A dual half-bridge blumlein type pfl nanosecond pulse generator

CN123034809APending Publication Date: 2026-11-03ANHUI UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611528394.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-30
Publication Date
2026-11-03

AI Technical Summary

Technical Problem

Existing Blumlein-type nanosecond pulse generators struggle to achieve bipolar output, fixed pulse width formation, and rapid port state recovery at high repetition frequencies, affecting the number of pulses and waveform consistency.

Method used

It adopts a dual half-bridge Blumlein type PFL structure, and generates bipolar pulses under the same main circuit topology by controlling the conduction sequence and relative conduction time of the four switches. It also achieves efficient utilization and fast switching through active recovery of port state and pulse width modulation.

Benefits of technology

It improves pulse formation efficiency and achieves stable output and waveform consistency of bipolar pulses at high repetition frequencies, making it suitable for high-speed device testing and SOT magnetic moment reversal excitation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123034809A_ABST
    Figure CN123034809A_ABST
Patent Text Reader

Abstract

The application discloses a double-half-bridge Blumlein type PFL nanosecond pulse generator and relates to the technical field of nanosecond pulse generators. A connecting point of a first upper bridge arm switch and a first lower bridge arm switch is connected to one end of a load through a first pulse forming line; a connecting point of a second upper bridge arm switch and a second lower bridge arm switch is connected to the other end of the load through a second pulse forming line; and two output pulses are formed on the load by controlling the conduction sequence and relative conduction time of the first upper bridge arm switch, the first lower bridge arm switch, the second upper bridge arm switch and the second lower bridge arm switch in one modulation period. The application can form two output pulses on the load by controlling the conduction sequence and relative conduction time of the four switches in one modulation period under the same main circuit topology.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of nanosecond pulse generator technology, and in particular to a dual half-bridge Blumlein type PFL nanosecond pulse generator. Background Technology

[0002] Nanosecond pulse sources, due to their rapid transient setup, flexible pulse width adjustment, and high repetition rate operation capabilities, have significant application value in electron gun gate control, fast pulse beam modulation, plasma excitation, advanced accelerator front-end devices, and dynamic testing of high-speed devices. In recent years, with the development of spintronic devices, magnetic moment reversal testing based on spin-orbit torque (SOT) has placed higher demands on nanosecond and even sub-nanosecond electrical pulse excitation. Existing research shows that SOT magnetic moment reversal can be achieved on timescales of 1 ns and sub-nanoseconds, and that pulse width, pulse shape, and polarity all affect the magnetic moment reversal process and energy efficiency. Therefore, how to achieve positive and negative pulse output in a compact structure, and further improve pulse width adjustment, pulse beam cutting, and high repetition rate burst operation capabilities, has become a key issue in nanosecond pulse source research.

[0003] Existing nanosecond pulse generation schemes mainly include avalanche transistor type, magnetic compression type, direct-drive switching type, and pulse-forming line (PFL) type technologies. Among them, the Blummlein-type pulse forming structure based on transmission lines can utilize the propagation and reflection of voltage waves in the transmission line to form a steep pulse leading edge and has good load impedance matching characteristics, thus becoming one of the important schemes for nanosecond pulse generation. With the improvement of the switching speed and repetition frequency operation capabilities of wide bandgap semiconductor devices such as gallium nitride (GaN), PFL or Blummlein-type nanosecond pulse sources based on fully controlled semiconductor switches show significant advantages in miniaturization, high repetition frequency operation, and waveform tunability.

[0004] However, existing PFL or Blumlein-type nanosecond pulse sources typically focus on unipolar output, fixed pulse width formation, or propagation under single-ended boundary conditions. A single triggering process often corresponds to one transmission line charge / discharge process and one main load pulse. After the pulse ends, the transmission line port state needs to be restored to the initial state required for the next pulse formation through the load, discharge path, or external recharging process. Under high-repetition-frequency burst operation conditions, the port state recovery time, repetitive charging process, and residual boundary response increase the minimum time interval required between adjacent load pulses and reduce the waveform consistency between consecutive load pulses, thus limiting the number of pulses that can be formed at the load end within a certain period. Summary of the Invention

[0005] In view of the defects of the existing technology, the present invention provides a dual half-bridge Blumlein type PFL nanosecond pulse generator, which solves the existing problems.

[0006] The present invention adopts the following technical solution: In a first aspect, the present invention provides a dual half-bridge Blumlein type PFL nanosecond pulse generator, comprising a first power supply circuit, a second power supply circuit, a first half-bridge circuit, a second half-bridge circuit, a first pulse forming line, a second pulse forming line, and a load; The first half-bridge circuit includes a first upper bridge arm switch and a first lower bridge arm switch connected in series between the positive and negative terminals of the power supply circuit. The connection point of the first upper bridge arm switch and the first lower bridge arm switch is the midpoint of the first half-bridge. The midpoint of the first half-bridge is connected to one end of the load through a first pulse forming line. The second half-bridge circuit includes a second upper bridge arm switch and a second lower bridge arm switch connected in series between the positive and negative terminals of the power supply circuit. The connection point of the second upper bridge arm switch and the second lower bridge arm switch is the midpoint of the second half-bridge. The midpoint of the second half-bridge is connected to the other end of the load through a second pulse forming line. Within one modulation cycle, by controlling the conduction sequence and relative conduction time of the first upper bridge arm switch, the first lower bridge arm switch, the second upper bridge arm switch, and the second lower bridge arm switch, two output pulses are generated on the load.

[0007] Preferably, when the first upper bridge arm switch is turned on before the second upper bridge arm switch, a positive pulse is generated on the load; when the second upper bridge arm switch is turned on before the first upper bridge arm switch, a negative pulse is generated on the load; when the second lower bridge arm switch is turned on before the first lower bridge arm switch, a positive pulse is generated on the load; when the first lower bridge arm switch is turned on before the second lower bridge arm switch, a negative pulse is generated on the load.

[0008] Preferably, when the midpoint of the first half-bridge transitions from a low potential to a high potential, the upper bridge arm switch is turned on; when the midpoint of the second half-bridge transitions from a low potential to a high potential, the lower bridge arm switch is turned on.

[0009] Preferably, within one modulation cycle, the polarity of the first output pulse is controlled by controlling the conduction sequence of the first upper bridge arm switch and the second upper bridge arm switch; the polarity of the second output pulse is controlled by controlling the conduction sequence of the first lower bridge arm switch and the second lower bridge arm switch.

[0010] Preferably, its output modes include positive-positive output, positive-negative output, negative-positive output, and negative-negative output.

[0011] Preferably, the pulse width of the output pulse is determined by the relative conduction delay of each switch in the first half-bridge circuit and the second half-bridge circuit.

[0012] Preferably, after the output pulse is established, the voltage across the load is brought back to zero in advance by controlling the relative conduction time of each switch in the first half-bridge circuit and the second half-bridge circuit. During the positive pulse output process, after the first upper bridge arm switch is turned on, the second upper bridge arm switch is turned on after a set delay, so that the potentials at both ends of the load are equal again, and the positive pulse is turned off in advance. During the negative polarity pulse output process, after the first lower bridge arm switch is turned on, the second lower bridge arm switch is turned on after a set delay, so that the potentials at both ends of the load are equal again, and the negative polarity pulse is turned off in advance.

[0013] Preferably, the first power supply circuit includes a first power supply, a first resistor, and a first capacitor, and the second power supply circuit includes a second power supply, a second resistor, and a second capacitor; the first resistor and the first capacitor are connected in series between the positive and negative terminals of the first power supply, and the second resistor and the second capacitor are connected in series between the positive and negative terminals of the second power supply.

[0014] Preferably, the first upper bridge arm switch, the first lower bridge arm switch, the second upper bridge arm switch, and the second lower bridge arm switch are all gallium nitride devices.

[0015] The above-mentioned at least one technical solution adopted in this invention can achieve the following beneficial effects: This invention, by setting two half-bridge circuits between the positive and negative terminals of the power supply circuit and connecting them to the two ends of the load via pulse forming lines, can generate two output pulses on the load within one modulation cycle under the same main circuit topology by controlling the conduction sequence and relative conduction time of four switches. This overcomes the technical defect of existing Blumlein-type pulse forming lines, which can only generate one load pulse during one charge and discharge cycle and require external discharge or recharging to restore the port state. Thus, it effectively improves the pulse utilization efficiency of the single pulse forming line in the charging state and provides favorable conditions for high repetition frequency burst operation. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a topology diagram of a dual half-bridge Blumlein type PFL nanosecond pulse generator according to the present invention; Figure 2 This is a schematic diagram of pulse generation methods corresponding to different triggering sequences under the dual half-bridge structure of the present invention; in, Figure 2 (a): Upper arm is conducting. Figure 2 (b): Lower bridge arm conduction; Figure 3 This is a schematic diagram of the driving timing and corresponding output waveform under the beam cutting control condition of the present invention; Figure 4 This is a schematic diagram of the beam-cutting control wave process of the present invention; in, Figure 4 (a): Positive pulse beam cutting process, Figure 4 (b): Negative pulse beam cutting process; Figure 5 This is the equivalent model of the path and the opposite boundary branch of the present invention; Figure 6 Stray inductors with different conduction paths according to the present invention Simulated waveforms under the given conditions; in, Figure 6 (a): Voltage at point C Figure 6 (b): Load voltage; Figure 7 Different non-conducting side junction capacitances of the present invention Simulated waveforms under the given conditions; in, Figure 7 (a): Voltage at point D, Figure 7 (b): Load voltage; Figure 8 For different branch stray inductors of the present invention Simulated waveforms under the given conditions; in, Figure 8 (a): Voltage at point D, Figure 8 (b): Load voltage; Figure 9 This is the experimental platform and prototype motherboard of the present invention; in, Figure 9 (a): Overall experimental platform Figure 9 (b): Prototype motherboard; Figure 10 These are the two consecutive positive pulse output waveforms of Embodiment 3 of the present invention; Figure 11 The negative and positive output waveforms are shown in Embodiment 3 of the present invention; Figure 12 These are the positive and negative output waveforms of Embodiment 3 of the present invention; Figure 13 The two consecutive negative pulse output waveforms are from Embodiment 3 of the present invention; Figure 14 A comparison of the output waveform under the influence of the parasitic parameters of the present invention and the waveform after beam cutting; Figure 15 The driving timing sequence of the four switches in this invention; Figure 16 This is the experimental waveform of the positive pulse width modulation of the present invention; Figure 17 This is the experimental waveform of negative pulse width modulation of the present invention; Figure 18 The output pulse waveforms of the present invention under different gate drive resistors; in, Figure 18 (a): Negative pulse output waveform, Figure 18 (b): Positive pulse output waveform; Figure 19 This invention provides bipolar alternating output within a 10μs time window. Figure 20 for Figure 19 The positive pulse amplification waveform; Figure 21 for Figure 19 The negative pulse amplification waveform. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] In actual hardware, GaN device output capacitance, parasitic inductance in switching circuits, stray capacitance on the PCB, and transmission line port impedance discontinuities all participate in nanosecond-level transient processes. These parasitic parameters not only cause pulse leading-edge overshoot, peak drop, and local oscillations, but also form residual pulse tail responses after pulse termination, affecting the formation and beam-cutting effect of narrow pulse widths. For bipolar nanosecond pulse outputs, positive and negative pulses typically require different switching sequences or energy paths, and their initial port states, boundary reflection conditions, and residual charge states are not entirely consistent, further increasing the difficulty of waveform consistency and pulse width adjustment.

[0020] Therefore, existing technologies still lack a high repetition frequency burst nanosecond pulse generation scheme that can simultaneously achieve configurable positive and negative polarity output, efficient utilization of PFL charging state, active recovery of port state, nanosecond-level pulse width adjustment, and pulse tail oscillation suppression under the same main circuit structure.

[0021] To address the shortcomings of traditional Blullein pulse forming lines (PFLs) in bipolar output, nanosecond-level pulse width modulation, and trailing edge control, this invention proposes a dual half-bridge Blullein PFL nanosecond pulse generator. This structure achieves output mode switching, pulse width modulation, and beam cutting control within the same topology by adjusting the switch on-time. Experiments show that the prototype can output alternating positive, negative, and bipolar pulses on the same hardware platform, with continuously adjustable pulse widths within 3ns to 10ns, and the beam cutting control effectively compresses pulse tails. Under a 200V DC voltage and a 50Ω load, the prototype stably outputs a peak current of 4A, providing a feasible solution for multi-mode nanosecond pulse modulation and trailing edge waveform control.

[0022] Configurable bipolar pulse sequences and port state recovery are achieved using two controlled voltage transitions within a single PFL charging state cycle. Secondly, a timing control method is established to achieve output mode selection, pulse width modulation, and pulse tail compression without altering the main circuit topology. Thirdly, a transient equivalent model related to switching states is established to distinguish between the excitation-side setup process and the high-impedance boundary response on the opposite side, and parasitic parameters are linked to leading-edge setup, flat-top hold, and pulse tail oscillation. Port state recovery refers to the recovery of the half-bridge port level to the initial low-level state required for the next operating cycle, and does not imply that all transient energy in the transmission line is completely dissipated within a finite time. Since the left and right half-bridge structures are completely symmetrical, subsequent analyses do not fixate different physical models on the left or right side.

[0023] This invention proposes a Blumein-type bipolar nanosecond pulse generator based on a dual half-bridge structure. This structure achieves output mode switching, pulse width modulation, and beam cutting control by adjusting the conduction sequence and relative conduction time of the same-polarity switches on the left and right sides. Compared with existing schemes that rely on single-sided excitation, fixed terminal boundaries, or external polarity switching stages, the proposed structure can complete positive and negative pulse formation and trailing edge waveform adjustment within the same main circuit topology. This provides an adjustable nanosecond pulse source solution for applications such as high-speed device testing, SOT magnetic moment reversal excitation, and fast pulse modulation at the accelerator front end. In terms of mechanism analysis, this invention attributes the changes in the output waveform to two parts: the active conduction path and the opposite boundary branch. It establishes the correspondence between relevant circuit parameters and waveform characteristics, thereby explaining the differences in waveform evolution during the leading edge establishment, top hold, and pulse tail oscillation regions. It also explains the mechanism by which beam cutting control compresses the effective pulse interval by adjusting the relative conduction time of the same-polarity switches.

[0024] Figure 1 The proposed dual half-bridge Blumlein-type PFL structure is presented. This circuit consists of two input power supplies. Bus capacitor and Current limiting resistor and Left half bridge and Right half bridge and Two pulse forming lines and and central load constitute.

[0025] The left half-bridge is the first half-bridge, including the first upper bridge arm switch connected in series. and the first lower bridge arm switch The right half-bridge is the second half-bridge, including the second upper bridge arm switch connected in series. Second lower bridge arm switch The two pulse forming lines include the first pulse forming line. Second pulse forming line The two input power supplies include a first power supply and a second power supply, along with a current-limiting resistor. and bus capacitor A current-limiting resistor connected in series between the positive and negative terminals of the first power supply. and bus capacitor It is connected in series between the positive and negative terminals of the second power source.

[0026] The midpoint of the left half-bridge switch is denoted as the midpoint of the first half-bridge, and its voltage is... The midpoint of the right half-bridge switch is denoted as the midpoint of the second half-bridge, and its voltage is... The midpoint of the first half-bridge passes through the line formed by the first pulse. Connected to the left end of the load, the midpoint of the second half-bridge is connected to the line formed by the second pulse. Connected to the right end of the load, the two pulse forming lines together with the central load form a double-ended pulse forming structure.

[0027] Compared to the traditional Blumein-type structure controlled by dual-ended switches, the topology of this invention retains half-bridge switching stages at both ends of the pulse forming line and places the load between the two pulse forming lines. Therefore, the conduction combination and relative timing of the left and right half-bridges directly determine the load-side voltage establishment method, thus enabling multiple output modes under the same main circuit topology. Furthermore, by adjusting the relative conduction times of the left and right switches of the same polarity, pulse width modulation and beam cutting control can also be achieved.

[0028] One-way propagation delay of pulse forming line Recorded as: (1); in, The length of the pulse forming line. Let be the propagation speed of the wave in the pulse forming line.

[0029] Figure 2 The pulse generation methods corresponding to different triggering sequences in a dual half-bridge structure are presented. The voltages at the midpoints of the left and right half-bridges are also given. and A transition occurs after the corresponding switch is turned on, and occurs along the pulse forming line respectively. and The voltage waves propagate towards both ends of the load. When the voltage waves from both sides arrive at the load simultaneously, it is difficult to establish an effective differential pulse voltage across the load; when there is a time difference in arrival between the voltage waves from both sides, a pulse voltage with a limited width is formed at the load. V L Therefore, the formation of a single load pulse in a dual half-bridge structure depends on the order of voltage transitions on the left and right sides and their relative trigger delays.

[0030] For the Each output pulse records the trigger times of the corresponding switches on the left and right sides as follows: and Then its relative trigger delay is: (2); When the two pulse forming lines have the same length and the propagation conditions are approximately the same... This can be used to approximately characterize the time interval between the voltage waves on both sides reaching the load end. When When the voltage waves on both sides act synchronously on the load end, it is difficult to establish a pulse voltage of finite width at the load end; when At this time, the voltage waves on both sides arrive at the load end at staggered intervals, thus forming a load pulse with a limited width. Therefore, it can be seen that the width of a single load pulse in a dual-half-bridge structure is mainly determined by the relative trigger delay. Decide.

[0031] The polarity of the load pulse is determined by the order in which the voltage waves from both sides arrive at the load end and the transition direction of the midpoint voltage. During the conduction process of the upper bridge arm, the midpoint voltage transitions from a low potential to a high potential; during the conduction process of the lower bridge arm, the midpoint voltage transitions from a high potential to a low potential. Therefore, when the left upper bridge arm switches... Pilot-on, right upper bridge arm switch During delayed conduction, a positive pulse is generated at the load end; if the triggering sequence is reversed, a negative pulse is generated. For the lower arm conduction process, the polarity relationship corresponds to that of the upper arm conduction process: when... pilot conduction, A positive pulse is generated when the delay is applied; when pilot conduction, A negative pulse is generated upon delayed conduction. This includes the first upper bridge arm switch connected in series. and the first lower bridge arm switch The right half-bridge is the second half-bridge, including the second upper bridge arm switch connected in series. Second lower bridge arm switch .

[0032] Table 1. Multimode output modes within one modulation cycle

[0033] Based on this, the output mode within a complete modulation cycle is jointly determined by the conduction processes of the upper and lower bridge arms, and representative multimode output modes are shown in Table 1. As can be seen from Table 1, the conduction sequence of the upper bridge arm determines the polarity of the first pulse, and the conduction sequence of the lower bridge arm determines the polarity of the second pulse. Therefore, the dual half-bridge structure can be implemented under the same circuit topology. , , and Four output modes.

[0034] The dual half-bridge ports can actively apply high and low level transitions within a controlled PFL charging state cycle. When the upper bridge arm moves, the corresponding port transitions from low to high, putting the PFL into a high-level charging state and participating in the formation of the first effective load pulse. When the lower bridge arm moves, the corresponding port transitions from high to low, using the established port voltage to form the second effective load pulse, while simultaneously restoring the half-bridge port to the initial low-level state required for the next duty cycle.

[0035] Therefore, the dual half-bridge PFL is not simply about increasing the number of switches, but rather about extending the PFL port from a single-trigger boundary to a voltage boundary capable of controlled charging and discharging. This structure can generate two effective pulses using two controlled voltage transitions after a single charging state is established, thereby improving the utilization of effective pulses during each charging recovery process and reducing the dependence on external charging recovery speed and passive discharge process during high-repetition-frequency burst output.

[0036] The propagation time of the fixed-length pulse forming line determines the natural pulse width of the dual-half-bridge Blumein-type PFL structure. Therefore, under nanosecond-level high-frequency operation, pulse width modulation mainly depends on the relative timing between the left and right switches of the same polarity. Meanwhile, influenced by the switching speed of GaN devices, interconnect parasitic parameters, and the consistency of the loops on both sides, reflections and waveform evolution after the falling edge are more likely to occur under high-frequency pulse conditions. Therefore, by adjusting the relative on-time delay of the switches of the same polarity on both sides, not only can pulse width modulation be achieved, but the waveform after the pulse falling edge can also be actively compressed. When this delay is further reduced, the corresponding post-pulse waveform will be compressed earlier, thus forming beam-cutting control.

[0037] Reference Figure 3 and Figure 4 For ease of description, the midpoint voltages of the left and right half-bridges are denoted as follows: and The voltages at the left and right ends of the load are denoted as follows: and , U gs Since this is a switch drive signal, the load voltage can be written as: = - (3); Positive pulse beam cutting process.

[0038] when At that time, the left upper bridge arm switch First, connect the node. The potential transitions from 0 to And transmit an amplitude of to the right transmission line The voltage wave. At this time, the system enters the positive pulse establishment process, while the upper bridge arm switch on the opposite side... It remains off.

[0039] StageI : for At the time of conduction, Let A be the moment when the voltage wave at point A reaches point C. During this stage, the voltage wave at point A is generated by the node. Launched The voltage wave propagates along the transmission line towards the load. Since the wavefront has not yet reached the load location, the node... ,node and nodes The potentials of all components remain unchanged, so the load voltage has not yet been established, and the system is in the initial propagation stage of positive pulse formation.

[0040] Stage II : When the voltage wave at point C reaches the load, After the voltage wave reaches the vicinity of the load, the node ,node The first voltage redistribution occurs between the node and the load. The potential rises to ,node The potential rises to Therefore, an amplitude value is established at both ends of the load. The differential-mode voltage indicates that a positive pulse has been formed across the load. Meanwhile, the amplitude is... The return wave is from the node To the node Propagation, and amplitude The forward wave is formed by the node To the node Continue to spread.

[0041] During this phase, after the delay time Then, switch on the opposite upper bridge arm. Conductivity, node The potential transitions from 0 to And by node To the node Launch a new amplitude The voltage wave. As a result, the opposite boundary changes from its original natural open-circuit state to a controlled high-potential boundary, and begins to reconstruct the original subsequent evolution process.

[0042] Stage III : The moment when the voltage wave from point B reaches node D is during this phase, determined by node D. issued The voltage wave arrives near the load and interacts with the previous stage caused by the node. The established wave processes act together at the nodes. and nodes .according to Figure 4 The wave process shown has nodes. With nodes The potential rises simultaneously to Since the potentials across the load terminals become equal again, the load voltage quickly returns to zero, thus the positive pulse is turned off prematurely, and the pulse width is shortened accordingly.

[0043] Stage IV : When the voltage wave at point D arrives at the load, during this stage, the node... With nodes Both sides maintain a high-potential conducting state, and the transmission lines on both sides are in a stable state under the combined effect of controlled boundaries. At this time, the node With nodes The potential remains consistent, so the load voltage remains zero. In other words, the system no longer continues to evolve along the natural propagation path of the later stage, but remains in a controlled cutoff state after the beam is cut off.

[0044] Negative pulse beam cutting process.

[0045] when At that time, the left lower bridge arm switch Conductivity, node The potential is from Quickly pull it down to 0, and transmit an amplitude of [value] to the right transmission line. The voltage wave. At this time, the system enters the negative pulse establishment process, while the lower bridge arm switch on the opposite side... It remains off. Since the positive pulse beam cutting process has ended, therefore... At the initial moment, node ,node and nodes The potentials are all maintained .

[0046] StageI During this stage, the nodes Launched The voltage wave propagates along the transmission line towards the load. Since the wavefront has not yet reached the load location, the node... ,node and nodes The potential remains unchanged, so the load voltage has not yet been established, and the system is in the initial propagation stage of negative pulse formation.

[0047] Stage II :when After the voltage wave reaches the vicinity of the load, the node ,node A new voltage redistribution occurs between the node and the load. The potential drops to ,node The potential drops to Therefore, an amplitude value is established at both ends of the load. The differential-mode voltage indicates that a negative pulse has been formed across the load. Meanwhile, the amplitude is... The return wave is from the node To the node Propagation, and amplitude The forward wave is formed by the node To the node Continue to spread.

[0048] During this phase, after the delay time Then, the switch on the opposite lower bridge arm. Conductivity, node The potential is from It quickly dropped to 0, and was controlled by the node. To the node Launch a new amplitude The voltage wave. As a result, the opposite boundary switches from the original high-potential state to a controlled low-potential boundary, and begins to reconstruct the original subsequent evolution process.

[0049] Stage III During this stage, the nodes issued The voltage wave arrives near the load and interacts with the previous stage caused by the node. The established wave processes act together at the nodes. and nodes .according to Figure 4 The wave process shown has nodes. The potential remains And nodes The potential is from Further decline to Since the potentials across the load terminals become equal again, the load voltage quickly returns to zero, thus the negative pulse is turned off prematurely, and the pulse width is shortened accordingly.

[0050] Stage IV During this stage, the node With nodes Both remain in a low-potential conducting state, and the transmission lines on both sides are in a stable release process under the combined effect of controlled boundaries. At this time, the node With nodes The potentials continued to decrease simultaneously, and eventually all reached [a certain value]. Since the load terminals remain at the same potential, the load voltage remains zero. In other words, the system no longer evolves along the natural propagation path of the later stages, but remains in a controlled cutoff state after the beam is cut.

[0051] As the above analysis shows, the main difference between the beam-cutting control condition and the natural positive and negative pulse modes lies not in the establishment of the pulse leading edge, but in the boundary evolution path of the pulse's latter segment. In the natural mode, the latter segment of the pulse is mainly determined by the superposition of the propagating wave, boundary reflection, and their return. However, in the beam-cutting control condition, by delaying the conduction of the opposite polarity switch after the active side switch is turned on, the opposite boundary is switched to a controlled state in advance. This compresses the latter segment evolution, which originally required a complete two-way propagation process, into a shorter controlled process, ultimately achieving pulse width adjustment and suppression of the latter segment fluctuations.

[0052] Example 1 Modeling of parasitic networks in power loops and waveform formation mechanism.

[0053] The output waveform of a dual half-bridge Blummlein pulse forming line (PFL) topology is determined by both the switching timing and the propagation process of the transmission line voltage wave. However, in actual hardware, switching commutation transients can excite parasitic networks in the power loop, causing the load-side waveform to deviate from the ideal rectangle. For this invention, the most significant transient distortions are leading-edge overshoot, top drop, and pulse tail oscillation. Therefore, this invention provides a unified analysis of the output waveform formation mechanism from three aspects: source-side dynamic impedance mismatch, stray inductance constraints in the conduction path, and energy transients at the high-impedance terminal on the passive side.

[0054] Within the nanosecond-level transient state established at the pulse leading and trailing edges, extremely high (Rate of change of current over time) and The rate of change of voltage over time gives GaN HEMTs a distinct dynamic characteristic, preventing them from being considered ideal switches. GaN stands for Gallium Nitride, and HEMT stands for High Electron Mobility Transistor. (See reference...) Figure 5 During the turn-on transient, the active-side switch can be equivalent to a time-varying channel resistor. and stray inductance of the active power circuit Series connection; simultaneously, the non-conducting side junction capacitance Connected in parallel to the input of the transmission line. Therefore, the complex frequency domain equivalent input impedance at the input boundary... It can be written as: (4); in, This is the characteristic impedance of the transmission line. During the commutation transient, the inductive reactance term... Instantaneous increase, causing Deviation , s Given a complex frequency. According to transmission line theory, the dynamic reflection coefficient at the source boundary is: (5); when At this time, dynamic reflection will occur at the source boundary, causing distortion of the leading edge waveform. It can be seen that the establishment of the leading edge in the dual half-bridge structure is not an ideal single voltage transition, but the result of the incident wave, reflected wave and parasitic parameters being jointly shaped under dynamic boundary conditions.

[0055] After the active-side switch is turned on, the step voltage wave propagates along the transmission line and is reflected and transmitted at the impedance discontinuity boundary. The node voltage inside the transmission line is determined by the superposition of each stage of reflection components. This leading-edge commutator circuit can be approximately equivalent to a second-order series RLC system, and its complex frequency domain output voltage at the load end satisfies: (6); in, , (7); In the formula, This is the total equivalent capacitance at the PFL input terminal. For the equivalent damping of the loop, This is the DC bus voltage. The frequency of undamped natural oscillation. The damping ratio is determined by changing the gate drive resistance. This directly alters the turn-on speed of the GaN HEMT, thereby adjusting the loop damping. When damping is weak, the high-frequency components of the leading edge are strong, easily triggering local parasitic network oscillations, manifesting as load leading-edge overshoot and voltage oscillations at nodes. As damping increases, overshoot is suppressed, but the rise time decreases and the effective flat-top width is compressed. Therefore, under extremely narrow pulse width conditions, damping should not be excessively high; instead, a slightly underdamped state should be maintained, allowing for controllable local overshoot while achieving a higher leading-edge build-up speed and a more effective flat-top width.

[0056] After entering the flat-top sustaining phase, the high-frequency oscillation decays, the switching switch enters the low-resistance conduction region, and the junction capacitance of the active-side cutoff transistor no longer generates significant dynamic current shunting. At this time, the power circuit voltage drop is mainly due to the series stray inductance in the conduction path. The dominant factor, the resulting transient pressure drop, can be expressed as: (8); In the formula, This refers to the current flowing through the conduction path. Because the output pulse width is extremely narrow, the setup time for the transition from the leading edge to the flat top is on the same order of magnitude as the pulse width. If the inductive voltage drop has not fully converged during turn-on, it will directly cause a drop in the flat top voltage. Therefore, reducing the current in the conduction path... It helps improve top retention and inhibit flat-top degradation.

[0057] During the falling edge of the pulse, the passive-side switch is in a high-impedance cutoff state and cannot form a fixed potential clamp. When the voltage wave reaches this high-impedance terminal, due to impedance mismatch, the energy cannot be quickly dissipated but instead transfers to the passive-side local capacitance. In, and with stray inductance Together they form a local resonant network, whose equivalent complex impedance is: (9); The voltage wave reflects back and forth between the mismatched termination and the transmission line, resulting in a continuous pulse tail oscillation with the following characteristic period: (10); As can be seen from the formula, and Together, they determine the time scale of local oscillations. Among them, the larger ones... It will lengthen the interval between peaks and troughs, making the oscillation unfold more slowly; larger ones This means that more local energy storage participates in redistribution, making the pulse tail disturbance amplitude more pronounced.

[0058] To suppress pulse tail oscillations caused by this local resonance, this invention employs an active pulse tail compression strategy, which involves adjusting the conduction timing difference between the two identical polarity switches. A voltage wave of opposite polarity is injected at the boundary. By utilizing the destructive interference of the two waveforms at the load end, the effective pulse tail interval can be compressed and local oscillations can be reduced. In other words, beam cutting control is not only a timing modulation method, but also an active compensation method for the energy transients of the high-impedance terminal on the passive side.

[0059] In summary, leading-edge overshoot, top drop, and pulse tail oscillation correspond to three mechanisms: dynamic impedance mismatch at the source, stray inductance constraint in the conduction path, and transient energy at the high-resistivity terminal on the passive side, respectively. Thus, the relationship between "high repetition rate achievement capability" and "parasitic waveform degradation" in the dual-half-bridge Blumein type PFL structure is uniformly explained. Further analysis will be conducted on... , and The mechanism of action was verified by simulation.

[0060] Example 2 To verify the power loop parasitic network model and transient waveform formation mechanism established in Example 1, this invention performs time-domain simulations of the proposed dual half-bridge Blumlein-type PFL pulse generation structure. The stray inductance of the conduction path is varied during the simulation. Non-conducting side junction capacitance and branch stray inductance The changes in the voltage waveforms of key nodes and load voltage can be observed. Through parameter scanning, the influence of parasitic parameters of the conduction path on the establishment of the leading edge and the maintenance of the top, as well as the effect of the parasitic network of the high-resistivity boundary on the pulse tail oscillation, can be analyzed.

[0061] Parasitic inductance Impact: With all other parameters remaining unchanged Figure 6 Different The waveforms of the voltage at point C and the load voltage are shown. Point C is located near the active side conduction path, and its voltage change reflects the influence of the stray inductance of the conduction path on the leading edge establishment process; the load voltage characterizes the effect of this parameter on the leading edge, top hold, and trailing edge characteristics of the output waveform.

[0062] Depend on Figure 6 As can be seen from (a), with As the voltage increases, the transient fluctuations at point C become more pronounced during the leading-edge setup phase. This indicates that stray inductance in the conduction path directly participates in the nanosecond-level commutation process. A larger... In high Under these conditions, a more pronounced transient voltage drop occurs, causing stronger oscillations and shifts in the node voltage during the establishment process.

[0063] Figure 6 (b) indicates that different Under these conditions, there is a significant difference between the establishment and maintenance of the load voltage leading edge. With... As the pulse width increases, the load-side holding capability decreases, and the waveform broadening and residual oscillations near the falling edge become more pronounced. This indicates that under extremely narrow pulse width conditions, the inductive reactance drop in the conduction path has not yet fully converged during the pulse sustaining phase, directly weakening the load-side holding capability.

[0064] therefore, It is an important parasitic parameter affecting the rise-lead-out and top-hold capability. Reducing the series stray inductance in the active-side conduction path helps to reduce transient voltage drop, improve the rise-lead-out speed at the load end, and improve the top-hold capability.

[0065] Non-conducting side junction capacitance Impact: With all other parameters remaining unchanged Figure 7 Different The waveforms of the voltage at point D and the load voltage are shown. Point D is close to the high-impedance boundary on the opposite side, and its voltage change reflects the influence of the junction capacitance of the cut-off switch on the local transient process; the load voltage characterizes the effect of this local oscillation on the output pulse tail characteristics.

[0066] Depend on Figure 7 As can be seen from (a), with As the voltage increases, both the amplitude and duration of the voltage oscillation at point D in the pulse tail region increase significantly. This is because the opposite switching transistor is in the off state during the pulse, and there is a lack of low-resistance potential clamping near point D. When the voltage wave propagates to this high-resistance boundary, some energy cannot be quickly discharged but is transferred to the junction capacitance of the off-state device. In the middle, it causes a partial charging and discharging process. Larger... The transient energy storage capacity at the boundary is improved, making the oscillation at the trailing edge of point D more pronounced.

[0067] Figure 7 (b) indicates that different Under these conditions, the leading and top waveforms of the load voltage are basically the same, indicating that... The impact on the active side establishment and top maintenance processes is relatively small. In contrast, the waveform differences in the pulse tail oscillation region are more pronounced. As the load increases, the amplitude of the tail oscillation increases, the fallback process slows down, and the duration of local disturbances lengthens.

[0068] therefore, Its main function is not to change the establishment or top level of the main pulse leading edge, but to enhance the transient energy storage at the high-impedance boundary on the opposite side, thereby amplifying the amplitude of the local disturbance in the pulse tail oscillation region and prolonging the fallback process. Reducing the junction capacitance of the cutoff device and its equivalent parallel capacitance helps to reduce local energy storage and improve pulse tail quality.

[0069] Branch stray inductance Impact: In fixed And with all other parameters remaining unchanged, Figure 8 Different The voltage waveforms at point D and the load voltage are shown. Since point D is located near the high-impedance boundary on the opposite side, its voltage change can reflect the influence of the branch stray inductance on the local transient oscillation; the load voltage characterizes the effect of this local oscillation on the output trailing edge waveform.

[0070] Depend on Figure 8 (a) shows that different Under these conditions, point D shows little difference during the main pulse establishment phase, but exhibits significant changes in the pulse tail oscillation region. With... As the voltage increases, the timing of the voltage oscillation peaks and troughs at point D gradually shifts later, the oscillation period increases significantly, and the duration of local oscillations also lengthens. This indicates that... Its main function is not to change the process of establishing the main pulse, but to change the time scale of local energy exchange at the opposite boundary.

[0071] Figure 8 (b) indicates that different Under these conditions, the leading edge and top regions of the load voltage are basically the same, indicating that... The impact on the output pulse build-up and peak hold is limited. In contrast, the pulse tail oscillation region has a greater impact on... More sensitive. With As the pulse increases, the peak-to-trough interval of the local oscillations in the later stage gradually increases, the oscillation period is lengthened, and the pulse tail fall-off process is correspondingly delayed.

[0072] therefore, Its main function is not to change the main pulse establishment process or top hold, but to adjust the time scale of local energy exchange at the opposite boundary, thereby changing the period and duration of local oscillation in the pulse tail oscillation region.

[0073] In summary, the simulation results verify the influence mechanism of the parasitic network in the power loop on the nanosecond pulse transient waveform. (Stray inductance of the conduction path) Primarily affecting the active-side commutation path, it increases the transient impedance drop during the leading edge build-up process and weakens the top-hold capability at the load end; non-conducting side junction capacitance It primarily determines the transient energy storage capacity at the high-resistivity boundary; its increase will enhance the amplitude of local oscillations in the pulse tail oscillation region; branch stray inductance This mainly changes the time scale of local energy exchange, thereby regulating the period and duration of pulse tail oscillations.

[0074] Example 3 To verify the working principle and nanosecond pulse output capability of the proposed dual half-bridge Blumlein type PFL structure, an experimental prototype was built and experimental tests were conducted.

[0075] Overall experimental platform such as Figure 9 As shown in (a), the experimental prototype consists of two symmetrical half-bridge switching boards, two pulse forming lines, and a center load. During the experiment, the DC bus voltage was provided by a JKDY300-10 power supply, and the switching logic signal was generated as a TTL signal by a signal generator and sent to the drive circuit. The high-voltage output signal at the load end was attenuated by a 50Ω, 40dB attenuator before being recorded on a Teledyne LeCroy WavePro 404HD oscilloscope with a bandwidth of 4GHz and a sampling rate of 20GS / s. Under the experimental conditions, the DC bus voltage was set to 200V, the load resistance was 50Ω, and the measured peak output current was approximately 4A. Other main experimental parameters are listed in Table 2.

[0076] Table 2

[0077] Prototype motherboard, such as Figure 9 As shown in (b). Considering the voltage withstand capability of the integrated GaN module under high voltage conditions, both left and right half-bridges are constructed using discrete GaN switches GS-065-004-1-L, with a withstand voltage of 650V. During high-frequency operation, parasitic parameters of the drive circuit can affect the waveform of the drive signal and the consistency of the switches. Therefore, the entire device adopts a floating structure, and each of the four switch drive units has an independent reference ground. The control signal is transmitted through an AFBR-2624Z opto-isolated device, and each drive unit is powered by a B1205S isolated power supply module. The drive chip selected is LMG1020YFFR, with a rise time of 375ps and a fall time of 350ps. The selected GaN device provides a Kelvin source pin, and the drive circuit can directly use the Kelvin source as the return path, thereby shortening the gate drive circuit and reducing the parasitic inductance of the drive circuit. Combined with a compact and symmetrical board layout, the stability of the nanosecond-level voltage transition process can be further improved.

[0078] It should be noted that a 40cm pulse forming line was used for the high-frequency output experiment; a 1.2m pulse forming line was used for the pulse width modulation and beam cutting experiments to facilitate observation of pulse width changes and beam cutting effects.

[0079] Output mode experiment.

[0080] Under the conditions of a DC bus voltage of 200V and a load resistance of 50Ω, the load voltage waveform under different switching sequences was tested, and the results are as follows. Figures 10-13As shown, the equivalent repetition frequency of the three waveforms is 20MHz. By changing the conduction sequence of the same polarity switches on the left and right sides, the load end can form three modes within one modulation cycle: two consecutive positive pulse outputs, positive and negative bipolar outputs, and two consecutive negative pulse outputs.

[0081] Figure 10 The results of two consecutive positive pulse outputs are presented. Within one modulation cycle, two positive pulses are generated sequentially at the load end, with the time interval between the two pulses corresponding to the set timing. The peak values ​​of the two pulses differ by only about 8V, which is a small difference relative to the pulse amplitude. A slight local dip is visible when the pulse falls back to near zero level, indicating that there is still a brief follow-up response after the pulse ends.

[0082] While keeping the conduction sequence corresponding to the second pulse unchanged, by adjusting only the conduction sequence of the switches corresponding to the first pulse, the output changes to... Figure 11 The diagram illustrates a bipolar form. In this case, the first pulse flips to negative polarity, while the second pulse remains positive, and the two pulses are still clearly separated in time. By adjusting only the switching sequence corresponding to the second pulse while keeping the first pulse's switching sequence unchanged, the output changes to... Figure 12 The bipolar form is shown. In this case, the first pulse remains positive, and the second pulse flips to negative, while the two pulses remain clearly separated in time. This result demonstrates that the switching of output polarity can be directly achieved through timing control without altering the main circuit connections.

[0083] Figure 13 The results of two consecutive negative pulse outputs are presented. The time interval between the two negative pulses also corresponds to an equivalent repetition frequency of 20MHz, and the amplitudes of the two pulses remain close. Similar to the two consecutive positive pulse outputs, a local transient exists near the zero crossover during the fallback process. However, the overshoot after the two negative pulses is not exactly the same; the overshoot after the first pulse is larger, while the overshoot after the second pulse is relatively smaller. This indicates that the local waveform details under continuous pulse conditions are related not only to the output polarity but also to the superposition state of the preceding and following wave processes.

[0084] comprehensive Figures 10 to 13 It can be seen that the dual-half-bridge Blumlein-type PFL structure can achieve three modes on the same hardware platform: two consecutive positive pulse outputs, positive and negative bipolar outputs, and two consecutive negative pulse outputs, by adjusting the conduction sequence of the same polarity switches on the left and right sides. All three output modes were clearly obtained in the experiment, and the pulse interval remained consistent with the set timing, indicating that the structure possesses stable multimodal output capability.

[0085] Experimental verification of pulse width modulation and beam cutting control.

[0086] Based on the output mode experiments, the pulse width modulation and beam-cutting control capabilities were further tested. A 1.2m pulse forming line was used in this section to more clearly observe the pulse width changes and the subsequent cutoff effect. The results are as follows: Figures 14-17 As shown.

[0087] Figure 14 A comparison of the output waveform under the influence of parasitic parameters and the waveform after beam clipping is presented. It can be seen that without beam clipping, a relatively long trailing waveform and local fluctuations remain in the pulse tail oscillation region, compressing the effective pulse range. After introducing beam clipping control, the later part of the pulse tail oscillation region is significantly truncated earlier, the original trailing waveform is compressed, and the additional fluctuations are also reduced. This result indicates that beam clipping control does not change the main circuit topology, but rather directly affects the effective pulse range at the load end through timing adjustment, thereby improving the later waveform.

[0088] The corresponding driver timings are as follows Figure 15 As shown. Among them, and Corresponding to the positive pulse formation process, and This corresponds to the negative pulse formation process. A specific relative conduction time is set between the two drive signals of the same polarity. This timing relationship determines the time interval between the voltage waves arriving at the load end on both sides, and forms the basis for positive and negative pulse width modulation and beam cutting control.

[0089] Figure 16 and Figure 17 Experimental results for pulse width modulation (PWM) of positive and negative pulses are presented separately. As the relative conduction time between switches of the same polarity decreases, the output pulse width gradually shortens. For positive pulses, the pulse width can be continuously adjusted from 10 ns to 3 ns; for negative pulses, continuous adjustment can also be achieved within the same range. Overall, the position of the pulse leading edge changes little, while the trailing edge shifts significantly forward with the control timing, indicating that the pulse width variation is mainly determined by the relative conduction relationship between switches of the same polarity. Under narrower pulse width conditions, local fluctuations in the pulse tail oscillation region are enhanced, but the main body of the pulse remains clearly distinguishable.

[0090] comprehensive Figures 14 to 17 It can be seen that the dual half-bridge Blumlein type PFL structure can not only achieve nanosecond-level pulse width modulation of positive and negative pulses, but also effectively compress and cut the local waveform of the pulse tail oscillation region through timing control.

[0091] The effect of gate drive resistor on pulse edge and narrowest pulse width capability.

[0092] To further analyze the impact of switching dynamic characteristics on narrow pulse beam-cutting capability, the gate drive resistance of the GaNHEMT was changed while keeping the main circuit structure and transmission line parameters unchanged. The effects of these tests on the positive and negative pulse output waveforms were then examined, and the results are as follows: Figure 18 As shown.

[0093] Depend on Figure 18 (a) and Figure 18 As can be seen from (b), different Under these conditions, the edge velocity and top waveform of both positive and negative pulses change. When At smaller values, the switching commutation speed is faster, and the pulse edges are steeper, but local overshoot and oscillation are also relatively more pronounced. With... As the voltage increases, the switching conduction process slows down, the pulse top overshoot and high-frequency disturbances are reduced, and the waveform smoothness is improved, but the switching speed of the leading and trailing edges also decreases.

[0094] For beam cutting control, the narrowest effective pulse width depends not only on the relative timing of the two switches but also on the switching edge speed. A larger... While it can suppress pulse overshoot at the top, it slows down the voltage wave's leading and trailing edges, making it difficult to form a sufficiently short effective pulse interval during the leading edge establishment and fallback process, thus limiting further pulse width compression. Therefore, The choice requires a trade-off between overshoot suppression and edge speed.

[0095] While existing literature proposes inductive energy storage line pulse generators that possess pulse width modulation capabilities and achieve adjustable pulse width output, their narrowest output pulse width is approximately 8 ns. In contrast, the dual half-bridge Blummlein-type PFL structure proposed in this invention, through dual-terminal timing control, compresses the narrowest effective pulse width of both positive and negative pulses to 3 ns in experiments, demonstrating that this structure has a stronger ability to achieve narrow pulse widths.

[0096] Bipolar alternating output experiment verification.

[0097] After completing the output mode and pulse width modulation experiments, the bipolar alternating output capability was further tested, and the results are as follows: Figures 19-21 As shown. Figure 19 The output waveform within a 10μs time window is presented. It can be seen that positive and negative pulses alternate at the load end, with a repetition frequency of 10MHz for both polarities, thus forming a continuous bipolar alternating pulse sequence at the load end. The pulse arrangement is clear throughout the time window, without obvious polarity misalignment or mode aliasing.

[0098] To further observe the details of a single pulse, Figure 20 and Figure 21Amplified waveforms of the positive and negative pulses are presented separately. It can be seen that the positive pulse width is approximately 3.5 ns, and the negative pulse width is approximately 3.6 ns, remaining close to each other. This indicates that the structure can maintain a relatively narrow pulse width even under bipolar alternating operation conditions. Simultaneously, both positive and negative pulses exhibit some post-pulse oscillation in the pulse tail oscillation region, with the local waveform change after the negative pulse's fall being more pronounced. This suggests that under continuous bipolar alternating output conditions, boundary response and parasitic parameters still affect the post-pulse waveform.

[0099] comprehensive Figures 19 to 21 It can be seen that the dual half-bridge Blumlein type PFL structure can not only achieve intra-period switching of bipolar pulses, but also stably form a positive and negative alternating nanosecond pulse sequence under continuous working conditions, and maintain a narrow pulse width under high repetition frequency conditions.

[0100] This invention proposes a dual-half-bridge Blummlein-type PFL nanosecond pulse generator. This structure achieves output mode switching, nanosecond-level pulse width modulation, and beam cutting control by adjusting the conduction sequence and relative conduction times of the switches on the left and right sides. Addressing the issues of pulse leading-edge establishment, top hold, and local oscillations after the falling edge, this invention establishes a power loop parasitic network model, which is verified through simulation and experiments.

[0101] The dual half-bridge Blummlein-type PFL structure proposed in this invention can realize positive pulse, negative pulse, and bipolar alternating pulse output on the same hardware platform. Experimental results show that by adjusting the conduction sequence of the same polarity switches on the left and right sides, a stable multimodal nanosecond pulse waveform can be obtained at the load end, and a bipolar pulse sequence of alternating positive and negative pulses can be formed under continuous operating conditions.

[0102] By adjusting the relative on-time of the switches of the same polarity on both sides, the prototype achieved nanosecond-level pulse width modulation. In the experiment, the widths of both positive and negative pulses could be continuously adjusted from 10 ns to approximately 3 ns. After further introducing beam-cutting control, the trailing portion of the pulse tail oscillation region was significantly compressed, and the output waveform quality was improved, indicating that dual-end timing control can effectively improve the narrow pulse generation capability.

[0103] Simulation and experimental results show that parasitic parameters of the power loop have a significant impact on the output waveform. The stray inductance of the conduction path mainly affects the rise-lead speed and the hold-at-the-top capability; the junction capacitance and branch inductance on the non-conducting side mainly affect the amplitude and time scale of local oscillations in the pulse tail oscillation region. The experimental phenomena are consistent with the simulation analysis results, indicating that the output waveform is jointly determined by the switching timing, parasitic parameters of the power loop, and the transmission line boundary response.

[0104] The gate drive resistance (Rg) has a significant impact on pulse edge speed and beam-cutting capability. A smaller Rg is beneficial for increasing switching commutation speed and obtaining a narrower effective pulse width, but it will enhance edge overshoot and local oscillations; a larger Rg can reduce top overshoot and high-frequency disturbances, but it will slow down the pulse edge, thus limiting further compression of the narrowest pulse width. Therefore, the drive parameters need to be compromised between edge speed, overshoot suppression, and pulse width modulation capability.

[0105] Under conditions of a DC bus voltage of 200V and a load resistance of 50Ω, the prototype achieved stable nanosecond pulse output with a peak output current of approximately 4A. Experimental results show that the proposed structure has good implementation capabilities in multi-mode output, bipolar pulse formation, nanosecond-level pulse width modulation, and beam cutting control. Future work will focus on optimizing parasitic parameters under high repetition frequency conditions, improving drive timing accuracy, and enhancing output waveform consistency.

[0106] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the invention.

[0107] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A dual half-bridge Blumlein-type PFL nanosecond pulse generator, characterized in that, It includes a first power supply circuit, a second power supply circuit, a first half-bridge circuit, a second half-bridge circuit, a first pulse forming line, a second pulse forming line, and a load; The first half-bridge circuit includes a first upper bridge arm switch and a first lower bridge arm switch connected in series between the positive and negative terminals of the first power supply circuit. The connection point of the first upper bridge arm switch and the first lower bridge arm switch is the midpoint of the first half-bridge. The midpoint of the first half-bridge is connected to one end of the load through a first pulse forming line. The second half-bridge circuit includes a second upper bridge arm switch and a second lower bridge arm switch connected in series between the positive and negative terminals of the second power supply circuit. The connection point of the second upper bridge arm switch and the second lower bridge arm switch is the midpoint of the second half-bridge. The midpoint of the second half-bridge is connected to the other end of the load through the second pulse forming line. Within one modulation cycle, by controlling the conduction sequence and relative conduction time of the first upper bridge arm switch, the first lower bridge arm switch, the second upper bridge arm switch, and the second lower bridge arm switch, two output pulses are generated on the load.

2. The dual half-bridge Blumlein type PFL nanosecond pulse generator as described in claim 1, characterized in that, When the first upper bridge arm switch turns on before the second upper bridge arm switch, a positive pulse is generated on the load; when the second upper bridge arm switch turns on before the first upper bridge arm switch, a negative pulse is generated on the load; when the second lower bridge arm switch turns on before the first lower bridge arm switch, a positive pulse is generated on the load; when the first lower bridge arm switch turns on before the second lower bridge arm switch, a negative pulse is generated on the load.

3. The dual half-bridge Blumlein type PFL nanosecond pulse generator as described in claim 1, characterized in that, When the midpoint of the first half-bridge transitions from a low potential to a high potential, the upper bridge arm switch is turned on; when the midpoint of the second half-bridge transitions from a low potential to a high potential, the lower bridge arm switch is turned on.

4. A dual half-bridge Blumlein-type PFL nanosecond pulse generator as described in claim 1, characterized in that, Within one modulation cycle, the polarity of the first output pulse is controlled by controlling the conduction sequence of the first upper bridge arm switch and the second upper bridge arm switch; the polarity of the second output pulse is controlled by controlling the conduction sequence of the first lower bridge arm switch and the second lower bridge arm switch.

5. A dual half-bridge Blumlein-type PFL nanosecond pulse generator as described in claim 1, characterized in that, Its output modes include positive-positive output, positive-negative output, negative-positive output, and negative-negative output.

6. A dual half-bridge Blumlein-type PFL nanosecond pulse generator as described in claim 1, characterized in that, The pulse width of the output pulse is determined by the relative conduction delay of each switch in the first half-bridge circuit and the second half-bridge circuit.

7. A dual half-bridge Blumlein-type PFL nanosecond pulse generator as described in claim 1, characterized in that, After the output pulse is established, the voltage across the load is brought back to zero in advance by controlling the relative turn-on time of each switch in the first half-bridge circuit and the second half-bridge circuit. During the positive pulse output process, after the first upper bridge arm switch is turned on, the second upper bridge arm switch is turned on after a set delay, so that the potentials at both ends of the load are equal again, and the positive pulse is turned off in advance. During the negative polarity pulse output process, after the first lower bridge arm switch is turned on, the second lower bridge arm switch is turned on after a set delay, so that the potentials at both ends of the load are equal again, and the negative polarity pulse is turned off in advance.

8. A dual half-bridge Blumlein-type PFL nanosecond pulse generator as described in claim 1, characterized in that, The first power supply circuit includes a first power supply, a first resistor, and a first capacitor; the second power supply circuit includes a second power supply, a second resistor, and a second capacitor; the first resistor and the first capacitor are connected in series between the positive and negative terminals of the first power supply, and the second resistor and the second capacitor are connected in series between the positive and negative terminals of the second power supply.

9. A dual half-bridge Blumlein-type PFL nanosecond pulse generator as described in claim 1, characterized in that, The first upper bridge arm switch, the first lower bridge arm switch, the second upper bridge arm switch, and the second lower bridge arm switch are all gallium nitride devices.