**3. Experimental Procedure**

The bead formation and droplet transfer are more affected by waveform parameters [25,26]. Q235, a typical mild steel, was selected by researchers to study the formation, arc behavior and droplet transfer, like Wu [15], Ghosh [27,28], and so on. In our work, 4-mm-thick mild steel (Q235) sheets were used as workpieces, whose dimensions were 150 mm × 50 mm. Thus, mild steel wire (ER50-6) with a diameter of 1.2 mm was used as the electrode wire and the CTWD (contact tube-to-work distance) was selected as 20 mm. As shielding gas, 82% Ar + 18% CO2 mixture gas with a flow rate of 15 L/min was selected. The speed of the wire feed and the speed of welding were 3 m/min and 3 mm/s, respectively. With the purpose of stable ODPP transfer, pulse current waveform showed in Figure 2 was selected according to Ref. [15]. In Figure 2, *t*0, *t*<sup>1</sup> and *t*<sup>2</sup> are the start point of the pulse peak current period, the droplet-detachment current period and the base current, respectively. Over 90% droplets are ODPP transfer at middle arc length (about 5 mm) in terms of corresponding high-speed photographs. Typical high-speed photographs of ODPP transfer mode are displayed in Figure 3.

**Figure 2.** Main parameters of pulse current waveform.

**Figure 3.** Typical high-speed photographs of "one droplet per pulse" (ODPP) transfer mode.

In the high-speed camera system, in order to observe the droplet process more clearly, a narrow-band filter was used to weaken the brightness of the arc. As a result, when the current was below 300 A, the arc was almost filtered out, and was difficult to observe (see Figure 3). However, the morphology of the arc can be seen clearly with the high current and arc brightness, such as in photographs near *t*1. Additionally, the anode spot of arc is always over the droplet when the current is above a critical current level, so the arc length is not affected by the droplet near *t*1. Therefore, for the purpose of more accurate measurement, the arc length near *t*<sup>1</sup> was selected to be measured in this pulse period. Then, the average arc length of a random 30 pulses is calculated as the arc length with these welding parameters.

Based on the pulse current waveform in Figure 2, the arc length was adjusted from less than 2 mm to more than 10 mm by only changing the base current time, without altering other parameters. Then, welding tests were performed on the piecework with different arc length. The parameters table of the pulse waveform and the experiment are presented in Table 1.


**Table 1.** The parameters table of pulse waveform and experiment.
