RU2549248C1 - Method to measure time intervals and device for its realisation - Google Patents

Abstract

FIELD: measurement equipment.

SUBSTANCE: method to measure time intervals is based on analogue n-fold conversion of measured first time intervals between start signals and (m₁+1) signals of reference frequency into p times higher second time intervals to generated stop signals and coding of second time intervals in integer m₁ periods τ₀ of the reference frequency. At the same time they use two parallel channels of conversion of the sum of measured time interval with different values of the number m₁₁ of the sample intervals in the first and m₂₁ in the second channels, with different coefficients of expansion in discharge converters $$$ in the first and $$$ second channels. The start signal of t measured time interval is supplied, the dynamic range of values of which makes m periods of reference frequency, at the inlet of both channels simultaneously. The summary time intervals are converted into codes of counts {x_1ƒ(t)} and {x_2ƒ(t)}, comprising n₁, digits for the first channel and n₂ digits for the second channel. Codes of counts are fixed. Start signals (n₁+n₂) are supplied to inlets of channels or more inlet signals τ_x(t)^-, $$$ . In a similar manner they convert these time intervals and fix codes of counts $$$ . Weight coefficients are determined for digits of codes of counts $$$ for the first and $$$ for the second channel by the solution of the system (n₁+n₂) or more linear algebraic equations of type $$$ , and digital values are determined for measured time intervals $$$ according to the formula $$$ , where $$$ . The substance of the invention-device for measurement of time intervals: it comprises a reference frequency unit (RFU) and digit converters of time intervals into a code with an interpolator (DC_ij), the outlets of which are connected to a control unit of an inlet blocking unit (IBU). The device also comprises a calculator (C) and two parallel channels of digit-by-digit coding (DDCC1 and DDCC2), where DC_ij with various coefficients of expansion are included in series by start inlets and stop outlets. The RFU comprises three outlets. The outlet of the start reference frequency of the RFU is connected to the first control inlets of all DC_ij. The outlet of the first reference frequency of RFU is t times less than its start reference frequency, is connected to the second control inlets of the DC_ij in the DDCC1. The outlet of the second reference frequency of RFU shifted by phase to the half of the period relative to its first reference frequency, is connected to the second control inlets of DC_ij in DDCC2. Outlets of stop signals of final DC_ij in DDCC1 and in DDCC2 are connected accordingly to the first and second control inlets of the calculator C. The control outlet of the C is connected to the control inlet of the device IBU. The inlet of the start signal of measured time interval via the IBU is connected to the inlets of the first DC_ij in DDCC1 and DDCC2. Code outlets of all DC_ij are connected to information inlets of the calculator C, the outlet of which is the outlet of the device.

EFFECT: increased accuracy of measurement of short time intervals due to identification of weight coefficients of digits of a count code of time intervals in working mode with accuracy of setting periods of reference frequencies.

2 cl, 3 dwg

Description

The invention relates to the field of measuring information technology and is intended for use in areas where accurate and high-speed analog-to-digital signal conversion is needed.

Many tasks of analog-to-digital signal conversion are solved through the intermediate conversion of signals into a time interval. All problems of converting short time intervals shorter than the reference frequency period into a code are reduced to the task of measuring in time fractions of the reference frequency the time interval between a certain starting pulse signal and the following pulse signal of the reference frequency.

A known method of solving this problem [invention "A method of measuring time intervals based on analog conversion of the measured first time interval", patent of Ukraine No. 40629, author Gaysky V.A., publ. 08/15/2001 - Bull. No. 7], based on the analog conversion of the measured interval to p times the time interval before the formation of the stop pulse signal and bit coding of the extended time interval in whole periods of the reference frequency using bit conversion, the formation of the next least significant bits by repeating this procedure several times by taking the stop signal as next starting one. Since the range of the converted time interval is stored in each bit, it is formally possible to obtain arbitrarily many bits.

This method for the totality of features most closely matches the proposed technical solution, therefore, it is adopted as a prototype for each of the inventions included in the claimed group.

The prototype is characterized by the following characteristics that are common with the claimed method of measuring time intervals: it performs an analog n-fold conversion of the measured first time intervals between the start signals and the (m ₁ +1) -th reference frequency signals p times larger second second time intervals to the generated stop signals and encoding of second time intervals in integer m ₁ periods τ _{0 of the} reference frequency.

The disadvantage of the prototype is that due to the analog implementation of the operation to expand the time interval during bit conversion, which is usually performed by a push-pull integrator, the accuracy of the conversion is limited.

For the current level of technology, the accuracy of the push-pull integrator is limited to 10 ^-4 -10 ^-3 , i.e. such analog-to-digital conversion does not provide more than 10 ÷ 13 exact binary digits due to the error and changes in the weights of the digits of the reference code.

On the other hand, the achieved stability of the reference frequency period reaches 10 ^-12 and potentially provides encoding of time intervals with an accuracy of up to 40 bits.

The basis of the invention is the task of creating a method for measuring time intervals and a device for its implementation, which provides a single technical result - improving the accuracy of measuring short time intervals.

на входы обоих каналов одновременно, преобразования сумм измеряемого интервала

с разными значениями числа m₁₁ образцовых интервалов в первом и m₂₁ во втором каналах в коды отсчетов

из. n₁ разрядов в первом и

из n₂ разрядов во втором каналах, фиксации кодов отсчетов, повторения процедуры преобразования для (n₁+n₂)или более входных сигналов

определения цифровых значений измеряемых интервалов времени

по формулеThe problem is solved by identifying in the operating mode the weight coefficients of the digits of the code of the reference time intervals with the accuracy of setting the periods of the reference frequencies by using two parallel conversion channels with different expansion coefficients p _1j and p _2j in the bit converters, supplying the start signal of the t-measured time interval

to the inputs of both channels simultaneously, the conversion of the sums of the measured interval

with different values of the number m ₁₁ model intervals in the first and m ₂₁ in the second channel in the sample codes

of. n ₁ digits in the first and

of n ₂ bits in the second channel, fixing the sample codes, repeating the conversion procedure for (n ₁ + n ₂ ) or more input signals

determination of digital values of measured time intervals

according to the formula

 Where

для первого и

второго сигналов определяются решением системы (n₁+n₂) или более линейных алгебраических уравнений видаweight coefficients of digits of codes of samples

for the first and

the second signals are determined by the solution of the system (n ₁ + n ₂ ) or more linear algebraic equations of the form

и

where mτ ₀ is the period of the reference frequency, in fractions of which the sample codes are presented

and

τ ₀ - the value of one model time interval.

The objective of the invention is also solved by the fact that in a device for measuring time intervals containing a reference frequency unit and a bit converter of time intervals into a code with an interpolator, the output of which is connected to the control input of the input blocking unit of the device, it is new that it additionally contains a calculator and two parallel channel bitwise coding, in which bit converters with different expansion coefficients are connected sequentially at the start inputs and stop outputs, while it has three outputs to the reference frequencies, the output of the zero reference frequency of the block of reference frequencies is connected to the first control inputs of all bit converters, the output of the first reference frequency of the block of reference frequencies, m times smaller than the zero reference frequency, is connected to the second control inputs of the discharge converters of the first channel, output the second reference frequency of the block of reference frequencies, phase shifted by half the period relative to the first reference frequency, is connected to the second control inputs of the discharge converters of the second channel, you The stop signal odes of the final bit converters in the first and second channels are connected respectively to the first and second control inputs of the calculator, the control output of which is connected to the control input of the device input blocking node, the start signal of the measured time interval through the input block block is connected to the inputs of the first bit converters in both channels, the code outputs of all bit converters are connected to the information inputs of the computer, the output of which is the output m device.

The invention is illustrated with the help of illustrations, which depict: FIG. 1 is a timing diagram of the expansion of p times the measured interval on the push-pull integrator and bit coding on the counter; FIG. 2 is a signal diagram for the claimed two-channel bitwise coding of time intervals; FIG. 3 is a structural diagram of the claimed device.

The invention consists in the following.

и кодируется целым числом периодов x_j (здесь x_j=1) опорной частоты.The known known method of encoding a short time interval using an interpolator based on a push-pull integrator, which provides an extension of the measured interval p times to a stop signal and encoding this interval in whole periods of the reference frequency, is illustrated in FIG. 1. Here, the measured time interval τ _xi between the start pulse t ₀ and the first following signal T _{1 of the} reference frequency with a period τ ₁ is converted to an extended p _j time interval

and encoded by an integer number of periods x _j (here x _j = 1) of the reference frequency.

и цикл преобразования интервала τ_x(i+1), повторяется.Next, the stop signal τ _{1 is} taken as a new start signal for the interval τ _{x (i + 1),} which is an addition to τ ₁ , of the segment

and the transformation cycle of the interval τ _{x (i + 1) is} repeated.

This coding method is used in each channel.

In the proposed method, two parallel channels with different weighting coefficients of the coding bits are used.

mτ₀, от последовательности Т₁. Последовательности Т₀, Т₁ и Т₂ синхронизированы. Преобразуемый интервал времени задается стартовым сигналом t₀, который возникает внутри периодов последовательностей Т₁ и Т₂ и длится до стоповых первых следующих сигналов Т1 и Т2 в первом и втором каналах. Эти интервалы далее кодируются в долях периода mτ₀.Timing diagrams of signals during two-channel implementation of the method of bitwise coding of time intervals are presented in FIG. 2. The following notation is used here: a sequence of pulses T ₀ (a) of the initial reference frequency with a period τ ₀ , T ₁ , is a sequence of pulses of the first reference frequency T ₁ (b) with a period mτ ₀ , T ₂ is a sequence of pulses of the second reference frequency T ₂ (2) with a period mτ ₀ , out of phase by a period, i.e. on $$\frac{\mathrm{one}}{2}$$

mτ ₀ , from the sequence T ₁ . The sequences T ₀ , T ₁ and T _{2 are} synchronized. The converted time interval is set by the start signal t ₀ , which occurs inside the periods of the sequences T ₁ and T ₂ and lasts until the stop first signals T1 and T2 in the first and second channels. These intervals are further encoded in fractions of the period mτ ₀ .

Consider the conversion process in the first channel, the plot of FIG. 2 (c). The start signal t ₀ starts the first integrator cycle, which lasts for the converted interval τ ₁₁ , to the first stop signal T ₁ . This interval contains m ₁₁ integer intervals τ ₀ , which are calculated (m ₁₁ +1) by signals T ₀ and are taken for the duration m ₁₁ τ _{0 of the} model interval τ ₀₁ , additive to the unknown interval τ _xl . Wherein

In the second integrator cycle, the interval τ ₁₁ is multiplied by the weight coefficient of the first discharge p ₁₁ , i.e. an interval of duration is formed

определяется из выражения In this time interval is embedded N ₁₁ , the pulses of the reference frequency T ₀ . In integer fractions of the interval mτ _{0, the} value x _{11 of the} first digit of the reference code

determined from the expression

which covers the segment x ₁₁ mτ _{0 of the} converted interval p ₁₁ τ ₁₁ , and the time interval Δ ₁₁ remains unknown, i.e.

The time interval Δ ₁₁ is an addition to the full interval mτ _{0 of the} interval τ ₁₂ so that it is possible to write

Then we can write

, т.е.The time interval τ ₁₂ is the initial one for the conversion of the second discharge, which is carried out similarly to the first with a weight

, i.e.

All subsequent discharges are formed similarly, therefore

пренебрегаем за малостью.with a residual member

neglected a little.

сигнала Т₂ опорной частоты относительно сигнала Τ₁, фиг. 1 (г), преобразуется в интервал времениThe waveforms in the second conversion channel are shown in FIG. 2 (d, d). The start signal t ₀ is common to two channels. However, in the second channel having a phase shift

the signal T _{2 of the} reference frequency relative to the signal Τ ₁ , FIG. 1 (g) is converted to a time interval

A known portion of this interval is taken as the reference interval.

and the unknown parts of the converted intervals τ _xl coincide in both channels.

By analogy with expression (11) for the second transformation channel, we write

For the converted intervals at different time instants τx (t) and two channels, we obtain

Subtracting the second from the first equation, we obtain

System of equations (17) contains (n ₁ + n ₂ ) unknowns corresponding to the weights of the digits of the samples of the converted intervals. The extended matrix of the system for the minimum (n ₁ + n ₂ ) number of equations has the form

The possibilities of solving system (17) are due to the formation of members of the common matrix (18) due to the variability of the converted signal or the variability of model signals.

разных разностей

образцовых сигналов или

разных сигналов. С другой стороны, если изменчивость матрицы обеспечивается только изменчивостью входного преобразуемого интервала, то достаточно, видимо, одной пары образцовых сигналов для одного момента времени.If the variability is provided only by changing the reference signals (their difference), then this will actually be the calibration mode and at least

different differences

reference signals or

different signals. On the other hand, if the variability of the matrix is provided only by the variability of the input transformable interval, then, apparently, one pair of model signals for one moment in time is sufficient.

The change of model signals can be performed by a phase shift by a discrete number of intervals τ _{0 of the} first pulses of the signals Τ ₁ and T ₂ , and a shift to the left decreases m _i1 , and a shift to the right increases m _i1 . In this case, the determination of m _i0 that is ahead of this operation allows you to set the upper boundary of the shift to the left to m _i0 = 0, it is equal to

The right shift boundary is limited by the maximum allowable duration of the first integrator clock cycle, i.e.

The bit coding operation can be performed using a single bit converter, closing the stop signal output to the start input. The weight of the j-ro discharge will be equal to p ^-j and the conversion time of all n bits will be up to (p + 1) n τ ₁ . Only after such a period of time can the start signal of the next measured interval be applied to the input.

а период времени, через который можно подавать на вход стартовый сигнал следующего измеряемого интервала, сократится до 2(p+1) τ₁.To improve performance, you can use n bit converters that are connected in series in the chain, with the first converter generating the code for the first bit, the second for the second, and the jth for j-ro. The weight of the j-ro discharge will be equal to

and the period of time after which the start signal of the next measured interval can be applied to the input will be reduced to 2 (p + 1) τ ₁ .

причем коэффициенты расширения в разрядах p_1j в первом 1 и p_2j во втором 2 каналах должны быть разными

Коэффициенты расширения в разрядах p_1j одного канала могут быть одинаковыми, но таковыми никогда не будут из-за технологических погрешностей, ухода от влияющих факторов и старения. Поскольку коэффициенты расширения p_1j задают веса разрядов

или основание позиционной системы счисления, в которой кодируется измеряемый потенциал, то при выборе их номинальных значений могут быть приняты известные соображения простоты реализации (р=2) или максимального быстродействия (p=4) [прототип]. Если принято для номинальных значений

и

для всех j, то число разрядов (и разрядных преобразователей) в первом n₁ и втором n₂ каналах может быть установлено из соотношения

которое обеспечивает примерное равенство разрешающей способности каналов.The composition of the claimed device (Fig. 3) includes the first 1 and second 2 channels of bitwise coding (KPK1, KPK2, respectively), a block of 3 reference frequencies — a generator (GOCH), a computer 4 (B), a device input blocking unit 5 (BWM). Bitwise coding channels 1 and 2 are formed by a serial connection of bit converters

moreover, the expansion coefficients in the discharges p _1j in the first 1 and p _2j in the second 2 channels must be different

The expansion coefficients in the discharges p _{1j of} one channel can be the same, but they will never be due to technological errors, avoidance of influencing factors and aging. Since the expansion coefficients p _1j determine the weight of the digits

or the base of a positional number system in which the measured potential is encoded, when choosing their nominal values, well-known considerations of ease of implementation (p = 2) or maximum speed (p = 4) can be adopted [prototype]. If accepted for nominal values

and

for all j, then the number of bits (and bit converters) in the first n ₁ and second n ₂ channels can be established from the relation

which provides an approximate equality of the resolution of the channels.

для

The reference frequency generator (GOC) 3 is used to generate pulse sequences T ₀ , T ₁ , T ₂ , the period of which is connected by the relations for Τ ₁ -

for

All sequences are synchronized by the sequence T ₀ , the sequence T ₂ is phase shifted by 0.5 mτ ₀ relative to the sequence Τ ₁ (figure 2).

с другой стороны - должно быть достаточным для формирования необходимого числа разных образцовых интервалов при идентификации

весовых коэффициентов разрядов в режиме градуировки, т.е.

The value of m must satisfy, on the one hand, the dynamic range of values of the measured time interval

on the other hand, it should be sufficient to form the required number of different model intervals for identification

weight coefficients of discharges in the calibration mode, i.e.

и

отсчетов последовательности преобразуемых интервалов τ_x(t),

решения системы линейных алгебраических уравнений (3) и восстановления цифровых значений τ_x(t) измеряемых интервалов по выражениям (1) и (2). По общей шине вход вычислителя 4 связан с выходами кодов всех разрядных преобразователей устройства. Узел 5 блокировки входа (УБВ) предназначен для отключения входов 1-го и 2-го каналов от стартового сигнала до окончания преобразования предыдущего сигнала в первых двух разрядах каждого из каналов в тех случаях, когда время его поступления не может быть синхронизовано с работой устройства. В аналого-цифровых преобразователях с промежуточным преобразованием в интервал времени такая синхронизация возможна и необходима и может выполняться этим узлом. Внешний вход устройства является входом узла 5 УБВ, а выход последнего подан на входы каналов 1,2.In the operating mode, when the variability of the converted interval is ensured by a sequence of different measured intervals, the value of m can be minimal (m = 2). The output of the reference sequence T _{0 is} applied to all the units of the device, the output T _{1 is} supplied to the bit converters of the first channel 1, T ₂ - of the second channel 2. The computer 4 (B) is designed to accumulate codes

and

samples of the sequence of converted intervals τ _x (t),

solving a system of linear algebraic equations (3) and reconstructing the digital values τ _x (t) of the measured intervals from expressions (1) and (2). On a common bus, the input of the calculator 4 is connected with the outputs of the codes of all bit converters of the device. Node 5 of the input blocking (BWM) is designed to disconnect the inputs of the 1st and 2nd channels from the start signal until the conversion of the previous signal in the first two bits of each channel is completed in those cases when its arrival time cannot be synchronized with the operation of the device. In analog-to-digital converters with intermediate conversion to a time interval, such synchronization is possible and necessary and can be performed by this node. The external input of the device is the input of the node 5 UBV, and the output of the latter is fed to the inputs of channels 1.2.

и

The device, according to the diagrams of the signals (Fig. 2), works as follows. The start signal t _{0 is} fed to the input of the BSS node 5 and, if it is open, passes to the inputs of the first bit converters of the first 1 and second 2 channels, respectively, RP ₁₁ and ΡΠ ₂₁ . In the discharge converters, the charge is accumulated and m ₁ and m _{2 are} counted for reference intervals τ ₀ in the first integrator cycle from the start signal to the first pulse T ₁ in the first 1 and T ₂ in the second 2 channels. These will be time intervals

and

In the second integrator cycle, the intervals τ ₁₁ are expanded by p ₁₁ times in the first channel 1 and the interval τ _{21 by} p ₂₁ times in the second channel 2 with simultaneous coding in integer numbers of the interval mτ ₀ x ₁₁ in the first channel and x ₂₁ in the second channel.

других разрядов в обоих каналах, поступают в вычислитель 4 и накапливаются. Причем каждый j-й разряд кодов отсчета интервалов τ₁₁(t)и τ₂₁(t) формируется j-м разрядным преобразователем РП_j.Codes m ₁₁ and m ₂₁ are formed only in the first digits. Further, the codes m ₁₁ , m ₂₁ , x ₁₁ x ₂₁ enter the calculator 4. Similarly, codes are generated

other discharges in both channels, enter the calculator 4 and accumulate. Moreover, each j-th digit of the reference codes of the intervals τ ₁₁ (t) and τ ₂₁ (t) is formed by the j-th bit converter RP _j .

After the first two bit converters in both channels are released, which is fixed by the calculator 4, the input blocking unit 5 opens the input for the start signal of the next measured interval τ _x (t + 1).

и

последовательности измеряемых интервалов

In the calculator 4 accumulates (n1 + n ₂ ) or more pairs of samples

and

sequence of measured intervals

по выражениям (2) определяются цифровые значения неизвестных частот

и

измеряемого интервала в первом и втором каналах и их среднее значение

по выражению (1).Next, the SPA is solved, according to expression (3), the weight of the discharges is determined

according to expressions (2), digital values of unknown frequencies are determined

and

the measured interval in the first and second channels and their average value

by expression (1).

Claims (2)

1. A method of measuring time intervals, based on analog n-fold conversion of the measured first time intervals between the start signals and the (m ₁ +1) -th signals of the reference frequency p times large second time intervals to the generated stop signals and encoding the second time intervals in whole ₁ m periods τ ₀ reference frequency, characterized in that the two parallel channels for converting the sum of the measured time interval with different values of the number m of intervals model ₁₁ in the first and in ₂₁ m sec m channels with different coefficients of expansion bit converters

in the first and

in the second channels, the start signal of the t-th measured time interval is fed, the dynamic range of which is m periods of the reference frequency, at the input of both channels simultaneously, the total time intervals are converted into reference codes

containing n ₁ bits for the first channel and n ₂ bits for the second channel, fix the codes of samples, apply to the inputs of the channels (n ₁ + n ₂ ) or more input signals of time intervals

similarly convert these time intervals and fix the code codes

determine the weighting coefficients of the digits of the codes of samples

for the first and

for the second channel by solving the system (n ₁ + n ₂ ) or more linear algebraic equations of the form

and determine the digital values of the measured time intervals

according to the formula

2. A device for measuring time intervals, containing a reference frequency unit and a bit converter of time intervals into a code with an interpolator, the output of which is connected to the control input of the device input blocking unit, characterized in that it additionally contains a computer and two parallel bit-coding channels, in which bit converters with different expansion coefficients are connected in series at the start inputs and stop outputs, while the block of reference frequencies has three outputs, the output of of the total reference frequency of the block of reference frequencies is connected to the first control inputs of all bit converters, the output of the first reference frequency of the block of reference frequencies, m times less than the initial reference frequency, is connected to the second control inputs of the discharge converters of the first channel, the output of the second reference frequency of the block of reference frequencies, shifted in phase for half the period relative to the first reference frequency, connected to the second control inputs of the discharge converters of the second channel, the outputs of the stop signals of the final discharge transducers in the first and second channels are connected respectively to the first and second inputs of the computer, the control output of which is connected to the control input of the input blocking unit of the device, the input of the start signal of the measured time interval through the input blocking input is connected to the inputs of the first bit converters in both channels, code outputs all bit converters are connected to the information inputs of the computer, the output of which is the output of the device.

Images