CS219315B2 - Method and apparatus for a transducer - Google Patents

Abstract

The invention relates to a method of correcting the output signals of a digital converter that converts physical quantities using a pulse train, wherein the characteristic of the converter frequency as a function of the physical quantity is linear in the working area and outside the working area (of the converter it intersects the coordinate indicating the frequency of the converter outside the origin of the coordinates, wherein the essence of the invention lies in the fact that the correction frequency is superimposed on the converter frequency (fg) in order to obtain a corrected converter frequency (fgkorr), wherein the correction frequency (fk) is set so that the resulting characteristic passes through the origin of the coordinates, wherein the correction pulses are fed to the device after the converter pulse has passed, while the converter and correction pulses that are identical or close in time are separated into two different pulses.

Description

The invention relates to a method and device for correcting the output signals of a digital converter that converts physical quantities using a pulse train, wherein the converter frequency characteristic as a function of the monitored physical quantity is linear in the working area and outside the working area of the converter intersects the coordinate indicating the frequency of the converter outside the origin of the coordinates.

In many cases, accurately working measuring instruments provide output signals that are in a linear relationship with the measured quantity. This is called the measuring characteristic of the measuring instrument, that is, a curve that shows the signal in relation to the measured values represented by a straight line. As a result of friction, some other physical influences and some properties of the measuring instruments, the output signals are actually inversely proportional to the measured variables. In other words, the characteristic of the measuring instrument, or its extension, does not occur at the origin of the coordinates of the graphic representation, but somewhat to the side. This fact causes difficulties especially when the converter sends a pulse train that is counted and registered by a adding device or a computer that works proportionally.

The invention aims to create a method and device by means of which it is possible to correct the sequence of pulses transmitted by the converter so that this sequence is directly and precisely proportional to the measured values in the entire working range of the converter.

This task is fulfilled and the shortcomings of known methods and devices are eliminated by the method of correcting the output signals of a digital converter according to the invention, the essence of which is that the correction frequency is superimposed on the converter frequency in order to obtain a corrected converter frequency, the correction frequency being set so that the resulting characteristic passes through the origin of coordinates, the correction pulses being fed to the device after the converter pulse has passed, while the converter and correction pulses that are identical or close in time are separated into two different pulses.

According to a further feature of the invention, the correction pulses that are superimposed on the converter pulse are limited to a maximum number if the frequency of the converter pulses falls below the minimum value of the operating range.

Furthermore, after each converter pulse, correction pulses are transmitted for a period greater than t _max = 1/fnnn, or after each converter pulse, correction pulses are transmitted with a number less than n = fk/fmin. Another feature of the invention is that after each converter pulse, correction pulses are transmitted for superposition, the number of which n is at most fk/fmin or equal to an integer N and K directly above or below this value, whereby the transmission of N and K correction pulses is alternately controlled so that the average value of the total number of these pulses per each correction pulse is equal to the number n.

The above method of correcting the output signals of a digital converter according to the invention can be carried out by a device comprising a pulse generator for producing correction pulses, also according to the invention, the essence of which lies in the fact that the correction pulse generator is connected to the first input of a gate, to the second input of which the converter is connected, while the output of the gate is connected to the first input of a correction circuit, to the second input of which the output of the converter is connected.

According to another feature of the device according to the invention, the gate comprises a filter, the output of which is connected to a product gate.

Furthermore, according to the invention, the correction circuit is formed by a combined: dividing and adding circuit, which contains a summing gate, connected by its one input to the output of the converter and by its second input to the output of the correction pulses of the correction circuit and by its output to the inputs of two pulse generators, wherein the first pulse generator is connected to the output of the combined dividing and adding circuit and the second pulse generator is connected by its output to the summing gate via the converter and the first product gate and simultaneously via the second product gate and a memory which is connected to the second pulse generator.

In addition, another feature is that converters of the supplied signals into needle pulses are connected before the inputs of the summing gate.

The advantage of the method and device according to the invention is that the circuit of the generator for correction pulses can be set to influence the frequency of correction pulses, and thus compensate for changed external conditions. This then makes it possible to eliminate the effects on the measured results. For example, in a flow meter, a change in temperature can affect the viscosity of the flowing liquid, and thus cause a measurement error. This error can then be corrected by the above-mentioned circuit, while the correction pulses are dependent on the temperature. Furthermore, it is possible to eliminate various manufacturing differences of converters of the same type by changing the frequency of correction pulses and to perform their most suitable settings.

The invention will be further described with the aid of drawings, where Fig. 1 shows an example of a linear measuring characteristic indicating the dependence of the frequency of the converter pulses on the measured quantity, Fig. 2 shows the correction of the measuring characteristic according to the invention, Fig. 3 shows the impulse diagram of superimposed converter and correction pulses according to the invention, Fig. 4 shows a simple block diagram of the correction circuit for carrying out the method according to the invention, Figs. 5 and 6 show the impulse diagram of superimposed converter and correction pulses corresponding to two alternative ways of carrying out the circuit according to Fig. 4, Fig. 7 shows the block diagram according to Fig. 4 in a detailed representation, Fig. 8 shows the impulse diagram from which the shape of the converter and correction pulses is derived, Fig. 9 shows the impulse diagram with superposition of converter and correction pulses, Fig. 10 shows the block diagram a separating and adding device, which is a detail from the circuit according to Fig. 7.

The disadvantages of the prior art, which was mentioned above, are evident from Fig. 1. Fig. 1 shows the measuring characteristic of the device in which the present invention can be used. This measuring device is a measuring transducer that produces a pulse train, the frequency of the pulses being directly proportional to the measured quantity. As an example, let us take a flow meter for measuring the flow rate q (m ³ /sec), in which the transducer is represented by a rotor or other rotating body. The frequency (Hz) of the rotor is then measured in a known manner. As a characteristic, we then obtain a straight line, shown in the figure as A. The origin of the straight line A is given by the coordinates f _min and Q min. This value shows the smallest flow rate that can be measured in practice. In other words, q _min indicates the lower limit of the measured working range. If we extend this straight line A (dashed line), then this intersects the vertical coordinate at the point p representing the hydraulic losses. It should be noted that there are also cases in which the linear characteristic intersects the vertical axis i above the origin of the coordinates, as shown in Fig. 1 by the straight line A* and the point p*. This case ⁱ is typical for various physical measuring transducers. As far as flow meters are concerned, this phenomenon applies to a hydrodynamic oscillator, for example a Wirkle flow meter, which has no moving parts and whose measuring characteristic passes above the origin of the coordinates. In the context of the invention, it is not essential whether the measuring characteristic intersects the vertical coordinate above or below the origin of the coordinates at the point p or p', respectively, which will be derived in the following text.

To reproduce the frequency f, an integrated, directly indicating device is used, for example a summation counter of converter pulses. The pulses are processed in this counter in a known manner. Since the line A is a straight line, the pulse frequency f is directly proportional to the flow rate q. The line A does not pass through the origin of the coordinate system, which means that the pulse frequency f is not directly proportional to the flow rate q. The counter registers the received, proportional signals, while its own characteristic passes through the origin of the coordinates. To compensate for this deviation, the counter is set in the normal way so that its characteristic (line S in Fig. 1) has such a slope a that it intersects the characteristic A of the converter at a point given by the coordinate q _med , which indicates the center of the desired working area. This means that the proportionality factor of the counter is

WITH

K = tan a.

However, this method causes errors in the measurement which are transmitted to the working area. Since in many cases the absolute magnitude of the errors is limited, the working area is narrowed to only a very small section ai (Fig. 1). This of course causes great difficulties in practice.

However, the invention allows for complete correction of the transducer signals. This means that the transducer signals are directly proportional to the measured quantity without affecting the accuracy of the measuring instrument. This* correction is achieved on the basis of a simple principle, according to which the transducer frequency is superimposed on a certain adjustable correction frequency, the characteristic of which has a positive slope. After the shift, the extension of the characteristic then passes through the origin of the coordinates. This relationship is shown in Fig. 2, where the uncorrected frequency characteristic of the transducer is shown by the straight line f _g , which corresponds to the straight line A of Fig. 1, while the corrected frequency characteristic is shown by the straight line fg _corr . The constant frequency, which is added to the transducer frequency, is given by the point f _k , so that:

fgcorr '= fg + fk ·

Within the working area, which is shown in Fig. 2 by line a from the lower value f _min to the desired upper limit f _max , the corrected frequency fg _cor r is directly proportional to the measured flow rate q. It follows from the above relations that it is not decisive at what point above or below the origin of the coordinates the uncorrected linear characteristic intersects the vertical coordinate. The term superposition and its derivation are used in this description and definition of the subject matter of the invention in its generally valid mathematical meaning.

To achieve the desired parallel shift of the converter characteristic, the pulse generator according to the method of the invention sends pulses with the desired correction frequency if _k . The pulse generator is then connected to the output of the converter. However, for this connection, certain conditions must be met in order to achieve a parallel shift of the converter characteristic, namely: no correction pulses may be sent at a time when no converter pulses are supplied, and this condition can also be formulated as follows: correction pulses may only be supplied after the converter pulses have been sent; furthermore, converter and correction pulses may not be supplied simultaneously, because the counter always registers only one pulse; therefore, correction pulses that are supplied simultaneously with the converter pulses must be divided so that the counter can register them; furthermore, it is not permissible to send correction pulses after setting the measuring converter or after exceeding the lower limit f _min of the working range, i.e., each converter pulse the impulse must be limited by the following correction impulse outside the lower limit.

These conditions will be discussed further in connection with the description of Fig. 3, which shows the converter and correction pulses in time sequence. The converter pulses are shown at the a coordinate, the correction pulses at the b coordinate and the resulting superposition of both pulses at the c coordinate. Fig. 3 shows the case in which the converter frequencies approach the lower value, with gi being the initial and g2 being the final pulse, which are transmitted with a frequency of f _min . The g3 pulse is introduced — if possible later — outside the actual working area. Correction pulses ki ... k„ are transmitted at regular time intervals corresponding to the respective reception. Due to the above conditions, the sequence of correction pulses after the last converter pulse g2 must be interrupted, and this interruption is carried out after the conversion of a certain number of correction pulses, corresponding to the time interval between the initial pulse gi and the final pulse g2. If the lowest converter frequency has the value f _min , the maximum time interval between converter pulses has the value f _max , which means that it is at the boundary of the working area, so the equation applies:

f = -1 ^A max «

Imine

If the correction frequency has the value f _k , then the number n of correction pulses during the time interval is given by the equation u = t _rnílx . f _k or by the equation

Imine

In Fig. 3, the perpendicular dashed line s shows that the transmission of n correction pulses, which are transmitted after the last converter pulse g2, is to be interrupted. Above the coordinate c in Fig. 3, the corrected pulse sequence is then shown immediately before the converter adjustment.

From the above it follows that f _min generally represents the lower limit of the linear characteristic of the converter. As already indicated, in certain cases it is possible to actually transmit converter pulses with a lower frequency than f _min , which, however, occurs in a nonlinear region outside the working range. In the used example of a flow meter, the medium may have such a high viscosity that at a reduced flow rate the converter transmits pulses that deviate from the linear characteristic and end at the point f _min , as can be seen from Fig. 2. In the practical application of this invention, f _mln is always calculated so that this point is at the end of the linear part of the characteristic. In the principle of the invention, nothing changes if the output signal of the converter is multiplied by a certain factor. The frequency f _k is then multiplied by a factor in the equation does not change.

Fig. 4 shows a block diagram of the electronic circuit by means of which the above-mentioned method according to the invention is implemented. The converter 1, as described above, transmits a sequence of pulses with a frequency proportional to the measured quantity, in our case the flow rate through the flow meter. The output of the converter 1 is connected to the input of the correction circuit 2 and simultaneously to the control input of the gate 3. The output of the gate 3 is connected to the second input of the correction circuit 2. The pulse generator 4 for producing the correction pulses is connected to the second control input of the gate 3. For simplicity, we present here these four basic circuit circuits, although in practice all these circuits can be combined into a single integrated circuit.

In accordance with the above, the pulses of the converter 1 and the correction pulses of the generator 4 are superimposed in the correction circuit 2 and transferred to a counter (not shown). As mentioned above, a simple superposition of the pulses is not performed, and the correction circuit 2 and the gate 3 have the task of maintaining only the stated conditions for superposition.

The so-called dividing condition will be described below, which stipulates that two pulses from the converter 1 and from the pulse generator 4 must not be fed as one pulse for further processing, even if they are transmitted simultaneously. Both of these pulses must therefore be divided. This condition is implemented by a correction circuit that contains a memory that records the simultaneously fed pulses and then passes them one after the other. If a pulse is sent at the same time as another pulse is fed to the correction circuit 2, then this new pulse is included in the memory in the sequence. This new pulse is fed to the counter only when a certain time interval has elapsed after the previous pulse has been sent. Thanks to this measure, the division of the two subsequent pulses fed to the counter is sufficient and the counter registers them with certainty.

The actual superposition is then carried out in the gate circuit. The coordinates between the two pulse sequences require that the correction pulses are not supplied arbitrarily, but continuously depending on the first incoming converter pulse, whereby the correction pulses are limited if the converter pulses occur so rarely that the lower limit of the measuring range is exceeded. The preferential setting of the condition for the converter pulses is ensured by the gate 3, which deflects the correction pulses and is opened only by the action of the converter pulses. For this case, the gate 3 is in such a state that the correction pulse sequence is interrupted if no converter pulse is supplied, i.e. if no further pulse g3 follows the final converter pulse g2. If at least one such late pulse is nevertheless supplied, only a certain number of correction pulses can be sent for this case. The gate can operate in two different ways, i.e. it reacts to the measurement time or to the number of pulses.

Gate 3 operates in such a way that it is controlled by each converter pulse and is open for a certain maximum time interval t _max . This time interval corresponds to the value f _min . If the converter frequency decreases above the value f min , then gate 3 remains open.

This condition is shown in Fig. 5, where the output signal of the converter 1 is shown at the coordinate a and the action of the gate at the coordinate a. The time interval t _max for which the gate 3 is open is shown at the coordinate a. The correction pulses are in this case shown at the coordinate b and the resulting output signal of the correction circuit is shown at the coordinate c. It should be noted that the first correction pulse ki passes through the gate only after the initial converter pulse gi has passed. Another condition is that the final converter pulse gž' must be applied somewhat later so that the gate has time to close. The pulse gž' then opens the gate again, so that the next correction pulse can pass. However, the next correction pulse k3 cannot pass, because the gate has again had enough time to close. Everything is then repeated again with the pulse g3.

Fig. 6 shows the function of a gate that performs pulse counting. This means that a certain number of correction pulses are always allowed to pass. This number is given by the equation:

f _k ^{11 =} fmin

In this case n = 3. As in Fig. 3, the converter pulses are shown at the a coordinate, the correction pulses at the b coordinate and the superimposed pulses at the c coordinate. The converter pulse g i opens the gate, then three correction pulses pass through the gate. Next, the gate input, which is shown by the vertical line s, closes because no further converter pulse passes through. If a further converter pulse g3 appears at the gate input later, it itself provides three correction pulses. This phenomenon is shown by the dashed line to the right of the vertical line s. In this connection, it should be noted, as can be seen from Fig. 3 and also from Fig. 6, that if the frequency of the correction pulses is greater than the frequency of the converter pulses, this results in greater accuracy. In practice, however, both frequencies are approximately the same, that is, n 1, sometimes even f _{k i} is smaller than f _g .

The method described above represents a certain approximate solution, because the number of pulses passing through the gate may not always correspond to the predetermined number n = f _k /fmin· The quotient must then be rounded to the nearest positive number. Then, without much difficulty, it is possible to set the gate so that the average number of passing correction pulses reaches the value n.

For this purpose, the gate is provided with a memory which allows a number of correction pulses to pass through which is sometimes greater and sometimes less than the average number n, so that this average number approaches the desired value n with a larger number of pulses.

This principle will be explained with a simple numerical example. For example, let us choose the quotient fk/fmin, that is, the number of correction pulses 2.6, expressed as a fraction 13/5. The gate will then pass three correction pulses three times and two correction pulses twice, based on five waveforms. The average value of this series is 2.6 correction pulses.

Fig. 7 to 10 show a practical example of the circuit of Fig. 4. Fig. 7 shows a circuit where the individual basic elements of the circuit shown in Fig. 4 are marked with a dashed line. The converter 10 outputs a signal 12, the size of which is given by the elements of the converter. The signal 12 is converted into a pulse in the amplifier and pulse shaper 14. In the correction generator 24, a correction frequency 22 is produced, which is set so that the correction pulses, when superimposed on the converter pulses, ensure a shift of the converter characteristic to the origin of coordinates, as previously described.

Fig. 8 shows the case with different pulses, the upper part of the figure showing the input signal 12 which is transformed into a rectangular pulse 16 by means of a pulse shaper 14. The pulse generator 24 supplies the correction signals 22 in rectangular form. The pulses can be defined as a series of potential changes, i.e. the signal fluctuates between a lower and an upper level, as shown in Fig. 8 by the values H and L. The duration of the pulses is understood as the time during which the signal is at the upper value H.

The above-mentioned gate 3 comprises, according to Fig. 7, a filter 18 connected to a product gate 28. This combination registers the priority of the converter pulses, that is, the correction pulses can only pass if the initial converter pulse has already passed. The converter pulses 16 are fed from the amplifier and pulse shaper 14 to the above-mentioned correction circuit 2, which is formed by a divider and a counter 20, which will be described in detail in Fig.

10. Each pulse is also controlled by a filter 18 which passes a signal with a specified maximum duration, corresponding to a predetermined time t _max , the filter 18 then supplies a continuous signal, i.e. the output remains at the upper level. This condition is shown in Fig. 9, where the time sequence of the various signals from Fig. 7 is shown. For better clarity, the coordinates in Fig. 9 are extended. First, we consider that no converter pulse is supplied. Signal 26 is at the lower level (level 0). When the converter pulse 16' appears, this pulse starts to control the pass filter, so that its signal level is brought to a higher value 26 ^t , which is maintained by the influence of the following converter pulse 16". If this following pulse 16" is not brought, the signal 26 drops to zero after a time interval t _max , as shown by the dashed line in Fig. 9. The signal 26 is brought to the product gate 28, to which the correction pulse 22 is also brought. In order for these pulses to pass through the gate 28, they must be at the upper level. This is easily done with gates of the type used. At the output of the gate 28, pulses 30 corresponding to the pulses 22 of the correction generator 24 then appear as long as the signal 26' is at the upper level. The resulting superposition is derived in Fig.

9. The correction pulses 22 cannot pass through the product gate 28 if the signal 26 is at a low level when the correction signals 22 are applied. The pulses 30 are applied to the summing circuit 20 if they are separated in time. If they pass together or are close to each other, they must be separated. The resulting output signal 32 is shaped as shown in Fig. 9, that is, the first converter pulse 16* corresponds to the first output pulse 32', the second output pulse 32" corresponds to the correction pulse 30 ^" and the third output pulse ³² " then corresponds to the final converter pulse 16" etc.

The separating and adding circuit 20, as shown in Fig. 10, is formed by connecting converters 40, 42, 44, a summing gate 46, two pulse generators 48, 50, a converter 52, a memory 54 and two product gates 56, 58. The circuit 20 processes the converter pulses 16 and the correction pulses 30, which it then processes into output pulses 32. The circuit 20 is set up so that it reacts to the leading edge of the received pulses, i.e. to a potential jump from a lower to a higher value. When the converter or correction pulses 16, 30 arrive at the circuit 20, they are passed through the converters 40, 42, where they are converted into needle pulses, the edges of which coincide in time with the edges of the supplied pulses. These short-term pulses then pass through the sum gate 46 and enter the two pulse generators 48, 50, which then send a pulse. The duration T2 of the pulse of the generator 50 is greater than the duration T1 of the pulse of the generator 48. After setting the two generators, their output signals are at the upper level and if another im12 pulse is now supplied, this previously supplied pulse is introduced into the memory 54 via the product gate 56: The product gate 56 is open at this time because both its inputs are at the higher level, i.e. the output signal of the generator 50 is also at the higher level. The memory 54, which receives the pulse at its input 54s, controls the next product gate 58 with its output. At this time, the input signal for this gate from the memory 54 is also supplied with a higher level. If the output pulse from the generator 50 drops from the upper level to the lower level after the time T2, then this potential jump is fed through the converter 52 to the summing gate 46, whereby the product gate 58 can conduct because both of its inputs are at the upper level. The signal enters the summing gate 46 via the converter 44, in which the supplied pulses are converted into needle pulses, as in the converters 40 and 42. This pulse, which comes out of the converter 44 and by which the memory 54 is reset via its second input 54c, can now pass freely to the output 32, while being separated from other pulses that are closely passing through. This separation is done based on the described feedback in the memory 54. It should be noted that the duration T2 of the pulses of the first generator 50 is less than the duration Ti of the pulses of the ^second generator 48 in order to avoid suppression of the pulses arriving at the generators at a time when one generator 48 is activated and the other generator 50 is not.

In the above-described circuit, the filter 18 operates in such a way that, depending on the converter pulses, the course of the correction pulses 30 entering the circuit 20 is determined depending on the duration T _max . As already mentioned, it is possible to feed a larger number of correction pulses into the circuit 20 in relation to one converter pulse, which is ensured by the above-mentioned filter 18.

In the described device, therefore, the superposition of the converter and correction signals is carried out by the sum of both frequencies. However, as was said in the introduction, there are also cases in which the correction frequency must be derived from the converter frequency. Even for this case, the circuit shown in Fig. 7 can be used after a simple modification. For example, a converter pulse eliminator can be connected to the previously used counter and if the counter is then set to a certain value n due to the correction pulses, for example 1, 2 or 3, then based on this value, the corresponding number of converter pulses is eliminated by means of a backward counting procedure in the aforementioned counter before the following pulses are fed back to the circuit output. The necessary division of the converter and correction pulses can then be carried out in the same way as in the circuit shown in Fig. 10, but pulse generators with a different pulse duration are used.

Claims (9)

Method for correcting the output signals of a digital converter which converts physical quantities by means of a pulse train, wherein the characteristic of the converter frequency as a function of the monitored physical quantity is linear within the working region and intersects outside the working region of the converter that the correction frequency (f _k ) is superimposed on the converter frequency (f _g ) to obtain a corrected converter frequency (fgkoor), wherein the correction frequency (f _k ) is adjusted such that the resulting characteristic passes through the origin of the coordinates, the correction pulses being fed to the device after the transducer pulse has passed, while the time-identical or near transducer and correction pulses are separated into two different pulses. Method according to claim 1, characterized in that the correction pulses which are superimposed on the converter pulse are limited by the maximum number if the frequency of the converter pulses falls below the minimum value [dmin] of the working area, Method according to claim 1 or 2, characterized in that after each converter pulse, correction pulses are emitted for a time greater than t _max = 1 / f _min . 4. Method according to claim 1 or 2, characterized in that after each converter pulse, correction pulses of less than n = fk / fmin are transmitted. 5. A method according to claim 1 or 2, characterized in that each transducer pulse sends correction pulses for the superposition, the number n of which is at most f _k / f _m in or equal to the integer N and K directly above or below this value. wherein the invention of the correction pulses N and K is alternately controlled so that the average value of the total number of these pulses for each correction pulse is equal to n, Device for carrying out the method according to claims 1 to 5, with a pulse generator for producing correction pulses, characterized in that the correction pulse generator (4) is connected to a first gate input (3) to which a converter (1) is connected. ), the gate output (3) being connected to the first input of the correction circuit (2), to whose second input the output of the converter (1) is connected. Device according to claim 6, characterized in that the gate (3) comprises a filter (18), the output of which is connected to the product gate (28). Device according to claim 7, characterized in that the correction circuit (2) is a combined dividing and adding circuit (20) comprising a summing gate (46) connected by its one input to the output (16) of the converter (1) and its a second input to the correction pulse output (30) of the correction circuit (2) and its output to the inputs of two pulse generators (48, 50), the first pulse generator (48) being connected to the output (32) of the combined divider and adder circuit (20) and the second pulse generator (50) is coupled to the summation gate (46) via a transducer (52) and a first product gate (58) and simultaneously through a second product gate (56) and a memory (54) which is coupled to the second pulse generator. generator (50). Apparatus according to claim 8, characterized in that the inverters (40, 42, 44) of the feed signals in the needle pulses are connected before the inputs of the summing gate (46).