JPH0371017A - signal calibration device - Google Patents

Abstract

PURPOSE:To remove the crosstalk of signals in a cosine and a sine function shape automatically by composing a device of a crosstalk removing means, an arctangent arithmetic means, a peak detecting means, and a crosstalk coefficient arithmetic means. CONSTITUTION:The crosstalk removing means 1 finds approximate values of costheta and Sintheta from a crosstalk coefficient (k) and input signals X0 and X1 and the arctangent arithmetic means 2 finds theta from the approximate values of costheta and sintheta. Assuming that theta is a fine angle, the peak value detecting means 3 finds the maximum value d0 of X0+X1 from X0+X1 when theta is (pi/4-delta)-(pi/4+delta) and the maximum value d1 of -X0+X1 from -X0+X1 when thetais (3pi/4-delta) - (3pi/4+delta). The maximum value d2 of -X0-X1 is found from -X0-X1 when theta is (-3pi/4-delta) - (-3pi/4+delta) and the maximum value d3 of X0-X1 is found from X0-X1 when theta is (-pi/4-delta) - (-pi/4+delta). Then the crosstalk coefficient arithmetic means 4 finds the coefficient (k) from equation I by using those maximum values and performs control 5 every time the signals X0 and X1 are obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an apparatus for detecting and automatically compensating for crosstalk fluctuations in a source signal having a cosine or sine function, and particularly relates to a device for detecting and automatically compensating for crosstalk fluctuations when the original signal is in the form of a cosine or sine function. The present invention relates to a signal preprocessing device suitable for detecting a minute angle by arctangent calculation from an encoder signal whose sine function-like signal is a function of position.

[Conventional technology]

Cosine and sine function signals are often used in measurement devices and the like. A typical example of this is a device for measuring minute angles, in which the analog output of the encoder is in the form of a cosine or sine function where one revolution is N periods (N is generally several hundred to several thousand). Using this method, the period is further subdivided by arctangent calculation.

[Problems to be Solved by the Invention] In such a device, if there is crosstalk (mixing of a cosine signal and a sine signal) in the original signal, the measurement result will contain an error. In particular, the cosine 1 sine signal obtained from the encoder described above has a detector setting position of 90
If the signal deviates from a position where the phase is different, a crosstalk component will be included. The cosine detector and the sine detector are often formed integrally, but in the case of a rotary encoder using a magnetic drum, when the distance between the detector and the drum changes, the circumference of the detection part changes. Even if the detectors are set at regular intervals, the phase will shift and crosstalk will occur. Further, when cosine and sine signal lines are provided in close positions, crosstalk occurs due to mutual interference.

The present invention was made to solve these problems, and an object of the present invention is to provide a signal calibration device that can easily detect crosstalk coefficients and easily correct crosstalk.

[Means to solve the problem]

The signal calibration device according to the present invention includes a crosstock removal means, an arctangent calculation means, a peak value extraction means, a crosstalk coefficient calculation means, and a judgment means.

[Effect]

In this invention, the input signal is x0=COSθ1k −sinθ.

x, = sin θ+k − cos θ (however,
lkl<<1. -π≦θ≦π), (a) Give an initial value to the crosstalk coefficient k, and (b) Input signal X to the crosstalk coefficient. +Xl or cosθ,
Find an approximate value of sin θ, C. The above cos θ
, to obtain θ from the approximate value of sin e.2.When δ is an infinitesimal angle, when θ is greater than or equal to π/4−δ and less than or equal to π/4+δ, then from the value of x(, +x, .+x1 maximum value d.

When θ is greater than or equal to 3π/4−δ and less than or equal to 3π/4+δ, find the maximum value d of −x0+x1 from the value of −x0 +xI, and θ is greater than or equal to 13π/4−δ. , and is less than or equal to -3π/4+δ, find the maximum value d2 of -x0-x from the value of -x0-x, and θ is greater than or equal to 1π/4-δ, and -π/ 4+δ or less, find the maximum value d3 of x0-x from the value of x0-x, e. Find the crosstalk coefficient ratio using the following formula, k-(do d+ +d2 ds )/4 ``By repeating the above steps 0 to E every time the signal X Or X l is obtained, crosstalk can be detected and corrected.

0 can be found by arctangent calculation from approximate values of cos θ and sin θ. Furthermore, the influence of noise can be minimized by calculating the maximum value in each angle range using the following equation.

d = d' + p (x-d') Here, d is the maximum value found in each range, d'
is the previously obtained value of d, and X is the input signal X. +X+
The value obtained by addition and subtraction of , p represents a constant. Furthermore, do and d2. By alternating the corrections of d and d3, even corrections are performed, and even if θ happens to stop within a certain correction range, shifts in the crosstalk coefficient due to unnecessary corrections are prevented.

〔Example〕

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention. In this figure, numeral 1 is a crosstalk removal means for input signal X. , xl to find approximate values of cos θ and sin θ. Reference numeral 2 denotes arctangent calculation means, which performs calculation to obtain θ from the value of COS θ sin θ obtained by the crosstalk removal means 1. 3 is a pic value detection means, and when π/4-δ≦θ≦π/4+δ, the maximum value d for XO+X.

When 3π/4-δ≦θ≦3π/4+δ, the maximum value of x, )+xl d.

-3π/4-δ≦θ≦-3π/4+δ when X (4X
The maximum value d2 of x0-x when -π/4-δ≦θ≦-π/4+δ is found. 4 is a crosstalk coefficient calculating means, and the maximum value d obtained by the peak value detecting means 3.

~d3 is used to calculate the crosstalk coefficient ratio. Reference numeral 5 denotes a control means for controlling each of the above-mentioned parts.

The signal calibration device of the present invention is practically constructed using a microprocessor. Each original signal is analog
It is converted into digital data and then manually processed by a microprocessor. The microprocessor sequentially and repeatedly executes signal input, crosstalk correction, arctangent calculation, angular range determination, and maximum value correction if necessary, according to a program written in a memory (usually ROM: read-only memory). An example of a program flowchart is shown in FIG.

(1) to (15) in this figure indicate each step.

Figure 3 shows the principle of this invention, in which two signals X o + X, which are original signals having cosine and sine functions,
This figure shows a diagram plotted on rectangular coordinates with + as the X and Y coordinates, respectively. The diagram emphasizes crosstalk, and the actual crosstalk is much smaller than what can be read from the diagram.

The operation of the embodiment shown in FIG. 1 will be explained using FIGS. 2 and 3.

First, in the crosstalk removal means 1, sin θ, co
Find sθ (1) to (5). Since the crosstalk coefficient ratio is generally sufficiently smaller than 1, this calculation can be performed using the following equation.

cos θ=x, - to -x.

stnθ=x, −to −x.

Using the values of 5ino and cosθ thus obtained, O is determined by the arctangent calculation means 2 (6). In this calculation, instead of simply calculating the ratio between the sine and the cosine to find the arctangent, it is necessary to find θ in the range of -π to +π based on the sign relationship between the sine and the cosine. Such a function is, for example, at in C language.
It is known as an2.

The obtained θ is π/4.3π/4. d as shown in the figure from the X O+ , d, , d, , d are determined by the peak value detection means 3 (11) to (14). Value of trigonometric function, cos (π/4) --cos (3yr /4)
--cos (-3rt /4)-cos (-π/4
) −(r/2 sin(yclo-5in(3yc/4)−-sin(
-37E/4) --sin (-π/4) -JT/2
From, [+, -JT+ k(dl-1dl-kFTd2
・JT+kJ'T, d3・bar kJ-d, and the crosstalk coefficient ratio is calculated by the following formula in the crosstalk coefficient calculating means 4 (15).

The signal correction from k= (do dl + d2 d3)/4ryr times can be performed using the newly determined crosstalk coefficient ratio.

θ is exactly π/4, 3π/4. -3π/4-π/4 is rare, so d in the vicinity of these (range of soil δ). , dr, d2+d3 are updated. δ is 1/
Range from 10 revolutions to 1/1000 revolutions, preferably 11 revolutions
The rotation speed is appropriately selected within the range of 50 rotations to 1/200 rotation. δ
Increasing the range increases the frequency of updates, but increases the error. The error is about 08% at 1150 rotations, 1/2
It is approximately 0.05% at 00 rotations. As described below, the actual error can be further reduced by updating the maximum and minimum values using a weighted average rather than using a single value.

It is preferable to update the maximum value using an exponential smoothing filter. This essentially corresponds to finding a weighted average in which the weighting coefficient decreases exponentially toward the past. The calculation formula is given by the following formula, where x1 is the value obtained from the human input signal and d is the value to be updated. Here, p is a small constant, and by setting this value to 2-n, multiplication can be replaced with a shift operation. Here, n is approximately 4 to 20, and may be appropriately selected based on the balance between update speed and value stability. If the value of n is small, the correction will be performed quickly, but it will be more susceptible to noise.

The zero point may change suddenly due to shock or other causes. In such a case, if θ is stopped within a certain correction range, only a specific maximum value will be corrected and the influence of the constant term will appear. This behavior is undesirable because the resulting errors are generally more harmful than crosstalk errors. To avoid this, after correcting do, d2
After correcting dl, it is sufficient to suspend the correction until d3 is corrected. This function is achieved by setting a correction flag in the microprocessor's memory.
It can be easily achieved.

(Effects of the Invention) As explained in detail above, the present invention configures a signal calibration device with a crosstalk removal means, an arctangent calculation means, a peak value detection means, and a crosstalk coefficient calculation means. Crosstalk of cosine and sine function signals often used in equipment is automatically removed, and errors can be reduced with a simple circuit using only one inexpensive component. This eliminates the need for extensive adjustments, making it possible to construct a highly accurate measuring device at low cost.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.
FIG. 2 is a flowchart for explaining the embodiment of FIG. 1, and FIG. 3 is a diagram showing the principle of the invention. In the figure, 1 is a crosstalk removal means, 2 is an arctangent calculation means, 3 is a peak value detection means, 4 is a crosstalk coefficient calculation means, and 5 is a control means. 2

Claims (1)

[Claims] Two signals with a 90° phase difference that are approximated by cosine and sine functions and include crosstalk components x_0=cosθ+k・sinθ, x_1=sinθ+k・cosθ, where |k|<<1, - cosθ, s from π≦θ≦π
A device for obtaining a signal approximated to inθ or a signal corresponding to θ, which obtains co from a crosstalk coefficient k and input signals x_0,
crosstalk removal means for obtaining approximate values of sθ and sin θ; arctangent calculation means for obtaining θ from the approximate values of cos θ and sin θ; when δ is an infinitesimal angle, θ is greater than or equal to π/4−δ;
and π/4+δ or less, calculate the maximum value d_0 of x_0+x_1 from the value of x_0+x_1, and find that θ is 3π/4.
−δ or more and 3π/4+δ or less, −
From the value of x_0+x_1 to the maximum value d_ of -x_0+x_1
1, and θ is greater than or equal to -3π/4-δ, and -3π
/4 + δ or less, from the value of -x_0-x_1 -
Find the maximum value d_2 of x_0-x_1, and θ is -π/4-
When it is greater than or equal to δ and less than or equal to −π/4+δ, x_
peak value detection means for the sum-difference signal that calculates the maximum value d_3 of x_0-x_1 from the value of 0-x_1, and the crosstalk coefficient k using these maximum values k=(d_0-d_1+d_2-d_3)/4√ 2 and the input signal x_0,
A signal calibration device comprising: a control means that repeatedly performs the above operation every time x_1 is obtained.