Note: Descriptions are shown in the official language in which they were submitted.
~LA~ 7~99
This invention relates to a method of calibrating a machining sensor
lying on a surface, a ~rst edge of which is to be machined from its initial
location to an ideal final position line having a predetermined spacing from
a feature line.
In certain manufacturing operations, particularly those for fabric-
ating disc memory thin-film magnetic heads in situ on the air bearing slider
to be carried by the head arm, it is desirable to machine the flying surface
until a precisely located line on another surface intersecting the flying
surface becomes the line of intersection of the two surfaces. In the thin-
film head example, the head is carried on an end face of the slider which
is approximately perpendicular to the flying surface, and the line is positionedto specify very accurately the thin-film head's throat height, that is the
dimension of the flux gap normal to the transducing surface. (The trans-
ducing surface, of course, is nearly parallel during disc memory operation,
to the medium surface.) Accuracy in throat height to within a few tens of
microinches is desirable to insure optimum electronic and magnetic character-
istics. Machining the flying surface until it coincides with the desired
line of intersection then automatically sets throat height to the accuracy
with which the line of intersection was set.
Controlling this dimension during fabrication has always been a
difficult problem because of the extremely small dimensions and tolerances
inuolved. Simply using the top of the slider prism as a reference surface
for controlling throat height was satisfactory when grinding ferrite heads,
see United States Patent No. 3,982,318. But tolerance and dimensions are
much larger in ferrite head technology.
Respecting thin-film heads, recent innovations allowing accurate
.
-- 1 --
~ ~ ~t7 ~ ~ ~
control of throat height involves the use of so called lapping guides or
machining sensors, e.g., as disclosed in IBM Technical Disclosure Bulletin
(rDB) Vol. 23, No. 6, November 1980, p. 2550. These guides or sensors are
deposited on conducting materials placed on the surface carrying the thin-
~ilm head. Two types of sensors are in general use. So-called discrete
sensors simply have their electrical continuity broken at some point during
nachining and hence, provide an indication o~ machining progress at only a
single instant. Analog sensors have an area of resistive material which is
slowly removed by machining and hence provide a continuous indication until
continuity is broken. With respect to discrete sensors, typically several
at different heights are employed. The continuity of each is successively
broken by the machining process, thereby providing a series of indications
of precisely how much more machining must yet occur to reach the desired
final position line. At the limits of or within the desired throat height
range, a last sensor's conductive path will be opened signaling that the
machining process should stop.
The use of these machining sensors drastically improves the accuracy
with which the edge can be positioned relative to the feature However, when
dealing with thin-film magnetic heads, one cannot form conventional machining
sensors with the same step which defines the throat of the gap. This is
because the throat is formed by the deposition of an insulating layer, whereas
thc machi.ning sensors are conductive patterns and hence are deposited in the
stops creating tlle magnetic legs of the head. It is a known difficulty that
successi.ve layers oE material deposited by the use of photo-optic masks and
Eornling a composite thin-film structure cannot be registered with respect to
eElcll other with perfect accuracy. That is, the masks or patterns which define
~)'71~
each of the -.Eeatures of successive layers such as khe bottorn
leg, the throat and the ~op ley, cannot be placed in precise
alignment with the pattern.s cxeated by pxeviou~ ma~kiny ~ep~
duriny typical ~lanu:Eacturiny operations. ThereEore, -~he -throat
heiyht o~ a typical thin-~ilm hea~ canrlot be controllecl to an
accuracy greater than the reyistration between kh~ throat in-
sulation-forming pattern and the maynetic leytmachininy sensor-
forming pattern. Experience shows that this inherent inaccuracy
results in a substantial percentage of head gaps which have
throat heights outside of the required tolerances. Worse
still, even though the throa-t height-defining step occurs
intermediately in the process, one cannot easily tell whether
or not the head is good until the manufacturing process ic com-
plete, making the relatively high number of reject heads an
expensive flaw in these previous systems.
The problem of aligning machining sensors with a
feature formed of insulating material such as -the throat de-
fining layer of a thin-film head is present for both discrete
and analog sensors. In a current manufacturing process, analog
~o sensors are used to indicate the progress of machining of a
workpiece carrying several thin-film heads. The machining s-tep
sets the throat heights for all the thin-film heads simultane-
ously. An analog sensor is interposed between each pair of
heads. It is necessary that the position of each analog sensor
vis-a-vis its adjacent heads be known very accurately so that
machining can be halted when the throat heights of as many
heads as possible are within the desired tolerances. (Due to
various inaccuracies in the process, it is possible that not
.~ J -3-
1~6.3~789~3
all throat heights can be re~uae~ to a ~a~ue ~Jlthin ~he t~l~x-
ance range at the same time.)
_M ~1DB Vol. 18, ~o. l, ~une 1975, p. ~7/ r~coyn~es
the dificu:Lty ~n aligniny featu.res o~ difEeren~ depo~it.ion
layers and apparently teaches depositiny the lappiny control
layer with the same step which forms the "registration of the
insulating layer forming the gap or covering the gap layer."
How an insulating layer can be registered in the same step with
depositing the lapping control layer isn't explained.
IBM TDB Vol. 23, No. 2, July 1980, p. 776, teaches a
method of calibrating an analoy lapping guide or machining sen-
sor to compensate for variations in bulk resistivity and film
thickness. This method is not involved with determining position
of the analog sensor relative to a feature of an insulating
layer.
The solution we propose to the problem is to create
the machining sensor or indicator with the same deposition mask
that defines the insulating feature to be precisely located
relative to the fi.nal position of the lapped edge. The way we
accomplish this is by providing a first conductive layer on
the face which is to carry the feature, which extends from near
the original position of the edge to be machined through the
allowable tolerance band which the final position of the edge
may occupy. This can, in the case of thin-film magnetic heads,
be conveniently included in the deposition step forming the
bottom leg of the thin-film head. During the step which creates
the feature from which the final position of the machined edge
~2~
is specifled, an additional barrier area of insulating material
is deposited on the first conductive layer and lying along a
so-called sensing line substantially parallel to the edge which
the machining will create, using the same mask to create both.
The sensing line defining the one edge of the barrier area is
-4a~
~3'~
precisely positioned relative to the feature because both are created wi-th
the same mask in the same deposition step.
Then a second conductive layer is de?osited on the barrier area
contacting the first conductive layer directly only between the initial loc-
ation of the edge and the sensing line. In the manufacture of thin-film heads,
this step will typically occur in conjunctlon with the depositing of the top
leg of the magnetic 1ux path. Those skilled in the art understand that each
of these three layers are produced by a series of steps including the use of
a precision mask, usually optical, to form the desired pattern in the layer
with very high precision.
The edge of the surface is then machined from its initial location
toward the sensing barrier line edge. I~hen the machined edge reaches the
sensing line, electrical continuity between the first and second conductive
layers is broken ~assuming a non-conductive machine tool). A continuity tes-
ter connected between the second conductive layer lying on the barrier area
and the first conductive layer will indicate an open circuit indicating pos-
ition of the machined edge. If the sensing line is intended to define the
ideal final position of the line of intersectian of the two surfaces, then
machining is halted.
In fact, the preferred application for this discrete machining
sensor is to calibrate a conventional analog sensor to precisely determine
its position relative to a feature line precisely defining the edge of an
insulating feature. This is accomplished by using one or more discrete
sensors, each having a different sensing line intersecting the sensing area
of the analog sensor, and each precisely positioned relative to the feature
line. At each point in the machining operation where a discrete sensor opens,
the resistance of the analog sensor is measured.~ These resistance values may
7~
be substituted in a general equation o-f the form h=K/R relating analog sensor
resistance R with spacing h of the top edge of the analog sensor from the
machined edge. The equation can then be solved to provide a value for the
constant K and any other constants to yield an equation directly relating
sensor resistance with machined edge spacing from the feature position line.
Accordingly, one purpose of this invention is to increase the
accuracy which machining of the edge of a surface can place the edge relative
to a feature carried on ~he surface.
A second purpose is to reduce the scrap rate during such machining
operation.
Another purpose is to combine the steps of forming the throat filler
material of a thin-film head with the step forming the machining guide when
machining a transducer assembly carrying a thin-film head.
Yet another purpose is to allow more accurate measurement of the
current status of the machining operation.
Thus, in accordance with a broad aspect of the invention, there is
provided a method of calibrating a machining sensor lying on a surface, a
first edge of which is to be machined from its initial location to an ideal
final position line having a predetermined spacing from a feature line, said
sensor being of the type having a discrete sensor comprising: a) a bottom
conductive area on the surface lying between the first edge's initial locat-
ion and extending to at least the ideal final position line; b) an insulating
barrier area having a sensing line boundary comprising at least one line seg-
ment each having precisely known spacing from the feature line and lying
between the ideal final position line and the initial location of the first
edge, said insulating barrier area extending away from the initial location
7~3~
of the first edge; and c} a layer containing at least one conductive path
lying entirely on the barrier area outside the area between the initial
location of the first edge and the sensing line ~oundary, each said
conductive path extending across one o the line segmcnts and making el~c-
trical contact with the conductive area; and d) a continuity tester connected
betweerl the end o~ each conductive path renlote :trom thc seJI~lng line boun~ary
and ~he bottom conductive area, cmd prov:idi.ng an i.ndication when continuity
ceases between a conductlve path and the bottom conductive area; ~nd an
analog sensor comprising a) a resistive conducting strip lying along the
first edge's initial location and intersected by i) extensions of each of
the sensing line boundary's line segments crossed by the conductive paths,
and ii) the ideal final position line; b) resistance measuring means
electrically connected across the resistive conducting strip for providing
a signal indicative of the resistance of the conducting strip; wherein
the method comprises a) machining the first edge of the su~face toward
the feature line until the c~ntinuity tester indicates that continuity
between a conductive path and the conductive area has ceased; then without
further machining b) analyzing the signal from the resistance measuring
means to determine the resistance in the conducting strip; and c) calculating
a constant of inverse proportionality K from an equation ~ the form h=KtR by
substituting i) the known spacing distance of the line segment crossed by the
conductive path whose continuity wlth the conductive area ceased, for the
distance h between the feature line and the first edge, and ii) the resist-
ance indicated by the resistance measuring means for the resistance R of the
conductive strip, and solving the equation for K; d) continuing machining o~
the first edge toward the ldeal final position line; e) while machining,
~z1~7~3~9
analyzing periodically the si~nal from the resistance measuring means to
determine the resistance of the conductive strip, and calculati~g therefro~n
the value h; and E) ceasing machining when the value of h ~g reducod to
within a predetermined range Oe th0 spacing between the ideal final position
and feature lines.
Okher objects and beneEits o~ this invention will be evident from
the ~ollowing explanation.
BRIEF DESCRIP~'ION OF T~IE DRAWINGS
Figure 1 is a perspective view o a prism having a surface on which
the subject inventive article is located, and showing an intermediate step
in the inventive method.
F~rè~ 2 and 4 are cross sections through one of the machining sen-
sors shown in Figure 1 before and after the machining step, respectively.
Fig~es 3 and 5 are cross sectional views of the feature relative
to which the edge positioned by the machining is respectively located before
and after machining.
Figure 6 discloses a structure incorporating this discrete sensor
in a preferred composite sensor to be employed in mass production of devices
such as thin-film heads, which have close tolerance dimensions based on the
position of an edge of an insulating area.
Figure 7a is a magni~ied perspective view of an individual thin-film
TesistOr of Figure 6.
Figure 7b is a circuit schematic of the analog sensor network of
Figure 6.
DESCRIPTION OF TIIE PREFERRED EMBODIMENTS
Since this discrete sensor has been developed specifically for the
l~S)~
purpose of controlling throa-t height of a thin-film head, the description is
based on an application in this area. It has identical applicability in any
case where such machining relative to a feature defined by deposited insulating
material must be controlled.
Figure 1 shows a greatly magnified perspective view of a machinable
prism or block 9 formed of a ceramic material, and comprising a thin-film
head air-bearing slider as it looks just before -the final machining of the
air-bearing face. Line 15 is the initial position of the edge of end face
10, defined by the intersection of the initial position of flying surface 26
(shown on edge in Figures 2 and 4) with face 10. Surface 26 is to be
machined until its intersection line with end face 10 r0aches its ideal
position coinciding with a sensing plane 13 defined by the two lines 13a and
13b.
On end face 10 there has been placed a machining sensor or guide
21 including a conductive layer or area 11 intersected by sensing line 13a
or plane 13, and having any convenient shape. Figure 2 shows this guide 21
in cross section prior to final machining. On top of conductive area 11 an
insulating layer comprising barrier area 12 is deposited, having one edge
lying along the sensing plane 13, extending away from the initial location
of line 15 at the e~ge of face 10 and lying atop conductive layer 11. Sensing
plane 13 should be substantially parallel to the initial location of line 15
at the edge of face 10. A preliminary machining step may be necessary to
configure prism 9 so that this relationship exists. Another deposited
conductive layer forming conductive area 14 is located entirely within barrier
area 12 on the side of the sensing line 13a and extends across line 13a,
contacting conducting surface 11 between the sensing line 13a and the initial
location of the edge at line 15. Thus, conductive layer 14 is completely
insulated from conductive layer 11 as to layer area 14b, i.e., the portion
above line 13a, and makes electrical contact with layer 11 in area 14a, below
line 13a.
For illustrative purposes here, a simplified diagram of a typical
thin-film head 20 is shown adjacent machining guide 21 ~nd in cross section
in Figure 3 before machining. This comprises a pair of magnetic flux paths
17 and 18 ~see Figures 3 and 5), a winding 19, and a deposited insulating
material 24 typically formed of al~inum oxide interposed between leg 17 and
leg 18 of the magnetic flux path, thereby creating the flux gap 25. A second
insulating layer 16 insulates turns 19 and defines the interior end of flux
gap 25. This interior end of flux gap 25 lies along one segment of a feature
line 27, shown on end as dots in Figures 2-5. The spacing between feature
line 27 and sensing line 13a is formed by the same deposition step and with
the same mask, and is therefore known with great precision, since no mask
alignment errors are present.
To provide a flux gap 25 of the proper throat height, it is
necessary to machine face 26 until it coincides with plane 13 on face 10
within a tolerance of 60 ~in. Flux gap 25 is physically formed by and essent-
ially comprises deposited non-magnetic insulating material. It will be
clear to one skilled in the art that by creatir,g the edge of barrier area 12
along sensing line 13a, which defines the point at which machining is to stop,
with the same mask and in the same deposition step defining the interior end of
gap 25 along feature line 27, gap 25 throat height will be very accurately
defined and much more accurately so defined than if the feature line 27 and
sensing line 13a were created during separate deposition steps or with
- 10 -
different masks. It will also be clear that control of throat height of a
thin-film head gap is only one o~ many possible applications where this
procedure may be used.
The machining is conventional, and can be perormed by lapping or
other high precislon operation, but must be perorme~ by a ~ool which does
not short betwecn layers 11 and l~. Con~inu:ity testers 22 ~rc conn~cte~ to
conductive surface 11 and conductive layer l~b by connectors 23.
The machining slowly erodes the material between plane 13 and
the initial location of the edge of face 10, line 15. When the material
between plane 13 and line 15 has been completely eroded, electrical con~act
between layers 14a and 11 is broken and continuity testers 22 indicate this
condition. The final configuration of a machining sensor 21 is shown in
Figure 5. The operator monitors testers 22 and can see the indication by
them and stop the machining. ~lternatively, the machining device can be
connected to testers 22 to automatically stop its operation once continuity
fails.
The reason the initial position of line 15 must be nearly parallel
to sensing plane 13 is now apparent. When the edge of face 10 is machined to
coincide with plane 13, if they are not parallel at that time, some material
past plane 13 will be removed, causing one corner of the sensor 21 to define
the end of continuity and the sensors 21 will lose continuity at different
times. Thus, at some point in the machining operation, edge 15 should be
approximately parallel with sensing plane 13. The position of edge of face
10 at that point can be considered its initial position. 2~achining to
achieve this relationship may be considered merely a prelimina-ry step. The
effect of such non-parallelism can be reduced by making layer 14 more narrow
~12~B~9
and by placing sensors 21 close together. However, the likeli-
hood of a defect in the electrical contact between -the~ which
totally destroys initial conkinui~y i~ then yre~ter. ~I'he in-
herent width o~ the ~eature and its appurteran~ ~t~ucture (hea~
20) limits the proximity between ~n~ors ~14
Wh:ile the sensors 21 and the associated process jusk
described function satisfactorily for certain requirements in
small production runs, the commercial requirement for many
thousands of magnetic heads 20 has led to a preferred use for
these sensors 210 To cheaply and efficiently manufacture these
heads 20, we prefer to place several on a single bar, and then
machine all of their flying surfaces 26 simultaneously.
A preferred use for this invention uses a workpiece
support capable of bending the bars on which the heads are plac-
ed, so as to place a greater number of the throat heights of
the heads on the bar within the tolerance range required. To
determine current status of each head's throat heightt frequent
measurements of each of these throat heights occur during the
final machining phase. Accurately calibrated analog machining
sensors are located adjacent each head on the bar. If indi-
cations from these sensors early in the final machining opera-
tion reveal that certain throat heights will be out of tolerance
when machining has placed all others within the desired toler-
ance, then the bar is bent to cause additional machining of the
flying surfaces of cer-tain heads to occur relative to the machin-
ing of other head's surfaces. By properly choosing the amount
and location of this bending, a much greater percentage of the
heads' throat heights can be caused to fall within the tolerance
range at the completion
-12-
, ", ~ ~
of machining. But of course, the sensors providing this information must
accurately measure throat height at requent intervals. ~ecause such analog
sensors have constituent elements formed by conductive deposit~, ~h~y su~'~'er
from the alignment errors which also plague conventional discrete sensors.
A composite machining sensor which includes an analo~ sensor 28
continuously providing a sigrlal speci~ying t'he position o~ the machi7le~
edge 15 is shown in ~ligure 6. The zero throat hei~ht or ~'ea~urc line 58
essentially defines the position of the eature relative to which line or
edge 15 is to be positioned by machining. The composite sensor is mounted
on end ace 10 of prism 9 ,md includes an analog sensing element 31 ormed
of a resistive conducting strip and three discrete sensors formed from
conductor paths 46-48, insulating barrier area 33 beneath them, and a
conductive area 49 below the barrier area 33 making electrical contact with
ends 50-52 respectively of conductor paths 46-48. Sensing line segments
38-40 form a staircase pattern along the bottom edge of barrier area 33 and
are offset with respect to each other, are approximately parallel to edge 15
as initially positioned, and have extensions which are a predetermined
distance from each other. ~ach of the sensing line segments 38-40 are
located at a precisely known spacing rom the zero throat height or feature
line 58 by virtue of their creation by the same process step and with the
same mask as that which produced the interior end of the flux gap of the
appurtenant head or other device. Conductor paths 46-48 have appreciable
electrical resistance and are commonly connected to connector pad or
terminal 43. Paths 46-48 cross line segments 40-38 respectively and all
make electrical contact with conductive area 49. Terminal 43 in turn is
connected to the upper selectable terminal of single pole double throw (SPDT)
- 13 -
lX~
switch 52, and to one terminal each of voltmeters 55 and 57.
Analog sensing element 31 is unitary with the conductive area
4g which forms part of the discrete sensors 29. The ends of sensing element
31 are connected by bridges 35 and 36 to resistive conductor paths 34 and 32
respectively. Element 31 has an appreciable amount of resistance, initially
Rl, between bridges 35 and 36. The nominal height hl and length L1 deter-
mine its resistance in large part, during machining. As the bottom edge
15 of end surface as face 10 is slowly machined away, the height hl of
element 31 decreases and, naturally, its resistance increases.
Paths 34 and 32 connect conductive bridges 35 and 36 to connector
pads or terminals 41 and 42 respectively. Conductor paths 34 and 32 them-
selves have in one preferred embodiment appreciable resistance, again depend-
ent on their lengths L4 and L2 and heigh-ts h~ and h2, respectively. Resistance
in conductive paths 34 and 32 is unavoidable because they too are mitary
with analog sensing element 31, which must have some resistance within it
to properly perform its sensing function. Connector pad 41 is connected to
the terminal of voltmeter 55 not connected to pad 43 such that voltmeter 55
measures voltage between pads 41 and 43. (Voltmeters 55 and 57, switch 52
and constant current source 53 are located remote from face 10.) Pad 41
is also connected to the lower selectable terminal of SPDT switch 52. Pad
42 is connected to one terminal of constant current source 53 and to the
terminals of voltmeter 55 and voltmeter 57 not connected to pad 43. The
terminal of constant current source 53 not connected to pad 42 is connected
to the center or co~mon terminal of SPDT switch 52.
We have developed an equation of the form h=K/R which relates the
value of sensor 31 height hl=h to the dimensions of conductors 34 and 32 as
- 14 -
~L~l3~
incorporated in the const~mt K, and to voltages measured by voltmeters 55 and
57 which provide a current indication o~ the analog sensor ~1 re~i~tance ~.
As is derived in the Appendix, sensor height hl=V2h2/Q(Vl - xV2), ~1 ~nd
V2 measured with switch 52 in the down posi~iorl shown. ~ is ~hu~ obvi~u~
that throclt heiLht =V2h2/Q~Vl-xv2)-Yo~ =hl - Yof~ whcre Oet
between the top of analog sensing eLement 31 and the zero throat height or
Eeature line 58 defining an edge of the feature relative to which discrete
sensor 29 is deposited. In these equations, Q=L2/Ll and x=1.4/L2. It is
relatively easy to control the deposition such that paths 34 and 32 have
nearly identical dimensions so that L4 = L2 and x=l to within ~ 2% or less,
and we prefer in one embodiment to do this. ~ven larger (+4%) errors affect
throat height measurements by only a microinch or so.
It is also possible to deposit path 34 with a very small effective
L~ (L4~<L2) by fo~ming path 34 with height and thickness substantially
greater than for path 32. By properly specifying the dimensions of path
34 formed by the deposition process, x can be set to fall in the range of
.01 to .1. Although the precision with which x is known in this case may
be no better than +10% or even +20%, since the value of x is quite small,
the overall impact on throat height measurement accuracy is similar to the
case where x=l and is known to +2%. Once the deposition process is stablized,
an average value of x can be determined by either calculations or direct
measurements of the resistance of paths 34 and 32 on representative prism
faces 10, allowing x to be treated as a constant thereafter.
There are therefore in either embodiment, two unknowns in the
throat height equation, h2/Q and Yoff. With Vl, V2, and x known, it is
possible to determine the values for h2/Q and Y ff by ~leasuring the values ~or
- 15 -
~Z~3'~
Vl and V2 at known throat heights. This is accomplished by reference to
discrete sensors 29. As machining o prism 9 begins, line 15 mo~es slowly
toward line 38, increasing resistance o~ and volkage across line 3~, ~rlc~ea~in~
resistance of and voltage across analog sensor elen~enk ~1. At some poink,
line 15 coincides with line 38 causing the sensor colllpr~sin~ con~uctor path
48 to open. IE switch 52 is in i~s up pos:i~ion near to k}a~t timc, ~hc vol~age
Vl measured by voltmeter 55 will undergo a sudden increase when continuity
ends since the resistance between conductive area 49 and pad 43 has increased,
while current flow Ic from constant current sowrce 53 has remained unchanged.
(Since voltmeter 55 is assumed to have very large resistance compared to the
resistance in path 34 and element 31, Vl/IC very precisely states the resist-
ance between area 49 and terminal 43.) At this time, throat height is known
with great precision as the preselected exact spacing between line segment 38
and zero throat height line 58.
As soon as the increase in Vl is detected, switch 52 must be moved
to its down position, allowing the value of Vl to be read for use in the
equation expressing throat height. V2 is also read at this time for use in
the equation. Although dimensions of the deposited resistors can not be
precisely set by the deposition process, Ll and L2 as well as h2 and Yoff
are known with reasonable initial accuracy, having been formed by the same
mask. At the time that line 15 coincides with line 38, throat height is known
with great precision. Substituting the approximations for Q(=L2/Ll) and h2,
the measured values for Vl, V2, and the exact throat height into the equation
for throat height above, yields a better approximation for Yoff, increasing
the precision of its value substantially.
With switch 52 again in the up position, machining continues until
- 16 -
~ z~)t~ ~3 ~
line 15 coincides with line 39, causing the discrete switch comprising path
47 to open and another jump in the value o~ Vl to occur, Again, a second
precise value for throat height is available. At thi.s pol-n~, with two values
for throat height known with great accuracy and wi~h two values each for Vl
and V~ for those throat height valuQs also clccurately knowrl, it is possible
to solvc two throat he:ight equations simultaneously for the value o~ h2/~ arld
Yoff. After this point, throat height will be known with great accuracy by
simply measuring the values of the Vl and V2 and calculating it using the
just-determined values for h2/Q and Yoff. Thus, voltmeters 55 and 57 function
as an ohmmeter in conjunction with the foregoing equation for throat height,
to determine resistance Rl after calibration.
For the particular application for which we have developed this
method, it is necessary that each composite sensor be particularly effective
in indicating when throat heights range from 20 to 80 ~in. ~Vith that
tolerance band, we have found it convenient to place a first sensing line 38
of barrier area 33 at 200 ~in. from the zero throat height line 58, a second
sensing line 39 at 80 ~in. from line 58, and sensing line 40 at 20 ~in. from
line 58. Recall that these discrete sensors can be placed at accurately
known distances from zero throat height line 58. Thus, during machining when
the individual sensor formed by line 39 and conductor 47 is severad, then the
operator knows that the upper limit for throat height has been reached by
the adjacent heads. When the sensor comprising sensing line 40 and path
46 opens, then the operator knows that the adjacent head has fallen out of
tolerance and must be discarded. The ideal final position line to which line
15 is machined, may be anywhere within the throat height range of 20 - 80 ~in.
Because of the relatively good accuracy with which Ll, L2, and h2
.!
- 17 -
~2~g~
are initially known being all de~ined by the same mask, in con~ras~ ~o the
lower initial accuracy with which YOf is known, -the great accuI~cy Wit}l
which thro~t height is known when tlle sensor comprisin~ conductor ~ and
barrier line 38 opens, allows one ~o dete~llirle YO~.e wi-th ~b~-tantiaLly i~lC~
reased accuracy. In our metho~ Yo~ :is .init:ially known ~o ~ 50 ~in. wherea~;
the value of h2/Q has an inherenk inaccuracy of only about ~ 10 ~in When
machining has proceeded such that line 15 coincides with line 39 and the
discrete sensor comprising path ~7 loses electrical continuityl then a
better value for h2/Q and Yoff can be calculated by solving for h2/Q and
Yoff simultaneously using the two values for throat height previously
measured. This yields a somewhat greater accuracy of around + 5 ~in. for
the final computations of throat height calculated by the throat height
equation as machining of prism 9 along line 15 occurs.
Accordingly, if a large number of these composite elements are
simultaneously employed during machining on a prism 9 carrying many thin-
film heads, it is possible to stop machining at a time which permits the
maximum number of heads adjacent to the sensors to have the correct throat
height. Alternatively, if one wishes to employ the aforementioned invention
permitting prism 9 to be bent during the machining process, one can sense
what direction of bending is necessary to result in the greatest possible
yield of good heads.
APPENDIX
Referring first to Figure 7a, the stylized thin-film resistor 32
is shown to have length, height, and thickness dimensions respectively of
L2, h2, and t2. Current flow is parallel to the length dimension.
The schematic diagram of Figure 7b reflects the electrical circuit
- 18 -
~Z(~'78~3
on surface 10 in Figure 6 and is amenable to mathematical analysis as follows
using the symbols:
R = resistance
P = resistivity o film
t = film thickness
h - resistor height
L = resistor length
A = cross sectional area of resistor
The conductor paths or areas o:f ~:igure 6 will hereafter in this
analysis be referred to as resistors~ but~the use of re:ference numerals will
be consistent from Figure 6 to Figures 7a and 7b.
We can write the following equations governing the resistance of
each resistor:
R4 = PL4/th4 = CL4th2
R2 = PL2/th2 = CL21h2
Rl = PLl/thl = CL~l/hl
(Assuming P and t are uniform across the entire surface of prism 9 and that
h2 = h4 allows C to be substituted for P/t. These are reasonable assumptions.)
We ne~t solve for hl in terms of resistance and resistor size:
~llR4 - C[(L4/h2)+(Llt~ll)]
Substituting the value of C=R2h2tL2 into this equation yields
Rl-~R4 = ~R2h2tL2) (L4/h2)'~ (R2h2tL2) ( lt 1
which can be rewritten as
hlL2(Rl+R4) - hlL4R2 = R2h2Ll.
Thus, hl = R2h2Ll/[L2(Rl+R4) - L4R2] ~1)
Since Ic is by definition constant, then Rl -~ R4 = Vl/Ic and R2 =
- 19 -
7~
V2/IC where Vl is the voltage drop across both resistors 31 and 34, as measured
by voltmeter 55 and Vz is the voltage across resistor 32 measured by voltme~er
57. Both measurements occur wi~h switch 52 in its "down" position, Vol~-
meters 55 and 57 both have internal resistances ve,ry lar~e co-npared to th~t
in the series path of area 49, paths 46 - 48, an~ pad 43 (~igwre 6). Thus
voltage across this series path :is negligible when measuring voltages between
pad 43 a,nd pad 41 or 42. Paths 46 - 48 serve double duty in a sense, fwnction-
lng as elements of discrete sensors 29 and also as connector paths between
voltmeters 55 and 57 and the junction between resistors 31 and 32. Once
machining reaches line segment 40, voltages Vl and V2 can no longer be
measured since the voltage adjacent brldge 36 is unavailable. Note that the
entire sensor 29 will typically be only a few thousandths of an inch wide.
Substituting these values for Rl and R2 into equation (1) yields
1 (V2/IC)(h2Ll)/[(vl/Ic)L2 ~ (V2/I )L4
or hl = V2h2Ll/(VlL2 V2L4) (2)
If we set x = L4/L2 and Q = L2/Ll so that L4 = xL2 and L2 = QLl,
then L4 = xQLl. Substituting these values of L2 and L4 into equation( ?)
above yields hl = V2h2/Q(Vl-xV2). (3)
In Figure 6, by definition hl = Yoff ~ throat height, where hl is
the current height of sensing element 31. Substituting the value of hl from
equation (3) into this equation above yields
throat height = [V2h2/Q(Vl-Xv2)] ~ YOff
- 20 -
