This invention relates to integrators for integrating linear voltages,
for example linear voltages which correspond directly to a mechanical movement
that in turn designates a particular process parameter such as flow rate.
It is known to employ voltage-to-frequency conversion and totalisation
techniques in integrators. A prior integrator, known as the Class Y integrator,
supplied by Bailey Controls Company (a division of The Babcock & Wilcox Company),
Wickliffe, Ohio, USA, consists of mechanical gears and pivot arms and uses no electronics.
The Bailey Class Y mechanical integrator suffers from various problems such as
poor reliability, poor versatility, large parts inventory due to the fact that
many versions of the integrator have been provided over several years, mechanical
wear resulting in inaccuracy, specialised labour required for assembly, and specialised
labour required for calibration.
One example of a known transducer which can be used to output a linear
voltage signal is the Type WM Recorder supplied by Bailey Controls Company. This
recorder includes a pen and pointer which designate a percentage range from 0 to
100% and can be used to measure fluid flow as a process parameter. An integrator
can be mechanically coupled to the pointer to provide a running integration of
the flow quantity over time.
US Patent No. US-A-4 380 757 discloses an electronic integrator for
integrating a linear voltage signal over time, the integrator comprising:
   input means for applying the linear voltage signal;
   voltage-to-frequency conversion means for converting the linear
voltage signal to an input frequency signal having a frequency proportional to
the linear voltage signal;
   a binary counter connected to the voltage-to-frequency conversion
means for counting pulses of the input frequency signal; and
   scaling means.
US Patent No. US-A-4 409 660 discloses an electronic integrator in
which an input 4 to 20 mA current signal is passed through a resistor to develop
a voltage signal. The voltage signal is passed via a virtual ground scanning junction
to a voltage-to-frequency conversion means. A pulse signal outputted by the voltage-to-frequency
conversion means is passed via an adjustable binary scaling circuit to a counter.
According to the present invention there is provided an electronic
integrator for integrating a linear voltage signal over time, the integrator comprising:
   input means for applying the linear voltage signal;
   voltage-to-frequency conversion means for converting the linear
voltage signal to an input frequency signal having a frequency proportional to
the linear voltage signal;
   a binary counter connected to the voltage-to-frequency conversion
means for counting pulses of the input frequency signal; and
   scaling means;
   characterised in that:
   the binary counter is operative to generate an operating frequency
signal and a calibrating frequency signal having a frequency which is high relative
to that of the operating frequency signal;
   binary rate multiplying means is connected to the binary counter
for receiving the operating frequency signal and for generating a counting rate
signal which has pulses that are at a scaled proportion of the operating frequency
signal, the binary rate multiplying means having said scaling means for changing
the proportion;
   a pulse counter is provided for counting pulses of the counting
rate signal and for counting pulses of the calibrating frequency signal; and
   switch means is provided having a first input connected to
the binary counter for receiving the calibrating frequency signal, a second input
connected to the binary rate multiplying means for receiving the counting rate
signal, and an output connected to the pulse counter, the switch means being selectively
switchable between a calibration position connecting said first input to the pulse
counter for incrementing the pulse counter rapidly, and an operating position
connecting said second input to the pulse counter for incrementing the pulse counter
more slowly and at a rate proportional to the linear voltage signal.
The binary rate multiplying means preferably comprises a plurality
of binary rate multipliers which are connected in series, each multiplier being
provided with a binary coded digital switch for setting a scaling factor in each
multiplier.
A preferred integrator embodying the present invention and described
hereinbelow is particularly suited to replace the Bailey Class Y mechanical integrator
and can be used, for example, with the Bailey Type WM 55 Recorder. However, the
present invention is in no way limited to this specific application. The present
invention can be used in any situation requiring totalisation (integration) of
a rate signal which is linearly related to a mechanical angular rotation. In addition,
integrators embodying the invention can be used for integration of any linear voltage
signal.
The preferred embodiment of the present invention described hereinbelow
avoids problems of the prior mechanical integrator in that it has high reliability
and versatility, suffers from virtually no mechanical wear and requires only relatively
unskilled personnel for assembly and calibration. The preferred integrator can
be used for an immediate conversion of a mechanical signal to an electrical signal,
resulting in high reliability that is inherent in electrical signal processing.
Adjustable scaling allows one unit to replace most previous integrators. Minimised
mechanical parts results in an extended life expectancy. Minimal skill and equipment
is necessary for field calibration and use of the preferred integrator results
overall in a significant cost reduction.
The preferred electronic or electromechanical integrator is simple
in design, rugged in construction and economical to manufacture, while at the
same time being easily calibrated and accurate.
The invention will now be further described, by way of illustrative
and non-limiting example, with reference to the accompanying drawings, in which:
- Figure 1 is a schematic circuit diagram of an electronic integrator embodying
the present invention; and
- Figure 2 is a plan view of the layout of components of the circuit shown in
Figure 1, showing the physical appearance of the electronic integrator.
The drawings show a preferred embodiment of the invention which comprises
an electronic integrator which can be mechanically linked to a movable arm or linkage
or other member of a transducer, movement of the arm or linkage or other member
being proportional to a process parameter to be measured. The electronic integrator
can, for example be connected to a moving pointer of the Bailey Controls Company
Type WM 55 Recorder mentioned above. Such recorders can be used for measuring flow
rate and the electronic integrator can read the total flow over time and thus
integrate the flow signal.
A power supply for the circuit of Figure 1, comprises a POWER STAGE
providing a single +12V dc supply V+ derived from a 117V ac input
20 applied to a transformer T1 at terminals 4 and 5. The ac power is rectified
by diodes CR2 and CR3 and then filtered and regulated by a capacitor C4 and a
resistor R27, and an integrated circuit U7, respectively. This provides all of
the power needed for the circuit or unit.
An INPUT STAGE of the integrator comprises a potentiometer R30 which
has a shaft thereof, and an lever 10 on the shaft, mechanically connected to a
linear mechanical motion device 30. The potentiometer R30 has two poles thereof
connected across positive and negative voltage supply terminals of the circuit
(shown as V+ and ground). A wiper arm of the potentiometer R30 provides an input
voltage VIN to a VOLTAGE-TO-FREQUENCY (V/F) STAGE which converts voltage
to frequency. This allows for mechanical movement of the motion device 30 to be
converted directly into an electrical representation of that movement, namely the
voltage VIN. The VOLTAGE-TO-FREQUENCY STAGE comprises potentiometers
R29 and R5, an LM 331 V/F converter U1, resistors R2, R3, R4, R6, R7, R28, part
of a resistor array R8 and capacitors C1, C2 and C3, all connected as shown in
Figure 1. The VOLTAGE-TO-FREQUENCY STAGE receives the above-mentioned linear electrical
signal (the voltage VIN) and produces a corresponding frequency output
fOUT' after proper adjustment of the potentiometers R29 (zero adjustment)
and R5 (span adjustment). The zero adjustment permits the variations between individual
components of the INPUT STAGE to be compensated for.
The zero voltage is the level at which the V/F STAGE produces no
output pulses if presented with a voltage less than or equal to that of the zero
adjustment. The span adjustment sets the voltage from the potentiometer R30 to
1220 mV ±10 mV. This corresponds to the maximum span voltage of the illustrated
integrator, but could be any voltage within the operating conditions of the LM
331 V/F converter U1 which is used in the VOLTAGE-TO-FREQUENCY STAGE. The frequency
ouput of the converter U1 is set by adjusting the potentiometor R5 to achieve a
period of 440± 20 microseconds at a port (connection) TP3 of a circuit U2 in a
following CALIBRATION AND SCALING STAGE. The ouput of the circuit U2 (which circuit
is a first stage binary counter) will, as a result of this adjustment, be a frequency
of 1.11 Hz at full scale input from the INPUT STAGE.
The frequency of 1.11 Hz is measured at a pin 1 of the circuit U2
and serves for establishing a full scale reading of 1.11 Hz for the scaling section
of the integrator. The value of 1.11 Hz is significant in that it will allow the
user to input a decimal scaling factor directly by movement of switches SW2, SW3
and SW4. The calibration accuracy and time it takes to calibrate the integrator
are both improved by using an output having a higher frequency fCAL,
derived by the same circuit U2, for a calibration input.
A pin 13 of the circuit U2 provides the output at the frequency fCAL
to
an output pulse counter M1 through a switch SW1, an amplifier comprising driver
transistors Q1 and Q2, a diode CR1 and resistors R24, R25 and R26, all connected
as shown, and terminals 2 and 3. This results in a count rate on the pulse counter
M1, which is an electromechanical counter, of 1066 counts per minute. This allows
for high accuracy calibration within a comparatively short amount of time compared
to normal count rate times of the integrator in service, which may vary from one
to 999 counts per hour as set by the scaling factor. The switch SW1 is a double-throw,
double-pole, centre OFF switch which either disables power to the CALIBRATION AND
SCALING STAGE while input V/F adjustments are being made, or is used to put the
integrator in operate (OPER) or calibrate (CALIB) positions. The calibrate (calibration)
position allows the scaling section to be by-passed and the faster rate ouput at
the frequency fCAL from the pin 13 of the circuit U2 to be applied to
the output pulse counter M1 as described above. Placing the switch SW1 in the operate
(operating) position will allow the counter M1 to receive pulses from the scaling
section of the integrator. This section allows an easy method of scaling an input
to read appropriately on the counter M1 of the integrator.
Rather than a complex set of changeable cams, as employed in the
prior mechanical integrator, the present integrator uses binary coded decimal (BCD)
switches, namely the switches SW2, SW3 and SW4, to directly input the scaling factor.
Each of the switches SW2, SW3 and SW4 is connected to a circuit U3, U4 and U5,
respectively, the junctions between respective switch elements of each switch and
the associated circuit being connected to ground via respective parts of the resistor
array R8 and of a resistor array R9 in the manner shown in Figure 1. The circuits
U3, U4 and U5 are connected in series. Each of the circuits U3, U4 and U5 is a
4527 binary rate multiplier.
Entry of a number by the user by means of the switches SW2, SW3 and
SW4, in such a manner that the setting of the switch SW2 represents hundreds of
units, the setting of the switch SW3 represents tens of units and the setting of
the switch SW4 represents units, provides for a scale selection between one and
999. This is to represent the maximum number of units per hour of the integrator
for a full scale input. The switches SW2, SW3 and SW4 allow the user to enter a
decimal number, and the number is converted by the switches to a BCD format for
the binary rate multipliers U3, U4 and U5. The binary rate multipliers U3, U4 and
U5 are arranged so that the output of the circuit U2 (as calibrated above) is applied
as an input to the multiplier U3. Then, an output of the multiplier U3 is presented
to the multiplier U4 in proportion to the setting of the switch SW2. An output
of the multiplier U3 goes into the multiplier U4 and is proportionately scaled
by the setting of the switch SW3. Lastly, the switch SW4 scales the output from
the multiplier U4 as it is counted through the multiplier U5. The result is a scaled
output of a frequency which is finally divided by four by a binary counter U6 before
being presented to the pulse counter M1 through the switch SW1 in its operate (OPER)
position.
Pin 4 of each of the multipliers U3, U4 and U5 is connected to the
junction of a network comprising a capacitor C5 and part of the resistor array
R8.
As shown in Figure 2, the counter M1 has a numeric display for visually
displaying a number corresponding to the integrated linear voltage signal VIN.
Figure 2 also shows the shaft and arm 10 of the potentiometer R30 which can be
mechanically linked to the movable member of a transducer such as the pointer of
the flow recorder mentioned above. Figure 2 also shows the relative positions of
the various elements shown in the circuit of Figure 1.
A particularly important feature of the present integrator is the
CALIBRATION AND SCALING STAGE shown in Figure 1. The INPUT STAGE may provide any
transduced signal with a linear voltage output and is not limited to a potentiometer
per se. The VOLTAGE-TO-FREQUENCY STAGE is conventional and can be replaced by an
equivalent circuit capable of converting a linear voltage into a frequency signal
with a frequency proportional to the linear voltage. By using the calibration and
scaling circuitry of the present integrator, all analog parameters except for the
input signal remains static after initial calibration. This means that the output
(counting rate) signal fOUT at full scale remains constant regardless
of the scale selection effected by using the switches SW2, SW3 and SW4. Any calibration
method which deviates fOUT can be shown to introduce analog error beyond
that of the present circuit.
Because the counting rate signal fOUT at full scale is
static, it is possible to optimise the counting rate signal with respect to accuracy,
linearity and calibration criteria.
Digital scaling results in an overall accuracy which is independent
of scale, and the ability to change scales without the need for recalibration.
The full scale frequency of the integrator may be calibrated to within
0.1% of full scale using only a time reference with an accuracy and resolution
typical of a common electronic wristwatch including a quartz timer with 0.1 second
resolution or better. This last advantage is achieved by making the counting rate
signal fOUT significantly larger than a typical calibration interval
(e.g. perhaps 1 to 5 minutes). The overriding source of error thus becomes the
uncertainty in the interval itself.
All of the advantages set forth above are achieved through the use
of the circuitry shown in Figure 1. The double-throw switch SW1 is used for calibration.
This makes the design very cost efficient, yet still versatile. The integrator
can be calibrated utilising a common wristwatch, which enhances its value for use
in field applications. In other words, when the switch SW1 is switched over to
its calibrate (CALIB) position, the high frequency signal fCAL is applied
to the pulse counter M1, changing the numerical display of the counter rapidly.
Calibration is achieved by observing the count of the counter and checking the
time it takes to reach that count. Adjustment can be made using the zero set and
scale set potentiometers R29 and R5 in the VOLTAGE-TO-FREQUENCY STAGE.
The INPUT STAGE may be any transducer or structure which generates
a useful linear voltage signal. Any voltage-to-frequency converter can replace
the one shown and the signal fOUT can be optimised for the particular
converter used.
The circuits U2 to U6 are digital counters and can be represented
in general by the transfer functions:
fOUT' = n/m(fOUT)
where n and m are integers and m is greater than or equal to n; and
fCAL = k(fOUT)
where k is an integer and is useful in the case where k is much greater than n/m
for the purpose of calibration.
The connection (port) TP3 and further connections TP1, TP2 and TP4
shown in Figure 1 are used during installation of the integrator to help make
initial settings for an initial calibration of the integrator to suit a particular
transducer. For example, when using the integrator with the Type WM55 Recorder,
a voltmeter having an accuracy of 0.1mV is connected to the connections TP1 (positive
terminal) and TP4 (negative terminal). The INPUT STAGE potentiometer R30 is then
turned until a reading of 200mV ± 50mV is reached. This procedure is followed with
the calibration switch SW1 in its OFF or neutral position. Power is thus removed
from all circuitry except for the input and zero potentiometers R29 and R30. This
sets an initial position for the lever 10 which can then be secured to the shaft
of the potentiometer R30. The negative voltmeter probe can then be moved from
the connection TP4 to the connection TP2. In this position, the potentiometer R29
is rotated until a reading of -3mV ± 1mV is read on the voltmeter.
Inputs of 50% and 100% are then applied artificially by rotating the
lever 10 and the voltage measurements are recorded. If the voltage span from 0
to 50% is less than the voltage span from 50 to 100% by more than 10mV, an adjustment
must be made in the linkage between the lever arm 10 and the pointer of the recorder
(not shown). Similar adjustments are necessary if the span from 0 to 50% is greater
than the span from 50 to 100% by more than 10mV.
To calibrate the integrator once the lever arm 10 and the potentiometer
R29 have been set as noted above, the switch SW1 is switched to its calibrate position
for 15 seconds. Beforehand, the number indicated in the display of the counter
M1 is noted. The 15 seconds can be verified by using a quartz watch with a sweeping
second hand. After 15 seconds, the switch SW1 is returned to its neutral or OFF
position. With the switch SW1 in the calibrate position, the counter operates at
high frequency using the signal fCAL from the counter U2.
The count now appearing on the counter M1 is then recorded and subtracted
from the previous reading to give the count rate for the 15 second period. If the
count rate is less than 133, the span or range potentiometer R5 is rotated clockwise
to increase the counting speed. If the rate is more than 133, the potentiometer
R5 is turned anticlockwise to decrease the speed.
For a course adjustment, the foregoing steps are repeated until a
15 second count of 133 ± 1 is achieved.
For a finer adjustment, a longer period can be taken. For example,
a count of 533 ± 1 should result if the switch SW1 is placed in the calibrate
position for 60 seconds.
As noted above, the scale is selected by manipulating the switches
SW2, SW3 and SW4, each being in the BCD format so that each can be switched for
providing a number from 0 to 9, and thus from 100 to 900 for the muliplier U3,
from 10 to 90 for the multiplier U4 and from 1 to 9 for the multiplier U5.