Research on ZVS Arc Ignition Circuit and Its Conducted Interference

Abstract

A zero-voltage switching (ZVS) push–pull self-oscillating arc ignition circuit was proposed, marking the first application of ZVS technology in welding arc ignition systems. The circuit’s working principle was analyzed, and time-domain waveforms of the switching transistors verified the realization of soft switching. A conducted interference test platform was established in order to assess the circuit’s electromagnetic compatibility under no-load and arc ignition transient conditions. In comparison with conventional domestic arc ignition circuits, the proposed ZVS circuit demonstrated substantially diminished quasi-peak interference levels, with a reduction exceeding 9.5 dB in both instances. Additionally, under no-load conditions, the ZVS circuit demonstrated interference levels comparable to those of a commercial Fronius system, while during arc ignition transients, it exhibited an over 5 dB reduction. The findings of this study demonstrate that the incorporation of soft-switching techniques into arc ignition circuits can effectively mitigate conducted interference, thus providing a promising and practical approach for industrial welding equipment.

1. Introduction

The year 2020 marked the release of the electromagnetic compatibility standard GB/T 15579.10 [1] for arc welding equipment in China, which represented a further increase in the requirements for electromagnetic interference of arc welding power sources. TIG (Tungsten Inert Gas) welding, a high-precision welding process, has found extensive application in the aerospace and nuclear industries due to its ability to produce excellent welding quality [2,3]. However, due to the existence of high-frequency and high-voltage arc ignition, the interference hazard is the greatest in the arc welding power source category. This has the potential to affect not only peripheral equipment, such as communication interruptions, data loss and precision instrumentation measurement errors, but also to cause serious safety hazards [4]. The existing research in this field has focused primarily on the development of arc ignition methods [5,6,7]. For instance, Pang [8] utilized high voltage arc ignition for micro-TIG welding, enhancing the stability of arc initiation through the optimization of the arc oscillation triggered by cathode spot (CSM) motion. Liu [9] proposed an arc ignition technique based on arc thermal conversion and elucidated its mechanism. This technique used the Joule heat generated by contact to form a metal bridge, which was evaporated by the electrode rotation and mechanical force to generate an arc. Tang [10] investigated the arc ignition process in TIG-MIG composite welding and examined the impact of altering parameters such as the welding torch tilt angle on the regularity of auxiliary arc ignition. Xia [11] investigated the TIG arc ignition method based on a 1000 W semiconductor laser. The research on electromagnetic interference of arc welding power supply mainly focuses on the simulation analysis of interference and filter design [12,13,14,15,16], and there is almost no research on the interference of arc ignition. Interference is primarily propagated by radiation and conduction in two ways. Conduction interference is directly coupled to sensitive equipment through power lines and other conductors. This form of interference is often more direct and difficult to inhibit. This is the focus of the optimization of electromagnetic compatibility design. In this paper, a novel arc ignition circuit for TIG welding machines is proposed, and the conducted interference is quantitatively analyzed with the objective of reducing the interference during the arc ignition process of TIG welding. This is of significant value for engineering applications. Moreover, future manufacturing systems may rely on satellite links for global connectivity [17]. In such environments, high-frequency interference from welding—especially during arc ignition—can compromise data integrity and communication reliability. Reducing EMI in the ignition stage is thus essential for ensuring stable operation in communication-sensitive applications.

2. Hardware Test Platforms

The conducted interference test platform has been constructed in accordance with the provisions stipulated in the standard GB 4824 [18]. This standard is instrumental in regulating the RF Emission characteristics, limits and measurement methodologies for industrial, scientific and medical equipment. It is notable that the standard does not provide a clear limit for the conducted interference in the arc ignition process.

The conducted interference test platform consists of a welding system, a linear impedance stabilization network (LISN), an isolation transformer, a spectrum analyzer, a digital oscilloscope, and a high-voltage probe, as demonstrated in Figure 1.

The welding machine is a single-phase TIG welding power supply model WS-250S manufactured in China. The linear impedance stabilizing network is the EM5040B from Zhiyong Company (Shenzhen, China), and the high pass filter at the output is set to 150 KHz. The oscilloscope employed is the SDS824X four-channel oscilloscope from Tingyang Company (Shenzhen, China). The interference receiver is a Goodwell GSP-818 spectrum analyzer, which has an EMI mode and complies with the international electromagnetic compatibility standard CISPR 16-1-1 [19]. The resolution bandwidth (RBW) and the video bandwidth (VBW) are set to 9 kHz according to the standard. The starting frequency is set to 150 kHz, and the ending frequency is set to 30 MHz.

3. Hardware Circuit Comparison and Design

3.1. Original Arc Ignition Circuit Design

The arc ignition circuit, a component of welding machines, is a subject of current interest in the domestic market. Despite the numerous variations, the fundamental differences among them are not significant. The original arc ignition circuit illustrated in Figure 2 will be employed for the purpose of comparison and analysis. It is composed of three primary circuits: the step-up main circuit, the step-up control circuit, and the oscillation coupling circuit. The step-up main circuit consists of the secondary center tap coil L₁ of the inverter transformer, diodes D₁-D₄, DC boost high voltage pack T₁, and R₁C₁ filter circuit. The step-up control circuit consists of an eight-legged relay K₁, a switching tube Q₁, and other associated components. The oscillating coupling circuit consists of a spark gap HF, a high-voltage capacitor C₅ and a coupling inductor T₂. The secondary side of the coupled inductor is connected to the main circuit of the welder via the secondary boost.

The operation of the circuit is facilitated by the switching of Q₁, resulting in an alternating current (AC) square wave voltage being generated on L₁. When the voltage is positive half-wave, D₁ and D₄ are conducting, and negative half-wave voltage is present in D₂ and D₃. When the arc ignition circuit functions, the control signal input to the eight-pin relay K₁ results in the closure of normally closed contacts ab, de and the opening of normally open contacts ac, df. The positive half-wave through the filter circuit R₁C₁ results in the flow of current through the diode D₄ (contact c, contact a). The gate starter resistor R₂ activates the switch Q₁, enabling the current to flow through D₄, R₄ into Q₁, and finally from the diode D₁ back to the L₁. The negative half of the work process is analogous. Following the T₁ DC high-voltage packet boost, the oscillator capacitor C₅ is charged, and the breakdown spark gap HF is established, in a manner analogous to the RLC attenuation oscillation.

3.2. Zero-Voltage Switching Arc Ignition Circuit Operation Principle

As illustrated in Figure 3, the proposed zero-voltage switching arc ignition circuit has been designed. The output of the inverter circuit, which is part of the transformer side square wave voltage, is rectified by the DC-DC module. The design utilizes the VRB4812LD-50W module manufactured by Jin Sheng Yang (MORNSUN). The input voltage of this module ranges from 60 V to 12 V DC, with an output current of up to 4167 mA and a maximum output power of 50 W. The power module’s primary parameters and indicators are delineated in

Table 1. Parametric indicators of the DC-DC module.

Parameter	Value Range
Power Efficiency	82%
Isolation Voltage	1500 V (DC)
Standby Power	0.048 W
Input Voltage	36~75 V
Insulation Resistance	100 MΩ

. The output of this is 12 V, which is connected to the VCC. The subsequent progression involves the utilization of a zero-voltage push–pull self-excited circuit, which is integrated with the DC high-voltage package T₁ to facilitate a boost. Following this boost, the capacitor C₃ undergoes a charging phase, subsequently leading to a breakdown spark gap HF, akin to the RLC attenuation oscillator. Finally, the coupled inductor T₂ provides a secondary boost to the main circuit of the welder.

The operation of the circuit is outlined as follows [20,21], where i_L1 denotes the current in the choke inductor, and the potentials at A and C represent the potentials of the drains of MOS tubes Q₁ and Q₂, which are denoted by u_A and u_C, respectively. The current flowing into the drain D of Q₁ and Q₂ is represented by i_Q1 and i_Q2, respectively. The potential of the center tap B is represented by u_B. The voltages at the extremities of capacitors C₁ and C₂ are designated as u₁.

The start-up phase is as follows: When the ZVS circuit is powered up, the current in the choke inductor L₁ is zero, and then the inductor current starts to increase. Concurrently, the current reaches the gate of the MOS tube via the gate startup resistors R₁ and R₂ and the voltage regulator diodes D₃ and D₄ after clamping. It is important to note that due to the difference in semiconductor devices, i_Q1 and i_Q2 are not equal. Assuming that i_Q1 > i_Q2, the excitation current on L₂ generates a positive induced voltage and a negative induced voltage. L₂ and C₁, C₂ resonance occurs in parallel. At this time, where u_A ≈ 0, the gate voltage of Q₂ is pulled down and cutoff.

The first operating state is defined as follows: Q₁ conduction, the previous state i_L2 direction from A to C, and the resonant capacitor voltage is zero. Given the impossibility of sudden alteration to the inductor current, the inductor charges the capacitors, resulting in a polarity of C positive and A negative. Consequently, the voltage at point A is pulled down to 0 V, thereby establishing a sinusoidal voltage at point C and inducing a sinusoidal state in i_L2. It is notable that when i_L2 = 0, the voltage of C₁ and C₂ reaches its maximum value. Q₁ is characterized by a constant state of activation.

Operating state 2 is characterized by the discharge of C₁ and C₂ from L₂, with i_L2 undergoing a sinusoidal increase from C to A. It is noteworthy that this sinusoidal increase in i_L2 is concurrent with the discharge of the resonant capacitor. The inductor current i_L2 reaches its maximum value when the resonant capacitor is discharged.

Operating state 3: L₂ is responsible for charging the resonant capacitor, and the voltage polarity is A positive, C negative. The gate voltage of Q₂ increases, and finally Q₂ conducts and Q₁ cuts off. In the context of a stable supply voltage, the alternation of conduction in Q₁ and Q₂ gives rise to a self-excited LC oscillation.

The maximum value of voltage u₁ across the capacitor is constrained by the following equation:

u₁ = VCC × π

(1)

The above equation demonstrates that the withstanding voltage of the resonant capacitor, in conjunction with that of the mosfet, is contingent on the supply voltage. In the context of mosfet selection, it is imperative to exercise caution with regard to the device parameters, as this particular topology necessitates a pronounced emphasis on the mosfet voltage withstand. The selection of the resonant switch is made with Infineon’s IRFP260NPBF mosfet. The device has been engineered to possess a 200 V drain source breakdown voltage and 50 A continuous drain current capability. This ensures that it meets the requirements of low-voltage high-current arc-inducing circuits on the voltage margin and current carrying capacity. The output capacitance Coss of 603 pF has been demonstrated to reduce turn-off losses in the ZVS mode and to control the resonant frequency within a reasonable range. In addition, the on-resistance of 40 mΩ has been shown to help suppress on-resistance losses.

The resonant frequency, f, of LC parallel resonance is expressed by the following equation:

$$f={\displaystyle \frac{1}{2\pi \sqrt{L_2(C_1+C_2)}}}$$

(2)

The peak current i_m2 on the resonant inductor L₂ is constrained by the following equation:

$$i_{\mathrm{m}2}=u_1\sqrt{{\displaystyle \frac{C_1+C_2}{L_2}}}$$

(3)

It can thus be concluded that the oscillation frequency of the circuit is set to approximately 15 kHz. The resonant capacitors C₁ and C₂ are 100 nF, and the resonant inductance is 560 µH.

3.3. Parametric Design of Oscillatory Coupling Circuits

The equivalent circuit of the RLC oscillation circuit is illustrated in Figure 4, which corresponds to the spark gap HF, the resonant capacitance C₃ and the primary inductance of the coupling coil T₂ in Figure 3. R represents the equivalent resistance of the circuit, including the resistance of the spark gap at the moment of discharge and the line resistance of the attenuating oscillation.

In the context of a series RLC circuit, the current i is expressed by Kirchhoff’s voltage law as follows:

$$L{\displaystyle \frac{d^{2}i}{d{t}^2}} + R{\displaystyle \frac{di}{dt}}+{\displaystyle \frac{i}{C}} = 0$$

(4)

As demonstrated in the aforementioned differential equation, the circuit underdamped response must satisfy the condition that the damping coefficient, denoted by ξ, is less than one. The expression for the damping coefficient is as follows:

$$\xi ={\displaystyle \frac{R}{2}}\sqrt{{\displaystyle \frac{C}{L}}}<1$$

(5)

It is evident that the line resistance in the actual circuit and the total resistance during spark discharge are not negligible [22]. The damped oscillation frequency, f_d, satisfies the following expression:

$$f_{\mathrm{d}}={\displaystyle \frac{1}{2\pi }}\sqrt{{\displaystyle \frac{1}{LC}}-{({\displaystyle \frac{R}{2L}})}^2}$$

(6)

The process resistance of the spark gap and the line resistance are not easily measurable parameters. Thus, Formulas (5) and (6) can be derived from the under-damped oscillation frequency, f_d, coupling inductance, L, and oscillation capacitance, C, in the under-damped oscillation conditions of the relationship expressed by Formula (7).

In the preceding analysis, the oscillation frequency is derived based on ideal lumped components. It is acknowledged that distributed parasitic capacitances and inductances have the capacity to introduce slight shifts to the practical resonant frequency. However, it is anticipated that their magnitude will be relatively negligible in this design.

$${\displaystyle \frac{1}{2}}\sqrt{{\displaystyle \frac{C}{L}}}\cdot \sqrt{{\displaystyle \frac{4L}{C}}{-16\pi }^2{L}^2{f}_{\mathrm{d}}^2}<1$$

(7)

According to the damped oscillation frequency, an increase in capacitance C will lead to a decrease in f_d. Concurrently, an increase in capacitance will enhance energy storage capacity, thereby amplifying the discharge strength and resulting in an under-damped oscillation. This, in turn, will increase the amplitude and duration of the input side of the conduction interference, and concurrently, it may cause the circuit to transition out of the under-damped response state. Conversely, a decrease in capacitance C will stabilize the under-damped oscillation, and concurrently, as f_d rises, the energy will be distributed to multiple oscillation voltages, thereby reducing the single peak voltage and assisting in the reduction of interference at the input side. However, it is imperative to ensure that the cold welding arc requirements are met. It is evident that decreasing capacitance C will stabilize the underdamped oscillation, while an increase in f_d will result in the energy being allocated to multiple oscillation voltages, thereby reducing the single peak voltage. This, in turn, is beneficial in reducing interference at the input side. However, it is imperative to ensure that the arc requirements of the cold welding are met. The analysis of the oscillation parameters reveals that the oscillation coupling capacitor should be constructed using 470 pF, 5 kV high-voltage ceramic capacitors, and inductors employing a ferrite core, with eight turns of wire measuring 5.14 µH.

The proposed ZVS arc ignition circuit demonstrates good potential for industrial implementation. Its simple topology allows for seamless integration with existing welding power supplies. By utilizing readily available components and avoiding complex control systems, the circuit remains cost-effective. Additionally, the design is compatible with standard PCB manufacturing and assembly processes, supporting scalability and mass production.

4. Circuit Verification and Test Results

4.1. Circuit Verification

Following a thorough analysis of the principle and the design of the hardware, the waveforms of the soft-switching arc ignition circuit are verified on the WS-250S welding machine.

4.1.1. Zero-Voltage Switching Circuit Waveforms

As demonstrated in Figure 5, the gate voltage of the MOS tube exhibits fluctuations, indicating that the circuit is operating in self-excited oscillation mode. The two switching tubes are alternately turned on and off, forming a symmetrical operating state.

The signal condition of the MOS tube is illustrated in Figure 6. The electrical signals under single turn-on and turn-off conditions demonstrate that the ZVS zero-voltage switching circuit exhibits a negligible turn-on loss during the turn-on and turn-off process of the switch. Furthermore, the switch has achieved a quasi-resonant state in the self-excited state. However, it should be noted that the switching loss remains unavoidable due to the limited resonant energy.

As demonstrated in Figure 7, the output voltage waveform on the primary inductor L2 of the DC high-voltage packet is shown, and the LC resonant frequency is 15.4 kHz. The resonant sinusoidal peak voltage is approximately 38 V, and the resonant frequency and the sinusoidal peak voltage are consistent with the result of Equation (2).

4.1.2. Oscillating Coupled Circuit Waveforms

As illustrated in Figure 8, the output voltage of the coupling inductor T₂ on the secondary side of the oscillatory coupling circuit is depicted under undamped response. The underdamped oscillation frequency f_d is approximately 2.17 MHz, and the peak voltage of the underdamped oscillation is 2400 V. It has been verified that the oscillation parameters and frequency conditions described above satisfy the conditions of the underdamped response in Equation (7).

4.2. Conducted Interference Test

The present study is based on the conducted interference test platform on the welding power supply under different conditions of the arc ignition circuit on the input side of the live wire and the neutral wire of the conducted interference test. The test is comprised of two distinct states: the no-load interference test, defined as the state in which the arc has not been initiated, although the arc ignition circuit functions within normal parameters. The second state is the arc ignition instantaneous interference test, which involves the use of a fixed torch to position the tungsten electrode at a distance of 1 mm from the workpiece during a standard welding arc test.

The detection mode is in accordance with the requirements of the GB/T 15579.10 standard, employing quasi-peak detection as outlined in [23,24]. The findings of the study indicate that the data obtained from the live wire and the neutral wire are almost identical. Consequently, the quasi-peak value of the live wire is selected as a representative interference characteristic parameter for analysis.

4.2.1. Interference Test with No Load on the Soldering Gun

In the experimental setup, the WS-250S welding power supply was employed in conjunction with the original arc ignition circuit in the argon arc welding mode for the torch no-load interference test. To facilitate a comparative analysis, the ZVS arc ignition circuit was utilized for a comparative test. The quasi-peak test results on the live wire of the spectrum analyzer are shown in Figure 9.

The experimental findings demonstrate that the maximum amplitude of the quasi-peak detected wave on the live wire of the original arc ignition circuit is 98.37 dBµV. Following the implementation of the ZVS arc ignition circuit for the conducted interference test, it was determined that the maximum amplitude of the quasi-peak detected wave on the live wire was reduced to 85.76 dBµV. The maximum decibel difference of the quasi-peak detected wave on the live wire is 12.61 dB, which is equivalent to the voltage amplitude reduced to about 23.4% of the original amplitude. This reduction corresponds to a reduction in the voltage amplitude to about 23.4% of the original amplitude. This finding indicates that the ZVS circuit is capable of effectively mitigating the conducted interference caused by the arcing voltage on the input side of the torch when the torch is unloaded.

4.2.2. Arc Ignition Instantaneous Interference Test

The instant of arc initiation is a very brief process, necessitating the setting of the maximum hold function of the spectrum analyzer for the purposes of testing. The distance between the tungsten electrode and the workpiece is 1 mm. The test involves a comparison between the original arc ignition circuit and the ZVS arc ignition circuit. The results of the quasi-peak detection on the live wire are shown in Figure 10. The maximum interference value of quasi-peak detection on the live wire is 79.78 dBµV for the original arc ignition circuit. Subsequent to conducting interference tests on the ZVS arc ignition circuit, it was ascertained that the maximum interference value of quasi-peak detection on the live wire had been reduced to 70.26 dBµV. The maximum decibel difference of quasi-peak interference on the live wire is 9.52 dB, which reduces the corresponding voltage amplitude to 33.4% of the original amplitude. Consequently, the ZVS circuit demonstrates superiority over the original circuit with respect to input interference at the moment of arc ignition.

4.3. Product Comparison Analysis

In order to verify the design reliability of the arc ignition circuit, a comparison is made between the WS-250S welding power supply equipped with ZVS arc ignition circuit and the iWave 230i from Fronius. The quasi-peak frequency spectrum of the live wire for the instantaneous interference between the no-load and arc ignition moment of the welding torch is shown in Figure 11.

As shown in Figure 11a, in the no-load interference test of the welding torch, the maximum interference value of the quasi-peak detection mode on the live wire is 88.1 dBµV. A comparison and analysis with the welding power supply equipped with the ZVS arc ignition circuit reveals that the interference level of the ZVS arc ignition circuit is lower than that of mainstream products in the no-load interference test, and the overall performance is comparable to that of the existing products, with the amplitude of the interference voltage in a similar range.

As shown in Figure 11b, the arc instantaneous interference test on the live wire quasi-peak detector interference maximum value was found to be 76.03 dBµV. A comparison of the ZVS arc ignition circuit with mainstream products in the input side of the interference spectrum found in the arc instantaneous test on the live wire quasi-peak detector reveals that the ZVS arc ignition circuit has an interference maximum value of the quasi-peak detector spectrum that is approximately 5.77 dB lower than that of mainstream products. Furthermore, the voltage amplitude is reduced to 51.5%, verifying that the design of ZVS is able to suppress high-frequency interference. It is noteworthy that Fronius incorporates a dedicated EMI filter on the input side, while domestic welders generally lack such a feature, thereby further substantiating the superiority of the arc ignition circuit.

5. Conclusions

This study innovatively integrates zero-voltage switching (ZVS) technology with oscillatory coupling principles to design a new type of arc ignition circuit, addressing the persistent issue of electromagnetic interference during arc initiation. This paper is the first to document the implementation of ZVS in welding arc ignition systems. A conducted interference test platform was established in accordance with the provisions of the GB 4824 standard, encompassing both torch no-load and arc transient conditions. A quantitative analysis was conducted in accordance with the provisions of GB/T 15579.10, with the objective of evaluating the input-side interference.

The experimental findings indicate that, under conditions of no load, the maximum voltage amplitude was diminished to 23.4% of its original value. This is equivalent to a 12.61 dB reduction in interference. During the occurrence of arc transients, the amplitude underwent a decrease to 33.4%, accompanied by a 9.52 dB reduction. The findings of this study suggest that the proposed circuit is capable of achieving a conducted interference level that is comparable to that of mainstream commercial systems. This outcome serves as a testament to the practical viability of the circuit. Moreover, the successful implementation of soft switching in arc ignition circuits introduces a new approach to improving electromagnetic compatibility in welding technology.