How TIG welding procedure affects penetration and slag island formation - part 1

TWI Bulletin, November/December 1988

David Harvey, MA, is a Senior Research Engineer in the Arc Welding Department.

Recent research at the Institute was conducted to reduce susceptibility of mechanised TIG welding to materials effects. ^[1,2] One particularly problematic stainless steel cast exhibited poor penetration characteristics and slag island formation using certain welding conditions. However, specific combinations of welding current, shielding gas composition and travel speed significantly increased penetration and eliminated slag island formation.

Increased use of mechanised and fully automated welding techniques has generally led to reduction in welding costs and significant improvement in weld quality and reproducibility. Despite advances in TIG welding process technology, many materials related problems remain which must be overcome if true weld reproducibility is to be achieved. Of these, cast to cast variation in fusion profile is the most commonly reported. ^[3,4] It is believed that during TIG welding, variations in minor element concentrations subtly affect both arc characteristics, by evaporation at the molten pool surface, and fluid convection through their effect on surface tension. ^[5,6] By comparison with other arc processes the TIG molten pool is relatively static, and minor changes in arc energy distribution across the pool surface, or in magnitude or direction of any convective flows, can exert considerable influence on the resulting fusion profile.

Samples of type 316 stainless steel sheet, supplied by a Member Company in two casts, were reported as having completely different welding characteristics. With a procedure developed on the more highly penetrating control cast, the second was impossible to weld with mechanised equipment using normal welding parameters and argon shielding, yielding lack of penetration or a weaving penetration underbead ( Fig.1). Although slag island formation is not generally considered to be a significant characteristic of TIG welding, initial penetration trials indicated that this was another property of the second problematic cast ( Fig.2). However, the behaviour was not found to be a simple inherent characteristic of this particular cast of stainless steel. Limited welding trials performed with different shielding gas compositions, welding currents and travel speeds revealed that the slag islands were formed only with certain combinations of these three welding parameters.

Research on materials aspects has been mainly concerned with determining the elements and their respective concentrations giving rise to variable penetration, and formulating a mechanism for their action. An alternative approach, offering a practical method of overcoming cast to cast variations, is to increase the tolerance of mechanised TIG welding to materials effects. This article outlines research by the Arc Welding Department into effects of basic procedure modifications on the cast displaying poor penetration and slag island formation.

Experimental

The experimental programme was divided into four distinct phases:

I To reduce the difference in penetration between the highly penetrating control cast (3D55) and the poorly penetrating cast (3D56) through modification of the welding procedure;
II Investigation of welding procedure influence on slag formation on cast 3D56;
III Remedial action to disperse slag islands;
IV Materials analysis.

Phase I: Welding procedure and penetration

Both casts (3D55 and 3D56) were supplied in the form of 2.8mm thickness sheet. Welding procedure tests consisted initially of autogenous melt runs placed along the centreline of 250 X 50mm strips, using the equipment shown in Fig.3. A precision stepping motor traverse, under quartz oscillator control, together with a transistor power source ensured excellent reproducibility of travel speed and welding current.

Five different shielding gas compositions were used:

1 Argon;
2 70% argon-30% helium;
3 80% helium-20% argon;
4 95% argon-5% hydrogen;
5 70% helium-25% argon-5% hydrogen.

Compositions 1-4 are commercially available cylinder gases, whilst the helium-argon-hydrogen combination was a specially mixed cylinder. Gases containing helium were chosen because of the extra arc energy input available, and thosewith hydrogen because of its chemically reducing properties.

Each gas was tested at travel speeds of 60 and 150 mm/min, with four welding currents used for each strip in one continuous melt run. The electrode tip, ground to a vertex angle of 60°, was maintained 2mm from the plate surfacefor each test.

Argon

At a travel speed of 150 mm/min the control cast 3D55 had a conventional rippled surface and distinct, evenly penetrated underbeads were produced through the current range ( Fig.4). Cast 3D56 exhibited a much wider molten pool with no significant solidification ripple, and full penetration was not achieved even at 140A. Transverse sections illustrate the dependence of fusion behaviour on weldingcurrent ( Fig.5). Whilst at 80A the penetration profiles are similar, increasing the current produces a remarkable difference, with the cast 3D56 molten pool widening considerably.

Reduction of welding speed to 60 mm/min resulted in considerable quantities of slag being deposited on cast 3D56. The arc was observed to wander from side to side in the vicinity of the slag islands, resulting in an irregular meltrun with a similarly erratic penetration underbead ( Fig.6).

70% argon-30% helium and 95% argon-5% hydrogen

Average underbead width on cast 3D56 was similar to that of the control cast at each current when using both these shielding gases. However, large slag islands were deposited on cast 3D56 with both, resulting in irregular, wandering melt runs and inconsistent, erratic penetration.

80% helium-20% argon

At the higher travel speed of 150 mm/min slag was again generated on cast 3D56, with consequent irregular melt runs and inconsistent penetration. By reducing travel speed to 60 mm/min a significant improvement in penetration wasachieved. Melt runs on both casts had comparable top surfaces with no trace of deposited slag on material 3D56. With the exception of the 40A condition, penetration on 3D56 closely matched that of the control cast.

70% helium-25% argon-5% hydrogen

This specially mixed gas was selected to combine the high arc energy obtained from helium with the chemically reducing properties of hydrogen. At 150 mm/min penetration behaviour of the two casts was virtually identical and no slagwas generated on 3D56 ( Fig.7,8). Similar results were achieved at 60 mm/min.

Main conclusions

Phase I demonstrated that the addition of helium and/or hydrogen could improve penetration behaviour of cast 3D56, although with the 80% helium-20% argon mixture, penetration was still poor at the higher travel speed of 150 mm/min.It appeared that although slag island formation caused irregular melt runs with locally poor fusion profile, this did not affect the average penetration bead width adversely. Slag formation, when using 80% helium-20% argon at thehigher travel speed only, suggested that the choice of welding procedure had a direct influence on this phenomenon.

Only the 70% helium-25% argon-5% hydrogen mixture eliminated slag island formation at both travel speeds. A more extensive investigative programme was required to establish whether a direct relationship existed between slag islandformation and the poor penetration behaviour of cast 3D56, and determine the influence of welding procedure on formation of slag islands. This was the purpose of Phase II of the project.

Phase II: Welding procedure and slag formation

Having established in Phase I that cast 3D55 was not susceptible to slag island formation, welding trials were restricted to cast 3D56 in Phase II. The five shielding gas mixtures used previously were tested again, but the range oftravel speeds was increased to include 240 mm/min. As before, autogenous melt runs were placed along the centreline of 250 X 50mm strips of material 3D56, each made at a single welding current. For each shielding gas/travel speedcombination, the welding current was generally increased in 10A steps between a minimum value, corresponding to the lowest current which resulted in plate fusion and a maximum, corresponding to an excessively penetrating weld bead. Thestep in welding current was reduced to 5A where changes in slagging behaviour were found. Each melt run was examined visually for slag island formation and plate penetration.

Results are illustrated graphically in Fig.9-13 by plotting slag island formation and penetration against welding current and arc energy (heat input). It can be seen immediately that there is no obvious relationship between slag island formation and plate penetration.

Argon

Use of argon shielding encouraged slag island formation at all three travel speeds. At 60 mm/min it was not possible to produce slag free, fully penetrated melt runs, and at 240 mm/min, whilst increasing the current eliminated slag formation, full penetration could not be attained within the operating range of the torch. However, slag free, fully penetrated melt runs were produced at 150 mm/min above a welding current of 110A. Figure 14 shows how critical a small change in welding current was to generation of slag islands. Increasing welding current from 95-100A at a travel speed of 150 mm/min eliminated their formation.

70% argon-30% helium

Whilst reducing the welding current required for full penetration to occur, use of this shielding gas increased the range of current over which slag islands were formed at each travel speed. It was, however, possible to attain full penetration at 240 mm/min with a minimum welding current of 160A.

80% helium-20% argon

No slag islands were generated at the lowest travel speed and the minimum current required for full penetration was reduced to 40A with 80% helium-20% argon. At both of the higher travel speeds slag was generated over a greatly reduced range of welding current values (when compared with the two previous shielding gas mixtures). However, slag free, fully penetrated melt runs were produced at the higher travel speeds by increasing welding current.

Argon-5% hydrogen

Addition of hydrogen to the shielding gas restricted slag generation to between 80-100A at a travel speed of 150 mm/min. Slag free, fully penetrated melt runs could be achieved above and below this range of welding current.

70% helium-25% argon-5% hydrogen

Use of 70% helium-25% argon-5% hydrogen mixture restricted slag island formation to the fastest travel speed above a welding current of 120A. This shielding gas resulted in the best combination of slag elimination and plate penetration of the five shielding gas mixtures.

Main conclusions

As previously reported, use of argon shielding resulted in poor penetration, coupled with slag island formation across a wide range of welding current at each travel speed. Addition of 30% helium to the shielding gas significantly improved penetration behaviour but did little to reduce slag island formation. The 80% helium-20% argon mixture improved penetration behaviour, and also greatly reduced the welding current/travel speed combinations generating slag islands. Inclusion of hydrogen in the shielding gas was found to be the most significant factor in reducing the incidence of slag island formation. The 70% helium-25% argon-5% hydrogen mixture resulted in the best combination of slag elimination and plate penetration.

Part 2

References