-
The Role of Pulsed GTA Welding Variables in Solidification and
Grain Refinement
The utility of the trailing thermal cycles in pulsed GTA welding
is demonstrated, but improved control of the solidification process
is shown to be minimal
BY D. W. BECKER AND C. M. ADAMS, Jr.
ABSTRACT. The effects of pulsed current GTA welding parameters
on solidification structure, segregation, grain size, and
mechanical properties were investigated. Bead-on-plate welds in t i
tanium alloys and mild steel sheet were employed.
Although moderate changes in so-lidification mode and cell size
were noted in t i tanium, no change in mechanical properties was
observed. The minimal changes in the solidifica-t ion structure
indicated that the pulsed current welding process did not
significantly enhance control of the GL/R ratio which controls the
solidification mode. Microprobe anal-ysis indicated, further, that
the intra-cellular segregation did not change, supporting the lack
of an observed change in mechanical properties.
Effects of pulsing conditions on grain size were nonexistent in
t itanium but very significant in the mild steel welds. Grain
refinement in the mild steel welds was attributed to mult iple
cycling of the fusion zone into the austenite region.
Introduction
The pulsed gas tungsten-arc welding (P-GTAW) process is a
variation of GTA welding which involves cycling the welding current
from a high level to a low level at a regular frequency. The
frequency of cycling and the two levels of current are selected on
the
Paper presented at the AWS 60th Annual Meeting held in Detroit,
Michigan, during April 2-6, 1979.
D. W. BECKER is a Metallurgist, loining Technology Group, Air
Force Materials Laboratory, Wright-Patterson Air Force Base, Ohio,
and C M. ADAMS, Ir., is a Private Consultant, Cincinnati, Ohio.
basis of their advantages for a given application. The peak
current is gener-ally selected to give adequate penetra-t ion and
bead contour whi le the low level of current is set at a level
suffi-cient to maintain a stable arc. These parameter selections wi
l l result in indi -vidual spot welds if the cycling frequency is
low. A welded seam results when the travel speed is selected to
permit the spot welds to overlap at least half of the spot
radius.
Extensive research has been per-formed on this process both in
the Soviet Union and to a lesser extent in the free wor ld.
Numerous advantages of this process have been identif ied in these
investigations. Reported physical advantages would include improved
bead contour, increased process con-trol, reduced distortion and
others. Metallurgical advantages which have been reported include
increased con-trol of weld microstructure, reduced grain size and
refined solidification structure.
The purpose of the investigation
described in this paper was to examine the metallurgical
advantages of the P-GTAW process in the welding of sheet material,
wi th the specific objec-tive of determining the feasibility of
using high current pulses to achieve increased control of the
fusion zone microstructures.
Background
A brief and general discussion of solidification theory1 as it
pertains to welding is appropriate at this time since certain
aspects of solidification theory wi l l be referred to in the
discus-sion section. This section is not intended to review
solidification theo-ry in detail but rather to highlight those
areas which wi l l be employed to evaluate the results of this
study.
Examination of the phase diagram illustrated in Fig. 1 shows
that the first material is to solidify from a melt wi l l have a
composit ion of kC0, where k is defined as CS/CL. The first
material to solidify rejects solute into the l iquid (for k < 1)
resulting in a solute di f fu-
FORMATION OF INTIAL TRANSIENT AND SOLUTE BOUNDARY LAYER
UJ
i -< cc LO
a.
Co/K
Co
KC0
SOLID
•
X " (
1 LIQUID
1
DISTANCE
CONCENTRATION-
Fig. 7—The development of the initial transient and the solute
boundary layer is illustrated with an ideal eutectic phase
diagram
W E L D I N G RESEARCH SUPPLEMENT | 143-s
-
2C0/K
<
UJ cr z>
4 CE
Q.
LfJ
ACTUAL TEMPERATURE
\ / ^ -L IQUIDUS
"/ 1 ,CONSTITUTIONALLY SUPERCOOLED REGION
0 DISTANCE FROM INTERFACE—•
Fig. 2-An illustration of constitutional supercooling is given.
The solid-liquid interface is located at the zero coordinate on the
abscissa
sion boundary layer in the l iquid at the interface. This
rejection of solute to the liquid causes the liquid solute
concentration at the solid-l iquid inter-face to increase to a
value of C0/k; at which composit ion the solid forming is of a
concentration of C„. No further solute bui ld-up results and a
steady-state condit ion is attained.
During the steady-state period the solute diffusion layer in the
l iquid is continually pushed ahead of the solid-ification front.
The width of this di f fu-sion boundary layer is inversely
pro-portional to the growth rate of the solid, R. If the growth
rate is changed after steady-state has been attained, the layer
must readjust. A solute-rich band wil l be observed in the solid
where a growth rate increase occurred, or a solute-lean band if the
rate was decreased. Solute banding can also occur from changes in
the convection boundary layer due to a washing of a portion of the
solute diffusion layer into the melt.
The planar solid-l iquid interface can become unstable if the
slope of the liquidus temperature at the interface is less than the
slope of the actual ther-mal gradient in the l iquid, G,. This
situation, which is depicted in Fig. 2, is referred to as
constitutional super-cooling. Increasing constitutional
su-percooling causes the planar interface to break down
successively to cellular and then dendritic growth. It was
previously pointed out that increasing growth rate decreases the
width of the solute boundary layer. The narrower boundary layer
translates into a steep-er l iquidus temperature which would in
turn push the interface condit ion toward instability. Steep
thermal gra-dients can be seen to suggest interface stability.
Interface stability is therefore, conveniently described by the
ratio,
SHPI II W .
Fig. 3—Current wave shape resulting from welding conditions of
200 A and a frequen-cy of 0.46 cycles/s. Major divisions on the
ordinate represent 120 A and on the abscis-sa they correspond to
0.5 s. Top number, or circuit inductance, is varying from "A" to
"D"
GL /R. When the planar interface breaks
down, solute is not only redistributed in the growth direction
but also normal to the growth direction. This redistribution is
described by an equa-t ion derived by Bower, Brody, and
Flemings:1
C = kC, fe ak t (k-1) + (1 --—)CKr k-1
where a = —DLG/MLRC„, fs = frac-t ion solid, D L = diffusion
coefficient in the l iquid, G = thermal gradient, and M, = slope of
liquidus curve.
All of the terms in this equation are material constants except
for G,. and R. The concentration difference between cell core and
cell boundary is therefore determined by the ratio G,,/R. The size
of the cells, however, is proportional to (GR).
To relate this general solidification theory to a
constant-current welding situation it is necessary to look at the
physical situation as it exists in the welding operation.
Immediately after the establishment of an arc, there is a period in
which the molten pool wi l l increase in size. As the electrode
trav-els over the workpiece a steady-state condit ion is
reached—that is, the shape of the isotherms and their location
relative to the electrode location wi l l not change. The
development of a solute boundary layer has already taken place and
is also in a condit ion of steady-state.
Fig. 4—Current wave shape for welding conditions ot 200 A and a
frequency of 20 cycles/s. Major divisions on the ordinate represent
120 A and on the abscissa they correspond to 0.01 s. Tap number, or
circuit inductance, is varying from "A" to "D"
The solid-l iquid interface at this t ime wi l l be located at
the isotherm corresponding to the solidus of the alloy. Variations
in the travel speed or the welding current w i l l cause a change
in the rate of movement of the solidus isotherm. This change in
veloc-ity of the solidus isotherm upsets the dynamic steady-state
condit ions in the solute boundary layer resulting in solute
banding, or fluctuations in solute concentration.3
As previously pointed out, the mode of solidification is
determined by the ratio of G,./R. Although the solidifica-t ion
rate in welding is constant when steady-state has been attained, it
wi l l vary w i th position on the pool peri-meter. The slowest
growth rate occurs at the side of the weld pool adjacent to the
base metal. It increases around the trailing edge of the pool to a
peak speed equal to the electrode travel speed at the weld
centerline. Also changing around the trailing perimeter of the weld
pool is the thermal gradient in the l iquid.
The gradient is largest at the edge of the pool and decreases to
a minimum at the weld centerline on the trailing edge of the pool.3
Therefore, the mini-mum value of the GL /R ratio on the edge of the
weld pool is located at the weld centerline; the maximum value is
located adjacent to the base metal on the side of the weld pool.
This indi-cates that there would be an increas-ing tendency for
interface instability as the solidification front approaches
the
144-s I M A Y 1979
-
Table 1-VVeld
Sample number
146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161
162 163 164 165 166 167 168 169
ing Conditions
Travel speed, ipm
1.2 1.2 1.2 1.2 1.2 1.2 1.2 9 9 9 9 9 9 9 9 9
15 .15 15 15 15 15 15 15
For Evaluation Of Sol
Fraction of t ime at
peak current
1.0 0.67 0.67 0.33 0.33 0.20 0.33 1.0 0.67 0.65 0.67 0.33 0.35
0.33 0.200 0.200 1.0 0.67 0.67 0.67 0.33 0.33 0.33 0.19
Lite Distribution In
Spot spacing, in.
-.005 .05 .005 .05 .005 .001
-.15 .05 .008 .15 .05 .008 .15 .05
-.15 .05 .013 .15 .05 .013 .15
Bead-On-
Frequency cycles/s
0 4 0.41 4 0.41 4
20 0 1 3
20 1 3
20 1 3 0 1.67 5
20 1.67 5
20 1.67
Plate Welds
Peak current, A
45 63 63
104 104 141 104 100 140 142 140 231 224 231 313 313 111 155 155
155 256 256 256 353
Heat input, k j / i n .
20 20 20 20 20 20 20
6 6 6 6 6 6 6 6 6 4 4 4 4 4 4 4 4
Probe trace length, microns
150 150 910 200
1020 315 100 360 150(5) '" 360 410 150(5) '" 920 260 150(5) '"
930 150 150(5) '" 920 360 150(5) '" 920 360 150(5) '"
" " Indicates 5 traces evenly spaced over one spot spacing.
centerline of the weld. Under condi-tions of high travel speed
it has been possible to reduce the value of G,./R sufficiently at
the weld centerline to obtain an equiaxed structure in the center
of the weld in some stainless steels and low alloys steels.3
Experimental Procedure
Equipment
The welding in this program was accomplished wi th a 400 ampere
(A) constant-current power supply. The controls included low
frequency puls-ing controls capable of pulsing from approximately
0.5 to 30 cycles per second(s). The welding was carried out in a
welding fixture wi th a grooved copper backing bar and copper
hold-
down tabs. The pulse current wave shape was
investigated before undertaking this program since this was
known to affect heat input, and therefore, growth rates and thermal
gradients. Several tap settings were available on the power supply
and were found to affect the wave shape by varying the circuit
inductance.
Figures 3 and 4 show the extremes of the current shape
variations wi th considerable ripple observed at all settings. It
is quite apparent from the two that the wave shape rapidly
dete-riorates to a saw tooth configuration which can overshoot or
fall short of the set value of current depending upon the time
constant of the circuit. Throughout the rest of the study the
Table 2-Fu
Sample number
173 175 177 179 181 183 185 191 187 193 189 195 203 199 201
II Penetration Bead-O
Frequency, cycles/s
0.0 0.41 0.41
20 20 0.0 1.33 1.33
20 20 0 3 3
20 20
n-Plate Welds
Fraction of Time
at l„
1.0 0.67 0.33 0.67 0.33 1.0 0.67 0.33 0.67 0.33 1.0 0.65 0.35
0.67 0.33
Travel speed,
ipm
1.2 1.2 1.2 1.2 1.2 4.0 4.0 4.0 4.0 4.0 9.0 9.0 9.0 9.0 9.0
Spot spacing,
in.
— .05 .05 .001 .001
-.05 .05 .003 .003
-.05 .05 .075 .075
Peak current, A
138 146 150 141 155 151 160 185 160 205 770 195 243 190 253
tap setting was set at a position corre-sponding to wave shape "
B " in Figs. 3 and 4.
Standard Welding Conditions
The pulsed GTA welding process employs high-current pulses
superim-posed on a background current which is generally very low.
A specific weld-ing operation can be described by specifying the
welding voltage "V," the travel speed "S," the level of peak
current " l p , " the level of low current " l L , " the t ime at
peak current " t „ , " and the time at low current " t L . " New
process variables are frequently de-fined by combining two or more
of the basic parameters. For example, the fraction of time spent at
peak current is defined as tp / ( tp + t,.). Another useful
combination of parameters is spot spacing "P" , which is defined
as:
P = S (tp + t,,)
The equipment employed in this study was designed such that the
value of I, was set as a fraction of lp. This fraction (e.g., I,, =
0.15 l„) was held constant wi th in each test series. The actual
value employed for a given test series was chosen to maintain a
stable arc whi le at low current, for the lowest level of peak
current selected.
All of the welds made in this study were bead-on-plate welds
using a standard 2% thoriated tungsten elec-trode. The electrode t
ip configuration was a blunt point w i th a 120 deg included angle.
Current density in the
WELDING RESEARCH SUPPLEMENT I 145-s
-
LONGITUDINAL WELD TENSILE 3.250" A
-i h 050
TRANSVERSE WELD TENSILE 75
;6302
500
5 0 -, h 050
F/g. 5—Tensile sample config-urations employed in this study.
Note the scale change between sample drawings
LONGITUDINAL FUSION ZONE TENSILE 1.000"
.125
1875
2.5
H h .050"
electrode was maintained as closely as possible to 20,000 A/
in.3 (12.9 X 106
A/mm 3 ) , and the arc length was set at 0.062 in. (1.59 mm).
Controll ing arc length, electrode tip configuration and electrode
current density resulted in an essentially constant voltage of
about 10 V for all welding condi-tions.
Intermittent travel speeds have been used and reported in the
literature. Wi th this arrangement the travel would be interrupted
during the high current portion of the cycle. The equipment uti l
ized in this study was not capable of intermittent travel;
therefore, the reported travel speed is not an average but a
uniform travel speed.
Solidification Studies
The effect that process parameter variations have on the weld
macro-structures, solute distr ibution, and so-lidification
structure was evaluated in a partial-penetration and
full-penetra-tion bead-on-plate weld test series. The welding
parameters for the par-tial-penetration weld tests are given in
Table 1, and the full-penetration weld
tests are given in Table 2. A back-ground current of 0.15 lp was
em-ployed for all welds in these tables.
The material selected for these tests was 0.063 in. (1.59 mm)
thick mil l annealed Ti-11.5Mo-6Zr-4.5Sn, com-monly referred to as
Beta III. This alloy was selected because the high temper-ature
beta phase can be retained in a metastable condit ion down to room
temperature. The molybdenum in this alloy segregates upon
solidification and this segregation is easily detected in the
microprobe.
An ETEC microprobe was used for collecting the solute
distribution data. Samples were prepared for micro-probe analysis
by machining 0.01 in. (0.25 mm) off the surface of the weld beads,
and metallographically polish-ing the machined surface through 0.05
micron alumina.
Line microprobe traces were taken along the weld centerline by
moving the instrument stage wi th a synchro-nous motor at 10
microns per minute. An accelerating voltage of 15 kV and a sample
current of 0.15 X 10~6 A were used for all the samples. A beam
current regulator was also employed to maintain a constant beam
current
over the length of each trace. A single spectrometer was used to
detect the Mo La signals from the samples. The length of the probe
traces were approximately equal to the spot spac-ing, except for
spot spacings of 0.15 in. (3.81 mm). For these samples, five
equally spaced line scans of 150 microns in length were used.
The data generated were in the form of a pen strip chart and
teletype output. The strip chart continuously monitored the sample
current and the X-ray output characteristic of the molybdenum La
line. The teletype printed out the total number of counts for the
molybdenum La l ine summed over 10 s by the f low-proport ional
X-ray counter, and other probe data which indicated the degree of
beam stability. The teletype also punched the date onto tape which
was subse-quently converted to computer cards. Although both the
strip chart funct ion and the teletype output were simulta-neous,
the teletype output was inter-rupted after every 10 s period for
about 5 s to print the totals of the previous 10 s count.
After transferring the teletype data to computer cards, a mean
and stan-
146-s I M A Y 1979
-
Table 3-Fu
Sample number
232 233 234 235 236 237 238 239 240 241 242 243 244 245 246
II Penetration Bead-O
Frequency, cycles/s
0 0.41 0.41
20 20 0 1.33 1.33
20 20 0 3 3
20 20
n-Plate Welds
Fraction of Time
at l„
1.0 0.67 0.33 0.67 0.33 1.0 0.67 0.33 0.67 0.33 1.0 0.67 0.33
0.67 0.33
For Grain Size Evaluation
Travel speed,
ipm
1.2 1.2 1.2 1.2 1.2 4 4 4 4 4 9 9 9 9 9
Spot spacing,
in.
-.05 .05 .001 .001
-.05 .05 .003 .003
-.05 .05 .075 .075
Peak current, A
138 146 150 141 155 151 160 185 160 205 170 195 237 190 253
- TOP WELD BEAD VIEW
^W^ Fig. 6—Three weld views employed for grain size
determinations.
Table 4—Evaluation of Grain Refinement in Mild Steel Welds
Frequency, cycles/s Fraction of t ime at lp Travel speed, ipm
Spot spacing, in. Peak current, A
Sampl
301
0.50 0.17 0.5
.017 200
e number
302 303
0.38 0.0 0.38 1.0 0.5 0.5
.022 -100 55
dard deviation was calculated for each sampie. The standard
deviation was anticipated to be higher for the samples w i th the
greatest solute con-centration difference between the cell core and
the cell boundary. Since it was not possible to maintain exactly
the same level of sample current from sample to sample, the average
count traces varied moderately from sample to sample making it
necessary to express the standard deviations as a percentage of the
mean number of counts in a 10 s interval. Following microprobe
analysis the samples were etched wi th Kroll's etch which re-vealed
the solidification structure and the probe trace.
A direct correlation of strip chart results and microstructure
was possi-ble. Selected quantitative point counts were performed on
samples 153 and 163 to compliment the qualitative data. Point
counts using the micro-probe were taken at several cell boundaries
and cell cores. The cells were made visible by lightly etching the
sample surface prior to microprobe examination. The composit ion of
a region including 30 — 50 cells was also quantitatively determined
as a check on the technique. The average chemi-cal composit ion of
this area as deter-mined by quantitative microprobe analysis was
compared to that of the analyzed composit ion.
Tensile tests were conducted for the welding conditions in Table
2. Longi-tudinal fusion zone, longitudinal weld tensiles and
transverse weld tensiles were prepared by making bead-on-plate
welds, aging at 950 F (510 C)/8 hours (h) and machining the test
coupons on all surfaces. The sample configurations are given in
Fig. 5.
Grain Size Effects
The original objective of this phase
was to measure the grain size of the full-penetration welds in
Beta III t i tan-ium level in Table 2. Delineating the grain
boundaries inside the solidifica-tion network, however, proved to
be impossible. The alpha-beta alloy Ti-6AI-6V-2Sn in the form of
0.063 in. (1.59 mm) sheet was subsequently selected for the grain
size measure-ments. This alloy was selected because the
transformation on cooling from the beta region served to break-up
the solidification structure and make for easier identification of
the grain boundaries.
Full-penetration bead-on-plate welds were made according to the
welding conditions shown in Table 3. Three views of the fusion zone
were metallographically prepared for exami-nation. These
metallographic views are shown in Fig. 6. A long etch of about 2
min with Kroll's etchant was used. After this length of time the
surface took on the appearance of steps at the grain boundaries. A
modification of ASTM standard method El 12 was employed for
determining the grain size in all views.
Further tests were run to evaluate the uti l i ty of employing
the cycling temperatures in the fusion zone and heat affected zone
to modify the as-welded microstructure. Low alloy steels are known
to exhibit grain refinement when cycled several times into the
austenitic region. If this same principle could be employed with
pulsed current welding, a finer grain size should be possible.
Weld samples for microstructural examination were prepared using
the welding conditions in Table 4. These welding conditions
represent roughly identical heat inputs, frequency, and spot
spacings; but offer different val-ues of t„. The effect of this
would be to
increase the volume of material taken into the austenite region.
The low level of current was set at 0.07 lp. The microstructures
were compared using the top weld bead surface after 0.01 in. (0.25
mm) of material was removed, and also a transverse view.
Results
Solidification Studies
The root weld-bead macrostructures were photographed and
compared to those obtained wi th constant current conditions. The
macrostructures for pulsed welds in general d id not differ from
those observed in constant current welds of similar travel speeds.
The only exception to this occurred at very low travel speeds,
where it was possible to obtain a radial solidifica-tion
macrostructure as can be seen in Fig. 7. A constant-current weld
made at the same travel speed exhibited a grain structure which was
curved in the direction of travel.
Table 5 shows the microprobe data taken as 10 s counts. These
standard deviation data are presented as a percentage of the mean
10 s count for each sample. There appears to be a slight trend in
these data which indi-cates that the standard deviation is reduced
by selecting small spot spac-ings and a low fraction of t ime at
peak current. In terms of the basic process parameters, decreasing
values of P and 0P correspond to a decreasing t ime at peak
current:
t„ =
t„ —
1-0P
p-st,
Interpretation of these data must
W E L D I N G RESEARCH SUPPLEMENT I 147-s
-
Table 5-Effected Pulsed GTA Parameters On Segregation
Fig. 7—Radial growth resulting from weld-ing conditions of 6„ =
0.5, S = 2 ipm (0.85 mm/s), P = 0.08 in. (2.03 mm) and /,, = 85
A
take the e x p e r i m e n t a l p r o c e d u r e a n d m ic
ros t ruc tu res i n to c o n s i d e r a t i o n . The e lec t ron
beam traverses the sam-p le at a rate o f 10 m i c r o n s / m i n
w h i c h results in a d is tance o f 1.67 m ic rons b e i n g t
rave led fo r each 10 s c o u n t i n g in te rva l . Th is l e n g
t h , p lus t h e f i n i t e size of t h e p r o b e spot , means
that t h e signals are averaged in each 10 s in te r -val over a d
is tance of several m i -c rons .
Figure 8 s h o w s t h e va r i a t i on in ce l l size w h i c
h can be obse rved over a d is tance o f o n e spot spac ing . T y
p i c a l -ly, t h e ce l l s ize var ied b e t w e e n 5 a n d 20
m ic rons for all of the c o n d i t i o n s tes ted. A m o r e
severe f l u c t u a t i o n in m i c r o p r o b e signals w o u l
d be expec ted in the large ce l lu la r reg ion in Fig. 8, s ince
t h e b e a m w o u l d no t cover as large a p o r t i o n o f
each ce l l .
This t e n d e n c y can be seen in Fig. 8 in b o t h t h e s
tandard d e v i a t i o n f igures l oca ted next t o each p r o b
e trace a n d also t he l ine scans des igna ted " B " a n d " C .
" Line scan " B " is a p o r t i o n of the t race w i t h a s
tandard dev i a t i on of 4.9%. L ine scan " C " is t aken f r o m
t h e coarse ce l lu la r reg ion . The ef fect o f t he pu ls ing
c o n d i t i o n s o n the actual i n t ra -ce l lu la r segregat
ion is in d o u b t b e -cause o f t h e averag ing of t h e c o m
p o s i -t i on by t he e x p e r i m e n t a l t e c h n i q u e .
Q u a n t i t a t i v e p o i n t c o u n t s us ing the m i c r o
p r o b e w e r e taken o n the ce l l cores a n d ce l l b o u n d
a r i e s for t w o w e l d i n g c o n d i t i o n s . These data
(Tab le 6) i nd i ca te tha t t he re was n o s ign i f i -cant c h
a n g e in t h e degree o f i n t race l -lu lar segregat ion b e t
w e e n the t w o samples tes ted .
The so l i d i f i ca t i on s t ruc tu re s h o w n in Fig. 8
is t yp ica l of t h e s t ruc tures seen unde r w e l d i n g c o
n d i t i o n s w i t h large tp.
Sample number
146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161
162 163 164 165 166 167 168 169 173 175 177 179 181 183 185 191 187
193 189 195 203 199 201
Travel speed, ipm
1.2 1.2 1.2 1.2 1.2 1.2 1.2 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0
9.0
15.0 15.0 15.0 15.0 15.0 15.0 15.0 15.0 1.2 1.2 1.2 1.2 1.2 4.0
4.0 4.0 4.0 4.0 9.0 9.0 9.0 9.0 9.0
Spot spacing, in.
.005
.05
.005
.05
.005
.001
.15
.15
.05
.0075
.15
.05
.0075
.15
.05
.15
.05
.0125
.15
.05
.0125
.15
.05
.05
.001
.001
.05
.05
.003
.003
.05
.05
.0075
.0075
Fraction of t ime at
peak current
1.0 0.67 0.67 0.33 0.33 0.204 0.33 1.0 0.67 0.65 0.67 0.33 0.35
0.33 0.200 0.200 1.0 0.67 0.67 0.67 0.33 0.33 0.33 0.194 1.0 0.67
0.33 0.67 0.33 1.0 0.67 0.33 0.67 0.33 1.0 0.67 0.33 0.67 0.33
Standard deviat ion,
%
3.0 2.4 5.3 1.5
1.4 6.0 5.1
4.5 5.4 4.1 4.0 5.2 4.0 5.4 5.9 4.3 4.1 5.5 3.8 4.8 3.9 2.3 2.3
4.2 2.8 1.9 4.2
4.1 3.0 2.5 3.6 6.6 3.0 3.1 2.8
As t h e t i m e at peak cu r ren t is r e d u c e d the w i d t
h of the large cel l reg ion decreases u n t i l it is v i r t ua l
l y e l i m i n a t e d as seen in Fig. 9. The o r i g i n o f t he
coarse ce l lu la r reg ion comes f r o m so l i d i f i ca t i on
t ak i ng p lace d u r i n g t he h i g h - c u r r e n t p o r t i
o n o f t h e cyc le . A r e d u c t i o n in tp mere ly reduces
the t i m e for th is so l i d i f i ca t i on to occur .
Tens i le tests us ing th ree d i f f e ren t test samp le c o n
f i g u r a t i o n s w e r e run o n t h e w e l d i n g c o n d i
t i o n s co r re -s p o n d i n g to samp le n u m b e r s 173 t h
r o u g h 203 in Tab le 2. Tens i le s t r eng th , y i e l d s t r
eng th , and e l o n g a -t i o n fa i led t o d i f f e ren t i a
t e b e t w e e n the var ious so l i d i f i ca t i on s t ruc
tures and w e l d i n g c o n d i t i o n s . Typ ica l t e n -sile
p roper t i es for t he l o n g i t u d i n a l f u s i o n z o n e
samples w e r e 150 ksi (1034 MPa) tens i le s t reng th , 140 ksi
(965 MPa) y i e l d s t reng th and a p p r o x i m a t e l y 6.5%
e l o n g a t i o n .
Grain Size Effects
As can be seen in Tab le 7, t he e f fec t o f p u l s e d - w e
l d i n g var iab les o n f u s i o n
zone gra in size is m i n i m a l . The observed var ia t ions
in gra in size are be l i eved d u e to i nhe ren t er ror ar is
ing f r o m the l i m i t e d n u m b e r o f grains in t he f i e
ld o f v i e w and f r o m c h a n g i n g d i r e c t i o n a l i
t y character is t ics . The e p i -taxial g r o w t h tha t occurs
in w e l d s does no t pe rm i t m a n i p u l a t i o n of f us
ion z o n e gra in size by va ry ing t he process parameters in l o
w f r e q u e n c y pu lsed GTA w e l d i n g .
T h e o n l y poss ib le m e t h o d o f o b t a i n -ing a
reduced fus ion z o n e gra in size w o u l d be by e m p l o y i n
g the cyc l i ng t h e r m a l c o n d i t i o n . Tests w e r e
run t o eva lua te t he feas ib i l i t y o f th is ap -p roach us
ing m i l d s teel . Figure 10 shows the m ic ros t ruc tu res o f
t he base me ta l , pu lsed cu r ren t w e l d f us ion z o n e , a
n d t h e cons tan t cu r ren t w e l d f us ion zone .
A l t h o u g h the ef fect o f t h e pu lsed cu r ren t o n gra
in size was q u i t e p r o n o u n c e d , it was a c c o m p l i
s h e d at an u n a c c e p t a b l e y l o w t rave l speed. Six m
o r e w e l d s w e r e m a d e us ing t h e same mater ia l to o b
t a i n a be t te r idea o f t h e m a x i m u m travel speed p e r
m i t t e d
148-s | M A Y 1979
-
while still refining the grain size. The welding conditions for
these addit ion-al welds along wi th comments on the
degree of grain refinement obtained are given in Table 8.
From the results presented in Table 8
it was estimated that, in order for grain refinement to occur,
the spot spacing had to be l imited to a maximum of
• S 4 . B %
20 Oy
Fig. 8—A is solidification structure of sample number 163 viewed
normal to the top weld bead surface. S = 15 ipm (6.35 mm/s), 6P =
0.67, p = 0.15 in. (3.81 mm), /„ = 755 A. 13 and C are microprobe
line scans of the Mo La radiation.
Table 6-Quantitative Microprobe Data, wt-%
Sample number 163 Sample number 153
M o Zr Sn Ti
Cell core
12.59 4.53 4.50
77.41
Cell boundary
10.14 7.86 4.01
77.47
Cell core
12.47 4.58 4.81
77.28
Cell boundary
9.93 7.84 4.84
76.34
Rastered area
10.93 5.77 4.79
78.23
Cell core
10.62 6.06 4.56
77.86
Cell boundary
9.10 8.03 4.67
76.74
Cell core
10.19 6.51 4.60
78.09
Cell boundary
9.46 7.74 4.64
77.12
Rastered area
10.22 6.97 4.67
77.82
Ba se metal composit ion
10.5 6.8 4.4
78.2
W E L D I N G RESEARCH SUPPLEMENT ! 149-s
-
- ^ 1
Fig. 9—Solidification structure oi sample number 166 viewed
normal to the top weld-head surface. The coarse cellular region is
virtually eliminated. S = 15 ipm (6.35 mm/s), 6P = 0.33, P = 0.15
in. (3.81 mm), /P = 256 A.
about 0.05 in. (1.27 mm), the time at low current had to be
greater than 1 s, and the t ime at peak current could go as low as
0.17 s. These values roughly define the conditions for a maximum
travel speed whi le still obtaining the desired grain refinement.
The maxi-mum travel speed was calculated to be about 2 ipm (0.85
mm/s).
A single additional weld was made using these conditions, but no
signifi-cant level of grain refinement was observed. The maximum
travel speed permitted whi le obtaining grain re-finement must
therefore be consider-ably less than 2 ipm (0.85 mm/s) and probably
on the order of 1.0 ipm (0.42 mm/s).
The grain refinement obtained in these welds is clearly due to
cycling the temperature above and below the austenitizing
temperature. Single cy-cles into the austenitizing region, as would
occur wi th large spot spacings, did not result in significant
grain refinement. Therefore, the number of cycles required to
obtain a reasonable amount of refinement must be greater than one
and is probably in the vicinity of three cycles. Wi th the
formation of each spot spacing a narrow region on the trailing edge
of the weld pool is heated into the austenite region.
Table 7 - G
Sample number
232 233 234 235 236 237 238 239 240 241
rain Size in Ti-6AI-6V-2Sn Welds
Fraction of t ime at
peak current
1.0 0.67 0.33 0.67 0.33 1.0 0.67 0.33 0.67 0.33
Normal grain size,
mm
.716
.668
.488
.635
.706
.536
.387
.494
.377
.810
Transverse grain size,
mm
0.638 0.645 0.748 1.22 1.11 0.762 0.500 0.604 0.570
-
fongi tudinal grain size,
mm
0.975 0.762 0.797 1.14 1.79 1.88 0.581 0.713 0.049
-
Average grain size,
mm
0.776 0.692 0.678 0.998 1.20 1.06 0.489 0.604 0.532
-
. *
3» • f y v / .' -
. » * • • • - ' -
•. j r .(fr
zmm &
M H M B Fig. 10—Transverse views of welds in mild steel from the
weld centerline on the right to the base metal on the left.
A-constant current weld, sample 303; B—pulsed current weld, sample
301
Table 8 -
Travel speed.
ipm
0.9 1.8 0.6 0.7 0.66 1.0
Evaluation Of Grain Refinement In Mild Steel Welds
Time at peak
current, s
1.00 1.00 0.50 0.17 1.00 1.00
Time at low
current, s
1.65 1.65 1.65 1.65 1.00 0.33
Peak current,
A
125 125 180 400 90 90
Low current,
A
10 10 10 10 10 10
Spot spacing,
in.
.040
.080
.022
.022
.022
.022
Extent of grain
refinement
Moderate Poor Extensive Extensive Poor Poor
To obtain the necessarv number of cycles, the distance covered
in one cycle cannot exceed one third of the width of the
austenitized region. This limits the spot spacing to a relatively
small value. Travel speed can be increased and the spot spacing
held constant by merely increasing the frequency of pulsing.
However, the need to heat a given region above the austenitizing
temperature and again to cool this region, places an upper limit on
the frequency and therefore the travel speed.
Very little freedom is seen in the variation of tp; but if the
cooling rate could be increased, the value of tL
could be decreased. The result would be a moderate increase in
travel speed, assuming a constant spot spacing. To test this idea
several partial penetra-t ion bead-on-plate welds were made on 0.25
in. (6.35 mm) thick mild steel plates. The result was that any
increase in travel speed realized whi le main-taining conditions
for grain refinement was very small and did not significant-ly
change the value found for the 0.063 in. (1.59 mm) sheet. This
would indi-cate that the use of a higher density heat input source
such as plasma arc welding would also have little effect on the
maximum obtainable travel speed.
150-s I M A Y 1979
-
•y •*":' '.:
Fig. 11-Solidification structures observed in sample number 766
viewed normal to the top weld bead surface. S = 15 ipm (6.35 mm/s),
6L = 0.33, P = 0.75 in. (3.81 mm), /p = 256 A. A—region which
solidified early in the tt portion of the cycle; B—region which
solidified late in the tL portion of the cycle.
Discussion
The fusion zone macrostructures in low frequency pulsed current
welds were not observed to deviate from that which occurred wi th
constant-current welds of the same travel speed, except at large
spot spacings and low travel speeds. Wi th these welding conditions
the weld macrostructure revealed a radial growth pattern. Slavin et
a/.3
investigated the weld macrostructures that result from pulsed
current welds and related these to hot tearing tendencies. They
reported that a large radius for the trailing edge of the weld pool
decreases the susceptibility to hot tearing. Although hot tearing
is generally not a problem in t i tanium alloys and was not a
subject of this study, it does seem logical that the trailing edge
of the weld pool serves as a stress raiser, the notch effect being
minimized by a large radius. The resulting radial growth patterns
would also serve to limit centerline segrega-t ion.
The solidification for all of the pulsed current welding condit
ions was epitaxial and varied moderately in the cellular to
cellular-dendritic range. The epitaxial situation dictated the
grain size in the fusion zone, since the fusion zone grain size
could be no smaller than the near heat-affected zone grain size.
The grain size of t itan-ium when taken above its beta transus is
dependent upon the peak tempera-ture that it experiences and to a
lesser extent on the exposure time.3 Since the near heat-affected
zones in all the samples would have been exposed to temperatures
just below the melting point of the alloy, the grain size in this
region would be expected to be the same from sample to sample, and
no
fusion zone grain size change would be expected under conditions
of epitaxial growth.
Pulsed-current welding conditions were varied over a very wide
range to evaluate the effect of the welding parameters on the
solidification pat-tern. The fact that the solidification mode
varied only moderately would indicate that there is little control
of the ratio, GL/R. Examination of a single spot weld for one of
the condi-tions of large spot spacing shows that the structure only
changes from cellu-lar at the periphery of the weld pool to
cellular dendritic much later in the solidification process.
Slavin etala measured growth veloc-ities of the solid during a
single cycle using high speed cinemotography. They found that the
growth rates were greatest at the beginning of the low current port
ion of the cycle and decreased rapidly as solidification
pro-gressed. The thermal gradients in the l iquid were noted to fo
l low a similar pattern. The gradient in the l iquid was greatest
at the start of the low current portion of the cycle and decreased
by an order of magnitude by the t ime the peak level of current was
restored. The similar changes which Gh and R go through apparently
do not permit much change in the Gr./R ratio.
Wi th both G and R maximum at the beginning of the low current
portion of the cycle, the factor (GR)-' which is proportional to
cell size would indi-cate that the min imum cell size should
correspond to the beginning of the low current portion of the
cycle. Regions of a weld corresponding to the beginning and end of
the tL portion of the cycle are shown in Fig. 11. The cell size is
observed to be moderately larger in the photograph
corresponding to the end of the t,, portion of the cycle.
As pointed out previously under "Background," intracellular
segre-gation is proportional to the ratio of GL/R. This ratio was
not observed to change sufficiently to cause a detect-able change
in the degree of solute segregation. The lack of a change in the
degree of solute segregation between the cell core and cell
bound-ary further supports the conclusion drawn from the
examination of the solidification structure—namely, that the ratio
of GL/R experienced only small changes wi th in a given spot weld
and over the entire range of welding parameters studied.
Reduction of the mild steel fusion zone grain size was possible
by employing the proper pulsing condi-tions to thermally cycle the
fused weld into the austenite range several times. The number of
times that a region is cycled into the austenitizing range affects
the grain size, and the number of cycles a region sees is
determined by the pulsing parameters. The re-quirement of cycling
each region of the fusion zone several times into the austenite
region limits the travel speed to a very low value.
Conclusions
1. Low frequency pulsed-current gas tungsten arc welding
afforded minimal control over the solidification structure observed
in the t i tanium alloy, Ti-11Mo-7Zr-4.5Sn.
2. No change in tensile properties in Ti-11Mo-7Zr-4.5Sn was
noted for the range of process parameters investi-gated.
3. Fusion zone grain size in Ti-6AI-§V-2Sn pulsed-current welds
was not
W E L D I N G RFSFARCH SUPPLEME NT | 131-s
-
obse rved t o change s ign i f i can t l y f r o m tha t o b t a
i n e d in c o n s t a n t - c u r r e n t w e l d s .
4. R e d u c t i o n o f grain size in m i l d steel was s h o w
n to be poss ib le by p rope r se lec t ion of p u l s e d - c u r
r e n t w e l d i n g parameters . The p rac t i ca l i t y o f t
he w e l d i n g p r o c e d u r e was severely res t r i c ted d u
e t o a n a r r o w process parameter o p e r a t i n g e n v e l o
p e w h i c h res t r i c ted t he t rave l speed to less than 1 i
p m (25.4 m m / m i n ) .
References
1. Flemings, M.C., Solidification Process-ing, McGraw-Hi l l
Book Company, New York, 1974.
2. Davies, G.|., and Garland, J.G., "Sol id-ification Structures
and Properties of Fusion Welds," International Metallurgical
Re-views, Vol. 20, Review 196, 1975, pp. 83-106.
3. Matsuda, F., Hashimoto, T., and Senda, T , "Fundamental
Investigations on Solidifi-cation Structure in Weld Meta l , "
Transac-tion of National Research Institute for Metals, Vol. 2, No.
1, 1969, pp. 43-58.
4. Slavin, G.A., Maslova, N.D., and Moro-zova, T.V., "The
Relationship Between Technical Strength and Solidif ication in the
Pulsed-Arc Welding of Creep Resisting Alloys w i th Non-Consumable
Electrodes," Svar. Proiz., 1971, No. 6, pp. 17-19.
5. Wil l iams, J.C, Private Communicat ion, Carnegie-Mellon
University, 1978.
6. Slavin, G.A., Maslova, N.D., and Moro-zova, T.V., "Some
Special Features of the
Solidification of Liquid Pool Metal in Pulsed Arc Weld ing Using
a Tungsten Elec-trode," Svar. Proiz., 1973, No. 6, pp. 7-9.
7. Petrov, A.V., and Birman, U.I., "Condi -tions under wh ich
the Metal Pool Solidifies during Pulsed Arc Weld ing w i th a
Tungsten Electrode," Avto Svarka, 1969, No. 8, pp. 24-26.
8. VanWinkle, D.B., "Soviet Technology in Pulsed-Current Weld
ing Processes," FTD-CS-01-03-70, October 1970.
9. "Pulsed Gas Tungsten Arc Welding: Research and Development
Status in the USSR," McGraw-Hi l l Information Systems Company, MHR
74-5, December 1974.
10. Becker, D.W., and Adams, C M . , " I n -vestigation of
Pulsed GTA Weld ing Parame-ters," Welding journal, 1978, Vol. 57
(5), May 1978, Research Suppl., pp. 134-s to 138-s.
The AWS Standard for Qualifying Welders and Welding
Procedures.
The AWS Standard Qualification Procedure (B3.0-77) covers
proce-dures and welders; pipe, plate and sheet metal, ferrous and
nonferrous metals; welded by all the major processes.
It's specifically intended for use by fabricators, contractors,
and others who use welding but have no appli-
cable welding product specifications. This Standard contains 112
pages
and is saddle-stitched, soft cover, 8V2 x 11 in., and three-hole
punched. AWS B3.0-77 is priced at $12.00 per copy. Available from
the Order Dept., American Welding Society, 2501 N.W. 7th Street,
Miami, FL 33125.
AMERICAN WELDING SOCIETY
| American Welding Society, 2501 N.W. 7th St., Miami, Florida
33125
p i p a q p Qpnrl me r n p i p c ; nf t hp A m p r i r a n
(NUMBER)
Welding Society Standard Qualification Procedure priced at $12
each.
Spnrl P.hpnk nr Total $
I money order (Add 4 % Sales tax for orders delivered in
Florida)
NAMF (PLEASE PRINT)
[ nr iMPANY NAMF
! A n n R F s s
OITY STATF 7IP
1
" 1
152-s | MAY 1979