Inductive Flow Meter

A preferred working example of the invention is explained more fully hereinafter with reference to the drawings wherein:

FIG. 1 shows the schematic design of an inductive flow meter which may be operated in accordance with the present invention.

FIG. 2 shows an electrical block diagram of an evaluation circuit according to the invention.

FIG. 3 shows different voltage curves illustrating the elimination of the voltage drift.

FIG. 4 shows different voltage diagrams to demonstrate the elimination of inductive and capacitive interference voltages.

FIGS. 5 and 6 show electrical block diagrams of alternative, optional circuitry for interference suppression.

As shown in FIG. 1, a measuring tube 10 of non-magnetic material is provided with a magnetic coil or solenoid 11 which generates a magnetic field passing through tube 10. Inside the tube there are provided two electrodes 12, 13 which are diametrically opposed. The imaginary line connecting the electrodes 12, 13 cuts perpendicularly across the magnetic field. Electrodes 12, 13 are connected to the input of an evaluation circuit 14. The measured voltage is read out at a meter or like instrument 15 connected to the evaluation circuit 14. To generate the magnetic field, the magnetic coil 11 is connected to the control circuit 46 which impresses on the magnet winding a current having a specific waveshape. In addition, the control circuit 46 establishes the sample points used by the evaluation circuit 14.

FIG. 2 shows a block diagram of the evaluation circuit 14 and of the control circuit 46. The input lines of the evaluation circuit 14 coming from the electrodes 12, 13 are connected via blocking capacitors 16, 17 with an input terminal of an amplifier 18. The output of the amplifier 18 is connected, via a low-pass filter 19 having a cut-off frequency below 16Hz, to two voltage scanning memories each including an amplifier 20, 21. At the input to each such amplifier, there is provided a holding capacitor 22, 23. Control of the sample points is performed by one or the other of a pair of switches 24, 25 operated by the control circuit 46. Switches 24, 25 are indicated as mechanical switches to simplify the illustration. In a practical implementation of the circuitry, electronic switches would be used.

The output lines of the voltage sample and hold circuit amlifiers 20, 21 are connected to the inputs of a difference amplifier 26, the output of which feeds a mean value former 27. The output of the mean value former 27 is the output of the evaluation circuit 14, and is supplied to the indicator 15.

The control device 46 is operated in synchronism with the 50 Hz power frequency. To this effect, the input of a multistage frequency divider 28 is connected to receive a clock signal at the power frequency. A rectangular voltage having a frequency of 50/64 Hz=0.78 Hz is obtained at one output of the binary frequency divider 28. This rectangular voltage is supplied to an integrator 30 consisting of a capacitive-feedback-coupled amplifier. A triangular voltage is generated at the output of the amplifier 30 which is supplied to a driver amplifier 31 having a high output resistance and being capable of impressing on the coil or solenoid 11 a triangular current with a frequency of 0.78 Hz.

The control of switches 24 and 25 is performed via a pair of AND-gates 32 and 33. In FIG. 2, each of the inverting inputs of the gates 32 and 33 is indicated by a dot; the other inputs are non-inverting.

At the inputs of the gates 32 and 33, the outputs of the frequency divider 28 are so combined that the output signal of each gate 32, 33 exists only in a short time interval within each period of the triangular magnetization current.

The frequency divider 28 could be also considered a counter, and the outputs associated with different counts can be supplied to a digital-to-analog converter which would take the place of the integrator 30.

For a better understanding and with reference to FIG. 3, it is first supposed that switch 24 is always closed on a short-time basis by the gate 32 at the upper peak value of the triangle current, and that the switch 25 is closed by the gate 33 at the lower peak value of the triangular voltage. It must be taken into consideration that the temporal development of the active voltage U.sub.net generated at the electrodes 12 and 13 is equal to the time variation of the triangular current generated at the output of the amplifier 31. To explain how the interference influences are eliminated due to slow changes of the direct voltage potential, suppose for the moment that inductive and capacitive couplings do not occur or are negligible, and that stray parasitic currents from external sources are eliminated by the low-pass filter 19. The remaining interference influences are polarization voltages which change very slowly in time, and the drift of the input amplifier 18.

In FIG. 3, curve (a) shows the development of the active voltage signal U.sub.net. This active voltage signal is superimposed on the interference voltage U.sub.drift shown in (b), composed of the polarization voltages and the amplifier drift, and changing very slowly in time. The combined signal, including both the interference voltage and the active voltage, which occurs at the output of the amplifier 18 is illustrated by curve (c) of FIG. 3. The sample points are shown in (d). These are the times at which the indicated switches 24 and 25 are closed briefly to feed into the voltage sample and hold circuits 20 or 21 the combined signal amplitude values. These values are retained until the next closure of the switch concerned. In FIG. 3 (e) the resultant voltage levels are shown for the two sample and hold circuits 20 and 21, if the input signal develops as shown in (c). It is apparent that stepped curves are formed at the outputs of the two voltage sample and hold circuits 20 and 21 which are offset with respect to each other as to amplitude and time.

In FIG. 3 (f) the temporal development of the output signal of the mean value former 27 is shown. The difference between the output voltages of the amplifier 20 and 21 varies about the mean value 39 in the form of a rectangular curve. This mean value is obtained in the arithmetic means value former 27 which is, for instance, an RC-element. The mean value 39 is proportional to the active voltage U.sub.net. It is readily evident that the rectangular curve in FIG. 3 (f) disappears and becomes the constant voltage 39 if there is no time variable interference voltage such as shown in (b). In such a case, the difference between the upper and the lower peak value of the signal to be evaluated always is constant.

Due to the temporal change of the interference direct voltage, the frequency of the magnet current cannot be substantially below about 1 Hz. Thus, it is still possible that inductive and capacitive interference voltages exist which will alter the measured result. In FIG. 4 (a), the temporal development of the magnetization current I is indicated as corresponding to the time variation of the flux density B and of the wanted signal U.sub.net. The inductive interference voltage, as known, is:

and, accordingly, it is proportional to the time derivative of the magnetization current. The magnetization current being a triangular function, the inductive interference voltage becomes a rectangular function. The same applies to the capacitive interference current I.sub.C, fo which the formula

is applicable. The two interference factors, which are of equal phase and rectangular, can be combined into the interference voltage U.sub.noise shown in FIG. 4 (b).

The combined active-interference signal is shown in FIG. 4 (c). This results from superimposing the triangular active voltage U.sub.net on the rectangular interference voltage U.sub.noise. To extract only the active portion from the combination signal shown in FIG. 4 (c), sampling points 42, 43 shown in FIG. 4 (d) are situated so that the evaluation of the combined signal always occurs only on the rising branch 40 or only on the descending branch 41. In the instant case, the rising branch 40 has been chosen. The inputs of gates 32 and 33 (FIG. 2) are switched so that the output signals 42 of the gate 32 always occur 50 msec after the lower reversal point of the triangle voltage, and so that the output signals 43 of the gate 33 always occur 50 msec before the upper reversal point of the triangle voltage. This ensures that the scanning is performed only on the rising branch 40 of the total signal. It is evident that subsequently, when the difference between voltages 44 and 45 is obtained, the interference voltage drops out because at both sample times it is always of equal amplitude.

The use of a triangular magnetization current is advantageous because the inductive and capacitive interference influences are rectangular and can be eliminated by a suitable selection of the scanning sample points. Moreover, the triangle voltage ensures that transient effects are avoided during the remagnetization. The triangle shape of the magnetization current can be implemented easily, and it can be derived from the power frequency, it being possible to perform the time control of switches 24 and 25 with the same control circuit 46.

FIG. 5 shows a first embodiment 50 of an interference suppression circuit that compensates for interference signals which do not linearly increase or decrease in time. The measurement signal from the electrodes 12, 13 is supplied via the low-pass filter 19 to a first differentiator circuit 51, and thereafter to a second differentiator circuit 52. The differentiators 51, 52 each consist, in a manner known per se, of an operational amplifier 51a, 52a having a feedback resistor 51b, 52b and an input capacitor 51c, 52c.

The output of each differentiator circuit 51, 52 is the time derivative of the input signal to that circuit. Thus, the output of the second differentiator circuit 52 corresponds to the second derivative of the electrode measurement signal supplied from the filter 19. If no interference voltages exist, so that the measurement signal increases or decreases linearly, the first derivative signal at the output of the circuit 51 will be constant, and the second derivative signal at the output of the circuit 52 will be zero. (Brief transients will occur at the output of the circuit 52 in correspondence with the knees or turning points of the triangular or trapezoidal measurement signal.)

If a non-linear interference voltage is superimposed on the measurement voltage, the second derivative of the combined signal will not be zero. In such case, the circuit 52 will produce an output signal indicative of the presence of such a non-linear interference voltage. This signal is supplied to one input of a differential amplifier 53 which receives at its other input a reference voltage set by a potentiometer 53a. This reference voltage establishes a threshold value which must be exceeded by the second derivative signal from the circuit 52 before the amplifier 53 will produce an output. The output of the differential amplifier 53 is supplied via a diode 54 having an output filter capacitor 55 to the control element of a switch 56 that is connected in the output line from the difference amplifier 26. For simplicity of illustration, the switch 56 is shown as a mechanical switch. In practice, an electronic switch such as a field effect transistor would be used. Connected between the switch 56 and the mean value former 27 is a voltage sample and hold memory 57 consisting of a holding capacitor 57a and an amplifier 57b.

If the output of second differentiator circuit 52 surpasses the set threshold value, the switch 56 will be opened so as to interrupt signal flow from the amplifier 26 to the mean value former 27. The last signal amplitude present just before opening of the switch 56 will be stored in the voltage sample and hold memory 57 so that the indicating instrument 15 connected to the mean value former 27 will not change its indication during the time of interruption of the switch 56. When the amplitude of the interference signal decreases below a certain value, the differential amplifier 53 will no longer produce an output signal. The switch 56 then will close, after a slight delay for discharge of the capacitor 55. The difference amplifier 26 then will again be connected to provide an input to the mean value former 27.

A light-emitting diode 59 provides a visual indication that an interference voltage is present. To this end, the output from the differential amplifier 53 and diode 54 also is supplied to a transistor 58 which then turns on current to the diode 59.

The transients which occur at the output of the second differentiator circuit 52 in correspondence with the knees of the triangular or trapezoidal measurement signal may exceed the threshold established at the differential amplifier 53. These transients will be filtered by the capacitor 55. However, even if they should cause opening of the switch 56, it will be for only a short time duration, and this will not effect the measurement indicated by the indicating instrument 15 since the last prior measurement signal will still be provided to the mean value former 27 by the sample and hold memory 57.

An alternative embodiment of an interference suppression circuit is shown in FIG. 6. This circuit 60 makes use of the fact that the output of the difference amplifier 26 (shown in FIG. 3 (f)) contains an alternating voltage component only if an interference signal is present. The amplifier 26 output is a constant or direct voltage if no interference signal is present. In the case of a slowly varying interference signal (drift), the output of the difference amplifier 26 contains a rectangular voltage superimposed on the effective dc signal. If interference voltages of higher frequency are present, they will cause the superposition of a corresponding higher frequency component on the output of the difference amplifier 26.

In the embodiment of FIG. 6, the alternating voltage component, present at the output of the difference amplifier 26 and indicative of an interference signal, is a.c. coupled via a capacitor 61 to an amplifier 62. The output of the amplifier 62 is rectified by a rectifier 63, the output of which is fed to one input of a difference amplifier 64. A reference voltage is supplied via a potentiometer 64a to the other input of the amplifier 64, so that the potentiometer 64a functions to establish a threshold value.

Similar to the example of FIG. 5, a diode 65 and a filter capacitor 66 are connected between the output of the difference amplifier 64 and the control element of the switch 56. Occurrence of an interference signal of sufficient magnitude to overcome the threshold set by the potentiometer 64a likewise causes the switch 56 to open. The last prior signal level from the difference amplifier 26 continues to be supplied to the mean value former 27 via the sample and hold memory 57 so that the indicated signal level is unchanged. Again, a light-emitting diode 59 gives a visual indication of the presence of the interference voltage.

The interference suppression circuit of FIG. 6 has the advantage that both slow drifts as well as high frequency interferences may be recognized and eliminated. The circuit 60 will handle voltages up to the threshold level set for the amplifier 64. The circuits 50 and 60 both have the benefit that the evaluation system will not supply incorrect values in the presence of interference.