Metal Transfer in Pulsed Current Gas Metal Arc Welding

Transcription

1 WELDING RESEARCH SUPPLEMENT T THE WELDING JURNAL, JULY 993 Spnsred by the American Welding Sciety and the Welding Research Cuncil Metal Transfer in Pulsed Current Gas Metal Arc Welding A static frce balance analysis was used t estimate the melting rates f the electrdes during pulsed gas metal arc welding BY Y.-S. KIM AND T. W. EAGAR ABSTRACT. In rder t achieve ne drp per pulse peratinal cnditins with pulsed current GMAW, it is necessary t cntrl bth the drp size at the peak current and the melting rate f the electrde. In this study, a static frce balance analysis was used t predict the drplet size at the peak current and a weighted sum f the melting rates measured under Direct Current Electrde Psitive (DCEP) welding was emplyed t estimate the melting rate with pulsed current. Cmbining the static frce balance analysis and the weighted sum methd, a mdel is prpsed t predict the ptimal cnditins f ne drp per pulse peratin. The mdel is fund t be in gd agreement with the experimental results when the base current and the lad duty cycle are small. When the base current increases abve 220 A and the lad duty cycle exceeds 0% using.6-mm-diameter steel electrdes, the predictin f the mdel deviates significantly frm the experimental results. The discrepancy between the mdel and the experimental results is discussed Y. S. KIM is Assistant Prfessr, Department f Metallurgy and Materials Science, Hng Ik University, Seul, Krea, and T. W. EAGAR is C-Directr, Leaders fr Manufacturing Prgram, Richard P. Simmns Prfessr f Metallurgy, Department f Materials Science and Engineering, Massachusetts Institute f Technlgy, Cambridge, Mass. and is shwn t be due t tapering f the electrde tip at high welding currents. Intrductin Since the intrductin f pulsed current Gas Metal Arc Welding (GMAW-P) in 962 (Ref. ), this methd f welding has been used widely bth in mechanized welding and in rbtic welding. With pulsed GMAW, a stable spray metal transfer mde can be btained at lw average currents that wuld therwise prduce glbu lar transfer with large spradic drps. Pulsing leads t stable spray metal transfer and frmatin f a unifrm bead shape KEY WRDS Pulsed Current GMAW Mdeling Metal Transfer Electrde Melt Rate Static Frce Balance Weighted Sum Methd Drplet Size ptimum Pulsing Freq. Electrde Tapering Melting Rates with shallw penetratin. Recent imprvements in pwer supply designs using transistr r frequency cnverter cntrls als prvide better cntrllability f the prcess (Ref. 2). Pulsing the current intrduces additinal peratinal parameters, which include peak current, base current, peak pulse time, and base pulse time, in additin t the variables f DC welding, which include electrde extensin, welding current and welding vltage. These extra variables cause difficulty in selecting ptimum perating cnditins fr pulsed current welding. A trial-and-errr methd is ften used t determine these cnditins. Hwever, the basic physics f metal transfer in pulsed current welding needs t be understd in rder t mre successfully cntrl the prcess. There have been several attempts t analyze pulsed current welding theretically (Refs. 3-5). Samati (Ref. 5) predicted the theretical pulsing frequency by dividing the electrde melting rate by the mass f the drp, and shwed gd agreement between these predictins and experimental results. Hwever, this agreement is anticipated since there is a range f wrking slutins instead f a single-valued pulsing cnditin as shwn experimentally by Allum (Ref. 3). In this study, a theretical framewrk is described fr predictin f the range f ptimum pulsing frequencies. The methd uses a cmbinatin f the WELDING RESEARCH SUPPLEMENT I 279-s

2 u (a) glbular (b)spray (c) streaming with tapered tip (d) streaming STICK-UT 35 MM * STICK-UT 26 MM STICK-UT 6 MM ~> ' ' ' T" 2= WELDING CURRENT (AMPERE) Fig. Schematic representatins f metal transfer mdes as welding current increases frm A t D. Fig. 2 Melting rates f steel electrdes (.6 mm diameter) shielded with Ar-2%0 2. drplet size predicted frm the static frce balance thery, and the melting rate frm the weighted sum f melting rates at the peak and base currents equivalent t DC welding. This theretical mdel is then cmpared with the experimental results btained using steel, Ti-6Al-4V and aluminum electrdes. Theretical Framewrk fr Pulsed Current GMAW As ne increases the current during DC welding in argn-rich atmspheres, the metal transfer mde changes frm glbular t spray. With further current increases in the spray current regime, the ande spt increases in size until it begins t climb the sides f the slid cylindrical electrdes. The cndensatin heat prduced by the current n these vertical surfaces causes melting f the cylinder edges (Ref. ). At sufficiently high currents, this prduces a tapered slid electrde tip as seen in Fig.. In rder t btain ne liquid metal drp with a size similar t the electrde a \0 M/S EC Insufficient pulse frequency - diameter at every pulse, the perating cnditins must be such that significant tapering des nt ccur at the tip f the electrde. If tapering ccurs, the pulsed current prcess degenerates int streaming metal transfer mde (Ref. 7) and it becmes difficult t btain ne drp with each pulse. Amng the fur pulsing parameters, which include peak current, peak time, base current and base time, the pulsing frequency and the lad duty cycle were used as the peratinal parameters f interest instead f the mre cmmnly used peak time and base time. Pulsing frequency is defined as /(peak time + base time) and lad duty cycle as (peak time) / (peak time + base time) X 00 (%). The emplyment f pulsing frequency and lad duty cycle as the peratinal parameters eliminates sme f the cmplexity f adjusting the prcess. Fr instance, if the lad duty cycle is kept cnstant, the pulsing frequency can be changed withut affecting the average welding current, which may lead t a relatively cnstant electrde melting rate. In this manner, it is pssible t determine a range f ptimum pulsing frequencies at a cnstant electrde melting rate. In pulsed current GMAW, a theretical pulsing frequency is btained by dividing the electrde melting rate with current pulsing by the mass f ne drp : theretical pulsing frequency where: m ru,,e V",. ('> () inputs, is the electrde melting rate with current pulsing,vdrp(lp) is the predicted vlume f the drp at the peak current, and p,j is the density f the drp. The average melting rate fr a square wave current may be estimated as the weighted sum f the DC melting rate at the peak current and at the base current. m P, =^m(/ ) + (l- y 8)m(/ B ) (2) P: lad duty cycle m[l\. DC melting rate at peak current ne drp per pulse -+ excessive pulse frequency M/SE&^Sw 2 a < cc ^ \ ^\. ^^^~^ ~~ natural base current frequency naximum puls frequency i i i, i WELDING CURRENT (AMPERE) Fig. 3 The equilibrium drplet size frm a.6-mm-diameter steel electrde calculated frm the static balance thery at tw different argn gas speeds (0 m/s and 00 m/s) arund drps. pulsing frequency (fp) Fig. 4 Schematic diagram f weld current pulsing. 280-s I JULY 993

3 L& 00- cling water Z?\ I. shielding gas~*-i A current regulate J5 Fig. 5 verall layut f welding equipment PEAK CURRENT (AMPERE) Fig. 5 verall layut f welding equipment. Fig. 6 Theretical pulsing frequency fr steel electrdes shielded with Ar-2%0 2 as a functin f peak current. m(l b ): DC melting rate at base current As shwn in ur previus wrk (Ref. 8), the melting rate underges a transitin as the welding current increases as shwn in Fig. 2. This transitin is related t frmatin f the taper. Since fully develped tapers have less tendency t frm in pulsed current welding, the DC melting rate measured in the pretransitin regin has been extraplated t the peak current levels in rder t estimate the melting rate at the peak current. The drplet size in pulsed current welding may be determined at the peak current using the static frce balance mdel. Figure 3 shws the results f this calculatin. The higher the peak current, the smaller will be the drplet size. The details f this calculatin can be fund elsewhere (Ref. 7). When the pulsing frequency is increased abve the theretical pulsing frequency f Equatin with ther peratinal parameters held cnstant, nt every pulse can detach ne drp. In ther wrds, the drplet size and the melting rate remain the same; theretically it is impssible t prduce mre drps than predicted by the theretical frequency given by Equatin. Therefre, the theretical pulsing frequency is the theretical maximum pulsing frequency (TMPF) that shuld be applied t the system. n the ther hand, as the pulsing frequency is decreased belw the TMPF, each pulse can still prduce ne drp ver a limited range f lwer frequencies, but the drplet size becmes larger than the equilibrium drplet size at the TMPF. If the pulsing frequency is decreased further, drplet transfer frequency at the DC base current will eventually becme fasterthan the applied pulsing frequency. Hence the drplet transfer frequency at the DC base current sets the lwer limit f the ne drp per pulse regin. When the pulsing frequency is lwer than the limit, the drp will be detached in tw mdes: ne cntrlled by the base current and the ther cntrlled by the peak current. Therefre, within ne cycle f pulsing, several drps may be detached and the size f the drplets will becme nnunifrm. Figure 4 schematically shws the cncepts f the preceding paragraph. The drplet transfer frequency t pulse frequency rati n the vertical axis is defined as the actual drplet transfer rate divided by the applied pulsing frequency. When the drplet t pulse frequency rati is equal t ne, each pulse prduces ne drp. This is the ptimum pulsing frequency regin fr practical welding. When the drplet t pulse frequency rati is larger than.0, the natural frequency becmes larger than the pulsing frequency, hence insufficient pulse frequency is present. Finally, when the drplet t pulse frequency rati is less than.0, pulsing becmes s fast that nt every pulse can prduce a drp, hence the pulse frequency is excessive. Experimental Prcedures Mild steel (AWS E70s-3), aluminum ally (AA 00, AA5356), and titanium ally (Ti-6Al-4V) were used in the experimental prtin f this study. The shielding gases were pure argn and argn-2% xygen. The welding equipment included a cnstant current-type pwer supply, a transistrized current regulatr, and a vltage-cntrlled electrde feed with a lw inertia mtr. The pwer supply culd prvide a ttal ut r~ LAD DUTY CYCLE 5% BASE CURRENT IMA PEAK CURRENT 300 A PEAK CURRENT «w A PEAK CURRENT 500 A : PREDICTED MAX FREUENCY BASE CURRENT. 80 A PEAK CURRENT : 500 A LAD DUTY CYCLE : 5 % 3 DL G tt. C cc ' I ' I ' I ' I ' ' PULSING FREUENCY (/SEC) Fig. 7 ptimum pulsing frequency regins fr steel electrdes shielded with Ar-2%0 2. The base current was 80 A and the lad duty cycle was 5%. PULSING FREUENCY (/SEC) Fig. 8 Drplet size variatin in the range f ptimum pulse frequency fr steel electrdes with Ar-2%0 2 shielding. The peak current is 500 A. WELDING RESEARCH SUPPLEMENT I 28-s

4 Table Cnditins used fr Pulsed Current Welding Peak current (A) Base current (A) Frequency (Hz) Duty cycle % Mild steel 300,400,500 80,200,220,260 5 t 300 Aluminum 300, 400, 500 (00) 70,200 3 t 50 Ti-6AI-4V 300,400,500 30,200 3 t 00 put pwer f ver 200 A. The transistrized current regulatr used in this study can supply DC current with less than % ripple (Ref. 9). This system uses transistrs t cntrl the welding current and is capable f pulsing the DC current t a maximum f 5 khz fr small superimpsed signals. The equipment can cntrl pulsing parameters, peak current, base current, peak time and base time, independently frm a functin generatr included with the cntrller. An alumina tube was inserted int the cntact tip f a cmmercial welding gun leaving nly 5 mm fr cntact length rather than the nrmal cntact length f 24 mm. A transversing weld table was used s that the weld gun culd remain at a fixed psitin. Figure 5 shws the verall layut f the welding equipment. Analysis f metal transfer was perfrmed using high-speed videgraphy with a backlighted shadw graphic methd (Ref. 0). This methd excludes mst f the intense arc light and transmits mst f the laser light by a spatial filter that is placed at the fcal pint f the bjective lens. The high-speed vide camera is capable f prducing images at a maximum 000 full frame pictures per secnd (pps). The drplet transfer 5,0,20 rate was measured fr 0 s and an averaged drplet transfer rate fr each welding cnditin was calculated. The drplet size was measured frm the still image n the screen nce every secnd fr 0 s and averaged. The variatin in drplet size and frequency is estimated t be + 5% in mst cases. Melting rates f the electrde were measured using a tachmeter that was in cntact with the mving wire electrde. The utput vltage f the tachmeter and f the current shunt, which was filtered by a lw band pass filter, was recrded with a high-speed recrder. The ranges f peratinal variables fr pulsed current welding used in this study are shwn in Tablel. Based n the initial pulsing frequency, which was determined using predictins frm the theretical mdel develped in this study, the pulsing frequencies were changed in rder t determine the range f pulsing frequencies f ne drp per pulse. This range f the pulsing frequencies was judged primarily frm the recrdings f arc vltage and pulse current n a high-speed recrder and was later analyzed mre accurately using high-speed videgraphy. With the high-speed videgraphy, the drplet transfer frequency and the drplet size were determined. 0 0 Pulsed Current GMAW Steel Electrdes Effect f Peak Current with Figure 6 shws the TMPF calculated frm Equatin 6 as a functin f peak current at varius levels f base current. The TMPF increases as the peak current increases because the melting rate f the electrde increases due t the increase in average current and the decrease in drplet size. Using this TMPF as a reference frequency, a series f pulsing frequencies was tested experimentally in rder t determine the range f ne drp per pulse with ther pulsing cnditins remaining cnstant. Figure 7 shws the regins f pulsing frequency with a 80-A base current at three different peak currents: 300, 400 and 500 A. The lad duty cycle used was 5%. As seen in the figure, as the peak current increases, the width f the nepulse-ne-drp (PD) regin increases. When the peak current is 300 A, the PD regin is very narrw (4 t 6 Hz). When the peak current increases t 400 A, the range widens t 4 t 2 Hz and, finally, when the peak current is 500 A, the range expands t 4 t 38 Hz. This expansin f the PD regin as the peak current increases was als bserved at different base currents. This expansin f the PD regin, especially the increase f TMPF with peak current, is due t the increase in the electrde melting rate and the decrease in the drplet size as the peak current increases. The lwer bund pulsing frequency, 4 Hz, was nt affected by the peak current. This value agrees well with the measured natural DC drplet transfer frequency f 3.5 Hz at a current f 0.90 r BASE CURRENT: 80 A LAD DUTY CYCLE : 5% LU CXI g UJ I rr CD : PREDICTIN : EXPERIMENTAL PEAK CURRENT (AMPERE) vn WELDING CURRENT (AMPERE) Fig. 9 Cmparisn between drplet size frm the static frce balance thery and minimum drplet size in pulsed current welding fr steel electrdes with Ar-2%0, shielding. Fig. 0 Melting rate f steel electrdes at three different peak currents with Ar-2%0 2 shielding. 282-s I JULY 993

5 LAD DUTY CYCLE 5% BASE CURRENT 200 A _. PEAK CURRENT 400 A 0 PEAK CURRENT SC A i : PREDICTED MAX FREUENCY < tr CL a a. - T PULSING 3CJ FREUENCY (/SEC) Fig. Pulsing frequency regins f steel electrdes at a base current f 200 A with Ar-2%0 2 shielding. PULSING FREUENCY (/SEC) Fig. 2 Pulsing frequency regins f steel electrdes at a base current f 220 A Ar-2%0 2 shielding. 80 A. Hwever,the TMPF, which are indicated by the arrws in the figure, d nt cincide with the measured maximum pulsing frequency. Figure 8 shws the variatin f the drplet size as the pulsing frequency changes at a peak current f 500 A. As the pulsing frequency increases, the drplet size decreases until it reaches a minimum value. This crrespnds t the equilibrium drplet size at which the detaching frces at the peak current are just equal t the retaining surface tensin frce. Figure 9 cmpares the equilibrium drplet size frm the static frce balance thery with the experimentally measured minimum drplet sizes at different peak currents. The predictin and the experimental results agree with errrs f less than + 0%. These results shw that the static frce balance thery can be used t predict the drplet size at varius peak currents. The discrepancy f the TMPF frm the experimental results may be caused by tw pssibilities as ne can see frm Equatin : either the equilibrium drplet size is in errr r the melting rate predicted frm the weighted sum methd is in errr, r bth. Since the predicted equilibrium drplet size agrees reasnably well with the experimentally measured minimum drplet size, the actual melting rate during pulsed current welding was measured t cmpare with the melting rate predicted frm the weighted sum methd. The melting rates measured at different peak currents alng with the melting rates predicted frm the weighted sum methd f Equatin are shwn in Fig. 0. As mentined in the previus sectin, the melting rate fr the peak current is calculated frm the curve extraplated frm the pretransitin melting rate curve f the DCEP welding prcess. As seen in the figure, the measured electrde melting rate is higher than the calculated melting rate predicted by Equatin. When the increased melting rate under current pulsing is used in Equatin, the TMPF at 80 A base current, 500 A peak current, and 5% lad duty cycle is calculated t be 30 Hz, which is clser t the experimentally bserved 37 Hz. Therefre, this increased melting rate under pulsed current welding must cause a significant prtin f the discrepancy between the TMPF and the measured maximum pulse frequency. Effect f Base Current Figures,2 and 3 shw the experimental results f the drplet t pulse frequency rati as a functin f the pulsing frequency at base currents f 200, 220 and 260 A, respectively. When the base current is 200 (Fig. ) and 80 A (Fig. 7), the predicted TMPF lie within the PD regin. As the base current is increased t 220 A as in Fig. 2, the TMPF starts t shift utside f the PD regin. With peak currents f 400 and 500 A, the TMPF are within the regin f PD, but with a peak current f 300 A the TMPF becmes smaller than the lwer frequency f the PD regin. When the base current is increased t 260 A, the TMPF f all peak currents becmes smaller than the measured lwer limit frequency f the PD regin. These large deviatins f theretical predictin frm the experimental measurements can be explained frm the results f the drplet size measurements in ur previus studies (Ref. 7). Arund 2 0 A in DCEP welding the measured drplet size becmes significantly smaller than the drplet size predicted by the static frce balance thery due t - \ \ \ * LAD DUTY CYCLE 5* BASE CURRENT 260 A PEAK CURRENT SC A ; PREDICTED MAX FREUENCY 5 LAD DUTY CYCLE 5% BASE CURRENT 80 A BASE CURRENT 200 A BASE CURRENT 220 A t 500 A X cr rv ' PULSING, FREUENCY (/SEC) \ 0- DC ~T~ 250 PEAK CURRENT (AMPERE) Fig. 3 Pulsing frequency regins f steel electrdes at a base current f 260 A Ar-2%0 2 shielding. Fig. 4 The minimum drplet size f steel electrdes at three different base currents Ar-2%0 2 shielding. WELDING RESEARCH SUPPLEMENT I 283-s

6 3 cr cv PEAK CURRENT (AMPERE) Fig. 5 Partial tapering at the tip f the steel electrde when shielded with Ar-2%0 2. The base current is 220 A and the peak current is 400 A. Fig. 6 The minimum drplet size f steel electrdes at three different peak currents when shielded with Ar-2%0 2. The base current is 80 A and the lad duty cycle is 0%. tapering f the electrde. Therefre, with pulsed welding cnditins in which tapering f the electrde ccurs, the TMPF, which are calculated by the drplet size predicted frm the static frce balance thery, will be smaller than the measured drplet transfer frequency. Figure 4 shws the drplet size measured at different base currents when tapering ccurs as in Fig. 5. It can be seen that the minimum drplet size is smaller than that predicted by the DC (nntaper) predictin at the 220 base current. Thus, it is believed that it is the frmatin f a taper that causes the predicted TMPF t be smaller than that measured experimentally. The tendency fr tapering f the electrde increases as bth base currents and peak currents increase. Effect f Lad Duty Cycle When lad duty cycle is increased t 0%, tapering f the electrde ccurs even at lw base currents. Fr instance, with 0% lad duty cycle, tapering is bserved at a base current f 80 A and a peak current f 400 A. With such a high lad duty cycle, the electrde tapers during the peak current perid and des nt return t a cylindrical shape immediately after the current is lwered t the base current. This phenmenn is especially easy t bserve at pulsing frequencies near the lwer bundary f the ptimum pulsing frequency regin. Since a small amunt f tapering can expand the PD regin by creating decreased minimum drplet sizes, the tapering f the electrde can be beneficial if the degree f tapering is small enugh such that drplet sizes similar t the electrde size can be btained. Figure 6 shws the measured decrease in drplet size due t the partially develped taper seen in Fig. 5. When there is partial tapering f the electrde, the PD regin is increased significantly as shwn in Fig. 7. The pulse frequency wrking range at a base current f 80 A and a peak current f 400 A with 0% lad duty cycle is apprximately twice as wide as that with 5% lad duty cycle, which prduces n partial tapering. As the lad duty cycle is further increased up t 20%, the PD regin increases significantly because the drplet sizes are further reduced by the well-develped taper n the electrde. LAD DUTY CYCLE 0% BASE CURRENT 80 A A PEAK CURRENT 30 A» PEAK CURRENT 4D0 A PEAK CURRENT 500 A 4 PREDICTED MAX FREUENCY < rr rr PULSING FREUENCY (/SEC) DC i! [ ] ' i ' i ' r PULSING FREENCY (/SEC) Fig. 7 The pulsing frequency regin f steel electrdes shielded with Ar-2%0 2 at three different peak currents. The width f the ptimum pulsing frequency regin has increased significantly at 0% lad duty cycle. Fig. 8 The pulsing frequency regin f steel electrdes at 20% lad duty cycle Ar-2%0 2 shielding. The base current is 220 A. 284-s I JULY 993

7 ZC c DC -. "~Tg- CC TAPER -^ - _ FRMED I I I I I ( WELDING CURRENT (AMPERE) Fig. 9 A fully develped taper at peak current during pulsed GMAW. The steel electrde is shielded with Ar-2%0-,. Fig. 20 The minimum drplet size f steel electrdes shielded with Ar-2%0 2 under different pulsing cnditins. Depending n the degree f tapering, the drplet size frms ver a range f drplet size. Figure 8 shws the PD range with base current f 220 A and 20% lad duty cycle. The minimum drplet sizes measured with pulsed current welding are as small as the drplet sizes btained with streaming transfer in DC welding. Therefre, it is nly pssible t achieve ne pulse per drp at very high pulsing frequencies under these high duty cycle cnditins. In this case, there is n advantage f using pulsed current welding since the drplet size is n lnger similar t the drplet size f prjected spray transfer; ne culd use DC streaming transfer just as well. In pulsed current welding, it is bserved that there are varius taper shapes depending n the pulsing parameters. Fr example, there is a fully develped taper as seen in Fig. 9 and a partially develped taper as seen in Fig. 5. As the base current and lad duty cycle increase, the tapering becmes larger and will decrease the equilibrium drplet size. Figure 20 shws the minimum drplet sizes measured under different pulse cnditins. As seen in this figure, when there is n taper, the drplet size is very clse t the theretical value predicted by the static frce balance thery. When the taper starts t develp with increases in the base current and/r the lad duty cycle, the drplet size begins t decrease depending n the degree f tapering. Finally, when the pulsing cnditin develps a fully develped taper (such as with the 20% lad duty cycle and 220-A base current), the drplet size appraches the drplet size f streaming transfer in DCEP welding. The frmatin f a taper causes varius prblems in analyzing pulsed current welding; hwever, tapering prvides ne imprtant advantage f prviding a wider range f the ptimum pulsing frequency regin. Therefre, by cntrlling the develpment f the taper, the PD range can be widened, which leads t easier cntrl f the prcess. The additin f helium t argn shielding gas has been knwn t suppress the tendency t frm a taper (Ref. ). Thus, the use f helium gas mixtures is expected t widen the range f frequency ver which PD can be btained. Effect f Materials Ti-6AI-4V and Aluminum 00 electrdes were als tested. With Ti-6Al-4V electrdes, the general trend f the PD regin is the same as with the steel electrde. As shwn in Fig. 2, the ptimum frequency regin increases with the peak current. Als, the minimum drplet size decreases with the peak current but is larger than the drplet size predicted frm the static < rc A frce balance thery, which may reflect the repulsive frces experienced in this prcess. With Ti-6Al-4V electrdes, tapering was nt bserved up t 240 A with DCEP welding. Hwever, Fig. 22 shws that a taper frms during pulsed current welding at 200-A base current, 500-A peak current, with 0% lad duty cycle, and 20-Hz pulsing frequency. This shws that tapering f the electrde may ccur with argn shielding in materials ther than steel, but the nset current f tapering is dependent n the prperties f the material. With the aluminum electrde, the PD range was measured fr 300-, 400- and 500-A peak currents using a lad duty cycle f 0%. With these cnditins, it was nt pssible t find any satisfactry ptimum pulsing frequency regin. With mst f the cnditins, tapering ccurs, leading t streaming transfer. When tapering des nt ccur, secndary small drps ccur after primary drplet detachment as shwn in LAD DUTY CYCLE 0% BASE CURRENT 30A PEAK CURRENT 300 A PEAK CURRENT 400 A PEAK CURRENT 500 A PREDICTED MAX FREUENCY PULSING FREENCY (/SEC) Fig. 2 The pulsing frequency regin with TI-6AI- 4V electrdes shielded with pure argn. WELDING RESEARCH SUPPLEMENT I 285-s

8 Fig. 22 Tapering f TF6AI-4V electrde with argn shielding. A Beginning f the tapering; base current is 200 A and the peak current is 500 A. The lad duty cycle is 0%. an established taper f the electrde. The Fig. 23. This may be due t the fact that the peak currents used in this study were t high fr aluminum. The lw surface tensin f aluminum causes the detaching frce t create t much disturbance in the liquid drp. With lwer peak currents and with a lwer lad duty cycle, there may be a range f ptimum pulsing frequencies fr aluminum welding. Further bservatins with Pulsed Current GMAW Frm the bservatins made in this study, several imprtant aspects f the pulsed current welding prcess can be identified. Firstly, peak current has the mst significant effect n the PD range, as seen in Fig. 7. In general, the higher the peak current, the wider the Fig. 23 Secndary metal transfer with aluminum electrdes in pulsed current welding. The shielding gas is pure argn. PD range. Hwever, when peak current is increased t much, tapering f the electrde will ccur, leading t a streaming transfer mde in which the drplet size is t small t cntrl. Tapering may be suppressed by using a shielding gas cnsisting f Ar-He mixtures. Secndly, when welding with steel electrdes using carbn dixide as a shielding gas, the applicatin f pulsed GMAW will nt prvide any advantages in cntrlling drplet size. Since the drplet size remains nearly the same and the mde f metal transfer is repelled transfer (Ref. ), pulsing f current will nt prduce prjected spray transfer when welding steel electrdes shielded with carbn dixide. Thirdly, when welding with steel electrdes using helium as the shielding gas, pulsed current GMAW may prduce prjected metal transfer in the nrmal DC range f repelled glbular transfer. The repelled metal transfer mde at lw welding current transfrms int the prjected spray transfer mde as welding current increases. Therefre, if the peak current used is greater than the transitin current f repelled-prjected transitin, pulsed current GMAW will prduce a prjected transfer mde. The same reasning can be applied when welding with titanium electrdes shielded with argn, which exhibit the same transitin phenmenn as the welding current increases. Cnclusins A theretical mdel f pulsed current welding is develped t predict ranges f ne pulse per ne drp pulse frequency. Experimental results cnfirm this apprach. The width f the ptimum pulsing frequency regin increases as the peak current increases. This is due t the fact that the range f drplet sizes available and the melting rate increase as the peak current increases. The static frce balance thery can predict the drplet size at a given peak current prvided that there is n significant tapering at the tip f the electrde. The melting rates under pulsing current cnditins are greater than melting rates calculated using a weighted sum f the melting rate (fr DC currents) at the peak current and at the base current. The wrkable ranges f base current and lad duty cycle can be expanded when tapering f the electrde can be suppressed. This may be achieved by adding helium and/r carbn dixide t the argn gas. 286-s I JULY 993

9 Acknwledgments This research was funded by a grant frm the United States Department f Energy under cntract number DE- FG02-85ER-333. References. Needham, J. C Cntrl f transfer in aluminum cnsumable electrde welding. Physics f Welding Arc, The Institute f Welding, Lndn, pp Shimada.W., and Ukai, J. Effects f pulsed current cntrl n welding quality imprvement. IIW Dcument # XII-B uintin, L., and Allum, C. J Pulsed GMAW: interactin between prcess parameters part II. Welding and Metal Fabricatin, Vl. 4, pp uintin. L., and Allum, C. ) Pulsed GMAW: interactin between prcess parameters part I. Welding and Metal Fabricatin, Vl. 3, pp Samati, Z Autmatic pulsed MIG welding. Metal Cnstructin, Vl. 8, N., pp. 33R-44R. 6. Lesnewich, A Cntrl f melting rate and metal transfer in gas-shielded metal-arc welding. Part I Cntrl f electrde melting rate. Welding lurnal 37(8):343-st353-s. 7. Kim, Y. S., and Eagar, T. W. Analysis f metal transfer in gas metal arc welding. Welding Jurnal 7 (6):269-s t 278s.. 8. Kim, Y. S., and Eagar, T.W Temperature distributin and energy balance in the electrde during GMAW. Prc. f Trends in Welding Research, Gatlinburg, TN. 9. Eickhff, S. T Gas-metal arc welding in pure argn. M.S. thesis, MIT, Cambridge, Mass. 0. Allemancl, CD., Scheder, R., Ries, D.E., and Eagar, T.W A methd f filming metal transfer in the welding arc. We/ding Jurnal 64(): Kim, Y. S Metal transfer in gas metal arc welding. Ph.D. thesis, MIT, Cambridge, Mass. Z LU. _ tu > UJ Plan nw t attend! Cmmercializatin f Advanced Jining Technlgy Thrugh Industry-Gvernment Partnering September 27-28, Bulder, Clrad This cnference, spnsred by the American Welding Sciety, will describe the mechanisms (CRADA, licensing, etc.) by which the gvernment can wrk with industry; make the welding industry aware f gvernment spnsred research that is available fr cmmercializatin; ffer the cncept f teaming (gvernmentindustry partnerships) t address remaining technical barriers; and, btain input frm industry (future R&D suggestins, cmments n gvernment prgrams). Fr a cmplete brchure, fill ut the frm belw and mail t: AWS Cnferences, 550 N.W. LeJeune Rad, Miami, FL 33 26, r call , x 278. Please send me infrmatin n the cnference, "Cmmercializatin Technlgy Thrugh Industry-Gvernment Partnering" f Advanced Jining Name Cmpany Address City Telephne, State Fax Zip. WELDING RESEARCH SUPPLEMENT I 287-s