Characterization of Spot Welding Behavior by Dynamic Electrical Parameter Monitoring The use of two parameters—dynamic resistance and critical expulsion energy—is proposed for controlling resistance spot welding BY D. W. DICKINSON, J. E. FRANKLIN AND A. STANYA ABSTRACT. A program was undertak­ obtained by comparing dynamic a continuous or dynamic monitoring en to develop techniques for studying resistance curves for various high system is needed. This system would the resistance spot welding process. A strength steels. monitor and record instantaneous dynamic electrical parameter monitor­ Expulsion is analyzed in terms of changes in the electrical functions ing device was designed to simulta­ power curves integrated to obtain during welding. neously record the instantaneous val­ total weld energy input. It was found Several authors156 report contin­ ues of voltage, current, power, and that expulsion occurs when the total uous variations in the electrical param­ resistance during spot welding. The useful energy into the weld exceeds a eters during welding of mild steel data obtained using this technique critical value. It is proposed that these sheet to be typified by the sketches have been analyzed in terms of the two parameters, dynamic resistance presented in Fig. 1. After approximate­ relationships of these parameters to and critical expulsion energy, can be ly the first cycle, an increase in voltage the phenomena occurring during the used to control the spot welding pro­ across the welding electrodes and a formation of a spot weld, to the effects cess. decrease in current flowing through of changes in welding variables on the weld zone occurs until a "peak electrical parameters, and to the Introduction region" is reached. Throughout the effects of variations in steel composi­ remaining portion of the weld cycle, tion and properties. It has long been known that the the voltage decreases to a constant The phenomena occurring during process of spot welding occurs value while the current increases to a spot weld formation (surface break­ through the localized melting and constant value. down, asperity collapse, heating of the coalescence of a small volume of These changes in voltage and cur­ work pieces, molten nugget formation, material due to the heating caused by rent can also be represented as instan­ nugget growth, and mechanical col­ the passage of electric current. This taneous or dynamic resistance. Several lapse) can be understood through heating is equivalent to the product of authors1,56 have calculated the dy­ analysis of dynamic resistance curves. the current squared times the total namic resistance changes throughout If expulsion occurs during a spot weld, resistance of the material to be spot the welding cycle by dividing the it also is readily detected from this welded. During the process of spot instantaneous voltage by the instanta­ parameter. A generalized resistance welding, however, both the current neous current (R = E/l). This dynamic curve consists of an initial very rapid and the resistance continually change resistance trace is also plotted in Fig. 1. drop in resistance to a minimum value as the material is heated and melts. After an initial drop, it too rises to a within the first few cycles of current Until recently, spot weld character­ peak in the first portion of the weld flow. This is followed by a rise to a istics have been determined by mea­ cycle, dropping off later in the cycle. maximum and finally a gradual de­ surement of the initial or static The appearance of this dynamic crease. resistance and the initial current surge resistance trace and hence the basic Dynamic resistance measurements at the onset of welding, or an average welding mechanisms can be altered by RMS current value obtained during changes in welding variables such as can be used in obtaining a better 6 understanding of lobe curves. The welding. Recent papers,'" however, weld time, electrode force, and overall shape of the dynamic resistance curve have indicated that in order to fully weld current. Bhattacharya and An­ 1 for welds near the lobe boundary is characterize the spot welding process, drews have made the following correlated to the observed characteris­ observations: tics of these welds. The effects of Based on a paper presented at the AWS 1. With low current values, the changes in welding current and elec­ 60th Annual Meeting held in Detroit, resistance trace does not show a prom­ trode force can also be related to the Michigan, during April 2-6, 1979. inent peak; with increased welding dynamic resistance curves. Spot weld­ current, however, a well-defined peak D. W. DICKINSON, I. E. FRANKLIN and A. ability varies as a function of steel STANYA are with the Research Center, appears. chemistry, and an improved under­ Republic Steel Corporation, Independence, 2. With higher current values, the standing of these material effects is Ohio 44131 maximum or peak resistance occurs 170-sl JUNE 1980 earlier during a weld. changes in these basic electrical mea­ 3. For welds produced with high surements thus confirming some ear­ currents, the resistance trace indicates lier speculations and developing new a lower value of resistance towards the insight. completion of weld time. 3. Characterize the dynamic electri­ 4. A splash weld (expulsion), caused cal measurements over the range of by excessive welding current, is char­ weld parameters giving acceptable acterized by a sudden step in the spot welds (i.e., over the lobe curve NO EXPULSION resistance trace. area) as a first step to using these Savage6 has attempted to explain the electrical measurements for control of shape of the resistance curve on theo­ the spot welding process. retical grounds. When analyzing the WELD TIME, CYCLES resistance after each half cycle of cur­ Fig. 1— Schematic representation of dy­ Materials rent, he noted rectification which is namic electrical parameters during spot believed to indicate the presence of welding of mild steel' The spot weldability of several heats oxide films for times as long as 6 to 8 of plain carbon and high strength hot cycles on some welds. He has also rolled and cold rolled steels was inves­ attributed the rise in resistance after •. • . " :• • tigated. Specific chemical composi­ the initial few cycles to be the result of ' • ' -: •:•' ' ' tions of the materials for this investiga­ LU a EXPULSION _l tion are listed in Table 1. These include the material heating. The latter de­ o ;:\ LEVEL >- plain carbon, rimmed, renitrogenized, crease in resistance appears to coin­ o Cb-added, Cb + V-added and Cb 4- cide with the growth of the fused UJ A^AJ -NUG zone. The sudden drop in resistance 2 Mn-added steels. Also, some welds were made on a Type 304 stainless upon expulsion was attributed to the o _i SMALL OR \ steel. increase effective contact area pro­ LU BRITTLE NUGGETS'fe- vided by the expelled metal trapped Except where specifically noted, between the sheets. V: III.•^-^yk most of the data discussed are related These observations lay a pathway for to the welding of a 0.032 in. (0.8 mm) RMS SECONDARY CURRENT, AMPS effective utilization of dynamic elec­ thick plain carbon aluminum-killed trical measurements in monitoring and Fig. 2—Typical spot weld lobe curve material (material A in Table 1). control of spot welding. However, before such control can be utilized, a Equipment characterization of these electrical hibit brittleness during tearing and are All welding was performed on two measurements throughout the entire likewise considered unacceptable. Taylor-Winfield air-operated electric range of acceptable spot weld parame­ Only welds made with weld currents resistance spot welding machines. ters must be made. and times lying within the lobe area are acceptable. These are equipped with Weldtronic One method for determining the Once the lobe curve is known, char­ controllers for electronic phase con­ range over which acceptable spot trol (heat adjustment) and Duffers cur­ welds are obtained on plain carbon acterization of the dynamic electrical measurement can be made at critical rent analyzers for simultaneous mea­ sheet is through the use of spot weld surements of weld time and secondary lobe curves. A typical spot weld lobe areas on the curve. Such critical areas are a) around the lower limit line, b) at RMS weld currents. Electrode forces curve is presented in Fig. 2. A lobe expulsion, and c) within the accept­ were accurately set at the recom­ curve is a graphical representation of 7 able region. mended force for the sheet thickness ranges of welding variables over which being welded and checked several acceptable spot welds are formed on a times throughout the welding program specific material welded with a prese­ Investigation Procedure using a universal force gauge. lected electrode force. The lobe curve Objective is determined by making spot welds Electronic Circuits for Monitoring The objectives of this investigation using different weld time/weld cur­ the Dynamic Electrical Parameters rent combinations. Welds made with were to: currents and/or times exceeding the 1. Build a spot weld monitoring sys­ A schematic of the circuits used to upper curve experience expulsion on tem capable of measuring dynamic develop the welding parameters is welding and are, therefore, considered values of voltage, current, resistance, shown in Fig. 3. unacceptable. Welds made with cur­ and power. The instantaneous voltage across rents or times below the lower curve 2. Relate the phenomena occurring the electrodes is sensed by