
circuitry described above drives the resonator
consisting of the probe member 12 and the. capacitor
arrangement 16 at or very close to resonance at all
times. This makes the debris detection system highly
sensitive to, and accurately indicative of, almost
instantaneous or in any event quite short-lived
-20 -
'~'(1~)f~'~~)~
perturbations in the resonator behavior (i.e. resonant
frequency) as caused by metallic debris particles
pasting through the gassage 11 of the probe member 12.
On the other hand, there is provided automatic and
continuous compensation for the effect of gradual
changes, such as those~accompanying temperature
changes, aging of the system components, or the like,
on the resonant frequency of the resonator. In
essence, such short-lived perturbation detection and
long-term change compensation.is the result of the use
and operation of the in-phase and quadrature detection
and feedback arrangement which automatically nulls out
a high frequency bridge circuit in response to
long-term drifts.in the resonant frequency and in the
quality factor Q of the resonator, and which provides
high sensitivity detection of transient changes caused
by the passage of metallic debris particles through
the passage 11. Experiments conducted with an actual
implementation of the above-described detection system
have demonstrated that such system provides reliable
temperature compensation and excellent detection
sensitivity over a wide range of temperatures (the
range of between 15 and 120°C having been actually
tested). The output of this detection system or
arrangement contains sufficient information in real
time for detecting metallic particles, for
discriminating between ferromagnetic and
non-ferromagnetic debris, and far at least coarsely
determining the sizes of the metallic debris ,
particles.
While the present invention has been illustrated
and described as embodied in a particular construction
of~a metallic particle detection arrangement, it will
_.21~_
,'~.'~~)fi'~;~)6
Be appreciated that the present invention is not
limited to this particular example; rather, the scope
of protection of the present invention is to be
determined solely from the attached claims.
-22
