A switch unit calibration method based on benes network

Abstract

The application provides a switch unit calibration method based on a Benes network, which comprises the following steps: selecting a switch unit in the middle stage of the network, selecting a route from the input port to the output port according to the switch unit, scanning the voltage of the measured unit, and obtaining the optimal working voltage of the measured unit; then, starting from the outermost switch unit, all switch units are calibrated to the middle stage, and the method is that other switches affecting the test switch crosstalk path are set to a state deviating from the crosstalk path, so that the optical power entering the crosstalk path is minimized, thereby approximately ignoring the influence of the crosstalk optical power, and the optimal working voltage of the measured unit is obtained by scanning. Due to the symmetry of the Benes network, the left and right switches are equivalent. The switch unit calibration method of the application does not need to set a monitoring point in the network, greatly reduces the packaging difficulty of a large-scale integrated switch array, and the application can match an automatic test scheme, thereby providing support for subsequent larger-scale optical switch chip manufacturing and testing.

Description

Technical Field

[0001] This invention relates to the field of silicon-based optoelectronic integration technology, and in particular to a method for calibrating switching units based on Benes networks. Background Technology

[0002] With the rapid development of photonic integration technology, large-scale optical switching chips are receiving increasing attention in cloud computing, high-performance computing, and data centers. However, due to errors in CMOS processes, the initial states of optical switching units at different locations within the entire Benes network are not the same, requiring testing and calibration of the optimal operating voltage for each optical switch. Currently, the mainstream calibration scheme involves setting monitoring ports after each switching unit or at specific locations for optical detection. These monitoring ports undoubtedly exacerbate the challenges of subsequent optical and electrical packaging of large-scale chips. This voltage calibration problem for switching units has become one of the factors limiting the number of I / O ports in large-scale optical switching chips. Therefore, there is an urgent need for a solution that eliminates the need for built-in monitoring ports and calibrates all switching units in the network solely through the characteristics and input / output ports of the Benes network. Summary of the Invention

[0003] In view of the shortcomings of the prior art, the purpose of this invention is to provide a switching unit calibration method based on Benes network to overcome the deficiencies in the prior art.

[0004] To achieve the above objectives, the present invention is implemented through the following technical solution:

[0005] A method for calibrating switching units based on Benes networks includes the following steps:

[0006] (1) In the Benes network, select the switch unit where the intermediate level is located as the first switch unit to be tested, and select a route from the input port to the output port in a fully cross state according to the switch unit;

[0007] (2) Set a pre-biased bias crossover state for each switch unit on the route, and input an optical signal with an intensity of P to the input port of the switch unit to be tested. i And connect an optical power meter to its output port to detect the output optical signal P. o ;

[0008] (3) Taking the cross-state of the switch as the transmission, its transmittance is T, and the straight-through state is taken as the crosstalk term X. Scan the voltage of the switch unit under test and obtain the point with the minimum output optical power as the point of the switch unit that deviates most from the cross-state under the selected cross-path, that is, the optimal operating point of the switch unit in the straight-through state; the point with the maximum output optical power is the optimal operating point of the switch unit in the cross-state; select the full cross-path corresponding to each switch unit in the intermediate stage in sequence, and obtain the accurate optimal operating voltage of each switch unit in the straight-through and cross-through states according to step (2);

[0009] (4) Select the outermost switch unit in the Benes network as the switch unit to be tested, and select a fully cross route from the input port to the output port according to the switch unit;

[0010] (5) Repeat step (2) and set the intermediate level switches through which the route passes to the accurate cross-state voltage that has been measured. In addition, set the switch units through which the other crosstalk optical signal path of the test switch unit passes to a state that deviates from the main route. Scan the voltage of the unit under test and measure its optimal working voltage in the pass-through and cross-state. In turn, calibrate the optimal working voltage in the pass-through and cross-state of each switch unit of the outermost level.

[0011] (6) Starting from the outermost switch, test each switch inwards, and finally select the switch unit closest to the middle level in the Benes network as the switch unit to be tested. Select a route from the input port to the output port that is fully crossed according to the switch unit. Repeat step (5) to set the tested switches to the accurate cross-state voltage and set the switch units through which the other crosstalk optical signal path of the test switch unit passes to the state of deviating from the main route. Finally, calibrate in sequence to obtain the optimal working voltage of each switch unit on both sides of the middle level in the through and cross states.

[0012] Specifically, in step (1), the route selected in the Benes network for the full cross state can encompass all switching units without repetition, and setting a voltage of +1V to the switching unit can make the switching unit biased towards the cross state.

[0013] Specifically, in step (2), the pre-biasing of the biased crossover state is set as follows: the upper and lower phase shift arms of the switch based on the Mach-Zehnder interferometer (MZI) structure are designed as two parallel PIN structures with opposite polarities in a push-pull working mode, and the initial phase of π / 2 is reduced on a certain phase shift arm, so that the initial state of the switch unit is in the middle state between crossover and through, and the positive and negative signs of the optimal working points of crossover and through are always correct and predictable.

[0014] Furthermore, in step (3), the transmittance T and crosstalk term X of a normal switching unit differ by about 20 dB, which is about two orders of magnitude. When calculating the output optical power, the optical signal is multiplied by the state corresponding to each switch it passes through.

[0015] Furthermore, in step (3), the output optical signal P o The transmission coefficient T of the measured unit is linearly related to the voltage. The transmission coefficient T of the measured unit changes with the voltage. Therefore, scanning the voltage and optical power of the measured unit can reveal the crossover and through states of the switching unit.

[0016] Furthermore, the output optical signal P o The linear relationship between the transmittance and the unit under test is due to the pre-bias setting maximizing the signal light required for our test. Setting the switch unit through which the other crosstalk signal path of the test switch unit passes to deviate from the main route minimizes the crosstalk light that interferes with the test results; thus, it ensures that the output optical signal P... o It has an approximately linear relationship with the transmittance of the measured unit.

[0017] Furthermore, in step (5), due to the left-right symmetry of the Benes network, the left and right sides of the Benes network are equivalent through input-output swapping.

[0018] The beneficial effects of this invention are as follows:

[0019] This invention only requires optical power detection at the input and output ports, and pre-biasing using the existing electrical ports of the switching units, ensuring that the transmittance of each switching unit in the path is approximately linearly related to the optical power. In subsequent applications of large-scale optical switching chips, the optimal operating point of each switch can be automatically tested using only an optical fiber array. This significantly reduces the difficulty of calibrating and testing optical switching units in large-scale optical switching chips. Attached Figure Description

[0020] Figure 1 These are test operation diagrams in an example of the present invention; wherein, (a) is the full cross-route diagram selected by the switching unit S4-8 in the embodiment of the present invention; (b) is the voltage preset diagram of the crosstalk optical power deviating from the original route when operating the switching unit S7-5 in the embodiment of the present invention; and (c) is the voltage preset diagram of the crosstalk optical power deviating from the original route when operating the switching unit S4-8 in the embodiment of the present invention.

[0021] Figure 2 This is a schematic diagram of the phase shift arm design of the switching unit in an embodiment of the present invention;

[0022] Figure 3The figure shows the relationship between optical power and voltage of a certain switching unit; where (a) is the curve result obtained by the method of the present invention, and the dashed line represents the optimal operating voltage in the switch-on state; (b) is the curve obtained by setting a monitoring port at a specific location. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. However, it should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of the invention. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concept of the invention.

[0024] This invention provides a method for calibrating switching units based on a 16×16 Benes network, comprising the following steps:

[0025] S1. In the Benes network, the intermediate stage switching unit S4-8 is selected as the first switching unit to be tested. Based on this switching unit, a fully cross-connected route is selected, with input at port 1 and output at port 9'. For example... Figure 1 As shown in Figure (a);

[0026] S2. Set a +1V pre-bias to each switching unit on this route to ensure a high optical power under this route. The optical signal intensity input to the input port is P. i And connect an optical power meter to the output port to detect the output optical signal P. o ;

[0027] S3. Taking the crossover state of the switch as transmission and the transmittance as T, the straight-through state is taken as crosstalk term X. In a normal switching unit, T and X differ by approximately 20 dB, about two orders of magnitude. Each time the optical signal passes through a switch, it is multiplied by the corresponding state of that switch. Therefore, the final output optical power is: P o =P i *T 11 *T 25 *T 37 *X 48 *T 58 *T 67 *T 75 +C0, where C0 is the crosstalk light from the upper sub-network that enters the switch S after multiple stages of transmission. 75 Up port;

[0028] S4. Scan the voltage of the measured unit S4-8 separately. Since C0 is a constant relative to the measured unit, the output optical signal P... o The transmittance X of the measured cell 48The relationship is linear. The point with the minimum output optical power is the optimal operating voltage for the S4-8 switching unit in the direct-on state, and the point with the maximum output optical power is the optimal operating voltage for the S4-8 switching unit in the cross-connected state.

[0029] S5. Following the same approach, calibrate the optimal operating voltage for each switching unit in the intermediate stage for its pass-through and cross-through states.

[0030] S6. Second step: In the Benes network, select the outermost switch unit S7-5 as the switch unit to be tested. The full cross-route remains as follows. Figure 1 As shown in Figure (a);

[0031] S7. Repeat steps S2 and S3 to set the switches on the router to apply a +1V voltage. Set the previously tested switches S4-8 to the accurate crossover voltage. At this time, P o =P i *T 11 *T 25 *T 37 *T 48 *T 58 *T 67 *X 75 +C1*T 75 The value of C1 is related to the crosstalk light intensity of the upper sub-network. It is only necessary to set the switching unit through which the other crosstalk light signal path of the test switching unit passes to a state deviating from the main route (to transmit the light to other places as much as possible), such as... Figure 1 As shown in Figure (b), the crosstalk light C1*T 75 ≈P i *X 4 *T 3 Compared to P i *T 11 *T 25 *T 37 *T 48 *T 58 *T 67 *X 75 ≈P i *X 75 *T 6 The difference is at least 6 orders of magnitude, so its effect on P is negligible. i *T 6 *X 75 Interference of the term, output optical signal P o The voltage of the measured unit (S7-5) is approximately linearly related to the transmittance of the measured unit. The voltage of the measured unit S7-5 is scanned, and its optimal operating voltage in both through and cross states is measured. The optimal operating voltage of each outermost switching unit in both through and cross states is then calibrated sequentially.

[0032] S8. At the same time, due to the left-right symmetry of the Benes network, simply replacing the left 1 input 9' output with 9' input 1 output is sufficient to test the optimal operating voltage of switch S1-1 in both the through and cross states.

[0033] S9. Test each switch sequentially from the outermost level inwards, finally selecting the switch unit S5-8 closest to the intermediate level in the Benes network as the switch unit to be tested. Then, select a fully cross-connected route from the input port to the output port based on the switch unit. Figure 1 As shown in Figure (a);

[0034] S10. Repeat steps S7 and S8 to set the tested switch to the accurate cross-state voltage. At this time, P o =P i *T 11 *T 25 *T 37 *T 48 *X 58 *T 67 *T 75 +C2*T 58 The value of C2 is related to the magnitude of the crosstalk light. Similarly, the switching unit through which the other crosstalk light signal path of the test switching unit passes is set to deviate from the main route (to transmit the light to other places as much as possible), such as... Figure 1 As shown in Figure (C). At this time, the crosstalk light C2*T 58 ≈P i *X 2 *T 5 The difference from the signal light is minimal, but still two orders of magnitude. Furthermore, since all other switches have been calibrated to their optimal operating points, the crosstalk light will have a larger difference compared to the signal light, but its impact on the signal path can still be ignored. The voltage of the unit under test S5-8 is scanned, and its optimal operating voltage in both through and cross states is measured. The optimal operating voltage in both through and cross states of each switch unit on both sides of the intermediate stage is then calibrated sequentially.

[0035] Due to the output optical signal P o The linear relationship between the transmittance and the unit under test is due to the pre-bias setting maximizing the signal light required for our test. Setting the switch unit through which the other crosstalk signal path of the test switch unit passes to deviate from the main route minimizes the crosstalk light that interferes with the test results; thus, it ensures that the output optical signal P... o It has an approximately linear relationship with the transmittance of the measured unit.

[0036] Figure 2The diagram shows the design principle of the phase shift arm of the switching unit in this embodiment of the invention. Specifically, the upper and lower phase shift arms of the switch based on the Mach-Zehnder interferometer (MZI) structure are designed as two parallel PIN structures with opposite polarities in a push-pull operation mode. The initial phase is reduced by π / 2 on one of the phase shift arms, so that the initial state of the switching unit is in an intermediate state between crossover and through. Even with manufacturing errors, the sign of the optimal operating voltage in the crossover and through states is always correct and predictable. This ensures that when setting the pre-bias voltage, the optical power of the main signal path can be as high as possible, while the optical power of the crosstalk path can be as low as possible.

[0037] Figure 3 The diagram shows the relationship between optical power and voltage in a certain switching unit. For example, ... Figure 3 Figure (a) represents the curve results obtained by the method used in this invention; Figure 3 Figure (b) shows the curve obtained by setting a monitoring port at a specific location as previously proposed. The dashed line represents the optimal operating voltage in the direct-on state of the switch. The optimal operating voltages in the two direct-on states differ by only 20mV. In actual testing, the same scheme also exhibits an error of ±20mV during repeated tests. Therefore, the scheme of this invention can be considered practically feasible.

[0038] Furthermore, due to the left-right symmetry of the Benes network, the left and right sides of the Benes network are equivalent by swapping the input and output.

[0039] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for calibrating switching units based on Benes networks, characterized in that, Includes the following steps: (1) Select the switch unit where the intermediate level is located in the Benes network as the first switch unit to be tested, and select a route from the input port to the output port in a fully cross state according to the switch unit; (2) Set a pre-biased bias crossover state for each switch unit on the route, and input an optical signal with a light intensity of at the input port of the switch unit to be tested. An optical power meter is connected to its output port to detect the output optical signal. ; (3) Taking the cross-state of the switch as the transmission, its transmittance is T, and the straight-through state is taken as the crosstalk term X. The voltage of the switch unit under test is scanned, and the point with the minimum output optical power is the point where the switch unit deviates most from the cross-state under the selected cross-path, that is, the optimal operating point of the switch unit in the straight-through state; the point with the maximum output optical power is the optimal operating point of the switch unit in the cross-state; the full cross-path corresponding to each switch unit in the intermediate stage is selected in sequence, and the optimal operating voltage of each switch unit in the straight-through and cross-through states is obtained according to step (2); (4) Select the outermost switch unit in the Benes network as the switch unit to be tested, and select a fully cross route from the input port to the output port according to the switch unit; (5) Repeat step (2) and set the intermediate level switches through which the route passes to the accurate cross-state voltage that has been measured. In addition, set the switch units through which the other crosstalk optical signal path of the test switch unit passes to a state that deviates from the main route. Scan the voltage of the unit under test and measure its optimal working voltage in the pass-through and cross-state. In turn, calibrate the optimal working voltage in the pass-through and cross-state of each switch unit of the outermost level. (6) Starting from the outermost switch, test each level inwards. Finally, select the switch unit closest to the middle level in the Benes network as the switch unit to be tested, and select a route from the input port to the output port that is fully crossed according to the switch unit. Repeat step (5), set the tested switches to the accurate cross-state voltage, and set the switch units through which the other crosstalk optical signal path of the test switch unit passes to the state of deviating from the main route. Finally, calibrate in sequence to obtain the optimal working voltage of each switch unit on both sides of the middle level in the through and cross states.

2. The method for calibrating a switching unit based on a Benes network according to claim 1, characterized in that, In step (1), the route selected in the Benes network for the full cross state can encompass all switching units without repetition, and setting a voltage of +1V to the switching unit can make the switching unit biased towards the cross state.

3. The method for calibrating a switching unit based on a Benes network according to claim 1, characterized in that, In step (2), the pre-biasing of the biased crossover state is set as follows: the upper and lower phase shift arms of the switch based on the Mach-Zehnder interferometer (MZI) structure are designed as two parallel PIN structures with opposite polarities in a push-pull working mode, and the initial phase of π / 2 is reduced on a certain phase shift arm, so that the initial state of the switch unit is in the middle state between crossover and through, and the positive and negative signs of the optimal working point of crossover and through are always correct and predictable.

4. The method for calibrating a switching unit based on a Benes network according to claim 1, characterized in that, In step (3), the transmittance T and crosstalk term X of a normal switching unit differ by 20dB, which is two orders of magnitude. When calculating the output optical power, the optical signal is multiplied by the state corresponding to each switch it passes through.

5. The method for calibrating a switching unit based on a Benes network according to claim 1, characterized in that, In step (3), the output optical signal The transmission coefficient T of the measured unit is linearly related to the voltage. The transmission coefficient T of the measured unit changes with the voltage. Therefore, scanning the voltage and optical power of the measured unit can reveal the crossover and through states of the switching unit.

6. The method for calibrating a switching unit based on a Benes network according to claim 4, characterized in that, The output optical signal The linear relationship between the transmittance and the unit under test is due to the pre-bias setting maximizing the signal light required for our test. Setting the switch unit through which the other crosstalk signal path of the test switch unit passes to deviate from the main route minimizes the crosstalk light that interferes with the test results; thus, it ensures a higher output optical signal. It has a linear relationship with the transmittance of the measured unit.

7. The method for calibrating a switching unit based on a Benes network according to claim 1, characterized in that, In step (5), due to the left-right symmetry of the Benes network, the left and right sides of the Benes network are equivalent by swapping the input and output.

Images