 |
Description  |
|
|
BACKGROUND OF THE INVENTION
This invention relates to apparatus for radiation absorption measurement
and, more particularly, to apparatus for determining the moisture content
of a moving web of material by measuring the relative reflectance of two
beams of infrared radiation.
A well-known method for determining the content of a substance such as
water in a material involves the measurement of relative infrared
reflectance. At certain characteristic wavelengths corresponding to
resonance of a particular substance, the absorption, and thus reflectance,
of the material being analyzed varies considerably with the content of the
substance, while at other wavelengths not coinciding with a resonant
wavelength, the degree of absorption is relatively insensitive to changes
in material composition. By measuring the ratio of the reflectance of the
material at resonant and nonresonant wavelengths, the content of the
substance can be determined simply and rapidly. In many practical
applications of this technique, both the resonant-wavelength beam and the
nonresonant-wavelength or reference beam are derived from a single
radiation source using a chopping wheel to obtain alternating pulses of
radiation. One such implementation is shown in U.S. Pat. No. 3,150,264,
issued to R. C. Ehlert.
Systems of the type described above are often used in on-line applications
to measure the moisture content of webs of paper pulp or the like. Such
systems, however, are susceptible to errors resulting from the
nonlinearity inherent in any practical detector used to sense infrared
radiation. Because of this nonlinearity, the reflectance measurement is
sensitive to spurious sources of infrared radiation such as the moving web
itself, which may be as hot as 300.degree. F. Variations in the absolute
amplitudes of the reflected resonant and nonresonant radiation pulses due
to sheet flutter and the like will also affect the measurement, even
though the amplitudes of the respective pulses vary proportionately.
Finally, the nonlinearity of the detector makes the reflective ratio
measurement sensitive to changes in operating point due to changes in the
ambient temperature.
Error may also result when the sheet flutter contains frequency components
near the chopping frequency, which is typically about 10 Hz. Since the
apparatus cannot distinguish between amplitude variations due to sheet
flutter and those due to changes in material composition, the measured
reflectance ratio will contain a spurious component at a frequency equal
to the difference between the flutter frequency and the chopping
frequency.
SUMMARY OF THE INVENTION
One object of my invention is to provide a moisture measuring apparatus
which may be used to measure moisture in a rapidly moving web.
Another object of my invention is to provide a moisture measuring apparatus
which is insensitive to spurious sources of radiation and changes in
ambient temperature.
A further object of my invention is to provide a moisture measuring
apparatus which is insensitive to web flutter.
Other and further objects will be apparent from the following description;
In general, my invention contemplates apparatus for analyzing a material by
measuring relative infrared reflectance in which a chopped radiation beam
is produced by arranging a tuning fork such that an oscillating element
alternately moves first and second spaced optical bandpass filters into
position to intercept a beam of source radiation. The chopped beam is
directed on the material being analyzed so that a detector receptive to
radiation reflected from the material generates spaced apart pulses
corresponding to the radiation pulses. The relative transmittances of the
filters are such that the radiation detector generates alternating pulses
of equal amplitude for a material having some standard or specified
composition.
By using a tuning fork rather than a motor-driven chopping wheel such as
used in the prior art, I can attain a substantially higher chopping
frequency and thus greatly reduce the sensitivity of the apparatus to
sheet flutter, since the sheet movement in any one chopping cycle due to
sheet flutter is minimal. In addition, the absence of any wearing surfaces
gives the tuning fork obvious mechanical advantages over shaft-driven
chopping wheels of the prior art.
Further, by choosing filter transmittances to provide alternating reference
and resonant-wavelength pulses of equal amplitude for a material having
some standard content of moisture, for example, I effectively minimize
errors due to nonlinearity in the radiation detector. Deviations in the
content of the material from the norm are translated into deviations in
the ratio of the amplitudes of the alternating pulses from unity. Such
extraneous influences as sheet flutter and stray or external radiation
produce negligible error in the amplitude ratio if such ratio is close to
unity. This is especially important in systems in which the moisture
content measurement is used for feedback control.
In another aspect, my invention contemplates an improved system for
controlling the temperature of the radiation detector to govern its
operating point. Prior art systems have typically used a temperature
sensor, separate from the radiation detection, which is coupled to a
suitable feedback temperature control. The temperature control receives no
input from the radiation detector itself. As a result, random variations
in manufacture produce corresponding variations in the "dark" current and
hence the operating point of such detectors, resulting in unreliable
operation. My apparatus avoids this defect inherent in prior art systems
by using the detector signal itself as a composite indication of the
temperature of the detector and of the magnitude of stray radiation from
the hot web. By using the detector itself rather than a separate sensor to
control temperature, I directly control the operating point of the
radiation detector and hence reduce errors in the measurement of amplitude
ratios differing appreciably from unity.
BRIEF DESCRIPTION OF THE DRAWINGS
In the accompanying drawings to which reference is made in the instant
specification and in which like reference characters are used to indicate
like parts in the various views:
FIG. 1 is a side elevation, shown partly in section, of my moisture
measuring apparatus.
FIG. 2 is a top plan view of the tuning fork.
FIG. 3 is a top plan view of the optical slit.
FIG. 4 is a schematic diagram of the signal processing portion of my
apparatus.
FIG. 5 is a graph of the displacement of the tuning fork tines as a
function of phase angle.
FIG. 6 is a graph of a typical waveform generated by the radiation
detector.
FIG. 7 is a schematic diagram of an alternative signal-processing circuit
for my apparatus.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
Referring to the drawings, my moisture measuring apparatus, indicated
generally by the reference numeral 10, is disposed facing the surface of a
moving web 12 of paper or the like. A collimating lens 16 directs
radiation from an incandescent bulb 14 or other suitable source onto the
surface of the web 12 along a radiation path 17. A tuning fork 18 with a
resonant frequency of 80 Hz and having an opaque filter mount 20 attached
to one of its tines is arranged with the filter mount 20 intercepting the
radiation path 17. A counterweight 21 is attached to the other tine to
balance the tuning fork 18. A spacer 22 separates the tuning fork 18 from
a supporting member 24 mounted on a frame 26. Disposed between the tines
is a coil assembly 28 wound on a permanent magnet core 30. The coil
assembly 28 comprises a drive coil 32 and a pickup coil 34 (FIG. 4).
Pickup coil 34 is coupled to a tuning fork control circuit indicated
generally by the reference numeral 36. In the control circuit 36, a
detector 38 coupled to the pickup coil 34 provides a signal representing
the envelope of the pickup coil signal. The envelope signal from detector
38 is fed together with a reference signal from a source 40 to a
comparator 42. The output of the comparator 42 is connected to the gain
control input of a variable-gain amplifier 44. Pickup coil 34 is connected
to the signal input of amplifier 44. The output of amplifier 44 is
connected to drive coil 32.
The control circuit 36 ensures a constant amplitude of vibration of the
tunning fork 18. When the apparatus is first turned on, the output from
coil 34, and thus detector 38, is zero. Comparator 42 provides maximum
gain from amplifier 44, and oscillations build up. When the output signal
from detector 38 becomes substantially equal to the reference signal 40,
comparator 42 reduces the gain of amplifier 44 to a value sufficient to
sustain oscillation at the desired amplitude, determined by reference 40.
The filter mount 20 holds two spaced, parallel slit-shaped filters 46 and
48. The filters are arranged such that they intercept the radiation path
17 when the tuning fork tine carrying the filter mount 20 reaches its
maximum displacement from equilibrium position. A fixed slit 50 arranged
parallel to the slits 46 and 48 is disposed in the radiation path 17. The
width of slit 50 is less than that of filters 46 and 48 to ensure that the
amplitude of the transmitted beam is relatively insensitive to variations
in the amplitude of vibration of the tuning fork 18.
The filters 46 and 48 are selected to pass respectively radiation having a
reference wavelength different from any resonant wavelength of the
substance being measured and radiation corresponding to a resonant
wavelength of such substance. In a typical application involving moisture
measurement in a paper web, reference wavelength filter 46 may have a
passband wavelength of 1.81 microns, while resonant-wavelength filter 48
may have a passband wavelength of 1.94 microns. Assuming that the desired
or normal moisture content is about 18%, the transmittance of
resonant-wavelength filter 48 should be about 25% higher than that of the
reference filter 46. For higher or lower desired moisture contents, the
25% figure should be raised or lowered proportionately.
Reflected radiation from the web 12 passes through an infrared-transmissive
quartz window 52 and is collected by an integrating sphere 54 arranged
concentrically with the incident radiation path 17. Integrating sphere 54
is formed with a diffusely reflective inner surface such as spun aluminum
which has been roughened by grit blasting or the like. A cylindrical
shield 56 arranged coaxially with the incident radiation path 17 prevents
interaction between the incident and reflected radiation and ensures that
no incident radiation impinges on the inner surface of the sphere 54. A
lead sulfide detector 58 disposed to one side of sphere 54, as seen in
FIG. 1, senses the reflected radiation after it has been diffused by the
inner surface of sphere 54. The detector 58 is cooled by a surrounding
water jacket 60 through which water or other liquid is pumped and is
simultaneously heated by a heating element 62 which is controlled in a
manner to be described. Preferably, the interior of the sphere 54 is
pressurized with dry nitrogen gas to eliminate inaccuracies due to changes
in the relative humidity and to protect the optical and mechanical
components. Accordingly, shield 56 is provided with a quartz window 53.
One terminal of detector 58 is connected to a line 64 supplying a constant
DC voltage. The other terminal of detector 58 is coupled to ground through
a temperature-compensated resistor 66. The output of detector 58 at its
junction with resistor 66 comprises a series of alternating pulses 68 and
79 (FIG. 5) corresponding respectively to the intensity of the reflected
radiation at the reference and resonant wavelengths and include a
relatively constant "background" signal 72 corresponding to the detector
output in the absence of any incident radiation. Pulses 68 and 70 occur
respectively at maximum positive and negative displacement of the tuning
fork tines (FIG. 2), when respective filters 46 and 48 are in position to
intercept the radiation path 17. For illustrative purposes only, I have
shown the amplitude of reference pulse 68 to be appreciably greater than
that of resonant-wavelength pulse 70, as would be the case if filters 46
and 48 were to have substantially equal transmittances. However, as
previously pointed out, filter 46 has less transmittance to approximate
the attenuation due to absorption in the resonant-wavelength beam. The
background signal 72 is observable when the tuning fork tines are at their
equilibrium position; the radiation path 17 is blocked by the opaque
filter mount 20. The opaque portion of filter mount 20 disposed between
the filters should have a width greater than that of aperture slit 50 to
insure complete interruption of the incident beam. The detector output is
connected to a buffer amplifier 74. Sample-and-hold circuits 76, 78, and
80 sample the output of buffer amplifier 74 at discrete instants in the
operating cycle. More particularly, a line 82 enables circuit 76 with a
sampling pulse at 90.degree., or maximum positive displacement, so that
circuit 76 provides an output corresponding to the reference-wavelength
pulse 68. Similarly, a line 84 enables circuit 78 at 270.degree., or
maximum negative displacement, so that circuit 78 provides an output
corresponding to the resonant-wavelength pulse 70. A third line 86 enables
circuit 80 at 0.degree. and 180.degree., or zero displacement of the fork
tines, so that circuit 80 provides an output corresponding to the
background signal 72. Sample-and-hold circuits 76 and 80 drive respective
positive and negative inputs of a differential amplifier 88 to provide a
corrected reference-wavelength signal compensated for the background
signal 72. Similarly, sample-and-hold circuits 78 and 80 drive the
respective positive and negative inputs of a differential amplifier 90 to
provide a corrected resonant-wavelength signal compensated for the
background signal 72. A divider circuit 92 responsive to the outputs of
amplifiers 88 and 90 provides an output representing the ratio of the two
signals. The output of circuit 92 is fed to a graphic or other recorder 94
and may be used to drive a suitable feedback control circuit (not shown).
The background component derived by circuit 80 also provides a signal for
controlling the electrical power supplied to the heating element 62. More
particularly, the output of circuit 80 is coupled to the negative input of
a differential amplifier 96. The positive input of amplifier 96 is coupled
to the common terminal of temperature-compensated resistors 98 and 100
which are series-connected between ground and line 64. The output of
differential amplifier 96 is coupled to a power amplifier 102 which drives
the heating element 62. Resistors 98, 100, and 66, in conjunction with
detector 58, form a bridge circuit. The bridge circuit is nulled by
controlling the temperature of detector 58. Any increase in the amplitude
of the background signal 72, due to an increase in temperature of either
the detector or the web, results in an increased output from
sample-and-hold circuit 80. This increased output in turn decreases the
output from differential amplifier 96. As a result, amplifier 102 supplies
less power to heating element 62 and allows the detector 58 to be cooled
by the water jacket 60. Similarly, any decrease in the amplitude of the
background signal 72 due to a decrease in temperature will cause more
power to be supplied to the heating element 62 and thus raise the
temperature of the detector 58.
The sampling pulses on lines 82, 84, and 86 are derived by a pulse
generator circuit indicated generally by the reference numeral 106. In
this circuit, a buffer amplifier 108 responsive to the pickup coil 34
drives a Schmitt trigger 110 and a differentiator 112. Schmitt trigger 110
drives a one-shot multivibrator 118 which produces an output pulse
whenever the output of amplifier 108 changes from negative to positive.
Schmitt trigger 110 also drives a second one-shot multivibrator 116
through an inverter 114. Differentiator 112 drives a Schmitt trigger 120.
Schmitt trigger 120 in turn drives a one-shot multivibrator 122 directly
and a second one-shot multivibrator 124 through an inverter 126.
Multivibrator 122 provides an output pulse whenever the output of
differentiator 112 changes from negative to positive. One-shot
multivibrators 122 and 124 are each coupled to inputs of an OR gate 128.
The output of the pickup coil 34 is a cosine signal proportional to the
time derivative of the displacement plotted in FIG. 5. Amplifier 108
provides this cosine signal to Schmitt trigger 110 and differentiator 112.
Differentiator 112 further differentiates the cosine signal to provide an
inverted sine signal to Schmitt trigger 120. At 180.degree. in the
oscillation cycle, when the fork tines are at their equilibrium position,
the inverted sine signal from differentiator 112 changes from positive to
negative, causing an output pulse from multivibrator 122 which is coupled
through OR gate 128 to line 86. Similarly, at 0.degree., with the fork
tines at equilibrium position, inverter 126 causes multivibrator 124 to
provide a pulse which is also coupled through OR gate 128 to line 86.
When, at 270.degree., the fork tines reach their maximum negative
displacement, the cosine signal provided by amplifier 108 changes from
negative to positive; and multivibrator 118 provides a pulse on line 84.
Similarly, at 90.degree., when the fork tines reach maximum negative
displacement, inverter 114 causes multivibrator 116 to provide a pulse on
line 82. OR gate 128 thus provides a pulse on line 86 whenever the fork
tines pass through their equilibrium position, while one-shot
multivibrators 116 and 118 provide pulses on lines 82 and 84 whenever the
fork tines reach their maximum positive and negative displacements,
respectively.
In the circuit shown in FIG. 4, the temperature control signal is derived
by sampling the detector output during the dark period. In practice, the
peak amplitude of the reference-wavelength pulse 68 and the
resonant-wavelength pulse 70 are sufficiently constant over a long period
of time that the heating element 62 can be controlled by a signal that is
simply the time average of the detector output. In FIG. 7, I show an
alternative signal-processing circuit in which the detector output is used
directly to control the heating element 62. More particularly, the
detector 58 is coupled between a line 130 providing a DC potential of -100
volts and the negative input of a differential amplifier 132. The positive
input of the amplifier 132 is coupled to a temperature-compensated
resistor 134, the other terminal of which is coupled to line 130. The
detector 58 and the resistor 134 in effect form the lower half of a
balanced bridge, the upper half being formed by the internal resistance of
each of the inputs of amplifier 132. Amplifier 132 is coupled through a
resistor 135 to the input of a high-gain inverting operational amplifier
136. The output of amplifier 136 is fed to an inverting power amplifier
138 as well as back to its input through a resistor 137 connected in
series with a capacitor 139. The values of capacitor 139 and resistor 135
are such as to produce a time constant of about 30 seconds. Amplifier 136
thus essentially acts as an integrator. Resistor 137 causes amplifier 136
to produce some proportional output to insure stability of the heater
control circuit. Power amplifier 138 is coupled to one terminal of the
heating element 62, the other terminal of which is coupled to a line 140
providing a DC potential of +28 volts.
The above-described circuit constitutes a negative feedback system for
controlling the temperature of heating element 62. Any rise in the
temperature of the detector 58 will result in a corresponding decrease in
the resistance of the detector. As a result, the signal applied to the
negative input of amplifier 132 becomes more negative. Amplifier 132 thus
provides a more positive output, causing amplifier 138 to provide a more
positive output and thus decreasing the power applied to the heating
element 62. Similarly, any decrease in the temperature of the detector 58
will produce a less positive output from amplifier 138, thus increasing
the power supplied to heating element 62.
Differential amplifier 132 is also coupled to the portion of the
signal-processing circuit which recovers the reference-wavelength pulse 68
and the resonant-wavelength pulse 70. More particularly, amplifier 132
drives a noninverting buffer amplifier 144 through a blocking capacitor
142. The output of amplifier 144, constituting a zero-time-average AC
signal having the same wave form as shown in FIG. 6, is applied through a
restoring capacitor 145 to a gated DC level restoring circuit 146. Circuit
146 may be of a conventional type in which a normally nonconductive, or
disabled, gate is selectively enabled by AND circuit 172 to couple line
147 momentarily to ground. Gating pulses applied to the gate input of
circuit 146 at 0.degree. and 180.degree. in the oscillation cycle charge
capacitor 145 to the level attained by the signal during the dark period.
Capacitor 145 thus provides an offset voltage sufficient to restore the
level of the signal during the dark period to zero. The signal on line 147
thus has a wave form identical to that shown in FIG. 6, but with a
sufficient offset to produce a zero output during the dark periods between
pulses.
Circuit 146 drives a pair of sample-and-hold circuits 148 and 150, which
are gated at 270.degree. and 90.degree. in the oscillation cycle to
recover the reference-wavelength and resonant-wavelength signals,
respectively. Sample-and-hold circuits 148 and 150 drive respective
noninverting amplifiers 152 and 154. Amplifier 152 is connected to the
numerator input of a ratio-determining circuit 156, the denominator input
of which is supplied by circuit 154. Circuit 156 provides a ratio signal
on line 158 which may be coupled to a suitable control system or to a
recorder.
The signal-processing circuit also includes a gate pulse generating circuit
indicated generally by the reference numeral 160. In circuit 160 a
noninverting buffer amplifier 162 responsive to the pickup coil 34 drives
one input of a phase-locked oscillator circuit 164. Oscillator circuit 164
is of a conventional type having an internal voltage-controlled oscillator
responsive to a control signal from an internal phase-sensitive detector.
The oscillator portion of circuit 164 is tuned to a nominal frequency of
320 Hz, or four times the frequency of the tuning fork 18. The
phase-sensitive detector portion receives one input from amplifier 162 and
another input from an A2 output of a frequency divider 168.
The output of circuit 164 drives a frequency divider 166 which divides the
input frequency by a factor of two. Circuit 166, which may comprise a
D-type flip-flop, has a normal output A1 and an inverted output A1.
Circuit 166 changes state in response to positive-going transitions of the
signal provided by oscillator 164. The normal output of circuit 166 drives
the second frequency divider 168, which also may comprise a D-type
flip-flop, and which divides the input frequency by another factor of two.
Circuit 168 has a normal output A2 and an inverted input A2. Like circuit
166, circuit 168 changes state in response to positive-going transitions
at its input. The normal output A2 of frequency divider 168 comprises an
80 Hz square wave which, as previously indicated, provides the other input
to the phase-sensitive detector portion of circuit 164.
The phase-sensitive detector portion of circuit 164 provides a control
signal corresponding to the time average of the product of its inputs. The
oscillator portion of circuit 164 decreases or increases its frequency in
response to positive and negative control signals, respectively, so that,
at equilibrium, the A2 signal lags the output from amplifier 162 by
90.degree. phase. Circuit 168 thus changes to logic 1 and 0 at 0.degree.
and 180.degree., respectively. Circuit 166 thus changes to logic 1 at
0.degree. and 180.degree. and to logic 0 at 90.degree. and 270.degree.,
while circuit 164 produces an output that changes from negative to
positive at integral multiples of 90.degree. phase as defined in FIG. 6.
Circuits 166 and 168 gate the output of a one-shot circuit 170, coupled to
oscillator circuit 164, to produce a pulse at integral multiples of
90.degree.. Thus, at 0.degree., circuits 166 and 168 provide A1 and A2
outputs of logic 1 to gate the one-shot pulse through AND gate 172, which
provides a gating pulse to circuit 146. At 90.degree., circuits 166 and
168 provide respective A1 and A2 outputs of logic 1 to gate the one-shot
pulse through AND gate 174 to enable sample-and-hold circuit 150. At
180.degree., circuits 166 and 168 provide respective A1 and A2 outputs of
logic 1 to gate the one-shot pulse through AND gate 172 again. Finally, at
270.degree., counters 166 and 168 provide A1 and A2 outputs of logic 1 to
gate the one-shot pulse through AND circuit 176 to enable sample-and-hold
circuit 148.
It will be understood that certain features and subcombinations are of
utility and may be employed without reference to other features and
subcombinations. This is contemplated by and is within the scope of my
claims. It is further obvious that various changes may be made in details
within the scope of my claims without departing from the spirit of my
invention. For example, while the filters 46 and 48 are preferably
disposed as shown in the radiation path between the source 14 and the web
12, the filters need not be so disposed and may, if desired, be disposed
in the reflected radiation path between the web 12 and the detector 58.
Also, while transmission-type filters are used in the embodiment shown, it
will be appreciated that reflectance-type filters could also be used. It
is, therefore, to be understood that my invention is not to be limited to
the specific details shown and described.
* * * * *
|
|
 |
|
 |
Description |
|
