RU2495376C1 - Emission source with low level of intensity noises for fibre-optic gyroscope - Google Patents

Abstract

FIELD: measurement equipment.

SUBSTANCE: device includes a source, the optic emission from the output of which is divided into several propagation channels of various length. Source output is connected to the first input of the first divider of row of N dividers. Every first output of the previous divider of the row is connected to the first input of every next divider. The second input of each divider is connected by means of a light-emitting diode with its second output. Length of each light-emitting diode is more than length of LK coherence of the source emission and more or less than length of the light-emitting diode of every next divider of that row by length of LK coherence of the source emission. Length of the light-emitting diode connecting the second input and output of the first divider of row of N dividers is more than NLK.

EFFECT: reduction of intensity noises of an optic emission source.

2 cl, 5 dwg

Description

The invention relates to the field of fiber optics and can be used in the design of fiber-optic gyroscopes and other sensors of physical quantities based on single-mode optical fibers.

The fiber-optic gyroscope (hereinafter referred to as VOG) contains a fiber optic ring interferometer and an electronic information processing unit. An optical fiber interferometer contains an optical radiation source, a fiber radiation power divider, an integrated optical circuit (hereinafter - IOS), a multi-turn sensitive coil and a photodetector. IOS contains in its composition a Y-divider of the power of optical radiation based on polarizing channel waveguides and a phase modulator located on the output arms of the Y-divider. Channel waveguides of the Y-divider are formed in a lithium niobate substrate using proton-exchange technology, which allows the waveguides to acquire polarizing properties. To the output channel waveguides of the Y-divider, the ends of the optical fibers of the sensitive gyro coil are docked.

An interference pattern is formed on the photodetector of a ring fiber-optic interferometer, formed by two optical beams that have passed the sensitive coil of the gyroscope in two opposite directions. When the ring interferometer rotates between these two beams, due to the Sagnac effect, a phase difference arises, which is expressed as follows:

ϕ _S = [4πrl / λc] × Ω

where R is the radius of the sensitive coil of the gyroscope;

L is the fiber length of the coil;

λ is the central wavelength of the radiation source;

c is the speed of light in vacuum;

Ω is the angular velocity of rotation of the gyroscope.

Thus, at the photodetector, the intensity of optical radiation can be represented as:

I _Ф = 1 / 2P ₀ (1 + cosϕ _S )

where P ₀ is the power of the rays interfering at the photodetector.

To increase the sensitivity of VOG near zero angular velocities, auxiliary phase modulation is used. To achieve the effect of phase modulation of rays in a ring interferometer using an IOS phase modulator, we use the time delay of the ray fronts interfering on the photodetector while passing through the IOS phase modulator. This time lag amounts to:

$$\tau =\frac{L{n}_0}{c}$$

where n ₀ is the refractive index of the material of the fiber of the sensitive coil.

When applying the following voltage pulses to the phase modulator with a frequency of 1 / 2τ t, a signal of the form is observed at the output of the synchronous detector:

$$U_{\mathrm{at}x}^{\mathrm{FROM}D}\sim P_0\left[\mathrm{one}+\cos {\phi }_m\cdot \cos {\phi }_S\pm \sin {\phi }_m\cdot \sin {\phi }_S\right]$$

To ensure a large dynamic range of measuring angular velocities and to obtain a high linearity of the FOG output characteristic in the optoelectronic information processing circuit, the so-called compensation method for reading the phase difference of the rays is used, the essence of which is that a compensating phase difference is supplied to the phase modulator along with the voltage of the auxiliary phase modulation Sagnac sawtooth step voltage [1]. As a result, the signal at the output of the synchronous detector takes the following form:

$$U_{\mathrm{at}x}^{\mathrm{FROM}D}\sim P_0\left\{\mathrm{one}+\cos {\phi }_m\cdot \cos \left[{\phi }_S-{\varphi }_K\right]\pm \sin {\phi }_m\cdot \sin \left[{\phi }_S-{\varphi }_K\right]\right\}$$

where φ _K is the phase shift introduced by the sawtooth voltage to compensate for the Sagnac phase difference.

Given that in the compensation mode ϕ _S -φ _K ≈ 0, the voltage at the input of the synchronous detector can be represented:

$$U_{\mathrm{at}x}^{\mathrm{FROM}D}\sim P_0\left\{\mathrm{one}+\cos {\phi }_m\pm \sin {\phi }_m\cdot \sin \left[{\phi }_S-{\varphi }_K\right]\right\}$$

It is known that the main components of VOG noise that determine its sensitivity are the following:

- shot noise of the photodetector;

- thermal noise of the preamplifier of the photodetector;

- noise intensity of the source of optical radiation. The sensitivity of VOG in terms of shot noise can be represented as:

$${\Omega }_{\min }^{dR}=2.05\cdot {10}^5\frac{c\lambda }{\mathrm{four}\pi RL}\sqrt{\frac{2hc}{\lambda }}\cdot \frac{\mathrm{one}}{\sqrt{P_0}}\sqrt{B}\cdot \frac{\mathrm{one}}{\sin \left({\varphi }_{m/2}\right)}\left[gR\mathrm{but}d./h\mathrm{but}\mathrm{from}\right]$$

- минимально обнаруживаемая угловая скорость по уровню дробовых шумов;Where $${\Omega }_{\min }^{dR}$$

- the minimum detectable angular velocity in terms of shot noise;

h is Planck's constant;

B is the bandwidth of the electronic information processing path [Hz].

The sensitivity of VOG in terms of thermal noise of the pre-amplifier of the photodetector can be represented as:

$${\Omega }_{\min }^{\mathrm{uh}le\mathrm{to}tR}=2.05\cdot {10}^5\frac{c\lambda }{\mathrm{four}\pi RL}\sqrt{\frac{\mathrm{four}kT}{R_H}+2e{I}_T}\cdot \frac{\mathrm{one}}{P_0}\sqrt{B}\cdot \frac{\mathrm{one}}{\sin \left({\varphi }_m\right)}\left[gR\mathrm{but}d./h\mathrm{but}\mathrm{from}\right]$$

- минимально обнаруживаемая скорость по уровню шумов предварительного усилителя;Where $${\Omega }_{\min }^{\mathrm{uh}le\mathrm{to}tR}$$

- the minimum detectable speed according to the noise level of the pre-amplifier;

k is the Boltzmann constant;

T is the absolute temperature in degrees Kelvin;

R _H is the load resistance of the pre-amplifier;

e is the electron charge;

I _T is the dark current of the photodetector.

The sensitivity of the FOG with a closed feedback loop by the noise level of the radiation source intensity can be represented as:

$${\Omega }_{\min }^{\mathrm{and}sl}=2.05\cdot {10}^5\frac{\lambda c}{\mathrm{four}\pi RL}\sqrt{\frac{{\lambda }^2}{c\Delta \lambda }}\sqrt{B}\cdot ctg\left({\varphi }_{m/2}\right)\left[gR\mathrm{but}d./h\mathrm{but}\mathrm{from}\right]$$

- минимально обнаруживаемая угловая скорость по уровню шумов интенсивности источника излучения;Where $${\Omega }_{\min }^{\mathrm{and}sl}$$

- the minimum detectable angular velocity according to the noise level of the intensity of the radiation source;

Δλ is the width of the radiation line of the source.

Below are formulas for calculating the optical noise of a fiber-optic gyroscope, depending on its main characteristics.

$$`{\Omega }_{\min }^{dR}=R{W}_{dR}=\left[6.42/RL\right]\cdot {10}^{-5}\cdot \mathrm{one}/\left\{{\left(P_0\right)}^{\mathrm{one}/2}\sin \left[{\varphi }_m/2\right]\right\}gR\mathrm{but}d./h\mathrm{but}{\mathrm{from}}^{\mathrm{one}/2}$$

$$`{\Omega }_{\min }^{\mathrm{uh}le\mathrm{to}tR}=R{W}_t=\left[\mathrm{1,61}/RL\right]\cdot {10}^{-5}\cdot \mathrm{one}/\left\{R_n{}^{\mathrm{one}/2}{P}_0\sin \left[{\varphi }_m\right]\right\}gR\mathrm{but}d./h\mathrm{but}{\mathrm{from}}^{\mathrm{one}/2}$$

$$`{\Omega }_{\min }^{\mathrm{and}sl}=\left[1.13/RL\right]\cdot {10}^{-5}{\left[\Delta \lambda \right]}^{-\mathrm{one}/2}ctg{\varphi }_m/2gR\mathrm{but}d./h\mathrm{but}{\mathrm{from}}^{\mathrm{one}/2}$$

For the FOG parameters RL = 40, P ₀ = 6 × 10 ^-4 W, λ = 1550 nm, Δλ = 20 nm, ϕ _m = 7/8 π radians, R _n = 100 kOhm, the shot noise level will be 6.4 × 10 ^-5 degrees / hour ^1/2 , the noise level of the photodetector current amplifier is 9.4 × 10 ^-6 degrees / hour ^1/2 , and the noise level of VOG due to noise intensity of the radiation source will be 4 × 10 ^-4 degrees ./hour ^1/2 . Thus, even with a large amplitude of auxiliary phase modulation, the closed-loop VOG noise level determined by the noise of the radiation source intensity significantly exceeds the shot noise of the photodetector and the thermal noise of the amplifier.

The closest in technical essence to the proposed invention to reduce noise intensity of the radiation source is the invention described in [1]. The radiation from the source output is divided into several independent propagation channels, and then the radiation is summed at the site of the photodetector. The length difference of each radiation propagation channel is consistently different from each other by no less than the length of the coherence of the radiation. As a result, several incoherent rays are summed at the photodetector, which leads to a decrease in the noise of the radiation intensity at the photodetector while maintaining its total power. The main disadvantage of the known method for reducing intensity noise is that the optical beam at the output of any interferometric sensor based on a single-mode fiber is divided into several beams that arrive at the site of the photodetector with a time delay relative to each other and therefore the interference pattern is recorded by the photodetector, carrying information about the measured value at different points in time, which can introduce significant errors, for example, when measuring the angular velocity of fiber-optic by the microscope. Thus, when using the known method of reducing intensity noise, the accuracy gain of the fiber-optic gyroscope due to this can be reduced due to the loss of its accuracy due to the delay of the rays carrying information about the measured value relative to each other when they hit the site of the photodetector. Another type of device proposed by the authors of the well-known patent is two fiber polarizing splitters and two segments of a fiber with large birefringence. When using this device, errors in the measurement of the angular velocity do not occur, since all the rays that carry information about the same angular velocity arrive at the photodetector site simultaneously. But this type of device fundamentally has a loss of optical power of at least 3 dB, which in turn leads to a decrease in the accuracy of the gyroscope due to an increase in shot noise of the photodetector.

The aim of the present invention is to improve the accuracy of a fiber-optic gyroscope by reducing the noise of the intensity of the optical radiation source while eliminating the delay of the rays relative to each other when they hit the site of the photodetector.

This goal is achieved by the fact that the source output is connected to a series of N pieces of optical power dividers, where N = 1, 2, 3, ... .. with a division coefficient of each of these dividers along the direct channel k, where k can take digital values greater than 0 and less than 1, and the source output is connected to the first input of the first divider from a number of dividers, and each first output of the previous divider of the series is connected to the first input of each subsequent divider, while the second input of each divider is connected using its fiber with its second output, and the length of each Each of these fibers is longer than the coherence length L _{to the} source radiation, and the length of the fiber connecting the second input and output of each previous divider from the number N of dividers is greater or less than the length of the fiber of each subsequent divider from this series, at least by the coherence length L _{to the} source radiation, and the length of the fiber connecting the second input and output of the first divider of a series of N dividers is greater than NL _to .

2. The radiation source according to claim 1, characterized in that the output of the source is connected to the first input of the first divider from a number of dividers, and each second output of the previous divider of the series is connected to the first input of each subsequent divider, while the second input of each divider is connected using a fiber with your first exit.

A reduction in intensity noise and, as a result, an increase in the accuracy of a fiber-optic gyroscope is achieved by dividing the radiation into a large number of independent sources when the radiation is circulated through the channels of the optical power dividers. The proposed technical solution involves the separation of the radiation of one source into several channels at the input of the optical circuit of a fiber-optic gyroscope, and thus several independent beams that carry information about the same angular velocity simultaneously fall on the photodetector site. Thus, when dividing one beam into several independent rays at the input of the optical circuit, the FOG does not introduce additional errors when measuring the angular velocity. An increase in the level of shot noise also does not occur, since the radiation propagates through the channels of the dividers with virtually no loss.

The invention is illustrated by drawings. Figure 1 presents the radiation source with the division of radiation into four propagation channels. Figure 2 presents the source of optical radiation with a reduced level of noise intensity based on one fiber divider type 2 × 2. Figure 3 presents a graph of the dependence of the noise reduction coefficient in a fiber optic gyroscope when using a radiation source with one divider. Figure 3 presents a graph of the dependence of the noise reduction coefficient in a fiber optic gyroscope when using a radiation source with one divider. Figure 4 presents the source of optical radiation with a reduced level of noise intensity based on N fiber dividers of optical power type 2 × 2. Figure 5 presents the source of optical radiation with reduced noise of intensity based on N fiber dividers of optical power type 2 × 2 with another connection of the fiber ends of the dividers.

Figure 1 shows the source of broadband optical radiation 1 for a fiber optic gyroscope. This source can be made based on erbium fiber. The advantages of such sources are its high output power, which provides a low level of shot and thermal noise of the photodetector current amplifier. But the width of the emission line is not large enough (Δλ = 20–25 nm), and therefore the intensity noise of these sources worsens the noise characteristics of fiber-optic gyroscopes by more than 10 times. One of the ways to reduce intensity noise [1] is the separation of radiation using a divider 2 (Figure 1), for example, into four channels 3, 4, 5, 6 of the propagation of radiation. The radiation from the outputs 7, 8, 9, 10 then comes to the site of the photodetector. The lengths of the distribution channels L ₁ , L ₂ , L ₃ , L ₄ differ from each other by more than the coherence length of the radiation source. Consider the usual addition of powers of many identical sources. This task will clearly illustrate the very idea of reducing noise when using multiple sources.

So, we consider the addition of powers of N identical sources, each of which has a constant (p _{0, n} ) and noise (δp _n ) components:

p _n (t) = P _{0, n} + δp _n (t).

We assume that the noise components from different sources do not correlate with each other. We also believe that each individual noise component δр _n from the nth source is white noise. The numerical characteristic of this noise in this case is its dispersion

, $${\sigma }_n=\sqrt{〈\delta p_n^2\left(t\right)〉}=\sigma$$

,

the same for all sources with numbers n = 1 ... N. As a simple calculation shows, if we summarize the powers of all these sources, then the variance σ _{0 of the} total noise component is

$${\sigma }_0=\sigma \sqrt{N}$$

Thus, the dispersion of the total power from N sources is N ^1/2 times greater than the dispersion of the power of any single source. However, at the same time, the total constant component is N times greater than that of a single source. Thus, the noise of the total power of all sources improves by N ^1/2 times. This resembles a picture of fluctuations in the density of an ideal gas in a certain volume of space, which also decrease as N ^1/2 , only here N is the average number of gas molecules in a given volume. Thus, when the radiation propagating through four channels is added to the photodetector, provided that their powers are the same, it is possible to reduce the intensity noise by a factor of 2 (when dividing the radiation propagation through four channels). The fiber optic gyroscope (FOG) consists of a fiber ring interferometer (FRI) and an electronic information processing unit. The radiation source is part of the FRI, which is made from single-mode optical fibers. Therefore, to improve the noise characteristics of the VOG, it is necessary to introduce radiation propagating through four channels into the input single-mode fiber of the optical scheme of the FRI VOG, but in this case, large losses of optical power of all four sources inevitably occur. Therefore, a decrease in VOG noise due to a decrease in intensity noise can be compensated by a negative increase in shot noise of the photodetector and thermal noise of the amplifier due to a decrease in the power of the useful VOG signal.

Figure 2 presents the source of optical radiation with a reduced level of noise intensity based on one fiber divider 11 type 2 × 2. The divider can be made according to the alloy technology of two segments of single-mode optical fibers. For this, two segments of the fiber are parallel to each other and in their middle part the fibers are fused in a small area with the simultaneous formation of a biconical constriction at the point of fusion of the fibers. Then the biconical constriction is placed, for example, in a quartz capillary. The divider has two inputs (first and second, located on the left) and two outputs (first and second, located on the right) in the form of segments of single-mode fiber optical fibers. Let us designate the optical power division coefficient on the forward channel as the ratio of optical power at the first output when it is input to the first input or as the ratio of optical power at the second output when it is input to the second input. The division ratio of the optical power in the cross channels can be defined either as the ratio of the radiation power at the second output when it is input to the first input or as the ratio of the optical power at the first output when it is input to the second input. We denote the division coefficient of the 2 × 2 divider on the forward channel as k, where k can take digital values greater than zero but less than 1, and the optical power division coefficient in the cross channels is defined as (1-k). The single-mode fiber of the second input is connected, for example, by arc welding 12 with the fiber of the second output of the divider. The radiation from the source is introduced into the fiber segment, which is the first input of the divider, then the part of the radiation, determined by the power division coefficient on the direct channel of the divider, goes to the first output of the divider 13, and the other part of the radiation, determined by the coefficient of division through the cross channels of the divider, from the second output of the divider arrives at the second input of the divider, after which this radiation partially enters the first output of the divider (this part of the power in this case is determined by the cross-section of the power -th channels), and the other part of the power goes back to the second output of the divider and so on. Thus, with multiple passage of the divider channels, several radiation sources are formed at its first output. If the length of the fiber connecting the second output of the divider with its second input exceeds the coherence length of the source radiation, several independent sources (incoherent to each other) are present at the output of the single-mode fiber, which is the first output of the divider. The number of sources is determined by the amount of radiation circulation through the channel, the second input - the second output through the connecting segment of a single-mode fiber. This configuration determines almost the entire output power of the source at the output of the single-mode fiber segment, which is the first output of the divider and, thus, when connected to the input section of the single-mode fiber of the FRI, the entire power of the source will be used almost without loss. The total output power, which is the sum of the powers of independent sources, is determined in this case by the radiation loss in the divider, at the junction of the radiation source with the length of the fiber of the first input of the divider and at the place of welding of the segments of the optical fibers connecting the second input and output of the divider. From the above consideration it also follows that the number of independent sources and the power of their radiation is determined by the coefficient of division of power along the direct channel of the divider, and their total power is determined by the loss of radiation during the passage of radiation in the elements of the given design of the divider, that is, in the places indicated above. Thus, suppose that the powers of independent sources add up with different weighting factors A _n > 0 (that is, none of the powers is subtracted from the resulting power). In this case, the constant and noise components of the total power are equal

, $$\delta P\left(t\right)={\displaystyle \sum _{n=1}^N{A}_n\delta p_n\left(t\right)}$$

. $$P_0\left(t\right)=p_0{\displaystyle \sum _{n=\mathrm{one}}^N{A}_n}$$

, $$\delta P\left(t\right)={\displaystyle \sum _{n=\mathrm{one}}^N{A}_n\delta p_n\left(t\right)}$$

.

In this way, the constituents simply add up. The dispersion of the total noise component is calculated as

. $$\sqrt{〈\delta P^2\left(t\right)〉}=\sigma \sqrt{{\displaystyle \sum _{n=\mathrm{one}}^{\mathrm{<figure-callout\; id="2"\; label="N\; A\; n"\; filenames="00000023.png,00000024.png"\; state="\{\{state\}\}">N\; </figure-callout>}}{A}_n^{}{}_2^{}}}$$

.

From this it is clear that the noise of the intensity against the background of the total power changes in comparison with the noise of a single source by n ₀ times, where

. $$n_0={\displaystyle \sum _{n=\mathrm{one}}^N{A}_n}/\sqrt{{\displaystyle \sum _{n=\mathrm{one}}^{\mathrm{<figure-callout\; id="2"\; label="N\; A\; n"\; filenames="00000023.png,00000024.png"\; state="\{\{state\}\}">N\; </figure-callout>}}{A}_n^{}{}_2^{}}}$$

.

For any positive A _n (see above), it follows from the Cauchy-Bunyakovsky theorem that n ₀ > 1, i.e. there is a gain in intensity noise. In particular, above, where we considered the addition of capacities with the same amplitudes, we have A _n = 1, i.e. n ₀ = N ^{l / 2} , where n ₀ is the coefficient (times) of noise reduction of the radiation source intensity. To increase the intensity noise reduction coefficient, it is necessary to have as many circulating waves as possible, and they should have close power values.

We first consider the case taking into account losses in the divider. For the deterministic power component after the splitter (signal):

. $$P\left(t\right)=t\left[\kappa +\frac{{\left(\mathrm{one}-\kappa \right)}^2{t}_{0}t}{\mathrm{one}-\kappa t_{0}t}\right]P_0\left(t\right)$$

.

Dispersion of the noise component of power after the splitter (noise):

, $$\delta P=\delta P_0\mathrm{<figure-callout\; id="2"\; label="t\; \kappa "\; filenames="00000023.png,00000024.png"\; state="\{\{state\}\}">t\; </figure-callout>}\sqrt{\frac{{\kappa }^{}}{}}\sqrt{\frac{{}^2+t_0^2{t}^2\left(\mathrm{one}-2\kappa \right)\left(\mathrm{one}-2\kappa +2{\kappa }^2\right)}{\mathrm{one}-t_0^2{\mathrm{<figure-callout\; id="2"\; label="t"\; filenames="00000023.png,00000024.png"\; state="\{\{state\}\}">t</figure-callout>}}^2{\mathrm{<figure-callout\; id="2"\; label="\kappa "\; filenames="00000023.png,00000024.png"\; state="\{\{state\}\}">\kappa </figure-callout>}}^2}}$$

,

Where

t = exp (-α _c ), t ₀ = exp [- (αL + α ₀ )].

Here κ is the power transmission coefficient along the forward channel of the splitter (1-κ is the cross), α _с is the loss in the splitter when it passes once, L is the length of the splitter, α is the loss in the fiber of the closed channel of the splitter, α ₀ is the welding loss when this fiber is closed, L is its length.

The signal-to-noise ratio after the divider is σ = P / δP, whereas before the divider it is equal to σ ₀ = P ₀ / δP ₀ (that is, the signal-to-noise ratio of the source itself). The reduction in intensity noise is thus equal to the coefficient n ₀ = σ ₀ / σ. Figure 3 shows graphs of this gain depending on the division ratio of the divider on the forward channel for two cases: 1. curve 14 in the absence of any kind of loss; 2. curve 15 for the case when the propagation loss through the divider is 0.05 dB, the welding loss is 0.05 dB, the losses in the closed fiber connecting the second input and output of the divider are 0.5 dB / km, the length of this fiber is L = 1 m. Dotted curve 16 characterizes the gain in noise figure 3 ^1/2 times.

The graph shows that these two cases differ slightly. In addition, both cases have a maximum at κ≈0.4, equal to approximately 1.78 (the case with losses is slightly less). In both cases, there is also a fairly wide range κ≈ (0.33-0.5), in which the gain n ₀ is no less than 3 ^1/2 times (in the case of losses, this is a slightly narrower range of coefficients κ).

Thus, for maximum gain in noise reduction coefficient, κ = 0.4 should be chosen. In this case, the losses turn out to be quite small in view of the fact that no more than 5–7 waves circulating through the second channel of the fiber splitter with rather rapidly decreasing amplitudes take real part in the operation of the device described here.

Figure 4 shows the source of optical radiation with a reduced level of noise intensity based on N fiber dividers of optical power type 2 × 2. The first output of the first divider by welding the optical fibers 17 is connected to the first input 18 of the second optical power divider 19. Moreover, the second input and output of the second divider are also connected by welding the optical fibers 20. The first output 21 of the second divider is also connected to the first input of the next fiber divider 22 of a series of N dividers. The second input and output of each of the N dividers are also connected to each other using a length of the fiber. The source output in this case will be the first output of the 23rd fiber divider. In order for all secondary sources generated due to the circulation of optical power along the optical fibers connecting the second input and output of each divider, it is necessary that the length of the optical fibers connecting the second input and output of each subsequent divider be either less than or greater than the coherence length of the radiation source. In the case where the length of the optical fiber piece connecting the second input and the output of each successive divider would be smaller by the length of coherence of the radiation than the length of the previous segment of the fiber, then in this case it is necessary that the length of optical fiber piece connecting the second input and the output of the first divider is greater than NL _to where L _to - the coherence length of the radiation source. If the power division coefficient of each divider of a series of N dividers is equal to 0.4, then the intensity noise reduction coefficient will be n ₀ = (3 ^1/2 ) ^N , thus, when using four fiber dividers, it is possible to reduce the intensity noise by at least 9 times, while the total loss of optical power of all incoherent rays when propagating through the channels of the dividers is not more than 0.6 dB.

Figure 5 shows a source of optical radiation with a reduced level of intensity noise based on N fiber dividers of optical power type 2 × 2, but with a different connection of the fiber ends of the dividers. The second output of the first divider 24 in the form of a segment of a single-mode fiber 25 is connected by an arc welding method 26 to a segment of a single-mode fiber 27 of the first input of the second divider 28. The first output and the second input of the first divider are connected by welding the optical fibers 29. The second input and the first output of the second divider are connected arc welding 30 of two corresponding segments of optical fibers. The first output in the form of a segment of a single-mode fiber 31 of the second divider is connected to the first input of the next of N dividers, and the second output of the Nth divider 32 serves as the output of the radiation source, while the second input and the first output of the Nth divider are interconnected by arc welding 33 matching optical fibers. The optimal division coefficient along the direct channel of fiber dividers using such a scheme for connecting dividers will be k = 0.6.

Claims (2)

1. A radiation source with a low level of noise of intensity for a fiber-optic gyroscope, containing a source, the optical radiation from the output of which is divided into several propagation channels of different lengths, and the difference in the lengths of these channels exceeds the coherence length of the radiation of the source, characterized in that the output of the source is connected to next to N pieces of optical power dividers, where N = 1, 2, 3 ..., with the division ratio of each of these dividers on the direct channel k, where k can take digital values greater than 0 and less than 1, moreover, the output of the source is connected to the first input of the first divider from a number of dividers, and each first output of the previous divider of the series is connected to the first input of each subsequent divider, while the second input of each divider is connected via a fiber to its second output, and the length of each of these fibers is longer the coherence length L _{to the} radiation of the source, while the length of the fiber connecting the second input and output of each previous divider from a number of N dividers is greater or less than the length of the fiber of each subsequent divider from of this series, at least, by the coherence length L _k of the source radiation, and the length of the fiber connecting the second input and output of the first divider of the series of N dividers is greater than NL _k . 2. The radiation source according to claim 1, characterized in that the output of the source is connected to the first input of the first divider from a number of dividers, and each second output of the previous divider of the series is connected to the first input of each subsequent divider, while the second input of each divider is connected using a fiber with your first exit.

Images