scispace - formally typeset
Search or ask a question
Journal ArticleDOI

Basic deposited integrated magnetic circuit element for fast computer circuitry

01 Sep 1965-IEEE Transactions on Magnetics (IEEE)-Vol. 1, Iss: 3, pp 211-217
TL;DR: In this article, a deposited configuration of a magnetic thin film and a coupling loop was studied with a view to the future development of integrated magnetic circuitry, and a line charge model, predicting flux linkages of coupling loops to an accuracy of about one percent was established.
Abstract: A deposited configuration, consisting of a magnetic thin film and coupling loop, was studied with a view to the future development of integrated magnetic circuitry. A line charge model, predicting flux linkages of coupling loops to an accuracy of about one percent was established. The almost complete linkage of the film flux with a deposited loop, due to the very close coupling, was verified. A decrease of about 20 percent in film flux at both ends of the easy axis was noted for the experimental assemblies used. Circulating loop currents were shown to be the chief parasitic factor which modified the switching of the magnetic film. The change in switching time due to eddy currents was small when the loop conductor size was of the same order as the magnetic film. For resistive loop loading, the average field during switching is a good measure of the slowing due to the loading. The film-loop assembly has good potentialities as a circuit element, with good transmission of both read-out and control signals occurring in the loop. The field calibration for these control signals was shown to be the same for both bias and drive field applications.

Summary (2 min read)

Basic Deposited Integrated Magnetic Circuit Element for Fast Computer Circuitry

  • Circulating loop currents were shown to be the chief parasitic factor which modified the switching of the magnetic film.
  • The film-loop assembly has good potentialities as a circuit element, with good transmission of both read-out and control signals occurring in the loop.
  • I n all tests, the magnetic drive field was applied perpendicular to the loop axes so that an air-flux compensation loop was not required for the external loop.
  • The frequency response of the measuring system was limited by the 0.8-ns risetime of the sampling oscilloscope circuitry, since the analog signal was slow compared to the response t,inw of the operational amplifiers in the integrator unit.

MAGKETIC F I L M FLUX DISTRIBUTION

  • The flux distribution of a rectangular magnetic film was investigated by observing t'he flux changes within various external and deposited coupling loops.
  • This method avoided errors due to inaccuracies in the forementioned quantities and also obviated the need for accurate calibration of the oscilloscope-integrator unit.
  • It also seems as if the two-dimensional approximation generally gives the better correlation with the experimental curve.
  • The variation of flux through the magnetic film and the average flux linkages of wide deposited loops were measured by flux switching experiments performed on a number of film assembly sets.
  • It can be seen that, of the four cases considered, the qua,dratic variation predicts the flux ratios with the best.

INTERSCTION OF R'IAGKETIC FILM AXD DEPOSITED

  • Film-loop interactions arise from eddy current's and circulating loop currents.
  • It is first assumed t'hat eddy currents have little effect on film switching since the loop conductors are not large compared to the film size and so flux-closure paths exist through nonconducting material.
  • In the preceding discussion, it was assunled that eddy current fields are small, since the loop conductors are about tjhe same size as the magnetic film.
  • The validity of this assumption was examined by the following experiment.
  • Figure 7 shows tJhe effect. of various loop conduct,ors on the switching time of the assembly.

FILM-LOOP ASSEMBLY AS A CIRCUIT ELEMEKT

  • Utilization of the film-loop assembly in high-speed circuitry depends on being able t,o use the loop bot'h as a highspeed read-out conductor and as a magnetic field source to control high-speed film switching.
  • Thus, although the presence of the deposited loop may have an appreciable effect on the film switching speed, t,he effect of the deposited loop transfer function on the observed waveform is quitre small.
  • In a few cases, the deposited loop swit,ching times were longer than t,he ext,ernal loop switching times and this result was attributed t'o large loop resistance, giving rise to excessive loop attenuation.
  • The result of such an experiment is illustrated in Fig. 8 , which shows the zero-bias swit'ching wavefornl and the two reduced-amplitude waveforms, obtained by separate application of the two bias fields.
  • The film flux output was obtained by taking the mean of the two flux outputs, when the film was switched from both directions of the easy axis.

CONCLUSIONS

  • The internal flux of magnetic films varies along the easy axis direction in a quadratic manner, a 20 percent decrease at both ends of the axis being observed for the assemblies investigated.
  • (Very close to the ends, departures may occur from this value because of the effect of local fields of reversed domains.).
  • Thus, significant miniaturization of integrated deposited circuitry is feasible without incurring appreciable flux loss by closure within the coupling loop (e.g., a film one mm2 having loop insulation layers 2.5 microns thick has about 1/4 percent flux closure within the loop).
  • The variation of film flux along the easy axis direction indicates the need for accurate deposition of the loops relative to the magnetic films if similar flux outputs are to be obtained from similar films.
  • This lather field decreases rapidly with film size and so is not a significant factor in miniaturized clrcuitry.

Did you find this useful? Give us your feedback

Content maybe subject to copyright    Report

WEE
TRANSACTIONS
ON
MAGNETICS
VOL.
MAG-
survey,”
IRE Trans.
on
Electronic Computers,
vol.
EC-9,
pp.
308-314, September 1960.
L.
C. Hobbs, “Comparison: major types of mass memories,’’
Data Systems Design,
vol.
1,
pp. 16-21, January 1964.
R.
Shahbender,
T.
Nelson, R. .Lochmger, and
J.
Walentine,
“MicroaDerture hiah-meed ferrlte memory,”
RCA Rev.,
vol.
23, pp. $39-566, 1362.-
V.
T.
Shahan and
0.
A. Gutwin, “Threshold properties of
partially switched ferrite cores,”
J. Appl. Phys.,
vol.
33,
pp.
C.
P.
Bourne and
D.
F.
Ford, “The historical development,
and predicted state-of-the-art
of
the general purpose digital
computer,”
1960 Proc. Western Joint Computer Conf.,
vol. 17,
”,
1049-1050, 1962.
pp. 1-21.
J.
I.
Raffel, “Future developments in large magnetic film
memories,”
J. Appl. Phys.,
vol. 35, pt. 2, pp. 748-753, March
1964.
C. Chong and
G.
Fedde, “Magnetic films-revolution in
computer memories,”
1961 Proc. Fall Joint Computer
Conf.,
E. Bloch, “The engineering design of the stretch computer,”
1959 Proc. Eastern Joint Computer Conf.,
pp. 48-58.
J.
K.
Shortle, et al., “TD memory,”
Elect. Equipment Engrg.,
vol.
11,
p. 16, June 1963.
J.
I.
Raffel, “Operating characteristics of
a
thin-film memory,’’
E.
Bittmann, “Thin-film memories: some problems, limitations,
and profits,”
Proc. 1963 INTERMAG
Conf.,
pp. 9-1-1 to
W. Jutzi,,, “Nondestructive read-out in thin magnetic film
memorles,
Proc. IEEE (Correspondence),
vol. 52, p. 875,
July 1964.
performance microelectronic,”
IBM
J.
Res. and Dev.,
vol.
8,
E.
M.
Davis, et al., “Solid logic technology: versatile, high-
pp. 102-114, April 1964.
R.
T.
Shevlin, Private communication on the utilization
of
n
storage matrix with a diode and a magnetic device per cell,
Courant Institute, New York University,
New
York,
N.
Y.
M. V.
Wilkes, W. Renwick, and
D.
J: Wheeler, “The,,design
IEE
(London),
vol.
105B,
p..121,1958.
of the control unit of an electronic digital computer,
Proc.
IEE (T,ondon),
vol.
110, pp. 157-166, January 1963.
D.
NI.
Taub, i‘A short revlew of read-only memories,”
Proc.
VO~.
22, pp. 213-224.
J. Appl. Phys.,
VO~.
30, pp. 605-615, 1959.
9-1-6.
.I,
NO.
3
SEPTEMBER,
1965
[26]
7-,
“Analysis of sneak paths and sense line distortion in an
Improved capacitor read only memory,”
Proc. IEEE,
vol.
51,
pp. 1554-1569, November 1963.
[27] L.
W.
Stammerjohn, “An evaluation of design and performance
of
the permanent magnet twistor memory,”
Proc. 1964 INTER-
MAG
Conf.,
pp. 8-4-1 to 84-6.
[28]
R.
E.
Matick, “Thick film read-only memory device,’’
J. Appl.
Phys.,
vol. 34, pp. 1173-1174, April 1963,
[29]
S.
W. Miller and
J.
L.
Haynes, “Investlgation of storage and
access techniques sultable
for
use in large-capacity digital
memories,’’ in
Large-Capaeity Memory Techniques
for
Computing
Systems,
M.
C. Yovits, Ed. New York: Macmillan, 1962.
[30] A.
V.
Pohm, et
,$.,
“Analysis of
lo8
element magnetic film
memory systems,
Proc. 1964 INTERMAG Conf.,
pp. 5-3-1 to
[31]
J.
D.
Carothers, et al., “A new high density recording system,”
5-3-5.
[32] G. Fan,
E.
Donath,
E.
S.
Barrekette, and A. Wirgin, “Analysis
1963 Proc.
Fall
Joint Computer
Conf.,
vol. 24.
of a magnet.0-optic readout system,”
IEEE Trans.
on
Electronic
Computers,
vol.
EC-12,
pp. 3-9, February 1963.
[33] W. E. Glenn and
J.
E,
Wolfe, “Thermoplastic recording,”
Internat’l. Sci. and Tech.,
June 1962.
[34]
R.
F.
M.
Thornley,
A.
V.
Brown, and A.
J.
Speth, “Electron
beam recording of digital information,”
IEEE Trans. on Elec-
tronic Computers,
EC-13, pp. 36-40, February 1964.
[35]
G.
W. King, “Data processing with the photoscore,” in
Large-
Capacity Memory Techniques for Computing Systems,
111.
C.
Yovits, Ed. New York: Macmillan, 1962.
[36] W. A. Baker, “The piggyback-twistor, an electrically change-
able non-destructive read-out twistor memory,”
Proc. 1964
INTERMAG Conf.,
pp. 8-5-1 to 8-5-6.
E371
J.
R.
Kiseda, et al., “A magnetic associative memory,”
IBM
J. Res. and Dev.,
vol.
5,
pp. 106-121, April 1961.
[38] A.
E.
Slade and
H.
0.
McMahon, “A cryotron catalog memory
system,”
1962 Proc. Eastern Joint Computer Conf.,
pp. 115-120.
[39]
17.
L.
Newhouse and R.
E.
Fruin, “A cryogenic data addressed
memory,”
1961 Proc. Southern Joint Computer Conf.
[40]
A.
Kaplan, “A search memory subsystem for a general purpose
computer,”
1963 Proc.
Fall
Joint Computer Conf.,
pp. 193-200.
[41] “Summary
of
investigation on associative memorles,” Com-
puter Command and Control Co., Philadelphia, Pa., Rept.
5-101-5, January 1964.
Basic
Deposited
Integrated Magnetic Circuit Element
for Fast Computer Circuitry
Abstract-A
deposited configuration, consisting of
a
magnetic
thin film and coupling loop, was studied with
a
view to the future
development of integrated magnetic circuitry.
A
line charge model,
predicting flux linkages of coupling loops to an accuracy
of
about
one percent was established. The almost complete linkage of the
film flux with a deposited loop, due to the very close coupling, was
verified.
A
decrease of about
20
percent in film
flux
at both ends of
the easy axis was noted for the experimental assemblies used.
Circulating loop currents were shown to be the chief parasitic factor
which modified the switching
of
the magnetic film. The change in
switching time due to eddy currents was small when the loop
conductor size was of the same order as the magnetic film. For
Manuscript received January 22, 1965; revised June
1,
1965.
versity College, Cork, Ireland. He was formerly with the Callfornia
B. C. Reardon
is
with the Dept. of Mathematical Physics, Uni-
Institute
of
Technology, Pasadena, Calif.
Pasadena, Calif.
F.
B. Humphrey is with the California Institute of Technology,
resistive loop loading, the average field during switching is a good
measure of the slowing due to the loading. The film-loop assembly
has good potentialities as
a
circuit element, with good transmission
of both read-out and control signals occurring in the loop. The field
calibration for these control signals was shown to be the same for
both bias
and
drive field applications.
T
HE
GENERAL TREND in computer design towards
machines having large memory and logic arrays and
having short operating times is further characterized by
the use
of
miniaturized circuitry with low power dissipa-
tion. The potentialities of magnetic thin films, having short
switching times for moderate drive fields, will be greatly
enhanced by the development of integrated circuitry
[l
1,
[a],
that is, circuitry of magnetic films and associated
electrical connections made on
the
same substrate.
A
funda-
211

212
IEEE
TKhNSlCTIONS
ON
MAGNETICS
mental element', consisting of a magnetic thin film closely
coupled by, but insulated from, a deposited loop, is the
basis of almost all of the more complicated circuitry.
It
is
important, therefore, to study this fundamental element in
detail in order to understand the more complicated cir-
cui
ts.
The close coupling of deposited
loops
around magrletic
films results in almost complete linkage with the magnetic
film flux but, owing to variation of the flux within the film,
the flux linkage varies with the relative position of the loop
and rnagnetic film. The interaction between a switching
magnetic film and a deposit>ed coupling loop is primarily
due to the effect of circulating loop currents. The effect of
eddy currents is quite small, unless t.he loop conductor area
is
large relative
to
the film area. Good transmission of both
read-out and control signals occurs in the loop. The loop
bias and drive fields can be accurat,ely predicted from the
loop dimensions.
FILM
ASSEMBLY
AKD
TEST
EQUIPMEYT
The magnetic film and coupling
loop
assemblies
were
made four at a time by vacuum deposition. With the glass
substrate heated to
300°C,
the assembly was formed from
layers of Pernlalloy
(83
percent nickel,
17
percent iron),
aluminum (conductor material), and silicon monoxide
(insulator material). Deposition of the five layers was
effect'ed without opening the vacuum in the system.
Static measurements were made
on
each film before it was
accepted for high-speed switching experiments.
It
was re-
quired that the static
B-H
loops along both the hard and
easy axes be similar to those for the other films of the set.
Deposited loop resistance was recorded with acceptable
values in the range
of
1
to
10
ohms. The insulation re-
sistance between the Pernlalloy and loop was determined;
in many cases, where the resistance between the
two
con-
ductors was less than
1
kQ, presumably caused by pinholes
in the silicon monoxide layers, the resistance could be
increased
t'o
an acceptable value
of
about
10
1tQ
by apply-
ing a potential difference of
20
volts across the insulation.
A
projective view of a complete film assembly is shown in
Fig. l(a). The standard film size
(0.45
inches
X
0.36
inches
X
1500
amperes) is shown, together with a loop
conductor
0.03
inches wide, which, because
of
equipment
limitations, was the minimum width loop attainable.
The t,est equipment, utilized a high voltage dc supply,
discharge ca,ble, and mercury relay system to generate a
fast-rise high current pulse. The magnetic field was
generated in a strip-line section having a center conduct'or
1.2
inches wide and
0.05
inches thick, situated symmetri-
cally between two ground planes
1
inch apart. Using a
maximum charging voltage of
5
kV,
a drive field of
6.5
Oe
was obtained in the test region centered direct'ly beneath
the center conductor. The duration of this field pulse was
500
ns.
It
had a rise time of about
1
ns
and a repetition
rate of
60
pps. The two loops, deposited and external,
were arranged as shown in Fig. l(b). The spring-mounted
deposited loop probe, consisting of a ceramic-covered
copper wire, pressed against the end tab
of
the deposited
LOOP
CONDUCTOR
4000A
25000A
MAGNETIC
FILM
1500A
LOOP
GROUND
PLANE
40006
i
-0.45"L/
(a)
ELECTROSTATIC
SHIELD
(b)
Fig.
1.
Film assembly and coupling
loop
configuration.
loop conductor, while the ground plane of the loop
was
held in contact with the strip-line ground plane by clamp-
ing the substrate into position. In all tests, the magnetic
drive field was applied perpendicular to the loop axes
so
that an air-flux compensation loop was not required for
the external loop.
The output of either loop, displayed on a sampling
oscilloscope, could also be integrated with respect to time,
by means
of
an integrator unit which compensated for
drift in the analog out,put of the sampling oscilloscope.
The frequency response
of
the measuring system was
limited by the 0.8-ns risetime of the sampling oscilloscope
circuitry, since the analog signal was slow compared to the
response t,inw of the operational amplifiers in the integrator
unit.
In all experiments, the magnetic drive field was applied
along the hard axis
of
the film, with a reset field applied
between pulses to set the magnetization in one direction
along the easy axis. When the drive field exceeded the
anisotropy field
Hg,
the final position of the film magnetiza-
t'ion, with t.he drive field still present, was assumed to be
along the hard axis. When complete rotation to Ihe hard
axis was desired, as in the meslsurement, of flux change mag-
nitudes, a drive field of about
2Hk
was used to ensure this
condit'ion.
MAGKETIC
FILM
FLUX
DISTRIBUTION
The flux distribution of a rectangular magnetic film was
investigated by observing t'he flux changes within various
external and deposited coupling loops. All experimental
flux change measurements mere expressed as flux ratios,
rather than relating them to theoretical flux values, derived
from film magnetization, thickness, and width. This
method avoided errors due to inaccuracies in the fore-
mentioned quantities and also obviated the need for ac-
curate calibration
of
the oscilloscope-integrator unit. (In

1965
REARDON AND HUMPHREY INTEGRATED CIRCUIT ELEMENT FOR FAST CIRCUITRY
213
a few cases not treated explic4tly in this paper, correlation
between experimental and theoretical flux values was
obtained to about
20
percent). Comparison of the flux
change within an external loop, whose
loose
coupling
results in an appreciable amount of flux closure within the
loop, to the flux change within a deposited
loop,
which
is assumed to have negligible closure flux, was used to
study the external flux distribution. Comparison of flux
changes within various deposited loops associated with the
same magnetic film
was
used to study the int'ernal flux
distribution.
A
model, previously used by Oguey
[3],
for the external
flux distribution of a single donlain rectangular magnetic
film having its magnetization directed parallel to one edge
of the film, can be obtained by representing the film mag-
netization by two uniform nlagnet,ic line charges of
op-
posite polarity. These line charges lie along the
two
op-
posite edges of the film, at either end of the single domain
and extend the length
of
the film, perpendicular to the
direction
of
magnet,ization. The flux linkage of a coupling
loop having its axis parallel to the film magnetization
equals the internal film flux
pm,
less the closure flux
'po
within the loop. When the film magnetization is switched
through
YO",
the change in flux linkage, taking account
of
the eddy currents
[4]
induced in the thick ground plane
[Fig. l(b)], is given by
(cp,
-
Zcp,).
Using the film flux
model just presented and the dimensions shown in Fig.
2,
the ratio
R
of flux change t,o film flux for a loop situated on
the
film centerline can be shown
[4]
to be given by:
R
=
1
-
2~c/~m
where
2 2hx
IA(x)
=
-
tan-'
1
1[(2/2)2
+
h2
+
2211'2
LINE
CHARGE
/
+MAGNETIC
FILM
Fig.
2.
Rectangular magnetic
film
and coupling loop.
ratio of the integrated output voltage to t'he integrated in-
duced voltage depends only on the steady-state attenuation
of the loop and its termination impedance, provided the
loop
is
linear. When the loop has zero shunt conductance,
this ratio is expressed by
TO/
(TO
+
rJ,
where
ro
is the steady-
state resistance
of
the loop termination and
rz
is the total
series loop resistance. The predicted value of the ratio of
the output fluxes of an external and deposited loop
F,
is
given by
where
ro
=
termination resistance
(53.5
ohms)
Td
=
deposited loop resistance
re
=
externd loop resistance (0.118 ohms).
The experimental values of
F,
were determined
for
a
number
of
film assemblies, having
a
narrow loop on the
film
centerline. These values, expressed as a function of the
By assuming a two-dimensional external flux pattern, an
approximate value for
R
is
obtained, which is within a
few
percent of t8he exact value over a wide range of film and
loop dimensions. This value
is
given by
As
can be seen from the foregoing equation, the flux clos-
ure wit'hin a deposited loop (where
h
<<
1)
isnegligible com-
pared to Dhe film flux, and
so
the value of
R
also gives the
ratio of flux change in an external loop to that occurring in
a
deposited loop, when both loops are situated on the film
centerline.
The change in flux linkage of a loop can be measured by
integrating the loop output voltage with respect to time.
Although the output and induced voltage waveforms may
differ, because
of
the effect
of
the
transfer voltage ratio
of
loop circuitry, it has been shown theoret.ically
[5]
that the
deposited loop resistance, with the film dimension
1
parallel
to the loop axis as a parameter, are shown in Fig.
3.
The theoretical values of
F7,
corresponding to both the
exact flux rat'io and its two-dimensional approximation, are
also shown. The dashed straight lines, representative of the
experimental data, have almost identical slopes to the
corresponding theoretical curves, thus verifying the de-
pendance of
F,
on the deposited loop resistance.
It
also
seems as
if
the two-dimensional approximation generally
gives the better correlation with the experimental curve.
As
will be shown later, this better
fit
arises from a cancella-
tion of the error due to the two-dimensional approximation
by the error due to the simple line charge model.
The variation of flux through the magnetic film and the
average flux linkages of wide deposited loops were measured
by flux switching experiments performed on
a
number of
film assembly sets. The various width loops and their loca-
tions relative to the magnetic film are indicated in Fig.
4.

214
IEEE
TRANSACTIONS
ON
NIAGSETICS
SEPTEMBER
X
31
J
LL
.................................................
o,4
....................................
a
2
0
3
1
_________
:-_-0-
__-----
a
------
}e=o09"
I-
0
0
I
I
1
,
,
, ,
1
2
3
4
5
6
7
DEPOSITED
LOO?
RESISTANCE
(a)
Fig.
3.
Output flux
rat,io
F,
as
a function of deposited loop resistance
with the film dimension
I
parallel to the loop axis
as
a parameter.
The solid and dotted curves are based on the exact
flux
ratio and its
two-dimensional approximation, respectively. The dashed curves
are representat,ive
of
the experimental data.
0.03"
0
03"
0
03"
Fig.
4.
Film assemblies for meauxlrement
of
flux variationLin
mag-
net.ic
film
and flux linkages
of
wide loops.
TABLE
I
Aksemblv
12
n
RE
1
2
3
4
1.000
0.780
0.818
0.521
0.973
0.782
0.43'7
0.396
Since the films of each set were deposited at the same time,
they are presumably identical
so
that flux measurements
on the four
films
of each set can be directly compared t'o
each other. The mean of the results obtained from a num-
ber of film sets are shown in Table
I
(all flux measure-
ments have been corrected
for
the effect of loop resistance).
The quantity
RD
is the rat'io of deposited loop flux change
for each assembly to the deposited loop flux change for
assembly
1.
The quantity
RE
is the ratio of the external
loop flux change for each assembly to the deposited
loop
flux change for assembly
1.
The values of
RD
indicate that the flux through the
magnetic film varies along the easy axis direction, being
about
20
percent less near the ends of the film relative
t'o
the value on the film centerline. The values of
RE
for
assemblies
1
and
3
indicate that t,he presence of the de-
posited loops has no effect on the flux change of the mag-
netic film, and the decrease in
RE
as the external
loop
moves from the centerline
t,o
the edge of the film is
con-
sist,ent with the line charge model predictions.
Assuming linear, quadratic, and cubic flux variations
along the easy axis direction [i.e., flux at distance
z
from
film centerline
=
$0
(1
-
Kz"),
where
40
=
flux on film
centerline,
n
=
1,
2,
3
for linear, quadratic, and cubic
variations, respectively], the average flux linkage of a loop
of any width can be determined. By calculating the
RD
value for assembly
2
for each type
of
flux variation and
using the experimental value of
R,
for assembly
2
to deter-
mine the appropriate values of
K,
the corresponding
values
of
RD
for assemblies
3
and
4
were calculated. In addi-
tion, by using systems of line charges which correspond
to the foregoing flux variations, the value of
RE
for
assembly
1
was determined. The results of the foregoing
calculations together with the corresponding
RD
and
R,
values for a constant easy axis flux distribution are coni-
pared witah the experimental values in Table
11.
TABLE
I1
FLUX
VARIATION
Quad- Experi-
Constant
Linear ratic Cubic ment.al
RD
(assembly
3)
1.000
0.916
0.940
0.955
0.9'73
RD
(assembly
4j
0.500
0.42'7
0.437
0.445
0
43'7
RE
(assembly
1)
0.818
0.767
0.785
0.791
0.780
It
can be seen that, of the four cases considered,
the
qua,dratic variation predicts the flux ratios with the best.
accuracy, being slightly better than the cubic variation in
this regard. In the quadratic case, the flux varies by
21.8
percent along the easy axis direction, and the line charge
model consists of two uniform line charges at both ends
of the easy axis, of the strength
A0.782
relative to the
total film flux, xnd a uniform line charge distribution of
strength 1.744
z/Z2
at a distance
z
from the film center-
line.
INTERSCTION
OF
R'IAGKETIC
FILM
AXD
DEPOSITED
COUPLING
LOOP
Film-loop interactions arise from eddy current's and cir-
culating loop currents. The importrance of these fields in
modifying film switching was investigated experimentally.
Switching experiments were first performed on the set of
film assemblies depicted in Fig.
5
which have similar mag-
netic films but various loop configurat,ions. The result,s are
shown in Fig.
6.
At
any given drive field, the swit'ching
time becomes progressively longer as additional loop cir-
cuitry is added. The change in switching time as the loop

1965
REARDON .4XD HUMPHREY: INTEGRATED CIRCUIT ELEMENT FOR
F.1ST
CIRCUITRY
215
L
0
45"
>
Fig.
5.
Film assemblies for study of effect
of
loop loading on film
switching. All loops are the full width
of
the film. Film A has no
loop conductor; film
B
has
a
loop conductor which is not connect,ed
to t,he ground plane conductor; film
C
has a loop conductor con-
nected to the ground plane conductor at one end; and film
11
has
a
loop conductor connected to the ground plane conductor
at
both
ends.
0
oe
I
oe
2
oe
3
OP
4
oe
.
5
oe
0
I
I I
I
3
4
5
6
7
DRIVE
FIELD
(Oe)
Fig.
6.
Variation of film inverse-switching time as
a
ftmction of
drive field with the loop loading as
a
parameter. The definition of
switching time is given in the inset, the total film
flux
being meas-
ured by switching the film at maximum drive field and the decreased
flux
switched being obtained for fields
of
the order
of
Hk(
~4.5
Oe)
due to incomplete rotation
of
the film magnetization to the hard axis.
circuitry is modified can be explained in the following
manner.
It
is first assumed t'hat eddy currents have little effect on
film switching since the
loop
conductors are not large com-
pared to the film size and
so
flux-closure paths exist through
nonconducting material. The change in switching time
due to the deposited loop can then be attributed to the
effect
of
circulating loop currents.
The electrical lengths of the loops are short compared to
the widths
of
the switching pulses, and
so
the loop loading
is considered to consist of
a
lumped impedance terminating
a short loop, which has a voltage induced along its length.
For assembly
Dl
this loading is considered to be purely re-
sistive, since the loop capacitance is ineffective in a shorted
loop and the
L/R
time constant is small
(<
0.05
ns). The
resistive field always opposes the change in film magnetiza-
tion
so
that t'he average field
is
a measure of the total
slow-
ing of the switching due to loop loading during the entire
switching time.
It
can be shown
[4]
that this average field
H(O
-
T)
is given by
H(O
-
T)
=
pT/TRw
(4)
where
pT
=
total flux change in the direction of the loop
axis,
T
=
total switching time,
R
=
loop resistance, and
21:
=
loop width. This field acts along the loop axis and thus
t'he conditions of assembly
D
can be simulated by switching
assembly
A
with various steady fields applied in this
direction. The results of such an experiment are shown by
the dashed lines of Fig.
6,
where it is seen that the simulated
average field
Ha,
decreases with increased switching time.
According t'o
(4),
the value of
Havr
should remain constant
as t.he drive field
is
changed. Values of
Havr,
corresponding
to various
1/~
values for assembly
Dl
were obtained by
interpolation between the bias curves for assembly
A.
These field values, together with the corresponding values
of
Hav7,
are shown in Table
111.
_e..
TABLE
I11
31/3
188 2.10 11.20
4
251 2.32 9.25
42/3 300 2.70 9.00
51/3
345
2.96 8.58
389
3.27
8.40
425
450
~~~
3.61
3.89
8.50
8.65
It
can be seen that the variation of
Havr
is quite snzall,
especially at' the higher field values where the loop loading
simulation is most accurate.
The switching curve for assembly
D
was predicted by
assuming that
Ha,7
remained const'ant at all switching
times and that, in the middle of the range
(Ha"
=
3
Oe),
the predicted switching time was identical wit'h the
measured value. The predicted curve for assembly
D
is
shown in Fig.
6
(dotted curve) and is seen to be close to
the experimental curve
over
the entire range
of
drive
field values. Using the measured value of
1/r
correspond-
ing to
Ha,
=
3
Oe, in
(4),
a loop resistance of
0.28
ohms
was obtained in comparison to a theoretical de value of
0.21
ohms. The discrepancy between these two resistance
values is chiefly attributable to the presence of oxide
layers, frequently encountered in the film assemblies,
between the loop conductor and loop ground plane, thus
increasing the effective loop resistance. (The increase of
resistance due to high frequency effects is negligible since,
at
1
K
Mc/s,
the skin depth of aluminum is
27
000
amperes, which is much greater than the
3500
amperes
thickness
of
the loop conductor.) The comparison be-
tween the two resistance values is considered t,o be reason-
able especially when the effects
of
possible additional

Citations
More filters
Journal ArticleDOI
TL;DR: In this paper, the authors investigated the effects of magnetization reversal on thin permalloy films using pulse techniques and vector locus configurations and found that wall motion is predominant when no transverse bias field is applied.
Abstract: Dynamic and nearly static magnetization reversal mechanisms in thin permalloy films are investigated experimentally using pulse techniques and vector locus configurations. At least for the driving field strength used, easy-axis switching waveforms indicate that wall motion is predominant when no transverse field is applied. At a given transverse bias field, the simultaneous pick-up signals from aligned and crossed loops show that the voltage-time integral at zero crossing time of the transverse signal becomes dominant for increasing drive field. The complicated irreversible magnetization phenomena on the astroid are illustrated experimentally on the coordinate system by the vector locus for a 10 kc/s sinusoidal driving field and pulse field having 0.5 ns rise-time. Wall motion and rotation during flux reversal are clearly distinguished on these configurations for various combinations of externally applied fields. The critical angle for coherent rotation is in good agreement with that derived from the Stoner-Wohlfarth model at a 10 kc/s sinusoidal field. However, for excess driving pulse fields, the dynamic vector locus suggests that until the walls nucleate and start to move, the coherent rotation continues over the critical angle suggested by the astroid. This gives a clear answer as to the cause of the nonlinearity on the plots of the inverse reversal time vs. driving field with the transverse bias field as a parameter.

10 citations

Journal ArticleDOI
TL;DR: In this article, the fraction of magnetic thin-film flux which links a rectangular coupling loop is calculated for rectangular, square and circular film configurations based on a magnetic line charge model for the film external flux distribution.
Abstract: The fraction of magnetic thin-film flux which links a rectangular coupling loop is calculated for rectangular, square and circular film configurations. The calculation is based on a magnetic line charge model for the film external flux distribution. The results are presented in a graphical dimensionless form.

1 citations

References
More filters
Journal ArticleDOI
01 Jan 1961
TL;DR: In this paper, a simple rotational model modified to include the effects of wall switching and dispersion of the preferred direction of magnetization provides a basis for describing properties of engineering interest.
Abstract: Thin magnetic films of permalloy have characteristics ideal for high-speed digital storage. A simple rotational model modified to include the effects of wall switching and dispersion of the preferred direction of magnetization provides a basis for describing properties of engineering interest. A selection system has been chosen which allows great latitude in film uniformity. Production of films with magnetic properties uniform to within ± 10 per cent is readily achieved. Specifications for operation in a destructive mode can easily be met by existing film arrays; the nondestructive mode is considerably more stringent unless very small signals can be tolerated. The first film memory has been in reliable operation since the summer of 1959. It has 32 ten-bit words and has been operated with a minimum cycle time of 0.4 , ?sec. Higher speed and larger capacities will require higher bit densities and improved techniques to minimize undesirable coupling between drive and sense lines. The use of 10 × 60 mil rectangles, balanced sense windings, and longer words will hopefully permit memories of about 200,000 bits with cycle time under 0.2 ?sec.

56 citations

Journal ArticleDOI
TL;DR: In this article, the authors examine the two main difficulties encountered in the design of a sensitive hysteresis loop tracer for thin magnetic films are the flux calibration and the reduction of noise.
Abstract: The two main difficulties encountered in the design of a sensitive hysteresis loop tracer for thin magnetic films are the flux calibration and the reduction of noise. The aim of this paper is to examine their nature and to show the possible solution for the design of very sensitive hysteresis loop tracers. The study of the flux distribution around a thin magnetic film specimen permits determination of the merits of various pickup coil configurations, as well as the form which optimizes the signal‐to‐noise ratio. The various disturbing voltages and the ways to eliminate them are examined. Optimization of the amplifier noise figure, proper choice of the integration network, dc restoration, and hum synchronization are described for the reduction of the output noise after integration and amplification. Two instruments built according to these principles are outlined. The first has a single wire pickup and is well suited for measurement of the flux distribution around a thin magnetic film and for experiments in vacuum at elevated temperatures; the second is more flexible and sensitive. By using different pickup coils covering a frequency range from 50 cps to 10 kc, its sensitivity is sufficient to measure flux values of 2×10−12 v‐sec at a frequency of 500 cps.

55 citations

Proceedings Article
01 Jan 1959

6 citations

Journal ArticleDOI
TL;DR: In this paper, it was verified that the extremely close coupling between a magnetic thin film and a surrounding deposited loop in this configuration results in 100% flux linkage with the loop, and the dependence of the observed flux linkage of the deposited loop on the loop circuitry was studied.
Abstract: Future applications of magnetic thin films to computer circuitry will require the use of integrated deposited circuitry, which involves depositing magnetic films and associated electrical circuitry on the same substrate. It is verified that the extremely close coupling between a magnetic film and a surrounding deposited loop in this configuration results in 100% flux linkage with the loop. The dependence of the observed flux linkage of the deposited loop on the loop circuitry is studied; the effect of width, thickness, and separation of deposited loop conductors on high-speed film switching is investigated; and the relative importance of circulating loop currents and eddy currents in the conductor materials in slowing film switching is evaluated.

2 citations