Note: Descriptions are shown in the official language in which they were submitted.
Z003143
SENSOR WITH ABSOLUTE
DIGITAL ~Ul~Ul
This invention relates to sensors producing digital out-
puts, and in particular to sensors that can sense multi-valued
positions or dimensions of a stationary or moving object and out-
put a digital signal that is an indication of the sensed parame-
ter.
Background of the Invention
Sensors for multi-valued parameters are known. A typical
sensor could, for example, output an analog voltage or current
signal whose value changes continuously with the value of the
sensed parameter. For digital processing of that signal, the
analog signal is typically converted by a known A/D converter to
its digital code. When digital ou~u~s are obtained in this man-
ner, each discrete ouL~uL is unique and thus an absolute in-
dicator of the sensed value. By "absolute" is meant that no two
outputs are the same over the desired range, so that each output
unambiguously identifies a particular analog value or particular
range of analog values.
To minimize errors in decoding sensed ou~Ls it is also
known to choose Gray coded digital outputs. The Gray code dif-
fers from other encoding schemes in that successive coded charac-
ters never differ in more than one bit. For example, in a shaft
position encoder that outputs a digital signal to indicate which
937.006.PIT-207 - 1 -
CA 02003143 1998-03-30
of the segments the shaft is in, when the shaft moves from
segment seven to segment eight, the code must change from that
for seven to that for eight. As the shaft moves across the
segment boundary, if more than one bit has to change, it is
possible due to slight merhAn;cal inaccuracies that not all
bits will change at exactly the same time. If, for example,
the most significant bit in a BCD code changed before any of
the other bits changed, a very large error, would result. With
the Gray enCoA; ng scheme, since only one bit is allowed to
change at a time, the error is minimized. Also, ambiguity is
re~ e~ when the shaft position is in the line that separates
any two segments.
It is also desirable to eliminate the A/D converter, an
pensive compo~nt, and construct a sensor with plural
detectors to directly output the digital signal. To the best
of my knowledge, no digital-signal-outputting sensor is known
that produces a Gray coded output, much less one that is
absolute, that is, without repetition of the coded outputs over
the operating range of the sensor. Nevertheless, even without
the absolute quality, such sensors can be used to obtain
absolute information by recording and/or trA~k; ng the sequence
of outputs to unambiguously distinguish between two outputs of
the same code. This requires additional electronics, which is
costly and consumes space.
2003 1 ~1 i
Brief SummarY of Invention
An object of an aspect of the invention is a sensor that
directly outputs a Gray encoded signal.
An object of an aspect of the invention i8 a ~ensor that
directly outputs an ab~olute Gray encoded signal.
An object of an a~pect of the invention is a sensor oper-
ating with magnetic fields that directly outputs an absolute
Gray encoded signal.
These and further objects and advantages of the invention
are achieved, briefly speaking, with a novel sensor comprising an
array of spaced detectors cooperating with a detector-actuating
medium configured such that relative motion of-the array and me-
dium produces over a given range a sequence of outputs that are
Gray encoded.
In accordance with a further aspect of the invention, the
configuration of detectors and actuating medium is such as to
directly output absolute Gray encoded signals.
In a preferred e~bodiment, the detectors are linearly
spaced such that their centerlines are spaced by a distance of 4
X dr , where d~ is the resolution accuracy desired. The actuat-
ing medium includes at least three segments alternately capable
of actuating and deactuating each detector, with the length of
the segments in the array direction being in the ratio of 5:2:3.
In accordance with still another aspect of the invention,
the detectors are magnetic detectors, and the actuating medium
'A i'
20û3 1 ~S
--is a magnet having at least three pole segments differing in
length.
Another aspect of thi8 invention is as follows:
A combination for determ;n;ng the relative positions
between a first and second object, comprising:
said first object having fixably mounted thereto an array
of detectors along a defined detector array pattern, each of
said detectors being constructed to output a logic unit "1" or
"O" such that said output of said array of detectors produces
a binary output;
said second object being displaceable relative to said
first object and having a generating portion aligned with said
detector array pattern such that displacement of said second
object causes said generating portion to displace
corre~pon~ingly along said detector array pattern;
said generating portion having generating means for
causing said binary output of said array of detectors to
change in a gray encoded manner with respect to successive
changes in position of said generating portion of said second
object; and
said detectors being ~all-Effect devices responsive to a
magnetic field, and the generating means being a magnet
divided up into plural North and South poles, wherein the
detector array is fixed with equal center-to-center spacings,
and the magnetic poles of said magnet having non-equal center-
to-center spacings.
-- 4
A ~
~ 2003 1 4~
Brief DescriDtion of Drawings
The invention will now be described in greater detail
with reference to the accompanying drawings wherein:
Fig. 1 schematically illustrates one sensor embodiment
using magnets in accordance with the invention;
Fig. 2 is a table showing the sensor outputs for various
actuating medium positions;
Fig. 3 is a view similar to Fig. 1 of a second embodiment
using magnets in accordance with the invention;
Fig. 4 is a table showing the sensor outputs as a func-
tion of actuating medium position for the embodiment of Fig. 3;
Fig. 5 is a view similar to Fig. 1 of a third embodiment
using magnets;
Fig. 6 shows still a fourth embodiment of the invention
using magnets:
Fig. 7 is a table showing the outputs for the embodiment
o~ Fig. 5;
Figs. 8-10 are schematic views of three additional em-
bodiments using optics in the sensor;
Fig. 11 is a block diagram showing the sensor of Fig. 1
used to measure the height of a moving object.
- 4a -
A~
2003143
Fig. 12 shows still a fifth embodiment of the invention
using two magnetic tracks;
Fig. 13 is a table showing the sensor outputs as a func-
tion of actuating medium position for the embodiment of Fig. 12.
DETAILED DESCRIPTION OF ~K~K~ED EMBODIMENTS
Fig. 1 illustrates one sensor embodiment in accordance
with the invention that will directly output an absolute Gray-
encoded signal. An array 8 of seven discrete detectors is pro-
vided. The detectors are arranged in a row and referenced 10-16.
Each is schematically shown as a rectangle, representing the ac-
tive detecting area of the detector. The detectors are equally
spaced, with a centerline spacing indicated by 18. The detectors
can, for example, be magnetic detectors of the Hall-Effect type
commercially available as inexpensive discrete electrical com-
ponents from many supply houses.
The magnetic field to actuate the detectors is an
elongated permanent magnet 20 divided into plural segments
referenced 21-25 and spaced a short distance or gap 19 from the
detector array. The magnet position referenced 20 is the zero or
start position. For this explanation, the array 8 is fixed and
it is assumed that the magnet 20 is movable to the right in a
line parallel to the array 8 in response to some se~e~ parame-
ter. Shown in dashed lines is the relative vertical position for
its fourth 20' and ninth 20" positions when moved by the sensed
parameter. For clarity's sake, they are ~hown offset, but ac-
937.006.PIT-207 - 5 -
200;~143
tually would be in line with the first position. In a practical
embodiment with the detector array fixed, the magnet would be
coupled to a suitable mechanism that causes it to move to the
right to sense, for example, a dimension of an object.
Each detector responds to the presence or absence of a
specific magnetic field. In the case illustrated, the detectors
are constructed to output a logic "1" when no field is present or
when it detects the field from a North (N) pole, and to output a
logic "0" when it detects the field from a South (S) pole. The
outputs from the seven detectors represent an absolute Gray code
when there is a specific relationship between the detector spac-
ing and the pole lengths of the magnet. In the arrangement shown
in Fig. 1, which is drawn to scale, the active area of each
detector, indicated by reference numeral 27, is one unit long,
the detector centerlines spacing 18 is four units long, the lead-
ing S pole segment 22 is five units long, the adjacent trailing N
pole segment 23 is two units long, and the trailing S pole seg-
ment 24 is three units long. Since the detectors produce "1"
when detecting no field or a N pole field, the end N pole seg-
ments 21 and 25 can be omitted, but it is preferred to include
them because it sharpens the transition between segments, focuses
the magnetic field more at the detectors as is wanted, and
reduces stray and fringing fields. The length of the end N pole
segments is not important, which is why they are shown with
937.006.PIT-207 - 6 -
200~1~3
.
broken lines at their ends. It is important to note, and a fea-
ture of the invention, with a one-unit active area detector,
centerline spacings between detectors of four units, the active
pole segments 22-24 starting from the leading segment 22 have
lengths in the ratio of 5:2:3 units.
In the start or zero position shown in solid lines in
Fig. l, detectors 10, 11 and 12, facing S poles, output a "0",
and the remaining detectors facing no pole or a N pole output a
"1". Treating detector 10 as outputting the most significant bit
(MSB) and detector 16 as outputting the least significant bit
(LSB), the detector output for the zero position of the magnet 20
in BCD is 0001111, which in Hexadecimal notation (Hex) is OF. By
the same reasoning, when the magnet is in the fourth position
20', positioned four units to the right, the ou~u~ is 1000111 =
47(Hex), and when the magnet is in the ninth position 20", posi-
tioned nine units to the right of the start position, the output
is 1101011 = 6B(Hex). The table in Fig. 2 shows the ou~uLs in
binary and in Hex for each of the magnet 20 positions of which
there are a total of twenty. It should be noted from a com-
parison of the binary outputs that Gray encoding exists, because
never for the twenty unit range shown is there more than a one-
bit change in the binary o~L~uL between adjacent magnet posi-
tions. Moreover, the equivalent Hex output column demonstrates
that an absolute code has been created because no two outputs are
alike.
937.006.PIT-207 - 7 -
- 2003143
In a ~pecific example with Fig. 1 geometry, one unit
equalling 0.05 inch, the detector spacing was 4 X 0.05 ~ 0.2
inch, and the magnet lengths 22, 23, 24 were, respectively, 0.25,
0.1 and 0.15 inch long. In this case, as previously noted, the
resolution (o~) desired was 0.05 inch -- thus the detector spac-
ing of 4 X c~. In the 4econd column in the Fig. 2 table are
listed the subrange of movements for each magnet position to pro-
duce the output indicated for that row. Thus, the system il-
lustrated in Fig. 1 will measure twenty positions each with a
resolution of 0.05 inch over a range of 0 - 1.0 inch.
The invention is not limited to a seven detector array
employing a magnet with the three segments depicted in Fig. 1 to
produce absolute Gray encoded ou~u~s. The design rules to fol-
low to select other arrangements are as follows:
1. Resolution will be + ol.c ~arter of the center dis-
tance between detectors.
2. The magnet will have its poles configured so that the
detector's o~ changes each time the magnet moves a distance
equal to one-quarter of the detector spacing.
3. The magnet must have at least one pole segment that
bridges two detectors -- in the Fig. 1 arrangement, segment 22
being five units long bridges two detectors spaced four units
apart.
4. Because of the one-quarter center pitch, the pole
pattern in the magnet is chosen ~uch that the bit output sequence
937.006.PIT-207 - 8 -
2003143
-
at the first detector will be repeated at every center distance
at the next successive detector. For example, in the Fig. 2
table, note that the ouL~u~ of detector 10 for magnet positions
0-3 is 0011, which is the same ou~ sequence at detector 11 for
magnet positions 4-7, which is the same sequence from detector 12
for magnet positions 8-11, and so on. There are four possible
magnet positions for each detector center spacing. This behavior
is characteristic of constructions according to the invention.
Other examples of sensors according to the invention are
described below. Fig. 3 shows part of a detector array 29 com-
prising twelve detectors 30-41. The magnet configuration 42 com-
prises the same three segment arrangement 44-46 with reversing
poles 47, 48 at either end to increase cut-off sharpness. In
this case, the effective magnet length of segments 44-46 is ten
fourths of the detector centerline spacing, in the same 5:2:3
ratio previously described. This arrangement with twelve detec-
tors will produce forty absolute Gray encoded positions, the out-
put sequence of which, in BCD and Hex, is listed in the table of
Fig. 4 for the first thirty-six positions. The first column on
the left in Fig. 4 uses position notation covering four positions
each, so that the shifting sequence of BCD outputs for each posi-
tion designated in the left column is made more apparent.
The invention is not limited to the use of a single ac-
tuating medium. Fig 5. shows an arrangement comprising five
937.006.PIT-207 - 9 -
- 2003~4~
detectors which with two actuators also produces twenty absolute
Gray encoded positions. In this embodiment, the detectors are
referenced 50-54, and the actuators are a first magnet 57 with
the 5:2:3 ratio of segment lengths coupled to a second magnet 58
with the 5:2:3 ratio of segment lengths displaced eleven units
from the first magnet. Both magnets move in unison to the right
in response to the sensed parameter. The output pattern for the
arrangement is displayed in the table of Fig. 7. Alternatively,
the two magnets can be combined into a single magnet, with the
connecting piece being a single N pole interconnecting the trail-
ing reversing N pole 59 of the first magnet and the leading re-
versing N pole 59' of the s~con~ magnet.
The total length of this connecting magnet is ten times
the detector resolution giving a pattern that would be
5:2:3:10:5:2:3. This pattern can be repeated for what ever num-
ber of times required by exten~;ng the total length of the magnet
or adding more magnets.
The invention is not limited to linear geometries. Figs.
6 and 12 show circular geometries for measuring rotation angles.
The Fig. 6 embodiment employs five detectors 60-64. The magnet
65 is formed in the shape of a circle as shown with the 5:2:3
ratio S-N-S, and will yield a circular pattern that repeats every
360 degrees.
The ouL~L code would be the absolute Gray code shown in
Fig. 7. The 5:2:3 pattern can be repeated any number of times
937.006.PIT-207 - 10 -
2003~43
spaced apart by 10 increments of no active pole or N poles yield-
ing a code that is not absolute for 360 degrees. The output code
will be repeated once every 360 degrees for each repeat of the
pattern. That is, o~ s of the 20th to 39th positions will
have the same sequence as the ouL~u~s from the 0th to l9th posi-
tions as shown in Figure 13.
The detectors are spaced every 360/RD degrees where R is
the number of pattern repeats and D is the number of detectors.
The detectors will be spaced 360/lx5 = 72 degrees if there is one
pattern and five detectors; if there are two patterns in 360 de-
grees the detectors will be spaced 360/2xS = 36 degrees apart.
The total number of repeats possible der~n~c on the diameter of
the circle and the minimum detectable pole size.
The o~ can be converted to an absolute code by plac-
ing a second pattern parallel and connected to the first as shown
in Figure 12 consisting of one assembly with an outer and inner
magnetic track and seven detectors. The outer track consists of
two 5:2:3 pattern repeats 130 and 131 with five detectors 132 -
136 that are spaced 36 deyLees apart opposite the outer track.
The inner track consists of one north 137 and one south pole 138
each of 180 degrees aligned to the outer track and two detectors
139 and 140 that would be used to identify the absolute value of
the code as shown in the table of Figure 13. Thus, when the Hex
o~u~ from the five detectors 132-136 begins to repeat, the out-
937.006.PIT-207 - 11 -
200~ 43
put from the two detectors 139, 140 will change providing ab-
solute determination. The dual track embodiment could also be
used with a linear magnet, and the lower resolution/inner track
could also use the 5:2:3 pattern of poles.
In both of the circular geometry emhoA;ments of Figs. 6
and 12, the detectors would be fixed in the positions shown, and
the magnetic pattern would rotate. The angular rotation would be
indicated by the detector outputs indicated, for example, in Fig.
13 which can thus measure 40 positions, or over the 360~, 360/40
= 9~ rotation per position. In Fig. 6, the S poles are single
hatched, the N pole double hatched, and the 20 positions shown by
the numbers 1-20 on the outside.
The preferred embodiment employs the Hall-Effect detec-
tors and magnetic actuators because they are readily available at
low cost, require little maintenance, and detecting magnetic
fields provides a sturdy sensor that can operate in dirty en-
vironments. But the principles of the invention are also ap-
plicable to other kinds of detectors that can respond to a mag-
netic field, as well as to any kind of sensor comprised of radia-
tion or field generating parts and an array of detectors capable
of responding in a binary manner to the presence or absence of
the radiation or field, which of course includes the pos-
sibilities of built-in thresholds; that is to say, radiation
above and below a threshold respectively actuates and de-actuates
the detector.
937.006.PIT-207 - 12 -
200314.~
Thus, for example, the radiation generators can be LED's
or any light source, and the detectors photo-detectors.
Fig. 8 depicts an arrangement similar to Fig. 1 with an
array of photo-detectors 70-76 and an actuator built up of as-
sembled LED's 77, 78 and spacers 79, which LED's are always ON
indicated by the vertical arrows. Partitions 80 between the ON
LED's avoid light spilling over to actuate adjacent detectors.
Fig. 9 depicts an alternative in which a single or multi-
ple light source 82, always ON, stretches the full length of the
photo-detector array 83-39. In this case, with a fixed array,
and with a fixed light source 82, the movable member is a mask 90
with holes or slots 91, 92 corresponding to the positions of mag-
net segments 22 and 24. These holes or slots 91, 92 allow light
through to the detectors in the same way that the moving segments
22, 24 interact with the Hall-Effect detectors in Fig. 1.
Fig. 10 shows still a further alternative wherein radia-
tion sources (LED's) 92 are each combined with its own photo-
detector 93. Such components are readily commercially available,
and commonly used to detect the presence of a reflecting medium
above the unit. When a reflector is present, the light 94 from
the LED 92 will bounce off the reflector and be detected by the
adjacent photo-detector, typically a silicon photo-detector. For
application to the invention, a mask 101 would be provided lo-
cated above the array and representing the movable part of the
937.006.PIT-207 - 13 -
200~
sensor. The mask 101 would be reflective and provided with holes
or slots that prevent reflected radiation, or be non-reflec~ive
and be provided with reflecting spots or areas where reflection
is desired. Choosing the latter alternative, the first three op-
tical units 95-97, correspo~ing to detectors 10-12 of Fig. 1,
are shown under reflectors 99, 100 in positions corresponding to
the S-N-S pole segment pattern in Fig. 1. The electrical behav-
ior would be the same.
While the embodiments depicted all used an active ac-
tuator pattern in the ratio of 5:2:3, though highly desirable,
this is not essential. It turns out that othe~ actuator patterns
can be devised following the principles of the invention, but
they are less desirable for mechAnical reasons within the current
state of the art. For example, for the encoded ou~u~ to be ab-
solute and a Gray code, the magnet must move one divided by a
power of two distance where the detectors are spaced a distance
equal to the same power of 2; the available possibilities are 21,
22 which is included in the examples given, 23, 24, etc. 21
can't be used because it will not produce an ouL~uL with in-
creased position resolution as the single detector could only be
on or off and would give a position resolution of 1/2 detector
spacing not 1/4.
23, 24 result in no improvement in detector spacing
resolution because it would still take three (23) or four (24)
937.006.PIT-207 - 14 -
2003~q3
detectors to distinguish the 8 or 16 bit patterns that result
giving no improvement in performance but increasing magnet com-
plexity. The choice of 22 turns out to be the most practical
producing a sturdy sensor with remarkable resolution, and very
reliable performance.
As mentioned, the sensor can be used in any application
wherein the detector array or the actuator is physically moved in
the course of making a measurement. For example, Fig. 11 depicts
a simple application for accurately measuring height of any ob-
ject using the Fig. 1 embodiment. In this case, the object 120
moves in the direction shown by the arrow beneath a roller 121.
The roller 121 is displaced upward until it rests on top of the
object 120. The roller 121 is mec~nically linked 122 to the
magnet 20 of Fig. 1, as shown, and moves the magnet 20 an equal
or proportionate distance upward. The seven ouL~u~s 123 from the
detector array 8 is sent to a conventional signal proceC~or 124.
The processing circuitry is easily designed as is well known in
this art not to accept ou~uLs from the detector array until
after the roller 121 has settled on top of the moving object. At
that point, the array 8 is polled and a 7-bit output is generated
that unambiguously indicates the precise subrange of values
within which the object height falls. Any height that falls at a
subrange boundary will cause an error no worse than the next sub-
range location. For the dimensions given for the example of
937.006.PIT-207 - 15 -
Z003~43
Fig. 1, this means an error no worse than 0.05 inch. The pro-
cessor 124 can either display the height measurement or use the
information in some other manner, for example, for sorting the
objects according to height.
In the example given, there is a 1:1 proportional rela-
tionship between the upward movement of the roller 121 and that
of the magnet 20. This is not nececsAry. The linkage 122 can be
changed so that the magnet 20 moves upward a multiple or sub-
multiple of the roller movements in order to enhance the accuracy
or increase the total range of measurement. Also, as mentioned,
the geometry is not limited to straight line geometries. For ex-
ample, the array of detectors can be arranged along the arc of a
circle as shown, be bent into a full circle, or oriented to fol-
low any curve. The only restraint is that the actuator must have
a similar shape or at least be able to actuate the detectors in a
sequence as described herein. As previously mentioned, either
the detector array or the actuator can be made movable. It is
preferred to move the actuator because the active part of it will
typically be shorter than that of the array. Moreover, when mag-
nets are the actuator, they can take more abuse than the Hall-
Effect detectors, which are more sensitive and typically include
integrated circuits.
It is understood that, in the embodiments disclosed, all
the detectors in the array are continuously energized with the
937.006.PIT-207 - 16 -
~ 200314.~
appropriate voltages and currents so that they all remain in a
continuous activated or on condition ready to output a "1" or "0"
depending upon the polarity of the sensed magnetic field. In
other words, the sensor ~uL~uL is in parallel. However, though
the ouL~uLs at each detector all appear simultaneously, they can
be polled and converted if desired into a serial stream that can
be transmitted to a remote location if desired by conventional
data communications equipment.
While the invention has been described and illustrated in
connection with preferred embodiments, many variations and modi-
fications as will be evident to those skilled ~n this art may be
made therein without departing from the spirit of the invention,
and the invention as set forth in the appended claims is thus not
to be limited to the precise details of construction set forth
above as such variations and modifications are intended to be in-
cluded within the scope of the ArrenAed claims.
937.006.PIT-207 - 17 -
