CN115792763A - Silicon-based NxN optical switch chip automatic calibration system and method - Google Patents
Silicon-based NxN optical switch chip automatic calibration system and method Download PDFInfo
- Publication number
- CN115792763A CN115792763A CN202211512218.6A CN202211512218A CN115792763A CN 115792763 A CN115792763 A CN 115792763A CN 202211512218 A CN202211512218 A CN 202211512218A CN 115792763 A CN115792763 A CN 115792763A
- Authority
- CN
- China
- Prior art keywords
- optical switch
- optical
- signal
- analog
- input
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
A silicon-based NxN optical switch chip automatic calibration system and method monitors the state of an on-chip optical path through a photoelectric detector on an expansion-level redundant port in an optical switch array chip, reads the optical power of the photoelectric detector to a microprocessor in a time-sharing manner through an analog selection switch, and completes the quick and automatic calibration on the states of all units in the optical switch array by utilizing an automatic calibration algorithm deployed in the microprocessor. The invention can effectively reduce the number of the on-chip photoelectric detectors and the peripheral analog-to-digital converters, realizes the quick calibration of the state of the silicon-based optical switch chip, and has the advantages of simple method, high calibration speed, low cost and the like.
Description
Technical Field
The invention relates to the field of optical communication, in particular to a silicon-based NxN optical switch chip automatic calibration system and a silicon-based NxN optical switch chip automatic calibration method.
Background
With the popularization of high-definition video application and 5G technology, the global data traffic is increased explosively. The International Data Center (IDC) predicts that global Data traffic will increase from 33ZB in 2018 to 175ZB in 2025. Wherein, the data center internal traffic exchange accounts for more than 71.5% of all traffic. The traditional electric exchange network can not meet the requirements of low power consumption, high bandwidth and low cost of the current data center.
The all-optical switching network directly completes the signal switching in the optical domain, is not influenced by the traditional electronic technology, and has the advantages of high bandwidth, low power consumption, low cost and the like. The optical switch can route high-speed optical signals to different output ports with low power consumption and low delay, and is one of the most core devices of an all-optical switching network. The nxn high-speed optical switch chip is the most basic and most core device in the optical switch.
The traditional optical switching system is mainly built by adopting 2 multiplied by 2, 4 multiplied by 4 and other discrete optical switch arrays, and has the problems of large volume, high energy consumption, low reliability and the like. The silicon-based optical switch chip has the advantages of high integration level, low power consumption, compatible manufacturing process with the traditional microelectronic CMOS (complementary metal oxide semiconductor) process and the like, and is suitable for a large-scale optical switch array of an all-optical switching system.
The topological structure of the silicon-based optical switch chip mainly comprises Benes, crossbar, switch & select (S & S) and the like. The Benes-based optical switch array can realize an N × N optical switch array by using fewer optical switch units, such as a 32 × 32 optical switch array, and the Benes structure can realize a non-blocking switch array by only 144 switch units. However, because the Benes switch is a reconfigurable non-blocking switch structure, the change of the current routing state can affect the transmission of other routing signals, and therefore, the control algorithm of the switch array based on Benes is more complex than a strict non-blocking switch structure.
The structure of the switch unit mainly includes a Mach-Zehnder interferometer (MZI), a micro-ring resonator, and the like. The MZI structure has the advantages of broadband response, temperature insensitivity and the like, so that the MZI structure has an application prospect. However, due to process errors, the initial state of the switching unit and the design value do not coincide, and initial state calibration of the chip is required. In addition, the phases of two output optical signals of the 3dB optical splitters in the MZI unit are not exactly different by pi/2, and the splitting ratios are not uniform, so that the working voltages of Cross and Bar under different input ports are slightly different. Under the same switch state, the voltages of optical signals input from different ports to realize the optimal extinction ratio are different, and the crosstalk of the switch array is further increased.
At present, a silicon-based optical switch chip mainly calibrates the state of a switch unit on a path by monitoring optical power change of a target output port, and since an interference signal of a crosstalk port inevitably exists in a switch calibration process, the state calibration can be realized only by the aid of the crosstalk switch unit. The storage problem (Qiao L, tang W, chu T.32323232 silicon electro-optical switch with build-in monitor and balanced-status units [ J ]. Scientific Reports,2017,7 (1): 1-7) at Zhejiang university directly calibrates the state of the optical switch unit on the target path by inserting the bi-directional coupler and the grating coupler at the right side of the middle stage of the Benes structured optical switch, but the use of the grating coupler requires off-chip optical fiber vertical coupling and off-chip optical power detection, resulting in further increase of the chip packaging difficulty and calibration time. The Huawei corporation (P.Dumais et al, "Silicon photo Switch Subsystem With 900monolithic Integrated Calibration photodes and 64-Fiber Package," in Journal of Lightwave Technology, vol.36, no.2, pp.233-238,15Jan.15, 2018.) directly achieves Calibration of the state of the Switch unit on-chip by embedding two weak directional couplers and an Integrated Silicon-germanium photodetector on each 2X 2 optical Switch unit output waveguide, but further increases the packaging complexity and control difficulty of the optical Switch chip because each Switch has two photodetectors connected, resulting in up to thousands of output electrodes of the entire 32X 32 optical Switch chip.
Disclosure of Invention
The method aims to solve the problems of difficult state calibration, high complexity, long time and the like of the silicon-based N multiplied by N optical switch array chip. The invention provides a silicon-based NxN optical switch chip automatic calibration system and a method, which can quickly and automatically finish the calibration of all unit states in an optical switch array through a photoelectric detector on a redundant port of the optical switch array, a peripheral control circuit and an automatic calibration algorithm in the system.
The technical scheme of the invention is as follows:
in one aspect, the present invention provides an automatic calibration system for a silicon-based nxn optical switch chip, comprising:
the light source is used for generating two paths of input optical signals with equal energy and single wavelength;
two programmable polarization controllers which are symmetrically arranged and used for respectively receiving the input optical signals and adjusting the polarization state of the input optical signals through the respectively received control signals;
two 1 × N optical switches symmetrically arranged for receiving a feedback signal and controlling the input optical signal to be input from any one of 1 input port and output from any one of N output ports, or controlling the input optical signal to be input from any one of N input ports and output from 1 output port;
the silicon-based NxN optical switch chip comprises 4M photoelectric detectors, a first-stage 2 x 2 optical switch, a second-stage 2 x 2 optical switch, a … …, a Tth-stage 2 x 2 optical switch, … …, a 2T-2-stage 2 x 2 optical switch and a 2T-1-stage 2 x 2 optical switch; the Tth-stage 2 x 2 optical switch consists of M2 x 2 optical switches, each 2 x 2 optical switch comprises 4 redundant ports, and the left end and the right end of each 2 optical switches are respectively connected with the photoelectric detectors;
the two 1 × 2M analog selection switches are symmetrically arranged, and 2M ports of each 1 × 2M analog selection switch are respectively connected with 2M photoelectric detectors, and are used for outputting real-time photocurrents of the photoelectric detectors on different paths and transmitting the real-time photocurrents to corresponding transimpedance amplifiers;
the two transimpedance amplifiers are symmetrically arranged and used for amplifying the received light current and converting the received light current into a voltage signal;
the two analog-to-digital converters are symmetrically arranged and used for reading the voltage signal output by the transimpedance and converting the voltage signal into a digital signal;
the microprocessor is used for realizing a hardware platform of an automatic calibration algorithm, outputting a feedback signal and a feedback voltage value to the digital-to-analog converter by receiving the voltage value fed back by the analog-to-digital converter, automatically switching an output port of the 1 xN optical switch unit according to the automatic calibration algorithm, outputting a port number by the analog switch and feeding back a new voltage value to the digital-to-analog converter;
the digital-to-analog converter is used for reading a feedback signal of the microprocessor, converting the feedback signal into an analog signal and feeding the analog signal back to the driving amplifier;
and the driving amplifier converts the digital signals fed back by the digital-to-analog converter into corresponding voltage values and loads the voltage values on the corresponding 2 multiplied by 2 switching units.
Further, the network topology structure of the silicon-based NxN optical switch chip comprises a Benes, an extended Benes, a Crossbar and a double-layer network (DLN) structure;
further, the 2 × 2 optical switches are arranged in a certain network topology structure, the number of the optical switches is determined by the topology structure, and the optical switches are 2 × 2 mach-zehnder interferometer structures or 2 × 2 micro-ring resonator structures;
further, if the T-th stage 2 × 2 optical switch has no redundant port, the 2 × 2 optical switch is configured by cross-connecting 4 2 × 2 optical switches.
Furthermore, the photoelectric detector is a silicon germanium PIN type detector.
On the other hand, the invention also provides an automatic calibration method of the silicon-based NxN optical switch chip, which is characterized by comprising the following steps:
outputting an optical signal: the microprocessor starts the light source and respectively introduces two paths of input optical signals with equal energy and single wavelength into the respective programmable polarization controllers;
selecting a calibration path: selecting a path I to be calibrated m -O n (I m And O n The optical switch chip comprises an mth input port and an nth output port of a silicon-based N multiplied by N optical switch chip, wherein m, N =1,2, …, N), optical signals are input into a 1 multiplied by N optical switch array from left and right paths, namely, the left optical signal is input into the N multiplied by N optical switch array chip from the mth port, and the right optical signal is input into the N multiplied by N optical switch chip from the nth port; path to be calibrated I m -O n Through the ith (i =1,2, … M) optical switch of the middle stage, the microprocessor sends control signals to the 1 × 2M analog switches on the two sides, so that the two left and right 1 × 2M analog switches are respectively switched to the 2 × i-2+p and 2 × i-2+q paths (p, q =1,2), and are connected with the p-th left and q-th right photodetectors of the ith optical switch unit of the middle stage; the microprocessor reads the voltage signals input by the two A/D convertersA signal to obtain an optical signal input to the intermediate stage optical switch;
initializing a calibration path: according to the read voltage signal, the microprocessor sends a control signal to the programmable polarization controller, and adjusts the polarization state of the introduced input optical signal to maximize the optical signal of the intermediate-level optical switch; sending a control signal to the digital-to-analog converter to treat a calibration path I m -O n All the optical switch units are loaded with required initial voltage, so that the voltage signal read from the analog-to-digital converter is larger;
a calibration switch unit: calibrating paths I to be calibrated in sequence m -O n The microprocessor sends a control signal to the corresponding digital-to-analog converter to change the driving voltage loaded on the optical switch unit to be calibrated; if the optical switch to be calibrated is on the right side of the middle stage, reading a voltage signal of the analog-to-digital converter on the left side, otherwise, reading a voltage signal of the analog-to-digital converter on the right side; continuously searching the driving voltage loaded on the optical switch unit to be calibrated through a certain search algorithm until a convergence condition is met; repeating the back and forth till the path I to be calibrated m -O n All the optical switch units reach the convergence condition;
selecting the next path to be calibrated, and repeating the steps until the driving voltage values of the working states of all the units of the NxN optical switch are obtained; and writing the obtained driving voltage into a lookup table, and finishing calibration.
Further, the switch initial voltage can be obtained through previous experience or testing of a testing device.
Further, the searching algorithm comprises one-dimensional golden section, linear scanning, gradient descent, particle swarm optimization and the like.
Further, the convergence condition is that the voltage signal read from the digital converter is maximum or minimum. The voltage signal is maximum, which represents that the unit is in a state required by a calibrated path; the minimum voltage signal indicates that the cell is in the opposite state of the nominal path.
Compared with the prior art, the invention has the following beneficial effects:
1. the on-chip photoelectric detectors on the redundant ports of the intermediate-stage optical switch units are used for monitoring the optical power in the routing path, the number of the detectors is small, and the size of a chip is reduced; the calibration complexity of the switch array unit is reduced, and the calibration of the working states of all units in the optical switch array is quickly and automatically completed by means of a peripheral control circuit and an optimization algorithm.
2. And the number of trans-impedance amplifiers and analog-to-digital converters is reduced by adopting an analog switch mode, so that the cost is effectively reduced.
Drawings
FIG. 1 is a schematic diagram of an automatic calibration system for a silicon-based NxN optical switch chip according to the present invention;
FIG. 2 is an extended Benes switch array of embodiment 1 of a silicon-based NxN optical switch chip;
FIG. 3 is a calibration diagram of an extended Benes switch array of embodiment 1 of a silicon-based NxN optical switch chip;
FIG. 4 is a flow chart of the method for automatically calibrating a silicon-based NxN optical switch chip according to the present invention;
FIG. 5 is a 2 × 2MZI optical switch
FIG. 6 is the transmission lines of two input ports
FIG. 7 is a single switch unit 'Cross' state calibration process
FIG. 8 is the transmission line of path 5-4
FIG. 9 shows the transmission lines for routing an optical signal from input No. 5 to Path 4
Detailed Description
The technical solutions of the present invention are further described below with reference to the drawings and examples, but the scope of the present invention should not be limited thereby.
The automatic calibration system of the structure of the invention is shown in figure 1 and comprises a laser: a single wavelength input optical signal is generated. 3dB optical splitter: the light equal energy is divided into 2 paths. Programmable polarization controllers, number 2: the polarization state of the input optical signal may be adjusted by a control signal. The number of 1 × N optical switches is 2, and an input optical signal can be output from any one of the N ports, or an optical signal input from any one of the N ports can be output from one port. Optional 1N optical switch includes 1N silicon-based optical switch, 1N MEMS optical switches, 1 × N mechanical optical switches, and the like are used in systems to select optical signals to be input from different input ports to an N × N optical switch array chip. In this embodiment, an extended Benes nx 0N optical switch array chip is adopted, and as shown in fig. 2, the chip includes (2 × 1log 2N-1) × 2N/2 × 32 optical switches in total, and is distributed in N/2 rows and (2 × 4log 2N-1) columns, and is arranged in a Benes topology structure. Wherein each 2 × 52 optical switch of the intermediate stage (log 2N column) is composed of 4 2 × 62 switch units, and the 4 2 × 72 switch units are connected with each other in a cross manner, as shown by a dashed box in fig. 2; the 2 × 82 optical switches in each of the remaining columns are each constituted by 12 × 92 switching unit. The 2 × 2 switching unit may be a 2 × 02 mach-zehnder interferometer structure, a 2 × 12 microring resonator structure, or the like, and the total number in the chip is (2 × log2N + 2) × N/2. Each 2 × 2 switch unit in the middle column has a redundant port, and is connected to one on-chip photodetector, and the total number of photodetectors is 4M (M = N/2). The on-chip photo-detector may be a silicon germanium PIN detector for monitoring optical power in the switch path. The 1 multiplied by 2M analog circuit selection switches are 2 in number, and each 1 multiplied by 2M analog switch is respectively connected with the N photoelectric detectors and used for reading the photoelectric current in the photoelectric detectors on the path. And the trans-impedance amplifiers are used for amplifying the photocurrent output by the photoelectric detector into voltage signals which can be collected by the analog-to-digital converter, and the number of the voltage signals is 2. The analog-to-digital converter reads the analog voltage signal, and reads the voltage output by the transimpedance in the system, wherein the number of the voltage output is 2. And the microprocessor is used for realizing a hardware platform of an automatic calibration algorithm, automatically adjusting the output port of the 1 xN optical switch unit, the output port number of the analog switch and assigning a new voltage value to the analog-to-digital converter by receiving the voltage value fed back by the analog-to-digital converter and outputting an algorithm feedback signal. Computer and Field Programmable Gate Array (FPGA) can be selected. The digital-to-analog converter is used for reading the analog signal fed back by the microprocessor, converting the analog signal into a digital signal and feeding the digital signal back to the drive amplifier (Driver), and the number of the digital-to-analog converter is (2+2 Xlog) 2 N). Times.N/2. The drive amplifiers convert the digital signal signals fed back by the analog-to-digital converter into corresponding voltage values and load the voltage values on corresponding optical switch units, and the number of the drive amplifiers is (2+2 Xlog) 2 N)×N/2。
The automatic calibration system based on the extended-stage Benes optical switch array is shown in fig. 1, the extended-stage nxnbenes optical switch array rapidly completes the calibration of the state of the switch unit by means of the on-chip photodetector, the peripheral control circuit and the automatic calibration algorithm, and the specific calibration process is shown in fig. 3 and 4.
The method for automatically calibrating the extended Benes nxn optical switch array chip of the embodiment comprises the following steps:
1) The microprocessor sends a control signal to the laser and outputs an optical signal, the signal passes through the 3dB optical beam splitter to form two beams of light with the same single wavelength energy and respectively enter the programmable polarization controller, and the whole process is the output optical signal in the figure 4.
2) Selecting a path I to be calibrated m -O n ，(I m And O n The mth input port and the nth output port of the silicon-based N × N optical switch chip, m, N =1,2, …, N), and a path to be calibrated is shown by a black solid line in fig. 3, where m = N-2 and N = N-3. The microprocessor sends control signals to the left and right 1 XN optical switches respectively, so that optical signals input from the left and right are input into the NXN optical switch chip from the Nth-2 (mth) and the Nth-3 (nth) ports respectively. If the path I to be calibrated m -O n Through the middle stage (T = log) 2 N-1 (i =1,2, … N) optical switch of N stages), specifically, the left 2 nd (p-th, p =1,2) 2 × 2 switch unit and the right 2 nd (q-th, q =1,2) 2 × 2 switch unit, the microprocessor sends control signals to the two side 1 × 2M analog switches, so that the two left and right analog optical switches are respectively switched to the 2N-2 (2 × i-2+p) and 2N-2 (2 × i-2+q) paths, and the middle stage (T = log) is connected to the second analog optical switch 2 N stages) of the left 2 nd and right 2 nd on-chip photodetectors of the N-1 (ith) 2 x 2 optical switches. An input optical signal is converted into photocurrent by a photoelectric detector, the photocurrent passes through an analog switch and then enters a transimpedance amplifier, and the input photocurrent is converted into a voltage signal by the transimpedance amplifier and is collected by an analog-to-digital converter and fed back to a microprocessor. At this time, the microprocessor reads the voltage fed back by the analog-to-digital converter and sends a control signal to adjust the state of the polarization controller, so that the polarization controller can be controlled to workSo that the photoelectric detectors on the two sides output larger photocurrent. The whole process is the 'selecting calibration path' in the figure four "
3) Next, the process of "calibrating all switch units in path" in FIG. 4 is performed, and the path I to be calibrated in FIG. 3 is selected m -O n And loading initial voltage required in the routing state to the optical switch unit on the path, so that the photodetectors on two sides of the intermediate-stage switch unit can detect larger photocurrent, wherein the initial voltage values of the "Bar" state and the "Cross" state of the switch unit can be obtained by a test device. In addition, since the photo detector is not on the black path, in order to make the photo detector obtain a large photo current for feeding back the change situation of the link optical signal, the 2 × 2 optical switch units on the left and right sides of the middle stage (T-th stage) need to operate in the opposite state ("Bar → Cross") to that in the black path, at this time, the 2 × 2 switch units on the left and right sides of the T-th stage on the black path are loaded with the voltage value in the "Cross" state, and thus the step of "loading the initial voltage" in fig. four is completed.
4) The "calibration switch unit" step in fig. 4 then begins. First, the load voltage of the first stage switching unit is calibrated. As can be seen from the path to be calibrated in fig. 3, the optical signal is routed to the photodetector (2 × i-2+q) on the 2 × 2 switch unit on the right side of the middle stage (T stage) through the first stage optical switch unit in the 'Bar' state. As can be seen from fig. 5, when the 2 × 2 optical switch unit has the phase difference between the upper and lower arm optical signals
When nearby changes, the optical power input from the port 1 to the port 3 is more sensitive to phase changes, so that the difference between two phase arms of the switch unit at the moment can be judged by whether the optical power output from the port 3 reaches the minimum value or not
If the current is 0, the calibration condition of the switch can be judged by detecting whether the photoelectric detector obtains the minimum photocurrent on the link. In order to quickly calibrate the voltage value of the state of the first-stage switch unit 'Cross', the initial voltage of the 'Cross' can be setAnd searching a voltage value for realizing an optimal Cross state by a one-dimensional golden section searching algorithm in a section near the voltage. The flow of the one-dimensional golden section searching method is shown as a four-dotted line in the figure, if the voltage value of the algorithm convergence is met in a given voltage range, the voltage value of the optimal 'Cross' state is found, and if the voltage value of the algorithm convergence is not met, the searching range is expanded, and the searching is continued until the voltage value is found. The specific implementation process is as follows, under the initial voltage range of the given switch working in the "Cross" state, the microprocessor calculates the next voltage value to be loaded to the switch unit through the digital-to-analog converter and the driving amplifier shown in fig. 1, the microprocessor compares the new feedback voltage signal with the last voltage signal, calculates the driving voltage of the unit next time, loads the driving voltage to the switch unit, repeats the reciprocating operation until the driving voltage value which enables the photoelectric detector to have the minimum photocurrent is found, and the voltage value is the optimal voltage of the first-stage switch working in the "Cross" state on the black path and records the optimal voltage. FIG. 7 is a diagram of a search process for calibrating the "Cross" state of a first-stage switching unit by using a one-dimensional golden section method, wherein the voltage value of the "Cross" state of the unit is 1860mV, and the calibration precision is 3mV, which are found by 24 times of iterative calculation. Next, the initial voltage of the Bar state is applied to the first stage optical switch unit again, so that the photodetector of the T +1 th stage obtains a larger optical current value again, so as to mark the state of the next optical switch unit, and thus the step of "calibrating the switch unit" in fig. 4 is completed.
As can be seen from the flowchart of FIG. 4, the calibration of the states of all the switch units (1 st to T-1 th stages, left side of T stage) on the left side of the path (1 st to T-1 st stages, left side of T stage) can be completed by sequentially calibrating the second optical switch unit and the remaining switch units on the left side by the above method. Similarly, the calibration of the states of all the switch units on the right side (the T + 1-2T-1 stage, the right side of the T stage) on the path is completed by the photoelectric detector on the switch unit on the left side on the T stage.
5) After all the switch units of the path are calibrated, returning to the second item of the flow chart in fig. 4 to select the next calibration path, and similarly in the synchronization steps 3-4), the microprocessor sends a control signal again to switch the output ports of the left and right 1 × N optical switches and the 1 × 2M analog selection switch, adjusts the programmable polarization controller to enable the photoelectric detectors on the two sides to have larger photocurrents, and repeats the step of "calibrating all the units of the path" in fig. 4 to complete the calibration of the states of all the switch units of the path to be calibrated. Repeating the round trip until the calibration of all the states of all the units of the optical switch array is completed, and the whole calibration flow is shown in the fourth diagram.
As shown in fig. 5, when the switch unit operates in a Cross state or a Bar state, it is defined that an optical signal is input from port 1 and output from port 4 as "Cross1" (1-4), an optical signal is output from port 2 as "Cross2" from port 3, an optical signal is output from port 3 as "Bar1" from port 1, and an optical signal is output from port 4 as "Bar2" from port 2. There is a little phase deviation in the phase of the light output through the 3dB coupler due to process errors. FIG. 6 is a definition
When other parameters are ideal parameters, the transmission matrix spectral line diagram of the 2 × 2 optical switch unit, as can be seen from fig. 6, after an optical signal is input from the port 1,2, the "Bar" and "Cross" states of the maximum extinction ratio cannot be realized under the same phase difference, if the phase difference under the highest extinction ratio is realized by any port alone to represent the working state of the whole switch at that time, the other port cannot realize a high extinction ratio, so the phase difference calibration method under the single-port optimum cannot embody the optimum working performance of the 2 × 2 optical switch unit. Therefore, in practical tests, the average value of the operating voltage of the "Cross1" for achieving the optimal extinction ratio and the operating voltage of the "Cross2" for achieving the optimal extinction ratio can be used to represent the operating voltage of the switch in the "Cross" state, so that the optical signal input from one port or two ports can simultaneously achieve a higher extinction ratio, as shown by the six black dots in the figure.
Therefore, the input signal only needs to pass through the paths of 'All-Cross' and 'All-Bar' (2N in total) to complete the calibration of All the required states of All the switches. By using the method, 4 driving voltages are obtained for each 2 × 2 switch unit, and the driving voltages are respectively in a circuit with states of 'Cross 1', 'Cross 2', 'Bar 1' and 'Bar 2'. If the routing of the left switch unit is changed to be in the states of 'Cross 1', 'Cross 2', 'Bar 1' and 'Bar 2', the working voltages in the states of 'Bar 1', 'Bar 2', 'Cross 1' and 'Cross 2' are actually obtained in the calibration method process by the method for obtaining the minimum photocurrent, and the working voltages in the states opposite to the routing state are obtained on the right side in the same way. The voltage values of all the cells in the N × N switch array in the operating state can be obtained by averaging the voltages in step 6, that is, each 2 × 2 switch cell has an operating voltage in two states of "Cross" and "Bar". All operating voltages are stored in the microprocessor by means of a look-up table.
FIG. 8 is a graph showing the magnitude of "Cross" and "Bar" state voltages after calibration of an expanded Benes 32X 32 optical switch array, where the "Cross" state voltage is mainly distributed around 2000mV and the "Bar" state voltage is mainly distributed around 3200 mV. And loading the calibration voltage to a switch unit of the transmission path and detecting the output power of the target port and the crosstalk port to obtain the transmission spectral line of the path. Fig. 9 shows transmission lines of an optical signal routed from the input No. 5 to the path 4, where the optical isolation between the target output port and the unrelated port can reach about 28dB under a bandwidth of 80nm, and is consistent with the extinction ratio of the device of the test switch unit, which indicates that the switch unit on the path is in a target state, and the calibration accuracy of the whole calibration system is relatively high.
Claims (7)
1. An automatic calibration system for a silicon-based N x N optical switch chip is characterized by comprising:
the light source is used for generating two paths of input optical signals with equal energy and single wavelength;
two programmable polarization controllers which are symmetrically arranged and used for respectively receiving the input optical signals and adjusting the polarization state of the input optical signals through the respectively received control signals;
two 1 × N optical switches symmetrically arranged for receiving a feedback signal and controlling the input optical signal to be input from any one of 1 input port and output from any one of N output ports, or controlling the input optical signal to be input from any one of N input ports and output from 1 output port;
a silicon-based NxN optical switch chip, which comprises 4M photoelectric detectors, a first-stage 2 x 2 optical switch, a second-stage 2 x 2 optical switch, a … …, a Tth-stage 2 x 2 optical switch, a … …, a 2T-2-stage 2 x 2 optical switch and a 2T-1-stage 2 x 2 optical switch; the Tth-stage 2 x 2 optical switch consists of M2 x 2 optical switches, each 2 x 2 optical switch comprises 4 redundant ports, and the left end and the right end of each 2 optical switches are respectively connected with the photoelectric detectors;
the two 1 × 2M analog selection switches are symmetrically arranged, and 2M ports of each 1 × 2M analog selection switch are respectively connected with 2M photoelectric detectors, and are used for outputting real-time photocurrents of the photoelectric detectors on different paths and transmitting the real-time photocurrents to corresponding transimpedance amplifiers;
the two transimpedance amplifiers are symmetrically arranged and used for amplifying the received light current and converting the received light current into a voltage signal;
the two analog-to-digital converters are symmetrically arranged and used for reading the voltage signal output by the transimpedance and converting the voltage signal into a digital signal;
the microprocessor is used for realizing a hardware platform of an automatic calibration algorithm, outputting a feedback signal and a feedback voltage value to the digital-to-analog converter by receiving the voltage value fed back by the analog-to-digital converter, automatically switching an output port of the 1 xN optical switch unit according to the automatic calibration algorithm, outputting a port number by the analog switch and feeding back a new voltage value to the digital-to-analog converter;
the digital-to-analog converter is used for reading a feedback signal of the microprocessor, converting the feedback signal into an analog signal and feeding the analog signal back to the driving amplifier;
and the driving amplifier converts the digital signals fed back by the digital-to-analog converter into corresponding voltage values and loads the voltage values on the corresponding 2 multiplied by 2 switching units.
2. The silicon-based nxn optical switch chip automatic calibration system of claim 1, wherein the network topology of the silicon-based nxn optical switch chip comprises Benes, extended Benes, crossbar, double Layer Network (DLN) structure;
the 2 x 2 optical switches are arranged in a certain network topological structure, the number of the optical switches is determined by the topological structure, and the optical switches are 2 x 2 Mach-Zehnder interferometer structures or 2 x 2 micro-ring resonator structures;
if the T-th stage 2 × 2 optical switch has no redundant port, the 2 × 2 optical switch is configured by cross-connecting 4 2 × 2 optical switches.
3. The silicon-based nxn optical switch chip auto-calibration system of claim 2, wherein the photo-detector is a silicon germanium PIN-type detector.
4. An automatic calibration method for a silicon-based NxN optical switch chip is characterized by comprising the following steps:
outputting an optical signal: the microprocessor starts the light source and respectively introduces two paths of input optical signals with equal energy and single wavelength into the respective programmable polarization controllers;
selecting a calibration path: selecting a path I to be calibrated m -O n (I m And O n The optical switch chip comprises an mth input port and an nth output port of a silicon-based N multiplied by N optical switch chip, wherein m, N =1,2, …, N), optical signals are input into a 1 multiplied by N optical switch array from the left and the right, namely, the left optical signal is input into the N multiplied by N optical switch array chip from the mth port, and the right optical signal is input into the N multiplied by N optical switch chip from the nth port; path to be calibrated I m -O n Through the ith (i =1,2, … M) optical switch of the middle stage, the microprocessor sends control signals to the 1 × 2M analog switches on the two sides, so that the two left and right 1 × 2M analog switches are respectively switched to the 2 × i-2+p and 2 × i-2+q paths (p, q =1,2), and are connected with the p-th left and q-th right photodetectors of the ith optical switch unit of the middle stage; the microprocessor obtains the optical signal input to the intermediate-level optical switch by reading the voltage signals input by the two analog-to-digital converters;
initializing a calibration path: according to the read voltage signal, the microprocessor sends a control signal to the programmable polarization controller, and adjusts the polarization state of the introduced input optical signal to maximize the optical signal of the intermediate-level optical switch; hair-like deviceSending a control signal to the digital-to-analog converter to treat the path I to be calibrated m -O n All the optical switch units are loaded with required initial voltage, so that the voltage signal read from the analog-to-digital converter is larger;
a calibration switch unit: calibrating paths I to be calibrated in sequence m -O n The microprocessor sends a control signal to the corresponding digital-to-analog converter to change the driving voltage loaded on the optical switch unit to be calibrated; if the optical switch to be calibrated is on the right side of the middle stage, reading a voltage signal of the analog-to-digital converter on the left side, otherwise, reading a voltage signal of the analog-to-digital converter on the right side; continuously searching the driving voltage loaded on the optical switch unit to be calibrated through a certain search algorithm until a convergence condition is met; repeating the back and forth till the path I to be calibrated m -O n All the optical switch units reach the convergence condition;
selecting the next path to be calibrated, and repeating the steps until the driving voltage values of the working states of all the units of the NxN optical switch are obtained; and writing the obtained driving voltage into a lookup table, and finishing calibration.
5. The method according to claim 4, wherein the switch initial voltage is obtained through previous experience or testing of devices.
6. The silicon-based N x N optical switch chip automatic calibration method according to claim 4, wherein the search algorithm comprises one-dimensional golden section, linear scanning, gradient descent, particle swarm algorithm, and the like.
7. The method as claimed in claim 4, wherein the convergence condition is the maximum or minimum voltage signal read from the digital converter. The voltage signal is maximum, which represents that the unit is in a state required by a calibrated path; the minimum voltage signal indicates that the cell is in the opposite state of the nominal path.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211512218.6A CN115792763A (en) | 2022-11-29 | 2022-11-29 | Silicon-based NxN optical switch chip automatic calibration system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211512218.6A CN115792763A (en) | 2022-11-29 | 2022-11-29 | Silicon-based NxN optical switch chip automatic calibration system and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115792763A true CN115792763A (en) | 2023-03-14 |
Family
ID=85443160
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211512218.6A Pending CN115792763A (en) | 2022-11-29 | 2022-11-29 | Silicon-based NxN optical switch chip automatic calibration system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115792763A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117081664A (en) * | 2022-09-05 | 2023-11-17 | 之江实验室 | Test method of optical switch unit based on Benes network |
-
2022
- 2022-11-29 CN CN202211512218.6A patent/CN115792763A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117081664A (en) * | 2022-09-05 | 2023-11-17 | 之江实验室 | Test method of optical switch unit based on Benes network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111427122B (en) | NxN silicon-based polarization independent optical switch system | |
| JP2828134B2 (en) | Optical switch and optical information transmission device | |
| CN113805641B (en) | Photonic neural network | |
| KR20230129447A (en) | Balanced photonic architecture for matrix computation | |
| Chiaroni et al. | First demonstration of an asynchronous optical packet switching matrix prototype for multi-terabit-class routers/switches | |
| CN115792763A (en) | Silicon-based NxN optical switch chip automatic calibration system and method | |
| Dupuis et al. | An 8× 8 silicon photonic switch module with nanosecond-scale reconfigurability | |
| CA3163597A1 (en) | On-chip monitoring and calibration circuits for frequency modulated continous wave lidar | |
| Testa et al. | Integrated reconfigurable silicon photonics switch matrix in IRIS project: Technological achievements and experimental results | |
| CN113406747B (en) | Wavelength division multiplexer and silicon optical integrated chip | |
| US20220123840A1 (en) | Optical redistribution layers for high-channel-count photonics | |
| WO2013177514A1 (en) | Coherent combiner for high power beams | |
| Feyisa et al. | Compact 8× 8 SoA-based optical WDM space switch in generic INP technology | |
| US5271074A (en) | Integrated optical waveguide apparatus | |
| US9755738B2 (en) | Crosstalk suppression in a multi-photodetector optical channel monitor | |
| JP4344145B2 (en) | Optical signal switching device and control method of optical signal switching device | |
| CN112383357A (en) | Power balancing device of space division multiplexing optical switching node | |
| CN117255967A (en) | Balanced photon architecture for matrix computation | |
| CN114336257B (en) | Accurate phase control method of all-fiber laser phased array system | |
| US11709317B2 (en) | Wavelength demultiplexer, optical transceiver front-end module, photonic circuit, and wavelength demultiplexing control method | |
| Rohit et al. | Multi-path routing at 40Gb/s in an integrated space and wavelength selective switch | |
| WO2022228136A1 (en) | Optical power adjustment system and optical power adjustment device | |
| Heck et al. | Integrated recirculating optical buffers | |
| CN115987387A (en) | Switch unit calibration method based on Benes network | |
| CN117454950A (en) | Neuron based on single-row carrier photoelectric detector |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |