The total voltage was measured between the contact tube of the welding torch, where a specific junction had been mounted, and the workpiece.

The two-colour pyrometer was placed on the left side of the torch in welding direction and at a distance of 0.55 m from lens to wire tip. It has been operated at a frame rate of 6000 fps and a typical exposure time of 160 us. The measurement range is between 1600 and 3400 K, which well covers the droplet temperature range. Droplet images were recorded during the base current phase (25 A) where the impact of arc radiation can be neglected. The surface temperature for every point of the droplet image was calculated from the corresponding points of the two images according to the principals of two-colour pyrometry. The optical system was calibrated using a tungsten ribbon lamp.

Determination of the droplet shape from side-on droplet images of the two-colour pyrometer using the canny edge detection method, a calibration of the images with the coefficient k=36 pixel/mm and the assumption of a circular cross section of the droplet at each vertical position.

Determination of the droplet mass and energy as the sum of masses and energies of horizontal cylindrical slices of the droplet with a height of 27.8 µm corresponding to one camera pixel. In particular, the change of the temperature and shape of each droplet along the vertical direction was considered. Therefore, the diameter of each slice was determined from the droplet image processing. A homogeneous temperature of each slice was considered as the average of the measured surface temperature of the slice. The mass of the slice corresponding to the slice volume and a temperature dependent density as well as the energy of each slice corresponding to a temperature dependent heat capacity were calculated.

The averaged power delivered to the wire anode was determined from the sum of the energies of all droplets in a time period of 0.15 s taken at the moment right after droplet detachment. The power delivered to the wire anode from the plasma was estimated considering the droplet energies as well as the Joule heating power of the wire. The latter have been taken from preceding studies based on pyrometry of the wire and a temperature dependent conductivity of the wire material. A voltage equivalent of the power delivered to the wire from the plasma was calculated from an energy balance taking into account the time dependent current and the time dependent resistances of wire and contact tube.