A programmable bipolar nanosecond pulse calibration method and generation system

Abstract

The application provides a programmable bipolar nanosecond pulse calibration method and a generating system. Target output parameters, working conditions and initial calibration parameters are obtained, a plurality of relative sampling delays are set, and positive and negative polarity calibration pulses are output alternately under the same load condition; paired sampling data is obtained through the same output node and shared observation channels, positive and negative polarity output waveforms are reconstructed, and actual absolute amplitude, actual pulse width and zero return residual are determined; the common error is determined according to the deviation between the average value of the actual values of the two polarities and the target value, the polarity differential error is determined according to the difference between the actual values of the two polarities, and then the common control parameter, the polarity adjustment parameter and the active zero return control parameter are updated; the verification pulse is output again, the calibration index is measured, the effective calibration parameters are stored when the threshold value is met, and iteration, parameter rollback or failure protection is performed when the threshold value is not met. The application can reduce the influence of observation link differences and working condition changes on the consistency of bipolar pulses.

Description

Technical Field

[0001] This application relates to the field of pulse generation and calibration technology, and in particular to a programmable bipolar nanosecond pulse calibration method and generation system. Background Technology

[0002] Programmable bipolar nanosecond pulse generation technology can be used in scenarios such as electronic device testing, sensor excitation, and high-speed pulse measurement. Bipolar nanosecond pulse generation systems typically need to output positive or negative polarity pulses according to application requirements, and the amplitude, pulse width, repetition period, and polarity arrangement of the pulses can be configured.

[0003] In a bipolar nanosecond pulse generation system, positive and negative pulses are typically generated by different output branches. Due to factors such as positive and negative supply voltage deviation, switching device characteristics, branch parasitic parameters, output matching status, load conditions, and operating temperature, the absolute amplitude and pulse width of the actual output pulse may deviate from the corresponding target values, and there may also be inconsistencies in amplitude or timing between the positive and negative pulses.

[0004] Due to the short duration of nanosecond pulses, high requirements are placed on the bandwidth, sampling rate, and trigger timing accuracy of the observation circuit. When the cost and sampling rate of the sampling circuit are limited, it is difficult to directly and completely acquire the nanosecond pulse waveform. If positive and negative polarity pulses are measured under different load conditions, observation links, or sampling times, load variations, channel differences, and time drift may be superimposed on the measurement results, thereby reducing the comparability between the actual characteristic values ​​of the positive and negative polarity pulses.

[0005] Existing calibration methods may also involve independent adjustments to the positive and negative output branches. In this case, the amplitude or timing deviations shared by the two output branches and the polarity differences between them can easily couple. Adjusting one output branch may simultaneously change the overall output level and the consistency of the positive and negative outputs, leading to mutual interference between absolute amplitude calibration, pulse width calibration, and polarity consistency calibration.

[0006] Furthermore, if the output node fails to return to the output reference level in a timely manner after the pulse ends, a zero-return residual may form, which in turn affects the baseline state of subsequent pulses and the output consistency of the pulse sequence. Therefore, accurately obtaining the actual absolute amplitude and actual pulse width of the positive and negative polarity nanosecond pulses, and coordinating and calibrating the deviations shared by the two output branches and the differences between the two output branches, remains a problem that needs to be solved in bipolar nanosecond pulse generation technology. Summary of the Invention

[0007] This application provides a programmable bipolar nanosecond pulse calibration method and generation system to solve the problems in existing bipolar nanosecond pulse generation technology, such as the difficulty in accurately determining the actual absolute amplitude and actual pulse width of positive and negative polarity pulses under the same load conditions and common observation reference, as well as the problem that common deviations and polarity differences are coupled with each other, affecting the calibration accuracy and stability.

[0008] In a first aspect, this application provides a programmable bipolar nanosecond pulse calibration method, comprising: acquiring a target absolute amplitude and a target pulse width; setting multiple relative sampling delays between a pulse trigger event and a sampling event; for each relative sampling delay, controlling a positive polarity output branch and a negative polarity output branch to output a positive polarity calibration pulse and a negative polarity calibration pulse respectively through the same output node under the same load conditions, based on the target absolute amplitude, the target pulse width, shared control parameters, and polarity adjustment parameters, and sampling through an observation channel shared by the positive and negative polarity calibration pulses to obtain positive and negative polarity paired sampling data corresponding to the relative sampling delay; determining the actual absolute amplitude and actual pulse width of the positive polarity calibration pulse and the negative polarity calibration pulse respectively based on the multiple relative sampling delays and their corresponding positive and negative polarity paired sampling data; and determining the common polarity calibration pulse based on the deviation between the average value of the actual values ​​of the positive and negative polarities and the corresponding target values ​​for the absolute amplitude and pulse width respectively. The error is used to determine the polarity difference error based on the difference between the actual values ​​of positive and negative polarities. The common amplitude parameter and polarity amplitude adjustment parameter are updated based on the common error and polarity difference error corresponding to the absolute amplitude, respectively. Similarly, the common pulse width parameter and polarity timing adjustment parameter are updated based on the common error and polarity difference error corresponding to the pulse width. The common control parameters include the common amplitude parameter and the common pulse width parameter, and the polarity adjustment parameters include the polarity amplitude adjustment parameter and the polarity timing adjustment parameter. The common amplitude parameter is used to make the actual absolute amplitudes of the positive and negative polarity calibration pulses change in the same direction, and the polarity amplitude adjustment parameter is used to make their actual absolute amplitudes change in opposite directions. The common pulse width parameter is used to make the actual pulse widths of the positive and negative polarity calibration pulses change in the same direction, and the polarity timing adjustment parameter is used to make their actual pulse widths change in opposite directions.

[0009] Furthermore, the identical load condition is formed by applying the positive polarity calibration pulse and the negative polarity calibration pulse to the same load. Under each relative sampling delay, the relative sampling delay remains constant, and the positive polarity calibration pulse and the negative polarity calibration pulse are output alternately in the polarity order of positive, negative, negative, positive. Multiple samples are taken for each polarity to obtain a positive polarity sample group and a negative polarity sample group. Statistical processing is performed on the positive polarity sample group and the negative polarity sample group respectively, and the obtained statistical values ​​are determined as the positive polarity sampling data and the negative polarity sampling data under that relative sampling delay. If the sample dispersion does not exceed the dispersion threshold, the number of effective samples reaches a preset number, and the observation channel is not saturated, the positive polarity sampling data and the negative polarity sampling data are determined as valid paired data.

[0010] Further, setting the multiple relative sampling delays includes: performing a coarse scan with a first delay step to determine the pre-pulse baseline interval, effective level interval, rising edge interval, falling edge interval, and tail interval after the pulse ends; performing a fine scan with a second delay step smaller than the first delay step to at least the rising edge interval, the falling edge interval, and the tail interval; determining the actual absolute amplitude and actual pulse width of the positive polarity calibration pulse and the negative polarity calibration pulse, including: determining the measured baselines of the positive polarity calibration pulse and the negative polarity calibration pulse respectively based on the sampling data of the pre-pulse baseline interval; reconstructing the positive polarity output waveform and the negative polarity output waveform respectively based on the sampling data obtained from the coarse and fine scans, and based on the positive polarity calibration... The polarity relationship between the pulse and the negative polarity calibration pulse relative to the output reference level is determined. Using the measured baseline corresponding to each output waveform as the zero offset reference, the positive offset of the positive polarity output waveform relative to the corresponding measured baseline and the negative offset of the negative polarity output waveform relative to the corresponding measured baseline are mapped to the same coordinate direction to obtain the polarity-normalized positive and negative polarity output waveforms. Based on the statistical values ​​of each polarity-normalized output waveform within the effective level range, the corresponding actual absolute amplitude is determined. Based on the corresponding actual absolute amplitude and a preset ratio, a pulse width determination threshold is determined. Based on the difference in the time when each polarity-normalized output waveform passes through the pulse width determination threshold in the rising edge range and the falling edge range, the corresponding actual pulse width is determined.

[0011] Further, updating the shared control parameters and the polarity adjustment parameters includes: for absolute amplitude and pulse width, determining the update amount of the shared control parameters based on the corresponding shared error and the corresponding shared control parameter sensitivity coefficient, and determining the update amount of the polarity adjustment parameters based on the corresponding polarity difference error and the corresponding polarity adjustment parameter sensitivity coefficient; limiting the single update amount and the updated parameter values; using the updated shared control parameters and polarity adjustment parameters, outputting positive polarity verification pulses and negative polarity verification pulses under the same load conditions, and re-determining the shared error and polarity difference error corresponding to the absolute amplitude and pulse width respectively; repeating parameter updates and verifications until all errors meet the corresponding calibration thresholds or the number of iterations reaches a preset upper limit; restoring the parameters before the update if the absolute value of at least one error increases by more than a preset allowable jitter amount compared to before the update, or if at least one calibration index that did not exceed the corresponding calibration threshold before the update exceeds the corresponding calibration threshold after the update.

[0012] Further, the current load condition parameters and the current temperature are obtained, the current load condition parameters being used to characterize the same load conditions; the sampling positions corresponding to the multiple relative sampling delays also cover the tail intervals after the positive polarity calibration pulse and the negative polarity calibration pulse ends; based on the sampling data of the tail interval, the zero-return residuals of the positive and negative polarity output waveforms relative to the measured baseline before the corresponding pulse are determined respectively, and the active zero-return control parameters for restoring the output node to the output reference level are updated based on the zero-return residuals, the active zero-return control parameters including at least one of the active zero-return start time and duration; in the actual absolute amplitude, actual pulse width If both the zero-return residual and the corresponding calibration pass conditions are met, the updated common amplitude parameter, polarity amplitude adjustment parameter, common pulse width parameter, polarity timing adjustment parameter, and active zero-return control parameter are associated with the target absolute amplitude, target pulse width, the current load condition parameter, and the temperature range to which the current temperature belongs and stored as valid calibration parameters. If the calibration fails, the historical valid calibration parameters that match the current target absolute amplitude, current target pulse width, current load condition parameter, and current temperature are restored. If no matching historical valid calibration parameters exist, the output of positive and negative polarity pulses is prohibited, and the output node is kept at the output reference level.

[0013] Secondly, this application provides a programmable bipolar nanosecond pulse generation system, comprising: a pulse configuration interface for acquiring a target output polarity and a target pulse characteristic value, wherein the target pulse characteristic value includes a target absolute amplitude and a target pulse width; a bipolar pulse forming circuit, including a positive polarity output branch and a negative polarity output branch connected to the same output node, for outputting a positive polarity pulse or a negative polarity pulse through the output node according to an amplitude control quantity and a timing control quantity; a shared observation circuit connected to the output node for acquiring pulse sampling data through a signal conditioning and sampling link shared by positive and negative polarity pulses; a programmable sampling delay circuit for setting a relative sampling delay between a pulse trigger event and a sampling event; and a controller connected to the pulse configuration interface, the bipolar pulse forming circuit, the shared observation circuit, and the programmable sampling delay circuit, respectively, for: in a calibration state, controlling the programmable sampling delay circuit to set multiple relative sampling delays, and under each relative sampling delay, according to the target pulse characteristic value and the shared control quantity, to output a positive polarity pulse or a negative polarity pulse through the output node; a shared observation circuit connected to the output node for acquiring pulse sampling data through a signal conditioning and sampling link shared by positive and negative polarity pulses; a programmable sampling delay circuit for setting a relative sampling delay between a pulse trigger event and a sampling event; and a controller connected to the pulse configuration interface, the bipolar pulse forming circuit, the shared observation circuit, and the programmable sampling delay circuit for acquiring pulse sampling data through a signal conditioning and sampling link shared by positive and negative polarity pulses; a programmable sampling delay circuit for setting a relative sampling delay between a pulse trigger event and a sampling event; and a controller connected to the programmable sampling delay circuit for setting multiple relative sampling delays, and under each relative sampling delay, controlling the programmable sampling delay circuit to set multiple relative sampling delays according to the target pulse characteristic value and the shared control quantity, to acquire pulse sampling data through a signal conditioning and sampling link shared by positive and negative polarity pulses; a programmable sampling delay circuit for setting a relative sampling delay between a pulse trigger event and a sampling event; and a controller connected to the programmable sampling delay circuit for setting a relative sampling delay between a pulse trigger The system uses control parameters and polarity adjustment parameters to control the bipolar pulse forming circuit to output positive and negative polarity calibration pulses under the same load conditions, and controls the shared observation circuit to perform paired sampling of the positive and negative polarity calibration pulses. Based on the paired sampling data, the system determines the actual characteristic values ​​of the positive and negative polarities of the pulses. It also determines a common error based on the deviation between the average of the actual positive and negative polarity values ​​and the corresponding target value, and a polarity differential error based on the difference between the actual positive and negative polarity values. The system updates the common control parameters and polarity adjustment parameters based on the common error and the polarity differential error, respectively. The common control parameters and polarity adjustment parameters are used to adjust the corresponding control quantities of the positive and negative polarity output branches in the same and opposite directions, respectively. In the output state, the system determines the amplitude control quantity and the timing control quantity based on the target output polarity, the target pulse characteristic value, the updated common control parameters, and the polarity adjustment parameters, and controls the bipolar pulse forming circuit to output the target pulse.

[0014] Furthermore, the bipolar pulse forming circuit includes a programmable positive voltage supply branch, a programmable negative voltage supply branch, a positive high-speed switch, a negative high-speed switch, an active zero-return switch, and an output matching network; the positive high-speed switch is connected between the programmable positive voltage supply branch and the output node, the negative high-speed switch is connected between the programmable negative voltage supply branch and the output node, the active zero-return switch is connected between the output node and the output reference level, and the output matching network is connected between the output node and the load interface; the controller is used to set the programmable positive voltage supply branch according to the amplitude control quantity. The power supply parameters of the power supply branch and the programmable negative power supply branch are used to control the positive high-speed switch and the negative high-speed switch according to the timing control quantity, and a shoot-through prevention interval is set between the positive high-speed switch, the negative high-speed switch and the active zero-return switch. The controller is also used to determine the zero-return residual of the positive and negative output waveforms relative to the measured baseline before the corresponding pulse according to the sampling data before and after the positive and negative pulse output obtained by the shared observation circuit, and adjust at least one of the start time and conduction duration of the active zero-return switch according to the zero-return residual.

[0015] Furthermore, the shared observation circuit includes a broadband attenuation network, a broadband buffer, a bipolar-to-unipolar level conversion network, a sample-and-hold unit, and an analog-to-digital converter connected in sequence. The positive and negative pulses share the broadband attenuation network, the broadband buffer, the bipolar-to-unipolar level conversion network, the sample-and-hold unit, and the analog-to-digital converter. The controller is used to convert the sampled value output by the analog-to-digital converter into the voltage value corresponding to the output node according to the pre-calibrated observation mapping parameters. The programmable sampling delay circuit includes a coarse timing unit and a fine delay unit. The controller is used to coarsely scan the pulse waveform through the coarse timing unit to determine the pulse pre-baseline interval, effective level interval, rising edge interval, falling edge interval, and tail interval, and to finely scan at least the rising edge interval, the falling edge interval, and the tail interval through the fine delay unit.

[0016] Furthermore, in the calibration state, the positive polarity calibration pulse and the negative polarity calibration pulse are applied to the same load through the output node; the controller is used to keep the relative sampling delay constant under each relative sampling delay, control the bipolar pulse forming circuit to output calibration pulses alternately according to the polarity sequence of positive polarity, negative polarity, negative polarity, positive polarity, and control the shared observation circuit to sample multiple times for each polarity; the controller is also used to perform statistical processing on the positive polarity sample group and the negative polarity sample group respectively, and determine the corresponding statistical value as the effective paired sampling data under the current relative sampling delay if the sample dispersion does not exceed the dispersion threshold, the number of effective samples reaches the preset number, and the shared observation circuit is not saturated.

[0017] Furthermore, the pulse configuration interface is also used to acquire at least one of the repetition period, pulse number, polarity arrangement, and pulse interval; the system also includes a parameter memory and an operating condition acquisition interface respectively connected to the controller, the operating condition acquisition interface being used to acquire the current load condition parameters and the current temperature; the controller is used to associate and store the common control parameters and polarity adjustment parameters that have passed calibration verification with the target absolute amplitude, target pulse width, the current load condition parameters, and the temperature range to which the current temperature belongs in the parameter memory, and before outputting the target pulse, it calls the valid calibration parameters that match the current target pulse characteristic value, the current load condition parameters, and the current temperature; if the calibration verification fails, the controller restores the matching historical valid calibration parameters; if there are no matching historical valid calibration parameters, it controls the positive polarity output branch and the negative polarity output branch to stop outputting and generates a calibration failure flag.

[0018] Compared with existing technologies, the above technical solution has at least the following beneficial effects: By performing multi-delay paired sampling of positive and negative polarity nanosecond pulses under a unified measurement benchmark consisting of the same load, the same output node, and a shared observation channel, the measurement deviations introduced by load differences, channel differences, and sampling timing differences can be reduced. This improves the comparability of positive and negative polarity sampling data, as well as the accuracy of determining the actual absolute amplitude and actual pulse width. Furthermore, it reduces the dependence of nanosecond pulse characteristic measurements on continuous high-speed sampling capabilities. Based on this, by separately characterizing the overall deviation and polarity differences of positive and negative polarity pulses, and using a combination of in-direction adjustment and reverse adjustment for correction, the target value calibration and positive / negative polarity consistency calibration can be decoupled, reducing mutual interference between different calibration targets and control parameters. This balances the absolute output accuracy and positive / negative polarity symmetry of bipolar pulses, improving the accuracy and efficiency of calibration.

[0019] The above technical solution also has the following advantages: By interleaving output, repeated sampling, statistical processing, and sampling validity judgment, the impact of time drift, random noise, and abnormal sampling on calibration results can be reduced; by combining coarse and fine delay scanning, polarity normalization, and feature extraction based on waveform intervals, waveform reconstruction efficiency and edge positioning accuracy can be balanced; by performing sensitivity conversion, amplitude limiting, verification, and error degradation backoff on parameter updates, excessive parameter adjustment can be suppressed, improving the convergence and stability of iterative calibration; by detecting zero-return residuals and adjusting the active zero-return process, the output node can be prompted to recover to the output reference level in a timely manner, reducing the impact of baseline offset on subsequent pulses; by setting anti-snap-through intervals, output matching, and shared broadband observation links, the risk of branch mis-passing can be reduced, and the consistency of pulse transmission and observation can be improved; by storing, recalling, and backoffing effective calibration parameters according to pulse targets and operating conditions, and stopping output when there are no available parameters, the convenience of parameter reuse under different load and temperature conditions can be improved, reducing the risk of abnormal pulse output due to calibration failure or parameter mismatch, while meeting the programmable configuration requirements of different polarity arrangements, pulse numbers, and output timing. Attached Figure Description

[0020] The present application will be further described below with reference to the accompanying drawings and embodiments: Figure 1 A schematic diagram of the programmable bipolar nanosecond pulse generation system provided in this application; Figure 2 A schematic diagram of the bipolar pulse forming circuit framework provided in this application; Figure 3 A schematic diagram of the shared observation circuit and programmable sampling delay circuit provided in this application; Figure 4 This is a schematic diagram of the programmable bipolar nanosecond pulse calibration method provided in this application.

[0021] The system includes: 100, Programmable Bipolar Nanosecond Pulse Generator; 110, Pulse Configuration Interface; 120, Bipolar Pulse Forming Circuit; 121, Programmable Positive Voltage Supply Branch; 122, Programmable Negative Voltage Supply Branch; 123, Positive High-Speed ​​Switch; 124, Negative High-Speed ​​Switch; 125, Active Zero-Return Switch; 126, Output Matching Network; 127, Load Interface; 128, Output Node; 130, Shared Observation Circuit; 131, Wideband Attenuation Network; 132, Wideband Buffer; 133, Bipolar to Unipolar Level Conversion Network; 134, Sample and Hold Unit; 135, Analog-to-Digital Converter; 140, Programmable Sampling Delay Circuit; 141, Coarse Timing Unit; 142, Fine Delay Unit; 150, Controller; 160, Parameter Memory; 170, Operating Condition Acquisition Interface; 200, Load. Detailed Implementation

[0022] The technical solution of this application will be further described below with reference to the accompanying drawings. It should be understood that the following embodiments are used to illustrate the technical concept and implementability of this application, and are not intended to limit this application to the specific structures, parameters, sequences or application scenarios listed. Where there is no conflict, the technical features described in the various embodiments can be combined with each other.

[0023] In this application, unless otherwise expressly defined, the term "connection" can refer to both a direct electrical connection and an indirect electrical connection formed through an intermediate unit; the term "output" can refer to applying a working pulse to an external load or applying a calibration pulse for measurement in a calibration state. Positive and negative polarities are relative to the output reference level; positive pulses are higher than the output reference level, and negative pulses are lower than the output reference level.

[0024] The nanosecond pulse referred to in this application refers to a pulse with a pulse width on the order of nanoseconds or a pulse that requires control or measurement with nanosecond or sub-nanosecond timing resolution. The following explanation uses a voltage pulse as an example; with corresponding adjustments to the control quantity, observation mapping relationship, and load interface, the calibration concept of this application can also be applied to other high-speed pulse signals that can be formed by positive and negative polarity branches.

[0025] In a calibration task, the controller acquires the target output parameters and operating conditions. The target output parameters may include the target absolute amplitude. Target pulse width The target amplitude is defined as follows: pulse repetition frequency, target polarity, or polarity output sequence; operating conditions may include at least one of the following: external load identification, impedance range, output matching configuration, temperature or temperature range, power supply status, and operating mode. The target absolute amplitude is the absolute value of the pulse relative to its preceding measured baseline, not the algebraic value of positive or negative voltage. Therefore, positive and negative polarity pulses can be compared and calibrated using the same positive target absolute amplitude.

[0026] The output reference level is denoted as It is the active zero-return branch that attempts to restore the output node to a preset level. It can be zero potential or a bias level set according to the requirements of the subsequent interface. The output node levels actually measured by the shared observation circuit before the positive and negative polarity pulse outputs are denoted as follows: and . , The measured baseline may include bias, noise, and drift caused by changes in operating conditions of the observation link; therefore, it is not included. , Directly equivalent to .

[0027] For polarity , Belongs to positive polarity "Or negative polarity" "Shared observation circuit in relative sampling delay" The sampled data is obtained at the sampling point. Data obtained by positive and negative polarity pulses under the same relative sampling delay, the same physical load, the same output matching configuration, and comparable operating conditions are called paired sampled data. Pairing does not require that positive and negative polarity pulses be output simultaneously, but rather that the corresponding sampling conditions can be compared with each other.

[0028] This application acquires paired sampling data of positive and negative polarity pulses through a single output node and a shared observation link, reducing the risk of gain differences, bias differences, and delay differences between two independent observation links being mistaken for pulse polarity differences. Based on this, a shared error is formed according to the average deviation of the actual positive and negative polarity values, and a polarity differential error is formed according to the difference between the actual positive and negative polarity values. Then, shared control parameters and polarity adjustment parameters are adjusted separately. The shared control parameters cause the corresponding output characteristics of positive and negative polarities to change in the same direction, while the polarity adjustment parameters cause the corresponding output characteristics of positive and negative polarities to change in opposite directions.

[0029] The calibration state and the normal output state are distinct. In calibration state, the controller can output calibration pulses multiple times at a controlled repetition frequency to obtain discrete sampling points and reconstruct the equivalent waveform over multiple cycles. In normal output state, the controller outputs working pulses using verified and valid calibration parameters. If the current calibration fails and there are no recoverable historical valid calibration parameters, entering or continuing into the normal output state is prohibited.

[0030] Figure 1 A schematic diagram of the programmable bipolar nanosecond pulse generation system provided in this application. See also... Figure 1 This embodiment provides a programmable bipolar nanosecond pulse generation system 100. System 100 includes a pulse configuration interface 110, a bipolar pulse forming circuit 120, a shared observation circuit 130, a programmable sampling delay circuit 140, a controller 150, and a parameter memory 160. In some embodiments, system 100 further includes an operating condition acquisition interface 170. An external load 200 is located outside the boundary of system 100 and is connected to system 100 via a load interface.

[0031] The pulse configuration interface 110 is used to receive target output parameters and calibration trigger commands. The pulse configuration interface 110 can be a register interface, bus interface, communication interface, or a software configuration interface executed by the controller 150. The operating condition acquisition interface 170 is used to obtain current load conditions and environmental conditions. Load conditions can be represented by load identifier, nominal impedance, impedance range, output matching network configuration, or a combination thereof, and are not limited to DC resistance values. Temperature conditions can be obtained from a temperature detection unit located inside or outside the system 100, or provided by a higher-level control device.

[0032] The controller 150 provides the bipolar pulse forming circuit 120 with supply voltage control, polarity switching control, pulse width control, and active zero-return control based on the target output parameters, operating conditions, and initial calibration parameters. The bipolar pulse forming circuit 120 generates a positive or negative pulse at output node 128, and the generated pulse is applied to the external load 200 via output matching network 126 and load interface 127. The shared observation circuit 130 obtains the observation signals of the positive and negative pulses from the same output node 128 and returns the sampled data to the controller 150. The controller 150 calculates the calibration error and updates the calibration parameters based on the sampled data, thereby forming a closed-loop calibration path.

[0033] The parameter memory 160 is bidirectionally connected to the controller 150 and is used to store factory preset parameters, observation mapping parameters, parameters before update, historical valid calibration parameters, parameter application conditions, and calibration quality information. The parameter memory 160 may include volatile storage areas and non-volatile storage areas: the volatile storage area is used to temporarily store parameters before update in the current iteration, and the non-volatile storage area is used to store calibration parameters that have been verified to be valid and can be recalled later.

[0034] Initial calibration parameters can be determined according to preset priorities. For example, first, search for historically valid calibration parameters whose target amplitude, target pulse width, load conditions, and temperature conditions all match the current calibration task; if no perfectly matching historically valid calibration parameters exist, select safe parameters that are close to the target and allowed as a calibration starting point; if none exist, use factory-preset or design-preset safe initial parameters. Parameters with close matching conditions are only used as the starting point for this round of calibration and are not directly used as normal output parameters before re-verification.

[0035] The controller 150 can be implemented by a field-programmable gate array (FPGA), a microcontroller, a digital signal processor (DSP), a system-on-a-chip (SoC), a dedicated logic circuit, or a combination thereof. For example, the FPGA can be used to implement high-speed pulse timing, coarse and fine delay triggering, and sampled data acquisition, while the microcontroller can be used to implement statistical processing, error calculation, parameter management, and fault handling. The functional division of the above-mentioned devices or modules does not constitute a strict limitation on the embodiments of this application, as long as the corresponding control and calculation functions can be performed.

[0036] Figure 2 A schematic diagram of the bipolar pulse forming circuit framework provided in this application. See also... Figure 2 The bipolar pulse forming circuit 120 includes a programmable positive voltage supply branch 121, a programmable negative voltage supply branch 122, a positive high-speed switch 123, a negative high-speed switch 124, an active zero-return switch 125, an output matching network 126, a load interface 127, and an output node 128. The programmable positive voltage supply branch 121 is connected to the output node 128 via the positive high-speed switch 123, and the programmable negative voltage supply branch 122 is connected to the output node 128 via the negative high-speed switch 124. The positive and negative output branches are independent of each other and converge only at the output node 128. The active zero-return switch 125 is connected between the output node 128 and the output reference level V_REF. The output node 128 is connected to an external load 200 via the output matching network 126 and the load interface 127.

[0037] Programmable positive voltage supply branch 121 and programmable negative voltage supply branch 122 are used to generate the adjustable voltage required for positive and negative polarity pulses, respectively. The voltage supply control quantity can be a digital-to-analog conversion code, a programmable power supply configuration word, a switching power supply duty cycle, an adjustable reference value, or other control quantities that can change the output level of the corresponding branch. Positive high-speed switch 123 and negative high-speed switch 124 select the corresponding polarity branch under the timing control of controller 150, and form the pulse width by the start and end times of conduction.

[0038] When a positive polarity pulse is output, the positive polarity high-speed switch 123 is turned on, while the negative polarity high-speed switch 124 and the active zero-return switch 125 are turned off; when a negative polarity pulse is output, the negative polarity high-speed switch 124 is turned on, while the positive polarity high-speed switch 123 and the active zero-return switch 125 are turned off; when active zero-return is executed, both the positive polarity high-speed switch 123 and the negative polarity high-speed switch 124 are turned off, and the active zero-return switch 125 is turned on at a preset time. The positive polarity high-speed switch 123, the negative polarity high-speed switch 124, and the active zero-return switch 125 are not turned on simultaneously.

[0039] In one embodiment, after the positive polarity high-speed switch 123 or the negative polarity high-speed switch 124 is turned off, the controller 150 first maintains a preset anti-straight-through interval, and then turns on the active zero-return switch 125. The adjustable parameters of the active zero-return switch 125 include at least the active zero-return start time and the active zero-return conduction duration. The above parameters can be shared by both positive and negative polarities, or they can be combined with polarity adjustment parameters to form corresponding start times and conduction durations for positive and negative polarities respectively.

[0040] When the active zero-return switch 125 is turned on, the output node 128 approaches through the controlled low-resistance path. This is to reduce tail residue, baseline recovery time, or polarity-dependent memory effects after the pulse ends. Active zero-homing does not require the output node to be instantaneously and error-free equal to... The actual recovery effect is evaluated by the zero-recovery residual of the tail interval relative to the measured baseline before the pulse.

[0041] Output matching network 126 provides predetermined impedance matching, current limiting, waveform shaping, or transmission path adaptation between bipolar pulse forming circuit 120 and external load 200. Different output matching configurations can correspond to different parameter records. The sampling position of shared observation circuit 130 is located at output node 128, which is used for both positive and negative polarities, without setting separate observation nodes for the two polarities.

[0042] In the safe stop state, the controller 150 keeps the positive high-speed switch 123 and the negative high-speed switch 124 off, and maintains the output node 128 in a safe stop state via the active zero-return switch 125 or other safety holding path. nearby. Figure 2 The dashed lines in the diagram indicate that the controller 150 sends control signals to each branch, and do not indicate that there is an electrical series connection or mutual control relationship between the positive high-speed switch 123, the negative high-speed switch 124, and the active zero-return switch 125.

[0043] Figure 3 A schematic diagram of the shared observation circuit and programmable sampling delay circuit framework provided in this application. See also... Figure 3 The shared observation circuit 130 includes a broadband attenuation network 131, a broadband buffer 132, a bipolar-to-unipolar level conversion network 133, a sample-and-hold unit 134, and an analog-to-digital converter 135 connected in sequence. The broadband attenuation network 131 scales the pulse voltage at output node 128 to the allowable range of subsequent stages; the broadband buffer 132 provides impedance isolation to reduce the load effect of the observation branch on output node 128; the bipolar-to-unipolar level conversion network 133... The bipolar observation signal, which varies in amplitude, is converted to the input range allowed by the sample-and-hold unit 134 and the analog-to-digital converter 135; the sample-and-hold unit 134 holds the instantaneous level when the sampling trigger arrives; the analog-to-digital converter 135 converts the held level into digital sampled data and sends it to the controller 150. The above units constitute an observation link that uses both positive and negative polarities.

[0044] The code value output by analog-to-digital converter 135 cannot be directly equated to the voltage of output node 128 without processing. Controller 150 uses an observation mapping relationship to convert the statistically processed sampled data into the output node voltage. (Regarding polarity...) and relative sampling delay It can be expressed by the following formula: ; in, To account for the corresponding polarity in relative sampling delay The statistical measure obtained after statistical processing of the valid sampled data at the location; To observe the mapping relationship; These are the observation mapping parameters. The observation mapping parameters can be pre-calibrated by applying multiple known levels to the observation link. In one implementation, This is a linear mapping that includes gain and bias; in other implementations, It can be a piecewise linear function, a polynomial fitting relation, or a lookup table. The same mapping version is used for both positive and negative polarities, preventing the mapping error from being artificially split into two sets of polarity errors.

[0045] The programmable sampling delay circuit 140 includes a coarse timing unit 141 and a fine delay unit 142. The controller 150 provides the coarse timing unit 141 with a cycle start event and delay configuration. The coarse timing unit 141 generates a coarse delay event, and the fine delay unit 142 adds a programmable fine delay to the coarse delay event and outputs a sampling trigger to the sample-and-hold unit 134. The coarse timing unit 141 can be implemented by a counter, a timer, or a sequential state machine, and the fine delay unit 142 can be implemented by a programmable delay chain, a dedicated delay unit, a multi-phase clock selection circuit, or a delay structure capable of providing a delay step smaller than the coarse delay step.

[0046] To avoid misinterpreting the sampling before pulse triggering as generating a physical negative delay, this embodiment uses the start time of the calibration cycle. To standardize the time reference, let the pulse trigger time be... Sampling time is The relative sampling delay Defined as: ; when When less than 0, Still relative to the start of the period The non-negative delay, with the sampling point located before the pulse trigger, can be used to obtain the measured baseline before the pulse; when When the value is greater than 0, the sampling points sequentially cover the rising edge, effective level range, falling edge, and tail range of the pulse. The value of makes the minimum value to be used Not less than the minimum delay achievable by the programmable sampling delay circuit 140.

[0047] In one implementation, the controller 150 first performs a coarse scan according to a first delay step to identify the pre-pulse baseline interval, rising edge interval, effective level interval, falling edge interval, and tail interval; then, it performs a fine scan on at least the rising edge interval, the falling edge interval, and the tail interval according to a second delay step smaller than the first delay step. The fine scan window for the rising edge interval and the falling edge interval can be determined based on the amplitude change rate obtained from the coarse scan, the threshold cross position, or a preset edge window; the fine scan window for the tail interval can be determined based on the active zero-return start time, the tail residual change interval identified by the coarse scan, or a preset zero-return observation window.

[0048] Each relative sampling delay typically corresponds to one or more re-output calibration pulses. Controller 150 changes during multiple calibration cycles. The system collects discrete sampling points and then reconstructs the equivalent waveform according to the order of sampling delays. Therefore, the analog-to-digital converter 135 does not need to continuously acquire the complete waveform during the duration of a single nanosecond pulse, and its sampling rate can be lower than the equivalent sampling rate corresponding to the reconstructed waveform. However, the sample-and-hold unit 134 and its preceding observation link should have bandwidth and setup time adapted to the edge to be observed.

[0049] To reduce the slow drift of the observation link, in a preferred embodiment, for the same relative sampling delay... The calibration pulses are output and sampled in a "positive polarity—negative polarity—negative polarity—positive polarity" sequence, ensuring that the time intervals between adjacent calibration pulses are approximately consistent. Therefore, the time centers of two positive polarity samples are the same as or close to the time centers of two negative polarity samples. Each One or more of the above sampling blocks can be repeated, and then statistical processing can be performed on the positive and negative polarity samples respectively. Other sequences that make the positive and negative polarity sampling symmetrical or staggered in time can also be used.

[0050] The statistical measure can be the median, truncated mean, arithmetic mean after removing outliers, or robust fit result. If the analog-to-digital converter 135 experiences upper or lower limit saturation, the dispersion of the same group of samples exceeds a preset threshold, the number of valid samples is insufficient, or the pulse trigger validity check fails, the corresponding sampling point is marked as invalid. For invalid sampling points, the controller 150 can reduce the observation gain, adjust the attenuation configuration, re-output the calibration pulse, or terminate the current calibration without updating the normal output parameters with invalid data.

[0051] Figure 4 A schematic diagram of the programmable bipolar nanosecond pulse calibration method provided in this application. See also... Figure 4The calibration method in this embodiment can be executed by the controller 150 calling the bipolar pulse forming circuit 120, the shared observation circuit 130, the programmable sampling delay circuit 140, and the parameter memory 160. The following follows... Figure 4 Steps S101 to S112 are described below. The step numbers are used for the convenience of describing the embodiments of this application. Where the technical logic allows, some data checking, statistical processing, and parameter preparation actions can be performed in an overlapping manner.

[0052] S101: Obtain target output parameters, operating conditions, and initial calibration parameters.

[0053] Controller 150 acquires the target absolute amplitude Target pulse width The controller 150 obtains the pulse repetition frequency, output polarity or polarity sequence, and acquires the current load conditions, output matching configuration, and temperature conditions. The controller 150 also acquires common amplitude parameters. Polarity amplitude adjustment parameters Shared pulse width parameters Polarity timing adjustment parameters Active zeroing parameters, observation mapping parameters, allowable thresholds for each calibration index, upper limit for the number of iterations, and upper limit for the amount of update per single update.

[0054] Before outputting the calibration pulse, the controller 150 checks the parameter range, branch interlock status, analog-to-digital converter range, and load connection status. If the target parameter or initial parameter exceeds the hardware's allowable range, or if the interlock status of the positive and negative polarity high-speed switches cannot be confirmed, the calibration pulse is not output and a corresponding fault flag is generated. After the check, the controller 150 copies the initial parameter to the current parameter and the recoverable parameter before the update.

[0055] S102: Set multiple relative sampling delays and output positive and negative polarity calibration pulses alternately under the same load conditions.

[0056] The controller 150 determines the scanning range with relative sampling delay based on the target pulse width, preset edge margin, and tail observation duration, and configures the coarse and fine scan step sizes. The scanning range covers at least the baseline interval before the pulse, the effective pulse interval, and the tail interval after the pulse ends. For the same calibration task, positive and negative polarity calibration pulses are connected to the same physical load 200, and the load connection status, output matching configuration, and main operating conditions remain unchanged.

[0057] At each relative sampling delay, the controller 150 sequentially turns on the positive polarity high-speed switch 123 or the negative polarity high-speed switch 124, forming positive and negative polarity calibration pulses that are not output simultaneously. An interval sufficient to allow the output node 128 to recover and meet the switching safety requirements is set between adjacent pulses. A sampling block of "positive polarity—negative polarity—negative polarity—positive polarity" is preferably used; when operating conditions change rapidly, the duration of the sampling block can be shortened or the number of symmetrical samples within the same sampling block can be increased.

[0058] S103: Acquire and statistically process positive and negative polarity paired sampling data through the same output node and shared observation channel.

[0059] During each calibration pulse output, the programmable sampling delay circuit 140 generates a sampling trigger at the set relative sampling delay, the sample-and-hold unit 134 holds the corresponding instantaneous level, and the analog-to-digital converter 135 outputs digital sampled data. The controller 150 performs range checks, dispersion checks, trigger consistency checks, and outlier handling on the sampled data, and obtains the output node voltage using the observation mapping relationship.

[0060] Only valid positive and negative polarity data obtained under the same relative sampling delay, the same load conditions, and the corresponding target parameters constitute corresponding paired sampling data. If a certain polarity is in a certain... If a location lacks a sufficient number of valid samples, the paired data at that location is incomplete. Controller 150 prioritizes resampling that location instead of using one-sided data to calculate the polarity difference error. For security checks that do not rely on polarity comparisons, valid one-sided data can be used alone.

[0061] S104: Reconstruct the positive and negative polarity output waveforms to determine the actual absolute amplitude, actual pulse width, and zero-return residual.

[0062] The controller 150 sorts the positive and negative polarity sampling points according to the relative sampling delay, and can use linear interpolation, local fitting, or other reconstruction methods that preserve the edge positions to form equivalent positive and negative polarity waveforms. The reconstruction process does not require the generation of high-order details in the unsampled intervals; it only needs to ensure that the amplitude plateau, threshold crossover positions, and tail residuals can be stably determined.

[0063] Controller 150 first according to Valid sampling points less than 0 are used to determine the measured baselines of positive and negative polarity pulses, respectively. and . , The median or truncated average of the sampled values ​​within the pre-pulse baseline interval can be used. Then, polarity normalization is performed on the waveforms of both polarities: ; thus, and Both change in the positive direction within the corresponding pulse's effective range. The actual absolute amplitudes of the positive and negative polarities... , The effective level range can be determined by the median, truncated average, or fitted amplitude of the normalized waveform within the corresponding effective level range. The effective level range can be determined based on the target pulse width and edge position, or it can be determined within a continuous range where the normalized waveform exceeds a preset proportional threshold and the slope is less than a stable threshold.

[0064] For polarity According to the preset ratio Compared with the actual absolute amplitude The product of these is used as the pulse width measurement threshold. The normalized waveform is determined to cross over [a certain threshold] on the rising edge. The moment and crossing at the falling edge The moment Actual pulse width for: ; A threshold of 50% can be used, or other ratios can be adopted depending on the pulse width definition, interface specifications, or noise reduction requirements. The threshold crossover time can be obtained by linear interpolation of adjacent sampling points or by fitting a curve to the edge. The same threshold determination rule is used for positive and negative polarities. If the effective sampling points in the edge region are insufficient to determine the crossover time, the corresponding pulse width index is invalidated and a supplementary fine scan is triggered.

[0065] Controller 150 in the tail section after the pulse ends Internally calculate the deviation of the output node relative to the corresponding measured baseline. Zero-return residual. It can be represented as: ; in, This indicates a preset statistical operation. A value greater than 0 usually indicates that a residual pulse in the same direction still exists after the positive polarity pulse ends. A value less than 0 typically indicates that a residual pulse in the same direction still exists after the negative polarity pulse ends. Controller 150 can also calculate independently. and The baseline deviation is used to identify cases where the overall baseline has not recovered to the reference level. Baseline deviation and zero-return residual are different indicators; the former reflects the state before the pulse, while the latter reflects the tail recovery state after the pulse ends.

[0066] S105: Determine the common error and polarity difference error.

[0067] The amplitude errors for positive and negative polarities are defined as the differences between the target absolute amplitude and the actual absolute amplitude, respectively. ; The common amplitude error is determined by the average of the positive and negative polarity amplitude errors, and the polarity amplitude difference error is determined by the half difference between the two. ; in, It reflects the degree to which the overall absolute amplitude of positive and negative polarities is too high or too low. It reflects the relative inconsistency between the two polarities. The half-difference definition is used so that the common quantity and the difference component can correspond to the two polarities respectively after inverse transformation.

[0068] The pulse width errors of positive and negative polarities, as well as the corresponding shared pulse width errors and polarity pulse width differential errors, are determined in the same way: ; In other implementations, the average and half-difference of the actual positive and negative polarity values ​​can be calculated first, and then compared with the target value. This is equivalent to the error decomposition described above when the sign convention is consistent. To avoid ambiguity in the update direction, a fixed error sign convention is used within the same calibration task, and the sign of the sensitivity coefficient corresponds to this convention.

[0069] S106: Update common control parameters, polarity adjustment parameters, and active zero-return control parameters.

[0070] This embodiment at least sets a common amplitude parameter. Polarity amplitude adjustment parameters Shared pulse width parameters and polarity timing adjustment parameters The amplitude control values ​​for the positive and negative branches can be formed by the following formula: ; The pulse width or edge timing control values ​​of positive and negative polarities can be formed by the following formula: ; Therefore, change or When the corresponding control quantities of positive and negative polarities change in the same direction; changing or At that time, the corresponding control quantities for positive and negative polarities change in opposite directions. This parameterization method allows the overall deviation and polarity inconsistency to be borne by different control degrees of freedom, reducing the possibility of mutual drift or repeated compensation when only the two polarities are closed independently.

[0071] For amplitude or pulse width characteristics , belong or It can be based on the local sensitivity coefficient and Calculate the update amount of shared parameters and differential parameters in this round. For example: ; in, and The update coefficient is less than or equal to 1; This indicates the local sensitivity of changes in the shared control quantity to the corresponding average characteristic; This indicates the local sensitivity of the polarity adjustment amount to the corresponding differential feature; This indicates that the calculation results will be limited to a preset single-round update range. If the control parameter is a discrete configuration word, the result after limiting will be rounded or quantized to correspond to the achievable control step size.

[0072] The local sensitivity coefficient can be obtained by looking up or interpolating from the pre-calibration parameter table saved according to load conditions and temperature ranges, or by applying small positive or negative perturbations near the current operating point and estimating it based on the measurement results before and after the perturbation. If the absolute value of the sensitivity is lower than the preset lower limit, the sign is inconsistent with expectations, or the perturbation measurement fails the validity check, the corresponding coefficient is not used to directly update the parameters. Instead, the operating range is narrowed, the sensitivity is re-estimated, or the current calibration is terminated.

[0073] Let the current iteration number be... The parameter updates after limiting and quantization can be expressed as: ; After the update, controller 150 simultaneously checks , And the final formation , Whether it is within the allowable range of the corresponding hardware. If any final branch control quantity goes out of bounds, the corresponding update quantity is reduced first; if an effective control quantity cannot be formed within the safe range, the update round is determined to be unexecutable.

[0074] For active zero-return control, controller 150 according to... , The magnitude and direction of the active zero-return start time and conduction duration can be adjusted. For example, when the tail residual is still along the original pulse direction, the active zero-return start time can be appropriately advanced or the active zero-return conduction duration can be extended, provided that the anti-straight-through interval is met; when a reverse overshoot occurs at the tail, the active zero-return start time can be appropriately delayed or the active zero-return conduction duration can be shortened.

[0075] In another embodiment, the controller 150 generates multiple candidate combinations of "start-time - duration" around the current active homing parameters, outputs verification pulses for each, and selects the candidate combination with the smallest homing residual that does not worsen the amplitude, pulse width, or overshoot parameters. The active homing parameters can be updated after the amplitude and pulse width parameters have basically converged, or they can be updated alternately with the amplitude and pulse width parameters.

[0076] S107: Re-output the verification pulse and remeasure all calibration parameters.

[0077] After saving the parameters before the update, the controller 150 applies the candidate parameters for this round, re-outputs the positive and negative polarity verification pulses, and re-measures according to the sampling, statistics, and feature extraction rules consistent with the aforementioned measurements. , , , , and It can also measure baseline deviation, overshoot, recovery time, or other safety-related metrics. Even if only the amplitude parameter is updated in this round, the pulse width and homing-to-zero metric are remeasured to identify coupling effects between control variables.

[0078] When the absolute values ​​of the positive and negative polarity amplitude errors do not exceed the amplitude threshold, the absolute values ​​of the positive and negative polarity pulse width errors do not exceed the pulse width threshold, the absolute values ​​of the positive and negative polarity zero-return residuals do not exceed the zero-return threshold, and other indicators set as mandatory inspection items all meet their corresponding thresholds, all calibration indicators are deemed to meet the thresholds. The thresholds can be configured according to the target parameter range, load category, and temperature range.

[0079] S108: Associate stored valid calibration parameters.

[0080] If all calibration parameters meet the threshold, the controller 150 marks the current parameters as valid and stores them in association with the target output parameters and operating conditions. Each valid record may include the target absolute amplitude or applicable range, the target pulse width or applicable range, the load identifier or impedance range, the output matching configuration, and the temperature range. , , , Active zeroing start time, active zeroing conduction duration, observation mapping version, final calibration error, quality index, generation time, and valid flag.

[0081] Before writing a new valid record, a checksum, version number, or dual-copy consistency check can be performed. Only after the parameter record has been fully written and passed the consistency check is the system allowed to use the record to enter the normal output state. Failure results in this round do not overwrite existing valid historical records.

[0082] S109: Restore the parameters before the update if any calibration error worsens.

[0083] If the verification results show that the absolute value of any calibration error increases by more than the preset allowable jitter amount relative to the value before the update, or if at least one calibration index that did not exceed the corresponding calibration threshold before the update exceeds the corresponding calibration threshold after the update, then the corresponding calibration error is considered to have deteriorated. The controller 150 restores the parameters before this round of updates and records the update direction, update amount, and operating conditions that caused the deterioration. If the maximum number of iterations has not yet been reached, the iteration count can be reduced. , Alternatively, it sets the upper limit for a single update round and returns to S105 to recalculate the next update round using the current valid measurement results. Restoring parameters before the update does not change the already saved valid historical records.

[0084] S110: After reaching the upper limit of the number of iterations, find the matching historical valid calibration parameters.

[0085] If all calibration parameters have not yet met the thresholds and the maximum number of iterations has been reached, the controller 150 stops further exploratory updates and searches for matching historical valid calibration parameters in the parameter memory 160. Matching requires that the target amplitude, target pulse width, load conditions, and temperature conditions fall within the applicable range of the historical records, and that the output matching configuration and observation mapping version meet preset compatibility conditions. When multiple candidate records exist, records with smaller conditional distances, smaller final calibration errors, or higher quality levels can be prioritized.

[0086] S111: Restore matching historical valid calibration parameters.

[0087] If a matching historical valid calibration parameter exists, the controller 150 restores that parameter and puts the system into the "historical parameter restoration complete" state. Historical parameter restoration complete indicates that the system has been restored to a set of previously verified valid parameters, but does not indicate that this round of calibration was successful. Depending on the safety level and the degree of change in operating conditions, the controller 150 may perform a reduced verification process after restoration; if the verification passes, it will enter the normal output state; if the verification fails, it will proceed to S112.

[0088] S112: Stop pulse output and generate calibration failure flag.

[0089] If no matching valid historical calibration parameters exist, or if the restored historical parameters fail the required verification, the controller 150 keeps the positive high-speed switch 123 and the negative high-speed switch 124 off, and maintains the output node 128 at a certain position via the active zero-return switch 125 or other safety holding path. Nearby, normal pulse output is stopped, and a calibration failure flag is generated. The calibration failure flag can include the failure stage, invalid index, number of iterations, last valid parameter identifier, and current operating conditions, for the upper control equipment to handle the fault.

[0090] To illustrate the coordination between the above systems and methods, a set of exemplary calibration procedures is given below. The listed values ​​are only used to demonstrate the calculation relationships of parameter acquisition, error decomposition, parameter updating, and verification, and do not represent the actual measurement results of a specific product, nor do they constitute limitations on amplitude range, pulse width range, load type, sampling step size, or threshold.

[0091] In this example, the nominal impedance of the external load 200 is 50Ω, and the output matching configuration is denoted as... During calibration, the temperature is 25℃±1℃, and the output reference level is [not specified]. Set to 0V. Target absolute amplitude. 5.000V, target pulse width The pulse repetition frequency is 10.000 ns, and the calibration pulse repetition frequency is 100 kHz. The amplitude error threshold is set to 0.050 V, the pulse width error threshold is set to 0.080 ns, and the homing residual threshold is set to 0.025 V. The above thresholds are the configuration values ​​for this example.

[0092] Controller 150 first searches for initial parameters based on the target parameters, load identifier, output matching configuration, and temperature range. Parameter memory 160 does not contain a completely matching record, but it does contain a safe start parameter under the same load category and matching configuration; therefore, this safe start parameter is used as... , , , The initial values ​​of the active zero-return parameters are also loaded, along with the linear observation mapping parameters corresponding to the shared observation circuit 130. These safety start parameters are not considered normal output parameters until this round of verification is complete.

[0093] Starting time of a calibration cycle Based on, Set to 4.000ns, relative sampling delay The coarse scan range was set to -2.000 ns to 16.000 ns, and the first delay step was set to 0.500 ns. It can be seen that the earliest sampling event corresponds to =2.000ns, therefore the programmable sampling delay circuit 140 does not need to generate a negative physical delay. After the coarse scan identifies the rising edge interval, falling edge interval, and tail interval, a fine scan is performed with a second delay step of 0.050ns within a window before and after each edge, and within the residual change window identified by the coarse scan in the tail interval.

[0094] In each At this point, controller 150 continuously executes four "positive polarity-negative polarity-negative polarity-positive polarity" sampling blocks. Each sampling block provides two samples for both positive and negative polarities, so theoretically each polarity has eight samples. If a sample experiences analog-to-digital converter saturation or deviates from the group median by more than a preset discrete threshold, the sample is discarded. At least six valid samples are retained for each polarity; otherwise, the sample is resampled. The controller 150 uses a truncated mean to form the effective sample value. .

[0095] After observation mapping and waveform reconstruction, the pre-pulse baselines are as follows: =0.002V =0.001V. The actual absolute amplitude is determined by the median of the normalized waveform within the effective level range, resulting in... =4.820V =5.140V; with The threshold of 50% determines the edge crossing time, resulting in... =9.700ns =10.240ns; obtained within the tail interval =0.045V =-0.038V. and Residual particles remain along the corresponding original pulse direction.

[0096] Based on the target value and the actual value, the amplitude error is: ; Therefore, the shared amplitude error and the polarity amplitude difference error are: ; Pulse width error is =0.300ns =-0.240ns, the corresponding shared pulse width error and polarity pulse width differential error are: ; This example obtains the parameters from a table near the current work point. =0.005V / LSB =0.010V / LSB =0.010ns / LSB =0.010ns / LSB, and will , , and All values ​​are set to 0.5. The upper limit of a single-round update is sufficient to accommodate the following calculation results. After calculation and nearest-neighbor quantization, the following is obtained: =2LSB、 =8LSB =2LSB、 =14LSB. Controller 150 generates new positive and negative branch amplitude control values ​​and timing control values ​​based on this, and checks... None of them exceeded the limits allowed by the hardware.

[0097] because For positive and Negative values ​​for both indicate that the tail portion still lingers in the original pulse direction. Controller 150, while meeting the anti-shoo-through interval, advances the active zero-return start time (shared by both positive and negative polarities) by 0.100ns and extends the active zero-return conduction duration by 0.200ns. If the adjusted active zero-return parameters exceed the allowable range, only the update portion within the allowable range will be applied.

[0098] Controller 150 saves the parameters before the update, applies the above candidate parameters, re-outputs positive and negative polarity verification pulses, and re-measures all indicators. The verification result is as follows: =4.960V =5.030V =9.940ns =10.050ns =0.020V = -0.018V. Therefore, the absolute values ​​of the positive and negative polarity amplitude errors are 0.040V and 0.030V, respectively; the absolute values ​​of the pulse width errors are 0.060ns and 0.050ns, respectively; and the absolute values ​​of the zero-return residuals are 0.020V and 0.018V, respectively, all of which do not exceed the thresholds set in this example.

[0099] Controller 150 will currently , , , The active zero-return start time and conduction duration are marked as valid and are associated with it. =5.000V =10.000ns, 50Ω load type, output matching configuration The corresponding temperature range and observation mapping version are associated and stored. After the parameter record is written and the consistency check passes, the system 100% allows the use of the record to enter the normal output state.

[0100] If in another round of calibration, after the parameters are updated The amplitude error is reduced and If the pulse width error exceeds the corresponding error before the update and exceeds the allowable jitter, then the calibration error is still considered to have deteriorated. Controller 150 restores the parameters before the update for that round, reduces the corresponding update coefficient or the upper limit of the single-round update amount, and re-executes S105 to S107 if the upper limit of the number of iterations has not been reached. If the upper limit of the number of iterations has been reached, then S110 searches for historical valid parameters; if no matching record is found, then S112 stops the pulse output.

[0101] In some implementations, the absolute amplitudes of the positive and negative polarities can be different. In this case, they are set separately. and First, calculate the error of each polarity's actual value relative to its corresponding target value. Then, use the same averaging and half-difference relationships to form a common error and a polarity differential error. As long as the common parameter and the polarity adjustment parameter still perform the functions of same-direction and reverse-direction adjustment respectively, the above calibration framework can be used.

[0102] In some implementations, the target pulse does not have a stable plateau; the actual absolute amplitude can be determined by the peak value, the fitted amplitude within a preset time window, or the scaling factor relative to a standard template. The actual pulse width can be defined by the crossover time of two preset levels, the integral equivalent width, or other width definitions specified in the interface specification. Positive and negative polarities should use the same or corresponding feature extraction rules to maintain consistency in differential comparisons.

[0103] In some implementations, coarse scanning and fine scanning are not strictly divided into two sequential stages, but rather the delay step size is adaptively reduced based on the rate of change of adjacent sampling points. Equivalent waveform reconstruction can be achieved as long as multiple relative sampling delays cover the baseline, edges, effective intervals, and tail intervals, and valid paired sampling data for determining calibration parameters can be obtained.

[0104] In some implementations, each relative sampling delay uses a "negative polarity—positive polarity—positive polarity—negative polarity" or other time-symmetric order. Changing the sampling order does not change the fundamental relationships of simultaneous output of positive and negative polarities, sharing the same physical load, sharing the same output node, and sharing the observation link.

[0105] In some implementations, observation mapping parameters Each mapping version is saved separately depending on the attenuation level, temperature range, or observation link configuration. After switching mapping versions, the controller 150 writes the corresponding version identifier into the calibration record to avoid directly comparing sampled values ​​obtained under different mapping relationships. If the mapping self-test fails, the update of normal output parameters using that observation link will be stopped.

[0106] In some implementations, the amplitude, pulse width, and active zero-return parameters are updated in stages according to a preset order, such as calibrating the amplitude first, then the pulse width, and finally the active zero-return. In other implementations, candidate update values ​​for multiple parameters are jointly calculated based on the local coupling matrix. Regardless of the update method used, all mandatory indicators are remeasured after each round of candidate parameter application, and the parameters before the update are retained so that they can be restored if the error worsens.

[0107] In some implementations, the applicable conditions for historical valid parameters are represented by discrete labels; in other implementations, they are represented by target amplitude range, target pulse width range, load impedance range, and temperature range. If multiple historical records meet the range conditions, records can be selected for recovery based on normalized condition distance, calibration quality level, record freshness, or a combination thereof. Recovery of historical parameters does not overwrite records of current failure states.

[0108] In some implementations, the parameter memory 160, pulse configuration interface 110, and operating condition acquisition interface 170 can be integrated into the chip containing the controller 150; the bipolar pulse forming circuit 120 and the shared observation circuit 130 can be integrated into the same circuit board or the same package, or they can be implemented by multiple interconnected circuit modules. The physical integration of the modules does not change their functional relationship in the signal path, observation path, and control path.

[0109] The above embodiments illustrate the technical solution of this application. Those skilled in the art can adjust the specific parameters, hardware implementation, statistical methods, scanning strategies, and parameter update strategies without departing from the technical concept of this application. The above embodiments are only for illustrating the technical concept and features of this application, and their purpose is to enable those skilled in the art to understand the content of this application and implement it accordingly. They cannot be used to limit the scope of protection of this application. It is obvious to those skilled in the art that this application is not limited to the details of the above exemplary embodiments, and that this application can be implemented in other specific forms without departing from the spirit or basic characteristics of this application. Therefore, all embodiments should be considered exemplary and non-limiting, and all changes within the meaning and scope of the equivalent elements of the technical features of this application are included within this application.

Claims

1. A programmable bipolar nanosecond pulse calibration method, characterized in that, include: Obtain the target absolute amplitude and target pulse width; Configure multiple relative sampling delays between pulse trigger events and sampling events; For each relative sampling delay, based on the target absolute amplitude, the target pulse width, the shared control parameters, and the polarity adjustment parameters, the positive polarity output branch and the negative polarity output branch are controlled to output positive polarity calibration pulses and negative polarity calibration pulses respectively through the same output node under the same load conditions. The positive and negative polarity paired sampling data corresponding to the relative sampling delay are obtained by sampling through the observation channel shared by the positive and negative polarity calibration pulses. Based on the multiple relative sampling delays and their corresponding positive and negative polarity paired sampling data, the actual absolute amplitude and actual pulse width of the positive polarity calibration pulse and the negative polarity calibration pulse are determined respectively. For both absolute amplitude and pulse width, the common error is determined based on the deviation between the average of the actual values ​​of positive and negative polarities and the corresponding target values, and the polarity difference error is determined based on the difference between the actual values ​​of positive and negative polarities. Based on the common error and polarity differential error corresponding to the absolute amplitude, update the common amplitude parameter and polarity amplitude adjustment parameter respectively, and based on the common error and polarity differential error corresponding to the pulse width, update the common pulse width parameter and polarity timing adjustment parameter respectively. The shared control parameters include the shared amplitude parameter and the shared pulse width parameter, and the polarity adjustment parameters include the polarity amplitude adjustment parameter and the polarity timing adjustment parameter. The shared amplitude parameter is used to make the actual absolute amplitudes of the positive polarity calibration pulse and the negative polarity calibration pulse change in the same direction, and the polarity amplitude adjustment parameter is used to make their actual absolute amplitudes change in opposite directions. The shared pulse width parameter is used to make the actual pulse widths of the positive polarity calibration pulse and the negative polarity calibration pulse change in the same direction, and the polarity timing adjustment parameter is used to make their actual pulse widths change in opposite directions.

2. The programmable bipolar nanosecond pulse calibration method according to claim 1, characterized in that, The same load condition is formed by applying the positive polarity calibration pulse and the negative polarity calibration pulse to the same load; Under each relative sampling delay, the relative sampling delay is kept constant, and the positive polarity calibration pulse and the negative polarity calibration pulse are output alternately in the polarity order of positive polarity, negative polarity, negative polarity, positive polarity, and multiple samples are performed for each polarity to obtain positive polarity sample group and negative polarity sample group. Statistical processing is performed on the positive polarity sample group and the negative polarity sample group respectively, and the resulting statistical values ​​are determined as the positive polarity sampling data and the negative polarity sampling data under the relative sampling delay. Under the condition that the sample dispersion does not exceed the dispersion threshold, the number of effective samples reaches the preset number, and the observation channel is not saturated, the positive polarity sampling data and the negative polarity sampling data are determined as effective paired data.

3. The programmable bipolar nanosecond pulse calibration method according to claim 1, characterized in that, Setting the multiple relative sampling delays includes: A coarse scan is performed with the first delay step to determine the pre-pulse baseline interval, effective level interval, rising edge interval, falling edge interval, and tail interval after the pulse ends. A fine scan is performed on at least the rising edge interval, the falling edge interval, and the tail interval using a second delay step smaller than the first delay step. Determining the actual absolute amplitude and actual pulse width of the positive polarity calibration pulse and the negative polarity calibration pulse respectively includes: Based on the sampling data of the pre-pulse baseline interval, the measured baselines of the positive polarity calibration pulse and the negative polarity calibration pulse are determined respectively; The positive and negative output waveforms are reconstructed based on the sampling data obtained from coarse and fine scans, respectively. Based on the polarity relationship between the positive and negative calibration pulses and the output reference level, and using the measured baseline corresponding to each output waveform as the zero offset reference, the positive offset of the positive output waveform relative to the corresponding measured baseline and the negative offset of the negative output waveform relative to the corresponding measured baseline are mapped to the same coordinate direction to obtain the polarity-normalized positive and negative output waveforms. The corresponding actual absolute amplitude is determined based on the statistical values ​​of each output waveform after polarity normalization within the effective level range; The pulse width determination threshold is determined based on the corresponding actual absolute amplitude and the preset ratio, and the actual pulse width is determined based on the difference between the times when each output waveform after polarity normalization passes through the pulse width determination threshold in the rising edge interval and the falling edge interval.

4. The programmable bipolar nanosecond pulse calibration method according to claim 1, characterized in that, Updating the shared control parameters and the polarity adjustment parameters includes: For absolute amplitude and pulse width, the update amount of the common control parameter is determined according to the corresponding common error and the corresponding common control parameter sensitivity coefficient, and the update amount of the polarity adjustment parameter is determined according to the corresponding polarity difference error and the corresponding polarity adjustment parameter sensitivity coefficient. The single update amount and the updated parameter value are limited. By adopting the updated common control parameters and polarity adjustment parameters, positive polarity verification pulses and negative polarity verification pulses are output under the same load conditions, and the common error and polarity differential error corresponding to the absolute amplitude and pulse width are re-determined. Repeatedly update and verify the parameters until all errors meet the corresponding calibration threshold or the number of iterations reaches the preset upper limit; if the absolute value of at least one error increases by more than the preset allowable jitter amount compared to before the update, or if at least one calibration index that did not exceed the corresponding calibration threshold before the update exceeds the corresponding calibration threshold after the update, restore the parameters before the update.

5. The programmable bipolar nanosecond pulse calibration method according to claim 3, characterized in that, Obtain the current load condition parameters and the current temperature, wherein the current load condition parameters are used to characterize the same load conditions; The sampling positions corresponding to the multiple relative sampling delays also cover the tail intervals after the positive polarity calibration pulse and the negative polarity calibration pulse end; Based on the sampled data of the tail section, the zero-return residuals of the positive and negative polarity output waveforms relative to the measured baseline before the corresponding pulse are determined respectively, and the active zero-return control parameters for restoring the output node to the output reference level are updated based on the zero-return residuals. The active zero-return control parameters include at least one of the start time and duration of active zero-return. If the actual absolute amplitude, actual pulse width, and zero-return residual all meet the corresponding calibration pass conditions, the updated common amplitude parameter, polarity amplitude adjustment parameter, common pulse width parameter, polarity timing adjustment parameter, and active zero-return control parameter are associated with the target absolute amplitude, target pulse width, the current load condition parameter, and the temperature range to which the current temperature belongs and stored as valid calibration parameters. If calibration fails, restore historically valid calibration parameters that match the current target absolute amplitude, current target pulse width, current load condition parameters, and current temperature; if no matching historically valid calibration parameters exist, prohibit the output of positive and negative polarity pulses and keep the output node at the output reference level.

6. A programmable bipolar nanosecond pulse generation system, characterized in that, include: A pulse configuration interface is used to obtain the target output polarity and target pulse characteristic values, wherein the target pulse characteristic values ​​include the target absolute amplitude and the target pulse width; A bipolar pulse forming circuit includes a positive output branch and a negative output branch connected to the same output node, used to output a positive or negative pulse through the output node according to the amplitude control quantity and the timing control quantity. A shared observation circuit, connected to the output node, is used to acquire pulse sampling data through a signal conditioning and sampling link shared by positive and negative polarity pulses; A programmable sampling delay circuit is used to set the relative sampling delay between a pulse trigger event and a sampling event; as well as The controller, connected to the pulse configuration interface, the bipolar pulse forming circuit, the shared observation circuit, and the programmable sampling delay circuit respectively, is used for: In calibration mode, the programmable sampling delay circuit is controlled to set multiple relative sampling delays. Under each relative sampling delay, based on the target pulse characteristic value, common control parameters, and polarity adjustment parameters, the bipolar pulse forming circuit is controlled to output positive polarity calibration pulses and negative polarity calibration pulses respectively under the same load conditions. The shared observation circuit is also controlled to perform paired sampling of the positive polarity calibration pulses and the negative polarity calibration pulses. The actual pulse characteristic values ​​of positive and negative polarities are determined based on the data obtained from paired sampling. The common error is determined based on the deviation between the average value of the actual positive and negative polarities and the corresponding target value. The polarity differential error is determined based on the difference between the actual positive and negative polarities. The common control parameter and the polarity adjustment parameter are updated based on the common error and the polarity differential error, respectively. The common control parameter and the polarity adjustment parameter are used to adjust the corresponding control quantities of the positive polarity output branch and the negative polarity output branch in the same and opposite directions, respectively. In the output state, the amplitude control quantity and the timing control quantity are determined according to the target output polarity, the target pulse characteristic value, the updated common control parameters and polarity adjustment parameters, and the bipolar pulse forming circuit is controlled to output the target pulse.

7. The programmable bipolar nanosecond pulse generation system according to claim 6, characterized in that, The bipolar pulse forming circuit includes a programmable positive voltage supply branch, a programmable negative voltage supply branch, a positive high-speed switch, a negative high-speed switch, an active zero-return switch, and an output matching network. The positive polarity high-speed switch is connected between the programmable positive voltage supply branch and the output node; the negative polarity high-speed switch is connected between the programmable negative voltage supply branch and the output node; the active zero-return switch is connected between the output node and the output reference level; and the output matching network is connected between the output node and the load interface. The controller is used to set the supply parameters of the programmable positive supply branch and the programmable negative supply branch according to the amplitude control amount, control the positive high-speed switch and the negative high-speed switch according to the timing control amount, and set a first-break-then-connection anti-straight-through interval between the positive high-speed switch, the negative high-speed switch and the active zero-return switch. The controller is further configured to determine the zero-return residuals of the positive and negative polarity output waveforms relative to the measured baseline before the corresponding pulse based on the sampling data obtained by the shared observation circuit before and after the positive and negative polarity pulse outputs, and adjust at least one of the start time and conduction duration of the active zero-return switch based on the zero-return residuals.

8. The programmable bipolar nanosecond pulse generation system according to claim 6, characterized in that, The shared observation circuit includes a broadband attenuation network, a broadband buffer, a bipolar to unipolar level conversion network, a sample-and-hold unit, and an analog-to-digital converter connected in sequence. The positive and negative pulses share the broadband attenuation network, the broadband buffer, the bipolar to unipolar level conversion network, the sample-and-hold unit, and the analog-to-digital converter. The controller is used to convert the sampled value output by the analog-to-digital converter into the voltage value corresponding to the output node according to the pre-calibrated observation mapping parameters; The programmable sampling delay circuit includes a coarse timing unit and a fine delay unit. The controller is used to coarsely scan the pulse waveform through the coarse timing unit to determine the pulse pre-baseline interval, effective level interval, rising edge interval, falling edge interval, and tail interval, and to finely scan at least the rising edge interval, the falling edge interval, and the tail interval through the fine delay unit.

9. The programmable bipolar nanosecond pulse generation system according to claim 6, characterized in that, In calibration mode, the positive polarity calibration pulse and the negative polarity calibration pulse are applied to the same load through the output node; The controller is used to keep the relative sampling delay constant under each relative sampling delay, control the bipolar pulse forming circuit to output calibration pulses alternately in the polarity order of positive polarity, negative polarity, negative polarity, positive polarity, and control the shared observation circuit to sample multiple times for each polarity; The controller is also used to perform statistical processing on the positive polarity sample group and the negative polarity sample group respectively, and determine the corresponding statistical value as the effective paired sampling data under the current relative sampling delay if the sample dispersion does not exceed the dispersion threshold, the number of effective samples reaches the preset number, and the shared observation circuit is not saturated.

10. The programmable bipolar nanosecond pulse generation system according to any one of claims 6 to 9, characterized in that, The pulse configuration interface is also used to obtain at least one of the following: repetition period, number of pulses, polarity arrangement, and pulse interval; The system also includes a parameter storage and an operating condition acquisition interface connected to the controller, the operating condition acquisition interface being used to acquire current load condition parameters and current temperature; The controller is used to associate and store the common control parameters and polarity adjustment parameters that have been verified by calibration with the target absolute amplitude, target pulse width, the current load condition parameters and the temperature range to which the current temperature belongs in the parameter memory, and to call the effective calibration parameters that match the current target pulse characteristic value, the current load condition parameters and the current temperature before outputting the target pulse; If the calibration verification fails, the controller restores the matching historical valid calibration parameters; if no matching historical valid calibration parameters exist, the controller stops the output of the positive polarity output branch and the negative polarity output branch, and generates a calibration failure flag.

Images