GB1575289A

GB1575289A - Compensating for time errors in measuring devices employing analog-stochastic converters

Info

Publication number: GB1575289A
Application number: GB18560/78A
Authority: GB
Original assignee: Heliowatt Werke Elektrizitaet GmbH
Current assignee: Heliowatt Werke Elektrizitaet GmbH
Priority date: 1977-05-09
Filing date: 1978-05-09
Publication date: 1980-09-17
Also published as: DE2721275A1; FR2390736A1; CH614783A5

Abstract

Timing errors in single-channel multiplexing of different measurement values can be compensated for by interpolation. When analog/stochastic-value converters are used for a number of input signals, this arithmetic averaging by addition is not suitable since the results of addition must not exceed the range of probability between zero and one in the stochastic technique. For this reason, alternating sampling is used for compensating for the timing errors when using analog/stochastic-value converters for a number of input signals, in such a manner that products of the associated value pairs are continuously formed after each sampling and conversion in that the value just obtained is multiplied by the associated value previously stored. As a result, instantaneous values are formed which produces a constant timing error of the samples with a constant sampling frequency of the input signals in comparison with the reference frequency of a standard. This constant value can be completely compensated for by calibrating the device.

Description

(54) COMPENSATING FOR TIME ERRORS IN MEASURING DEVICES EMPLOYING ANALOG-STOCHASTIC CONVERTERS (71) We HELIOWATT WERKE ELEKTRIZITATS GESELLSCHAFT MBH, a German company, of Wilmersdorfer Strasse 39, 1000 Berlin 12, Federal Republic of Germany, 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 compensating for time errors in measuring devices which employ analog-stochastic converters for a number of input signals which are applied by means of a multiplexer to the converter.

The invention is particularly although not exclusively concerned with analogstochastic convertors in electronic energy meters or power meters.

In single-channel multiplexing, there occurs between the sampling instants of two associated values, for example the voltage and the current of one phase, a time offset whose magnitude depends upon the multiplexing frequency. It is known to compensate for the time error in single-channel multiplexing by the principle of interpolation. Since this arithmetical average is effected by additions, but the results of additions in the stochastic technique must not exceed the range of probability from zero to one, the method of averaging by addition is less suitable for analog-stochastic converters.

Usually, in known analog-stochastic converters, analog input signals are compared with signals from a noise source. At the output of the converter, there appear signals on which are superposed positive and negative comparison noises. An offset is added to one of the signal currents so that there is an overlap of the positive and negative parts at the zero-point of the signals. The time of the overlap then forms a correction quantity.

According to the present invention, there is provided a method of compensating for time errors in an analog-stochastic converter, in which a plurality of input signals are applied successively to the converter through a multiplexer, to provide a succession of respective output values from the converter, each output value of the converter is stored, a product is formed of each output value and an associated stored preceding value, there is set up a constant time error dependent upon the known fixed frequency of the multiplexer and a known fixed control frequency of the measuring device, which time error is compensated by calibration of the device, and successive products formed from successive said output values are added in an integrator.

By "control frequency" of the measuring device is meant the frequency (such as a clock frequency) of a control signal used to control operation of the device.

The integrator may be a forward and reverse counter.

Some embodiments of the invention will now be described, by way of example.

In an electronic energy meter or power meter, voltage and current input signals of one phase are fed alternately to an analogstochastic converter, via a multiplexer. Each output value from the converter is stored, and a continuous product is formed of the instantaneous output value with the stored immediately preceding value. In this way, there is continuously formed the product of the alternately sampled current and voltage input signals. The successive values are added, with correct sign, in an integrator, and in this way, averaging is effected, which compensates for the time error introduced by the multiplexer.

We have found that this method, which can be particularly simple from the point of circuitry, can completely compensate for the time error if the multiplexing frequency of the meter is made constant, and a control signal of the meter is derived from the reference frequency of the alternating input signals. Thus, the time error arising is constant and, knowing the fixed multiplexing frequency and the fixed reference frequency of the meter, time error can be completely compensated for by corresponding calibration of the meter.

In another example of the invention a polyphase (in this case three-phase) electronic energy meter comprises an analogstochastic converter, which is fed by a multiplexer. Firstly, the three-phase voltage input signals are sampled by the multiplexer, and successively converted. Each corresponding output value from the converter, being the stochastically coded voltage value of the measured signal, is then fed into a respective one of three shift registers, and also to a first input of a respective one of three multipliers. Then, optionally, a zeropotential input signal is fed to the converter, and a zero-point correction of the meter is effected.

Following this, the three phase current input signals are then sampled by the multiplexer, and converted to provide stochastically coded output values. These three output values are similarly fed into the three respective shift registers, and also to the first inputs of the respective multipliers.

The length of each shift register is such that, as a new value is introduced into the shift register from the convertor output, the preceding value in the shift register is displaced to the shift register output. The shift register output is connected to a second input of the respective multiplier. Thus, the two inputs to a given multiplier at any one time comprise the new value just fed into the input of the respective shift register, and the immediately preceding associated value at the output of the shift register. These two signals in the shift register always correspond to a voltage value and an immediately preceding current value (or vice-versa) of one phase. Each multiplier therefore continuously forms the product of the associated current and voltage values of one phase. It will be appreciated that each sampled current and voltage is used twice for the formation of a corresponding product, once when the value is at the shift register input, and again when the value is at the shift register output. The continuously formed products from the multipliers are fed, with correct sign, to respective integrators. The time error is compensated for by the formation of the mean value, in the integrators.

Optionally, a zero-point correction of the meter can be effected both prior to sampling the three voltage input signals and prior to sampling the three current input signals.

Advantageously, the output values from the converter are fed to the shift registers via bipolar two-conductor connections, for forwards and reverse counting.

WHAT WE CLAIM IS: 1 A method of compensating for time errors in a measuring device employing an analog-stochastic converter, in which a plurality of input signals are applied successively to the converter through a multiplexer, to provide a succession of respective output values from the converter, each output value of the converter is stored, a product is formed of each output value and an associated stored preceding value, there is set up a contant time error dependent upon the known fixed frequency of the multiplexer and a known fixed control frequency of the measuring device, which time error is compensated by calibration of the device, and successive products formed from successive said output values are added in an integrator.

2. A method according to claim 1, wherein the integrator is a forward and reverse counter.

3. A method according to claim 1 or 2, wherein the measuring device is an electrical energy meter, and products are formed from pairs of said output values corresponding to voltage-current pairs of input signals.

4. A method according to claim 3, wherein the measuring device is a threephase electrical energy meter, and wherein three phase voltage input signals are first sampled and converted, each corresponding output value from the converter is fed into a respective one of three shift registers and a first input of a respective one of three multipliers, an output of each shift register being connected to a second input of the respective multiplier, three associated phase current input signals are then similarly sampled, converted and fed to the first inputs of the respective multipliers and into the respective shift registers so as to displace the previous associated voltage values to the shift register outputs, each multiplier being operative to multiply together the values at its first and second inputs to form a product which is added, with correct sign, to the contents of a respective integrator, and the three phase voltage input signals are again sampled, converted and fed to the shift registers and multipliers, such that products of the phase current-voltage pairs of input signals are continuously formed and averaged.

5. A method according to any preceding claim, wherein the multiplexer periodically feeds a zero-potential input signal to the converter, and a zero-point correction of the measuring device is then effected.

6. A method according to claims 4 and 5, wherein a zerd-point correction is

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

Claims

**WARNING** start of CLMS field may overlap end of DESC **. the meter is made constant, and a control signal of the meter is derived from the reference frequency of the alternating input signals. Thus, the time error arising is constant and, knowing the fixed multiplexing frequency and the fixed reference frequency of the meter, time error can be completely compensated for by corresponding calibration of the meter. In another example of the invention a polyphase (in this case three-phase) electronic energy meter comprises an analogstochastic converter, which is fed by a multiplexer. Firstly, the three-phase voltage input signals are sampled by the multiplexer, and successively converted. Each corresponding output value from the converter, being the stochastically coded voltage value of the measured signal, is then fed into a respective one of three shift registers, and also to a first input of a respective one of three multipliers. Then, optionally, a zeropotential input signal is fed to the converter, and a zero-point correction of the meter is effected. Following this, the three phase current input signals are then sampled by the multiplexer, and converted to provide stochastically coded output values. These three output values are similarly fed into the three respective shift registers, and also to the first inputs of the respective multipliers. The length of each shift register is such that, as a new value is introduced into the shift register from the convertor output, the preceding value in the shift register is displaced to the shift register output. The shift register output is connected to a second input of the respective multiplier. Thus, the two inputs to a given multiplier at any one time comprise the new value just fed into the input of the respective shift register, and the immediately preceding associated value at the output of the shift register. These two signals in the shift register always correspond to a voltage value and an immediately preceding current value (or vice-versa) of one phase. Each multiplier therefore continuously forms the product of the associated current and voltage values of one phase. It will be appreciated that each sampled current and voltage is used twice for the formation of a corresponding product, once when the value is at the shift register input, and again when the value is at the shift register output. The continuously formed products from the multipliers are fed, with correct sign, to respective integrators. The time error is compensated for by the formation of the mean value, in the integrators. Optionally, a zero-point correction of the meter can be effected both prior to sampling the three voltage input signals and prior to sampling the three current input signals. Advantageously, the output values from the converter are fed to the shift registers via bipolar two-conductor connections, for forwards and reverse counting. WHAT WE CLAIM IS:

1 A method of compensating for time errors in a measuring device employing an analog-stochastic converter, in which a plurality of input signals are applied successively to the converter through a multiplexer, to provide a succession of respective output values from the converter, each output value of the converter is stored, a product is formed of each output value and an associated stored preceding value, there is set up a contant time error dependent upon the known fixed frequency of the multiplexer and a known fixed control frequency of the measuring device, which time error is compensated by calibration of the device, and successive products formed from successive said output values are added in an integrator.

6. A method according to claims 4 and 5, wherein a zerd-point correction is

effected prior to sampling the three voltage input signals and prior to sampling the three current input signals.

7. A method according to claim 4 or to claim 5 or 6 as appendant thereto, wherein the output values from the converter are fed to the shift register via bipolar twoconductor connections, for forwards and reverse counting.

8. A method of compensating for timeerrors in a measuring device employing an analog-stochastic converter, the method being substantially as hereinbefore described.

9. A measuring device comprising a multiplexer, an analog-stochastic converter, a multiplier, and an integrator, and being arranged to operate in accordance with a method according to any preceding claim.

10. A measuring device comprising a multiplexer, an analog-stochastic converter, three shift registers, three multipliers, and three integrators, and being arranged to operate in accordance with a method according to claim 4, 5, 6, 7 or 8.