Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
-`- 1082945
,:
In the prior art it is known to obtain an indication
of color and brightness characteristics of a paper web during
manufacture by an on-line measurement of reflectance value (Rg),
but this measurement is decidedly different from that necessary
for actual color and brightness characterizations. Accordingly,
such a measurement must be accompanied by very frequent off-
line testing, so as to enable an adequate empirical calibration
of the measuring instrument. Further, a separate set of
calibration parameters is required for each grade and weight of
paper. Off-line instruments which adequately measure these
characteristics require that a pad of several thicknesses of
paper be exposed to the light source aperture so that a different
reflectance value (Roo) is obtained. Obviously this is impossible
with an on-line instrument unless the far more inaccessible reel
itself is tested.
Only where the on-line measured reflectance value (Rg)
approaches the off-line value (Roo), as in instance of paper of
extremely high opacity such as heavily coated or heavily dyed
paper, can the above problems be minimized to the point where
'; 20 accuracy becomes sufficient for control purposes.
This invention relates to an optical device and
method for sensing optical properties of a paper sheet material,
and particularly to an on-the-paper-machine device and method
~ for simultaneously sensing both transmitted and reflected light
; so as to obtain measurements from which the optical properties
of interest can be calculated substantially independently of
grade and weight of paper involved.
In accordance with the invention there is provided
apparatus for controlling the production of sheet material
including an optical measuring device having a receiving region
, -
',,' .~ - 1-
.
,. ,.
082945
for receiving in operative relation thereto a single thickness
of sheet material, said optical measuring device having an
"~ optical system with at least two substantially independent
photometric sensors and at least two distinct light energy paths
each including at least light source means and a respective one
of said photometric sensors, and each intersecting said receiving
. region prior to the respective associated photometric sensor,
f~ said light source means supplying a spectrum of light energy to
.- the light energy paths from one side of the receiving region
accommodating characterization of any one of a plurality of
optical properties comprising at least two of color, brightness
and opacity, said paths being respectively arranged for
collecting reflected and transmitted light energy from the
receiving region after impingement of the light energy on a
single thickness of sheet material at said region to provide
: reflectance and transmittance measurements which essentially
characterize a plurality of said optical properties.
Accordingly it is an object of the present invention
to provide an optical monitoring device and method for sensing
. 20 optical properties
`.
: '
: `
~ la -
1082945
.
based on measurements made on a single thickness of partially trans-
Iucent paper sheet material and which measurements sufficiently
characterize the actual properties of interest that a minimum of
empirical calibration is required regardless of changes in grade and
S weight of paper.
Another object of the invention is tO provide such an optical
monitoring device and method capable of accurately sensing optical
properties such as brightness, color, opacity and/or fluorescent
contribution to brightness.
While such an optical monitoring device is useful off-line for
sensing optical properties of a single thickness sample, it is a further
important object of the present invention to provide such an optical
monitoring device which is of sufficiently light weight and compact con-
. ~
struction so as to be adapted for on-line monitoring of the
desired optical properties.
Another and further object of the invention is to provide an
on-the-paper-machine optical monitoring device of sufficient
flexibility and accuracy to enable control of desired optical properties
~; during the paper making process.
~ .,
A unique feature of the on-line optical monitoring device is
its ability to simultaneously measure both reflected and transmitted
.,.j
'~ Iight. By measuring two independent optical parameters it is possible
~,
'~, to thoroughly characterize the paper optical properties of a partially
~, !
~, translucent web with a minimum of empirical correction for factors
; 25 s uch as pa per grade a nd we ight .
:, ~
~ Other objects, features and advantages of the present inven-
`, tion will become apparent from the following detailed description
.
taken in connection with the accompanying drawings.
~,.j,
s~ 2
.,;,
.
108294S
ON THE DRAWI~GS
Fig. 1 is a fragmentary somewhat diagrammatic longi~udinal
sectional view of a paper machine showing in ou~line a side view of
an op~cal monitoring device in accordance with ~he presenr invention
operatively moun~ed on line with the machine;
Fig. 2 is a fragmenrary somewhat diagrammatic cransverse
sectional view of the paper machine of Fig. 1 and taken generally as
indicated by the line Il-II of Fig. 1 and looking in the direction of the
arrows (toward the wet end of the paper machine), the view being taken
so as to show in outline a direct front view of the optical monitoring
device of Fig. l;
Fig. 3 is a diagrammatic longitudinal sectional view of an on-
the-paper-machine optical monitoring device in accordance with the
present invention;
Fig. 4 is a partial diagrammatic plan view of the filter
wheel asse~T~bly utilized in the monitoring device of Fig. 3;
Fig. 5 is a somewhat diagrammatic view illustrating an
optical analyzer unit in electrical association with the optical monitoring
device of Figs. 1-4 and with a power supply unit;
Fig. 6 is an electric circuit diagram illusrrating the
electrical connections between the various components of Figs. L-~;
-
i
~'~
r~, ~ 3~
;'
1082945
.:
.
.: Figure 7 is a flow chart illustrating an existing
direct digital control analog point scan program which has
been adapted to allow for the collection and temporary stor-
age of the reflectance and transmittance data acquired from
the system of Figures 1 - 6;
Figures 8-16 when arranged in a vertical series
: represent a program fourteen which is designed to read the
reflectance and transmittance values stored pursuant to
Figure 7 and generally to control the operation of the sys-
~ 10 tem of Figures 1 - 6 and to apply correction factors to the
::.; raw reflectance and transmittance data;
... :
s Figures 17-20 when arranged in a vertical sequence
represent a data reduction program forty-two whose purpose is
-.~ to reduce the corrected reflectance and transmittance data as
.. . .
produced by the program of Figures 8 - 16 into terms with
which papermakers are familiar and upon which paper optical
,.,
' specifications are based, e.g. brightness, opacity, color and
; fluorescence;
.
~'
, . . .
''~
J~
,~
~082945
. .
DESCR~ION OF THE PREFERRED EMBODlMENTS
Detailed Descrip~ion Of The Apparatus
Of Fi~s. 1 and 2
, Figs. 1 and 2 will serve to illustrate the mcdifications of an
S existing paper machine which are re~uired for carrying out a preferred
embodiment of the present invention. Referring to Figs. 1 and 2, an
on-the-paper-machine optical monitoring device is diagrammatically
indicated at 10 and comprises an upper sensing head 11 and a lower
sensing head 12 which are maintained u~ precise relact~e alignment
10 and disposedfor operative association and transverse scanning move-
ment relative to a paper web loca~ed as indicated at 14 in Figs. 1 and
2. .As will be described hereinafter with reference to Figs 3 and 4,
in a particular design of the optic~l monitoring device, upper head 11
includes a light source for projecting light onto the web such that a
15 portion of the light Ls reflected parallel to an optical axis indicated at
15, while a further portion of the Iight is transmitted through the paper
web for collection and measurement by means of the lower sensing
head 12.
For purposes of illustra~ion, Figs. 1 and 2 show portions of an
20 existing we~ scanner construction which is utilized to scan the we~ 14
for conventional purposes. The conventional scanner construction
includes fixed ~rame components such as 20, 21 and 22 forming what
is known as an "O" type scanner frame. The conventional scanning
, . .
,-' s1rucNre further includes upper and lower slides 25 and 26 for joint
,. . .
2S horizon~al movement along the horizontal beams 21 and 22. Associa~ed
.~ .
, with the slides 25 and 26 are mova~le assemblies 27 and 2~ carried
by the respective slides 25 and 26 and including verticaUy disposed
plates 31 and 32 and angularly disposed flange members such as
-5-
108Z945
indicated al 33 and 34 in Fig. 1. These flange ?ortions 33 and 34
have broad surfaces Iying in planes generally parallel to the plane of
the web 14 and are utilized for mounting of the monitoring device 10
of the present invention. In particular a top head mounting braclcet is
indicated at 41 in Figs. 1 and 2 and is shown as being secured to the
existing flange part 33 so as to mount the upper head 11 for scanning
; movement with the assembly 27. Similarly a lower head mounting
, ~ bracket is indi'cated at 42 and is shown as being secured to flange part
34 of the lower movable assembly 28 so as to mount the lower sensing
head 12 for scanning movement joindy with the upper sensing head 11.
For the purpose of electrical connection with the monitoring
' device 10 during its traverse of the web 14, electric cables are indicated
~,~ at 51 and 52 for electrical connection with the components of the upper
'I sensing head 11 and lower sensing head 12 of the monitoring device 10.
The cable 51 is shown as being fastened by means of straps 53 and 54
~ to a top carrier slide bracket 55. The bracket is shown as being.~ secured by means of fasteners 56 and 57 to the upper portion of vertical
;~ plate 31. As indicated in Fig. 2, successive Ioops of cable 51 are
secured to swivel type ball bearing carriers such as indicated at 61.
A trolley track 62 is supported f~om existing channels such as indicated
at 63 and mounts the carriers 61 for horizontal movement as required
,
tO accommodate the scanning movement of the monitoring dence 10
, across the width of the web 14. Similarly, successive Ioops of the
cable 52 are fastened to the eyes such as indicated at 71 of a lower
25 series of carriers 72. As seen in Fig. 1 each of the carriers such
~j as 72 includes a pair of rollers such as 73 and 74 riding in the trolley
track 75 which is secured directly to the lower flange 22a of beam 22.
, A lower carrier slide brackel 31 is secured to vertical plate 32 by
~- ~082g45
means of fasteners 82 and 83 and is provided with a horizontally extend-
ing flange 84 for engaging with the first of the series of lower carriers
72. In particular, carrier 72 is provided with a shank 85 which extends
into a longitudinal slot 84a of flange 84. Thus, the first carrier 72 is
S interengaged with the bracket 81 and is caused to move with the lower
assembly 28 and the lower sensing head 12. The remaining lower carriers
such as that indicat~dat 83 move along the trolley track 75 as necessary
to accommodate movement of the monitoring device 10 transversely of the
,; web 14.
. .,
While Figures 1 and 2 have illustated the optical monitoring
device of the present invention as being mounted on line withthe paper
machine and have further illustrated the case where the monitoring device
is to be scanned transversely of the web, it is considered that the opti-
cal monitoring device of the present invention would also be of great value
.,
15 if redesigned for bench mounting. By placing a single sheet of paper in
a sample mount of the device, a technician could simultaneously test the
sample for color, brightness, fluorescence, and opacity in a matter of
seconds.
In the illustrated embodiment, however, it is contemplated
., I
20 that the monitoring device 10 will be mounted on line with the paper
machine and will be capable of movement to a position clear of the edge
'.;
of the web as indicated in Fig. 2 at the end of each hour of operation, for
example. W~n the end of a production run for a giv~ we b 14 has been
reached, or when a web break occurs for any other reason (such as
25 accidental severance of the given web), the monitoring device 10 will be
moved clear of the edge of the web path as indicated in Fig. 2. Each time
the moni~oring device 10 is moved tO the off-web position shown in Fig. 2
it is preferred that readings be taken of the reflectance and transmittance
--7--
~082945
,~
values (without the web in theoptic path) for the purpose of obtaining
an updated calibration of the monitoring device. Thus, such updating
of calibration may take place automat<ically (for e2~ample under the
i' co~rol of aprocess control computer controlling the paper manufac-
,.
5 turing operation) at hourly intervals and also after web breaks. The
monitoring device can, of course, be retracted manually any time
desired by the operator for the purpose of checking calibration. By
~i way of example, the monitoring device 10 may be capable of a normal
~;~ scanning travel over a distance of 115 inches with provision for an
10 additional travel of 16 inches to the position shown in Fig. 2.
A fla~ge is i~dicated at 87 which serves to insure proper re~ngagement
of the 3ensing head with the web at the operator's side of ~he illustrated
'ii paper machine (opposite the side indicated in Fig. 2).
'J The lower head 12 is designed to contact the web 14 during
~t 15 scanni~ thereof. The design spacing between the upper and lower
heads 11 and 12 is 3/16 inch. The optical opening in the upper head 11
is aligned with the optical a~cis 15 and is to be maintained in alig~ment
~j with the center of the window in the lower head 12. Four ad3usring
screws such as those indicated a~ 91 and 92 æ~provided for accurate
~; 20 positioning of the upper head 11. Similarly four position adjusting
^.,
screws such as 93 and 94 serve for ~he accurate positioning of the
,. 10WOE head in conjunction with set screws such as indicated at 95 and
- 96. The adj~ting screws are located at each corner of mounting
` brackets 41 and 42.
~- -8-
1082945
~.
Modifications of Figs. 1 and 2 To Insure Accurate Scanning
.
i Where the web is not perfectly horizontal, but instead is curved
across, its width, it is desirable to provide a web deflecting guide bar
as indicated at 97 in Fig. 3 for insuring stable contact between the web
14 and the web engaging surface 98 of the lower sensing head 12 By
way of example the guide bar may protrude from the lower surface of
, . .
the upper sensing head a distance of S/16 inch so as to overlap with
respect to the vertical direction a distance of 1/8 inch relative to the
lower sensing head web contacting surface 98. The guide bar 97 may
have a width to force down at least about four inches of the width of the
web at a section of web centered with respect to web engaging surface
98 of the lower sensing head relative to the machine direction. This
insures a minimum of a 1/8 inch bellying of the sheet as it travels
;; over the lower sensing head in all lateral positions of the sensing head.
,," 15 In order to minimize changes in the 5/16 inch thickness dimen-
... .
sion of the guide bar 97 due to wear, the guide bar is provided with a
, I .
', flat web engaging surface 97a which has a dimension in the direction of
web movement of about one inch. By way of example, the guide bar
may be made of Teflon (trademark).
Since the guide bar 97 is not necessary when the web is fed from
the calender stack to the reel in a relatively planar configuration, it
has not been shown in Figs. 1 and 2.
Various modifications may of course by made tO adapt the moni-
toring device of the present invention to various types of paper ma-
chinery, and to secure any desired degree of accuracy in the joint s~-
ning movement of the upper and lower s~n~g he~ls reL?live to the paper.
:
9 _
...
1082945
Structure Of The Optical
Monitoring Device As Shown in Figs.
3 and 4
Referrring to Fig. 3, the upperq~nsing head 11 is shown as
comprising a casing 110 having suitable connectors 111 and 112 for re-
,, ceiving suitable internally threaded fittings 114 and 115, Fig. 1, asso-
ciated with the electric cable 51. The casing 110 receives a top head
shoe 120 including an interior open rectangular fra~ 121 having a base
flange 121a spot welded to shoe plate 122. The upstanding portion
121b engages the ad~acent wa11 of casing 110 along all four sides thereof
~' .
and lS secured to the casing 110 by suitable fastening means such as
indicated at 124 and 125 in Fig. 3. An edge 122a of shoe plate 122 is
bent up at an angle of 45~ at the side of the sensing head ~;1 facing the
;? wet end of the paper machine, and a sim1l~r inclined edge 122b, Fig. 1,
~,j
j 15 is provided at each of the sides of the sensing head ~o as to prese~
it~j smooth faces to the paper web during scanning movement of the sensi~g
head. The shoe plate 122 is provided with a circular aperture of less
,1
~`~; than one inch diameter as indicated at 130 centered on the optical a~is la
of the device. I~ a present embodiment aper~ure 130 has a diameter of
abcut t/8 inch. This aperture 130 is preferably of mL~imum diameter
necessary to accommodate the light paths of the instrument. In the
''il
i31ustrated embodiment the light path for the incident light beam as indi-
i cated at 133 is di~ected at an angle of approximately g5~ and is
, . .
focused to impinge on a window 135 at the optical a~s 1~. A reilected
light path 2S indicated at 137 is normal to the web engaging surface 98
(which is the upper surface of window 135), and is coincident with tt~
optical axis lj, while light transmitted through the web 14 and through
,. ..
~; the window 135 is directed as indicated by rays 1~ 3, for example,
into an integrating cavity 145 of lower head 12.
: `
; -10-
;,.';
108Z945
The lower head 12 comprises a casing lS0 having an annular
dished plate 151 secured thereto and providing a generally segmental
spherical web-contactilgsurface 151a surrounding window 135. The
window 135 is preferably formed by a circular disk of translucent
5 diffusing material. In the illustrated embodiment the window 135 is
made of a polycrystalline ceramic material available under the trade-
mark "Lucalux" from the General Electric Company. This material
has physical properties similar to that of sapphire. The opposite
faces of window 135 are flat and parallel and the thickness dimension
.,'! 10 iS 1/16 inch. A lip is indicated at 153 for underlying an annular edgeportion of window 135. This lip provides a circular aperture 154
having a diarneter of about ~S~irch0that the effective viewing area
for the transmitted light is determined by the diameter of aperture
154. The casing 150 is shown as being provided with an electrical
15 connector terminal 155 for receiving a suitable internally threaded
fitting 156, Fig. 1, of cable 52.
'
. .,
~,' -11-
,.. .
108X945
. :
(,
,.
~'
As diagrammatically indicated in Figs. 3 and 4, the upper
,, sensing head 11 includes a light source 201, incident optical path rr.eans
including lenses such as indicated at 202 and a photocell 203 for
measuring reflected light returning along the reflected light path 137.
5 A filter wheel 210 is shown diagrammatically as being mounted on a
shaft 208 for rotation by means of a low torque motor i~dicated at 209.
~,~! As best seen in Fig. 4, the filter wheel includes an outer series of
~'~ apertures 211-217 for selective registry with the incident light beam
path 133, and includes a series of inner apertures 221-227 for selec-
10 tive registry with the reflective light beam path 137. The various
, apertures may recei~esuitable filter elements as will hereinafter be
;~,
~; ' explained in detail such that a series of measurements may be tak~
by successively indexing the filter wheel 210 to successive operating
,.,~
positions. In each operating position one aperture such as 211 is in
15 alignment with the incident beam path 133 and a second aperture such
as indicated at 221 is in alignment with the reflected light beam path 137.,
. -, .
$~ By way of example, the motor 209 may be continuously energized
,
.
."
",~
- -12 -
'.
.'
108~g45
during operation of the monitoring device, and the filter wheel may be
retained in a selected angular position by engagement of a ratchet ~m
230 with one of a series of cooperating lugs 231-237 arranged generally
as indicated in Fig. 4 on the filter wheel 210. A solenoid is indicated
at 240 as being mechanically coupled with ratchet arm 230 for momen-
tarily lifting the ratchet a,.L. 230 out of engagement with a cooperating
lug such as 231 so as to permit the filter wheel to index one posidon.
; Immediately upon release of the energization of solenoid 240,the force of
gravityreturns the ratchet arm 230 to the position shown in Fig. 3 so
as to be disposed in the path of the lugs and thus to engage the nex~
lug in succession such as lug 232 as the motor 2~9 moves the filter wheel
210 into the next operating position.
As will hereafter be explained in greater detail, reed switches
are mounted in circles on respective switching boards 241 and 242,
Fig. 3, and the filter wheel shaft 208 carries a magnet 243 for
acmating a respective pair of the reed switches in each operating posi-
tion of the filter wheel 210. Thus the position of the filter wheel 210
determines which of the switches on the switching boards 241 and 242
are closed. As will be explained hereinafter, the reed switch on tAe
upper switching board 241 which is closed determines tAe gain setting
of an upper head amplifier at a level appropriate for the set of filrers
which are in the operating position. The reed switch on the lower
switching board 242 which is closed activates a relay on a circuit board
245 in the lower head 12, and such relay in turn sets the lower head
amplifier gain at ;he proper level. As will be explained in connection
with the electric circuit diagram for the monitoring device, cer~ain
conductors of the cable ~1 may be interconnected at a remote location
~' -13-
:' .
.~
` ~082945
; . .
`::
,; so as to cause an indexing movement of the filter wheel 210. This
external command serves to momentarily energize solenoid 240 and
: lift the ratchet arm 230 about is pivot point 250, allowing the motor
209 to rotate the filter wheel 210. The ratchet arm 230 returns to the
5 position shown in Fig. 3 to catch the next lug on the filter wheea stalling
the motor 209.
Four heate~ssuch as indicated at 251 are mounted around photo-
cell 204 ~o as to minimize the temperature variations of the photocell.
A circuit board for mounting an amplifier for photocell 2û3 and for
10 mounting the gain se,~ting ~esistances associated withthe reed
j ~witches is indicated at 255 in Fig. 3.
::
Referring to the lower head 12, Fig. 3 indicates a photocell
` 260 for receiving light from the intergrating cavity 14; and a series
of heaters such as 261 mounted around the photocell 260 to minimize
15 the temperature variations of the photocell. Circuit board 245 may
mount a suitable amplifier for photocell 260, the gain of which bei~
controlled by the relays previously mentioned.
~ The heaters 251 and 261 in the prototype unit were Pennsyl-
; vania Electronics Technology Type 12raa. (These are positive tem-
20 perature coefficient thermistors with 55~. switching temperatures.)
,, .
These heaters will t~nd to stabilize the te;nperature si~ e their
ability tO provide heat decreases as the ambient temperature increases.
Above ~5~., they pro;Tide essentially no heat at all.
~,,
' ~1
.,~
-14-
..
~08Z94S
Discussion of Illustrative Operating Details for the MonitoAng
Device of Figs. 3 and 4
A basic feature of the illustrated embodiment resides in its ability
to measure simultaneously both reflected and transmitted light. While
S in the illustrated embodiment, the reflected light path 137 and the
transmitted light path intersect the web 14 essentially at a common
point, reflected light could be obtained from a point on the sample or
web offset from the point where light is transmitted through the sample.
For example, a backing of some specified reflectance such as a black
body of zero or near zero reflectance could be located on the lower
sensing head just ahead of or behind the transmitted light receptor
compartment (with respect to the machine direction of the sample or
the direction of movement of the web). In this case the upper sensing
head could contain the light source as well as a reflected light receptor
for receiving light reflected from the sample or moving web at a point
directly above the backing of specified reflectance. Both the reflected
light receptor in the upper sensing head and the transmitted light recep-
tor in the lower sensir~ head could then supply signals simultaneously
and continuously during measurement operations. Many other varia-
;~ 20 tions in the arrangemement of the optics for measuring both reflected
, and transmitted light will occur to those skilled in the art.
Referring to the details of the illustrated embodiment, however,
.
and to the case where it is desired to measure brightness, color, opa-
city and fluorescent contrihni~to brightness, li~t s~e 201, ~Lg.3, may
- 25 consist of a Mode3 1962 Q~artzline (trademlr~) la~ ope~ated at 5. 8 volts as
~-~ measured at the lamp terminals. The 45 incident beam pQth l~and~e
normal reflected beam path 137 corres~nd to those a~ a standa~d bright-
ness tester, and a castulg (not shown) from a bench type star~lard bright-
- 15-
:':
.
~ ` ~
1~82g4S
` ness tester was used in constructing a prototype of the illustrated
, embodiment to give rigid support for the optical components such
as indicated at 202 and 271-276 in Fig. 3. In the specific
prototype unit, a stock thickness polished Corning type 4-69 glass
S fflter 271 and a second type 4-69 filter 272 ground and polished
to an appropriate thickness were used in the incident beam path
to absorb most of the infrared as well as to give proper spectral
' response.
The reflected light path 137 included a pair of lenses 273
and 274 which focus the light on a 3/8-inch aperture in the plate
275 of the casting. A piece of diffusing glass 276 is located on
the 3/8-inch aperture so that the light distribution over the sur-
face of photocell 203 will be reasonably uniform. A Weston
, ~ .
model 856 RR Photronic (trademark) cell was employed.
The filter wheel 210 is designed and located in such a way
' that either the incident or the reflected beam or both can be
filtered as desired. In the prototype, the wheel 210 was driven
....
by a small motor 209 operated at reduced voltage so that it could
operate continuously in a stalled condition.
... .
~.
.:
,,; . .
. .
,'
''':
- 1 6-
~ ~082g45
Commerically available color and brightness meters are
usually manufactured with the spectral respDnse $1ters located in the
reflected beam. In the prototype device, and in the later on-machine
version here illustrated as well, however, the filters which determine
5 the spectral response of the first six filter positions are located in the
incident beam. There are two basic reasons for this choice of desig~.
(1) Both the reflected and transmitted light have the same
incident inten,sity and spectral response against which
each can be compared. The alternate would necessitate
two sets of identical filters, one set located in the re-
flected beam and another in the transmitted beam--a dif-
....
ficult design to achieve in practice. -
(2) Filters in the i~cident beam can be used to absorb all
ul~aviolet light and prevent it from striking the specimen.
Thus, fluorescence, a phenomenon not accounted for by
Kubelka-Munk theory is avoided.
.. . .
For reasonsexplained shortly, the seventh filter poæition is
an exception to the a~ve in that s~bsta~ial ultraviolet light is inten-
Al tionally permitted to exist within the incident beam. Outside of the
` 2û phenomenon of fluorescence the spectral response is independent of
whether such filters are located in the i~cident or the reflected beams.
The spe~al response provided by the respective positions
of the filter wheel 210 were as follows: (1) papermaker's brightness
¢TAPPl brightness), (2) blue ponion of the Ecx function, (3) red por-
',,! 2S tion of the Ecx function, (4) E ~ function without f~uorescence (~) Ecy
function, (6) Eay fun.tion, and (7) EC2 function, with fluorescence.
-17-
. .
` 1~8Z945
As is understood in the art, the symbols E x, E y, Eay,
and Ecz refer to tristimulus functions of wavelength as defined by the
Commission Internationale c l'Eclairage which is identified by the
abbreviation C.I.E. and is also known as the International Committee
on Illumination. The subscript a in the function designation E y
indicates that the function is based on a standardized illumination
designated as C.I.E. Illuminant A, while the subscript c in the other
function designations refers to a somewhat different standardized
illumination which is designated as C.I.E. Illuminant C. -
Filters for providing the above spectral response character-
istics in the respective operating positions of the filter wheel 210
have been indicated in Figure 4 by reference numeral 281-288. In the
specific example under discussion, apertures 221-226 are left open.
Filter 281 is a standard filter for use in measuring TAPPI brightness,
. TAPPI referring to the Technical Association of the Pulp and Paper
,..~
Industry. This filter transmits a narrow band of wavelengths in the
vicinity of 457 nanometers.
~ Filters 282-285 are standard filters for a four-filter
6 colorimeter and are conventionally designated X (blue), X (red)~ Z,and Yc~ These filters provide the wavelength distributions required
~: for the measurement of the C.I.E. X, Y, and Z tristimulus values
under Illuminant C.
Filter 286 is conventionally designated as a YA filter and is
required by the TAPPI standard method for opacity measurements. This
is a broad band filter producing the C.I.E. Y wavelength distribution
for Illuminant A, in conjunction with the source 201 previously describ-
ed in this section. A discussion bearing on the feasibility of this
type of measurement is found in a paper by L.R. Dearth, et al entitled
~`~ "Study of Instruments for the Measurement of Opacity of Paper, V. Compari-
,:
- - 18 -
::
1082945
son of Printing Opacity Determined with Several Selected Instruments,
Tappi, volume 53, No. 3 (March, 1970).
With respect to position No. 7 of the filter wheel 210, filters
287 and 288 are convendonally designated as Z (blue) and Z (yellow).
S As prenously indicated, the purpose of the filters is to provide for a
determination of the C.I.E. Z tristimulus value with the fluorescenoe
component included. In filter posidon No. 4, filter 284 serves to re- -
move the ultraviolet component frcmthe incident beam so that a measure
o~ the Z tristimulus value without fluorescence is obtained. In position
10 No. 7 of the filter wheel, however, filter 287 in the incident beam is
designed to transmit the ultraviolet component, so that the fluorescent
component if any will be transmitted to photocell 203. The ultraviolet
absorbing component of the Z type filter means is located in the re-
flected beam 137, whOEeas this component is in the incident beam for
' 15 the No. 4 position. T~e fluorescent component is lineally related to
the dfflerence be~ween the Z tristimulus values deterrnined in the No. 4
and No. 7 posirions of the filter wheel 210.
Filters 281-288 have been shown in Fig. 4 with different types
of hatching which have been selected to represent generally the dif-
20 ferent lighttransmission properties of the filters. In particular, thebatching f~r flltOEs 281^288 are those for representing wbite, blue, red,
.: .
blue, green, orange, blue and yellow light transmission p~pOEties. The
selection of hatching is primarily for purposes of graphical illustration
and is not, of course, an exact representation of the light transmission
25 properties of ~he respective fllters.
:'~
-19-
"
~08Z945
Detailed Description of Figs. 5 and 6
Fig. S illustrates diagrammatically the optical monitoring device
10 of Figs. 1-4, and illustrates by way of example an optical analyzer
unit 300 which may be electrically associated with the monitoring device
5 and serve as an operator's console to be disposed at a convenient loca-
; tion adjacent the paper machine. By way of example, the optical analy-
zer unit may be mounted near the dry end of the paper machine, andmay
receive conventional alternating current power from the paper machine
dry end panel. The optical analyzer unit 300 is illustrated as being
coupled with the monitoring device 10 via a power supply unit 301
which is mounted adjacent the vertical column 20, Fig. 2, of the "O"
frame along which the monitoring device is to travel in scanning the
width of the web. For purposes of diagrammatic illustration,power supply
unit 301 is shown as being provided with a mounting plate 302 which is
~' lS secured by means of a bracket 303 to an end of horizontal beam 22 which
has been specifically designated by reference numeral 304 in Figs. 2
and 5. Referring to Fig. 2, it will be observed that the ends 305 and
~,
306 of cables 51 and 52 are adjacent the end 304 of beam 22 so that
this is a convenient location for mounting of the power supply 301.
The electrical interconnections between the power supply unit 301 and the
optical analyzer unit 300 are indicated as extending via a signal conduit
31] and a control conduit 312. By way of example, the signal conduit
311 may contain shielded electric cables for transmitting millivolt
signals from the analogue amplifiers of the upper and lower sensing
` 25 heads 11 and 12. The control conduit 312 may contain conductors which
are resFectively energized to represent the angular position of filter
:
wheel 210, and may also contain a conductor for controlling the indexing
movement of the filter wheel as will be explained in detail in connection
with Fig. 6.
-20 -
)82g45
Referring tO the optical analyzer uniI 300 of Fig. 5, the front
panel of the unit has been diagrammatically indicated at 320 as being
provided with a series of lamps 321-327 for indicating the angular posi-
tion of the filter wheel 210 within the upper sensing head 11. I~le
lamps 321-327 have been numbered 1 through 7 in correspondence with
the seven positions of the filter wheel, and the color of the lamps, for
example, may be selected so as to signify the characteristics of the
Slters located in the openings of ~he filter wheel such as those indicated
at 211-217.
In order to provide a visual indication of the amplitude of the
millivolt signals supplied from the sensing heads 11 and 12, a suitable
meter i~ indicated at 330 and a selector switch is indicated at 331 for
selecti~ely supplying to the me~er the analogue signal from the upper
sensing head 11 or from the lower sensing head 12. A switch 332 is
indicated for controlling thesupply of conventional alternating current
power to the meter, and a second switch 333 is indicated for controlling
the supply of energizing power for the lamps 321-327. Another switch
334 may be momentarily actuated so as to index the filter wheel 210 to a
desired station. The switches 331-334 may, of course, take any desir-
ed form, and have merely been indicated diagrammatically in Fig. S.
Referring to Fig. 6, various of the components previousay refer-
red to have been indicated by electrical sym~ols, and for convenience of
c~rrelation of Fig. 6 with Figs. 1 through 5, the same reference charac-
ters have been urilized. In particular, Fig. 6 shows symbolically a
light source 201, associated photocells 203 and 260, fiker wheel drive
motor 209, control solendd 2~0, and permanent magnet 2~3 which rota~es
with the filter wheel 210 so as to represent the angular position of the
filter wheel. Also shown in Fi~. 6, are the four heaters 2S1 associated
' : --~1--
;",
~082945
with photocell 203, and the four heaters 261 associated with the photocell
260. Further, lamps 321-327, millivoltmeter 330 and switches 331-
. 334 of the optical analyzer unit 300 have been symbolically indicated in
. Fig. 6.
Referring first to the components associated with the upper sensing
head 11, there is illustrated in the upper left part of Fig. 6 a diode 340
connected across solenoid 240. For diagrammatic purposes, permanent
~' magnet 243 is shown arranged between two series of reed switches 341-
347 and 351-357. A further reed switch 358 is indicated for actuation
~r; 10 in the number 1 position of the filter wheel 210 along with switches 341
and 351. The conductors 359 and 360 associated with switch 358 may
be connected with the optical analyzer unit 300, and may be connected
via the optical analyzer unit 300 with a rernote computer, where the
illustrated apparatus forms part of a computer control system for con-
lS trolling the associated paper machinery.
The reed switches 341-347 are shown as being associated with
an operational amplifier 361, so that switches 341-347 serve to select
the desired val~ue of feed back resistance for the amplifier in each posi-
tion of the filter wheel 210. Thus, switches 341-347 served tO selective-
Iy connect in parallel with resistance 370, additional resistance values
371-377, respectively, for adjusting the total resistance between the
input and output terminals of the amplifier 361. Thus, in the number
1 position of the filter wheel, permanent magnet 243 is in a position to
actuate switch 341, and connect resistance value 371 in parallel with
resistor 370. As will hereinafter be explained, resistance means 371-
. .,
377 may include variable resistors for adjustment so as to provide the
desired gain of amplifier 361 in the respective filter positions, or fixed
resistance values may be inserted as indicated, once the desired values
-22 -
,.,,~
~08:~945
have been determined for a given filter whe_l. As indicated in Fig. 6,
the output of amplifier 361 may be transmittined by means of shielded
s cables 381 and 382. These cables form part of the overall cable indi-
cated at 51 in Fig. 5 leading from the upper sensing head 11 to the power
supply unit 301.
Also forming part of the cable 51 would be the conductors such
as indicated at 383 from the respective reed switches 351-357. These
conductors such as 383 would connect with respecffve condu~tors 391-397
of cable 52 leading from the ~ower supply 301 to the lower sensing headl2.
Included as part of the power supply unit 301 would be compo-
nents such as relay actuadng coil 401, associated normally open contact
402, and resistors 403 and 404 shown at the upper left in Fig. 6. Further,
~-~ the power supply would include an adjustable direct current lamp power
supply component 410 for supplying a precisely adjusted or controlled
electrical energizadon for light source 201. Further, of course, the
power supply would supply the required direct current operating potentials
for the upp~ sensing head as indicated in Fig. 6.
The lower left section of Fig. 6 illustrates the electrical com-
~ ponents of the lower sensing head 12,.In the lower sensing head, conduc-
;;~ 20 tors 391-397 control energization of the opOEatdng coils of respective
,~ relays K1 through K7. With the permanent magnet 243 in the number 1
posidon, reed switch 351 is closed, and~operating coll 420 of relay K1
, ,,;
is energized closing the associated relay contact 421. ~he remaining
relays K2 through K7 are deenergized, so that the respective associated
`' 25 contacts 422-427 remain open. The contacts 421-427 se ve to cont~l
path
the resistance in the feedback/of operational amplifier 429 in conjunc-
- tion with resistor 430 and resistance means 431-437. As explained in
reference to the upper sensing head, resistance means 431-437 may
.,
-23 -
.. .
. . .
~08Z945
include adjustable resistors, or fixed resistors as shown selected to
provide the desired gain of amplifier 429 for the respective positions
of the filter wheel 210. The shielded cables 441 and 442 from the out-
put of ampliHer 429 connect with power supply unit 301 as part of cable
52. The outputs from the amplifiers 361 and 429 are conducted from the
,.,
. ~ power supply unit 301 to the optical analyzer unit 300 via signal conduit
311, and within the optical analyzer unit connect with respective terr.ninals
of the selector switch 331 as indicated at the lower part of Fig. 6.
Thus, io the upper position of the selector 331, the output of amplifier
361 is connected with the meter 330, while in the lower position of
selector 331, the output of amplifier 429 is supplied to the meter 330.
Of course, the optical analyzer 3~0 may further include analogue to
digital converters for converting the outputsof the amplifiers 361 and 429
to digital form for transmission to a remole computer, for example.
It will be apparent to those skilled in the art that the remote computer
could be programmed to control the sequendal actuation of relay 401
during each incremenr of scanning movement of the monitoring device
10 so as to obtain readings from each desired sampling region of the
web 14 for each of the seven positions of the filter wheel 210. The remote
2~ computer would then be ul a position to correspondingly determine the av-
erage optical characterictics of a given length section of the paper web
14, for example, and control suitable inputs to the paper machine so
a~ to maintain desired optical characteristics of the paper being manu-
factured~ Alternatively, of course, the arrangement of Figs. 1-6 can
be utilized simply to take readings from the me~er 330 for each filter
~: wheel position during scanning of the web, so as to obtain readings re-
. .,
aecting the optical charac~eristics of ~he length sections of the web
so scanned. Still further, of course, the circuitry of Figs. ~ and 6 can
. -24 -
~08Z945
be udlized either with the monitoring device located in a fixed position -
relative to the width of the web (by means of a C-type frame), or with
the device off-iine fro~ the paper machine, so as to obtain desired
readings from the meter 330 for each position of the filter wheel 210
5 during optical excitation of a single sheet sample of the web held in a
sample holder so as to be disposed essendally as indicated for the web
~ 14 in Fig. 3.
,' .
- .
:
; ' ' .
. ~
'~
:.
.'~ ' .
,
,. ~
. -25 ^
,
108Z945
i;
: '
Computer Interfacing
In preparing the moni~oring device for on^line operation on
the paper machine, the zero to 140 mi~livolt DC sig~als from the
l;
sensing heads will be supplied to respective emf-to-current converters
5 of component 501, Fig. ~. As an example, Rochester Instrument
Systems Model SC-1304 emf-to-current converters may be used. Such
a converter will provide an output of 10 to 50 m~ amperesDc suitable
for driving an analog to digital converter at the computer. The emf-
- to-current converters will provide an isolated input and outFut so that
groundi~g w;ll not be a problem.
The c~nverters of component 501, will be housedwi~optic
analyzer 300, Fig. 5, and will connect with respective points thirty
one of Groups five hundred and six hundred (not shown) at the co~rol
computer analog signal input via conductors such as indicated at 502
and 503 in Fig. 6.
Conductors S05 and 506, Fig. 6, associated with filter
wheel inde~ing solenoid 240, Figs. 3 and 6, may extend within coukol
conduit 312, Fig. 5, and connect with the control computer output tenn-
iDals at a location designated Groupforty two hundred and six, point
; 20 nineteen (not shown). (Switch 334 should remain open (off) during
computer operatioll of Figs. 1-6.)
Conduclars 3~9 and 360, Fig. 6, may connect with an input
- of the con~rol computer at a location designated Group fourteen hundred,
~ point twenty-three (not shown).
: -
`'`
-26-
''
~Ol~Z945
DISCUSSION OF AN EARLIE~ PROTOTYPE
SY~TEM
Structure and Operation of a Prototype Optical Monitorina Device
:-~
A pro~o~rpe optical monitoring device was first constructed so as -
S to test the feasibility of the concep~s of Ihe present invention. As a re-
sult of the experimental work with the proto~ype system, a preferred
sys~em has been designed and wiU hereinafter be described in greater
detail. Since the operation of the prototype system is somewhat differ-
ent from that of the later designed system, a description of the proto-
10 type system will serve to iUustrate alternative features and an alterna-
tive method of operation in accordance with the present invention.
In the original setting up of the prototype system~ the upper
'A~` ant lower sensing heads should be brought into proper alig~ment a~d
spacing. The spacing should be just under 1/4-inch between the case
". 15 110 and the surface of the diffusing glass of window 13~. (In the proto-
~pe unit, there were no additional parts between the case 110 and
, window 135 such as the shoe plate 122 shown in Fig. 3.) The lower
; sensing head should be moved lateraUy in a~U directions to locate the
, point where the maximum reading occurs from photocell 260 as well as
' 20 the point of least sensitivity to relative movement of the upper and
;, . .
. lower sensing heads. In an inidal calibra~ion of the prototype monitor-
. ~ ing device, potentiometers are included as part of the resistance means
., .
t^! 371-3~7 ~nd 431-437 and are adjusted for the respective positions of the
filter wheel 21'~ ~o give the correct readings for the reflectance and
2S transmi~tance of the diffusing glass 135 (in the absence any paper sample
between the upper and lower sensing heads). The values which were
~, used in this ini~ial calibration are indicative of percentage absolute reflec-
, . ~
.~ tance and transmittance on a scale of 100, and are as follows:
.~ ,.
~7
~ 1082g45
Table 1
Table Showing Exemplary Calibration
for the Prototype System-
Diffusing Glass Reflectance and Transmittance
S Values With No Paper Specimen Present
Filter Wheel Reflectance Transmittance
Position Value, RSD Value, TSD
No. (Millivolts) (Millivolts)
35.4 54.0
2 35.0 56.1
3 34.4 '56.9
: 4 34.6 56.6
34.7 56.4
6 34. 5 56. 6
7 34.8 0.6 *
The r~ings in millivolts can be converted to other desired units
. .
by comparing ~he readings in millivolts for a given paper specimen
with the readings obtained with a standard laboratory instrument, measur-
ing the reflectance of the specimen with the laboratory instrument while
backing the paper sheet with a piece of Lucalug and a black body. By
measuring the reflectance of the sin~le sheet backed with a black body
(no fluorescence), the value of transmittance for the specimen can be
calculated and this calculaled v~lue utilized for calibrating the lower
sensing head. lf the fluorescent component is included in the labora-
. .
25 tory inst~ument, and if fluorescence is involved, the fluorescsnce com-
ponent can be determined by means of a standard reflection meter, and
~he fluorescent component can then be subl:racted from the measured
data before making the calculation af transmittance.
s,
The labora~ory testing of the prototype system confirmed that a
30 monitoring device sllch as illustrated in Figs. 1-4 should have a po~en-
dal accuracy equal ~o that of comparable off-line ~esters provided certain
.~1
we~ scanning requirements are met.
:;..
* The transmittanc value of the l~o. 7 filter position is not
needed~ and consequentlv a low amplificarion of this signal was
35 arbitrarily selected.
:,~
-28-
.
, 101~2945
Laboratory tests wOEe run on color srandard samples of the
grades and colors usually run on the paper machine s~own in Figs. 1
and
and 2. In addition, a variety of opaques,/a variety of colored ~0 pound
and 70 pound offsets were included in the ;ests. A four centimeter
S diame;er circle was scribed on each sample to insure that all tests
would be done within the same 12 square centimeter section of the
sample. Values of RoJ Roo, and TAPPI opacity measurements were
made on the available standard laboratory instruments. All test were
made on the felt side of tiTe sample with the grain in the standard
direction. For Roo measurements, the samples were backed by piles
of tabs cut from the edge of the same sheet of paper. In addinon to
the TAPPI opacity measured on the standard opacimeter, TAPPI opacity
was calculated via ~l~a-Munk theory from data ob~ained with a stan-
dard automatic color-brightness tester.
The ~ me paper samples were clamped into a holder which
held the sample under tension with the lower head of the monitoring
device bellying y8-inch tO 1/4-inch into the sheet. The grain of the
.~,,
;;- s heet was oriented parallel to the longitudinal axis of the upper s^nsing
. .
head (that is the machine direction of the sheet was in the same orienta-
20 tion as would-occur on the paper machine as indicated in Figs. 1 and 2).
The felt side was always up. Care was taken to make sure that the
" tested area was within the twelve square centime~er circle scribed on
the sample.
The transmittance and reflectance readings were taken ~rom a
digital vol~ me~er attached tO ~he OUtpUI terminals of ampliriers 361 and
~.~
429. Calibration data was ~akenoff the Lucalux ~ith no shee~ presen~.
~ .
Test values w~e ~aken on all fikers with ~he she~t in place. The trans-
mittance and reflectance values were keyed into a standard calculator
,. with the calibration data. The calculator was programmed to calculate
~ -29-
1082945
the color (in C.l.E. X, Y, Z, for example), fluorescent component,
brightness, TAPPI opacity and printing opacity tbased on Yc). By sup-
plying the basis weight, the computer could also be requested to calcu-
late s, the scattering coefficient (an index of the effect of pigment
5 efficiency and fiber surface area), and k, the absorption coefficient (an
index of the effectiveness of dyes in the sheet). The coefficicnts s and
lc are essentially independent of basis weight. Kube3~a-Munk theory is
the basis of the calculations used.
All of the samples were tested without changing the relative
10 position of the two se~sing heads. One set of data was obtained with
the heads in a variety of positions to determine the effect of geometric
variations.
Since fluorescence is not compatible with Kube~a-Munk
theory, the prototype system was carefully designed so that all data used
15 for Kubelka-Munk analyses have excludcd fluorescence. The pro~ocype
system measurcs fluorescence separately. A fluorescent contribution
is determined from the prototype data by subtracting the Z distribution
reflectance without fluorescence (filter wheel position No. 4) from the Z
distribution reflectance with fluorescenc~ (filter wheel position l~to. 7),
20 and multiplying by the appropriate factor.
An independent check on fluorescence measurements, a
modified brightness tester wasut1lized which had a filter wheel allowing
~ for standard brightness and Z distribution filters ~o be pUt in the reflect-
; ed beam. In addition, the filter wheel contained brightness and Z distri-
25 bution filters which had been modified by removing the ultraviolet absorb-
ing component of these filters. A special mount allows the operator to
put the appropriate ultraviolet absorbing filter in the incident beam.
Thus, measurements of brightness an~ C.I.E. Z tristimulus, with and with~ut
fluorescence, could be made. Fluorescent contributions were calculated
-30-
- 108Z94S
:.
by difference. Some measurements were made on single sheets with a
standard backing. Most of the samples were measured with an inffnite
pack of tabs The incident beam filter of the prototype's No. 7 positia
was such that it permitted about twice the standard quantity of ultra-
violet light to strike the specimen. Consequently, measurements of the
fluorescent contribution measured on the modified brightness tester
and the prototype system correlated well (correlation coefficient of
.992) but the modofied brightness tester value is only 0.528 as large
; as that measured by the prototype system, Calculations of prototype
:,
data now involve calculation of the fluorescent component by multiply-
ing the difference of filter positions No, 7 and No. 4 by 0 528.
Because only one fluorescent dye (Tinopal, a trademark) in all of
,~ the paper specimens was used, the fluorescent contribution needed to
be measured only once. The prototype data provides a basis for mea-
suring the fluorescent component Z. Measurements by an independent
laboratory showed that the paper specimens do not fluoresce signifi-
l,i "
~ cantly in the X (red) or Y distributions; therefore, fLuorescent contri-
,
butions need only be determined for the blue colored distributions. A
linear regression was run on the independent laboratory data which
demonstrated that the fluorescent component for X (blue) can be pre-
`, dicted by multiplying the fluorescent component for Z by 1.204. A re-
gression run on fluorescent data from the modified brightness tester
t~' shows that the fluorescent contribution for bnghtness can be calculated
;,:
by multiplying the fluorescent contribution for Z by 0. 864. In summa~r,
fluorescent contributions are calculated by the following formulas:
, .~
~;; Fz=0.528 (Z reflectance with fluorescer~e minus Z reflectance without
i~ fluorescence. )
Fx(blue)= 1.204 Fz Fgrightness = 0-864 Fz
3 1-
~,,
~08Z945
These fluorescent contributions are added to the respe~ive
calculated Roo values when calculating optical properties from protorype
data. The test results for fluorescent and~Dn-fluorescent papers a~ree
with values measured on the standard automatic color-~rightness tester.
i:,
.` ,, :
.,.~ .
.,
' '
. ~ .
,, .
-32 -
08~4S
:'
'
Discussion of the Results of Mechanical Life Testing of the Prototype
System and Design Features Selected for t~e Preferred Sys~em In Light
of Such Life Testing
. _ . .
The following details concerning the resul~s of life testing of
the prototype system are considered to reflect minor problems of con-
struction and operation whrh considered individually are readily correct-
ed for by those ski~led in the art. ~n order to minimize the burden ~
the total number of such minor problems, and thus to expedite practice
of the prototype syste~n, solutions to the various problems which were
encountered are briefly referred to.
The iilter wheel is advanced by a low torque stallable motor.
A timing belt links sprockets on the motor and the filter wheel shaft.
The original timing belt had a dacron core. The core of the original
belt broke in two places resulting in stretching and eventual loss of
teeth. Une~en rate of rotation of the filter wheel occurred due to bind-
ri; ing of the bek. Eventually, the plastic drive sprocket broke. Bothsprockets were replaced with stainless steel sprockets and the timing
belt was replaced with a belt containing a steel core. Installation of
!t the steel sproc~ets and steel core belt revealed that excessiue belt ten-
c. 20 sion could st~ll the motor. The motor mount holes were slotted allowing
the motor to pivot slightly around one mounting screw. Belt tension was
` adjusted by pivoting the motor. It is concluded that future models
. .
should include an idler wheel or some other means of adjusting the
tension of the timing belt.
Some problems were experienced with respect to indexing of
the filter wheel with the ratchet arm sticlcing on the too.h so that the rat-
che; arm does not clear the tooth when a command is given tO index the
filter wheel. The remedy has been to redlce the rou3hness ofthemating
-33-
~082945
surfaces by filing on the tooth, or smoothing the tooth with a stone.
In future models, the shapes and/or smoothness of the ratchet arm and
the teeth should be altered to minimize sticking. One solution would
be to provide the ratchet arm and the teeth with highly polished mating
surfaces.
The ratchet arm is lifted by a 24 volt d~rect current solenoid~
A~tOE some time, the plunger of the solenoid became magnetized and
would stick to the inside of the coil. This "hanging up" would p~event
the ratchet arm from catching the nexc tooth. A resistor was installed
in sefies with tb solenoid coil to reduce the strength of the magnetic
field. The plunger of the solenoid was coated with a special materiaL
The coated plunger worlced we11 for about three months before it, too,
magnetized enough to hang up. The solution adopted was to provide
;:i,.,
' the solenoid with a flat topped plunger which is stopped at the end of
j
15 its stroke by a bumper of rubber-like material.
The response of a photocell ~s somewhat temperature sensitive.
For thi9 reasons,it is necessary to keep the photocells at a constant
temperature. Ambient temperatures on the O-frame of the No. 6 papOE
(48C. )
;; machine indicated in Figs. 1 and 2 have ~een measured as high as 118F/
20 in the summer. The photocells in both heads a~e mounted in massive
metal blocks. Each Iretal block has four thermistor heaters mou~ted in
close proximi~y to the photocell. These thermistors have switching
;
temperatures of 55C,(~hat is about 130~). The intenrion of this de-
sign was to add enough heat to the instrument to hold the ;emperature
25 steady at about 35~. During bench studies, this temperature was
never reached due to the low capacity o. the heaters. At machine room
temperatures, however, ;he instrument temperature may reach 55C.
-34 -
~,
,
108Z945
During the bench studies, it was found that the heaters did
minimize temperature variations. The few degrees of temperature
variation that were obsOEved during normal operation usua)ly occurred
;~ slowly. Changes in instrument temperature affected the cutput signal
5 less than acticipated. Based Oll this experience in the ~boratory, the
maximum variation in head temperature should be less than 3~F per
hour. Temperature variations of this magnitude will not have a signi^
ficant effect on th~o output signal. Long term tempe~ature changes
; would be corrected for by the calibrations each time the head goes off
10 web.
In the laboratory, there was a minimum of dirt problems. On
the machine, however, Ihe hole could aLlow dirt to enter the upper head.
Up to a point, dirt on the lenses and filters will be correctet for by
the periodic calibration routine. Excessive dirt, however, will reduce
15 the sensitivity of the instrument and may even affect its accuracy. Peri-
odic cl~aning of the lenses and filters will be required. If dirt accumu-
lates too rapidly, it may be necessary to attach an air purge to the
upper head.
The lower head of the proto~pe system is completely sealed so
20 that no tirt pro~lem is anticipated inside the low~rhead. Because tte
Lucolux window is in contact ~th the sheet, friction will keep it clea~.
.,
. .
-35 -
::'
1082945
Most of the filters consisted of two or three component parts.
There have been some problems with dirt getting be~ween the components
of the filters.
The case on the lower head as well as the case on the UppOE
S head should allow most general maintenance and trouble-shooting to
be done without dismounting the head. A completely removable
case would be desirable. At a minimum access should be provided for
the following: (1) convenient light bulb c'nange, (available on the proto-
type), (2) cleaning of lenses, (available on the prototype), (3) cleaning
10 of the filters. (Access is presently available to one side of each ~ilter.
' The side which is most likely to collect dirt is not accessible in the
prototype.) (4) The amplifier. The amplifier is a standard plug-in
; module. In the event of a breakdown it could be replaced in seconds
if it iS accessible. Furthermore, it is necessary to remove the
15 amplifier to do any trou~le-shooting on the gain circuitry. (S) The
circuit board holding all of the galn control resistors. The choice of
gain circuitry is controlled by reed switches which are not accessible
on the prototype ~ithout a partial dlsassembly of the i~strume~.
Malfunctions of the reed switches, however, can easily be diagnosed
20 by removing the amplifier and taking resistance measurements on the
gain control circults. There is also the possibility of mechanical or
- electrical damage to a resistor or a potentiomecer mounted on this
circuit board. With proper access a damaged part could be replaced
in five to rwenty minutes. (6) The ptlotocell. With proper access,
25 the photocell could be replaced quickly and easily. (7) The heater.
~; The heater are adjacent to the phococell and are generallv just as
~; easily serviced. (8) Indexing mec'nanism. The present accessibility
3 6
1082945
~
to the ratchet teeth, rachet arm and solenoid is adequate but not very
convenient on the prototype. A certain amount of access to these parts
is needed to correct chronic indexing problems such as sticking and
?,~, "hanging up".
The filters are presently mounted in the filter wheel of the
prototype by spring clips. Most of the filters are compound filters
. . .
containing as many as four component pieces of glass. During laboratoDy
trials, increases intheoptical density of a filter were frequently observed
~ which could not be corrected by cleaning the surfaces of the filter. Upon
- 10 removing one of the filters, it was discovered that foreign material was
~ collecting between tb components ofthe compound filter. The use of lens
. . .
~ cleaning solution on the filters may have accelerated the problem if
.,
,~ capillary action drew foreign material between the components. A set of
gaskets and some type of threaded mount should be used to mount the
15 filters in such a way as to minimize forelgn material (including cleaning
solutions) from getting between the components of compound filters.
, ~ In mounting the prototype sensing heads on an O~frame, it is
necessary to bring the geometric alignment of the heads as close to their
optimum relationship as possible. The original intention was to set the
20 gap between the heads with the aid of a spacer; however, flexibility of the
`1 sheet metal case of the prototype upper sensing head prevented the use of
...
a spacer for setting the gap. Accordingly, the shoe plate 122 of the new
upper sensing head shown in Fig. 3 has been made of a thickness and
consequent rigidity so as to enab~e the use of a spacer gauge to set the
25 gap between the upper and lower heads. ~The gap is reduced by 1/16
, .
` inch to 3/16 inch because of the thickness of shoe plate 122. )
~ .
;''
'';'
- 1 -37 -
..,
1082945
The gap becween the heads is a most critical dimension as far
as calibration and reproducability is concerned. In the pro~otype it
was intended to calibrate relative to an average gap, thus correcting
the readings for variations in the gap from the average gap.
One of the criteria used in designing the prototype was minimum
head length in the machine direction. Unfortunately, ~he upper head
was turned 90 in order to give the prototype unit the same geometry as
the General Electric Brigh~ness Meter, Automatic Color-Brightness Tester,
and Hunterl~b Color Meter. In this new position, the prototype head
is 12 1/4 inches long in the machine direction plus 2 1/2 inches for
cable connectors. Redesign should be possible tO reduce ;he machine
direction dimension to about 8 inches and to relocate the position of the
cable connections.
The lining of the case for the upper head should be matte as
well as blacl; to prevent reflection o~ ambient 1ight within the case and a
pos6ible spuriou9 e~ect on ~he pho~ocell reading.
. . .
. .
,~, .
., .
~.,
--3 8--
108~94S
~.
-~' Conclusions from Mechanical Testing of the Prototype System
Following the correction of miscellaneous start up problems
- the prototyFe system was found to function well mechanically. As a
test of its durability, the prototype system was placed in continuous
5 operation for a period of over ten months and no serious mechanical
problems resulted except the failure of the solenoid. The solenoid
failure was expected and the replacement solenoid is of a design which
~ is expected to give a long sOEvie life. The light application of sili-
; cone lubricant spray to the indexing control ratchet arm and cooperatiing
10 teeth corrected a problem of malfunctioning of the filter wheel inde~ing
, mechanism (which occurred on two occasions during the ten months).
!,';,,~ The ~rototype system was not intended to be a low maintenance instru-ment; however, the experience during the durability test with the proto-
type in continuous operation indicates that the prototype system should
15 operate on a paper machine with an acceptably smaLl amount of down
time.
:. ,
<;.,,
, .
~, .
. ..i
. ~, . . .
~i
;::
~,...
'
..;
~i
~ -39 -
.,
~;
lO~Z5~45
DISCUSSION OF LABORATORY TESTING
OF FIGS. 3-6
LaboratorvOperation of ~he System of Figs. 3-6
In the prototype svstem, potentiometers are included as
5part of the resistadce means 371-377 and 431-437 and are adjusIed for
the respective positions of the filter wheel 210 to give desired values
such as given in the fcregoingTable 1. In the preferred system of Figs.
3-6, Ihese potentiometers for adjusting amplifier gain are omitted
- and are replaced with fixed resistors 371-377 and 431-437 selected to
10 give scale readings from meter 330 in ~he respective filrer wheel
positions which are well above the values given in the preceding
TableL This is intended to improve the stabili~y and increase the
sensitivity of measurement.
In calculating optical parameters from measurements relative
. 15 to various samples, values were first established for the reflectance
RD of the diffuser 135, Fig. 3, in the absence of a paper specimen,
; for eash filter wheel position. Initially ca~culated values for RD were
; used in a first computation of optical values, and ~hen the values of
RD were adjusted slightly ~o give the best agreement with the correspond-
20 ing optical measurements by means of the standard automatic color-
.. brightness tester. The following table shows the reflectance values
which were estabaished for certain laboratory testing of the system of
:.,,
Figs. 3-6.
;:
. .
~,it
-4~-
8294S
` Table 2
Table Showing Reflec~ance
of the DiffusiDg Glass With
No Paper Specimen Present
S in a Laboratory Test of the
. System of Figs. 1-6
. . .
Filter Wheel Sym~ol Diffusing Glass Reflec-
.- Position No. tance Value
,,, .... _
~: 1 RDl 0. 349
. 10 2 RD2 O. 347
; 3 RD3 0. 355
4 RD4 0.349
RD5 O. 354
6 RD6 O. 354
15 . 7 RD7 - O. 349
The transmirtance of the diffusing glass 135 ~eed n~t be
known since the ratio of the transmittance of the diffusing glass and
paper (in series) to the transmittance of the di~using glass is employed
in calculating the desired optical parameters.
,'~
,,. ~ . .
,. ?
. ~ .
,~ .
' 1
i,
.....
. . .
.',
.
-41 -
~82945
A computer program was developed to process the data
collected during laboratory operation of the monitoring device 10 as we
as to compare the calculated reflectance value R and the calculated
CD
fïuorescen~ components with the data collected withthe standard auto-
5 matic color-~rightness tester. A listing of the symbols employed in a
symbolic state~nt of the computer program in the Fortran computer
language utilized in this laboratory study is set forth in Table 3 on the
:: following pages.
,i
.;
;~3
,,,1,
.. . f
.
.` .,
.,.; .
, ,i .
r
:
'`
. -42 -
.
:
108294S
Table 3
Listing of Symbols (Including Input Data Symbols and
Output Data Symbols With a Brieî Indication of Their
Significance).
S Input Data Symbols
RSD OMOD scale reading for reflectance with no
paper specimen iQ place. (Filters 1 through
RSP OMOD scale reading for reflectance with
paper specimen in ~idcn.(Filters 1 through
6.)
TSD OMOD scate reading'for transmittance with
no paper specimen in place. (Filters 1
through 6. )
TSP OMOl~ scale reading for transmittance wilh
paper specimen in position. (Filters 1
through 6. )
RSD7 OMOD scale reading for reflectance with no
' specimen in place. (No. 7 filter~
RSP7 OMOD scale reading ~or reflectance with
paper specimen in position. (No. 7 filter. )
AR~oFC ACBT reflectance including the fluorescent
component.
AFC AC8T fluorescent component.
RSD4 OMO~ scale reading for reflectance with no
paper speclmen in place. (No. 4 ~lter. )
RSP4 OMOD scale reading for reflectance with
paper specimen in position. (No. 4 filter.)
GC Grade Correction as determined by the
~ 30 difference berween R FC and AR FC
fl' for each sample and ea~ch filter.
:;~
~',
. .
.'
,''
~ -43-
` 10~Z94S
.
Table 3 - Listing of Svmbols-continued
Output Data Symbols
R Reflectance of a single sheet backed with a
blac~ body (no fluorescence) as calculated
from OMOD data.
T Transmittance of a single sheet backed with
a black body (no fluorescence) as calculated
from OMOD data.
R Reflectance of an opaque pad (no fluores-
oo cence3 as ca~culated from OMOD data.
.,
R FC Reaectance of an opaque pad (including
j ~ fluorescence) as calculated from OMOD data.
. AR FC Reflectance of an opaque pad (including
fluorescence) ACBT.
DIFF Difference between R FC and AR FC.
oo oo
FC Fluorescent component OMOD.
?i~?l AFC Fluorescent component ACBT.
GC 5radeCorrection as dete.l~ined by ~he
, difference between R FC and AR FC for
r, ' 20 each sample and eac~.filter.
',,
~.
';.`
,. .
."
.'
:'
...
'',,'
., ~
' :
:,
:'
-44 -
..
.~
. .,
\ ~0~294s
Table 3-~isting of Symbols-continued
Additional Symbols (lJsed LQ the Computation
of the Output Data from
tne Input Data) -
RK Reflectance correction factor
:; (assigned a value of 1.000 for
. laboratory operation).
j TK Transmittance correction factor
~ (assigned a value of 1.000 for
i 10 laboratory operation).
;.
:/ RD Value representing the absolute
. reflectance of the di~user (on a
: scale of zero to 1.000) as ad3usted
to give best agreement with opti-
:~ 15 cal measurements by means of
. the standard automatic color-bright-
ness tester. ~The values given
~, in Table 2 are used for laboratory
operation~
:' 20 RP~ Re31ectance of paper specimen
,i when backed with the diiguser, as
:1 calc~llated from current values of
;i RK, RD, RSD, and RSP.
Tl~ Transmittance o~ paper specimen
2C and diffuser in series, as calcu-
~3 " lated from current values of TK,
TSD, and TSP.
,
v,
"
",:,
.
.`.
''
;
''
,
~ .
` -45-
~'
,
1082945
In the foregoing listing of symbols, the letters of the symbol
OMOD are taken from the phrase on-machine optical device; however,
this particular section of the specification refers to a system essentially
conforming to the sys~em of Figs. 3-6 operated tO measure optical pro-
perties of indi~Tidual paper sheets under laboratory conditions. ~rhe
laboratory work here reported was with an earlier version of the moni-
toring device designed for on-rnachine operation, prior to adoption of
a thickened shoe plate 122. The standard spacing between the upper and
- lower sensing heads for the eariier vOEsion was 1/4 inch, rather than
3/16 inch as with the final version of on-machine device as specifically
shown in Fig. ~ The OMOD scale readings are obtained from the me~er
330, Figs. 5 and 6, with the filter wheel 210, Figs. 3 and 4, inthe
respective~sitions IO activate the respective filters 281-286 (indicated
as "i~ilters 1 through 6" in the p~eceding listing) and to activate filters
lS 287 and 288 (indicated as "No. 7 filter" in the listing), and with switch
331, Fig. S, in its upper position to measure reflectance, and in its
lower position to measure transmittance. As to reflectance measure-
ments, the cavity 145 is considered to form essentially a black body
backing for the diffusing glass 13S.
The sym~ol"ACBT" in the foregoing listing of symbols is used
to designate a measurement made on the standard commerically avail-
able automatic color-brightness tester. The brightness measurement
obtained from the ACBT represents a value accepted as standard in the
U.S. Paper industry. A fur~her appreciation of the importance of the
fact that the OMOD measurements can closely con~orm to this industry
standard is gained from a consideration of ;he article ~y L.R. Dearth
et al "A ~tudy of Photoelectric Instrument3 for ~he ~Ieasuremenr of Color
; Reflectance, and Transmittance, XVI. Automatic Color-Brightness Tester",
Tappi, The Journ~l of the Technical
:
-46-
~082945
Association of the Pulp and Paper Industry, Vol. 50, No. 2, Fe~ruary
1967, pages 5LA through 58A. As explained in this article, the ACBT
is photometrically accurate, and the spectral response is correct for
the measurement of both color and standard brightness. The spectral
response of the AC8T very nearly matches the theoretical CIE functions
as indicated by the special technique for determining spectral response.
This involves the determination of the tristimulus values for deeply
saturated colored glass filters a very rigorous checl~ on the spectral
response, especially when it is noted that colored papersare less
~t: 10 saturated.
,.,
The symbols in the foregoing Listing of SyTnbols which as
..,
shown include lower case characters may also be written exclusively
with capital letters. Ths form of the symbols is convenient for com-
puter printout. The alternate forms of these symbols are as follows:
ARooFC or AROOFC; Ro or RO; R ~ ROOa~l R FC or ROOFC .
~,
'
''''
"''
, .;
, .
.
-47 -
- ~082945
Table 4
Symbolic Statement of the ComputRr Program
(Used for Processing the Data Obtained During the
: Laboratory Operation of the System of Figs. 3-6)
6PS FORTRAN D COMPILER
C OMOD (220)
:~ S. 0001 WRITE (6, 2001)
S. 0002 2001 FORMAT (lH, 'SAMPLE', 6X, '
RD', 12X, 'T', 12X, 'ROO.', 9X,
'ROOFC', 9X, l'AROOFC', 10X,
'DIFF', 7X, 'FC', 7X, 'AFC', 7X,
'GC', /)
S. 0003 READ (5, 1000~; RK, TK, RDl,
RD2, RD3, RD4, RD5, RD6
S. 0004 102 M=O
.~ S.0005 READ (5,1000~ RSD4, RSP4
S. 0006 1000 FORMAT (lOF8. 0)
S.0007 100 READ (5,1001) LA, IN, ID, RSD,
RSP, TSD, TSP, RSD7, RSP7,
AROOFC, AFC, R
., S. 0008 1001 FORMAT (I2, I2, A4, 9F8, 0)
S. 0009 GO TO (11, 12, 13, 14, 15, 16),. IN
S. 0010 11 RD=~Dl
S.0011 GO TO V
S. 0012 12 RD=~D2
S.0013 GO TO 17
~ S. 0014 13 RD--~D3
,'`! S.0015 GO TO 17
S. 0016 14 RD=RD4
. 30 S.0017 GO TO 17
S. 0018 15 RD=RD5
` S.0019 GO TO 17
-48-
,,
~-` 101~Z945
. ~
S. 0020 16 RD=RD6
S. 0021 17 RPD=(~RD~SP*RK)/RSD)
S. 0022 RPD4=l~D4~RSP4*RKtRSD4
S. 0023 TPDOTD=(TSP~TK)/TSD
S. 0024 RO=(RPD-~RD*~TPDOTD~*2)))/(1. -
(R~* TPDOTC)~2)
' S. 002S T=tTPDOTD*(l. -~RD*RPD)))/(l. -
(~D"TPDOTD)**2)
S. 0026 A=((l.+~RO**2))~ *2))~RO
S.0027 ROO=~A/2.)-SQR~(((A/2.)~*2)-1.)]
S. 0028 RPD7=~D4 *3~SP7*RK/RSD7
S. 0029 IF (IN-2)1, 2, 3
S. 0030 3 GO TO (7, 7, 7, 4, 7, 7), IN
S. 0031 1 FC={RPD7-RPD4)*. 450
; 15 S.0032 GO TO 6
S. 0033 2 FC=(RPD7-~PD4)*. 570
S.0034 GO TO 6
S. 0035 4 FC=(RPD7-RE~C)4)~. 510
. S.0036 6 ROOFC=ROO+FC
S.0037 GO TO 30
.~ S.0038 7 ROOFC=ROO
S.0039 FC=0. 0
S. 0040 30 IF (IA -2)18, 19, 19
, .,
S. 00~1 18 ROOFC--ROOFC+R
S.0042 GO TO 20
,:
S. 0043 19 ROOFC--ROOFC-l
;~ S. 0044 20 DIFF=ROOFC-AROOFC
- S.0045 GO TO (21,22), L~
-49-
-' ~082945
. . .
S. 0046 21 WRITE (6, 2000)ID,RO,T,ROO,
ROOFC, AROOFC,DI~F,FC,A~C,R
S.0047 2000 FORMAT (IH A4,7X,2(FB.6,4X),4
(F 10. 6, 4X), 2(F5. 4, 4X), '+', F4. 3)
S.00~8 TO TO 23
- S.0049 22 WRITE (6,2002)ID,R~,T,ROO,
ROOFC, AROOFC, DE; F, FC, AFC, R
S. 0050 2002 FORMAT (IH, A47X, 2~F8. 6, 4X), 4
(F 10. 6, 4X), 2(F 5. 4, 4X), ' -', F4. 3
S. 0051 23 M=~l
S. 0052 IF (M-6) 100,102,102
S.0053 END
; SIZE OF COMMON OOOOO
PROGRAM 01930
; 15 END O~ COMPILATION MAIN
In the foregoing Table 4, the symbols repre-
senting basic mathematicl operations were as follows:
..1
ODeration Symbol Example
., ~ .
~ Addition + A~B
:,
Subtraction - A-~
Multiplication ~ A*B
Division / A/B
E2ponentiation ~* A*~3(A )
:,
;~ Equality = A~
.
~, i
,...
, ~
',.'~
i;"
, .
: ~.
--5 0--
;:5,.
, .
;.~
lO~Z945
To indicate more concretely the calculations which a~e
pcrformed, the followi~g Table 5 will illustratc exemplary input and
output data for a given sample. The meaning of the various symbols
'; will be apparent from the listin_ of the symbol3 of Table 3:
.
S Ta~le 5 - Table Showing Exemplary
Input and Output Data for a
Given SamDIe
Sample No. 1, white Nekoosa Offset-60 pound paper,
specimen A RK=1.000, TK=1.000
10~ilter Wheel _
; Position No.
Input Data 1 2 3 4 5 6
RD 0.349 0.347 0.3Si 0.349 0.354 0.354
RSD 0.515 0.529 0.583 0. 636 0.525 0.596
, 15 RSP 1. 161 1. 187 1.339 1.422 1. 191 1.357
TSD 1.422 1.625 1.627 1.702 1. 625 1.546
,~, .
TSP 0. 236 0.256 0.354 0. 277 0.335 0.326
, RSD7 0.568 0.568 0.568 0.568 0.568 0.568
:::
RSP7 1.381 1.381 1.381 1.381 1.381 1.381
20AROOFC 0.837 0.829 0 847 0.830 0 839 0.844
AFC _ _ 0.034 0.034 0.0 0. 036 0.0 0. 0
RSD4 0.636 0.636 0.636 0.636 0.636 0.636
RSP4 1.422 1.422 1.422 1.422 1. 422 1.422
... ~ . .
GC -~.006 ~. 014 -0.021 -0.007 -0.009 -0. 01~
-.~
r'
: ''
51
108Z945
;T~Tr
.,, ~o 1~ u~ ~x x~ 1 C~
o o o o o ~ o o
~' o ~
e~ æ~ 8 0 ~ ~ ~
~' 0, ~ . _ i__
~ El ~ o~ 8 1 =~
e~, I o~o'~o~ ~
1~81~~81 ;3~~
,~ ~ ''~ - - - I ' -
,. Y ~ ~ =~1~` 0~ ~`O 8 8~ 3 8
h
,~i;
" -52 -
1082945
.
In the foregoing table showing exemplary input and output data,
the input and output data symbols have been shown as they are actually
printed out by the co~uter with all letters capitalized. In the text,
certain of the input and output data symbols are shown in a more conven-
5 tional manner with subscripts since the symbols are more farn~ r insuch form.
The data such as exemplified in Table S are based on a single
determination for each specimen. The "grade correction" GC is based
on the average d ference between RooFC and AR FC for two specimens,
10 specimens A and B.
The data as exemplified in Table 5 show that~here is generally
good agreement between the calculated RooFC and ARooFC values. The
spread in values for the duplicate specimens (A and B) is good with the
exception of several samples. Some ~lifflcul~y was experienced in
lS positioning the spec~men on the monitoring device 10 to give reproducible
resul~s. The d~Eicul~ should be minimized when the unit is placed
; "on-machine". The grade correction GC takes this discrepancy into
consideration so the correction should be established "on-machine".
The RD v~lues shown in Table S were punched into the first
;~ 20 data card along with the values for RK and TK for input to the computer
in advanc~ of a desired computation. The factoris RK and TK were
included as factors in the computations so that the transmittance and
reflectance values could be adjusted independently, if desired. In this
evaluation, RK and TK wOEe left at 1.000. (Calculated values for RD
25 were used in a first computer run and then the values were ad3usted
slightly to give the best agreemen. with the standard au;omatic color-
brightness tester. Th_ values for RD shown in Tabie ~ are the sligh~ly
- -53 -
i
, .,
108Z945
adjusted values utili~ed in obtaining the data discussed in this section of
the specification. )
A second set of data for the same fourteen samples W2S
co11ected using the monitoring device in the same condition as for the
;~ collection of the data previously given. All of the variables were left
the same to see how closely the datacal}lbe reproduced for the identical
specimens. The agreement was quite good except for samples 8 a~ 14.
It appears that theE~er may not have been lying flat in one or the other
tests. The grade correction GC on some of the grades was changed and
10 the second set ~ data was again calclllated for samples l, 2, 4, S, 6, 8
and 14. This improved the agreement between the monitoring device and
the standard automaticalor-~rightness tester.
The reflectance head of the monitoring device was then lower-
ed 0.025 inch and another set of data was collected for the same seven
15 samples. The same ACBT data was used. The data show that lowering
the reflectance head reduces the reflectance while transmittance remains
essentia~ly unchanged. The effects are not as large as was expected
,,.
and could be corrected through adjustment of RK; however, the variables
, RK, TK and GC were again held constant.
The reflecta~ce head was then raised to a spacing of 0.050
inch (0.025 inch above the ~ormal position for these tests), and another
;~?i;~ set of data was collected for the same seven sa~nples. The effects
~I were larger than when the reflectance head ll was lowe~ed. Again, an
..~i
~ ad~ustment of RK would improve the agreement.
.,,
!
.,,
" -54 -
1082945
It was concluded from these test results that a change of
plus or minus Q 025 inch from "normal position" is larger than can
~e tolerated. An estimale of a resonable tolerance, based on this and
earlier worlc, would be plus or minus 0.010 inch from "normal position".
S All of the variables used Ln calculating the data for samples
1, 2, 4, 5, 6, 8 and 14, after the in~ change in the grade correction
GC, werebeld the same to determine the effects of changing the reflec-
tance head position. The sars input data for the case of the reflectarce
head being raised 0.02S inch were processed again but with RK e~ual to
0.975 instead of 1.000. This reduces the reflectance value to the
proper level. The data o~tained in this way show good agreement
between the monitoring device and the standard automatic color-bright-
~ess tester. App~rendy the factor RK can be used quite effectively in
ad~usting for some variation in the geometric relationship of the upper
;15 and lower sensitlg heads. It would be preferred, of coursP, to maintain proper alig~ne~t and spacing.
A second set of samples were evaluated after returning the
re~ectance head tO its normal spacing from the transmittance head.
,.. .
8eIore calculating new output data, the computer program of Table 4 was
corrected i~ statements S. 0022 and S. 0028 by changing RD to RD4. The
; corrected computer program has been shown herein since the error in
the previously referred to data was insignificant in most cases. Thus
\ r.j
;' with the corrected computer program, the input data for the second set
of samples were processed. The values RK and TK were set to 1. 000
2S and the same grade correctionswOEe used as for samples 1, 2, 4, ~, 6,
8 and 14 previo~lsly referred to.
-55 -
10~2945
Conclusions drawn from all of the data are that the grade
correction GC will handle errors resulting from less than ideal
characteristics of the monitoring device 10 such as the relatively wide
bandwidth of light transmitted in the various filter positions in compari-
5 son to the requirements of KubeLka-Munk theory and the fact that this
theory applies strictly only to diffuse light rather than collimated light as
actually employed in the illustrated monitoring device 10. This correction
must be established "on~chine". Use of the diffusing glass 135 to cali-
brate the monitoring device 10 will handle changes in light level, photo-
10 cell sensitivity and amplifier gain. The reflectances RD of the diffusingglass 135 for the various filters as established in the present work are
set forth in the previous Table 2 entitled "Table Showing Reile~tance of
,1 the Diffusing Glass With No Paper Specirnen Present in a Laboratory Test
of the System of Figs. 1-6".
As previously mentioned, the transmittance of the diffusing
glass 135 need not be known as the ratio of the transmittance of the
diffusing glass and paper (in series), identified by the sym~ol TSP, to
the transmittance of the diffusing glass 135, identified by the symbol TSD,
is employed as will be apparent from ~he egplanation of the cakulations
;;
~; 20 employed set forth hereina*er.
~,
The fluorescent component is handled through the difference
, in reflectance as measured with the number 4 and the nurn~er7 filters
,' (RPD7 minus RPD4). The factors used in the subject computations, forfilters number 1, 2 and 4, are 0.500, 0.600 and 0.550 respectively.
.~ 25 This means of determining the fluorescent contribution FC appears to be
succe3sful.
-56-
:'
108Z945
The factor RK whereby the reflectance can be adjusted
to account for misalignment or incorrect spacing seems to function ~etter
. than was e~pected.
:~ The IOllOwing examples will serve to explain the calcu-
S la~ions of the output data for the different filter positions in greater detail.
.,
Table 6- Table Showing
E2emplary Calculation of Paper
Optical Parameters
ion of Ro, T, Roo.FC and RooFC from OMOD
, 10 da~a with the No. 1 filter in positio~
Input: RSDl, RSPl, TSDl, TSPl, RSD7, RSP7. TK,
RK,RSD4, RSP4, RDl, RD4, and GCl
Calculation:
RPDlz(RDlxRSPlxRK)/RSDl
^ 15 RPD4=(RD4xRSP4xRK)/RSD4
RPD7=(RD4gRSP7xRK)/RSD7
TPD/ID=(TSPLxTK)/TSDl
. ~ Ro=~RPD l -(RD l(TPD,tTD~ )]/[ l -(RD l~TPD~I~2)~
T={(TPD,~TDX~(RD lxRPD 1))]/[1 -(RD lCTPDfI .D)2)]
'~ 20 A=(1~o2 - T )/Ro
R =(A/2)-1 (A/2)2 - 1
oo
FC=0.500 (RPD7 - RPD4)
ooFC ROO+FC~GC1
Calculation of Ro,T,Roo, FC and RooFC from OMOD
data with the No. 2 filter in position
Input: RSD2, RSP2, TSD2, TSP2, RSD7, RSP7, TK,
- RK,. RSD4, RSP~, RD2 and GC2.
-57 -
- 108Z945
. '
Calculation:
RPD2=(RD2xRSP2~RK)/RSD2
, . .
.~ RPD4=~RD4xRSP4xRK)/RSD4
~. RPD7=~RD4xRSP7xRK)/RSD7
: S TPD/ID=~SP2XIK),~TSD2
Ro~RPD2 - (RD2~TPD~ )3/~1-(RD2(TPD~I~ )3
T=~¢TPD/TD)(l-(RD2X~PD2))]/[1-(RD2~ ID)2)]
i A=(l ~ Ro . -T )/Ro
=~A/2) - (A12) - 1
FC=0.600(RPD7 - RPD4)
RooFc~ oo~c~Gc2
Calculation of Ro,T,Roo, FC and RooFC from OMOD
3~ data wlth the No. 3 filter in position
Input: RSD3, RSP3, TSD3, TSP3,TK,l~K, RD3 and
Calculation
RE~D3=(RD3xRSP3xRK)~5D3
T~D=(TSP3xTK)/TSD3
~! R =[RPD3--(RD3~TPD/I~D) )]/[l--(RD3~TPD~ ]
2~ T=~PD/TD)(l -OE~D3xRPD3))]1[ 1-~RD3~ D)2)~ .
A=(l~Ro -T ) O 2
Roo=~A12) _ (A12) - 1
FC-~. 0
~c
,',~,' RooFC ROO+FC~GC3
~ 25 Note The calc~llations for Filters No. 5 and 6 are
,,,
carried out in the same man~er as for filter No. 3
except that the appropriate filter data are e~nployed.
FC is made equal tO zero for filrers No. 3, 5 and 6
for all samples.
-58-
',
-~ 108Z945
C culation of Ro,T,Roo,FC and RooFC from OMOD
data with the No. 4 filter in position.
Ilput: RSD4, RSP4, TSD4, TSP4, RSD7, TK, RK,
RD4 and GC4.
~, Calculation:
RPD4=~RD4xRSP4xRK)/RSD4
RPD7=(RD4~SP7~RK)/R~D7
TPD~TD=~TSP4xTK)/'I SD4
R =[RPD4 - ~RD4~T~),fI~ )]/~1-(RD4~TPD/TD) )]
T=~TPD/TD.Xl-(RD4xRl~D4))~ RD4~TPD~TD)2)]
A=(l+R 2 T2)~
o ~
"' Roo=(A/2) - ~ (A/2)2--1
FC=0.550~RPD7 - RPD4)
I R FC--R ~C~GC4
.~ 15 O oo
., .
:
:. :
"
.: _59_
' '
-- ~0~2945
.
On the basis of further experimental data, the factors relati~
the fluorescent component, as measured on the monitoring device, to the
fluorescent component as m~asured with the standard automatic color-
'~ brightness tester, have the following presently preferred values for
filter wheel position numbers 1, 2 and 4: 0.528, 0.636, and 0.456,
; ~ respectively.
~ '
... .
" ' .
~,
,. . .
~"'.;
i~J
;~'1 ' '
"','t
,,,
.~' .
,'' ,
;'~v.'~'i
~'~
: -60-
.: .
~O~Z945
OISCUSSION OF THE ON-MACHINE
SYSTEM OF FIGS. 1-6
Set Up Procedure For the System of Figs. 1-6
In the prototype system, potentiometers were included as part
S of the gain control resistance means and were adjusted for the respective
positions of the filter wheel 210 to give values correlated directly with
absolute reflectance and tra~smittance of the diiYusing glass, such as
given in the foregoing Table 1. In the preferred system of Figs. 1-6,
however, these poter~iometers for adjusting amplifier gain are omitted :
: 10 a~d are replaced with f~ed resistors 371-377 and 431-437 selected to
; give scale readings from meter 330 in ~he respective filter wheel posi-
, tions which are well above the values given in Table 1. The higher
gain values selected for the amplifiers 361 and 429 in the preferred
system are intended to provide improved stability and increased sensitivity
15 Of measurement.
, .
,~. :
,'' . .
. . .
:
.,
-61 -
-- ~O'~Z945
The upper and lower sensing heads are placed at a spacing of
: 3/16 inch by means of a gauging plate made of 3/16 inch Teflon
(trademark). The incident beam 133 forms a light spot of elliptical
configuration on the planar upper and surface 98 of the dif~sing
~; 5 window 135. The major axis of the elliptical light spot has a
length of about 5/8 inch and is parallel to the direction of web
movement, i.e.!the machine direction, while the minor axis has
- a length of about 3/8 inch and is at right angles to the machine
direction. The reflected beam 137 consists of the total light
' 10 reflected from a circular spot of approximately 3/8 inch diameter.
This viewed area lies substantially within the elliptical illuminated
; area on surface 98; however, the two essentially coincide in the
,~ .
r7 direction of the minor a xis of the illuminated spot.
Since the effective optical aperture 154, Fig. 3, of the lower
sen~ing head is of a diameter of about 15/16 inch, the system will
l i
- be insensitive to a certain amount of lateral offset between the
:!
optical axis 15 of the upper sensing head and the optical axis
515 of the lower sensing head.
'
. ,
s ~j
'~..1
,.:
.
:.;
, .
-62 -
~, .
- ~08~945
In setting up the sys~em, the position of the lower sensing
head may be ad~usted laterally so that the spot formed by the ihlcident
beam 133 is essentiallyoe~lered on the surface 98 of window 135.
The optimum relationship between the upper and lower sensing
' 5 heads can be precisely detectedbrobserving the reflectance output from
the upper sensing head (in any position of the filter whe_l 210) as the
heads are moved relative to one another while maintaining the spacing
of 3/16 inch between t~ heads. When the correct geometrical relation-
ship is attained between the incident beam 133, the reflected beam path
.- 10 137 and the plane of the surface 98 of the window 135, the renectance
signal will have a maximum value.
Wlth the upper and lower sensing heads in the optimum geome-
tric relationship~ and with the incident beam impinging on the cen~ral
part of surface 98, it is considered that relative shifting between the
upper and lower heads in theplæofsurface 98 over a range of plus or
minus 1/8 inch in any lateral direction should have an insignificant
ef~ect bscause oi ~he i~at platar coniiguratioD oi surface 98.
`.~
,
.,
-- .,
-63~
108294S
An important aspect of the disclosure relates to the measurement
of the basis weight of the moving paper web concurrently with the sLmul-
taneous measurement of reflectance and transmittance values for essentially
a common region of the web. Using the calculated value of infinite reflec-
tance R (including the grade correction factor) and the value of t~nsmit-
tance T, for example, for the same sample region, along with a concur-
rently obtained, average value for basis weight, essentially accurate values
of scattering coefficient s and the absorption coefficient k are obtained.
Such coefficients will exhibit essential independence of any variations
in the basis weight of the paper sheet material under ~hese circumstances.
The measurement of both a reflectance and a transmittance value
for a common sample region has an advantage over the measurement of
two reflectance parameters under conditions such as found in the paper
manufac~uring process since the transmittance measurement is relatively
insensitive to misalignment or tilting of the optical axis ~15 of the backing
assembly or lowerænsinz head 12, FIG. 3, relative to the optical axis
lS of the sensing head 11. This advantage is eSpecially important for
sheet material of relatively high opacity where two reflectance parameters
would tend to be relatively close in value.
. 20 Generally the results of laboratory tests discussed herein are
expected to be applicable to the on-line system. Thus the spread between
values of RooFC (See Table 3) obtained by the illustrated on-line system
and the corresponding values of AR FC taken as standard should not diffOE
.~ oo
~ by more ~han about plus or minus two points on a scale of zero
b~ 25 to one hundred, prior to any grade correction, for a wide ;ange of paper
sheet materials of different color and basis weight.
h.
' ~
~OB2945
The samples for which such accuracy was obtained in the labora-
tory included a range of basis weights of from 60 grams per square met~
to 178 grams per square meter for white paper. Without the use of a
correction fac~or, calculated R values which fell within two points of the
oo
S measured value included samples of paper colored white (several tints),
green, blue, canary, russett, ivory, gray and buff. Colors including pink,
gold, salmon, and cherry required a significant correction factor for the
XR~ YC and YA functions. All of the calculated R values involving the
XB and Z functions fell within 0.77 units of the m,easured value ona scale
10 of zero to 100, again without the use of any correction factor and regard-
less of color ~r basis welght,
.'' .
;,
.,. i
".
., ~
.. ~
~' ,
:.
., .
'''
1~)82945
The term quantitative measure of paper optical properties
as used in the claims refers to output quantities of a numerical
; nature such as supplied by the on-line digital computer system 996,
Figure 6, programmed as explained herein with reference to Figures
7 - 20. Examples of such quantities are those indicated in block
990, Figure 20; these quantities are identified with the correspond-
ing conventional paper optical properties in Table 21.
The term on-machine optical monitoring device is intended
generically and refers to the device 10, Figures 1 and 2, and other
comparable devices for sensing two essentially independent optical
response parameters such that a paper optical property is character-
ized prior to use of any correction factors with substantially improv-
ed accuracy in comparison to any characterization (prior to correction
factors) of such paper optical property from either of such optical
response parameters taken by itself. Such a monitoring device may be
used as an aid to manual control of the paper making process or may
be used as part of a closed loop automatic control system. Thus
"monitoring" does not exclude active control in response to the output
signals from the monitoring device.
Within the scope of the present subject matter, one or more
of the following paper optical properties may be sensed: brightness,
color, fluorescence, andtor opacity. Control of brightness and fluore-
scence offers a very substantial potential for cost reduction in the
production of a significant range of paper types. Color control, on
the other hand, may have important consequences regarding flexibility
of manufacture, product uniformity, and grade change flexibility.
....
.,
- 66 -
,'
082g4s
The value o~ on-line opacity control has already been
demonstrated to a large degree in a prior closed loop analog opacity
controller. In this installation, the average opacity across the web is
controlled almost exactly at any given desired value. In previous
S manually controlled operations, the PKT (Pigmentary Potassium
Titamate, K20-6Ti02 by du Pont) flow was set to some value chosen by
the beater engineer and usually held to such value for the duration of
the run of a given grade and weight. In the meantime, the paper
opacity varied up and down, depending on process conditions at the
time. Since the installation of the analojg opacity controller, the
opacity set point is adjusted rather than the PICT flow, thus holding
opacity constant at the desired level. Instead of opacity, the PKT flow
now varies up and down to compensate for other presently unavoidable
;j process upsets resulting from variations in broke richness, PKT
'J~ 15 solids, dye usage, save-all efficiency, and other machine retention
conditions. For a complete discussion of the installation of the analog
opacity controller, reference is made to F. P. Lodzinski article
"Experience With a Transmittance-Type On-Line Opacimeter for
.i .
Monitoring and Controlling Opacity", Tappi, The Journal of The
Technical Association of The Pulp and Paper Industry, Vol. 56, No.
2, Feb. 1973.
Existing on-line color meters have two serious disadvantages
as follows:
1. Each measures a reflectance value (Rg) which is decided-
ly different from that necessary for actual color and brightness char- -
acterizations. Off-line instruments, which adequately measure these
properties, r~quire that a pad of several thicl~nesses (Roo) of the same
paper be exposed to the light souroe aperture. Obviously, this is im-
-67-
1082945
. .
possible with an on-line instrument, unless the far more inaccessible
reel itself is tested. The use of Rg instead of Roo requires very
frequent off-line testing, and constant updating of an empirical
calibration procedure to maintain adequate accuracy. A separate set
S of calibration parameters for each grade and weight is also required.
Only in instances of extremely high opacity such as heavily coated,
or heavily dyed colors where Rg approaches Roo, can the above
problems be minimized to the point where accuracy becomes sufficient
for control purposes.
;~ 10 2. Existing color instruments are not equipped to measure
transmitted light which is much more sensitive to differences and, so
far, the only commercially proven method for the continuous
monitoring of opacity.
v' To assist in indicating the scope of the present disclosure,
15 the subætance of excerpts from an early conception record with
respect to the present subject matter are set forth in the following
paragraphs, headed "Proposa l" and "Proposed Instrument Des ign"
having reference to the defects of existing on-line color meters just
discussed:
20 Proposal:
An instrument built to the general specifications disclosed in
; the followlng section headed "Proposed Instrument Design" avoids the
above described defects and, at ~he same time, provides for a con-
cise, but extremely versatile, nearly total optical property monitor
2S and controller. Highly trained specialists in all fields required here,
including paper optics, color theory, photometry, computers, and
others, if needed are available. As an example, exact specifications
for the filters, photocells, and Light sources are essentially ready for
-68 -
~082945
manufacture now. Such specialists are also aware of factors impor-
tant to optical characterization frequently ignored by commercial
producers of optical instruments.
Proposed Instrument Desi~n:
S An instrument made up of two scanning sensing heads, one
above and one below the moving paper web, and a dedicated computer
with appropriate couplers for input and output, is envisioned. The bot-
tom head would receive light transmitted throlgh the sheet and subsequently
analyzed for its X, Y, and Z tristimulus components. It would also
contain a backing Oe some specified effective reflectance (possibly a
black body of zero, or near zero, reflectance) located just ahead or
~,
behind (machine di~ection) of the transmitted light receptor compartment~s~.
The upper head oDuld contain the light source, as well as a
reflected light receptor. The latter occurs after reflection from the
,:q
moving web at a point just above the backing, on the bottom head and
!~ would a lso be analyzed for its X, Y, and Z tristimulus components.
Both light receivers and, for that matter, the light source itself could
be integrating cavities of a type. This would be one way to insure the
;.
i~i uniform distribution of emitted, transmitted, and reflected light in the
20 X-direction in addition to providing identical samples of light going to
each phDtoelecric cell installed with filters within the cavities themselves.
Thermcstatically contrdled heaters or coolers would likely be desirable
- for temperature control. The flux of the light source could be moni-
tored or controlled by a third partial, or full, set of filter-photocell
25 combinations. The availability of both the transmitted (T) and
reflected (Rg) light signals described above allows for precise com-
putation of the reflectance with an infinite, backing (R~. It is the
latter, R.~value, which is required to characterize color, brightness,
-69 -
.,.
108294s
and an index of fluorescence. In addition, it would eliminate the needfor any grade corrections in measuring either printing or TAPPI
opaeity, both of which eould be made available if desired.
A small, rather low-cost, dedieated computer with ap?ropriate
5 interface equipment, could be used ro receive all signals, eompute all
p-rtinent optieal properties, and determine the signal for direct, closed
loop control of:
a. 2-5 separate eonventional dye additions
b. fluoreseent dye feed to the siz~ press: and
e. PlCT, TiO2, or other slurry flow:
so that brightness, opaeity, color (L, a,b) and fluorescenee could be
maintained almosr exaetly as ehosen by, perhaps even a master
eomputer, if desired.
Kubelka-Munk equations, quantitative color descriptions, and
15 their inter-relationships, reeently aequired wet end mathematieal models,
along with existing eontrol theory, are all presently available in some
form or other to eonvert the input signals from the seanning heads to
optieal measurements and flow feeds with whieh paper manufaeturers are
familiar. The eombined mathematieal technology above is also suffieient
20 for adequate deeoupling of this otherwise eomplieated information so
that overlapp-d eontrol is avoided.
Use of a dedieated computer wculd eliminate mosr of the
eleetronics now assoeiated with op~ieal measuring equipment. It eould
also be used to integrate resuks across the web and simplify and/or
25 maintain calibration. The package ~ould lend itself to rather universal
applieation and minimize Lhe ~ime and effort on -he p~r~ of the purchaser.
~ The key feature of this proposed instrument, which distinguishes
it from existing on-line op ical testers, is that it calls for the
-70 -
~```` 108Xg45
measurement of both transmitted and reflected light without undue
complications. This, in turn, can cause a great deal of improvements
regarding sensitivity, accuracy7 flexibility, and thoroughness of a
continuous optical property measuring device.
The following Table will serve to identify the computer
symbols used in Figs. 17-20 with the corresponding conventional
, symbols and terminology used in the text.
~:s .
`~ Table 21 - Identification of Computer Symbols Used in
F igs. 17 -20
Computer Conventional Conventional Term
Svmbol Symbol for Symbol
SGCF GC(Table 3) specific grade
correction factor
RG RD(Table 3) Nominal reflectance
of the diffuser
window 135
TD Td Nominal transmit-
tance of the diffuser
window 135
RZERO R0(Table 3) reflectance with
(RZTABL) black body backing
, for each filter wheel
position I equals
zero through five.
T T(Table 3) Transmittance with
(TTABL) black body backing
; for each filter wheel
position I equals
- zero through five.
RINF Roo (Table 3) infinite backing
(R ITA BL) reflecta nce for f ilter
~- wheel positions I
~;; equals zero tO five.
~ .~
S S scatter coefficient
(STABL) for each filter
"
;'
~ -71-
.
08~945
wheel position
I equals zero
through fit.re.
IC K absorDtion
(lCTABL) coefficien~ Lor
each filter wh-el
position I ~ ~als
zero throu~h
five.
ZFLUOR - fluoresrent
contri~ution to
tristimulus Z
refle tance
ZRINF - tristimulus Z
infinite backing
reflectance with
fluorescence
, ., X~RINF - triatimulus X
infinite backi~g
reflectance with
fluorescence
; BRRINF - TAPPI brightness
. (see Table I in
the first section
~, 25 of this Topic ~or
spectral distribu-
; tion of ~he first
filter wheel
; position)
POPAC R o/Roo pri~tin~ opacity
`; YAR89 R~, 89 tristimulus Y
:~ reflectancc ~virh
89 baclcing
TO~AC RJRo 89 TAPPI opacity
XTRI X C I.E. tristimulus
.. coordinate X
,,:
YTRI Y ~. I. E. ~ristimulus
.. coordinate Y
.,~ . .
~
~' - ZTRI Z C. I. E. trii;imulua
coordinate Z
LH L Hun~er coordin~;e
: L
: AH, BH a, b Hun.e r coor~in~tea-
a, b.
-72-
~082g45
Scope of the Early Conception of This Invention
Given the foregoing conception, it is considered that many
modifications and variations will be apparent to those skilled in the
art. The basic conception claimed herein is the sensing of two
S essentially independent optical response parameters of a single
, thickness relatively homogeneous sheet material such that any other
desired parameter or paper optical property can be accurately
ca lculated .
The present subject matter is limited on the determination
of the specified optical properties of single thickness relatively
homogeneous sheet material with the use of filters of appropriate
spectral response characteristics as explained in the present specifi-
cation. For the case of opacity measurement, for example, the
present invention, is particularly applicable to an optical system
wherein system spectral response essentially simulates the C. I. E.
tristimulus Y filter with either illuminant A or C and to near-white
papers as explained in the Lodzinski Article of ~ebruary, 1973.
`:~
'
- -73 -
1082945
'
.., '
PROPOSED OFIlCAL CONTROL ~TRATEGY
_ _
While the on~ine automatic control of paper optical properties
is an ultimatc objective of the work reported herein, the claimed
subject matter relates to onmachine monitoring of paper optical
5 properties whether used as an aid to conventional manual control or
for other purposes. Nevcrtheless, in order to provide a disclosure
of the best mode presently contemplated for automatic control as a
separate but related area of endeavor, the following discussion is
presented.
The optical properties of a sheet of paper are dependent upon
all of ~he materials of which it is made but primarily upon the furnished
pulp, fillcrs, pigments, dyes, and some additives. It is often very
` difflcult to maintain the optical attributes of the pulp, fillers and
additi~es constant within a given production run. Such variation is
15 even greater between runs. The optical properties of the finished paper
may, however, be rcasonably controlled to specified standards by
varying the additions of dyes and fillers and pigments until ~he desired
,; compensations are achieved. The problem is that each furnished
,, .
ingredient affects each of the resulting paper optical properties in a
. .
i`~ 20 rather complicated manner. Indeed the intuition of experienced
` papermakers has essentially been the sole method of optical property
;, control. Unfortunately, this approach is inefficient, resulting in
considerable off-standard paper and/or waste of costly materials.
Accordingly, a dire need exists within the paper industry for a highly
reliable and continuous optical property monitor coupled with a closed
loop computer control syscem.
The v~lue o4 s~lch closed loop c ontrol, ~ased ~n a feedbac!~
-74-
.
1082945
` color detector, has already been demonstrated for the )ntinuous
addition of two and sometimes three dyes. (1) (2)Target dye concentra-
tions changes of up to three dyes can be determined by solving threc
simultaneous equations containin~ three unknowns. (1) One disadvantage
5 of such control is that accurate color monitoring is not presently
j~;
available unless large and frequent empirically determined correction
factors are applied to the original output results. A sccond dis-
advantage arises when opacity and the fluorcscence m~lst also be
simultaneously controlled. In this case the nurr~er of independcntly
10 controlled continuous additions incrcases from three to five. An
optical brightener and an opacifying pigment constitute the two
additional factors.
An object of this invention is to demonstrate a method by
which fluorescence can be continuously monitored. A means by which
lS the optical brightener addition can be separately and independently
controlled is inherently implied. The paper color is also analyzed
without the fluorescent contribution. It is, of course, this latter
characterization (without fluorescence) which should be, but which
has not in the past been, used to determine the required addition of
20 the conventional dyes. In other words, the effect of the optical
brightener is decoupled from the three conventional dyes making
possible the simukaneous control of all four dyes.
Another portion of this invention dcmonstrates a means of
continuously determining the scattering coefficient of the moving web
:~,
25 for each of the six available light spectrums. It is possible to determine
the scattering coefficient required tO ~chieve a given opacity specification
whenever the basis weight and absorp~ion coefficient are known. When
-75 -
.,
~082945
:
':
the latter are set equal to a given set of product specificadons, then
the calculated scatterin~ coefficient becomes the target scattering
coefficient. (The absorption coefficient can be acquired by off~ine
testing of a sample of the standard color to be matched. In reality,
this becomes a target absorption coefficient as well.) Thc dyes have
little, if any, effect on the scattering coefficient but the effect of the
slurry pigment is very large. Thus the target scattering coefficient
is used as the sole feedback variable to control the slurry pigment feed.
This will insure that the opacity is at or near the spccification as
long, as the absorption coefficicnt and basis weight are also on ~arget.
The absorption coefficient should, of course, be on target by virtue
of the independent color control. A completely independent system -
controls the basis weight.
A method by which the decoupling of three conventional dyes,
one cptical brightener and one opacifying pigment has hereby been
e~plained. Heretofore, such decouplirg as revealcd in the prior
art has been limited to three absorptive dyes and thereby neglecting
the need to also achieve a specified degree of fluorescence and opacity.
References
1. The development of dynamic color control on a paper
. machine by H. Chao and W. Wic~strom; Automatica, Vol.
. .
`` 6 PP 5-18, Pergamon Press, 1970.
2. Another consideration for color and formation by Henry
H. Chao and Warren A. Wic~;strom, color engineerin~, Sept/
Oct. 197 1.
-76 -
10~2945
The present invention is for the purpose of obtaining a
quantitative measure of an optical property such as brightness,
color, opacity or fluorescence of single thickness sheet material.
The sheet material is substantially homogeneous in its
thickness dimension such that the optical property of interest can
be reliably calculated from reflectance and transmittance measure-
ments on the basis of existing theory. Thus the present invention
is not applicable to the sensing of localized surface effects (such
as due to surface migration of light absorbing powder particles, for
example). To the contrary the present invention is concerned with
the average or bulk optical characteristics of the sheet material
considered as a whole, and especially is concerned with the character-
istics of paper sheet material as it is delivered from a paper machine
:
after completion of the paper manufacturing process.
The present invention in its broader aspect does not require
strictly homogeneous material since empirical correction factors can
; be applied for cases where theory is less effective. For example,
, the paper optical properties of calendered and coated papers may be
; effectively measured by the system of the present invention using
grade correction factors to correlate on-machine results with the
measurements obtained by standard off-line instruments.
j,.t The optical system of the monitoring device includes
,.
components such as those shown in Figure 3 which define or optically
.~ affect the incident, reflected and transmitted light paths such as
~` indicated at 133, 137 and 141-143 in Pigure 3. For the case of a
. filter wheel as indicated in Pigure 4, each filter wheel position may
be considered to define a separate light energy path with its own
predetermined spectral response characteristics.
- 77 -
t.,'~,
' .i
^ :
'
''
j . '' ,, ' ' . ' .
108Z945
In each filter wheel position, there are two distinct light
energy paths for measuring a reflectance value and a transmittance value,
respectively. In the illustrated embodiment e ch such light energy path
includes a common incident light path 133, but the paths diverge, one
5 coinciding with the reflectance sensing light path 137 and the other
' including the transmittance sensing light path. The photometric sensors
203 and 260 thus provide simultaneous reflectance and transmittance
output signals with respect to essentially a common region of the web.
The reflectance sensing light path collects light from a circular region
10 with a diameter of about 3/16 inch, and the transmittance sensing light
path collects light from a total elliptical region which includes substan-
t;~lly the same circular region as mentioned above. Because of sheet
formation effects and other localized variations in web characteristics
it is considered valuable that the reflectance and transmittance output
15 signals are based on readings from essentially a common region of the
web.
By taking at least one reading in each traverse of the web, and
taking such readings at different points along the width in successive
l:raverses, it is considered that accuracies equal to or superior to those
20 of an off-line sampling of a finished reel can be achieved, while at the
same time the readings are available immediately i~;tead of after comple-
` tionofamanufacturing run.
:~ By way of example, in the illustrated system a traversal OI
the web by the sensing head takes about forty-five seconds, so ~hat the
25 sensing head operates at a rate of at least one traversal of the width
of the web per minute in ;he time intervals b-tween the hourly off-sheet
standardizing operations.
-78 -
.
108Z945
an improvement to
I~ accordance with the teachings oythe present invention, the
optical window 13~ is itself selected as to its optical characteristics so ~s
to provide the basis for off-sheet standardization. To this ena it is advan-
tageous that the optical window exhibit an absolute reflectance value as
5 measured by the standard automatic color-brightness tester of at least about
thirty-five per cent (3S%). The corresponding absolute transmittance value
- as measured on the G.E. Recording SpecLlophotometer with conventional
optics is about fifty-six per cent (56~o)~ With the illustrated embodiment,
once the system is properly adjusted with respect to the zero reflectance
readings ~as by the use of a black sheet of known minimal reflectance) the
system maintai~s such zero adjustment quite stably; accordingly the higher
the reflectance value of the optical window, the more effective is the re-
flectance standardization by means of the optical window. On the other hand
.,
a transmittance value which is of a reasonable magnitude is also desirable,
so that the provision of an optical window with substantial values of absolute
reflectance and transmittance is advantageous.
~ ~ '
With the illustrated embodiment, the transmittance readings for the
moving web are relatively more nearly ~ndependent of misalignment of the
upper and lower sensing heads than the reflectance readings. Further it is
; 20 considered that tilting of the lower sensing head relative tO the optical axis
of the upper sensing head has less effect on transmittance readings than on
reflectance readi~gs. Thus it is considered that it would be ad~antageous
to have an optical window such as 13~ with an absolute reflectance value of
seventy per cent (70~3~O) or more. A value of reflectance as high as nine~y
.i
2S per cent (90~O) would not be unreasonable and would generally still permit
a transmir.ance value of a substantial magnitude to give reasonably com-
parable accuracy of reflectance and transmirance readinas for on-line opera-
tion as herein described.
, .
-79 -
lO~Z~45
. .
While separate photometric sensing means for the reflectance
and transmittance readings have been shown, it is possible with the
use of fiber optics, for example, to use a common photometric sensor
and alternately supply light energy from the reflectance and trans-
mittance light paths thereto, providing the response time of the sen-
sor enables reflectance and transmittance readings to be obtained for
essentially the same region of the moving web. Generally the possi-
bility of such time multiplexing of reflectance and transmittance
~ readings will depend on the speed of movement of the web and the degree
; 10 of uniformity of sheet formation and the like.
It is very desirable that the system of the present invention
be applicable to sheet materials having a wide range of characteristics
such as basis weight and sheet formation, and operable at high speeds
; of movement such as 100 to 3000 feet per minute. Further, for maximum
accuracy, it is necessary that a region of the sheet material being
sampled have substantially uniform opacity. Accordingly, especially
for sheet material of relatively low basis weight and relatively poor
sheet formation, greater accuracy can be expected when the response
of the photometric sensor is relatively fast, and when reflectance
and transmittance readings are taken simultaneously and are a measure
of the characteristics of a common sampling region of minimum area
(consistent with adequate signal to noise ratios). Thus multiplexing
of reflectance and transmittance readings is not preferred for the
: case of high speed paper machinery and comparable environments, nor
is it desirable to use reflectance and transmittance light paths
which intersect the web at spacially offset regions.
With respect to speed of response of the photometric sensing
means, substantial improvements over the previously described components
, .
~; - 80 -
~o82945
are deemed presently available. If the spectral response and other
necessary characteristics are suitable, a sensor with such a higher
speed of response is preferred for the illustrated embodiment. Good
experience has been had with a silicon photocell presently considered
5 as having an appropriate spectral response characteristics for color
and other measurements in accordance with the present invention. The
specific silicon cell referred to is identified as a Schottlcy Planar Diffuse
Silicon Pin 10 DP photodiode of a standard series supplied by United --
Detector Technology Incorporated, Santa Monica, California.
i'
,.,
,;,
~, ,
~,
~,
,
1. ,.
. .
'.
,:,
'.~
-81 -
~082~45
In place of a rotatable filter wheel arrangement as shown
in Figures 3 and 4, a set of twelve fiber optic light paths may define
six simultaneously operative reflectance light paths in upper sensing
head 11 and six simultaneously operative transmittance light paths in
lower sensing head 12. The six reflectance fiber optic paths would
include respective filters corresponding to filters 281-286 and respec-
tive individual photocells and would be located to receive respective
portions of the reflected light which is reflected generally along
path 137 in Figure 3. The six transmittance fiber optic paths would
also include respective filters corresponding to filters 281-286 and
respective individual photocells, and would be located to receive
respective portions of the transmitted light which is transmitted
generally along paths such as 141-143 in Figure 3. The filter means
in the incident light path such as indicated at 133 in Figure 3 might
include a filter in series with filters 271 and 272 for filtering out
the ultraviolet component from the incident beam, so that the twelve
simultaneous photocell readings corresponding to those designated
RSDl through RSD6, and TSDl through TSD6 (when the device is off-sheet),
and corresponding to those designated RSPl through RSP6, and TSPl
through TSP6 (when the device is on-sheet~ will exclude a fluorescent
contribution. (See Table 3 where this notation is introduced).
. . .
If a reflectance reading corresponding to RSD7 (when the
device is off-sheet) and corresponding to RSP7 (when the device is on-
sheet) is desired so as to enable computation of fluorescent con-
tribution to brightness, it would be necessary to mechanically remove
the ultraviolet filter from the incident light path, or otherwise
introduce an ultraviolet component of proper magnitude, and obtain
another brightness (Z) reading, for example from the number four
reflectance photocell.
'~'
- 82 _
.:
. .
~o8Z945
As an alternative to the above fiber optic system with a common
incident light path, seven fiber optical tub`es incorporating filters corres-
E~;ndi~ to 281-287 of FIGS. 3 and 4, respsctively, at say the light exit
points of the tubes, could be used to supply the incident li~ht to seven
5 different points on the paper web. The reflected ligh~ from each of these
seven points could be monitored by seven different systems, each involving
lenses and a photocell, and the numbsr seven reflected light path including
also a filter corresponding to filter 288, FIG. 4. The transmitted light
from the f~rst si~c points would also need to be kept separately, and this
10 could be accomplished by six integrating cavities and slx photocstls.
As a further alternative the seven fiber optical tubes defining the
seven incident light paths could have a second set of seven fiber optical
tubes and photocells respectively disposed to re~eive reflected light from
the rQspective illumi~ated points. Another set of six fiber optical tubes
,. .
, 15 and photocells could be associated with the first six illuminated points
:.
for receivi~g transmitted light. This could eliminate the need for the
light collecting lenses in the upper sensing head and the integrating cavi-
ties in the lower sensing head.
The last two mentioned alternatives with seven fiber optical tubes
20 defining the incident ligh~ paths appear to be rather complicated systems,
but they do offer means of eliminating both the mechanical filter wheel
as well as any mechanical device to control the presence of ultraviolet
light in the incident beam.
Still another alternative ls to use "screens' in addition ~o the
:,.
25 filters in the embod~nent of FIGS. l-~. The new pho~odiodes are con~ider-
ed sensitive enoughtomeasure reduced li~h~ intensites so ~hat sc~eenswi~h
differenttransmittance values could be used with Si~Y of the incident b-am
-83 -
10~2945
filters so that the net photocell output for each reflectance light path,
and for each transmittance light path, would be similar enough so that
separate and invidual pre-amplification for the respective reflectance
outputs would not be necessary, and so that separate and individual pre-
5 ampli~ication for each transmittance output would not be necessary.
This means that reed switches 341-347 and 351-357, and ~elays K
through K7 in FIG. 6 could be eliminated, and that the feeback paths
for amplifiers 361 and 429 could have the same resistance value in each
filter wheel position. A means of sensing filter wheel position would st
10 be necessary, butthis could be done in a number of simple ways, one of
which would be a single reed switch such as reed switch 3S8 shown in
FIG. 6. The number of necessary conductors in the cables 51 and 52,
FIG. 5, would, of course, be reduced in this modification.
' The term "rscreen is understood in the art as referri~to a net-
15 worlc of completely opaque regions and intervening openings or completely
translllcent regions, such that light energy is uni~ormly attenuated over the
., .
entire spectrum by an amount dependent on the proportion of opaque to
~; transmitting area.
The device of Figs. 1 and 2 has been tested on a machine
20 operating at about 1000 feet per minute, and no pro~lems ha~e appeared
in maintaining the nec~ssary uniform and stable contact geometry be-
' tween the head and the moving we~. -
It will be apparent Ihat many further modifications and variations
may be effected without departing from the scope of the novel concepts
25 of the present invention.
,.:
'~
' ! -84-
