scribes current and weld time effects on the thinning seen across more than 60 spot welds from DP18 and DP20. Exactly like in Fig. 2, the results for both DP18 and DP20 were nearly identical. In Fig. 4, note that we superimposed a dash line to show regions of expulsion and no expulsion (left to the dash line). At the onset of expulsion, thinning never exceeded 30%, and was less than 20% for most welds. With this limited indentation, numerous studies have shown that the indentation effect is negligible compared to that of weld diam- eter (Refs. 15, 20, 21). Consequently, the effect of process parameters on weld di- ameter (measured after peeling to pro- mote weld buttons) had to be examined before a link with the types of weld frac- ture could be researched.
Figure 5 shows that weld (button) di- ameter increases with the current at a rate that is gradually and monotonically de- creasing, as observed elsewhere (Refs. 20, 22–24). When the current exceeded 11 kA, the weld button diameters in both steels stabilized slightly above 6.0 mm. At lower currents, the welds were noticeably larger in DP18, particularly when button diameters were less than 5.5 mm. This re- sult revealed that welds in DP18 formed at lower heat inputs than in DP20, and that could be reasonably well explained by the fact that DP18 was 10% thinner than DP20. With thickness emerging as the main explanation for the discrepant weld diameters in DP18 and DP20, we were ex- pecting a number of spot welds from DP18 and DP20 to exhibit different types of fracture, even with identical process parameters. The effects of process para- meters on weld mechanical properties, in- cluding types of weld fracture, were there- fore important to clarify, first under well-controlled conditions, and then using the chisel test; i.e., as if welds were quality controlled on factory floors.
Tensile-Shear Test Results
Many welds, fabricated with identical parameters as in Fig. 5, were tensile- sheared to quantify their load-carrying ca- pabilities, and, despite loading conditions different than in the chisel test, determine if one steel was also more prone to spot weld interfacial fracture than the other. In Fig. 6, peak force recorded during testing (i.e., the weld fracture force) is repre- sented as a function of current. Figure 6 shows that the weld fracture forces were consistently greater in DP18. Although the trend lines for the two steels were alike, the trend line for DP18 was also shifted to the left of that for DP20; a con- firmation that the welds began to form at smaller currents in DP18, and that the
greater fracture forces seen in the welds of DP18 were primarily due to its thinner gauge.
Among the welds in Fig. 6, six out of the 38 (all in DP18) created buttons. With several, the fracture forces were seen to reach a maximum at intermediate cur- rents (i.e., 10 to 10.5 kA). At these cur- rents, interfacial fracture was also seen to fully disappear. With a further increase in current, interfacial fracture continued to be avoided, but the sheets were also alarmingly indented. As confirmed by Fig. 4, with currents near 10 kA, thinning across welds was about 20%. For the welds produced with greater currents, thinning not only exceeded 30%, but the welds were also observed to form buttons along the indentation. The possibility that in- dentations over 20 to 30% influenced weld fracture during tensile-shear testing, and especially during chisel testing, was there- fore raised.
Quality (Chisel) Test Results
Figure 7 compares chisel test results for spot welds also made with a 1200-lb (5.3 kN) force. The two types of fracture, interfacial (white squares) and button- pullout (black squares), are mapped as a function of current and weld time. Since each data point was generated by testing 3 to 12 welds depending upon the observed repeatability, results from more than 400 welds are summarized in Fig. 7. While both steels exhibit regions where only a single type of fracture occurred, the DP20 steel is characterized by having a region with the two types of weld fracture. This region is bound by currents between about 10 and 14 kA and weld times between about 15 and 25 cycles. As currents and/or weld times were increased, the percentage of each type of fracture varied from zero to one hundred, and vice versa. For DP20, interfacial fracture was also found to van- ish when expulsion started. Of these two observations, the first indicated that the chisel test was largely reproducible, and the second inferred that zinc expulsion was potentially related to weld fracture, as hypothesized previously.
To determine if zinc expulsion had in- fluenced the type of weld fracture, some welds were produced by varying currents and forces, which, as seen in Fig. 3, both affect expulsion remarkably well. The welds, made with selected forces of 600 lb (2.7 kN), 900 lb (4.0 kN), 1200 lb (5.3 kN), and 1700 lb (7.6 kN), were subsequently chisel tested, and correlations between type of weld fracture and occurrence of ex- pulsion were searched. The results of these tests are summarized in Fig. 8, where percentage of a given type of weld fracture
is represented as a function of current and force. For a given force, Fig. 8 shows that the occurrence of weld interfacial fracture gradually decreased with the current. More specifically, at 600 lb (2.7 kN), ex- pulsion of zinc started before the current could be raised high enough to fully elim- inate interfacial fracture. In contrast, at 1700 lb (7.6 kN), zinc expulsion was de- tected after interfacial fracture had been prevented. For the process conditions of Fig. 8, it became clear that zinc expulsion was not primarily related to weld fracture, and the common boundary in Fig. 7 for the regions of mixed fracture and expulsion could only be coincidental.
Figure 8 also points out that the high weld forces reduced the minimum current to prevent interfacial fracture. Indeed, once weld force exceeded 900 lb (4.0 kN), weld buttons were produced at smaller currents (10 kA vs. 12.5 kA at 600 lb, or 2.7 kN). This result confirmed that the level of applied force affected weld formation, presumably by affecting contact area and contact electrical resistance (Refs. 18–20, 24, 25). Since Fig. 8 also shows that weld current was negligibly affected by forces over 900 lb (4.0 kN), we concluded in agreement with another study (Ref. 19) that the contact resistances were practi- cally unchanged once a certain force, or pressure, is exceeded.
In Fig. 8, also note that the minimum weld diameter to eliminate interfacial fracture was consistently in the vicinity of 5.5 mm regardless of weld force and cur- rent. This 5.5-mm value, measured after chisel testing, was confirmed by some of the test data presented earlier. For cur- rents of 9 and 10 kA, Fig. 7 showed that weld interfacial fracture fully disappeared once weld times exceeded 25 cycles in DP18 and about 30 to 35 cycles in DP20. With these process parameters, Fig. 5 then revealed that interfacial fracture would not occur for weld button diameters over 5.5 mm in DP18 and 6.0 mm in DP20. Based upon this analysis and the data of Fig. 8, we therefore confirm that some minimum weld diameters could be recom- mended to prevent weld interfacial frac- ture, as further investigated in Part II of this work (Ref. 4).
As complementary remark resulting from a comparison of Figs. 6, 7, and 8, note that many of the tensile-sheared welds of Fig. 6 that fractured interfacially would have produced buttons if instead they would have been chisel tested. This demonstrates that the chisel test is less se- lective than the quasi-static tensile-shear test, and that any weld identified as a “good” weld by the tensile-shear test would also be a “good” weld if tested in production using the chisel test.
WELDING JOURNAL 179 -s