Note: Descriptions are shown in the official language in which they were submitted.
2~ 196
DETECTION LOGIC AND SICNAL PROCESSING
~ETHOD AND APP~RATUS FOR THEFT DETECTION SYSIE~
TECHNICAL FIELD
This invention relates in a broad sense to detection systems
and apparatus, and more particularly to de~ection systems principally
used in "anti-pilfering", i.e., theft-prevention, systems; more
particularly still, the invention relates to that type of detection
system in which an alternating electromagnetic field is monitored
to unobtrusively and invisibly detect the presence ~ithin the field
of a small strip of permalloy or like highly-magnetizable (ultra-ls~Y
coercivity) metal foil which is hidden upon or in articles such as
consumer merchandise whose theft or othenYise-impermissible ta~in~
is to be detected and prevented.
BACKGROUND OF THE INVENTION
Many prior efforts have been made toward deterring or pre-
venting ~hefts in the nature of shoplifting~ or other undesired removalof "contraband" articles or goods, or example, unchecked librar~
books or the like, and such prior efforts have given rise to a variety
of different systems and approaches, based upon dierent technological
phenomena including, for e~amplel detection o~ permanent magne~ pieces,
a variety o electromagnetic field applications, micrswave systems,
infrared or ultraviolet, etc. These rather extensive prior efforts
have, quite understandably, advanced the general state of the art in
these diferent fields, and have in general enhanced the degree of
success available; however, ~he desired end is exceedingly difficult
from a technological point of view, since the areas ~o be moni~ored
(in general, doon~ays or like points o~ egress~ are large in a physical
sense, whereas the articles ~mder surveillance are usually relatively
small, requiring a proportionally ~iny detection element or "marker".
Generally speaking, this requires e~ceedingly high system sensi-
tivity, but it is not only important to d~tect the illicit passageo contraband material; it is almost equal~y as important to a~oid
"~alse alarms", in which bona fide customers or other innocen~ persons
are wrongly pointed out as carrying stolen or contraband goods through
the portal, since this not only leads to i~m~diate wTongful embar-
rassment of the individual involved, but also is likely to cost themerchant or other proprietor the loss of sllbstc~ntial goochYill and,
-2- ~L~2~3~
potentially, possible litigation by those claiming to be damaged by
such incidents.
Accordingly, real progress satisfying both of the afore-
mentioned requirements of high-sensitivity attended by great
selectivity has been di-fficult to achieve and slol~ in coming. This
conclusion is evidenced by the issuance of various patents over a
long period of years, each asserting the achievement of improvement~,
but each followed in time by another patent directed to still a further
improvement in a seemingly continuous sequence. By way of example,
perhaps the most frequently-employed, and probably the most success-
full system concept, relates bacX to the often-noted French Patent of
P. A. Picard, No. 763,681, issued in 1934, in which the technological
phenomenom is described as involving electromagnetic field perturba-
tions resulting from the insertion or presence within the field of a
piece of magnetic material. In particular, Picard noted the field
effects created by the presence of highly-magnetic (high penneability)
material such as pe~nalloy, which creates the presence of a number of
the higher-order odd hannonics of the fundamental frequency of the
applied field (e.g., Picard referred to the presence of the ninth and
eleventh harmonic). While a period of almost 50 years has elapsed
since the appearance of this pa~ent to Picard, various patents con-
tinue to issue from time to t~ne asserting advances in Picard's theories
and findings in the area of '~ilferage detection" systems of the type
noted hereinabove; foT example, reference is made to a number of paten~s
issued to Edward Fearon (including U. S. Nos. 3,631,442, 3,754,226,
3,790,945, 3,820,103, 3,820,104) and to Peterson (U. S. No. 3,747,086),
Elder et al. (U. S. Nos. 3,665,44~ ~d 3,765,007~ as well as U. S.
Patent No. 3,983,S52 to Bakeman. Indeed, a very recent such patent
is that issued to Robert Richardson, U. S. No. 4,300,183, which is
directed to and describes various attributes of the underlying-concept
relating back to Picard.
As stated above, the seemingly continuous advance in the
general state of the ar$, as evîdenced by the aforementioned patents,
has undoubtedly provided new insights ~nd improvements in the general
level o the art, the requirements of truly satisfactory detection
systems are very severe and demanding, and the need therefore continues
to exist, and in some ways becomes even more pronounced, for truly
reliable systems ~hich will unerringly detect relatively small '~arker"
elements or indicia, while at the same time being essentially immune to
a practically endless number of widely-varying metal devices, objects,
S articles, and components, all of which cause perturbations in the
magnetic interrogation ield, with resulting detection-actua~ing results
being inevitably present.
BRIEF SU~MARY OF THE INVENrION
The present invention provides new and highly significant
improvements in electromagnetic field-type detection systems of the type
noted above, whirh improvements substantially enhance both the sensi-
tivity and the selectivity of such a system, pursuant to which pre-
viously-unappreciated detrimental effects such as ield-perturbing metal
structural components in the environment of the egress ~ortal (e.g.,
field-perturbing eeiling grids overhead and/or field-perturbing rein-
forcing rods or mesh in structural concrete nearby, etc.) are substan-
tially eliminated as error sources. The improved system provided
hereby thus makes it possible to accurately, consistently, and reliably
detect ~he presence of tiny markers or ~ags of magnetic mateTial and
reject, or not detect, the presence of other field disrupting metal
elements or components as, for example, keys, pocXetXnives, wristwatches~
metal containers such as beYerage cans or the like~ baby strollers and
shopping carts, and a host of other widely-differing apparatus and
objects
In accordance with the disclosure a detection system and
method is provided with greatly enhanced processing of the marker-
detection signals, incorporating a summing and diferencing procedure
for substantially impro w d signal-to-noise ratios, in accordance with
which a co~paratively low frequency component band and a comparatively
high requency component band are separately det0rmuned, and utilized in
a multiple-step comparative manner to dynamically control the detection
alarm threshold. In this manner, a balancing of the frequency
components or bands is utilized, to produce alarms only when the ratio
of requency bands is in the appropriate order, representatiYe of
the actual marker indicia, thereby avoiding false alar~s produced
by prior systems in response to metal articles whos~ field-perturbation
.. . ~
~~ -4- ~L~2 ~3~
effects happen to be very similar to those o the authen~ic marker,
even including those articles which produce similar frequency components
but ~Yhich are distinguishable by the relative amounts of diferent
frequency bandsa i.e., the ratio of the signal strength representa-tive
of different frequency component bands.
Further in accordance with the disclosurethe method and
apparatus provided operates to additively, or construetively, sum
representations of detection signals indicative of marker presence
within the field, regardless of and continuously consistent with, field
alternation phase changes and differences; additionally, the method and
apparatus provided dif~erences or subtracts signals representative
of non-marker presence (i.e., nois~) in order to accurately portray non~
marker effects. In this manner, the marker-charac~erizing signals
are comparatively analyzed by reference to the non-marker signals, thus
substantially enhancing selectivity.
Somewhat more particularly, in accordance with the prese~t
disclosure detection logic and processing methods and apparatus ar~
provided for examining representations of electrical signals which ar~
produced initially by receiver means that monitor an alternating elec-
tromagnetic interrogation field in which a predetermined marXer may bepresent, so as to accurately and reliably determine the presence,
and/or the non-presence, o such 2 marXer in such a ~ield, In accord-
ance herewith, such signal-examining is carried out in a manner having
the efect of determining, and using, a first and second composite
analysis signal, the first being used to set a first value of a com-
parison operation and the second such signal being used to set a second
such comparison value, such two cornparison values being ef~ectively
compared in such comparison operation. Preferablyl this is accom-
plished by developing an integrated ambient-representative signal and
using the latter to vary a preset threshold in comparison stages whose
primary comparison inpu~s are signals representative of marker presence
and marker absence, with the results of such comp~risons being summed
against each other, and the resultant su~ation level used to trigger
an alarm or lndicator showin~ the verified presence of the marker within
the interrogation field. In particular, the inventor provides pre~erred
methods and appara~us for selective pre~mplification and processing of
,~ .
-5- ~L~ 3~3~
the signals produced initially by the receiver means9 prior to the
opera~ion of the preferred detection logic and processing methods and
apparatus, both of which are further disclosed in the following de-
scrip~ion of particular preferred embodiments.
In accordance with one aspect of the invention there is
provided, a me~hod for determining the presence of a predeternlined
marker member within an alternatlng electromagnetic int~rrogation field,
o~ the type using one or more monitors ~hich produce ~ecerical signals
o containing indicia indicative of the pres~nce o'~ such a marker withinsuch field, the improveme~t comprisin~: subjecting the slgnals frQm
different ones of said monitor means to preamplification and subtracti~e
processing before the signal anal~sis which determines and in~icates
marker presence or absencej said preamplifica~ion compr~sing the
separate application of the signals ~r,om different monitor means to
separate preamplification channels, and said subtractive processing
comprising subtracting a monieor mea,ns signal produced by a monl~or
means relatively further ~rom a mar~er within ~he interrogation field
from a different marker means signal produc~d by ~ moni~or means IYhich
is relatively closer to that marker, thereby at least partially
cancelling out ambient noise and/oF other undesired effects from the
monitor mean~ signals prior to urthe~ processing thereaf.
In accordance w1th a second aspece of the ln~ention there is
a method for det~cting the presence of a partlcular marker me~ber within
In a mothod of det~ctlng thc pres~c~ o~ a partlcul~r m~r~or ~e~bc~ wlthln
~n a1tern~tln~ electroma~netlc lnt~ro~tlon fl~ld ~st~b1lah~d ~croas a
portal, whercln sald ol~ctrom38notlc Pl~ld 13 monltor~d by at lca3t flr~t and
second recelver mean3, eAch for accessln~ the ~leld ~o~ a dlf~3s~nt poaltlon
and ~or producln~ an alectrlcal u~gna1 roprasontatl~a of ~lgna1 ,indlcla cau~qd
~y such a markcr ln ~aspons~ to thc alto~natlons o~ the flelt, th~ l~prove~3nt
for use in detectln2 s~d mar~or whSch co~prl~s~ the ~teps o~: ~roducin~ a
first proccssin~ si~nal for marker-pro30nce ans1ys~3 by uso Oe ~sld alactrlcsl
si~nals produced by at 1cast ona o~ ~ald recelver~; producln~ ~ socond
procossln~ slBnal eOr U9~ ln ~arkor-pr0senca ~n~1ys~ by dle~ar3ncln~ th~
c10ctricQ1 slgna1s producod by ~ald flrst ~nd sacond recel~ars, to roduce
common-mode nols~ or othor unda~l~ed sl~na1 char~ctsrl~tlc~ ln that o1ectrlc~1
si~nal; Bnt usln~ tho second procaq~in~ sl~nal In co~pcratlvo an~1ysl3 wlth
~aid flrst procassln8 al~ctrlc~ ns1 to doteraln4 pr~3ance and ~b~enc3 o~ 8
msrk~r ~lthln said ~i31d.
- 5a
A number of additional improvements and adYan~ages are pro-
vided in accordance herewith, as described in more detail hereinafter
in conjunction with certain preferred embodiments of the invention as
depicted in the attached drawings and specifically noted in conjunction
therewith for a more meaningful disclosure.
BRIEF DESCKIPTION OF ~HE ~RAl~INGS
-
Fig. 1 is a simplified, schematic-fonm block diagram of the
overall detection system in accordance with the invention;
Fig. 2 is an enlarged and schema~ic circuit diagram of the
preamplifier portion of the system shown in Fig. l;
Fig. 3 is an enlarged schematic circuit diagram of the inhibit
amplifier portion of the system shown in Fig. l;
Fig. 4 is an enlarged, simplified system block diagram il-
lustrating the preferred detection logic and processing circuitry;
Fig. 5 is a schematic circuit diagram showing a first portion
of the detection logic and processing circuitry of Fig. 4; and
Fig. 6 is a schematic circuit diagram showing the second
portion of the detection logic and processing circuitry of Fig. 4.
DETAILED DESCRIPTION OF PREFERRED Eh~ODIMENTS
. . . _ . .
The general nature o~ the overall system is illustrated in
Fig. 1, in which a typical two-portal system is depicted. Generally
speaking, such "portals" should be understood as being egress passages,
e.g., doo~aysl on the opposite sides of each of which are maintained
electromagnetic interrogation field sources (e.g , induction coils
constituting part of an oscillating L-C tank cirouit) together with a
receiving antenna which monitors the electromagnetic field from that
particular side of the portal. As will be understood, many of the prior
patents referred to hereinabove depict and discuss systems ~ing such
interrogation coils and antennae; for example, Richardson ~U. S. No.
4,300,183) depicts a system and components whose general nature may be
taken as more-or less standard, dating back to the work of E. Fearon
whose prior patents are also noted above~ As illustrated in the
~ 3''3~;
aforementioned ~ichardson patent, wherein the "portals" are designated
"doorways", and whèrein one of a number of different possible coil
shapes and orieneations are illustrated, the field-inducing coils are
physically large~ such that the interrogation field which they produce
occupies a physical area which is more than sufficient for a human being
to readily pass through. Inasmuch as the general characteristics,
attributes, and parameters of such systems, including their field-
inducing coils and receiving antennae, have long been known and have, in
effect, resulted from the work of a number of individuals working at
various points in time, the afore~entioned prior patents of Fearon,
Elder, Richardson, et al. should be consldered as portraying the known
state of the art and describing both general system characteriatics and
circuitry, componentry, and the liXe; consequently, these patents should,
to the extent dee~ed necessary or desirable for environmental disclosure
or otherwise, be considered.
Referring further to Fig. 1 herein for a very general
illustration of the overall syste~, it will be noted that "portal 1" has
a pair of oppositely-spaced sides, designated "side 1" and "side 2", and
the same is true with resp~ct to ~ortal 2, which may be considered a
substantial duplicaee of portal 1. E~ch side 1 of each portal preferahly
receives the same type of drive, i ~., interrogation coil-excitation or
drive current and each side 2 coil i5 also driven like its counterparts,
although as explained hereinafter the side 2 excitation preferably
changes in phase periodically whereas the side 1 excitati~n does not~
With contlnued reference to Flg. 1, it will be seen that the
signal path from each side of each portal is indlvidually coupled (via
channels or paths A and A', B and B') to a prea~p 1, from whlch two
outputs are separately processed, one being directed to an amplifier/
filter 2 and the other to an inhiblt amplifier 3, whose respecti~e
outputs are coupled on paths E and F to a detection lo~ic and processlng
module 9~ whlch also receives control signals on path G from a ti~in~
generator 4. Generally speaking, the detection logic module 9 funccions
to provide indicAtor and/or alarm signal~ indicative of the presence,
within the alternating interrogacion field maintalned between the
respec~lve sides of a glven portal, of the desired field-affecting
marker member, such indicator or alarm outputs being depicted in Fig. 1
as coupled along a pa-th H to an alarm module which is so marked. Power
supply paths are indicated in Fig. 1 as being directed from an outside
"power in" source and along a con~on bus 11 to a power supply 8. The
latter provides various power levels and types to the preamp 1, the
detection logic module 9, the amp/filter 2, the inhibit amplifier 3, the
timing generator 4, the phase driver control 5, and the phase driver 6.
The outputs of the phase driver 6 are coupled along the aforementioned
paths C and D to portals l and 2, to drive the oscillating interrogation
coils located there. ~Yhile state of the art circuits and components
which are generally usable as the foregoing functional units are cer-
tainly knol~n and available at the present point in time, and are re-
ferred to in the aforementioned prior patents, for e~ample~ certain
preferred new versions or improvements o such are disclosed hereinafter
lS It should be understood that in accordance with the present
invention the interrogation coils at the portals are preferably driven
at a nominal oscillation frequency o 10 kHz, and that in order ~o
maximize detection capabilities in a broad sense, it i9 desirable to
drive the two interrogation coils on opposite sides of ~he same portal
in an alternating in-phase and out-of-phase sequence, in "bursts" which
continue over a desired number of cycles. Thus, for a first such period
both sides of portal 1 and portal 2 will be driven in phase, wher0as or
the next ensuing such period side l of each will be driven with the same
phase as before but side 2 of each will be driven with directly out-of-
phase excitation, the effect of which will be to re-direct the resultant
direction of the interrogation field by gO~, thus affording detection
c~pabilities for particular marker crientations within the field which
might possibly be missed or produce very weak detection signals i by
chance oriented essentially orthogonal with respect to the direction of
flux within the interrogation field. A particular example of a pre-
ferred phase-reversal sequencing comprises alternating bursts of 160
cycles ~i.e., 16 msec) of the nominal 10 kHz fundamental alternation,
separated by a "dead time" or "inter burst gap" of 4 msec, with the
first such 160 cycle burst applied with the s~ne phase condition ~e.g.,
"phase A") on both sides l and 2 of each portal (i.e., "A-A" phasing),
and ~he second such burst applied with l'phase Al' on side 1 and the
8- ~L~ 3~
opposite ("phase B") applied to side 2 of each portal ci.e., "A-B"
phasing). Accordingly, the an*enna at each portal side (constituting
the ini~ial "receiving means" herewith) will return in-phase signal
components for marker-present conditions within the portal during the
first such in-phase drive condition, and out-of-phase marker-present
signals during the next such drive condition.
As indicated in conjunction with Fig. 1, the detection signals
from the antennae at the various portal sides are coupled along signal
paths A, A' and B, B' to the preamplifier 1, a detailed illustration of
a preferred embodiment of which is set forth in Fig. 2, to which refer-
ence is now made. In the preamp 1, signal pa~hs A and A' from the
receivers, or antennae, which monitor side 1 of both portals l and 2,
are coupled respectively to preamp inputs P-l and P-2. Conversely,
signal paths B and B' from side 2 of both portal l and portal 2 are
coupled, respectively, to preamp inputs P-3 and P-4. As may be observed,
each such preamp input feeds into an identically-configured amplifying
and filtering network branch, located generally within the circuit
portion on the left, designated l-A, and each of the four such pre-
amplifier/filter network portions feeds into a summation circuit portion
on the right, designated l-B.
Generally speaXing, the interrogation field~generating coils
are driven, in the alternating-phase sequence noted above, with current
pulses on the order of magnitude of approximately 50 amps, preferably
once every several oycles of tank circuit oscillation (e.g., every
fourth cycle of oscillation), resulting in an oscillation of approx-
imately three hundred volts (peak to peak) in amplitude. Each receiving
antenna, therefore, would nominally detect a very strong 10 ~J~ si~nal,
and for this reason the antennae are pre~erably figure-eighted in
winding configuration, so as to null out as much as possible of the 10
l~l~ component. The field perturbations caused by the presence within one
of the portals of the permalloy strip or other such marker element are
miniscule in comparison to the tan~ drive level, thus presenting very
substantial signal-processing difficulties in order to achieve lligh
sensitivity, to avoid missing contraband-carried m~rkers, while at the
same time achieving a high degree of selectivity, to avoid erroneous
contraband or theft-indicatiYe ala~ brought about by any o~ a variety
~ ~ .
.
9~ V,~%~
of metallic objects or articles which also cause perturbations in the
interrogation field.
Toward the foregoing end, certain characteristics
of marker detection have become knonn which greatly facilitate the
sensitivity-selectivity requirements, for example, the drive excitation
pulses applied to the field-inducing coils are highly disrLtptive in and
of themselves, and it is thus desirable to blank out all or part of the
recei~ing circuitry during the time such drive pulses are being applied
to the interrogation coil. Furthermore, the actual permalloy or other
such low-coercivity markers create field perturbations by switching
their magnetic domain orientation each half-cycle of alternation of the
interrogation field, i.e.> on each positive-going half-cycle as well as
upon each negative-going half-cycle, ~ith magnetic domain switching
occurring during the first 90 of current flow in the coils for each
such half-cycle. Accordingly, if the antenna ~"receiYer means") signals
are examined to deternune the presence of a marker within the field only
during the current-rise portion of the cycle (i.e., the first 90~>
other non-marker perturbations may be screened out. Furthermore, i a
sample of the antenna/receiver means signals is exanuned during other
por~ions of the cycles of interro$ation field alternation, i.e., when
marker perturbations are not anticipated (i.e., during the current-
falling portion of each half-cycle), a representative ~mbient field
condition may be established for comparison with the receiver means
signals obtained or examined during those periods when marker signals
are to be anticipated if indeed a marker is present within the portal,
i.e., within the interrogation field.
In addition to the Eoregoing, the treatment afforded the
receiver means signals prior to actual analysis efforts, whose purpose
is to deterntine whether or not a marker is present, beco~.es very im-
portant to successful processing. That is, while it has heretofore beenreco~tized that the interro~ation field fundamental frequency (here, 10
kHz~ must be eliminated to the fLtllest extent possible, the counter-
vailing consideration is to maintain the integrity (fidelity) o the
actual signal rom the antenna to the ~reatest extent possible. This
desired result is greatly facilitated by the circuit configuration shown
in Fig. 2 for the preerred fo}~ of preamp l, in which each separate
~.
-10-
9~3~6
pre~mp circuit path proceeding from the differen-t antenna inputs P-l; P-
Z, etc., is identical, thus mc~king a description of only one such path
necessary. Referring to path P-l, it may be seen ~hat the signals first
encounter a Pi-type RC filter 40, l~hich applies an initial attenuation
i"f 5 f 6 DB centered upon the lO kHz drive frequency, but does not introduceany appreciable noise content as other filtering might. Next, the
receiver signals encounter an amplifying stage 42, which is preferably a
low-noise voltage amplifier having a gain on the order of about ten,
designated UlO0. In a particular preferred embodiment, the latter is
implemented by use of an integrated circuit operational amplifier such
as that designated as IC5534 coupled into the circuit in the manner
indicated, with resistive feedback. Accordingly, the receiver signals
from the antennae essentially encounter strong low-noise amplification
prior to operational filtering. Such filtering occurs after the first
stage of amplification 42, in the t~in-T notch filter 44 comprised of
resistors RlO7, Rl08, RlO9, and capacitors ClO4, Cl05, and Cl06, in
which it will be observed that resistors Rl08 and RlO9 are variable in
na~ure, to provide for precise setting of the notch characteristics.
Notch filter 44 is centered upon the lO kHz interrogation field fre-
quency, and supplies at least 40 DB of rejection for such frequency.
Follol~ing the notch filter 44 in the preamp circuit pa~hs is abuffer stage of amplification 46, which may be implemented by another
integrated circuit No. 5534 operational amplifier, conigured to provide
unity gain. It is to be noted that both amplifiers 42 and 46 should be
wide band amplifiers, so as to accommodate all of the frequencies ~ithin
the range e.Ytending from the fundamental of the interrogation field out
to at least the fifteenth harmonic thereof. IYhile set forth more fully
hereinafter, it is to be noted that the high-order and low-order harmonic
content of these antennae signals are utilized as important determina-
tive factors in accordance here-~ith by observing the ratio or relative
amounts of these bands of frequencies. The lo~er-fre~uency band com-
prises primarily the third and fith harmonic range, and to some exten~
the seventh, since this range is highly representative of interrogation
field perturbations brought about by non-marker metal objects of many
and different particular natures. That is, the actual permalloy or like
marker produce5 a significantly different ratio of the higher-order
~9~
harmonic band with respect to the lo~er-order harmonic band, even though
both the authentic marker and other non-marker objects may produce
varying amounts of both frequency bands in the responses detec~ed from
the field perturbations which they cause. Indeed, some particular and
relatively unusual metal objects (such as certain plated keys and
certain loop-form or mesh-type metal objects) may to a considerable
degree "mimic" (that is, resemble) the response of an authentic perm-
alloy or like marker, although in essentially every instance the actual
ratio of the high order harmonic band to the lo~ order band (as defined
above) will be at least somewhat different than those brought about by
the authentic marker element.
Each of the preamp circuit paths commencing at the inputs P-l,
P-2, etc., thus produces a relatively noise-free and significantly
amplified version of the antennae/receiver means signals, with the
fundamental 10 kHz si.gnal substantially reduced but ~ith harmonics of
this signal present. Each such circuit path has a pair of output
resistors Rlll/R112, R211/R212, etc., which are coupled into the summing
portion of the preamp l-B in the following manner. First, it will be
noted that the Rlll-R211 outputs are ganged together and fed to the
inverting side of a differential amplifier V102. This same circuit path
is coupled to the output side of an upper switch portion Sl in a four-
stage C~IOS analog switch S100, through which signals from preamp path P-
3 output resistors R311 and R411 may also be coupled, upon appropriate
actuation (excitation) of s~tch control terminal SCl/4, which also
controls switch stage S4 of the ~S switch S100.
Output resistors R112 and R212 in preamp paths P-l and P-2 are
ganged together and coupled to the inverting input of a second differ-
ential amplifier U202, and that signal path is also coupled to the
output side of a third switch stage S3 of switch S100. Similarly,
output resistors R312 and R412 of preamp paths P-3 and P-4 are ganged
together and coupled to the input side of switch stage S3 in switch
S100. These same two output resistors, R312 and R412, are also coupled
to the input of the fourth switch stage, S4, of the CMOS switch S100,
just as output resistors R311 and R411 are additionally commonly-coupled
to the input side of the second switch stage, S2, of switch S100.
From the foregoing, it may be seen that the output of commonly
. i:
-12- ~ 3~3~g~
coupled, preamp paths P-l and P-2, which repres nt side 1 of both
portals 1 and 2 (Fig. 1), are coupled to the inverting (i.e., "-") side
of both differential amplifiers V102 and U202, and that signals from
preamp paths P-3 and P-4, representing side 2 of both poTtals 1 and 2,
will also be applied to this same side of differential amplifiers Ul02
and U202 ~as outputs from resistors R311 and R411) when either the first
stage (Sl) or the ~hird stage (S3) of switch S100 are actuated, through
energization of their respective different control terminals (i.e.,
SCl/4 or SC2/3). At the same time, the non-inverting (i,e., "~") side
of di~ferential amplifier U102 is coupled to receive the outputs from
side 2 of portals 1 and 2 (preamp paths P-3 and P-4, from resistors R311
and M 11) whenever the second switch stage (S2) of s-~itch S100 is
triggered by a signal on s~itch control terminal SC2j3, and the analogous
(non-inverting) side of differential amplifier U202 will receive the
side 2 ~paths P-3 and P-4~ output signals (from resistors R312 and R412)
upon actuation of the fourth stage (S4) of switch S100, by an appro-
priate signal on switch terminal SClJ4. For purposes of this specifi-
cation, these control signals applied to switch terminals SCl/4 and
SC2/3 may merely be considered as comprising appropriately-timed gating
signals produced by the timing generator 4 of Fig. 1 which are closely
synchronized to the frequency and phase of the oscillations actually
present at the interrogation field.
Accordingly, by supplying the aforementioned gating signals to
the switch terminals of CMOS switch SlO0, the antenna signals from the
opposite sides of the two portals will be constructively s~m~ned (magni-
tudes instantaneously added) in one path and con~ersely, "destructively
summed" or differenced in another path, to provide t-~o quite different
but nonetheless related signal outputs cn preamp output terminals P-5
and P-6, leading from differential amplifiers U102 and U202, respectively.
hlore particularly, when both sides of each portal are being driven in-
phase with one another, the outputs from their respectively-associated
antennae are in phase and are directly added (summed) by operation of
the first st~itch stage Sl of switch S100, an appropriate control signal
being suppled at that time to control terminal SCl/4. Under these
conditions, differential amplifier U102 has all four such amplified, in-
phase antenna signals applied to its inverting input, and none applied
''
~2~
to its non-inverting input. ~he control signal applied to terminal
SCl/4 also actuates the fourth switch stage, S~, thus applying the Yery
same output signal from paths P-3 and P-4 to the second differential
amplifier, U202, but in the reverse manner, i.e., applying such signals
to the non-inverting inputs, whereas the other two antenna signals are
applied to the inverting inputs, so that these two sets o signals are
differenced, or subtracted, at differential amplifier U202.
When the opposite phase relationship occurs at the inter-
rogation fields (i.e., side 2 of both portals driven out-of-phase with
side 1 thereof), a similar end result is obtained through opposite
switching of the analog swi~ch S100. That is, additive summation occurs
at differential amplifier U102 and subtractive sun~ation at amplifier
U202, i.e., signals from side 2 of both portals are arithmetically added
together at amplifier U102 (directly out-of-phase signals applied to
opposite inputs of the differential amplifier~ while the same signals
are arithmetically subtracted at amplifier U202 (i.e.~ directly out-of-
phase signals resistively combined and applied to the inverting input,
and no signal applied to the non-inverting input). Accordingly, the
output appearing at preamp ouput terminal P-5 represents the algebraic
difference but arithmetic sum of the antenna signals from sides 1 and 2
of the two portals, whereas the output at preamp terminal P-6 represents
the algebraic summation but arithmetic difference of the antenna signals
from opposite portal sides, taking into effect the alternating phase
conditions present within the interrogation field. In this connection,
it is to be noted that the control signals applied to terminals SCl/4
and SC2/3 and the C~S switch S100 are preferably o sufficient duration
to maintain switch actuation througilout tlle entire time interval from
the initiation of one phase condition t~Dr example A-A) to the initiation
of the next succeeding phase condition ~continuing the example, A~B).
~lat is, the "on" time for the s~itch should preferably continue through
the aforementioned dead space or interburst gap, rather than ending at
the immediate conclusion of the ongoing phase condition, since by
continuing the summing and differencing operation on into the "dead
space", additional signal information will be obtained with respect to
interrogation field perturbations which will contribute meaningfully to
system sensitivity and selectivity.
-14-
Quite clearly, ~he t~o signals appearing on preamp output
terminals P-5 and P-6 will have substantially different çharacteristics,
the first such output representing combined antenna outputs providing
the highest possible signal-to-noise characteristics, for maximum
S sensitivity, whereas the output on the second such terminal has had
eliminated from it "common-mode~ noise and other such undesired fre-
quency components. The first aspect is particularly important ~ith
respec~ to the weakest likely marker-present conditions, i.e., a marker
whose particular metallurgy and/or physical characteristics produce very
wea~ perturbations of the interrogation field, and ~hich is located in
the middle of the portal, midway between the two sides where the re-
ceiver means antennae are located, a set of conditions likely to be
missed in prior systems. Of course, by use of a separate preamp circuit
or path for each portal side receiver~ i.e., antenna, sensitivi~y is
optimized in any event, and even this basic actor has been lost upon
certain of the prior systems; this is all the more true when the par-
ticular preamp circuit path configuration, as described above, is taken
into consideration, since this optimizes the desired signal ~i.e.9
hannonic frequency bands with the least introduction of noise). Of
course, the summing (additive and subtrac~ive) described above is a
further and very substantial enhancement for systems such as those in
use or proposed heretofore.
Some of the reasons underlying the above-described differ-
encing of the co~non-mode noise signals from the two different sides of
a single port~l result from the fact that, in contrast to the true or
real ~e.g., pel~alloy) marker, most other objects or materials which
would have a low enough coercivity to generate harmonics of in~erest
(i.c., tending to mimic or mask a true marker) also have lo~ permea-
bility and only cause a perturbation in the interrogation field when in
close proximity to the portal sidesl i.e., to an interrogation ~ield-
generating coil or to an antenna or rcceiver. lhus, a non-marker object
carried through a portal causes a pertur~ation in the field as the
Gbject passes near the interrogation coil. At the same time, there are
many objects or structures ~hich may be present beneath the floor, or
example, wire mesh, concrete reinforcing rods, etc..., or in the ceiling
~e.g., ceiling grids) whicll cause receiver signals in the 30 to 50 kH2
. r,
!r ' .,
'` -15- 9L~ 3~3~
region, i.e.~ the third and fifth harmonic of the 10 kHz drive fre-
quency fundamental. Addi~ionally, even distortion in the capacitor
geometry and the coil geometry of the L-C field drive circuit are
likely, due to the high magnetic fl~Y densities, ~o produce receiver
signals in the 30 to 50 kHz region. I~ is desirable to lower this
background or ambient noise level, so tha~ field perturbations caused
by non-marker items passing through the portals very near the
interrogation field coil at one side would have the greatest detectable
effect, i.e., would be more easily and more reliably detected, and thus
discriminated out of alarm-causing effect.
To this end, the receiver tantennae) signals from opposite
sides of the same portal are "summed" (i.e., combined) in accordance
herewith such that during each particular interrogation field phase
condition such signals are arithmetically subtracted from one ano~her,
or differenced, that is, they will (at least partially) cancel out one
another. The signals coming from the two ante~lae are much alike, and
if the two signals, properly phased, are summed together so that one
cancels with the other, the result will be to lo~er the signal portion
attributable to "background" or environmental noise, thereby enh~ncing,
or highlighting, perturba~ion effects from objec~s located near one
portal side. On the other hand, a true marker causes a significant
perturbation even as it passes down the center of a portal, and it is
desirable to enhance those perturbation effects. To accomplish such
enhar.cement, the antenna signals from opposite sides of the same portal
are constructively ~algebraically~ added in the preamp and processor to
produce a dif~erently-constituted second composite or resultant signal
for subsequently processing, containing all of the perturbation effects
present at either antenna and therefore maximizing the result produced
by a real marker.
Thus, the present invention provides an appreciation and
realization that the only time a non-marker object i5 likely to produce
perturbations with a ha~onic response fairly closely mimicing that
caused by an actual marker is when the object is close to one of the
portal sides. Tha~ is, because of the comparatively low permeability of
t` 35 most non-marker objects, they do not have as large an ~ fect on the
interrogation field as the permalloy strip o a true marker does, and so
~- -16~ 9~
only generate significant harmonics when interrogated with ~ very strong
field. When so interroga~ed, a non-marker may ac~ually generate some of
the same harmonics as a marker, but not to the same extent, and not in
the same ratio of harmonic orders, and non-marker objects Yill thus be
S detected to a greater ex~ent when the object is close to one portal
side. It is important to realize, moreover, that perturbations caused
by non-markers do not have the same distribution or ratio of harmonicsJ
and that is why it is desirable to produce, and compare, the two dif-
ferent pre~np output signals, as done in accordance herewith. Further-
more, while a non-marker will theoretically produce the same distribu-
tion of hanmonics, or the same harmonic content, ~hether it is in the
middle of a portal or close to one side, the permeability of such an
object is likely to be such that its magnetic domains do not even
undergo switching by the interrogation field if the object is near the
center of the field, whereas a true marker will still undergo substan-
tial saturation and domain-s~itching under such circumstances. That is,
the strength of the interrogation field does differ across its width~
but the perturbation effect of any object is really a function of two
factors: first, coercivity, which for non-markers is mos~ likely not as
low as that of the real marker, requiring a stronger field to cause
domain-switching; second, non-marker objects do not have as high a
permeability as real markers, ~Id non-marker objects do not disrupt the
field as much when they do undergo switching. ~IUS, there is a double
effect as an object moves away from a sid~ of the portal, and the
effec~s caused by non-markers fade away very quickly.
The above-described separate outputs from preamp terminals P-5
and P-6 are separately and respectively appli.ed to the i~libit amplifier
3 and the ~nplifier/filter 2 noted above in conncction s~ith Fig. 1,
w~lere each such output is separately processed (basically, amplified and
frequency-shaped), and the resulting outputs are then separateIy supplied
to the detection logic and processin~ ~it 9.
~ lore particularly, the output signal from pre~p ter~inal P-6
is applied to input terminal A-l of the inhibit amplifier 3J a preferred
embodiment of which is illustrated for convenience in Fig. 3. BasicallyJ
inhibit amplifier i is preferably a tlio-stage band pass ~mplifier whos~
pass ba~d encompass~s primarily the third and fifth ha~onic, and
-17- ~L~ 98~3~
preferably the seventh as well, of the alternating interrogation field
frequency ~in the preferred embodiment already noted, 30 to 50, and up
to abou~ 70 kHz), which generally characterizes the lower frequency
spectr~n in the ratio used to critically identify the particular marker
element within the interrogation field~ as e~plained more fully herein-
after. It will be noted that both the input terminal A-l and the output
terminal A-2 are subject to switching by being coupled through the
complementary halves of a Ch~S analog switch S300, which may advan-
tageously be implemented by a single four-stage such switch, the two
complementary halves of which are shown for purposes of illustration at
different positions in Fig. 3 (i.e., one at the input and one at the
output). Additionally, the input tenninal A~l, after being switched
through swi~ch portion S5 of Ch~S switch S300(a), is coupled to a twin-T
notch filter 310 preferably having a variable resis~ance in both its
series-connected and parallel-connected branches. Like the twin-T notch
filter 44 noted above in connection with the pre~np 1) notch filter 310
is used for the purpose of further removing, i.e., diminishing, ~he lO
kHz fundamental frequency o~ the interrogation field, since the effects
of the latter are very strongly present in the receiver signals picked
up by the various antennae, and require substantial effort ~o properly
filter out for optimum sensi~ivity and selectivity in marker det~ction.
By the variable-resistance twin-T filtering CQncept no~ed, another 40 DB
of rejec~.ion in the residual level of the 10 kHz signal may be accom-
plished, with desirable results.
As noted, the input to the inhibit ~nplifier 3, and the output
from such amplifier and circuit, are both subject to switohing by the
c~nalog switch S300. ~lis switching is provided for bl~nking purposes,
during which the inhibit ~nplifier may be effectively removed from
operation at certain critical points in the operation of the system,
i.e., when the interrogation field-generating coils are receiv~ng their
drive pulses. Such blal~ing is accomplished by appropriately timed
inputs on C~S s~itch control tenninals SC-10 and SC-12, the first o
~hicll blanks the input cmd the 5econd of whicil bl~ks ~he output.
Signals for these two control te~inals are provided from the timing
generator 4 noted in connection with Fig. 1~ and may generally be
considered as pulse-t~e blanking signals whose pulse-w~dth determines
;~,
-18-
iL~ 9 ~3~
the time of circuit shutdo~n, the timing of the bl~nking signals being
synchronized to the application of the aforementioned excitation or
drive to the field-producing coils. In a more particular sense, the
blanking signal applied to control terminal SC-12, at the output of the
inhibit amplifier, is preferably about 50~ longer in duration than the
signal applied to switch control terminal SC-10, which blc~nks the input
of -this amplifier. In a particular sense, where the interrogation field
f~mdamental frequency is 10 kHz and one quarter-cycle ~during which time
the drive pulse is actually applied) has a duration of 25 microseconds,
a preferred input blanking period is on the order of 100 microseconds,
and a preferred output blanking period is 150 microseconds, both signals
synchronized to the drive pulse. By so doing, transients produced in
the amplifier as a result of switching will have been avoided, and both
the amplifier and the LC oscillating circuit will have undergone sub-
lS stantially complete settling, thus avoiding distortion effects which canbe very significant.
The output from the inhibit ~nplifier 3 comprises carefully-
timed bursts of the frequency range representing primarily the third,
fif~h and seventh harmonic of the interrogation field fund~nental, as
noted above, and this output from terminal A2 of the inhibi~ amplifier
is applied ~o input terminal DL-4 of the de~ector logic circuit 9 ~Figs.
4, 5 and 6), to be described further hereinafter.
The second output from the preamp l, namely that appearing on
its output terminal P-6, is applied to the amplifier/filter 2, noted
previously in connection with Fig. l. Although not illustrated specifi-
cally in the drawings, it should be understood that the ampffiltcr 2
has an input terminal to which the preamp signal (from output terminal
P-6) is applied. Preferably, the amp/filter 2 i5 a three-stage band-
pass device, having a single-ended output which is coupled to the
detector logic circuit 9. With respect to the preferred characteristics
of the amp/filter 2~ the three stages of amplification may all be
implemented by use of an L~-318 integrated CilCUit operational c~mplifier,
connected in a multiple-pole c~mplifying configuration with appropriate
frequency-shaping capacitance, centered upon the desired pass band
comprising the fifteenth hannonic of the fundamental frequency at which
the interrogation field is driven, in the embodiment contemplated here
T
~ -19~ 9~3~
approximately 140 kHz. In a preferred configuration, the first stage is
a high pass stage, Lhe second stage is a band-pass stage, and the third
stage is essentially a gain stage with both high and low alts. l~here
integrated circuit amplifier stages are used, each succeeding stage is
preferably coupled in complementary conductance configuration, ~ h
appropriate positive-negative-positive reference or biasing voltages.
As already indicated, the output from the amplifier/filter unit 2 is
coupled to the detector logic network 9, where it is inputted on termi-
nal DL-l.
Referring now to the detection logic net~ork 9, and initially
to Fig. 4 which illustrates the general nature of a preferred form
thereof, it ~ill be observed that this system has five discernible
branches, designated by the numerals 900, 910, 920, 930, and 940, which
are set apart from one another in this figure by dashed lines, for
purposes of illustration. 0f these, branches or sectors 900 and 930 are
essentially the same as one another from the standpoint of componentry,
although having very definite operational differences ~o be noted
subsequently. That is, both branches 900 and 930 embody a control
switch 10, 1?, respectively, a reference control ancl threshold compara-
tor set 14, 1$ and 16, 22, respectivelyJ and a driver, timer, and
indicator unit or circuit portion 22 and 74, respectively, each of the
latter having respective output terminals 21 and 25 as well as LED
signal elements ("~FD 2" and "LED 3", respectively). As further seen în
Fig. 4, the respective outputs from the threshold comparators 18 and 22
are also ~irected to an integrator 34, and thus are seen to be summed
with respect to one another; however, the particular manner in ~Ynich
such summing is carried out is an importc~nt aspect and is described in
much greater detail hereinafter
~ith continuing reference to the block diagram of Fig. 4, and
to the general at-tributes of detector logic unit 9, the center.circuit
portion 920 includes an ampli-fying and integrating, or integrating-
de-tector, circuit portion 30, ~hich receives an input from teI~linal DL-4
ancl has an output directed to a comparator and alarm 32 having ~n LED
indicator ("LED 1") as one oE its outputs The output from this alanm
is also fecl as an input to the lo~er circuit branch 940, more particular-
ly, to a driver, timer and alarm ~mit 38, whicll as indicated provides
-20-
~2~39~
an "Alarm Output No. 2". This same input terminal of the alarm unit 38
receives control signals on an input lead 39 connecting to the "signal
gate" and "noise gate" inputs fed to control switches 10 and 12 from
circuit input terminals DL-2 and DL-5. The second (upper) inpu~ terminal
of the driver, ~imer and alarm unit 3S is coupled back to the input side
of a discharge clamp 26 in path 910, whose primary input is f~om circuit
terminal DL-3. The output of the discharge clamp 26 is coupled to, and
directly affects, ~he integrator 34, and the integrator is coupled to,
and actuates, a comparator, timer and alarm 28 having a primary alarm
output directed to a lamp driver 29, which also provides a s~itched
alarm Output, labeled Alarm Output ~ \lso, timer and alarm 2~ controls
an indicator LED ("LED 4") coupled to its output.
Referring not~ in more detail ,o the detector logic circuitry
as depicted in Figs. S and 6, it will first be noted that the upper and
lower portions of the circuit, comprising channels 900, 910, 93Q and 940
in Fig. 4, are depicted in Fig. 5, ~hereas the central portion of the
circuit, comprising the path designated by the numeral 920 in Fig. 4, is
depicted separately in Fig. 6. In the preferred embodiment shot~n in
these Figures, the elements identified as "control switch 1" ~nd "control
switch 2" in Fig. 4, and desic~nated by the numerals 10 and 12 therein~
are sho~n to comprise input s~ntchin~ transistors Ql and Q2, whose bases
receive control inputs through resistors R8 and R9, respectively, from
circuit input terminals DL-2 and DL-5. Also, the bases of switching
transistors Q, and Q2 are coupled together through resistors R4 and R10,
and the junction of the latter two resistors is coupled to the low
voltage side of a pull-up resistor R21, and then through conductor 39 to
the positive or non-inverting side of an amplifier Ull in path 940. The
primary signal inputs to be switched by transistors Ql and Q2 are received
Oll circuit input terminal DL-l, which is coupled to the collector of
each such transistor through resistors R6 and R5, respectivelyr
The "reference control" componeIlts or units 14 and 16 of Fig.
4 are seen in Fig. 5 to comprise switches, e.g. transistors, Q4 and Q3,
respectively, ~hich are connected in emittel-follower configuration, and
whose bases are coupled together by a lead 49 so as to re`ceive a common
input, to be described subsequently. Ihe respective outputs from
transistors Q4 and Q3 are coupled as reference inputs to threshold
-21- ~L~ 3~39 6
comparators U-la and U-14a, and it is to be noted ~ha~ the circuit
arrangement of paths 900 and 930 is of an inverted configuration, i.e.,
the output ~rom transistor Q4 in path 900 is applied as an inverting
input to comparator U-la, whereas the output from transistor Q3 in path
930 is applied to the non-inverting input of comparator U-14a. Each
such comparator input also receives a particularly-set reference
voltage obtained from ~he junction of voltage-divider resistors R14 and
Rl, and applied through input resistances R16 and R3, respectively. The
respective opposite input terminals of threshold eomparators U-la and U-
14a receive inputs from the collectors of switching transistors Ql andQ2, respectively. These inputs are also supplied to comparators U-lb
and U-14b (which may be half of the same double integrated circuit
amplifier comprising comparators U-la and U-14a, respectively, for
example, an integrated circuit comparator No. 339). In essence, the
second comparators U-lb and U-14b are used as drivers for ensuing timers
and indicators U-2a and U-2b, whQse primary function is merely to time
out or an indicator drive signal of desired duration on respective
signal lamps LED 2 and LED 3, as described hereinafter. As indica~ed,
the two timers U-2a and U-2b are interconnected to one another9 and they
may in fact be implemented as the complementary halves of an IC5S6
timer, which is a double unit.
The lowermost circuit portion 940 of the detector logic
network 9 comprises in effect a comparator, acurrent source which drives
a ganged double-timer, and an amplif;ed l~np-driver output for alarm
signal purposes. More particularly, the initial comparator comprises
the aforementioned comparator unit U-ll, which may be implemented by use
of a 339 integrated circuit component. The comparator output is diode-
coupled to a transistor QS disposed in grounded-collector configuration
to act as a timed current source whose timing cycle is detennined by the
charge rate on capacitor C14. This current source drives the double-
timer U-i2a and U-12b, w}lich may advantageously be the two halves of a
No. 556 inte8rated circuit timer whose terminals are ganged in the
mam~er illustrated~ The first half of the timer, U-12a, is diode-
coupled ~D7) to a final amplifier or driver U-13 and lamp driver Q6,
driver U-13 being a further comparator component which may be imple-
mented by use of a 339 IC whose non-inverting input is supplied by the
22- ~L~ 3~3~
same signal ~ihich is applied to the inverting side of the first-stage
amplifier U-ll. Further, the first timer stage U-12a is connected to
(diode OR'd with) the second timer stage U-12b such that the first
stage, upon its initial excitation, immediately co~nences a continuous
lamp-driving operation of amplifier U-13 and switch Q6, as a "pot~er on"
indicator; ho~ever, whenever the current source comprising transistor
Q5 and its ~iming capacitor C14 reaches full charge, the second timer
(U-12b) is gated in and assumes control of the output signal, causing a
blinking of the signal lamp driven by driver Q6, for purposes noted
subsequently.
Generally speaking, the operation of that portion of the
detector logic circuitry described above is as follo~Ys. Input tenninal
DL-l receives the above-described output from the amp/filter 2, IYhich as
already pointed out comprises the arithmetically-summed antelma signa:Ls
from both sides of a given portal, or group of portals, after band-pass
amplification centered upon the fifteenth harmonic of the interrogation
field fundamental frequency. This signal is supplied equally to the
control switches Ql and Q2, whose switching operation determines whe~her
or not any portion of the supplied signal is gated ~hrough the switching
transistors to either path 900 or path 93Q. The latter two channels are
gated into and out of operation by timing signals applied to inputs DL-2
and DL-5, as supplied from ~he timing generator 4. The first such
input, to transistor Ql, is representative of the "signal gate" or
"marker signal ~indow", i.e., those particular increments of time
represen~ing an increasing-current condition ~both positive-going and
negative-going) in the interrogation field drive coils; conseq-lently,
these gate signals represent times l~hen a marker-present signal ~s
likely to be present in the signals from the portal antennae, if a
marker is in fact present ~ithin the interrogation field. Conversely,
the gating signals applied to terminal DL-5 and transistor Q2 represent
the opposite portion of the interrogation field alternations, i.e., when
marker-present signals are not likely to occur in the antemlae sign31s
even il a marker is present in the portal. Conse~uently, the gate
signals applied to te~ninal DL-5 define a "noise gate"~ i.e., a period
of time during which the signals received by the portal antennae~ on an
instantaneous basis, represent an actual measure of the e~is~ing noise
~_ -23-
~g~
level in the antennae signals.
It should be noted tha~, in accordance with this invention,
the duration o the aforementioned "noise gate" is shorter than the
duration o the "signal gate", and that there is a gap or interval
between the two. More particularly~ assuming the interrogation field
fundamental frequency to be 10 kHz, so that the duration of each quarter-
cycle is 25 microseconds, the marker-present signals are likely to occur
during the quarter-cycles when the current is increasing, eitheT posi-
tively or negatively, whereas the current-decreasing quarter-cycles
represent the condition when marker-present signals are not likely to
occur in the antennae signals. By maintaining the "signal gate" for a
full 25 microseconds but maintaining the "noise gate" for only approxi-
mately half that time, thus providing a gap of approximately 12 micro-
seconds between each noise gate and ensuing signal gate, distortion and
transients which otherwise would "ring through" the circuit will be
eliminated, thus further enhancing sensitivity and reliability. Of
course, the particular timing and synchronization of such signals are
also highly important. While the general state of the art includes
circuits and components well able to provide representative gating or
blanking signals of this type, a preferred form of timing generator is
a digital clock and divider, synchronized to the actual oscillation
conditions of the interrogation field.
Since the inputs to terminals DL-2 and DL-S occur at different
points in time, and in effect represent an alternating sequence, circuit
paths 900 and 930 of the detector logic 9 in effect alternate in opera-
tion, and during the period each is in operation it applies an input to
the aforementioned thresholcl comparators U-la (and U-lb) tin channel
900) and U-14a (and U-14b) (in channel 930). In so dGing, each such
circuit path functions to alter the charge state of an integrating
capacitor C9, and it is important to note that the two circuit paths act
oppositely from one another in that regard. That is, the switched input
from transistor Ql to comparator U-la in path 900 is applied to the non-
inverting (i.e., positive) input, whereas the opposite is true in path
930, where the signals gated through by switch Q2 are applied to the
inverting side of comparator U-14a. Consequently, the two such CiTCUit
paths act to raF)idly and sequentially apply increments of charge to, and
-
-24- ~L~ 9 ~3~
draw increments of charge from, integrating capacitor C9, on an alterna-
ting, increment-by-increment or pulse-by-pulse basis. As will be seen
hereinafter, these added and subtracted increments of charge are not
necessarily equal in magnitude, and the resultant charge state on the
integrating capacitor is thus cumulative with respect to time during
each "burst" of pulses, so long as they are of the same phase, as
described more fully hereinafter.
It is very i~nportant to note, in conjunction with the alter-
nating operation of circuit paths 900 and 930 noted just above, that the
signals gated through by transistors Ql and Q2 to comparators U-la and
U-14a work against variable reference levels, and ~hat these variable
reference levels are applied to the opposite-polarity input terminal of
each such comparator. That is, in circuit path ~00 the inverting input
of comparator U-la receives the variable reference level~ ~hereas in
circuit path 930 it is the non-inverting input of comparator U-14a which
receives the other such variable reference level. As already indicated
above, these variable reference levels both operate f~om the same
nominal or steady-state reference levelJ obtained from the junction of
voltage-divider resistors R14 and Rl~ through identical series resistors
R16 and R3. This steady-state reference level is subject to variation,
ho~ever, by the operation of transistors Q4 and Q3, which constitute the
"reEerence controls" 14 ~nd 16 noted in connection with Fig. 4. ~lat
is, in channel 900 the base of transistor ~4 ls controlled, in a manner
described more particularly hereina~ter, so as to vary the resulting
reference level applied to the inverting te~ninal of diferential
amplifier U-la. In channel 930, the steady-state re~erence level is
applied to the non-inverting input of comparator U-14a, and this nominal
reference level is made subject to variation by reference control 16,
i.e., transistor switch Q3, ~}liCh receives the same varying input as
transistor ~4, i.e., the base of each of these transistors is cor~nonly
coupled to receive the same control input signal (frcm the output of the
amplifier, peak-detector and integrator 30 in pat}l 920, sho~nn in Fig.
5). In the case of both circuit paths '900 and 930, the second-stage
comparatorsU-lb and U-14b, respectivelyl may be considered as in cssence
duplicative of the first such stage, insofar as inputs are concerned,
except that instead of applying and subtracting charge from integrating
-25- ~ g~9~i
capacitor C9, they are utilized to drive indicators LED 2 and LED 3,
which are pulsed by timer units U-2a and U-2b, to indicate the opera-
tional status of each such circuit path.
The second portion or channel 910 of the detector logic
network 9 is also illustrated in detail in Fig. S, and will be seen to
include a pair of inputs, a first one of which is provided by circuit
input terminal DL-3 which is coupled to the inverting input of a compara-
tor U-3, comprising the "discharge clamp" 26 noted in connection with
Fig. 4. The outpu~ of this comparator connects to the conductors 21 and
23 by which charge is applied to and removed from integrating capacito~
C9. Therefore, when an appropriate gating signal is applied to the
inverting side o~ comparator U-3, under general system conditions to be
noted subsequently, this comparator/amplifier will clamp integrating
capacitor C9 to ground, thus f~ly discharging the integrator. This in
lS effect terminates, and dissipates, the incrementally-accumulated charge
effect carried on for the duration of each different phase condition
present in the interrogation ield, as noted above. Therefore, each
time the interrogation field-inducing coils are to be switched from one
phase condition to another (for example, from an in-phase or phase A-A
condition to an out-of-phase, or phase A-B condition), an appropriate
pulse supplied from the timing generator 4 is applied to input terminal
DL-3, to fully discharge the inte~rating capacitor C9. Durin~ the time
each opposite phase condition exists (described previously as preferably
on the order of 16 msec, representing lS0 cycles of altema~ion) the
charge state existing on integrating capacitor C9 is continuously subject
to pulse-by-pulse change, depending upon the operational levels of
circuit paths 900 and 930, described above.
Animportant function of the detection logic and processor 9,
involving that portion thereof designatet generally as channel 910, and
particularly of that portion of the circuitry disposed to the ~ight of
portal point 911, is the production of a desired alarm or signal upon
the detected presence of the particular marker within the interrogation
field~ hlore particularly, it will be noted that the node or junction
911 ~here comparator U-3 interconnects with conductors 21 and 23, which
lead to the integration capacitor C9, comprises the signal input to a
comparator U-4, which receives a predetermin~d bias or steady-state
-26- 3L~ 3~3~
reference on its positive (non-inverting) input terminal from voltage-
divider resistors R13 and Rl5. Therefore, at any time the charge level
on integrating capacitor C9, representing the relative proportions o
higher-order harmonic content in the antenna signals versus lower-order
harmonic content therein, caused by an object producing perturbation of
the interrogation field, rises to a predetermined level~ established by
the reference applied to comparator U-4, this comparator triggers and
applies an alarm-causing output signal to tlle timer U-5 (which may be an
IC No. 555). Cne output of timer U-5 actuates an alarm signal ~LED ~)
and is also coupled to an output driver Q7, which may be used to drive a
signal lamp, sound an audible alarm, or the like, utilizing an ou~put
taken at terminal 912 connected to ~he collector of transistor Q7.
Furthers a switched output signal of timer U-5 which is representative
of the control signal applied to the base of transistor Q7 is available
on the output terminal designated 914.
Perhaps the most important of the many important functions of
the detection logic and processing network or unit 9, is carried out on
circuit path 910, sho~ in more detail in Fig. 6. As sho~n there, this
circuit path receives an input on terminal DL-4, which input comprises
the amplified, frequency-selective output from the inhibit ampliier 3,
noted generally in connection with Fig. 1 and more particularly described
in connection with Fig 3. This signal from the inhibit amplifier 3
comprises sequential, time-gated, synchronized bursts of the subtracted
(differenced) signals from the portal antennae, afler low-pass selective
amplification thereof in the inhibi~ ~nplifier. Consequently, this
input to the~detection logic circuit is representati~e o~ the low-
frequency component band (in the range of the third, fifth, and Up to
the seventh harmonic of the interrogation field, here on the order of 30
to 50, and approachir.g 70, ~Iz), which signal is attributable to an
object ~ithin the interrogation field. Whether that object is-an actual
marker, or merely some non-marker element causing perturbations in the
interrogation field, remains to be determined, but as already indicated,
the true or actual markers will have a relatively unique ratio or
balance of the high frequency component band with respect to ~he low
frequency band. This low frequency band is used in the detection logic
and processing unit 9 as a determinant which must be satisfied by the
-27-
lZ2g89S
magnitude of the high frequency band produced by ~he same object within
the interrogation field before a marker-present signal or alarm is
sounded; i.e., the amount (magnitude) of the low frequency band actually
encountered, as represented by the magnitude of the input applied to
terminal DL-4, is used to determine the required level ~hich the high
frequency band produced by the same object in the field must equal or
exceed if it is indeed an actual marker; the ratio or balance of these
frequency components for true markers being relatively uni~ue.
To achieve the above result, the aforementioned input on
terminal DL-4 is coupled to one end of a variable resistance or poten-
1:iometer R2, whose movable contact is coupled through a series resistor
R56 to the inverting input of a differential amplifier U-6 coupled into
the circuit as an inverting amplifier, whose gain is thus set by po-
tentiometer R2. Inverting amplifier U-6 ~which is preferably imple-
mented by use of a 3240 integrated circuit operational amplifier) formspart of the amplifier, detector and integrator unit 30 no~ed brie1y
al)ove în connection with Fig. 4; thusy the output of inverting amplifier
U-6 is diode-coupled through a series resistor R48 to the parallel
combination of a second inverting ampli~ier U-7 and an R-C integrating
network consisting of resistor R50 and capacitor C28. This over~ll
network in effect comprises a combination amplifier, peak-detector and
integrator, or in effect an integrating detector. That is, the charging
time-constant for capacitor C28 is a function of the voltage drop across
series resistor R48. Thus, the charge on capacitor C28 ~ill build
during the continuation of each burst of input signals applied to
te~inal DL-4 and passed by inverting amplifier U 6, with capacitor C28
integrating only the peaks of the negative excursions of the incoming
signals ~i.e., that portion of a cycle wllich exceeds the preset refer-
ence ]evel).
The peak-integration or inte~rating detector effect just noted
is preferably accomplished by maintaining the integration time constant
or capacitor C28 of very short duration, for example by utilizing a .1
microfarad capacitor for ~28 ~nd a 4.6 K-ohm resistor for R48. This
will produce a very fast~acting integrator which will operate in the
manner of a current source, i.e., integrating for only the irst few
excursions and tracking the applied signal very accurately and clo5ely~
-2~ 3~3~
yet reclucing the effects of narrow, high-amplitude spikes due to
switching tr~tsientS from the blanking ~gate-generating) and other
related circuitry. This type of detector is preferred since the band-
pass stages preceding it allow some of tile higher order components ~for
example in the range of 14n k}lz) to pass through, usually in the form of
spikes. Additionally, spikes may be created by the blanking circuitry,
as just indicated. If a more conventional peak-detector was used, such
spikes would result in a high level of detected signal, whereas the
preferred integrating detector responds more to the average value above
the diode (D-10) voltage drop. The level of the signal so integrated
appears on conductor 48, on the output side of inverting amplifier U-7,
and this signal level is not only coupled forward to a ccmparator U-8,
but is also reflected back (on conduc~ors 50, 47 and 49) as a threshold-
changing signal to transistors Q3 and Q4, (i.e., "reference contro j" 14
and 16) noted above in connection with Figs. 4 and 5.
The forwardly-coupled signal from inverting amplifier U-7 is
applied to comparator U-8 and, when this signal rises to a predetermined
level constituting a system override condition, comparator U-8 switches,
thereby energizing an indicator labeled "~FD 1", through a series
resistance RS2 and a level-setting resistor R53. This pTovides a
visual indication that the charge level on integrating capacitor C~8 has
- reached the override threshold voltage determined by comparator U-8.
Furthermore, the output of comparator U-8 is coupled to the inverting
input o differential amplifier U-9, to whose output is also coupled the
anode side of LED 1, and the resulting output from ~mpli~ier U-9 is
coupled to one input of a second inverting amplifier U-lO. ~le output
of amplifier U-10 is coupled back, on conductor 901, to the non-inverting
input of the aforementioned amplifier U-ll in path 940 (Fig. S~, whose
function has been described previously, and also coupled back ~on
conductor 39) to the bases of switching transistors Ql and Q2,-to bring
about system override, or lockoutJ as will bs noted subsequently.
Accordingly, it will be seen that the input to terminal DL-4
of channel 920, representing the lower-fre~uency spectrum produced by
the interrogation field-monitoring antennae, is utilized, with appro-
priate processing, to accomplish two distinct purposes. First, thepeak-detected and integrated reflection of this input is coupled back to
-29- ~ 9~3~
the bases of threshold reference-setting transistors Q3 and ~4, to
change the threshold level of comparators U-la and U-14a as a direct
function of the instantaneous level of the low frequency spect~m
produced by an object detected in the portals. Of course, the effect of
this is to chan~e in a very significant way the amo~ts of charge
applied to and accumulated on integrating capacitor C9. This directly
changes the relative conditions under which a logical decision is made
to either produce or not produce a marker-detection alarml through ~hat
portion of circuit path 910 coupled to node 911 and including comparator
U-4, timer U-5, and output driver Q7. That is, in the manner already
described in a qualitative sense above, the determination that an object
within the interrogation field is a genuine marker is made to be directly
dependent upon the relative proportion of the high frequency spectrum
(in the neighborhood of the fifteenth hannonic) with respect to the low
frequency spectrum (primarily third and fifth harmonic~ which that
object is producing in the interrogation field. In this manner, by
using the ratio of the detected frequency component bands as the re-
quisite detection criteria, substantial and accurate discrimination is
accomplished betwe~n actual markers and the myriad of other objects
which produce more-or-less analogous interrogation field perturba~ions
and which, if detected and indicated as being real markers, would
provide a false and erroneous output indicating the~t~ pilfering, or the
like where none was in fact taking place.
In accordance with the foregoing, it will now be appreciated
that the present detection system provides a multiple-step or multi-
layered approach for highly sensitive and yet hi~hly selective detection
of the lo~-coercivity permalloy or other such marker within the interro-
gation field, based upon the inevitably characteristic ~nd relatively
unique balance or ratio of low-order hanmonics versus high-order har-
monics caused by the magnetic domain-switching of the ma~ker in response
to each ensuing half-cycle of alternation of the interrogation field.
IYhereas many or even most metal objects will have some of the low-order
harmonic band, and may even have an appreciable quantity of the high-
order band, ew if any non-marker objects will have the same charac~er-
istic ratio of high order to lo~ order harmonic bands or componentgroupings; generally speaking, the higher-frequency harmonic b~ld
-30- ~L~ 9~3
will be deficient in objects and articles which are not true markers,
even through in a general sense substantial quantities of the higher-
order harmonics may indeed be present, particularly in objects and
articles ~hich provide multiple magnetic paths or loops and which
include at least some arcing points, i.e., gaps in the magnetic circuits.
l'hus, the invention provides a method and means to determine
the low-frequency components or band and the high-frequency components
or band of an object within the interrogation field, and these lo-~-
frequency components are used to dynamically control the detection
threshold of the high-frequency components which produce an alarm
signal. In so doing, the signals from the antennae monitoring the
interrogation field are carefully processed to p~oduce two different
types of signal output: one which represents the summation of the
signals from the antennae, for maximum sensitivity, and the other of
which represents the differencing of the signals from opposite sides of
the interrogation field, for maximum selectivity. These two signals are
separately processed to emphasize their respective high- and low-order
harmonic content9 and the signal with the high-order harmonic band is
time-sampled in a manner such that the resulting samples are likely to
accurately portray marker-presence signals on the one hand and marker-
absence or ambient-level (noise-level) signals on the other hand. The
resulting s~mples are then separately compared to a varying threshold
reference ~hich is pro~ided by a peak-integrated si~nal representative
of the detected object's low-order harmonic band, such that the higher
~he level o tlle latter signal, the higher the level which the marker-
present signal must have in order to bring about a threshold-crossing in
either of the two marker-present or marker-absent signal channels.
I~hatever threshold crossings do result from the foregoing
process are then in effect differenced and the result integrated cumu-
latively over the repeated cycles o~ the interrogation frequenc~y duringeach successive phase-related burst thereof. Should the resulting
integrated level exceed that indicative of the presence of a genuine
marker t~ithin the interrogation field, an alann is sounded~ Conver5ely,
if the peak-integrated signal representative o~' the low-order harmonic
band becomes sufficiently large to exceed a predetermined threshold,
indicating that the variable reference to ~ihich the high-order frequency
i` -
-31- ~L~2t38~3~
samples are compared has become prohibitively large and is, in effect,
blocking the detection channels, an indication of that status is given.
Initially, this indication results from energizing an indicator light
and, should the condition exist for a time period exceeding that at-
tributable to some unusual but nonetheless expectable occurr~nce, aflashing alarm is enabled (via detector logic channel 940).
The condition just described, indicative of an unusual and
undesirable situation prevalent within the interrogation field which is
causing a substantial overbalancing of the detection circuitry by way of
excessive levels of the low-frequency harmonic band, is a severe aber-
ration in the detection circuit parameters, and thus indicates the
advisability of fully inhibiting the detection circuitry, in addition to
the flashing lamp indication just noted which shows the existence of the
condition. Thus, the signal indicative of the low-frequency overbalance
which is coupled back to channel 940 for the purpose o~ enabling and
driving the flashing lamp indicator is also coupled back, on conductor
39, to each of the control switches 10 and 12 (transistors Ql and Q2)
such that ~hey latch out and bloc~ the input from terminal DL-l, thcreby
preventing any build-up on integrator 34 (capacitor C9) which might
otherwise result in an erroneous detection alarm.
In connection with the function and operation of detector
logic channels 900 and ~30, it is to be noted that the level of the
instantaneously-variable detection threshold or reference on comparators
U-la and U-14a, set initially by voltage divider R14 and Rl, and varied
by proportional or relative conduc~ion of transistors Q3 and Q4, is in
effec~ stored for a short time interval on capacitors C7 and C8, coupled
to the emitters of transistors Q4 and Q3, respectively, through a time
constant-setting resistor R16 and R3~ respectively. That is, the peak
levels of threshold variation due to conductance o transistors Q3 and
Q4 in response to elevated inhibit signals from integrating capacitor
C28, are stored on capacitors C7 and C8 between phases; thus, these
threshold peaks ~ill be held briefly when the interrogation field switches
its resultant fluY direction in response to reversal in the phase o~ the
drive excitation applied to one of the interrogation field-inducing
coils. Thus, the system will not be susceptible to error as a resul~ of
detection harmonic content levels which vary substantially ~rom one
~ _,, .
. ' .
'~ -32- ~L~ 33~;
phase condition to the ne~t. What is desired is to have enough stoTage
in the system so that relatively high inhibit levels built up during one
phase condition, which have effectively raised the threshold at the
comparators to a substantial degree, will be maintained after the change
in phase condition for at least the first half-cycle of the ne~t inter-
rogation field alternation and resulting detection signal, representa-
tive of a change in phase condition. For example, assuming the inter-
rogation field fundamental frequency to be the aforementioned 10 XHz,
utilizing a bleed-off time constant on the order of about 100 msec for
C capacitors C7 and C8 (the combined resistance of resistors R16 and Rl
for C7 and R3 and Rl for C8), storage will be provided for an interval
reasonably representative of the aforementioned period.
Of course, it is to be understood that the above is merely a
description of certain preferred embodiments of the invention, and that
various changes and alterations can be made without departing from the
underlying concepts and broader aspects of the invention as set forth in
the appended claims, which are to be interpreted in accordance with such
underlying concepts and broader aspects and by application of a full
range of equivalents.
