DD217036A1 - PROCESS FOR REDUCING FAULTS IN PERIODIC MEASUREMENT SIGNALS - Google Patents

Abstract

The invention relates to a method for reducing interference in periodic measurement signals, which can be used in metrology, especially in devices that include a digital signal memory. The aim and the object of the invention is to substantially reduce the effort required to carry out the process. In a method for reducing disturbances in periodic measurement signals, in particular by automatically running digital mean value, the object is achieved by a digital difference between a last calculated, stored digital average An 1 and an instantaneous digital sample Sn of the periodic measurement signal resulting transfer Ue is inserted after the digital division by a factor X in the subsequent formation of a sum Sn An 1 X An 1 in the least significant bit of the digital value of this sum, wherein in the difference equality of all bits of the digital samples Sn of the periodic measurement signal must be avoided. figure

Description

Title of the invention

Method for reducing interference in periodic measuring signals

Field of application of the invention

The invention relates to a method for reducing interference in periodic measurement signals, which can be used in the measurement technique, in particular in devices that include a digital signal memory.

Characteristic of the known technical solutions

The basics of a method for reducing interference in periodic measurement signals are in the article "Calibrated Real-Time Signal Averaging" ν. 3. Eran Deardorff and Charles R. Trimble in the Hewlett-Packard Journal April 1968 pp. 8-13. ,

The method uses the following recursion formula:

SA ₄

Herein mean:

A '- new mean value η to be calculated

A _ . - last calculated mean value S - 'sample of the periodic measurement signal X - fixed division factor for all η

'·' - 2 ^λ - "'

This recursion formula is processed digitally in four known automatically running successive steps. In the first step of the process of determining the difference between the sample value of the peri- carried ^Λ * 5 ödischen measurement signal and the last calculated mean value. In the second step of the procedure becomes the digital division. the difference between the Abbe value of the periodic measurement signal and the last calculated average by the division factor X = 2 performed. N represents a variable ν integer. The digital division is executed by "shifting to the right" of the digital value of the difference by N digits. In order to detect even changes in small digits of the difference, the width of the signals to be processed must necessarily be extended by at least N places for the two last steps following the division, the addition and the storage of the newly calculated mean value. This leads to a very disadvantageous, considerable increase in the expenditure on digital technology, which has a negative effect on the usability of the method in measuring devices with digital memory, their size and price. Since only with the largest possible value of N also a large improvement in the signal-to-noise ratio at periodic measurement signals achieved - is, it must be used for a large process engineering effort to implement the method.

25 ". ,,. ''''''''' Aim the invention,

The object of the invention is to improve the known method in the. Way that the effort required to carry out the process is substantially reduced, although only slight changes in the implementation of the known method should occur.

- 3 - Presentation of the essence of the invention

It is an object of the present invention to improve the existing method of digital averaging, which leads to the reduction of disturbances in periodic MeS signals, by extending the digital word width and the facilities required to process the digital values by N digits, after execution of the second step of digital averaging, can be omitted. Despite these limitations, however, a division up to the value of N = K-1 should be possible, where K is to represent the number of digits or word width of the digital values of the samples of the periodic measurement signal. This corresponds to a reduction of interference with periodic Meßsigna ^v len by a factor

, 2 ^N = Va.

^K -

In a method for reducing interference in periodic measurement signals, in particular by automatically running digital averaging, in which, after the digital difference between a last calculated, stored digital average ^ _n-1 and an instantaneous digital sample S of a periodic measurement signal, a digital division of this difference with a, in particular digitally adjustable factor X is made, the digital result with the last calculated stored digital mean summed results in a new storable digital mean, the object is achieved in that one in the digital difference between the last calculated , stored digital average A and the instantaneous digital sample S of the periodic measurement signal resulting transfer 0, after the digital division

by a factor X in the subsequent formation of a

S-A '

Sum η η - 1 ₊ . in the least significant bit of the

X η - 1 · ...

digital value of this sum is inserted, wherein the difference in the equality of all bits of the digital samples S of the periodic measurement signal must be avoided.

embodiment

The method according to the invention will be explained using an exemplary embodiment

 The accompanying drawing shows:

a block diagram of an arrangement for carrying out the method according to the invention

In a first method step, the digital value of the voltage applied to the input terminal 1 periodic measurement signal passes through a comparator 2 and a multiplexer 3 to a second input of an arithmetic unit 4. It is, by supplying a coming out of the result output of the comparator 2 logic signal, for equality all bits of the digital samples S. the bit following the least significant bit from the bus output of the analog-to-digital converter 5 by means of a controlled inverter 6, which may be advantageously formed by an exclusive-OR gate, negated and fed back to the first bus input of the multiplexer 3 ., Thus, never all bits of the first bus input of the multiplexer 3 can assume the same logical value.

Out. In a first memory step 7, the calculated stored digital mean value A _± also arrives at the first input of the arithmetic unit 4 in the first method step. The arithmetic unit 4 is intended to operate by means of a control circuit (not shown)

'.': - ⁵ -

be switched to subtraction. Thus lies at the output of the arithmetic unit 4, the digital value of the difference S -A,

Π Π "· J.

which is taken over in a shift register 8. In a second method step, the latter difference is divided by the factor X = 2 by shifting the digital value of the difference S-A "to the right by N digits. there

η η - l.

is negated fed over the Serieneingarig the shift register S of Oberteil 0 resulting in the arithmetic unit 4 negated. The same Obertrag C is in a third step

IQ into the least significant bit of the bus output of the

   · 'S' · »'A ^

gisters 8 digital results η η -, 1

via, an electronic switch 9, ¹ in particular, may consist of NAND members inserted. The electronic switch S is in the third step to the coming from the calculator 4 O.bertrag 0 switched by means of a control circuit, not shown. The logical value of the transfer 0 in the difference formation is: '..''.·.'''

'' S _n > A _n _ _± , O = -I '' -.

S _n ο 'A _n _ ₁ ü = »0

In η repetitions of the method, therefore, the logical value of the transfer 0 is added to it in the third method step until an alignment of the calculated digital average A '^ ₁ with the instantaneous digital sample S takes place. In this case, the averaging is always carried out. Thus, the change in the points lost by the division of the difference is detected via the transfer 0 of the difference and, at the same time, averaging is reduced by averaging Since the multiplexer 3 returns to its state with the bus output of the shift register after completion of the first method step 8 connected second bus input is switched,

is located at the bus output of the arithmetic unit 4 of the calculated new digital average A, since this is to be switched back to the calculation type addition after completion of the first process step. With a control signal delayed by the computing time of the arithmetic unit 4 from a control circuit, not shown, the calculated new digital average A_ can still be recorded in the shift register 8 during the third method step. In a fourth and final method step, the calculated new digital mean value A, which is present at the end of the third method step at the bus start point of the shift register 8, instead of the calculated digital average A _. transferred to the main memory 7. At the same time, the analog-to-digital converter 5 is started via a control circuit (not shown) for converting a new sample of the periodic measuring signal applied to the input terminal. After the end of the implementation of the analog-to-digital converter 5 again, the implementation of the method in the described four process steps in motion. This creates an automatic procedure of the procedure. The analog periodic measurement signal whose interference is reduced by means of frequent application of the method to the sampling points of the periodic measurement signal, by reading the main memory 7 when switching the arithmetic unit 4 to the mode "first bus input" + "0" via a digital Anälogwandler 10th be removed at an output terminal 11.

Claims (1)

•. · ^ - 7 - ' Erfindunqsanspruch Method for reducing periodic disturbances; Measuring signals, in particular by automatically running digital averaging, in which, after the digital difference between a previously calculated stored digital average A ^ ₁ and an instantaneous digital sample S of the periodic measurement signal, a digital division of this difference with a, in particular digitally adjustable factor X made whose digital result is summed with the previously calculated, stored digital mean value yields a new digital average which can be stored, characterized in that, in the case of a digital subtraction between the last calculated, stored digital average A _ .. and the instantaneous, digital Sample S _{n of} the periodic measurement signal resulting Obertrag O, after completed digital division by a factor X in the subsequent formation of a sum η * "η - 1 , , ^{Χ + η} - ^Χ is inserted into the least significant bit of the digital Vertes this sum, in the difference formation, the equality of all bits of the digital samples S of the periodic measurement signal must be avoided. For this 1 sheet drawing