Note: Descriptions are shown in the official language in which they were submitted.
~ 653~
Background of the Invention
This invention pertains generally to scales and
more particularly to a counting scale for determining the
number of articles in a group.
Heretofore, there have been some attempts to
~xpedite the counting of large numbers of articles, such as
electronic components and other small parts, by first
weighing a known quantity or sample to determine the average
weight of the articles, then weighing the unknown quantity
and combining the weights to determine the number of articles.
In order to provide sufficient resolution to accurately de-
termine the weight of the sample and sufficient range to r
measure anything but a very sm~ll unknown quantity, it has
generally been necessary to use two separate scales, one
a low capacity scale for weighing the sample and the o~er
a high capacity scale for weighing the unknown quantity.
The use of two scales has a number of obvious
disadvantages. In addition to the inconvenience of havLng
to work with two scales and make two separate readings,
both scales must be calibrated to a high degree of accuracy.
This can be difficult, particularly with electronic scales
utilizing amplifiers which are seldom perfectly linear
in practice. In addition, there are operational problems
~uch as determining whether the sample should be added
to the unknown quantity before the latter is weighed.
Summary and Objects of the Invention
The invention provides a counting scale of wide
range and high resolution utilizing a single load cell for ;
determining the weights of both a sample and an unknown
quantity of articles to be counted. The weights are
~ .
-2-
,. .
'
-
~04~S;~3
combined to determine the number of articles, and this number is displayed.
The platform of the scale is mounted on and supported by the load cell in a
system requiring no levers, pivots or other moving parts. Weight readings
are averaged on a moving average basis, and with no load on the scale, any
drift is cancelled by increasing or decreasing the weight readings in small
increments during successive counting cycles to maintain an accurate zero
reference.
It is in general an object of the invention to provide a new and
improved counting scale and method for determining the number of articles
in a group~
Another object of the invention is to provide a co~mting scale
and method of the above character utilizing a single load cell.
Another object of the invention is to provide a counting scale
and method of the above character in which weight readings are averaged on ~ ;~
a moving average basis to provide a more accurate reading than would be ~ ;
possible with individual weight readings.
Another object of the invention is to provide a counting scale and
method of the above character in which the zero reference level is accurately ;
maintained.
Thus, in accordance with one aspect of the invention, there is
:,
provided, in a counting scale for determining the number of articles in a
group: single load receiving means for receiving the articles to be counted,
a load cell mounted on the base and connected to the load receiving means
for providing an electrical signal corresponding to the number of articles
on said load receiving means, means responsive to the load cell signal for
providing a digital signal corresponding to the weight of the articles on
the load receiving means, means responsive to the digital signal when a
predetermined number of the articles is placed on the load receiving means
~or storing a signal corresponding to the weight of the predetermined number
T -3- : .
`. . . . .
- ~
: . ~ .. ..
~0~t;538
of said ar-ticles, and means for combining the stored signal with the digital 1 :
signal produced when the group of articles is placed on the load receiving
means to provide an output signal corresponding to the number of articles
in the group.
In accordance with another aspect of the invention there is
provided, in a counting scale for determining the number of articles in a
group of the same: a load receiving member, means including a load cell
connected to the load receiving member for providing an electrical signal
corresponding to the weight of a load on said member, means responsive to
the electrical signal for providing successive signals representative of
the instantaneous weight of the load, means for storing a predetermined
number of the successive signals on a rotating basis whereby the oldest of
the stored signals is replaced with the most recent of the successive
signals, means for averaging the stored signals to provide a digital signal
representative of the average instantaneous weight of the load, means for
iteratively checking the digital signal and changing said signal by a pre-
determined increment each time that the signal differs from a reference
level by an amount less than a predetermined amount corresponding to a
weight less than the expected weight of the load7 the predetermined increment
being substantially less than the predetermined amount whereby the digital :
signal is returned to the reference level in a plurality of discrete steps,
means responsive to the digital signal produced when a sample comprising a ~ ;
predetermined nu~ber of the articles is placed on the load receiving member
for storing a signal corresponding to the weight of the predetermined number
of articles, means for combining the stored sample signal with the digital -~
signal produced when the group of articles is placed on the load receiving
member to provide an output signal corresponding to the number of articles
in the group, and display means responsive to the output signal for indicat~
ing the number of articles in the group.
~ -3a-
:
,:
~0~;538
.;; ' .
According to another aspect of the invention there is provided, ;~.
in a method for determining the number of articles in a group placed on a
scale, the steps of: providing an electrical signal corresponding to the
. .
weight of a load on the scale, processing the electrical signal to provide
successive signals representative of the instantaneous weight of the load
at a predetermined clock rate, storing a predetermined number of the
successive signals on a rotating basis whereby the oldest of the stored
signals is replaced by the new signal provided during each clock cycle,
determining the average of the stored signals to provide a digital signal ~
representative of the average instantaneous weight of the load maintaining .
the digital signal at a reference level in the absence of a load by
iteratively checking the signal and changing the signal by a predetermined
increment each time that the signal differs from the reference level by an
amount less than a predetermined amount, the predetermined amount correspond-
ing to a weight less than the expected weight of the load and the predeter~
mined increment being substantially less than the predetermined amount, ~^ :
storing a sample signal corresponding to the digital signal wherein a pre-
determined number of articles are placed on the scale, combining the stored
sample signal with the digital signal produced when the group of articles~
: . .
placed on the scale to provide an output signal corresponding to the number ~ :
of articles in the group, and displaying a count corresponding to the output -~
signal to indicate the number of articles in the group.
In accordance with a further aspect of the invention there is
provided, in an electronic counting scale: a base7 a load cell mounted on
the base for providing electrical signals corresponding to the forces
applied thereto, a platform connected to the load cell for successively ~
receiving a sample consisting of a predetermined number of articles and a ~~
load consisting of an unknown number of the articles, and signal processing
means including a microprocessor responsive to the load cell signals
B ~ -3b-
. ~ , . .:
~4~53~
producecl by the sample and the load for delivering an output signal cor-
responding to the n~ber of artieles in the load.
Additional objects and features of the invention will be apparent
from the following description in which the preferred embodiments are set
forth in detail in conjlmction with the accompanying drawings.
Brief Descri ~ t~b`~ D~v-
~
Figure 1 is a perspective view, partially exploded, of oneembodiment of a counting scale according to the
, .
~ .
~ . .
1C~4~S3~3
invention.
Fig. 2 is a ragmentary sectional view,
partly broken away, of a portion of the scale of Fig. 1,
illustrating the load cell and the manner in which the
S platform is mounted thereon.
Fig. 3 is a cross-sectional view taken along
line 3-3 in Fig. 2.
Fig. 4 is a block diagram of the computing system
of ~he counting scale of Fig. 1.
Fig. 5 illustrates the organization of a weight
register in the computing system of Fig. 4.
Fig~ 6 is a flow chart of a program for maintain-
ing an accurate zero reference in the system of Fig. 4.
FigO 7 is a flow chart of a program for
determining weights and the number of articles in an ~ ~
unknown quantity on the scale of Fig. 1. ~`
Fig. 8 is a flow chart of a program for determin-
ing the weight of a sample size and computing the number
of articles in an unknown quantity on the scale of Flg. 1
Description of the Preferred Embodiments
As illustrated in FLgs. 1-3, the counting scale
includes a console or cabinet 11 having a bottom wall or base
12 on which a load cell 13 is mounted. A platform 14 for
receiving objects to be weighed is mounted on and
supported by the load cell. As illustrated, the platform
comprises a horizontal deck 14a, a depending skirt 14b
which extends along the front and side edges of the deck,
and an upstanding flange 14c which extends alDng the rear
edge of the platform. The platform rests upon a load
spider 18 having a plurality of diagonally extending
~,
-
53 !3
arms 19 and a depending stem ?1 which is secured to the
load cell by screws 22. The platform is affixed tG the
spider arms by screws 23.
A keyboard 24 is provided toward the front of
S cabinet 11, and the controls for operation of the scale
are mounted on the keyboard. A row of keys 26 provides
means for selecting the number of articles to he used as
a sample. In the embodiment illustrated, samples consisting
o~ 8, 16, 32 or 64 articles can be selected. Other
controls on the keyboard include a ZERO or TARE key 27, a
RESET key 28, and a PRINT key 29. In one presently
preferred embodiment, the keys are fabricated of a
transiuscent material, and lamps mounted beneath the keys -
are illuminated to indicate the functions which have been
selected.
A WEIGHT display 31 and a COUNT display 32 are
provided for indicating the weight of a load on the scale,
and the number of articles therein. In the preferred
embodiment, each display includes a plurality of 7-segment
display elements mounted behind a transluscent window 33 ;
in a sloping panel 34 above keyborad 24. A STOP MOTION/
CONTINUOUS DISPhAY switch 36 is mounted on the front apron ~;of the cabinet. When this switch is in one position, ~
the weight and count are not displayed until a steady ~ ;
state is reached after a change in the load. In the
other switch position, the running values of the weight
and count are displayed. A POWER ON-OFF switch 38 is
mounted on the apron next to switch 36.
As illustrated in Figs. 2-3, load cell 13
includes a generally rectangular framework comprising
-5-
'~ . , :
~O~S3~il
a base 46, a side frame 47, a load arm 48, an upper arm 49,
and a lower arm 51. Arms 49 and 51 are substantially equal
in length, and they are generally parallel to each other
and to base 46. The frame is a unitary structure, and it is
~ormed to include flexures 52-55 between arms 49, 51, side
~rame 47 and load arm 48.
Base 46 is affixed to the bottom wall or base 12
of cabinet 11 by screws 57, and the stem 21 of load spider
18 is mounted on load arm 48 of the load cell. Load arm
4B is provided with an upwardly facing shoulder 58 which is
engaged by a downwardly facing shoulder 59 on the load
spider stem. A load applied to platform 14 is thus
transmitted to load arm 48 and produces a downward deflection
of this arm corresponding to the weight of the load. Stops
61, 62 formed on base 46 and load arm 48 limit the travel
of the arm to provide overload protection.
A beam 63 is rigidly affixed at one end to a
block 64 which is secured to side frame 47. The other end
of the beam is affixed to a block 66, and tie rod 67
extends between this block and a protrusion 68 formed on
load arm 48. As best seen in Fig. 3, the tie rod is
narrower than arms 49, 51 and flexures 52-55, and it
extends in a direction generally parallel to the direction
in which load arm 48 is deflected by a load on the platform. ~-
When load arm 48 is deflected by a load, the
force exerted by the load is transmitted.to beam 63 by tie
rod 67, causing the beam to bend by an amount corresponding
to the weight of the load. The amount of bending is
detected by strain gauges 71, 72 which are mounted on beam
63 and have electrical resistances which depend upon tha
.
38
forces applied to ~he beam. An electrical current is
applied to the strain gauges by conventional wiring, not
shown, to provide an output voltage corresponding to the
weight of the load.
- Cover plates 73, 74 extend over the sides of the
load cell to enclose the same. These plates are secured
to side frame 47 by suitable means such as rivets 76. Even
though the platform area is substantially larger than the
load cell, the reading of the load cell is generally not
affected by the placement of the load`on the platform
or side-loading effects.
Referring now to Fig. 4, the output of load cell
- 12 is connected to the input of an analog-to-digital
converter 79 which conver~s th_ output voltage from the
load cell to digital signals corresponding to the weight
of the load at a suitable clock rate such as 15 times
per second. In the preferred embodiment, the A/D converter
is similar to a conventional dual slope digital voltmeter
dified to provide output signals in straight binary form,
rather than binary coded decimal form. The binary form
is partDcularly suitable for use in a scale in which
a computer is used to process the data because of the
relative case of binary arithmetic compared to decimal arithmetic
Means is provided for processing the signals
from A/D converter 79 to determine the weight of a load
on platform 14 and the number of articles in a group
on the platform. This means includes a microprocessor,
designated generally by reference numeral 81, having a
~ datë input line 82 and a data output line 83. One
suitable microprocessor is a four-bit parallel system ~;
5~8
available from Intel Corpora~ion, Santa Clara, California,
under the trademark MCS-4, and includes a model 4004
central processor unit (CPU), model 1702 programmable
read-only memories (PRO~'s), model 4002 random access
memories (RAM's).
The microprocessor receives a signal on line 84
from converter 79 each time a conversion is made, and it
delivers clock pulses and clock address signals to a clock
control 86 on lines 87 and 88. Read address signals are
applied to a read enable controller 89 on an address line
91 .
The binary output signals from A/D converter 79
are read into a shift register 92 upon receipt of a clock
pulse on line 93. The signals in the shift register are
read into the microprocessor through input line 82
upon receipt of a READ signal on line 94.
Sample size selection keys 26 are scanned by
a keyboard scanner 96 which receives control signals on
line 97 from the microprocessor and clock signals on line
93 from the clock control. Upon receipt of a clock
signal, the scanner transfers the data from the keyboard
to a bufer register 99 from which it is read into the
micxoprocessor through input line 82 upon receipt of a
RE~D signal- on line 101. The keyboard lamps, designated
generally by reference numeral 102, are illuminated in
accordance with output data on line 83 upon receipt of
clock pulses on line 103.
Weight display 31 and count display 32 receive
data from output line 83 and display this data in response
to clock pulses on lines 104 and 106, respectively.
-8-
~lO4~5313
A llnearity and scale factor control 111 is con-
nected to the microprocessor through a buffer 112 and
i~put line 82 to provide compensation for any non-linearities
in the load cell and associated circuitry and to set a
desired relationship between the weight readings and the
load cell output. This control is generally set during the
initial calibration o the scale and thereafter changed only
in the event that recalibration is necessaryO
Switches 121-124 are provided for selecting the
mode of operation of the scale. Closure of switch 121
conditions the scale to read gross weight, and closure of
switch 122 condition~ the scale to read net weight. These
switches are generally used only for test purposes, and
they can be located inside cabinet 11. Switch 123 represents
the contacts of switch 36 in Fig. 1, ~nd when closed
this switch inhibits the display of weight and count data
until the scale stops moving ~ollowing a change in the load.
Switch 124 is operated by ZERO key 27, and closure of this
switch serves to reset the weight reading to zero. Data
indicating the conditions of the switches is present in a
buffer register 126, and this data is read into the
microprocessor through input line 82 upon receipt of a
READ signal on line 127.
The microprocessor is connected to a printout
control 128 by a line 129, and the microprocessor can control
the operation of a printer, recorder or qther device to
which the printout control is connected. Depressing PRINT
key 29 causes the current weight and/or count reading
to be transformed to the external device.
As illustrated in Fig. 5, weight data is
_g_ .
,
..
ti53~
registered in a weight register 131 which, in the preferred
embodiment, comprises a random access memoxy. In order to
enhance the accuracy of the scale, weight determinations are
based on averaged readings, rather than individual readings.
S For this purpose, the eight most recent readings from
converter 79 are read into memory cells in register 131 on
a rotating basis, whereby each new reading replaces the oldest
reading in the register. The register is divided into four
sections, designated 0-3, and the memory cells for the
incoming weight readings~ designated WTl-WT8, are located in
sections 0 and 1. Each of the cells contains 16 bits
arranged in four 4-bit bytes, and up to 64,000 counts can
be registered in each cell. In order to provide a resolution
of 1 count in 100,000, both positive and negative count~
are used, and a suitable count such as -50,000 is chosen
: as the zero reference. As illustrated for WTl, the
least significant four bits in each cell range in
significance from 1 count (.001% of full scale) to 8 counts
(.008% of full scale), the next most significant four
bits range from 16 counts (.016% of full scale) to 128
counts (.128% of full scale), the next most significant
four bits range from 256 counts (.256~ of full scale) to
2,048 counts (2.048% of full scale), and the most
significant four bits range from 4,096 counts (4.096~
of full scale) to 32,772 counts (32.772% of full scale).
The polarities of weight readings WTl-WT8 are
entered into a polarity register in section 2 of register
1310 In this register, positive polarities are indicated
by the number 0000, and negative polarities are designated
by the number 0001.
--10-
4~;538
The weight readings in register 131 are
averaged each time a new reading is taken to provide
a moving average of the weight readings. This a~erage is
registered in a ~otalizing register in section 3 of register
131. The polarity of ~he average is registered in a computer sta-
tu~ registèr in aform similar to the manner in which the
polarities of the weight ratings are registered in the
polarity register.
In the preferred embodiment, the averaging
is done by addi~g ~he least significant four bits of the
first two weight readings together and registering the
sum in byte A of the totalizing register. The process is
continued for four bit groups of increasing significance
until the number stored in the totalizing register is the
sum of the first two weight ce'ls. The polarities of the
two readings are then checked, and if they are the same,
this polarity is assigned to the sum registered. If the
polarities are different the polarity of ~he sum is
determined by the presence or absence of a carry from the ,'~,
addition. If there is a carry, the sum is positive, and
- if there is no carry, the sum is negative. The correct
polarity is assigned to the sum. This process is
continued until all eight weight readings have been added
together and the number in the totalizing register
represents the total of the eight readings and the
polarity registered in the status, register is the polarity
of the total.
Since the sum of eight binary numbers is equal
to one-half of the average of the numbers, the number
registered in the totalizing register upon completion of
5313
the addition prGcess is equal to one-half of the average
--~ of the weight readings. To get the average, the total must
be multiplied by two, and this is done simply by shifting
the number in the totalizing register one place to the left,
following which it will be in the form illustrated in Fig.
6. It will be noted that the resolution has increased from
.001% of full scale for the individual readinys to .000125
for the average.
The number registered in the totalizing register
represents the average gross weight of the load auring
eight successive clock intervals. The average net weight
c~n be determined by subtracting the average gross weight
at a reference time, e.g., before the load is placed on
the scale, from the current value of the average gross
weight. In the preferred embodiment, this subtraction is
effected by storing a number Z which is the complement
of the average gross weight G at the reference time. This
number is stored in section 2 of register 131, and its
polarity is stored in another status register. The average
net weight is then determined by adding the numbers G and
Z together, and the polarity of the net weight is stored
in the shift register. As a result of the averaging and doubling
process, the least significant bit of the number Z repre-
sents .0000625~ of full scale.
As illustrated in Fig. 6, means is provided for
increasing or decreasing the net weight reading in small
increments during successive counting cycles to maintain
an accurate zexo reference. The net weight is checked
during each counting cycle, and if the absolute value is
greater than zero but less than .004% of full scale, the
-12-
1C~4~53~3
net weight is made 1/16 of .001% closer to zero by
incrementing or decrementing the number Z by that amount.
If the net weight is positive, the nurnber Z is increased
by 1/16 of .001% of full scale, and if the net weight is
negative, the n~nber Z is decreased by this amount. The
adjusting process continues in successive cycles until the
net weight is exactly zero. If the absolute value of the
net weight is greater than .00~% of full scale, it is
assumed that there is a load on the scale, and no adjustment
is made. If desired, a value other than .004% can be
utilized, but the value chosen should be less than the
expected weight of articles to be weighed on the scale.
~ikewise, a different value can be used for the increment
by which the net weight is adjusted. ~owever, this value -
is preferably substantially less than the range in whi~h
adjustments are made.
The flow chart of Fig. 7 illustrates the manner
in which weight readings are made and the number of
articles on the scale is determined. Instantaneous weight
readings from converter 79 ar~e read into the microprocessor
where they are averaged in the manner described above. -~
The keyboard is then scanned, and the weight readings are
checked to determine whether the scale is in motion. If
the scale is not in motion, and ZERO switch 124 is closed,
the number Z is replaced by the complement of ~he number
then in the totalizing register. The net weight is then
checked for the zero reference level, and if necessary
it is increased or decreased in the manner described
above. If the scale is in motion, the zeroing and net ;;
wei~ht adjustment steps are bypassed. If gross weight
~13-
.
s~
switch 121 is closed, the gross weight reading is
tr~nsformed from bi~ar~ to binary coded decimal form
and displayed. If the net weight switch is closed, the
net weight is transformed to binary coded decimal form
S and displayed. If the net weight switch is not closed, the
magnitude of the weight is checked, and if the weight
exceeds the capacity of the scale, an overscale indication
such as a plurality of E's is displayed by the weight
display.
If the net weight switch is open and the weight
reading is within the capacity of the scale, the scale factor
set by control 111 is read, and the net weight is multiplied
by this factor. Thereafter, the net weight is transformed
to binary coded decimal form and rounded off to the desired
number of digits. If the net weight is negative, it is
~ displayed at this point and the COUNT subroutine which
determines the number of articles on the scale is bypassed.
If the net weight is positive, the number of articles is
determined in the-manner described below. If switch 123 is
open, changes in the weight and count~readings are displayed
dynamically, and if the switch is closed, the count is not
displayed until the scale stops moving and a steady stage
is reached.
Thenanner in which the number of articles is
determined is illustrated in Fig. 8. If none of the sample
size selection keys 24 has been depressed, the remainder
of the COUNT subroutine is bypassed, and only the weight is
displayed. When a new SAMPLE SIZE key is depressed, the weight
of the sample on the scale is checked to make certain that it
is large enough to be determined accurately. In the
14-
~O~t;538
embodiment illustrated, the net weight of the sample
must be at least .032 per cent of the full scale value,
or the sample will be rejected and the count display will
display the word "ADD", indicating that a larger sample
is re~uired.
If the sample is of sufficient size and the scale
is not in motion, as indicated by the latest gross weight
reading being within +.004 per cent of the average gross
weight, a counter is incremented. When 16 successive weight
readings are within the specified range, th~s weight is stored
as the sam~le weight, and a KEY flag is set to indicate that
the sample has been established. If any of the weight
readings differ by more than .004 per cent of the average
during the determination of the sample weight, the counter
is reset to zero, and the process continues until 16
successive readings are within the specified range.
Once the sample weight is determined, the net
weight on the scale is divided by the sample weight, and
the quotient is multiplied by the number of articles in
the sample, as defined by the key depressed to select
the sample size. The resulting number corresponds to the
number of articles on the scale. This number is rounded
off to the nearest integer and transformed from binary
form to binary coded decimal form. This number is then -~
displayed unless it is greater than 100,000, in which ~;`
case an over-scale indication is given, for example,
by displaying a plurality of E's in the. count display.
Operation and use of the counting scale can be ;
described briefly. A sample consisting of 8, 16, 32 or
64 articles of a type to be counted is placed on platform ~ ;~
-15-
3~3
14, and the corresponding SAMPLE SIZE button 26 is
depressed. The weight of the sample is indicated by display
31, and the number of articles in the sample is indicated
by display 32. If the sample is too light to provide a
satisfactory count, display 32 will display the word "ADD"
to indicate that a larger sample should be used. Once a
s~mple has been established, articles can be placed on or
removed from platform 14, and display 32 will indicate the
number of articles on the platform at any given time. A
new sample can be entered either by depressing a new SAMPLE
SIZE key or by depressing RESET key 28 and then depressing
a new SAMPLE SIZE key.
Depressing ZERO or TARE key 27 resets the weight
and count displays to zero and causes the weight then on the
scale to be substracted from subsequent weight readings so
that subsequent weight and count indications will represent
~et values. The 2ERO key does not affect the weight stored
for the sample, however.
Switch 36 permits the operator to determine whether
the weight and count are displayed on a continuous basis or
only when a steady state has been reached following a change
in the load on the scale. Depressing PRINT key 29 causes
the current weight and count information to be transferred
to an external device connected to the scale.
In many counting applications, it is not necessary
to know the weight of the articles being counted/ and the
weight display can be omitted if desired.. Omission of this
display will reduce the number of components required and
permit the use of a less expensi-le load cell since absolute
weight readings are no longer necessary. Even without a
.
~4~538
separate weight display, weight readings can still be
obtained by using a sample having a weight which
corresponds to one of the sample sizes, e.g. 8, 16, 32 or
64 pounds. For example, by placing a weight of .32 pounds on
the scale and depressing the key for a sample size of 32,
the scale will be calibrated ~o read in units of .01 pound.
Likewise, by placing 64 grams on the scale and depressing
the key for a sample size of 64 pieces, the COUNT display
will give weight readings in grams.
It is apparent from the foregoing that a new and
improved counting scale and method have been provided. While
only certain presently preferred embodiments have been
described, as will be apparent to those familiar with the
art, certain changes and modifications can be made without
departing from the scope of the invention as defined by the
following claims.
''. ~'" ' '
';' ~ ' :
i ,, ;
-17-
