RU2507681C2 - Method and apparatus for detecting nonlinear distortions caused by analogue-to-digital converter - Google Patents

Abstract

FIELD: radio engineering, communication.

SUBSTANCE: invention is based on change in the average value of a random process as a result of nonlinear conversion thereof. After passing through a device with a monotonous nonlinear characteristic, a random process with a zero average changes its spectral composition such that a constant component which depends on the degree of manifestation of nonlinearity arises therein. Thus, the technical result is achieved by measuring and analysing the average value of a digital code at the output of the investigated analogue-to-digital converter. The apparatus for detecting nonlinear distortions has a subtractor unit, a unit for measuring the magnitude of the average value of the digital code and a decision unit. In another version, the apparatus consists of a unit for measuring the average value of the digital code and a decision unit.

EFFECT: high accuracy of detecting arising distortions by detecting distortions before clipping begins.

6 cl, 3 dwg

Description

The invention relates to the field of radio measurements and is intended to control the operation of analog-to-digital converters without the use of special test signals.

A known method that can be used to detect non-linear distortion introduced by an analog-to-digital Converter (ADC) in the converted signal. The method provides a statistical assessment of one of the indicators of the output signal (digital code) for a certain time interval and comparing it with a predetermined threshold level, if the threshold level is exceeded, a decision is made on the presence of distortions. An informative indicator may be, for example, the variance of the output signal converted to a digital code. To implement the method, a device is used, consisting of a unit for statistical estimation of signal level and a comparator [US Pat. RU 2274868. Publ. 04/20/2006. Bull. No. 11].

The mentioned method, selected as a prototype, allows you to control the appearance of nonlinear distortions in the ADC without using test signals, extracting information from a working random signal with an arbitrary spectrum. However, its disadvantage is the low accuracy of detecting distortions, due to the fact that the method allows to detect non-linear distortions caused only by the amplitude limitation of the output signals, that is, in situations where the distortions reach sufficiently large values. In fact, it is a method of statistical control of the output signal level.

The technical result achieved by using the present invention is to increase the accuracy of detecting distortions by detecting distortions before the start of the amplitude limitation.

According to the first embodiment, the technical result is achieved by the fact that in the method for detecting non-linear distortions introduced by an analog-to-digital converter into a signal to be converted, comprising supplying an alternating signal with a zero mean to the input of an analog-to-digital converter, according to the invention, the digital output code is reduced by an amount equal to the value of the output code with a zero input signal, after which they measure the module of the average value of the digital code, compare it with a threshold value other than zero, and if enny module mean value of the digital code exceeds the threshold, then give a decision about the presence of distortions introduced by the analog-digital converter into the processed signal.

In this case, the threshold value is set based on the requirements for an acceptable level of nonlinear distortion.

In addition, to achieve a technical result according to the first embodiment, the device for detecting nonlinear distortions introduced by an analog-to-digital converter into a signal to be converted, according to the invention, comprises a subtraction unit, a unit for measuring the average value of the digital code and a decision unit, the output of which is the information output of the device , the first input of the subtraction block is the input of the device, the output of the subtraction block is connected to the input of the measurement block of the module of the average value of the digital code, the output of which connected to the input of decision block, the second input of the subtractor is input to a fixed value.

According to the second embodiment, the technical result is achieved by the fact that in the method for detecting non-linear distortions introduced by an analog-to-digital converter into a signal to be converted, comprising supplying an alternating signal with a zero mean to the input of the analog-to-digital converter, according to the invention, the average value of the digital code at the output of the analog-digital converter, compare it with a predetermined reference value, if the measured average value of the digital code differs from the reference value for more than If the permissible value is reached, then a decision is made on the presence of distortions introduced by the analog-to-digital converter into the converted signal.

In this case, the reference value is determined by a digital code at the output of the analog-to-digital converter with a zero input signal.

In addition, to achieve a technical result according to the second embodiment, the device for detecting non-linear distortions introduced by an analog-to-digital converter into a signal to be converted, according to the invention, comprises a unit for measuring the average value of a digital code and a decision unit, the output of which is the information output of the device, the output of the measurement unit the mean value of the digital code is connected to the input of the decision block, the input of the mean value measurement unit of the digital code is the input of the device.

The invention is illustrated graphic material.

Figure 1 is a graph of voltage versus time showing three cases of distortion in a device with a non-linear characteristic of a random signal represented by normalized values of U _o / U ₀ (U _o is the output voltage of the distorting device, U ₀ is the output voltage of the device at zero input signal).

Figure 2 and figure 3 shows the functional diagrams of devices for detecting non-linear distortions, highlighted by dashed lines.

In figure 1, three sections T ₁ , T ₂ and T ₃ (time intervals) are highlighted, each of which corresponds to a certain level of distortion.

The circuit of FIG. 2 contains a monitored ADC 1, a subtraction unit 2, a digital code average value measurement unit 3, and a decision unit 4, the output of which is the information output of the device, the first input of the subtraction unit 2 is the input of the device, the output of the subtraction unit 2 is connected to the input of block 3 of the measurement module of the average value of the digital code, the output of which is connected to the input of block 4 of decision making, the second input of block 2 of subtraction is the input of a fixed value Code "zero".

The circuit of FIG. 3 contains a monitored ADC 5, a block 6 measuring the average value of a digital code and a decision block 7, the output of which is the information output of the device, the output of a block 6 measuring the average value of a digital code is connected to the input of the decision block 7, the input of the block measuring the average The digital code value is the input of the device.

The basis of this group of inventions is the fact of changing the average value of a random process as a result of its nonlinear transformation. A random process with a zero mean after passing through a device with a monotonous nonlinear characteristic changes its spectral composition in such a way that a constant component appears in it, depending on the severity of the nonlinearity. The theoretical rationale for this is given in [Avanesyan G.R. Determination of nonlinear distortion of signals from measurements in the time domain // Journal of Radio Electronics, No. 2, 2011]. As an illustration, FIG. 1 shows three cases of passing a random process with a zero mean through a device with a non-linear characteristic. Various non-linear distortions were created with a constant non-linear characteristic by changes in the input signal level. As a result, in addition to the expected change in the level of the output signal, a change in its average value was recorded. Figure 1 shows that with an increase in non-linear distortions caused by an increase in the signal level, an increase in the average value of the output signal is observed. The smallest distortion, conventionally taken as the zero level, corresponds to the average value of the output signal U _cp1 obtained on the interval T ₁ . The signal in section T ₂ corresponds to the average value of U _cf2 , and the signal in section T _{3 corresponds to the} value of U _cf3 . Moreover, it is easy to see that U _cp1 <U _cp2 <U _cp3 .

, в котором N - количество значений выходного кода, участвующих в процессе усреднения. Сравнение с порогом Н, отличным от нуля, необходимо для задания некоторого коридора допустимых значений, что вызвано наличием на практике как погрешностей аналого-цифрового преобразования, так и погрешностей усреднения. Значение Н может составлять несколько единиц (например, соответствовать коду числа из диапазона 2…4) и зависит как от вышеуказанных погрешностей, так и от требований, предъявляемых к порогу выявления искажений. Чем выше допустимый уровень нелинейных искажений, тем выше должен выбираться порог H. Оптимальным является определение порога экспериментальным путем на этапе разработки устройства с учетом всех вышеуказанных факторов.It follows from the foregoing that the task of detecting nonlinear distortions arising in the ADC process of converting the input analog signal u _in with a zero average to a digital code u (n) can be reasonably reduced to measuring and analyzing the average value of the code at the output of the controlled ADC. Two options are possible. According to the first, before measuring the average value of the ADC output code, the zero level u ₀ is subtracted from the output code — the code at the ADC output corresponding to the zero input signal (“zero” code). This is necessary to remove the DC component at the ADC output, which is present at its output when working with bipolar signals and, as a rule, equal to half the range of possible values of the output code. For this, a subtraction unit 2 is provided (see FIG. 2), to the input of which is subtracted a code corresponding to the zero input signal is supplied. The specified code can be determined either experimentally before the start of the control of the ADC 1 in the absence of an input signal, or obtained from the documentation for the controlled converter. After the average value is measured in block 3 of the module, the result is compared in decision block 4 with a predetermined threshold N. If the threshold is equal or exceeded, a decision is made about the presence of distortion in the output code of the ADC controlled. That is, the principle of operation of the device is reduced to checking the inequality: $$\frac{\mathrm{one}}{N}{\displaystyle \sum _{i=\mathrm{one}}^N|\; |\; |u(n)-{\mathrm{<figure-callout\; id="0"\; label="u"\; filenames="00000004.png"\; state="\{\{state\}\}">u</figure-callout>}}_0|\; |\; |}\ge H$$

, in which N is the number of output code values involved in the averaging process. Comparison with a threshold H other than zero is necessary for specifying a certain corridor of acceptable values, which is caused by the presence in practice of both analog-to-digital conversion errors and averaging errors. The value of H can be several units (for example, correspond to a number code from the range 2 ... 4) and depends both on the above errors and on the requirements for the threshold for detecting distortions. The higher the permissible level of nonlinear distortion, the higher the threshold H should be selected. The optimal determination of the threshold experimentally at the device development stage, taking into account all the above factors.

или нет. Если указанное неравенство соблюдается, то выносится решение об отсутствии искажений, в противном случае выносится решение о наличии нелинейных искажений. Присутствующая в приведенном неравенстве величина Δu₀ призвана учесть допустимый диапазон изменений среднего значения, в пределах которого решение о наличии искажений еще не выносится. Правила выбора значения Δu₀ те же, что и величины Н.In the second embodiment, the operation of the device (see Fig. 3) is reduced to measuring the average value of the output code u (n) of the controlled ADC 5 and determining it in decision block 7: a double inequality $$u_0-\mathrm{<figure-callout\; id="0"\; label="\Delta \; u"\; filenames="00000004.png"\; state="\{\{state\}\}">\Delta \; </figure-callout>}{u}_{}{}_0<\frac{\mathrm{one}}{N}{\displaystyle \sum _{i=\mathrm{one}}^{N}u(n)<{\mathrm{<figure-callout\; id="0"\; label="u"\; filenames="00000004.png"\; state="\{\{state\}\}">u</figure-callout>}}_0+\mathrm{<figure-callout\; id="0"\; label="\Delta \; u"\; filenames="00000004.png"\; state="\{\{state\}\}">\Delta \; </figure-callout>}{u}_{}{}_0}$$

or not. If the specified inequality is observed, then a decision is made on the absence of distortions, otherwise a decision is made on the presence of non-linear distortions. The Δu ₀ value present in the given inequality is called to take into account the allowable range of changes in the average value, within which the decision on the presence of distortions is not yet made. The rules for choosing Δu ₀ are the same as N.

и логический нуль при его соблюдении. Что же касается программной реализации блока сравнения, то этот вариант предпочтителен при исполнении всего устройства на базе универсального микропроцессора. В этом случае проверка соблюдений вышеприведенных неравенств осуществляется достаточно просто с использованием стандартных алгоритмов.Regarding the implementation of the decision block, we note the following. When implementing the functions of this unit, in the first case, unit 4 can be a binary code comparator, one of the inputs of which (signal input) receives the code from the output of unit 3, and the other (reference input) is the threshold code N. The comparator inputs are selected as so that when threshold H is exceeded, the comparator is triggered. The output of the comparator will be the output of block 4. In the second case, concerning block 7, it is necessary to use two comparators, one of which will be triggered if the code at the signal input is larger than the code corresponding to the value of u ₀ -Δu ₀ supplied to the reference input and the other when exceeding the value of u ₀ + Δu ₀ supplied to its reference input. When combining the outputs of the two specified comparators according to the logic, EQUIVALENCE (EXCLUSIVE OR NOT) at the output of the logic element performing this operation, a logical unit will appear if the inequality $$u_0-\mathrm{<figure-callout\; id="0"\; label="\Delta \; u"\; filenames="00000004.png"\; state="\{\{state\}\}">\Delta \; </figure-callout>}{u}_{}{}_0<\frac{\mathrm{one}}{N}{\displaystyle \sum _{i=\mathrm{one}}^{N}u(n)<{\mathrm{<figure-callout\; id="0"\; label="u"\; filenames="00000004.png"\; state="\{\{state\}\}">u</figure-callout>}}_0+\mathrm{<figure-callout\; id="0"\; label="\Delta \; u"\; filenames="00000004.png"\; state="\{\{state\}\}">\Delta \; </figure-callout>}{u}_{}{}_0}$$

and logical zero if it is respected. As for the software implementation of the comparison unit, this option is preferable for the execution of the entire device based on a universal microprocessor. In this case, verification of compliance with the above inequalities is quite simple using standard algorithms.

Claims (6)

1. A method for detecting non-linear distortions introduced by an analog-to-digital converter into a signal to be converted, which comprises supplying an alternating signal with a zero mean to the input of an analog-to-digital converter, characterized in that the output digital code is reduced by an amount equal to the value of the output code with a zero input signal, then measure the average value module of the digital code, compare it with a threshold value other than zero, and if the measured average value module of the digital code exceeds the threshold ovoe value, then judge whether the distortions introduced by the analog-to-digital converter in the processed signal. 2. The method according to claim 1, characterized in that the threshold value is set based on requirements for an acceptable level of non-linear distortion. 3. A device for detecting non-linear distortions introduced by an analog-to-digital converter into a signal to be converted, characterized in that it contains a subtraction unit, a unit for measuring the average value of the digital code and a decision block, the output of which is the information output of the device, the first input of the subtraction unit is the input of the device, the output of the subtraction unit is connected to the input of the measurement unit of the module of the average value of the digital code, the output of which is connected to the input of the decision unit, the second input of the unit Itani is input to a fixed value. 4. A method for detecting non-linear distortions introduced by an analog-to-digital converter into a signal to be converted, which comprises supplying an alternating signal with a zero mean to the input of the analog-to-digital converter, characterized in that the average value of the digital code is measured at the output of the analog-to-digital converter, and it is compared with specified with a reference value, if the measured average value of the digital code differs from the reference value by more than an acceptable value, then a decision is made on whether there is a distortion s introduced by analog-to-digital converter into the processed signal. 5. The method according to claim 4, characterized in that the reference value is determined by a digital code at the output of the analog-to-digital converter with a zero input signal. 6. A device for detecting non-linear distortions introduced by an analog-to-digital converter into a converted signal, characterized in that it comprises a unit for measuring the average value of a digital code and a decision block, the output of which is an information output of the device, the output of the unit for measuring the average value of a digital code is connected to the input of the decision making unit, the input of the unit for measuring the average value of the digital code is the input of the device.

Images