**1. Introduction**

Short circuit gas metal arc welding (S-GMAW) has various advantages such as low heat input, small heating area, and high thermal stability. Benefiting from advances in digital control technology, the power sources can control the voltage and current and output specific shapes of the arc curve which aim to handle the molten material transfer and control the spatter. Kah [1] proposed a classification of control techniques for S-GMAW: Natural metal transfer, current controlled dip transfer, and controlled wire feed short circuit mode. On account of the controlled wire feed short circuit mode represented by cold metal transfer welding (CMT) introducing external mechanical forces on the wire it will not be discussed here. Different shapes of the current and voltage waveform change the droplet transition, which leads to the different weld pool behavior, weld shape and microstructure.

Currently, there are a variety of S-GMAW waveform control methods on the market, which can be divided into two types. The first method is represented by Surface Tension Transfer (STT) [2] and Cold Arc process [3]. This type of methods reduces the circuit current at the beginning and end of the short circuit period to permit a smooth touch and break of the bridge of the molten metal, preventing spatter. The other type is represented by Cold MIG (Metal-Inert Gas Welding) process [1] and Low Spatter

Control (LSC) [4]. In this type of methods, considerable increase of the current gradient to accelerate the droplet detachment during the short-circuit period meanwhile the short circuit period, dramatically reduces the short-circuit period. The current in this period is reduced and occurs faster compared to a conventional short arc. However, the welding circuit of this process maintains partial current when the bridge breaks compared with the first type processes. Figure 1 compares the waveform of the conventional, the Cold Arc, and the LSC.

**Figure 1.** Comparison of (**a**) conventional, (**b**) Cold Arc, and (**c**) Low Spatter Control (LSC) waveform.

The behavior of weld pool is a direct reaction of weld pool state, the dynamic variation of weld pool has a great influence on the weld bead shape and microstructure. Weld pool behavior may contain sufficient information to understand the mechanisms of welding bead formation and control the stability of welding process [5–7], so it is of great significance to study the weld pool behavior of S-GMAW using different waveforms.

Many researchers have studied the weld pool behaviors in recent years. In gas tungsten arc welding (GTAW), for acquiring the amplitude and oscillation frequency of the weld pool, Yu Shi et al. [8] used line laser to illuminate the surface of weld pool. It is found that the oscillation frequency and amplitude of GTAW pool change abruptly in the process of partial penetration to full penetration. Liu [9] investigated the pulse frequency on fluid flow behavior of the weld pool in pulsed current GTAW. The result showed that weld pool oscillations triggered by pulse current lead to more heterogeneous nucleation sites, and the resonance between the movement of the weld pool and pulse current frequency greatly promotes grain refinement.

Compared with GTAW, there are complex interactions among arc plasma, droplet transfer, and pool behavior in GMAW. Richardson et al. [10] found that current pulses could not be used to trigger weld pool oscillation effectively for GMAW, the interactions between the transferred droplets and the weld pool can trigger the weld pool into oscillation. Tang et al. [11] developed a filter-reflection observation system to acquire the weld pool profile during double-pulsed gas metal arc welding process. It was found that the weld pool oscillation caused by low frequency pulse can effectively reduce the porosity and refine the weld structure.

To date, most investigations of weld pool behavior mainly focuses on GTAW and pulsed GMAW processes. However, no much research has been done in the area of weld pool characteristics with different S-GMAW waveform control methods. In this paper, a high speed photograph-images analysis system for weld pool observation was formed to capture the dynamic behavior of S-GMAW weld pool with the aim to reveal the internal relationship among the S-GMAW current waveform, the behavior of weld pool, the geometry of weld bead, and the microstructure.
