Note: Descriptions are shown in the official language in which they were submitted.
The present invention relates to non-destructive
testing of ferr~magnetic articles, and in particular to
non-destructive testing of elongated ferromagnetic articles
such as steel wire ropes. ~is invention is also applicable
to the non-destructive testing of ferromagnetic components of
more complex assemblies, such as the steel wire armouring on
certain types of electrical cables, ferromagnetic elements
in laminated assemblies, and so on.
Whilst not exclusively directed to the non-destruct-
10 ive testing of steel wire ropes, it will be convenient to base
the description of this invention on steel wire rope testing.
The testing of steel w~re ~opes of the kind used in heavy
industry and mining operations is important from both a safety
viewpoint and an economic viewpoint. Steel wire ropes of the
kind mentioned are subjected to-arduous workin~ conditions,
including in many cases exposure to corrosive atmosphere.
Defects which occur in the ropes during service fall into two
categories. ~irstly there is highly-localised darnage of the
kind exemplified by broken or nicked wire strands; this kind
of damage may occur both at the surface of the rope and inter-
nally. Secondly there is a loss of cross-sectional area, due
to wear and corrosion; this kind of defect will generally be
distributed randomly over the length of the rope, with some
portions more damaged than others. Both kinds of defect
govern the remaining strength of the rope. Ropes of the kind
mentioned are expensive so that it becomes important not only
to detect defects in ropes in service, but to be able to
evaluate those defects reliably so that the working life of a
rope can be gauged with safety and economy.
The non-destructive L~ ing of such erromdgnetic
articles as steel wire ropes has in the past developed along
two paths. In the first approach, commonly called the direct
current (DC) method, a powerful magnetic field created by a
permanent magnet of electr~ma~net is caused to ac~ axially
along the rope under test as the rope passes through holes in
the pole faces of the magnet. Between the pole faces, and
close to the steel wire rope under test, are placed curved
pick-up coils connected to high gain amplifiers, and in turn
to pen recorders, to draw a trace on a chart while the rope
under test is drawn through the pole faces of the magnet.
Broken steel wires in the rope under test create strong polar
fields about the breaks under the magnetic influence. These
polar fields intercept the coils and induce voltages therein
as they pass, causing the pen of the recorder to register a
pulse in response to the break in t:he wire of the rope under
test. The system works well if the breaks are fairly close to
the surface and the rope is moving moderately fast. As the
rope speed falls, the signal picked up by the coil, which is
dependent upon time rate of change of magnetic field, decreases
towards zero. The method responds to some types of corrosion
as noise, but is in general unresponsive to chdnges in magnetic
cross section of the steel wire rope, such as are caused by
wear and many forms of corrosion.
In the second approach, commonly called the alter-
nating current (AC) method, a test coil of insulaled wire is
applied around the steel wire rope Wit}l su~ficient clearance
for the rope to slide through the coil. The coil is connected
-- 2 --
,~, .
"' ' .~ 6
to form part of an AC bridge circuit, driven by a low-frequency
source of alternating current. The bridge is balanced to read
the apparent inductance of the coil of wire placed abou-t the
steel wire rope under test by comparison with sta~le electrical
S components. Loss angle is also obtained for the apparent
inductance. Changes in the magnetic cross sectional area of
the rope caused by wear, corrosion and the like change the
apparent inductance of the coil as these worn or corroded
portions pass through the coil. Variations in loss angle have
some diagnostic value in the detection of corrosion and
crushing, but in general they are not simPly related to
common faults. With increased diameter of steel wire ropes
under test, it is necessary to use lower and lower frequency
bridge drive currents, and slower and slower transit times
for the passage of the steel wire rope under test through the
coil, in order to use this method. This method is very poor
at detecting broken wires, either on the surface or inside
; the steel wire rope. The need for very low frequencies in
testing thick ropes makes the equipment massive.
Both of the earlier methods described here have
marked disadvantages, as set out above. Neither method will
operate satisfactorily with both categories of defect, as
identified earlier. It is noted in particular that both
methods impose restraints on the speed at which the article
under test can be passed through the test arrangement. It is
also noted that the characteristics of the DC and AC methods
are mutually exclusive in so far as sensitivity and speed of
the article under test are concerned.
- 3 -
~,r ~
A possible point of departure from the above impass
lay within a paper entitled "RECENT DEVELOPMENTS IN THE
ELECTROMAGNETIC TESTING OF WINDING ROPES" presented to the
South African Institute of Electrical Engineers, and appearing
in their Transactions of February, 1971 by a Mr. H.W. Kruger,
who stated on Page 37, paragraph 5.4 -
"SUGGESTED IMPROVEMENTS TO THE DC METHODIf the present search coils could be replaced
by suitable Hall detector elements, speed
dependency could be eliminated, measurements
on a stationary rope could be made, and
record traces would be repeatable with the
desired degree of accuracy."
This suggestion appears to have set the direction of all e~fc,rts
in the later art, to develop a practical, stable, sensitive,
and compact instrument for testing steel wire ropes and
similar objects. Certainly all make use of the Hall effect
magnetic field strength detector element known for over eighty
years.
The Applicant is aware of more recent developments in
the test~ng of wire ropes. Two examples are USA patent
specification 4,096,437 and United Kingdom patent specification
2,012,966. These disclosures reveal the use of Hall sensors
in a radial configuration for rope cross sectional area
indication, while Hall sensors in various fonns of "flux leal:-
age sensor configurations" are used for localized damage or
broken wire indication. Both specifications are rather similar,
and reveal devices much improved on the original AC and DC
- 4 -
~'~' .
'6
instruments previously described, yet the present Appli.cant
believes that they suffer a number of serious limitations. One
limitation arises from the fact that both specifications call
for Hall sensors to be placed radially on the inner surface of
at least one hollow cyl~ndrical pole face opposite the surface
of the rope under tes~. If the axis of the rope ~oves alon~
the axis of the hollow cylindrical pole face, and hence that
of the Hall sensors, the magn~tic flux will distribute itself
evenly around the annular gap between rope anù inner surface
of the pole face. Under these conditions the flux flowing
j~
through the sensitive area of the Hall sensor -typifies the flux
density of the whole gap, so that cross sectional area of the
rope estimates can be made from the Hall sensor outputs. If
the axial alignment between wire rope axis and that of the
hollow cylindrical pole face, becomes less than perfect - and
as a result the rope surface approaches one of the ~all sensors,
the misalignment will cause a drop in reluctance at the point
of approach and this will in turn cause a crowding of flux
lines in the vicinity of the sensor - so that the output of the
r 20 sensor is unduly increased by the intense local flux density -
not typical of the whole annular flux flow. q`he output of a
diametrically opposite se.nsor will suffer only a small decrease
in flux density and hence output, because the intense local
crowding of lines, or high density, is counteracted with a
much more extensive area of slightly reduced ~lux density in
the whole annular gap. Thus the reduction in c,utput from one
Hall sensor will not adequately compensate for the increase in
output of the other. A model ~est was made with two diametric-
ally placed Hall sensors within a hollow cylindrical pol~ face,
-- 5 --
~r~
through which a steel bar, representing the rope under test,
was passed, returning ~he flux through a similar pole face
arrangement to the magnet. A thir~ Hall sensor, in a cross
sectional gap in the bar filled with solidified plastic cement
recorde~ an indication of the total flux through the bar and
hence the annular gap as a whole. With the bar centred in
the gap both Hall sensors read the same output. Presumably
this was the even flux density condition. Then the bar was
moved a little towards one of the Hall sensors (and hence
away from the other); this caused a small change in the outpu~
of the bar flux sensor, and a very much larger and quite non
linear change in the sum output of the two diametrical Hall
sensors. Typical figures obtained were:-
Percentage rise in Total error in the summed
flux flowing through output of two Hall detectors
rope or bar under test mounted radially and in
caused by an intent- diametrically opposed
ional misalignment positions. When one of
toward one of the them is approached by the
Hall detectors rope
3.0% +5.8%
1 5.1% +23.5%
7.7% +64.~,6
The Applicant believes that the reasoning and model
tests show that a relatively large error can be produced in
the "output indication of area" for small errors of alignment
between the rope axis and pole face axis, in instruments made
to the beforesaid specifications. The practical importance of
; the area error caused by misalignment is in the need for great
stability of measured results if the slow errosion of rope
-- 6 --
IJ~
~: b
ti3~ .
area under wear is to be monitored over the life of a rope,
which can be as long as ei~ht years. Mechanical means of
assuring the alignment help but create difficulty in accommo-
dating the large number of different rope si~es met with in a
practical installation.
The localized rope damage caused by nicks and broken
wires is detected in both of the above-mentioned specifications
by the use of Hall sensors as leaka~e flux detectors. This
form requires the use of a "saturation" flux densit~ in the
steel wire r~pe under test for its operation. This need is
t.a~ht hv hn~h ;nvpnt-ors~
This method of detecting localized faults is
sensitive and works well, but its need for saturation flux
density to exist in the rope under examination~places a severe
restriction on the cross-sectional area detection system
which needs an unsaturated rope for accurate linear area
estimation. As saturation is approached the linearity and
~ensitivity of the cross-sectional area detection system becomes
progressively less linear and less sensitive to reductions in
the cross-sectional area of the wire rope under test. At full
saturation the cross-sectional area detection no longer resPonds
to reduction~ in the cross-sectional area of the roPe under
test.
A specification by V.A. KALANDADZE et al at the
TSULUKIDZE INSTITUTE in the S~viet Union, PATENT No.
SU 410,305, declared June 18, 1969, and published
September 12, 1974, entitledl "FAULT DETECTOR FOR STEEL ROPES"
discloses a method of improving the sensitivity of Hall-effect
sensors when used to examine the flux about a magnetised steel
wire rope. Two ferromagnetic rings of similar diameter
7 -.
i6~;
surround the rope with their axes common to that of the rope
to be tested, the rings adjoin one another and have small
depressions opposite one another in which Hall-effect sensors
are placed. The inventors claim a n~mber of Hall-effect
sensors located around the rings, typically 8, all oonnected
additively in series and leading to a single channel chart
recorder drawing a fault pattern from the rope being passed
through the rings. The nature of the fault pattern is not
revealed, nor is the intensity of magnetisation of the rope
lo under test. Presumably the response i8 an area sensitive curve.
The inventors claim that their method increases the sensitivity-
of the sensors of magnetic fields about the rope through its
ring coupling method. The rings as used certainly improve
the sensitivity of coupling, and are tolerant to lack of
centricity, but they are ill-suited to "localized fault or
broken wire detection" unless a saturating field is used, which
leads to the contradictions described in the previous
8 pecification.
The applicant Qolves the above dilemma, by
making use of two (2) independent "area detection systems" in
lieu of the combination of a "leakage flux detector" and an
;"area detection" sy~tem. m e two (2) area detectlon sy~tems both
need the sa~e ideal flux density in the rope under test, namely
in the first linear zone, above minimum value (in order to
overcome interference and noise); and well below the "knee"
of the BH characteristic curve (in order to avoid unlinearity
in response). Definitions and details will be expanded later.
An output sensitive to localized rope damage,
such as that caused by the presence of nicks, corrosion,
pits and broken wires is obtained, by subtracting the
two "area sensitive detector" outputs, 6~ that an overall
~ - 8 -
,~ , .
sensitivity compatible with that of a leakage flux detector is
easily obtained. An output channel sensitive to the cross-
sectional area of the steel wire rope under test is obtained
by adding the outputs of both of the "area sensitive detectors"
in an amplifier system.
It i~ an object of the present invention to provide
a simpler (and more reliable) cheaper mean~ whereby the
detection of 1088 of cross-sectional area of a wire rope under
test i8 more sensitively and accurately estimated at lower
than saturation levels.
It is an object of the present invention to provide
a sensitive indication of sudden discontinuities in the cross-
- sectional area of a wire rope under test, such as those pro-
duced by nicks, corrosion pits and broken wires, without a
need to mainta~n a saturation flux density in the wire rope
under test.
It is another object of the present invention to
provide for the same magnetic means being used to detect
both of the above-mentioned kinds of fault simultaneou~ly.
It is an ob~ect of the Present invention to provide
mean~ whereby the susceptlbility of the arrangement to errors
in estimation of the cross-sectional area of the wire roPe
under test, due to skew or non-central passage of the wire
rope under test, through the pole faces is substantially
etiminated.
Other advantageous aspects of the present invention
will appear from the detailed description which follows.
An arrangement for the non-destructive testing of
elongated ferromagnetic articles in accordance with this invent-
ion comprises a magnet system, consisting of a magnet and polepieces, the pole pieces being adapted to surround a part of the
elongated ferromagnetic article under test and to de~ine
".
~'
between the said pole pieces, a section of the said article,
and to produce in that said section, a substantially con~tant
magnetising force extending along, and substantially parallel
to, the length axis of the said section. The magnetising
force so produced, being such as to generate in the said
section, a magnetic flux flow in the "initial linear" portion
of the BH curve appropriate for the material, from which the
section is made. At least two defect sensing systems inter-
mediate the pole pieces. each defect sensing system c~mprising
lo at least two ferr~magnetic rings, adapted to surround the said
section, with at least one magnetic field sensor between the
rings, or where there are more than two rings, between each
adjacent pair of rings, and where possible symmetrically disposed
about the said section, with their axes of maximum sensitivity
parallel to the length axis of said section, or where the rings
are to be openable, and so must be cut into at least two sectors
per ring, then one magnetic field sensor per pair of longitudin-
ally ad~oining sectors. The pole faces are necessarily sectored
when the ring is openable. The sectored pole faces and rings
are adapted to surround the said section under test, when closed,
and means are Provided for combining and evaluating the outputs
of all field sensors, where the numbex employed for openable rings
is two or more.
The means evinces a pair of simultaneous outputs, one
in terms of cross-sectional area at the point of examination,
; and the other reflecting the degree of inhomogeneity or
localized cross-sectional variation over the sampled section
or length in the arrangement, but not that portion which is
not so enclosed.
B
-- 10 --
This invention will now be described with reference
to the accompanying drawings, of which Fig. 1 shows in simplified
form an axial cross section of the essential element~ of part
of the novel arrangement in accordance with this invention,
S F;g. 2 shows a longitudlnal cross sectional view of the essential
elements of the same part of the novel arrangement, Fig. 3 shows
a schematic ar~àngement of the basic defect ~ensing system used
in this inven~tion, and Fig. 4 is a schematic diagram of one
example of an arrangement in accordance uith this invention.
Like reference numbers identify like elements i~ all three
drawing~.
The essential elements of the portion of the arrange-
ment shown in Fig~. 1 and 2 will now be identified. ~ magnet
system i9 shown, consisting ~n th~ 8 example of a generally
cylindri~al permanent magnet 1, having its poles at opposite
ends of the cyllnder, as denoted by the letters N,S, and two
pole pieces 2,2'. The magnet itself, and more generally the
magnet sy tem, may take any form suitable to the application,
the essentlal ~eature of the magnet system being that the pole
pieces encompass the article under test and produce in that
article a magnetic force which is essentially longitudinal
between the N and S pole pieces, and which is sufficient to
produce in that section of the article under test, which is
between the pole pieces, a value o~ magnetic flux density B which
lies in the rising portion of a typical ~H curve and below the
characteri~tic "knee" o~ such a curve. The pole pieces may be
integral with the magnet itselfS the magnet may be other than o~
generaily cylindrical shape; and may be an electromagnet instead
of a permanent magnet.
B
ti 6~
It is assumed for the purpose o~ thi~ explanation
that a steel wire roPe is under test, and a ~ection of this roPe
is shown, marked 3. It will be seen that the pole pieces 2,2'
are provided with openings to accommodate the article under test,
and to allow the article under test to be drawn through that part
of the arrangement shown in Figs. 1,2. Alternatively, of course,
the arrangement may be moved relative to the article under test.
Disposed around the article under test, and intermediate
the two pole pieces 2,2' in this example two defect sensing
systems, each cDmprising in part two ferromagnetic rings. In
this example there are three such rings, marked 4, 5, 6, the
ring marked 5 being common to both defect sensing systems. These
rings are of similar size, located parallel with each other and
substantially coaxial with the article under test. Between each
adjacent pair of rings associated with a defect sensing system
is located at least one magnetic field sensor, and where the
rings are openable, and hence sectored, at least one per pair of
opposite and adjacent sectors. ~ecause in this example there are
two defect sensing system6, using three rings, each sectored into
two half-rings to be openable, there will be at least four
magnetic field sensors, 7, 8, 9, lo. These magnetic field
sensors, preferably Hall-effect devices, are positioned so that
their axes of maximum magnetic sensi~ivity are substantially
parallel with the major axis of the artlcle under test, and
intersect the rings. Advantageously, ferromagnetic ~ield
concentrators 11 of ~uitable shape, for example in the form
of frustums of cones, are used to concentrate the fields
between the rings through the magnetic field sensors. The
magnetic field sensors, and of course the ferromagnetic field
concentrator~, are symmetrically disposed relative to the circum-
ference of the rings, and thereby arranged substantially
_ 12 -
symmetrically around the article under test. Whilst Fig. 2~hows two magnetic field ~en~ors and four ferromagnetic field
concentrators associated with each ad~acent pair of rin~, op-
ti~nally more than two magnetic field sensors may be used between
each adjacent pair of rings, together with associated magnetic
field concentrators.
By avoiding the use of magnetic field sensors in the
gaps between the pole pieces and the article under test, errors
introduced by movement of the article under test normal to its
length axis within the pole pieces no longer occur in the present
invention, as they did in the Prior art of, for example, USA
patent sPecifiCation 4,og6,437. m e total flux density between
each pole Piece and the article under test will remain substan-
ially ~onstant, even though its distribution in the gaps will
vary according to the relative positions of the pole pieces and
the portion of the article under test within them. It follows
therefore that the magnetic flux in the article under test will
remain substantially constant for a constant condition of that
article. ~y placlng the magnetic field sensors of thls present
lnvention where they will not develop errors dependent on the
position of the article under test, by using the def~ct sensing
system as described elsewhere in this specification, the difficul-
ties and disadvantages attendant upon the prior-art sensitivity
to the relative positions of the pole pieces and the article5 under test, are substantially eliminated.
m e defect sensinc3 system of this invention, consisting
essentially of at least two ferromagnetic rings surrounding the
article under test intermediate the two pole pieces, together
with at least two magnetic field sensors and associated ferro-
- 13 -
~1~ 6 ~i
magnetiC field concentrators, all as previously described, will
now be,described in greater detail with reference to Fig. 3 of
the accompanying drawings to disclose the mode of operation.
Fig. 3 shows in schematic form one half of a cross section of the
magnet system, sensing system and an article under test, bounded
by the centre line of the article under test. Thus Fig. 3 shows
part of the (in this example) permanent magnet 1, parts of the
pole pieces 2,2', and part of the article under test 3. Shown
intenmediate the pole pieces 2,2' i9 the sensing system,
consisting in this example of part of two ferromagneti~ half
rings, 4,5, a magnetic field sensor 7, and associated ferromag-
netic field concentrators 11.
The output of this single Hall sensor assembly will be
influenced by the strength of the exciting magnet (1), the
reluctances of the pole faces of this magnet (2), the rel~ctance
of semi-circular ferromagnetic ring elements (4), the shaPe of
th~ concentrators (11) and the disposition of the air gaps in the
system, and the quality and cross section of the steel rope under
test. Many of these factors are substantia~ly fixed for a given
construction and rope material. The practical variable is in the
rope under test. If no rope is in position a certain output
voltage will be observed. If now steel ropes or steel test bars
are introduced the output of the Hall sensor will be reduced
witn their presence, and the output will fall substantially in
an inverse relationship to the steel cross sectional area of the
bar introduced. Thus the output reflects changes in the steel
cross sectional area in an inverse fashion especially for small
variations in cross sectional area (say ~20% area changes) when
compared with a "standard or healthy unworn samPle".
- 14 -
, ~
`
The basic arrangement described above d~tects only
cross-sectional area changes in a rope of homogenous steel (one
of material free from permeability variations). However, by the
use of two such arrangements~of single Hall cell sensors, local-
ized changes in the cross-sectional area can be detected with
great sensitivity with a "subtractive combination~' of the indiv-
idual outputs of the single Hall cell sensors. The "subtractive
combination" of the outputs of the two single Hall cell sensors,
is obtained by feeding one sensor output into an INVERTING input
of an operational amplifier and the other sensor output into a
NON-INVERTING input of the same operational amPlifier.
If the outputs of the same pair of single Hall cell
sensors, are combined with both outputs going into either
INVERTING inputs, or both going into NON-INVERTING inputs of
another operational amplifier, an ~addditive combination" of
these outputs will appear at the output of that operational
amplifier. To summarise, ~additive combination" produces an
average of the cross-sectional area sensed by each single Hall
cell sensor. "Subtractive combination" produces an output pro-
portional to the amount of dissimilarity in cross-sectionai area,
sensed by subtracting the properly adjusted outputs, of two
single Hall cell sensors.
A practical way in which to express the ideal flux
density range i8 by expressing it as a percentage of the
"minimum saturation flux density~' which i9 defined as that flux
density at which the permeability of the rope under test, falls
to equal twice that of free space in the units employed.
i.e. "minimum saturation flux density"
= flux density in material when dB/dH
= 2 x free space permeability (in same unit system)
Correct values for flux density for operation are
defined as percentages of the "minimum saturation flux density".
~n upper range for flux density is 7o% of "minimum saturation
flux density" and locates a region in the BH curve where approach-
ing saturation is ju~t reducing the signal output and linearityof the area response. A lower limit to flux density can be
expressed as 10% of "minimum saturation flux density". Densi-
ties less than this in ordinary industrial si-tuations tend to
reduce the ratio of signal amplitude to magnetic and electro-
~0 magnetic noise below that needed for efficient wor~ing.
There are a number of ways in which the pair of half-
ring single Hall sensor assemblies can be arranged in relation
to the roPe or other extended test object. At least three of
these are notable. one where the pair are located on the same
side of the rope, in line. This arrangement compares a half
cross section of one length of the rope with that of another
length of the same rope on the same side at some distance ahead
or behind. The lengths over which the comparisons are made is
approximately that between the rings of each assembly. This is
u~ually the preferred arrangement for steel wire rope testing.
The use of rings which extend completely around the roPe will
allow the whole cross section of each length of rope to be com-
pared rather than the half cross sections encountered above.
Rings extending completely around the rope are generally
not acceptable in practice, where a couple of joints are
necessary in each ring in order to allow the rings to be
"opened" to admit the rope under test. Variations in the upper
half cross section of the rope will be induced as magnetic
15 - 1 ~ 6 6 6~ 6
65~6
signals into the upper ring segment, where the Hall sensor i:;
situated, and wherein they will give rise to electrical sign.lls.
Identical variations occurring in the lower half cross section
of the rope will be induced as magnetic signals into the lower
ring segment, then they will have tc pass through the joints in
the ring to reach the upper section thence the Hall sensor. In
so doing they are attenuated by the reluctance of the joints so
that the signal is weakened. To remedy this situation Hall
sensors can be used in both upper and lower ring segments, and
their outputs combined in an operational amplifier or the like
so as to give the same effect as a complete (unhinged) ring in
responding to the whole cross section of the rope.
In another arrangement, the openable pair of half-ring
Hall assemblies, just mentioned, can have their upper and lower
Hall sensor signals not only added in one operational amplifier
to give a whole rope cross sectional area signal as described,
but also can have the normally equal Hall signals from top and
bottom half-rings set against each other (subtracted) in another
operational amplifier, the output of which will simultaneously
respond to clifferences in area between top and bottom cross
sections. The output of this arrangement responds fairly well to
corrosion pitst broken wires and nicks, which are unlikely to
co-exist on both sides of the rope. The arran~ement is insensit-
ive to certain fault locatio~s (those on the plane of joint open-
ing and in the axis of the rope). It is inferior to the preferreclarrangement for wire rope testing, but could be of value for
examining ferromagnetically plated strip and laminates.
Another arrangement of the basic half-ring simple Hall
sensor assemblies is possible - in which the assemblies are in
_ 16 -
line but on differen~ sides of ~he rope, so tha~ a length
on the upper side of the rope can be compared with a similar
length on the lower side of the rope. Generally this is not by
itself a useful arrangement for rope testing.
Yet another arrangement would be to ~lace the assemblie6
in line put with one further around the rope than the other. Suc~
an axrangemen~ if extended in double pairs of half-rings Hall
sensor assemblies can be used to cover a whole surface of a rope
when very large ropes necessitating large circumferential dist-
ances and hence reluctance in the ring may tend to leave aninsensitive zone near the axis of partition where the half rings
open. By a multiplicity of rings and staggering of opening
planes such an effect can be reduced at ~ill.
Having described the means for magnetising the article
under test and the arrangement of sensors for detecting defects
in the said article, the processing and utilisation of the outputs
from the maynetic field sensors will now be described. Fig. 4
shows in basic schematic form one example of an arrangement in
accordance with this invention including elements already
described with reference to Figs. l and 2. For reasons of
clarity, however, the magnet system is represented in this draw-
ing by the pole pieces 2,2' only, and the article under test is
not shown. other elements carrying reference numbers up to and
including ll have already been described.
The preferred method of processing and evaluating the
outputs of the n1agnetic field sensors 7-lO is illustratèd in
Fig. 4. The output of each magnetic field sensor i.s applied to
the input of a respective one of four conditioning amplifiers 12-
15. Typically these conditioniny amplifiers are such that the
gains of the amplifiers can be individually preset to compensate
!~ ~
~J _ 17 -
ti6`~i
for tolerances in the magnetic field sensors and in the condit-
ioning amplifiers themselves. Each sensor/amplifier combination
is set up so that with no test article in the axrangement, the
outputs of allthe amplifiers are equal.
The outputs of these four amplifiers are applied to the
inverting and non-inverting inputs of two summing amplifiers 16,
17 through resistive networ~s as shown. With a steady magnetic
field within the space defined by the pole pieces 2,2', such as
would be experienced with the arrangment closed and no article
under test, the outputs of magnetic field sensors 7,8 add in
the output of the summing amplifier 17 in a negative-going sense,
whilst the outputs of the magnetic field sensors 9,10 reinforce
thP output of the summing amplifier 17, again in the negative-
going sense. A stabilised reference voltage source is connected
to the potentiometer 18. Adjustment of this potentiometer is
made to cause the output of the summing amplifier to increase
positively, so as to counter the effect of the inputs from the
magnetic field sensors 7-10. This is an initial setting up
procedure, to obtain a zero reading xePresenting the condition of
no ferromagnetic material placed in the arrangement.
The insertion into the arrangement of an article under
test, will cause a Positive output from the summing amplifier 17,
the output being substantially proportional to the ferrous cross
sectional area of the article under test, as it passes through
the arrangement. The ou~put of the summing amplifier 17 may be
applied to an indicator represented by 19, for example a chart
pen motor.
The outputs of the magnetic field sensors are also
applied to a further summing amplifier 16. Il~ this case the
- 18 -
outputs o~ the magnetic ~ield sensors 7,8 oppose each other a
the output of the summing ampli~ier 16 for the aforementioned
steady magnetic field within the space defined by the pole pieces
2,2', whilst the magnetic field sensors 9,10 also oppose each
other, but with the overall effect that the outputs of the sensors
associated with each adjacent pair of rings, that is, 7 and 9,
8 and lO, add. The output of the summing amplifier 16 under
these conditions will undergo little or no change when the cross
sectional area of an article under test changes very gradually,
or when one article is changed for another of different cross
sectional area. However, this method of processing the outputs
from the magnetic field sensors renders the output of the summing
amplifier 16 sensitive~ to rapid changes in magnetic cross section
of the article under test, or highly-localised changes in the
magnetic field adjacent the article under test of the kind
associated with broken wires. The Output of the summing ampli-
fier 16 may be applied to an indicator represented by 20, for
example a chart pen motor.
Taking an adjacent Pair of ferromagnetic rings, say
A and 5, and the magnetic field sensors arranged between them,
the outputs of the sensors will depend primarily on the spaces
between the inner edges of the rings and the magnetised article
under test. For the relatively slow changes in cross sectional
area of the article under test, the spaces at both rings will
tend to change uniformly, or will vary only slightly and gradual-
ly. However, with a sharp magnetic discontinuity of the kind
associated with a broken wire arriving at one of the pair of
rings, the field linking that ring with the article under test
will be rapidly weakened, compared with that linking the other
_ 19,--
. ~ .
, . . .
ring of the pair to the article under test. A strenqtheninq of
the field would be caused by a splice. Such a field imbalance
between the rings disturbs the balanced situation previously
described in relation to the processing of the outputs of the
magnetic field sensors, and relative]y large pulses will occur
in the output of the summing amplifier 16.
Additional circuits to be described may be connected
to the outputs of summing amplifiers 16 & 17.
One such arrangement would accept the output of
lo amplifier 17 and apply it to an amplitude window comparator
which may be formed by two sections of a dual comparator such
as LM711 or LM193 or a quad comparator LM139. One comparator
is eonnected in the 'inverting' mode and the other in the 'non-
inverting' mode. SeParate reference voltage sources are provided
for each section. If the signal from amPlifier 17 shauld dev~ate
beyond pre~determined limits, which are represented by the
reference voltage sources, the output of the comparator will
change state in a binary manner. ~oth sections of the comparator
may be connected in a wired 'OR' configuration and the output
of this connection may be utilised to operate an alarm of the
visual or audible type and if so desired may operate via well-
known circuits a solenoid valve to release a splash of Paint on
the rope under test, to indicate the position at which the ropP
area has deviated from the preset limits.
Another arrangement would be an analogue to digital
encoder, connected to the output of amplifier 17. The encoder
'; would be for maximum utilisation a pure binary coded type.
Analogue to digital conversion is a process well described and
known to those versed in the art. The sampling or encoding rate
would be set to reflect the likely frequency of area changes.
B
- 20 -
The encoded di~lital information could be s-tored on magnetic
tape, in magnetic memory or in solid state memory. From such
storage devices it may be retrieved through the u~e of a
Digital to Analogue converter with the further advantageous
facility of time flexibility where the 'replay' may be at a
rate quite different from the 'record' rate. Sampling could
be under control of a rope derived signal obtained from a
distance transducer rotated by the rope. Such distance ~ignals
would be stored in the same manner as the digitally encoded
signal and would be available to identify the location of area
deviation during the retrieval operation.
A further application of well understood electronic
techniques is to connect to summing amplifier 16 a shaping
circuit using a Schmitt trigger or similar well-known circuit
and to apply the output of the shaping circuit to one or a
plurality of digital counting circuits. These counting circui~s
may be advantageously used in all or some lesser combinations
of the following ways~ to display on a panel display the
number of localised cross-sectional area variations over the
length of the rope, to be reset at fixed distance intervals
from a distance transducer (such as was referenced in the
previous paragraph), to provide a continually updated panel
display of significantly abnormal local cross-sectional area
variations per unit length, to provide to another section or
address of the storage medium referenced in the preceding
paragraph such continuously or periodically updated information,
to use the shaped Pulses to provide the sample rate ref~rred
to in the preceding paragraph, to use the shaped pulses to
initiate a storage of a distance transducer code to store a
- 21 -
string of distance addresses indicating -the pogition of local
cross-sectional area variations.
Such methods briefly described may be incorporated
in any imPlementation of the invention with various advanta-
geous aspects and are readily available to anyone practisedin electronic circuit design.
Control current leads to the magnetic field sensors,
power supplies and other well-known and well-understood
provisions are not shown or described, as they are within the
lo competence of the skilled addressee. An optional feature of
the arrangement described is the alternation of orientation
of the magnetic field sensors, as shown in Fig. 4 by a
thickened border. Inter alia, this step offers advantages
in reducing the effects of environrnental temperature changes
upon the outputs of the arrangement. The arrangement in
accordance with this invention is for all practical purposes
independent of the speed of the article l~nder test through
the arrangement, and provides in a relatively simple manner
efficient detection of both forrns of defect identified
earlier.
- 22 -
