SU993284A1 - Function generator - Google Patents

Description

The invention relates to analog computing and can be used for non-linear transformations of instantaneous values of signals.

A functional transducer is known that reproduces polynomial dependencies by means of connected in series links consisting of a multiplying block and an adder; i.

A disadvantage of the transducer is the complexity of the calculations and adjustment of the coefficients when playing high degree polinschms.

The closest to the present invention is a functional converter containing an output adder, n multiplication units, the first inputs of which are interconnected and with the second input of the first multiplication unit and are the input of the inverter, the second input Xc1 of the next multiplication unit connected to the output of the previous, n input outputs the adder is connected to the outputs of the respective multiplication blocks, and (n + 1) -th and (n + 2) -th inputs - respectively with the input of the converter and the voltage reference bus. In the converter, the input signal passes sequentially through all the multiplication blocks, the output signals of which are summed 2j,

A disadvantage of the known converter is the accumulation of errors in series-connected multiplication units, especially when using the approximation by Chebyshev polynomials, at which large values are often observed: summation coefficients of high degrees of the input signal. Practically very difficult to reproduce.

15 way polynomials of degree higher than sixth,

The purpose of the invention is to improve the accuracy of the conversion.

The goal is achieved

20 in that the functional converter containing the output adder, n multiplication units, the first inputs of which are interconnected and with the second input of the first oscillator unit and are the input of the inverter, the second input of each next block (and the multiplications are connected to the output of the previous, n inputs output adder connected to the outputs

Claims (2)

The 30 corresponding multiplication units, and (n + 1) -th and (n + 2) -th inputs, respectively, with the input of the converter and the reference voltage bus, additionally contain a limiter, an additional multiplication unit, and an additional adder, whose inputs are connected to the inputs output adder, the output of the additional adder is connected to the first input of the additional multiplication unit, the second input of which is connected to the output of the limiter, the input is limited to l connected to the input of the first multiplication unit, the output, the additional multiplication unit connected to (n + 3) input of the output adder. The drawing shows a block diagram of a functional converter, the functional converter contains an output adder 1, an additional adder 2, n blocks 3. . 3p multiplication limiter 4, optional multiplication block 5. Functional Converter works as follows. The input signal X is fed to the inputs of the first multiplication unit 3 at the output of which the signal X is formed. This signal is fed to the input of the next multiplication unit 3, where it is multiplied by X, and the output signal Y is output. This signal is again multiplied by X, etc. . until a signal is received at the output of the last nth block of multiplication 3, from the outputs of blocks 3. . . 3j, the signals are multiplied with the given weights at the n inputs of the output adder 1, where they are added to the desired sign. There, the (n + 1) and (n + 2) -th inputs receive signals from the converter input and from the voltage of the reference voltage, corresponding to the values of X and X °. Limiter 4 is adjusted so that the voltage at its output is zero at X O (or X 0) and coincides in magnitude with the input voltage at X 5 O (respectively at X 7/0). Let, for certainty, there be a restriction from above. Then, when X / O, zero voltage is applied to one of the inputs of the additional multiplication unit 5, while its voltage out is also zero and the additionally introduced part of the circuit does not affect the operation of the functional converter. In this case, the output of ycTpo is iCTBa. a polynomial is obtained, where a is the specified values of the transfer coefficients of the first (n + 2) -th inputs of the output adder. When X О, one of the inputs of the additional multiplication unit 5 receives the input signal X, and the second receives a polynomial of the form G, where Cf is similar to the previous one, the given coefficients. It can be shown that in this case the functional converter reproduces a function of the form: J J Cx)) ,, 1 t, -x W Similarly, if the limiter is set to limit from below, the output of the device is the view function; 1 ht-i; fCx) M а- (. t2} 1-0 1 u ada va values of the coefficients and, it is possible to approximate the functions by shifted Chebyshev polynomials with independent coefficients for positive and negative values of the input signals. Compared to the prototype device that reproduces the usual Chebyshev polynomials, the proposed approach allows an increase in the accuracy of the approximation by reducing the methodological error in reproducing smooth functions (on the order of 2, where n is the number of multiplying blocks of the prototype device). In most cases, the degree of the approximating polynomial and, therefore, the instrumental error is reduced. In addition, the coefficients with which the input signal degrees are summed, in this case, are usually much lower than when approximating the usual Chebyoev polynomials, which also contributes to reducing a certain increase in the approximation error in comparison with the theoretical one, due to the fact that the coefficients with a zero degree are not. change when the sign of the input quantity changes, is compensated by increasing the degree of the approximating polynomial to (n + 2) on one of the approximation sites. In addition, this additional error can be almost completely eliminated by varying the approximation coefficients (changing the type of approximation). By applying to the input of the limiter a signal of the form (Xa), it is possible to approximate functions on segments that are asymmetrical with respect to zero. In addition, it is obvious that using two limiters, additional multiplying blocks and auxiliary adders, it is possible, by applying signals (X-a) and (X-b) to the inputs of the limiters, to produce an approximation of functions by high-order splines with independent coefficients on three approximation intervals. In this case, the gain in accuracy can be achieved. Further complication of the functional converter in the direction, although it may improve accuracy, is impractical .-. Thus, the positive effect is to improve the accuracy of the approximation due to the possibility of using a polynomial approximation in the proposed functional converter with independent assignment of coefficients in% of the reproducible function definition areas. Claims of Functional Inverter Containing Output Adder | And the multiplication units, the first inputs of which are interconnected and with the second input of the first multiplication unit and are the input of the converter, the second input of each next multiplication unit is connected to the output of the previous one, the inputs of the output adder are connected to the outputs of the corresponding multiplication units -and +1) -th and (t2) -th inputs - c6io BetcT-i with the input of the converter ши bus of the reference voltage, о T: l fe h ay y and so that in order to improve the accuracy of the conversion, it additionally contains a limiter, an additional block multiplied by neither and the additional adder enter, which is connected to the inputs ayhoDg | totalizer, the output of the additional O) 6 adder is connected to the first BXOjtioM of the additional multiplication unit, the second input of which is connected to the output of the limiter, the input of the limiter is connected to the input of the first 6LOK11 multiplication, and the output of the additional multiplication unit is connected to (1h4-3) 4th input of the output multiplexer adder Sources of information taken into account in the examination 1. Reference nonlinear circuits. Under pesDi. D. Sheingold, M., Mir, 1977, p. 55. 2. Instruments and elements of automation and computing. RJ, YaKSPress-information, 1975, issue 7, ref.36f s, 3-5 (prototype).  r KlHxlr-HZlH2 Wop L / f3ff-i X dboroi