SU1555825A1 - Digital filter - Google Patents

Abstract

The invention relates to radio engineering. The purpose of the invention is to simplify the filter bandwidth adjustment. The digital filter contains algebraic adders 1, 3, 4, 5, 7 and 10, adders 2, 6, 8 and 9, delay lines 11-14, multipliers 15-19. This implementation of the filter ensures its operation in modes like a bandpass filter, and notch filter. In this case, all operations of summation and multiplication are performed in the time interval between two adjacent samples of the input signal, i.e. in the period of its quantization, equal to the delay time of the delay lines. The goal is achieved by providing a triple win in the amount of ROM coefficients. 2 Il.

Description

RF Entry

1

The invention relates to radio engineering and can be used in multi-tier filters for fixing the processes defined by a sequence of digital codes.

The aim of the invention is to simplify the adjustment of the bandwidth (notch) of the filters.

FIG. 1 shows a block diagram of a digital filter j in FIG. 2 - graph of digital filter.

The digital filter contains the first algebraic adder 1, the first adder 2, the third 3, the fourth 4 and the second 5 algebraic adders, the second adder 6, the sixth algebraic adder 7, the third 8 and the fourth 9 adders, the fifth algebraic adder 10, the first 11, the second 12 , the third 13 and fourth 14 delay lines, the first 15, the second 16, the fifth 17, the third 18 and the fourth 19 multipliers, the first 20 and second 21 inputs, the first 22 and second 23 outputs.

The digital filter works as follows.

If you use a digital filter as a bandpass filter, then the input signal samples come from the first input 20 to the first input of the adder 8, the second input of which receives the signal samples previously recorded in the third delay line 13 and weighted on the second multiplier 16 From the output of the adder 8 half-summed the samples are fed to the second input of the adder 9 and to the summing input of the algebraic adder 4, where they

folded twice in the invert

roving signal readings from the fourth 14 delay line. The value of the signal samples at the output of the adder 4 is fed to the first input of the adder 9, as well as through the adder 10 to the subtracting input of the algebraic adder 7 and to the input of the multiplier 17 "After weighting on the multiplier 17, the signal reads to the first output 22 and to the first summing input of the algebraic adder 7. In addition, they are sequentially weighted by multipliers 18 and 19, and then by. arrive at the third input of the adder 6 where they add up twice

pfi) Hpq.)

 +

z-)

A (1 - - - - /

1-2ct (2-2A-ABpZ-t + 2 () (1-2etl) 2o6t2-3ABy-2A) Z- + (1-2ABJ) Z "

(1-А-АВУ) (+ Z) ,,}

(2-2A-AB | Ug-i; 0-ABy -A) tl -2oit- A Z 4-2o6 (2-ZAVu-2A) (T-2AB J) Z n

0

five

0

five

...

Have

0

35

45

50

the signal coming directly from the output of the multiplier 17. From the output of the multiplier 19, the samples are also fed to the second summing input of the algebraic adder 7 and to the first subtractive input of the algebraic adder 1, where they add up to inverted counts coming from the delay line 14 and then are recorded on delay line 11. Half-summed in the algebraic adder 7, the samples through the adder 5 are fed to the input of the multiplier 15. Then the signal samples stored in the delay line 13 are rewritten into the delay line 14. The summed counts from the output of the adder 9 are fed to the first summing input of the algebraic adder 3, where they are added to the inverted readings from the adder 6 and the counts that come from the delay line 12. The result of the addition is recorded in delay line 13. Finally, the signal samples weighted on the multiplier 15 are added in the sweep 2 with the samples coming from the delay line 11, and the result of this addition is recorded in the delay line 12. After that, the filter is ready to process the next sample of the input signal. In this case, in order to calculate the next sample, it is necessary to store the signal samples recorded in delay lines 11-14 (zero values should be written to them before the first signal sample arrives). If you use a digital filter as a notch filter, then the input signal samples from the second input 21 are fed to the subtractive inputs of algebraic adders 10 and 5, and the output signal samples are fed to the second output 23 from the algebraic adder output 7. All other processing procedures are similar to the previous case. All the described operations of summation and multiplication are performed in the time interval between two adjacent samples of the input signal, i.e. in the quantization period, equal to the delay time of the 11-14 delay lines. j The transfer functions of the digital filter, found using the graph (Fig. 2), have the form

frequency:

and (M - 0-A-AVSH1-4o 4sL2-4 -I). P k; | -A-SetW1) + (1 -ABJ) (4-8oi + 4

ABJ) (4-8Ы + 4о / г)

515558256

where oi is the multiplication factor - (for a notch filter) by zero

teli 15 and 16;

And - multiplication factor of the multiplier 18;

B is the multiplication factor of 19;

X1 is the multiplication factor of the multiplier 1 7.

To estimate the frequency properties of the proposed digital filter, the transfer functions (1) and (2) are transformed into the frequency domain by means of a known transformation.

.b;

Expression (6) shows that the value of the scaling factor set at the filter input does not depend on the values of the transfer function coefficients (6) (unlike the prototype) and does not change during filter tuning.

Z - (1 - jtg) / (1 + jtg) (3

where T is the quantization period of the signal.

Further, solving the equations of the form | HB9 (jca) | 4 1 and ln№ (jO) | 2 0, you can get an expression for the center frequencies (band-pass and notch filters. As with the known filter, the coefficient d uniquely determines the center frequency of the entire filter generally:

QO -sr arccosoi.

(four)

The coefficient adjusts the band of the entire filter in accordance with the expression

DSO - arcctgjf

(five)

Therefore, regardless of the number of cascade - -, gadov in the multi-link filter, built on the basis of the proposed filter, the values of J1 are the same for all links of the passband (rejection) and provide automatic scaling of the signal, which in turn, significantly reduces the amount of ROM storage coefficients, and simplify the implementation of the filter.

Claims (1)

Invention Formula EV, which simplifies the process A digital filter containing the band adjustment, and also to reduce the 40 therefore connected first al- volume of the ROM allocated for storing the hebraic adder, the first line of coefficients. Delay, the first adder, to the second The coefficient A is determined by the value of the input of which is connected to the output of the second coefficient J, the parameter of the ana algebraic adder via changes in the 45 first multiplier, the second delay line, the third algebraic adder, the third delay line, the output ko. The second is connected to the input of the second intelligent prototype b and the law A 1 / (yy + y + 1). Since the parameter of the analog filter prototype in the general case is not the same for different units, the value of the coefficient A for each link is different. The coefficient B in the proposed filter is entirely determined by the parameter of the analog prototype filter and therefore, when rebuilding both the band and resident, and the fourth delay line, 50 whose output is connected to the first subtractive input of the first algebraic adder, to the first and second subtractive inputs of the fourth algebraic adder, in series with the central frequency, the value of the coefficient is connected to the third multiplier and the even number for each link is constant. vertex multiplier whose output Apply the method of limiting evaluation, connected to the second subtraction input, get the value of the amplitude-first algebraic adder module, the frequency response of the fifth algebraic adder device, and the deceleration frequency: and (M - 0-A-AVSH1-4o 4sL2-4 -I). P k; | -A-SetW1) + (1 -ABJ) (4-8oi + 4 ABJ) (4-8Ы + 4о / г) .b; It can be seen from expression (6) that the value of the scaling factor set at the input of the filter does not depend on the values of the transfer function coefficients (6) (unlike the prototype) and does not change during filter rearrangement. So, if a multi-tune bandwidth (notch) filter is required, consisting of M links with N discrete bandwidths, then for the proposed technical solution, it is necessary to store in the ROM (M + 1) N coefficient values, while for the prototype you will need to store 3MN coefficient values. For M.1, the proposed filter provides a gain in the amount of ROM coefficients three times compared with the known ,, Thus, the proposed filter allows to simplify the process of bandwidth restructuring (notch) and provides an automatic scaling of the signal, which, in turn, allows to significantly reduce the amount of ROM allocated for storing coefficients and simplify the implementation of the filter. Invention Formula resident, and the fourth delay line, the melting input of which is the second input of the digital filter, the sixth algebraic adder, the output of which is the second output of the digital / filter, is distinguished by AND and in order to simplify the bandwidth adjustment, a fifth multiplier is inputted to the input of which the fifth algebraic adder, to the subtractive input of which the subtractive input of the second algebraic adder is connected, and the subtractive input of the sixth algebraic adder, the output of which is connected to the summing input of the second algebraic the adder, the second adder, the output of which is connected to the subtractive input of the third algebraic adder, is connected to the first and second inputs of the second adder of the fifth multiplier, which is the first output of the digital filter, the third multiplier input and the first summing input of the sixth algebraic adder, to the second totalizer the input of which is connected to the third input of the second adder, serially connected to the third adder, the first input of which is the first input of the digital filter, and the fourth total op, the output of which is connected to the second summing input of the third algebraic adder, the first input is to the summing input of the fourth algebraic adder, the output of which is connected to the summing input of the fifth algebraic adder and the second input of the fourth adder, and the second input of the third adder is connected to the output of the second multiplier . VkodPF about one about x dhadrf 1 Output PF s -one yo 1 1 Exit DR 20- FIG. 2