GB1602385A

GB1602385A - Self-calibrating processing unit

Info

Publication number: GB1602385A
Application number: GB4775677A
Authority: GB
Original assignee: Water Research Centre
Current assignee: Water Research Centre
Priority date: 1978-05-15
Filing date: 1978-05-15
Publication date: 1981-11-11

Description

(54) IMPROVEMENTS IN AND RELATING TO A SELF-CALIBRATING PROCESSING UNIT (71) We. WATER RESEARCH CENTRE, a British company limited by guarantee, of 45 Station Road, Henley-on-Thames, Oxfordshire RG9 1BW, do hereby declare the invention, for which we pray that a patent may be granted to us, and the method by which it is to be performed to be particularly described in and by the following statement: This invention relates to a self-calibrating processing unit for the output of a sensor which has a linear response. The invention has particular, but not sole, application to the measurement of chemical properties of liquids, particularly in the water industry.

Transducers employed for measuring chemical properties of liquids often have linear characteristics which vary with time. Even so, such transducers generally still exhibit a linear response after extended periods of use. A two-point calibration is therefore sufficient to restore the accuracy of the output measurement data. However, frequent manual cleaning and calibration of sensors is time-consuming and therefore expensive.

This invention provides a self-calibrating processing unit for the output of a sensor which has a linear characteristic. comprising an offset correction amplifier having a transfer function P = y + C followed in series by a gain correction amplifier having a transfer function Q = (P - x1) G +xl, controlling means for comparing the output P of the offset correction amplifier with a reference signal xl when the sensor is applied to a first calibrating reference substance and adjusting the offset C of this amplifier until P = xl, and controlling means for comparing the output Q of the gain correction amplifier with a reference signal x2 when the sensor is applied to a second calibrating reference substance and adjusting the gain G of this amplifier until Q = x2.

An embodiment of this invention will now be described, by way of example only, with reference to the accompanying drawings, in which: Figure 1 is a representation of the electrical equivalent of a poor sensor having a linear characteristic: Figure 2 is a schematic circuit diagram of a self-calibrating processing unit in accordance with this invention: and Figures 3A and 3B joined together form a detailed circuit diagram of a preferred processing unit.

Figure 1 is a representation of the electrical equivalent of a "poor" sensor having a linear characteristic (i.e. the sensor characteristic is subject to variation with time). Thus it is supposed that an input parameter P, (for example dissolved oxygen concentration or pH value) is applied to a perfect probe which feeds a perfect probe amplifier, providing a voltage x = AP + B, where A and B are probe and amplifier constants. Owing to probe and amplifier errors, represented as shown, there is produced an output voltage y = mx + c, where m and c are probe/amplifier error and offset, respectively. For no error, m = 1 and c = 0.If xl. x2 are the equivalent input voltages for respective calibrating fluids or references P and P2, and y1, y2 are the corresponding output voltages, then the general output y is given by:

Figure 2 shows a self-calibrating processing unit in accordance with this invention. The output y from the probe amplifier is applied as input to an offset correction amplifier the output P of which is applied to a gain correction amplifier which provides an output Q. Two potentiometers are provided for presetting low and high calibration reference voltages x, and x2.

The transfer function of the offset correction amplifier is given by: P=y+C where the offset C is programmable or adjustable by an up/down counter 1, as shown, or alternatively by a successive approximation register. The up/down counter 1 is driven by a clock 2 when a switch 3 is closed for a low-point calibration step, and is controlled by the output of a comparator C1 which has its inputs connected to the amplifier output P and to the low-point calibration reference voltage xl.

The transfer function of the gain correction amplifier is given by: Q = (P-x1) G + xl where the gain G is programmable or adjustable by an up/down counter 4 (or alternatively by a successive approximation register). Counter 4 is driven by clock 2 when a switch 5 is closed for a high-point calibration step, and is controlled by the output of a comparator C2 which has it inputs connected to amplifier output Q and to the high-point calibration reference voltage x,.

In operation, the voltages x1 and x2 are determined to provide the required full-scale deflection of a read out meter for the measurements to be performed. These voltages xl and x2 are set by adjusting the two potentiometers. In order to calibrate, or re-calibrate. the unit, the probe is first cleaned and then placed in the low-point calibrating fluid with the switch 3 closed and switch 5 open. with the result (as will be explained) that the offset C of the offset correction amplifier is properly adjusted. Next, the probe is cleaned again and placed in the high-point calibrating fluid with switch 3 open and switch 5 closed, with the result (as will be explained) that the gain G of the gain correction amplifier is properly adjusted. The unit will now be properly calibrated.

It will now be shown how the above steps properly calibrate the unit. It has been mentioned that the transfer function of the offset correction amplifier is: P = y + C (where C is programmable) (1) Also. the transfer function of the gain correction amplifier is: Q = (P-xl)G + xl (where G is programmable) (2) With the probe applied to the low-point calibration fluid and switch 3 closed, the clock will drive the up/down counter 1 until P = xl (hence C = x1 - y1). Therefore, after the low-point calibration step, the general function is: P = xl - yl + y (3) Next, with the probe applied to the high-point calibration fluid and switch 5 closed, up/down counter 4 will be driven by the clock until Q = x2.Also, from (3) it follows that P = xl - yl + y2. Substituting these values of Q and P into (2), we have: x2 = (y2 - y1) G + xl therefore

Therefore, after the high-point calibration, by (2) and (4):

But it has been mentioned, with reference to Figure 1, that in general:

Substituting (6) into (5): Q=x Therefore the olltplll h:ls been properly calibrated.

The above desclibctl sclt-calibration can be supervised manually, i.e. manually closing the switches 3 nnJ 5 in turn by means of a 3-position control, or by automatic control which, at regular intervals. directs the two calibrating fluids in turn to the probe within mechanical fluid handling equipment, whilst simultaneously operating the switches 3 and 5. The self-calibrating processing unit will include a front panel with a meter displaying the corrected output reading of each measurement and also displays of the probe offset and slope errors. The same front panel will include the potentiometer controls. The processing unit may drive a pen recorder which then gives a continuously calibrated output.

It will be noted that the up/down counters effect digital control, whereby in practice counter 1 will firstly adjust the output P to P = xl+1 bit and then counter 4 will adjust the output Q to Q = x.+ 1 bit.

We have built a unit which will perform a two point calibration on an input signal in the range 0-5 volts with an accuracy of + (). / full scale deflection. The unit can automatically calibrate a sensor at any two preselected points within this range so the calibration fluids need not correspond to zero or full scale deflections. The transducer errors can be displayed in analogue or digital form and may be provided as an output in Binary Coded Decimal or analogue form. The unit can compensate for offset errors of + 50% (2.5 volts) and slope errors of 1 50C/c.

A detailed circuit diagram of a preferred unit is shown in figures 3A and 3B when joined together.

WHAT WE CLAIM IS: 1. A self-calibrating processing unit for the output of a sensor which has a linear characteristic, comprising an offset correction amplifier having a transfer function P = y + C followed in series by a gain correction amplifier having a transfer function Q = (P - x1) G + x,. controlling means for comparing the output P of the offset correction amplifier with a reference signal xl when the sensor is applied to a first calibrating reference substance and adjusting the offset C of this amplifier until P = xl, and controlling means for comparing the output Q of the gain correction amplifier with a reference signal x2 when the sensor is applied to a second calibrating reference substance and adjusting the gain G of this amplifier until Q = x2.

2. A processing unit as claimed in claim 1, in which each of said controlling means includes an up/down counter.

3. A processing unit as claimed in claim 1, in which each of said controlling means includes a successive approximation register.

4. A processing unit as claimed in any preceding claim, comprising an automatic control effective firstly to subject the sensor to the first calibrating reference substance and actuate the offset correction controlling means whilst maintaining the gain correction controlling means dormant, and then to subject the sensor to the second calibrating reference substance and actuate the gain correction controlling means whilst maintaining the offset correction controlling means dormant.

5. A self-calibrating processing unit substantially as herein described with reference to Figure 2 of the accompanying drawings.

**WARNING** end of DESC field may overlap start of CLMS **.

Claims

**WARNING** start of CLMS field may overlap end of DESC **. Substituting (6) into (5): Q=x Therefore the olltplll h:ls been properly calibrated. The above desclibctl sclt-calibration can be supervised manually, i.e. manually closing the switches 3 nnJ 5 in turn by means of a 3-position control, or by automatic control which, at regular intervals. directs the two calibrating fluids in turn to the probe within mechanical fluid handling equipment, whilst simultaneously operating the switches 3 and 5. The self-calibrating processing unit will include a front panel with a meter displaying the corrected output reading of each measurement and also displays of the probe offset and slope errors. The same front panel will include the potentiometer controls. The processing unit may drive a pen recorder which then gives a continuously calibrated output. It will be noted that the up/down counters effect digital control, whereby in practice counter 1 will firstly adjust the output P to P = xl+1 bit and then counter 4 will adjust the output Q to Q = x.+ 1 bit. We have built a unit which will perform a two point calibration on an input signal in the range 0-5 volts with an accuracy of + (). / full scale deflection. The unit can automatically calibrate a sensor at any two preselected points within this range so the calibration fluids need not correspond to zero or full scale deflections. The transducer errors can be displayed in analogue or digital form and may be provided as an output in Binary Coded Decimal or analogue form. The unit can compensate for offset errors of + 50% (2.5 volts) and slope errors of 1 50C/c. A detailed circuit diagram of a preferred unit is shown in figures 3A and 3B when joined together. WHAT WE CLAIM IS:

1. A self-calibrating processing unit for the output of a sensor which has a linear characteristic, comprising an offset correction amplifier having a transfer function P = y + C followed in series by a gain correction amplifier having a transfer function Q = (P - x1) G + x,. controlling means for comparing the output P of the offset correction amplifier with a reference signal xl when the sensor is applied to a first calibrating reference substance and adjusting the offset C of this amplifier until P = xl, and controlling means for comparing the output Q of the gain correction amplifier with a reference signal x2 when the sensor is applied to a second calibrating reference substance and adjusting the gain G of this amplifier until Q = x2.