SU1449923A1 - Method of determining deviation of frequency from rated value - Google Patents

Abstract

The invention relates to a measurement technique and makes it possible to increase the resolution and shorten the measurement time. The device, which implements the method, contains dividers 1 and 2 with variable division ratio, relay-phase detector 3, driver 4 for control pulses, control blocks 5 and 6, phasing determiner 7, time counter 8 and frequency error calculator 9. The method is used when it is necessary to measure within a short time a very small frequency deviation from the nominal, and the frequencies of the measured and reference oscillator signals are different and their largest divider is much less than any of the frequencies. I il.

Description

4: 4; 0 CO ND CO

The invention relates to a measurement technique, is intended for comparing stable frequency sources with high accuracy and can be used in radio metering technology, communication technology and metrology, in cases when it is necessary to measure a very small frequency deviation from the nominal during a short time, and the frequencies of the signals of the measured and reference oscillators are different.

The purpose of the invention is to increase the resolution and reduce the measurement time.

The drawing shows a block diagram of an apparatus for implementing a method for determining a frequency deviation from a nominal value.

The output of the measured generator, f, is connected to the counting input of the first divider 1 with a variable division factor (DPKD), and the output of the reference generator .f, to the second SchKL-2, the outputs of the same PDKD are connected to the inputs of the relay-phase detector 3 (RFD) and the driver 4 control pulses (FIU), and the outputs of the RFD 3 are connected to other inputs of the FIU 4, the outputs of which are connected to the clock inputs of the control blocks 5 and 6, as well as to the skew determiner 7

ten

15

25 ZO

499232

(OCR), to the inputs of the time counter 8 (SchVr, the DCPD 1 and additional outputs of blocks 5 and 7 are connected, and the outputs of blocks 5 and 6 are connected to the DCPD control inputs, and the outputs of the Schvr 8 are connected to the control inputs of the control blocks 5 and 6 and a special output of the control unit 5 is connected to the OCR 7 input, the outputs of which are connected to the inputs of the frequency error calculator 9 (HF),

The method and operation of the device for its implementation is considered on the example of comparing the reference frequency f-5 MHz and the measured frequency f 8448 KFu-vuf, where Af is the frequency deviation from the nominal value, is determined with an accuracy of W fc. The DPKD division factors are pre-determined. To select these factors, use the Euclidean algorithm,

20

Obviously, the frequency comparison

g, and iv, di to pour

q ± ftf can occur at their common reference frequency f. In this case, f 8 kHz, To get the frequency

Otit

fc.

fg must be divided by N

 625, and G „e on M 1056,

The incomplete quotients q, determined by the Euclidean algorithm, (r - remainder)

M q, N + g. ,, 1056 1x625 + 431, q 1 N + r. 625 1x431 + 194, q 2. 1 r q, r, + r. 431 2x194 + 43, q 2 194 4x43 + 22, q4 4

q

to I +1

+ i-R.z 3 1x22 + 21, q.y 1 22 1x21 + 1, q 1 21 21x1 + 0, qy 21 With the help of incomplete quotients, the division factors at which the difference in the periods of the divided signals is minimal and equal to one quantum are determined.

one

one

C. f.

M-f.

b25 8448-10

one".

1056 5 10

L Oh, 19 NA.

The division ratio is by. solving an integer equation of first degree with two unknowns

N, m - Mn ± 1, where num is the desired value.

The value m is obtained from the recurrent value P, and n from the recurrent value Q,

P d ;; Q, 1; PC pg q.2 + 1;

Qj, 2 "

 P.-, q.K + P

k-2

Qi, QR -, - qk- + Q.-i at Kyz.

in this case

R,

 I

P4

Pf P

one.

2x2 5x4 + 2

Oh .. 1

0

O., 1x2 + 1 3 ZxA + 1 13

 2 2.

five;

22; 0.4

22x1 + 5 27; 0.5 - 13x1 + 3 - 16 .g - 27x1 + 22 49; Qg 16x1 + 13 29

because R 1056, and Q7 625, then, therefore, m 49, and p 29, Indeed: 625x49 - 1056x29 30625-30624 1

106: 4.17-25, but less than 26, so that after the 26th division of the front, the Front of the divided f will overtake

ten

Scientific research institute front, received from fit, and at the same time p

4.17x26 - 106 2.42 no.

Now, the division coefficients 16 and 27 are set once, which together with the coefficients 13 and 22 form the division coefficients 29 and 39 and then set the coefficient

So, the convergence of the fronts divided 5 divisions 29 and 49 and produced

exact convergence of fronts. 16 „„ 27

signals by the amount of one quantum occur when the division factor for f ,, is 29, and for f ,, equals 49.

But it is inconvenient to phase with such a small step and only in one direction, therefore coarse phasing conditions are selected by determining compensating division factors for phase drift over the comparison period with accurate phasing.

N .n-K + n, IM tK + t ,.

N is divided by n and the value of K 21 and the remainder of n, 16 are obtained, and, similarly to m, 27. When dividing by 16 and 27, the leading edge shifts by 21 quanta and in the opposite direction compared to dividing by 29 and 49. It is advisable to determine coarse phasing conditions and in the other direction pn, n t. mm; A2 13, and Pig 22. With these division factors, the front-edge shift is 22 quanta.

The process of measuring the frequency deviation from the nominal begins with a preliminary approach of the edges of the divided signals (phasing).

Let there be dephasing, is equal to 106 not, and the leading edge obtained from dividing fnjs is ahead of the leading edge of the divided reference signal by (- 106 not).

First, there is a rough approximation of the leading fronts and the division coefficients 13 and 22 are set in the DPCD (since the front f is ahead). After each division, the divided fronts draw together 4.17 not, since

 3o

--- 2.6

1e

ISS,

 2,60417 ISS.

1449923

106: 4.17-25, but less than 26, so that after the 26th division, the anterior Front of the divided f will precede

Scientific research institute front, received from fit, and at the same time p

4.17x26 - 106 2.42 no.

Now, the division coefficients 16 and 27 are set once, which together with the coefficients 13 and 22 form the division coefficients 29 and 39 and then set the coefficients

 divisions 29 and 49 and produced

exact convergence of fronts. 16 „„ 27

4 - 9

 3.2 ISS,

f

and

 3.19602 ISS or the shift of the fronts in the opposite direction is - 3.98 n.So that in the transition to division by 29 and 44 the shift is equal to

-3.98 + 2.42 -1.56 NS, 1.56: 0 ,, but less than 9.

In the ninth division, it is determined that a preliminary approach has occurred and the actual skew is less than 0.19:

0.19 X 9 - 1.56 1.71 - 1.56

 0.15 NS.

The next operation begins - the formation of the interval of exposure. In order for the measurement accuracy to be no worse than f ,, when measured with quanta equal to 0.19 NS, the exposure interval must be equal to 10 quantum 19 s.

When forming the dwell interval, it is necessary that the periods obtained from f and from n 8448 kHz be equal, i.e. The shutter speed interval must be a multiple of the comparison frequency period ---

e

 125 ISS. In the interval of exposure of such periods 152 thousand chi, i.e. 152 thousand times DPDD divided by 625 and 1056, respectively.

After the end of the second operation — the formation of a time interval — a third operation begins — a reapproach of the leading edges of the divided signals with an accuracy of one quantum.

X

Let the frequency f ...

10-v.

 f

P

- 37

X

„, Then after the end of the measurement interval

split the frequency f ,, and the loan through time & front divided

Frequency f.

vtcJfHt

S9 C + L

ilQ56ii52:; ig},

.f,

"1056 -152

- 19 s (1 + 37-10

L - (

(I-37

) + 0.15 topMc

and -ti

YU-)

NA or 7.18 NA,

where 0.15 NS is the initial error.

Since the front of the divided frequency f has entered, the division of the PDCD by 16 and 27 is established, at which the difference of the periods is -3.98 NA. After dividing twice (K 2), the position of the divided fronts will change when the phase difference is

-2x3.98 + 7.18 -7.96 + 7.18

 -0.78 NS.

After that, the division into 29 and 49 is established and K divisions are made until a new change in the relative position of the K 5 fronts.

5 X 0.19 - 0.78 0.17 no.

The next step is to calculate the magnitude of the frequency deviation by multiplying the magnitude of the difference between the periods of the divided pulses by the number of repeated divisions and the summation of the multiplication data

-K, - 2lS + KjS - 425 + 5S o - 37S 2 - (- 3.98ns) + 5 X

X 0.19 NS,

those. just the value that fibma eadan at the beginning, where 5 is the quantum value, is related to the measurement interval.

Thus, it is determined that if - 37-.

In a device that implements a method for determining the frequency deviation from the nominal value, the necessary division factors are set by commands from blocks 5 and 6, respectively. These commands are formed by decrypting signals from FUU 4 and Scr 8. Scr 8 determines the starting time of the approach, and the sequence of signals from Fi 4 determines the commands generated in blocks

4499236

5 and 6. In the FUU 4, control signals are formed in phase with the pulses arriving at the DPKD 1 and 2 outputs according to the RFD commands 3. If the leading edge of the divided pulse with the PDKD 1 advances the corresponding front with the PDKD-2, then set the period of divided pulses 10 in DPCD 1 is greater than that in DPCD 2, when the relative position of the fronts of the divided areas change. the pulses produce an exact approach of the fronts. With an accurate approach of the fronts, the periods of the divided 15 signals differ by one quantum. When changing the time interval of the mutual position of the fronts of the divided signals in the process of exact approaching, the end of the process of approaching is fixed. The convergence is performed at the beginning and end of the measurement interval, which is determined by ScVr8, which starts counting the number of DPDC 1 divisions by signals from block 5. 25 OCR 7 multiplies the data on the difference in periods of divided signals that come from block 5 to the number of such the periods supplied from the FIU 4. The start time of the smart-OQ quantization and summation is determined by the ScVr 8. The frequency deviation of the divided signals produces the HF 9 from the LFS data.

0

five

0

five

Claims (1)

Invention Formula The method of determining the frequency deviation from the nominal value, based on measuring the change in the time interval between the pulses of the reference and measured signals by the end of the lag interval generated from the pulses of the reference signal, by dividing the frequency of the reference signal f D °, is an integer number times lower than the nominal frequency of the measurement signal f ,, and the subsequent calculation of the frequency deviation as the ratio of the measured change in the time interval to the exposure interval, characterized in that, in order to increase azreshayuschey ability and shortening the measurement time, before the beginning of the form Rovani vscherzhki phasing interval beginning vscherzhki interval measured emym signal by sequentially extracting pairs reference pulses and the measured signals and opredele714499238 nor the sign of the relative position of the i-front signals with the minimum vyt-fronts, moreover, the sequential measured k, and when measuring the change division of a pair of pulses produced by the time interval between the pulse through n periods of the reference signal of the reference and measured signals and for the n periods of the measured signal, the pin is repeated at the end of the spacing interval. mN is pM ± k ,, where k is an integer, phasing of the measured and reference greater than or equal to one, and the signals are measured and summed at  , - end the phasing of the values of the differences in the intervals the moment of change of the sign of the compared Qfigrated pulse pairs.