CN101127568A - Detection device for optical channel of dense wave division multiplexing transmission system - Google Patents

Abstract

The utility model relates to a detection device in a light transmission system, in particular to an optical channel detection device for a DWDM light transmission system; According to the technical proposal of the utility model, the optical channel detection device for the DWDM light transmission system comprises a processor, a 16 bits digital-to-analog converter, a driving amplification and conditioning circuit, a tunable optical filter and a feedback voltage signal filter circuit; wherein, a feedback voltage sampling analog-digital converter, a light signal preceding stage sampling analog-to-digital converter and a light signal hind stage sampling analog-digital converter are controlled by the drive pulse of the tunable optical filter, which is periodically output by the processor, so as to realize the sampling of the feedback voltage and the optical power signal; the power of each optical channel, the signal-to-noise ratio of each optical channel and the communication request response to user are acquired by calculating the sampling signal. The power of the optical channel, the central wavelength and the signal-to-noise ratio are real-time monitored by recovering signals with special algorithm, so as to achieve the monitoring function.

Description

The checkout gear that is used for optical channel of dense wave division multiplexing transmission system

Technical field

The present invention relates to the checkout gear in a kind of optical transmission system, especially, relate to a kind of optical channel checkout gear of the DWDM of being used for optical transmission system.

Background technology

21 century is multimedia epoch, and the amount of information demand of rapid growth impels dense wave division multipurpose based on 10Gbit/S by overall application.Trunk transit network develops towards the direction of single 800G/1600G information capacity.It is commercial that the 40Gbit/S system also moves towards in a large number.Exactly because the capacity of the magnanimity of optical fiber, in case optical fiber ruptures, loss also is huge.The breaking-up of China's optical cable is very serious, calendar year 2001 Sino-U.S.'s submarine fiber cable fracture several times, caused tremendous loss.Can service quality be very important concerning communication common carrier, guarantee service quality to be related to company and to win the user.The cable network that each big communication common carrier such as China Telecom, China Mobile, UNICOM, Tie Tong is laid now all is furnished with simultaneously long-range optic cable automatic monitoring system (RFTS) basically.But existing RFTS system also can't monitor the performance of dwdm system.If can't monitor the runnability of dwdm system effectively, communication common carrier also just can't guarantee service quality, and this also is one of present many communication common carriers reason that dwdm system is waited and seen.Compare with traditional optical communication system, dwdm system need be monitored more parameter, and these parameters and wavelength change relation.Relevant major parameter has: 1. passage luminous power: the luminous power of each passage must accurately be measured, can know so just whether each passage normally works in the system, and whether the amplification bandwidth of the erbium-doped fiber amplifier that adopts in the system (EDFA) has covered each passage.2. the centre wavelength of passage: in DWDM, channel wavelength changes will have influence on adjacent passage.Therefore, the centre wavelength of each passage must accurately be measured, and judges whether the wavelength drift of light source is qualified.3. signal to noise ratio (OSNR): be one of most important parameter in the dwdm system, the overall performance of its reflection passage.

Measure when these performances can be built in system by OSA, and system configuration is adjusted according to the result.But owing to have various accidents in the real time execution, and finally influence the system transmissions performance.So this low cost, multi-functional channel monitoring solution that can the real time monitoring network light signal become particularly important.

Summary of the invention

The objective of the invention is to design a kind of checkout gear that is used for optical channel of dense wave division multiplexing transmission system, adopt particular algorithm to carry out channel power, centre wavelength, the signal to noise ratio of the also original monitoring in real time of signal light signal, thereby reach monitoring function.This product is integrated in the fiber amplifier, and system board space is not had extra demand, is convenient to existing system and carries out upgrading.

According to technical scheme provided by the invention, the checkout gear that is used for dense wave division multipurpose (DWDM) transmission system optical channel comprises,

Processor: by the driving pulse of cycle output tunable optical filter, level sampling analog to digital converter is to realize the sampling to feedback voltage, optical power signals behind Control and Feedback voltage sample analog to digital converter, light signal prime sampling analog to digital converter, the light signal, and the signal that sampling obtains is calculated the power of each optical channel and signal to noise ratio (OSNR) and response user's communication request;

The 16bits digital to analog converter: being connected in the output of processor, is the SPI interface, is used for the driving digital quantity of receiving processor to tunable optical filter, and output drives analog voltage amount;

Drive to amplify modulate circuit: be positioned at the output of 16bits digital to analog converter, be used for the voltage analog of 16bits digital to analog converter output is amplified, and be the magnitude of current with the voltage linear transformation;

Tunable optical filter: received current driving pulse, owing to be subjected to the heating of current drive pulses, the voltage drop that tunable optical filter output current driving pulse produces on tunable optical filter, be called feedback voltage signal, and along with the accumulation of drive current to the tunable optical filter heat energy, tunable optical filter output is each wavelength signal strength signal intensity current signal from the short wavelength to the long wavelength, is called the luminous power current signal;

Feedback voltage signal filter circuit: be positioned at the output of tunable optical filter, feedback voltage signal carried out Filtering Processing by filter circuit;

Signal processing circuit: be positioned at the output of feedback voltage signal filter circuit, be used for feedback voltage is carried out processing and amplifying;

Luminous power current signal filter circuit: be positioned at the output of tunable optical filter, be used for the luminous power current signal is carried out Filtering Processing;

Prime amplifying signal treatment circuit: be positioned at the output of luminous power current signal filter circuit, tunable optical filter luminous power current signal conversion position voltage signal;

Back level amplifying signal treatment circuit: be positioned at the output of prime amplifying signal treatment circuit, be used to amplify the luminous power voltage signal of prime amplifying signal treatment circuit output;

Feedback voltage sampling analog to digital converter: be positioned at the output of signal processing circuit, under the control of processor, feedback voltage signal sampled; The output of feedback voltage sampling analog to digital converter is connected with the input of processor;

Light signal prime sampling analog to digital converter is positioned at the output of prime amplifying signal treatment circuit, under the control of processor, the luminous power voltage signal of prime amplifying signal treatment circuit output is sampled; The output of light signal prime sampling analog to digital converter is connected with the input of processor;

Level sampling analog to digital converter is positioned at back level amplifying signal treatment circuit behind the light signal; Under the control of processor, the luminous power voltage signal of back level amplifying signal treatment circuit output is sampled; The output of level sampling analog to digital converter is connected with the input of processor behind the light signal.

Native system adopts U.S. Aegis company tunable optical filter to finish the optical passage signal and the ASE detection computations of 100G dwdm system, by the meticulous design of testing circuit and algorithm, has realized the detection performance of high-precision luminous power and OSNR.Its basic principle is for to drive tunable optical filter by electric current, the light of different wave length is believed screened by changing refractive index, photodetector converts light signal to the signal of telecommunication then, utilize DSP that signal is handled, remove noise, the inter-channel crosstalk influence restores true light signal and calculates channel power, centre wavelength, signal to noise ratio.

Description of drawings

Fig. 1 is based on the hardware block diagram of the optical channel checkout gear of TOD.

Fig. 2 is in an input pure light list ripple signal (no ASE) TOD luminous power output sampling, through wavelength energy is resolved, and obtains the performance number of each wavelength value correspondence.

Fig. 3 is wavelength and the graph of a relation that drives energy.

Fig. 4 is energy and error of fitting graph of a relation.

Fig. 5 is through energy and error of fitting graph of a relation behind the 2 rank least square fittings.

Fig. 6 is luminous power and the relation that drives energy.

Fig. 7 is the deviation of signal power and OSNR calculated value and actual value.

Embodiment

Based on Aegis mosquito ^RTod 050 is mainly used in the detection of optical channel in the dense wave division multipurpose optical transmission system of 100G, sees Fig. 1 based on the hardware block diagram of the optical channel checkout gear of TOD.

Basic function module comprises processor, 16bits digital to analog converter, drive and amplify level sampling analog to digital converter behind modulate circuit, feedback voltage signal filter circuit, signal processing circuit, luminous power current signal filter circuit, prime amplifying signal treatment circuit, back level amplifying signal treatment circuit, feedback voltage sampling analog to digital converter, light signal prime sampling analog to digital converter, the light signal.

1. processor:

Driving pulse by cycle output tunable optical filter, level sampling analog to digital converter is to realize the sampling to feedback voltage, optical power signals behind Control and Feedback voltage sample analog to digital converter, light signal prime sampling analog to digital converter, the light signal, and the signal that sampling obtains is calculated the power of each optical channel and OSNR (signal to noise ratio) and response user's communication request.

2. 16bits digital to analog converter:

The SPI interface, receiving processor drives digital quantity to tunable optical filter, and output drives analog voltage amount.

3. drive and amplify modulate circuit:

Voltage analog to 16bits digital to analog converter output amplifies, and is the magnitude of current with the voltage linear transformation.

4. feedback voltage signal filter circuit:

By filter circuit feedback voltage signal is carried out Filtering Processing.

5. signal processing circuit:

Feedback voltage is carried out processing and amplifying.

6. luminous power current signal filter circuit:

The luminous power current signal is carried out Filtering Processing.

7. prime amplifying signal treatment circuit:

Tunable optical filter luminous power current signal conversion position voltage signal.

8. level amplifying signal treatment circuit after:

Amplify the luminous power voltage signal of prime amplifying signal treatment circuit output.

9. feedback voltage sampling analog to digital converter, level sampling analog to digital converter behind light signal prime sampling analog to digital converter, the light signal, under the control of processor, the luminous power voltage signal of feedback voltage signal, prime amplifying signal treatment circuit and back level amplifying signal treatment circuit output is sampled.

Processor is converted to current drive pulses by 16bits digital to analog converter output voltage driving pulse after driving the amplification modulate circuit, and loads this current drive pulses to tunable optical filter (aegismosquito Tod 050).Owing to be subjected to the heating of current drive pulses, the voltage drop that tunable optical filter output current driving pulse produces on tunable optical filter, be called feedback voltage signal, and along with the accumulation of drive current to the tunable optical filter heat energy, tunable optical filter output is each wavelength signal strength signal intensity current signal from the short wavelength to the long wavelength, is called the luminous power current signal.

Feedback voltage signal amplify by signal processing circuit and filtering after send into feedback voltage analog-to-digital conversion device, sample and read sampled value by processor control.

The luminous power current signal is converted to voltage signal by prime amplifying signal treatment circuit with current signal, is sent in the light signal prime sampling analog to digital converter, is controlled sampling and is read sampled value by processor.Consider signal detection accuracy when input signal is very faint, this signal by back level amplifying signal treatment circuit processing and amplifying again, is sent to behind the light signal in the level sampling analog to digital converter, by processor control sampling with read sampled value.

Groundwork principle and step are as follows:

(1), sweep current drives:

According to the driving standard of TOD, drive current:

$i\left(t\right)=\mathrm{DCA}*\left[\frac{2-e^{\left(\frac{-t}{0.12}\right)}-e^{\left(\frac{-t}{1.2}\right)}-0.02t}{1.8864}\right]---1-1$

Unit is mA, and wherein DCA is the drive current amplitude, should calibrate according to working temperature, and t is the time, and unit is ms.Drive current is the pulse of 5.5ms fixed pulse width, and the pulse period is 500ms, and on t ∈ [0,5.5] ms time interval, processor passes through the pulse of 16bits digital to analog converter output driving current with the equidistant time step of 4us by formula 1-1 at each drive cycle.

(2), the sampling of feedback voltage signal and optical power signals:

In each current drives step-length, feedback voltage and optical power signals are sampled by the feedback voltage level sampling analog to digital converter of sampling behind analog to digital converter, light signal prime sampling analog to digital converter, the light signal with sampling period of 1us, in 5.5ms, level after obtaining 5500 feedback voltage values and 5500 luminous power primes respectively and amplifying output signal and 5500 luminous powers is amplified output signal thus.

(3): the calibration of wavelength and the identification of optical channel:

TOD at the uniform velocity exports the optical power value from the shortwave to the long wave under the drive current heating, set up the relation of heat energy and wavelength thus.Because the feedback voltage value is the voltage drop that electric current produces by heating resistor, heat energy is calculated as thus: 4 $E\left(t\right)={\∫}_0^t(u*i)\mathrm{dt},t\∈\left[\mathrm{0,5.5}\right]\mathrm{ms},$ At u, the i dimension is respectively V, and under the condition of mA, the dimension of heat energy is mJ, is converted to digital integration and is: $E\left(t\right)=\underset{k=0}{\overset{5500}{\mathrm{\Σ}}}u\left(k\right)*i\left(k\right)*\mathrm{\Δt}.$

Can calculate the wavelength value of 5500 luminous power sampled point correspondences according to the relation of wavelength of calibrating and energy.According to ITU standard wave length, can inquire the pairing performance number of each ITU wavelength.

(4): the calculating of luminous power:

Processor calculates the performance number of ITU wavelength points and ASE (spontaneous emission noise) value of this wavelength correspondence, can calculate the optical channel performance number of system requirements and the OSNR of this passage.Consider the small-signal resolving accuracy, when level output does not reach back level sampling analog to digital converter maximum range after luminous power, the data that the rated output data use the sampling of back level sampling analog to digital converter to obtain.

Though system's optical channel spacing is 0.8nm, in order to improve the signal noise performance, design has filter between optical channel in system, like this except the ASE in each passage ± 0.2nm interval not by filtering, the ASE of remainder is all decayed.The calculation level of ASE can only be asked in the interval at ITU wavelength ± 0.2nm like this.Fig. 2 is in an input pure light list ripple signal (no ASE) TOD luminous power output sampling, through wavelength energy is resolved, and obtains the performance number of each wavelength value correspondence.Middle red-label place is an ITU wavelength place performance number, and the yellow mark in both sides is the performance number at ITU wavelength ± 0.2nm place.As seen from the figure, in both sides ± the 0.2nm place is even without ase signal, because the broadening of TOD filter, crest is also big unusually to crosstalking of this two place, no matter gets that and resolves ASE, if accurately do not deduct amount of crosstalk, the OSNR calculation accuracy will not have collateral security.

Though it is a lot of to use TOD to finish the product man that optical channel detects, calculation method is also varied to resolve almost not having of OSNR and calculation accuracy ± 3.5dB in 2 seconds, and reason has following 3:

1, the ASE selected point is too near from peak value, and ASE itself is smaller, extracts a less value from a very big amount, if a suitable extracting method is not taked in the extraction of ASE power, no matter how complicated algorithm is, and precision does not satisfy index request all the time.

2, wavelength to energy concern that fitting precision is not high enough, cause wavelength location inaccurate.

3, power signal has passed through the amplification conversion process of electronic circuit, inevitably introduces electromagnetic interference, if interference volume is not done suitable consideration, precision also can't improve.

4, this is one has the system of requirement to real-time, if algorithm is very complicated, real-time does not satisfy requirement yet, so algorithm must be effectively succinct.

The present invention is for addressing the above problem, and proposes the completely new approach of a kind of brand-new energy and wavelength relationship match, optical channel power, optical channel OSNR on the basis of hardware designs.Comprise that energy extracts compound least square fitting, the equivalent point method extraction of optical channel power, the ASE bilateral integration of wavelength, the ASE self-interference is crosstalked with optical channel power and is separated, and these three aspects run through the calibration of whole OPM module and the computational process of optical channel power and OSNR.At calculated signals bright dipping channel power and the OSNR that each scan period processor obtains according to sampling, calculating desired parameters calibration in advance also is burnt in the processor parameter list.

One: energy is to the compound least square fitting of wavelength:

No matter calculating the OSNR that optical channel power still is optical channel, accurately find the pairing power of each ITU wavelength, is to calculate correct foundation stone.If should concern not accurate enough, two ASE selected points produce relative to the ITU wavelength location and move about will causing among Fig. 2, the OSNR computational accuracy is brought influence, to produce the serious consequence of wrong passage when serious, in order to guarantee the precision of this match relation, adopted compound least square method that energy is carried out match to the relation of wavelength.

Because TOD do not provide wavelength information, but can calculate and wavelength indirect correlation amount, drive cumlative energy by drive current and the feedback voltage that obtains by the sampling of feedback voltage sampling analog to digital converter.In calibration, in working frequency range, import single ripple of 22 ITU standard wave lengths respectively, note the wavelength X of each single ripple _k(1≤k≤22) find out the pairing driving energy of each single ripple e _k(1≤k≤22) then obtain Fig. 3, and along with the accumulation that drives energy, TOD will be from the shortwave uniform speed scanning to long wave.

, to choose 4 rank least square methods earlier and carry out match because the error of fitting that measure error is brought has been done corresponding balance and considered between degree of fitting and match flatness in order reducing, promptly ask for:

E＝a ₀+a ₁λ+a ₂λ ²+a ₃λ ³+a ₄λ ⁴ 1-2

Fitting coefficient [a among the formula 1-1 ₀, a ₁, a ₂, a ₃, a ₄], according to the normal equation of least square method:

$\left\{\begin{array}{c}n{a}_0+a_1\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k+a_2\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^2+a_3\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^3+a_4\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^4=\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{e}_k\\ a_0\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k+a_1\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^2+a_2\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^3+a_3\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^4+a_4\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^5=\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{e}_k*{\mathrm{\λ}}_k\\ a_0\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^2+a_1\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^3+a_2\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^4+a_3\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^5+a_4\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^6=\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{e}_k*{\mathrm{\λ}}_k^2\\ a_0\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^3+a_1\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^4+a_2\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^5+a_3\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^6+a_4\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^7=\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{e}_k*{\mathrm{\λ}}_k^3\\ a_0\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^4+a_1\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^5+a_2\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^6+a_3\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^7+a_4\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{\mathrm{\λ}}_k^8=\underset{k=1}{\overset{22}{\mathrm{\Σ}}}{e}_k*{\mathrm{\λ}}_k^9\end{array}\right\}---1-3$

Claim group with measured value substitution 1-2 side, the group of solving an equation then can calculate and intend coefficient [a ₀, a ₁, a ₂, a ₃, a ₄], bring fitting coefficient into formula 1-1, it is long to bring the single wave-wave of 22 ITU that import above into, calculates pairing 22 and drives energy, and calculated value and measured value are compared, and obtains fit error curve, sees Fig. 4.

From error curve, the error of fitting maximum reaches-0.0018mJ, corresponds to the wavelength position error that will cause 0.12nm on the wavelength, and error is still very big, in calculating, if use this relational expression, brings the OSNR error of calculation will reach 4.6dB.

In order to solve the influence that error of fitting is brought, error of fitting has been made one time 2 rank least square fitting

Error＝b ₀+b ₁λ+b ₂λ ² 1-4

Use the 1-3 formula that the 1-1 formula is revised, then the 1-1 formula is modified to:

E＝(a ₀+b ₀)+(a ₁+b ₁)λ+(a ₂+b ₂)λ ²+a ₃λ ³+a ₄λ ⁴ 1-5

Then this moment, error of fitting was seen Fig. 5, and visible fitting precision has obtained bigger improvement.Wavelength energy is carried out twice fitting has brought bigger convenience to algorithm design:

1.: the high-order match solves the degree of fitting problem if adopt more, then match flatness variation, and with high frequency measurement noise introducing match relation.

2. because high-order will spend longer computing time in real-time is calculated.

Two: the calibration of calculating parameter:

The optical channel power of 44 passages and OSNR (power noise ratio) calculating formula is as follows:

p _s＝A ^-1P 1-6

N＝A ^-1N ^*-C*p _s 1-7

PEAKPOWER＝p _sB 1-8

ASE＝ND 1-9

After obtaining the linear power P eakPower of signal and noise power ase of each passage, according to following formula:

P _dB＝10*log(PeakPower-ase) 1-10

ase _dB＝10*log(ase) 1-11

OSNR＝P _dB-ase _dB 1-12

Can obtain the logarithmic signal power P of each passage _DB, logarithm OSNR.

Among the formula 1-6, p _sIs 44 * 1 matrixes for deducting other channels to its net signal intensity of crosstalking, and P is for sampling obtains signal strength signal intensity, and comprising all the other channels is 44 * 1 matrixes to its amount of crosstalk, and A is that convolution matrix is 44 * 44 matrixes, A ^-1Be the A inverse matrix.

Among the formula 1-7, N is the clean noise intensity of channel, is 44 * 1 matrixes, N ^*Comprising for actual samples is resulting that other channel signals are crosstalked to it and other channels ASE crosstalks to it, is 44 * 1 matrixes, and it is crosstalked is 44 * 44 matrixes to C for other channel signals.

B is a calibration matrix among the formula 1-8, be expressed as power level and be converted to the actual signal power value coefficient, and be 44 * 1 matrixes.

D is a calibration matrix among the formula 1-9, and being expressed as the ase signal intensity-conversion is actual ase signal performance number coefficient.

The calibration of calculating parameter comprises the calibration of A matrix, C matrix, B matrix, D matrix.

1, compute matrix calibration

In should using, in the bandwidth of system works, all-wave is by being 44 passages, and the spacing of each passage is 100G.For any point power, other wavelength places always disturb it and exist, for the total crosstalk components value in this place $P_r=\stackrel{L}{\∫}{p}_r\mathrm{dl},$ P _rBe the crosstalk summation of other wavelength places to this point, p _rFor wavelength points is to the crosstalk components of this point everywhere, integrating range is the whole operation wavelength of traversal, then should the actual power value P of place _s=P-P _rThe signal strength values that P measures for this wavelength points.

In real work, owing to will calibrate the interference components p of everywhere _rIn fact can not operate, the real-time of calculating requires also not allow, in fact, the crosstalk strength maximum point is present in the signal cross-talk of each ITU wavelength points, owing to there is signal, ASE itself is also little relatively, and ASE can ignore the interference volume at the ITU wavelength place that will calculate.Can be expressed as P for 44 any one i passages thus _i=p _i+ k _I1p ₁+ k _I2p ₂+ k _I3p ₃++ k _I-1p _I-1+ k _ip _i++ k ₄₄p ₄₄, P wherein _iBe the actual signal strength signal intensity that records of this point, p _i(i=0...44) be the signal strength signal intensity that 44 ITU import single ripple, comprehensive 44 ITU wavelength places are calculated as P=A*p, P is 44 * 1 matrixes, represent 44 signal strength values that ITU wavelength place records, p is 44 * 1 matrixes, represent that 44 ITU do not have the signal strength values of other channels crosstalks, A is 44 * 44 matrixes, and i passage of i (i=0...44) line display is to the crosstalk coefficient of 44 passages.

Must be a diagonally dominant matrix still no matter, must have inverse matrix, the counter then p=A that separates from the mathematics matrix A from physical significance ^-1* P, according to this formula, promptly can calculate 44 ITU wavelength places does not have the crosstalk signal intensity level.

1. the equivalent point of signal strength signal intensity extracts:

Because signal is all handled the back through electronic circuit and is produced, the unavoidable electromagnetic interference that is subjected to, to signal power still is that reading of ASE power all needs to consider this interference signal, otherwise the computational accuracy of OSNR must be affected, and all adopts integration to extract performance number on calculating.

Sampled signal is converted to the relation of signal power to wavelength after overdrive energy and wavelength convert, the effective range of data is ITU ± 0.2nm, sees Fig. 2.In order fully to use this segment data, it is done equal divisional processing, ITU ± 0.1nm is the signal power data district, and ITU-0.2nm-ITU-0.1nm is ASE data field, left side, and ITU+0.1nm-ITU+0.2nm is ASE data field, right side.Integration is carried out to wavelength in these three zones, obtain respectively, the power spectrum signal integral area is designated as S ₁, ASE left side power spectrum integral area, ASE right side power spectrum integral area.Their equal equivalences are signal strength signal intensity, with ASE left side power spectrum integral area and right side addition, for Equivalent A SE intensity is designated as S ₂Equivalent point extracts signal strength signal intensity and the extraction of ASE bilateral integration has following 3 meanings to the computational accuracy that solves electromagnetic interference and OSNR:

A. adopt the equivalence of power spectrum integration when calculating is crosstalked, more to approach and consider crosstalking of all-wave scope for the method for signal strength signal intensity.

B. the equivalent power spectral integral of power is converted to average power spectra to the actual power spectrum, is equivalent to multi-point average in filter action, is equivalent to ask 0.2nm bandwidth energy spectrum on the result.Thus, reasonable filtering electromagnetic noise interference owing to adopt energy spectrum to weigh power, can overcome laser centre wavelength again and slightly be offset the error of calculation of bringing.

C.ASE bilateral integration extracts and has overcome because the ASE error of calculation that the centre wavelength error of calculation is brought is seen Fig. 2.Left and right sides ASE integrated value is designated as S ₁, S ₂Overgauge appears if centre wavelength is calculated, because the distance of the relative centre wavelength in both sides can think error free in this interval, and then obvious S ₁Calculated value is bigger than normal, but this moment S ₂Will be less than normal, though because the decision of the spectrum waveform of TOD filter can't guarantee S ₁+ S ₂Be certain value, but overcome the calculation deviation that the centre wavelength position error is brought to a great extent.

2. the calculating of compute matrix:

With 44 pure light of ITU passage (no ASE) input TOD, this passage S is calculated in input to any i (1≤i≤44) passage respectively ₁, be designated as S _i, the S of (1≤i≤44) and 44 each passages of passage ₂, be designated as S _1i, S _2i... S _Ki... S _43i, S _44i, S _Ki, the flashlight of (1≤k≤44) expression i passage is to the amount of crosstalk of k passage ASE power spectrum area, with S _1i, S _2i... S _Ki... S _43i, S _44i, divided by S _i, i.e. c _{K, i}=S _Ki/ S _i, represent the crosstalk coefficient of the flashlight of i channel to k passage ASE.Comprehensive 44 passages then can obtain signal among the formula 1-7 to the crosstalk matrix C of ASE.

Under the initial conditions, calculate the power spectrum signal integral area S of 44 passages in the above ₁, be designated as s _1i, s _2i... s _Ki... s _43i, s _44i, divided by S _i, i.e. a _{K, i}=s _Ki/ S _i, represent the crosstalk coefficient of i channel signal light to k channel signal, comprehensive 44 passages then can obtain the convolution matrix A among the formula 1-6.Measure the linear power p of each single ripple of input with power _i, calculate calibration of power coefficient with corresponding area

b _i=p _i/ S ₁, (1≤i≤44) can calculate signal power calibration matrix B.

Import pure ase signal, scan the ASE performance number P that obtains each wavelength points with the OSA instrument _Ase, the power spectrum signal integral area that calculates each ITU passage according to the equivalent point method is designated as S ₁, ASE power spectrum integral area S ₂, according to front formula 1-7,1-8 calculates each passage corresponding A SE area under spectrum integration n that the pairing nothing of each passage is crosstalked _i, then can calculate ASE calibration factor d _i=P _Ase/ n _i, comprehensive 44 passages promptly can be calculated the ASE calibration matrix.

2. the calculating of luminous power and OSNR:

The top parameter matrix of calibrating as calculating parameter, according to formula 1-7 noted earlier, 1-8,1-9,1-10,1-11,1-12 promptly can carry out luminous power and OSNR and calculate.Being input as single wave power to one below is-26dBm---32dBm, and OSNR is 16-18dB, and bandwidth of operation is that the input condition of 1529.55nm-1564.50 carries out calculation specifications, and sampling gained luminous power vs drives the relation of energy and sees Fig. 6:

1. signal and ASE intensity are extracted in the integration equivalence

According to the calibration relation of wavelength to energy, calculate the corresponding energy position of each ITU wavelength, according to wavelength and energy relationship, the relation that the interval luminous power vs of each channel ITU ± 0.3nm is driven energy is converted to luminous power vs wavelength, sees Fig. 1.Calculate the pairing power spectrum signal integral area of each ITU passage according to the equivalent point extracting method of signal strength signal intensity noted earlier and be designated as S _i, ASE power spectrum integral area n ^* _i, (1≤i≤44).Obtain 44 * 1 sampled signal power spectrum integral area matrix P respectively, ASE sampled signal power spectrum integral area matrix N ^*

2. luminous power calculates with OSNR

P calculates when sampled signal power spectrum integral area matrix, then according to formula 1-6, and 1-7,1-8 can calculate the signal strength signal intensity of each ITU passage.

Because ASE trimmed book body is very little, thus very responsive to crosstalking, consider based on this point, not only considered each channel signal light to the crosstalking of ASE, and crosstalking of having considered that each passage ASE brings, and with these two separation of crosstalking.So handle and have following 2 very strong realistic meanings.

Ignore the amount of crosstalk of ASE when 1, calculating luminous power, avoided making convolution matrix A to be enlarged into 88 * 88 flashlight.Though considered the ASE cross talk effects when ASE calculates, thereby increased by one 44 * 44 matrix, total data length has still reduced half.

2, because data length has reduced half, obviously the amount of calculation ratio has reduced one times with 88 * 88, and the real-time of calculating has been alleviated pressure.

Fig. 7 is the deviation of signal power and OSNR calculated value and actual value, and curve top among the figure is the signal power error of calculation, and following surface curve is the OSNR error of calculation.As seen from the figure, the signal power error of calculation is [0.84,0.5] dB, and the OSNR error of calculation is [0.22 ,-2.18] dB, if do not consider not have the error of optical channel, then precision is respectively [0.2,0.5] dB and [0.22 ,-1.97] dB.For a 100G TOD, to have only in the data effective range ± system of 25G, this computational accuracy has far satisfied customer requirements.

Claims (1)

1. be used for the checkout gear of optical channel of dense wave division multiplexing transmission system, it is characterized in that,

Processor: by the driving pulse of cycle output tunable optical filter, level sampling analog to digital converter is to realize the sampling to feedback voltage, optical power signals behind Control and Feedback voltage sample analog to digital converter, light signal prime sampling analog to digital converter, the light signal, and the signal that sampling obtains is calculated the power of each optical channel and signal to noise ratio and response user's communication request;

The 16bits digital to analog converter: being connected in the output of processor, is the SPI interface, is used for the driving digital quantity of receiving processor to tunable optical filter, and output drives analog voltage amount;

Drive to amplify modulate circuit: be positioned at the output of 16bits digital to analog converter, be used for the voltage analog of 16bits digital to analog converter output is amplified, and be the magnitude of current with the voltage linear transformation;

Tunable optical filter: received current driving pulse, owing to be subjected to the heating of current drive pulses, the voltage drop that tunable optical filter output current driving pulse produces on tunable optical filter, be called feedback voltage signal, and along with the accumulation of drive current to the tunable optical filter heat energy, tunable optical filter output is each wavelength signal strength signal intensity current signal from the short wavelength to the long wavelength, is called the luminous power current signal;

Feedback voltage signal filter circuit: be positioned at the output of tunable optical filter, feedback voltage signal carried out Filtering Processing by filter circuit;

Signal processing circuit: be positioned at the output of feedback voltage signal filter circuit, be used for feedback voltage is carried out processing and amplifying;

Luminous power current signal filter circuit: be positioned at the output of tunable optical filter, be used for the luminous power current signal is carried out Filtering Processing;

Prime amplifying signal treatment circuit: be positioned at the output of luminous power current signal filter circuit, tunable optical filter luminous power current signal conversion position voltage signal;

Back level amplifying signal treatment circuit: be positioned at the output of prime amplifying signal treatment circuit, be used to amplify the luminous power voltage signal of prime amplifying signal treatment circuit output;

Feedback voltage sampling analog to digital converter: be positioned at the output of signal processing circuit, under the control of processor, feedback voltage signal sampled; The output of feedback voltage sampling analog to digital converter is connected with the input of processor;

Light signal prime sampling analog to digital converter is positioned at the output of prime amplifying signal treatment circuit, under the control of processor, the luminous power voltage signal of prime amplifying signal treatment circuit output is sampled; The output of light signal prime sampling analog to digital converter is connected with the input of processor;

Level sampling analog to digital converter is positioned at back level amplifying signal treatment circuit behind the light signal; Under the control of processor, the luminous power voltage signal of back level amplifying signal treatment circuit output is sampled; The output of level sampling analog to digital converter is connected with the input of processor behind the light signal.

Images