 |
Description  |
|
|
BACKGROUND
The present invention relates to temperature control systems, and more
particularly to fluid temperature control and chemical reaction analyzers
and the like.
In chemical reaction analyzers it is essential to maintain precise
temperature control of the chemical reactions. In some analyzers, it is
also required that a substantial quantity of fluid be fed into a reaction
test cell, the fluid being introduced at the same temperature as the cell
and its contents for preventing fluctuations in the cell temperature when
the fluid is introduced. Further, the temperature control of a plurality
of reaction cells is often required in a single analyzer.
Thus there is a need for a temperature control system that precisely
controls the temperature of a plurality of reaction cells while
simultaneously controlling the temperature of a liquid that is to be fed,
on demand, into one of the cells without upsetting the temperature of the
receiving cell.
SUMMARY
The present invention is directed to a temperature control system that
meets this need. The system includes a thermoelectric element coupled to a
reaction cell, power means for variable bidirectional electrical drive of
the element for selectively heating and cooling the cell in response to a
control signal, sensor means for measuring the temperature of the cell,
and control means for generating the control signal in response to the
sensor means.
In one embodiment, the control means has a first operational mode for
driving the measured temperature to within a coarse temperature range. The
control means also has a second mode in which a predicted temperature
signal is generated, the second mode additionally changing the control
signal by an amount proportional to the difference between the measured
temperature and the desired temperature according to a comparison between
the predicted temperature and the desired temperature. The control means
has a further third mode for maintaining the desired temperature by
limiting the rate of change of the control signal.
In another embodiment, the system includes buffer means for controlling the
temperature of the liquid before the liquid enters the cell, and ambient
sensor means for measuring ambient thermal loading, the control means
compensating the buffer means in response to the ambient sensor means.
DRAWINGS
These and other features, aspects, and advantages of the present invention
will become better understood with reference to the following description,
appended claims, and accompanying drawings where:
FIG. 1 is a fragmentary sectional elevational view of apparatus according
to the present invention;
FIG. 2 is a fragmentary schematic diagram of the apparatus of FIG. 1;
FIG. 3 is a fragmentary elevational view of an alternative configuration of
the apparatus of FIG. 1 within region 3 in FIG. 1;
FIG. 4 is a computer flow diagram for the apparatus of FIG. 1;
FIG. 5 is a computer flow diagram of an alternative configuration of the
diagram of FIG. 4 within region 5 in FIG. 4; and
FIG. 6 is a graph showing operation of the apparatus of FIG. 1.
DESCRIPTION
The present invention is directed to a temperature control system for a
chemical reaction analyzer. With reference to FIG. 1, a reaction analyzer
8 having a temperature control system 10 includes primary and secondary
cells for conducting chemical reactions, designated primary reactor 12 and
secondary reactor 13, the reactors 12 and 13 being connected by respective
conduits 14 and 15 to metering valve means 16. The metering valve means 16
is connected by a tube 17 to a reservoir 18.
A portion of the conduits 14 and 15 form a bifilar coil 20 of a buffer or
incubator unit 22, the coil 20 being attached with close thermal coupling
to an aluminum slug 24 for equalizing the temperature of the coil 20. The
metering valve means 16 is capable, by means not shown, of feeding, on
demand, a selected volume of a feed fluid 26 from the reservoir 18 into
the primary reactor 12 and/or the secondary reactor 13. The purpose of the
incubator unit 22 is to conform the temperature of the feed fluid 26 to a
desired temperature prior to passage into the respective reactors 12 and
13. The coil 20 is appropriately sized for holding at least the greatest
selectable volume of the fluid to be delivered from the metering valve
means 16 to the respective conduits 14 and 15. Thus the entire volume of
the feed fluid 26 that enters each of the reactors 12 and 13 incubates in
the coil 20 for a finite period of time between successive demand feedings
thereof. The analyzer 8 also has means, not shown, for draining the
reactors 12 and 13.
The slug 24 of the incubator unit 22 is attached to a thermoelectric
element 28, the element 28 being thermally bonded on opposite sides
thereof to a heat sink 30 (schematically shown) and the slug 24. The
element 28 comprises dissimilar conductors, such as antimony and bismuth,
having a high Peltier electromotive force (voltage) for selectively
heating and cooling the slug 24 in response to a corresponding direction
of current in the element 28. It should be understood that the heat sink
30 is capable of carrying heat from an external source into the element 28
during heating of the slug 24, as well as carrying heat out of the element
28 during cooling of the slug 24.
A temperature measuring first sensor 32 is bonded to the slug 24 at a
location proximate the element 28 for providing a short response time
between a change in the temperature of the slug 24 at the element 28 and a
corresponding electrical response in the first sensor 32. The first sensor
32 is a conventional thermistor or other suitable device.
A second temperature measuring sensor 34 is provided for permitting control
compensation, described below, responsive to ambient thermal loading. In
one version of the present invention, shown in FIG. 1, the second sensor
34 is bonded to the slug 24 at a location distant from the element 28 for
measuring a temperature gradient in the slug 24.
The first sensor 32 and the second sensor 34 are located proximate to
opposite ends of the coil 20, the first sensor (and the element 28) being
associated with an entrance end 36 of the coil 20, the second sensor being
associated with an exit end 38 of the coil 20. Thus the feed fluid 26
enters the coil 20 at a location proximate to the element 28 for
relatively direct, rapid heat conductivity between the entering feed fluid
26 and the element 28. During progress of the feed fluid 26 through the
coil 20, the temperature of the feed fluid 26 becomes more or less
equalized with the temperature of the slug 24 such that by the time the
feed fluid 26 arrives at the exit end 38 of the coil 20, there is little,
if any, heat transfer between the feed fluid 26 and the slug 24.
It should be understood that the relative positions of the reactors 12 and
13, and the coil 20 are arbitrary in FIG. 1, for clarity. For example, a
terminal portion 37 of the conduit 14, having a length L between the
primary reactor 12 and the exit end 38 of the coil 20, is not necessarily
different in length than a corresponding portion of the conduit 15.
In an alternative embodiment of the present invention, the second sensor,
designated second sensor 34' in FIG. 1, is relocated on a frame member 35
of the reaction analyzer 8. The purpose of relocating the second sensor
34' is to overcome ambient loading of the terminal portion 37 of the
conduit 14 (or conduit 15) between the exit end 38 of the coil 20 and the
primary reactor 12 (or secondary reactor 13). In practical versions of the
analyzer 8, the length L can be up to approximately one foot. Thus ambient
loading of the terminal portion 37 is a significant problem, especially
when there is a large difference between the ambient temperature and a
desired reaction temperature. The location of the second sensor 34' is
selected for providing a response to ambient temperature changes that
approximates the ambient loading of the terminal portion 37 and, to some
degree, the exit end 38 of the coil 20. The second sensor 34' may be
coupled directly to the terminal portion 37; however, it is expected that
temperature fluctuations of the terminal portion 37 during the demand
feeding of the feed fluid 26 would unduly complicate the required
compensation.
In a further version, shown in FIG. 3, the element 28 is relocated on the
slug 24 proximate to the exit end 38 of the coil 20. Thus the feed fluid
26, to the extent that there is a temperature gradient in the slug 24, is
"preequalized", then increasingly close-coupled to the temperature of the
element 28 as the feed fluid 26 approaches the exit end 38 of the coil 20.
Correspondingly, the first sensor 32 is relocated proximate the element 28
and the exit end 38 of the coil 20, and the second sensor 34 is relocated
proximate the entrance end 36 of the coil 20.
The primary reactor 12, shown schematically in FIG. 1, includes a cell body
40, a base 41 of the cell body 40 being equipped with a thermoelectric
element for heating and cooling the reactor 12, the element being
designated element 42. The element 42 is attached to a heat sink 48 as
described above in connection with the incubator unit 22. Similarly, the
primary reactor 12 is equipped with a first sensor 44 at the base 41 and a
second sensor 46 at a location remote from the base 41 for electrically
monitoring respective temperatures of the cell body 40.
Correspondingly, the secondary reactor 13 includes a cell body 50, a base
51 of the cell body 50 being equipped with a thermoelectric element 43 for
heating and cooling the reactor 13. Similarly, the element 43 is attached
to a heat sink 52 as described above in connection with the incubator unit
22. Also, the secondary reactor 13 has a first sensor 45 at the base 51
and a second sensor 47 located remote from the base 51 thereof.
In alternative versions of the present invention, one or more of the second
sensors 34, 46, and/or 47, is omitted, for a reduction in cost and
complexity of the system.
The elements 28, 42, 43, and the sensors 32, 34, and 44-47, are connected
to control electronics 60. Although the electronics 60 may be implemented
in various ways, it is expedient to incorporate a programmable element
such as a computer for a high degree of control flexibility and reduced
hardware development costs. Accordingly, the following description is
based on a microcomputer implementation of electronics 60.
With further reference to FIG. 2, the element 28, and the first and second
sensors 32 and 34 of the incubator unit 22 are connected in the
electronics 60. The elements 42 and 43, and the sensors 44-47 are also
connected to corresponding circuits, not shown, of the electronics 60. The
electronics 60 includes a microprocessor 62 and a buffer-decoder interface
64 for transmitting data and timing signals into and out of the
microprocessor 62.
The microprocessor 62 incorporates appropriate memory circuits, both
programmable and non-volatile (not shown) for storing and executing a
temperature control program as further described below. As shown in FIG.
2, the microprocessor 62 provides 16 data lines, 16 address lines, and a
plurality of control lines for use by the interface 64, the microprocessor
62 comprising commercially available circuits that are known to those
skilled in the art. For example, an "8001" microprocessor circuit suitable
for use in the present invention is available from Zilog Corporation,
Campbell, Calif.
The electronics 60 is connected by a semiconductor device or power switch
70 to a source of DC power 72 for enabling operation of the element 28
when the control program of the microprocessor 62 is operating. For this
purpose, a "heartbeat" signal 74 is decoded and received from the
microprocessor 62 and directed to a retriggerable monostable multivibrator
76 by the interface 64. The multivibrator 76 is connected for controlling
a gate 78 of the switch 70, turning the switch 70 "on" when the
multivibrator is triggered by the heart beat signal 74. As long as the
heart beat signal 74 keeps retriggering the multivibrator 76 under program
control, the switch 72 remains on, connecting the power 72 to a switched
power bus 80.
The interface 64 decodes and directs an incubator current byte 82 from the
microprocessor 62 to a digital to analog converter, designated DAC 84, for
producing a corresponding analog current signal 86. The current signal 86,
buffered by an operational amplifier 88, drives another switch designated
"PosCool" switch 90 through a comparator 92, the comparator 92 being
referenced by an amplifier to a current sensing resistor, designated
positive resistor 96 for limiting a positive or cooling current in the
element 28. The positive resistor 96 is connected between ground and a
"CoolGnd" switch 98 for sinking the cooling current to ground through the
positive resistor 96. The CoolGnd switch 98 is directly controlled from
the amplifier 88 through an amplifier 100, the CoolGnd switch 98 being on
whenever the current byte 82 is positive.
It should be understood that the current byte 82 is a "two's compliment"
number having a range of from -127 to +127. Thus, when the current byte 82
is positive, the element 28 cools the incubator unit 22 at a rate
proportional to the magnitude of the current byte 82.
The amplifier 88 also drives an inverting amplifier 102 for producing a
negative or heating current in the element 28 as described herein. The
inverting amplifier 102 drives another switch, designated "PosHeat" switch
104, through a comparator 106, the comparator 106 being referenced by an
amplifier 108 to a current sensing resistor, designated negative resistor
110, for limiting a negative or heating current in the element 28. The
negative resistor 110 is connected between ground and a "HeatGnd" switch
112 for sinking the heating current to ground through the negative
resistor 110. The HeatGnd switch 112 is directly controlled from the
inverting amplifier 102 through an amplifier 114, the HeatGnd switch 112
being on whenever the current byte 82 is negative. Thus, when the current
byte 82 is negative, the element 28 heats the incubator unit 22 at a rate
proportional to the magnitude of the current byte 82.
In a similar manner, the elements 42 and 43 are connected for heating and
cooling the respective primary and secondary reactors 12 and 13 in
response to a primary current byte 116 and a secondary current byte 117 by
other circuits (not shown) as described above.
The interface 64 also communicates to the microprocessor 62 a twelve-bit
temperature word 118 from an analog to digital converter or ADC 120, the
ADC 120 being controlled by a converter signal 121 from the interface 64.
The ADC 120 receives an analog temperature signal 122 from an analog
multiplexer or MUX 124.
The MUX 124 receives a plurality of input signals 126, producing the analog
temperature signal 122 from a selected one of the input signals 126 as
defined by a three-bit select code 128 from the interface 64.
The first sensor 32 is connected in a sensor circuit 130 for producing a
first sensor signal 132 corresponding to a measured temperature of the
slug 24 proximate to the entrance end 36 of the coil 20. The sensor
circuit 130 includes an operational amplifier 134 having the sensor 32
connected as a feedback element, a reference network 136 supplying an
offset current for producing variable output voltage responsive to a
changing resistance of the first sensor 32. The amplifier 134 drives a
buffer 138 for isolating the first sensor signal 132 from the first sensor
32.
Similarly, the second sensor 34 is connected in a sensor circuit (not
shown) for producing another of the input signals 126 for the MUX 124.
Further, other sensor circuits (not shown) are provided for the first and
second sensors 44, 45, 46, and 47 of the primary and secondary reactors 12
and 13, each of the circuits providing another of the input signals 126
for the MUX 124.
The temperature control program of the microprocessor 62 interacts with the
interface 64 for sampling the input signals 126 from the sensors 32, 34,
and 44-47. The program appropriately generates the incubator, primary, and
secondary current bytes 82, 116, and 117 for driving the thermoelectric
elements 28, 42 and 43 for controlling the temperature of the primary and
secondary reaction cells 12 and 13.
With further reference to FIG. 4, the program communicates with the
interface 64 at intervals of 20 ms for successively controlling the
primary reactor 12, the incubator unit 22, and the secondary reactor 13.
Thus the program operates at a sample interval of 60 ms. The program,
implemented in the conventional Pascal programming language, also permits
other analyzer operations by further means, not shown.
For example, the 3-bit MUX signal 124 and the heartbeat signal 74 are
combined and sent from the microprocessor 62 to the interface 64 as a
single byte of data, the program alternately complementing the
corresponding heartbeat bit every 20 ms for keeping the power switch 70 on
while the program runs as described above.
The program, as shown in FIG. 4, omits provisions for the second sensors 46
and 47 of the primary and secondary reactors 12 and 13. Based on FIG. 4
and the description herein, instatement of these provisions is within the
skill of ordinary programmers.
The program, beginning at MAIN 150, performs an INIT 152 step for setting
variables to appropriate initial values and, from a reentry point 154,
enters a WAIT 156 step for synchronization with an interrupt timer, not
shown, the reentry point being encountered every 20 milliseconds (ms). In
successive reentries, the program is directed to the primary reactor 12,
the incubator unit 22, and the secondary reactor 13, by a channel index H
having respective values 1-3. Thus the program controls each of the
elements 28, 42 and 43, at intervals of 60 ms.
Following WAIT 156, an INITEST 158 step tests whether other initialization,
not shown, has been accomplished. If not, control is transferred to the
reentry point 154; otherwise, a SAVE 160 step temporarily stores previous
values of T[H] and C[H], described herein, as LT[H] and LC[H]. Next, a
READ 162 step interrogates the ADC 120, storing a new value of T[H]. T[H]
corresponds to the measured temperature from the temperature word 118 of
the respective first sensor 32, 44, or 45. In the program, temperature
values range from zero, corresponding to about 32.degree. C., to 4095,
corresponding to about 22.degree. C. The desired temperature nominally has
a value of 2048, corresponding to approximately 27.degree. C. Thus a
change of one degree centigrade is represented by about 409.5 temperature
value increments. For simplicity in the following discussion, temperature
is treated as synonomous with temperature value.
Once T[H] has been obtained, a STARTEST 164 step tests whether control
operation, by means not shown, has been started. If not, control is
transferred to an exit point 165 for indexing the next channel (H) and
calling a SECONDSENS 200 procedure described herein. Otherwise, a SAVE 166
step computes and temporarily stores a deviation value DS[H], a rate value
DT[H], and a direction value SGNDS. DS[H] is the difference between a
setpoint or desired temperature SET[H] and the measured temperature T[H].
DT[H] is the difference between the measured temperature T[H] and the most
recent previously measured temperature LT[H]. SGNDS is set to -1 unless
DS[H] is greater than zero, in which case it is set to +1.
Based on the above, the program computes a control signal current value
C[H] for producing the corresponding current byte 82, 116, or 117, for
each of the elements 28, 42 and 43. In a first mode of operation, when
T[H] is outside a coarse temperature range of 2 to 4093, the associated
C[H] is driven to a maximum control level for full current operation of
the associated thermoelectric element 28, 42 or 43. Thus a COLDTEST 168
step tests whether T[H] is greater than 4093, in which case C[H] is set by
a MAXNEG 170 step to -127, the maximum negative value for the current byte
82; otherwise, a HOTEST 172 step tests whether T[H] is less than 2, in
which case C[H] is set by a MAXPOS 174 step to +127, the maximum positive
value for the current byte 82.
In a second mode of operation, when the first mode is inactive and DS[H] is
outside a fine range R[H] that defines a third mode of operation, if the
magnetude of DS[H] is increasing, then C[H] is increased algebraically by
the product of SGNDS and one-sixteenth the magnetude of DS[H], otherwise,
if the magnetude of DS[H] is less than sixteen times the magnetude of
DT[H], then C[H] is decreased algebraically by the product of SGNDS and
one-sixteenth the magnetude of DS[H]. In this second mode, sixteen times
the magnetude of DT[H] represents a predicted temperature change magnetude
over a time period of sixteen 60 ms time intervals, or about one second.
The comparison with DS[H] is effectively a prediction of whether the
measured temperature will overshoot the desired temperature after a period
of about one second. Thus the second mode changes the control signal in
response to the predicted temperature comparison.
In particular, as further shown in FIG. 4, a negative outcome of HOTEST
172, the first mode being inactive, transfers control to a DECTEST 176
step that tests whether the magnetude of DS[H] is greater than R[H], the
fine range of the third mode of operation, and whether the product of
DS[H] and DT[H] is less than zero. If so, the magnetude of DS[H] is
increasing, T[H] moving away from SET[H]. (If DS[H] is positive and DT[H]
is negative, or vice versa.) Thus an INCREASE 178 step changes C[H] by
algebraically adding the product of SGNDS and one-sixteenth the magnetude
of DS[H] to C[H]. If the result of DECTEST 176 is negative, a PREDICT 180
step tests whether the magnetude of DS[H] is greater than R[H] and the
magnetude of DS[H] is less than sixteen times the magnetude of DT[H]. If
so, a predicted overshoot condition exists, DS[H] decreasing so fast that
T[H] would, at the present rate of DT[H], pass through SET[H] within
sixteen time intervals. Thus a DECREASE 182 step changes C[H] by
algebraically subtracting the product of SGNDS and one-sixteenth the
magnetude of DS[H] from C[H].
By changing C[H] by an amount proportional to DS[H], rather than by a
"predicted DS" based on the predicted temperature change, the second mode
advantageously combines a high degree of control stability with quick
response.
In a third mode, when neither of the first two modes is active and the
magnetude of DS[H] is less than R[H], if DS[H] is less than 20 of R[H],
C[H] is not changed. Otherwise, C[H] is incremented by one in a direction
for decreasing the magnitude of DS[H]. Thus, as shown in FIG. 4, control
is transferred from the MAXNEG 170, the MAXPOS 174, the INCREASE 178, the
DECREASE 182 and the PREDICT 180 (when its outcome is negative) steps to a
RANGETEST 184 step for testing whether the magnetude of DS[H] is less than
R[H]. If so, a SUBRNGTEST 186 step tests whether the magnetude of DS[H] is
within 20 of R[H], in which case an INCREMENT 188 step changes C[H] by the
value of SGNDS, one count only in a direction for decreasing the magnetude
of DS[H]. Otherwise, C[H] is not changed, DS[H] being sufficiently within
R[H] that the control signal need not be changed.
Control is transferred, following the INCREMENT 188 and the RANGETEST 184
(when its outcome is negative) steps, to a CLAMP 190 step for forcing C[H]
to a magnetude of 127, when it otherwise would fall outside of that range,
without changing its polarity. Next, a DIRTEST 192 step tests whether the
polarity of C[H] is opposite the polarity of LC[H] (C[H] and LC[H] being
both non-zero), in which case a PAUSE 194 step sets C[H] equal to zero.
Thus the program prevents sudden changes in the direction of C[H] by
forcing C[H] to zero in any interval that would otherwise produce
successive nonzero values not of the same polarity.
After the DIRTEST 192 (or PAUSE 194), an OUTPUT 196 step sends C[H] to the
interface 64 for generating the appropriate current byte 82, 116, or 117,
then an INDEX 194 step points H the next control channel, and a CALL 196
step invokes the procedure SECONDSENS 200, described below. Finally,
control is transferred to the reentry point 154 and the process is
repeated as described above.
The procedure SECONDSENS 200, invoked by CALL 196 as described above,
compensates SET[2], the desired temperature of the first sensor 32.
SECONDSENS 200 averages, at half-second intervals, 60 samples of the
temperature word 118 from the second sensor 34 over a 30 second interval,
the ten most recent such averages, if within a predetermined stable range,
are averaged as a five-minute running average and, if a gradient between
the sensors is sufficiently small, a settled condition of the incubator 22
is identified. When the settled condition is identified, if SET[2] is not
equal to the running average, the difference between T2 and a setpoint
constant SETC2 (2048) is scaled and used to adjust SET[2] in a direction
dependent on the difference between the average and T[2], the measured
temperature of the first sensor 32.
After the settled condition of the incubator unit 22 is detected,
SECONDSENS 200 continues to update the ten averages, but the settled
condition is not tested until after another five minute time period.
Accordingly, the desired temperature of the first sensor is never adjusted
twice in any five minute period. Thus the program responds quickly to a
settled condition, yet the adjustments are sufficiently spaced in time
that the short term control dynamics are affected only slightly, if at
all.
In particular, with continued reference to FIG. 4, SECONDSENS 200 first
performs a HALFTEST 202 step for testing whether a time period of 0.5
second (25 of the 60 ms intervals) has elapsed. If not, a RETURN 246 step
increments a counter for the 60 ms intervals and transfers control to the
reentry point 154; otherwise, a READ 204 step resets the intervals counter
and interrogates the ADC 120, storing a new value of T2[2]. T2[2]
corresponds to the measured temperature from the temperature word 118 of
the second sensor 34.
Next, a STARTEST 206 Step tests whether the control operation has been
started, as in STARTEST 164. If not, control is transferred to the RETURN
246 Step; otherwise, a SPIKETEST 208 Step tests whether T2[2] differs from
| | |
