Gas Metal Arc Welding of

Carbon Steel

TABLE OF CONTENTS

Section/Topic
Section 1

Page
1

Introduction
Section 2

Section/Topic
Section 6

Page
28

Metal Transfer
2

Base Metals
A. Alloying Elements

B. Carbon Steels

C. Alloy Steels

A. Short-Arc

28

B. Globular Transfer

30

C. Spray Transfer

31

D. Pulsed Spray Transfer

34

Section 7

35

Welding High Strength Steels

Section 3

Electrical Characteristics
A. Power Supply Basics

B. Constant Voltage Power

Supply Controls

C. Electrical Stick Out

A. Select the Proper Filler Metal

36

B. Minimize Hydrogen Contamination

37

C. Control Heat Input

38

D. Use the Correct Technique

40

Section 8

42

Technique and Equipment Set-Up

D. Constant Voltage Power

Supply Characteristics

11

A. Torch Angle

42

1. Slope

11

B. Feed Roll Tension

43

2. Inductance

12

C. Burnback

43

3. Heat Input

13

D. Arc and Puddle Position

44

E. Vertical Down Welding

44

F. Gaps

45

G. Crater Filling

46

H. Arc Starting

46

Section 9

47

Section 4

14

Shielding Gases
A. Shielding Gas Functions

14

B. Flow Rates

16

C. Gas Losses

17

D. Shielding Gas Selection

20

Section 5

23

Electrodes

Weld Discontinuities and Defects

A. Lack of Fusion

47

B. Porosity

48

C. Burn-Through

49

A. Alloying Additions

23

D. Undercut

50

B. 1. Solid Wire Designations

24

E. Spatter

51

F. Cracking

51

Section 10

52

and Chemistry
B. 2. Metal-Cored Wire Designations

25

and Chemistry
C. Flux-Cored Tubular Wire

25

Designations
D. Slag and Gas Formation

26

E. Solidification of the Weld Puddle

27

Conclusion

SECTION 1

1
Introduction

This training program was written to give

As you learn more about GMAW, it will

you a better understanding of the MIG

become apparent that this is a sophisticated

welding process. MIG is an acronym for

process. Welders that have used stick

Metal Inert Gas, which is not technically

welding (Shielded Metal Arc Welding or

correct for steels, because shielding gases

SMAW) are sometimes of the opinion that

for steels contain an active gas such as

the GMAW process is simpler; but to deposit

oxygen or carbon dioxide. The correct term

a high quality bead requires as much know-

according to the American Welding Society

ledge, or probably more, than the SMAW

(AWS) is Gas Metal Arc Welding (GMAW).

process. The reason for this is the number of

We will use the correct terminology as

variables that affect the arc and the degree of

defined by the AWS and also explain the

control the operator has over those variables.

slang used so that you will be familiar with

all the terms applicable to this process.

The purpose of this manual is to make you a

better welder by increasing your knowledge
of how the GMAW process works. A more
knowledgeable welder can be more productive by working smarter, not harder. Figure 1

Wire
10%

shows why your company is interested in

Gas
5.0%

your education. Your labor and overhead

account for about 85%of the cost of depositing weld metal. Any knowledge you gain
from this course not only helps you, but also
helps to make your company more competitive in a very tough marketplace. If you
should have any question in the future that
this manual or your supervisor cannot ans-

Overhead
and Labor
85%

wer, please free to have him contact your

Praxair regional engineering staff for further
assistance.

Figure 1
Breakdown of
welding costs

SECTION 2

2
Base Metals

The two main categories of steel fabricated

today are carbon and alloy steels. Carbon

Unit cell
(9 iron atoms)

steels basically contain carbon and manganese. Alloy steels contain carbon, manganese, and a variety of other elements to give
the base metal the required properties. Alloy
steels cover a wide range of materials, such
as the stainless and tool steels.
As weight becomes an issue in the transportation industry, there has been a substitution
of high strength low alloy steels (HSLA) in
applications that previously used carbon
steels. The higher tensile and yield strengths

Figure 2 Iron body centered

allow the cross-sectional area weight of

Cubic Unit Cell

the structural members to be reduced. The

resulting weight reduction allows better fuel

areas between the iron atoms. In carbon

economy and a stiffer body in the case of

steel, manganese is the other alloying ele-

automobiles, or an increase in payload in

ment. In the matrix, a small portion of the

the case of heavy trucks. Stainless and tool

iron atoms is replaced by manganese atoms.

steels are two more familiar categories of

The layers above and below the first layer

alloy steels. This manual will concentrate on

are arranged identically, except that they are

the low alloys such as the construction

shifted on a 45 degree angle to fall into the

steels, (A514, for example, referred to by

areas where the first layer of balls intersect.

USX as their T1 alloys). In order to under-

Now each ball in the second layer is touch-

stand alloy steels, it helps to understand a

ing four balls in its layer and four balls in

little bit about the metallurgy involved with

the layers directly above and below it. The

the elements added to increase the strength

orderly manner in which metals are arranged

of these materials.

in crystals is one of the reasons that they are

so strong. When a metal yields, or deforms

Metals are crystals, which mean that the

atoms are arranged in an ordered matrix. An
easy way to visualize a metal is to think of
layers of balls with each ball in the layer
touching its four neighbors (see figure 2).
The balls represent the iron atoms of the
metal. Carbon atoms are much smaller than
the iron atoms, and they fit into the open
2

plastically, the planes of atoms slip in relation to the adjacent planes and at the grain
boundaries. The only single crystal materials
used today are for turbine blades. They are
extremely strong due to the orderliness of
the matrix. They are also extremely expensive to make.

The steels that are used in fabrication are

When carbon is added to a steel, the strength

actually made up of grains or groups of

of the steel goes up dramatically. The car-

crystals. During welding, grains begin to

bon forms a compound called iron carbide

grow into the molten puddle from the solid

(Fe3C). The size of the iron carbide molecule

base metal at the edge of the weld. Each

(a molecule is a combination of atoms held

grain continues to grow until it meets an-

together to form a compound) is consider-

other grain. The area where they meet is

ably larger than the surrounding iron atoms.

called a grain boundary. In the steels of

When the atomic layers begin to slip, the

most interest, a lot of the slip or shifting

larger carbide molecule resists this slip by

of atoms, occurs at these grain boundaries.

pinning the layers together due to their

Because of grain boundaries, the actual

difference in size. There is a number of other

strength of these steels is typically 25%

carbide forming elements that work in a

to 50% of the theoretical strength of a

similar manner to strengthen alloy steels.

single crystal of iron.

The following section lists some of the
elements that are added to steel to yield
the desired properties.

A.
Alloying
Elements

Carbon

The manganese reacts with oxygen and

As mentioned previously, carbon in steel

forms manganese oxide (MnO). The man-

forms iron carbide (Fe3C) in the matrix.

ganese also combines with sulfur to form

The larger size of the carbide compound pins

manganese sulfide (MnS) (these are some-

the layers within the metal matrix and makes

times known as stringers and can cause

it much more difficult for the material to

welding problems in certain steels). The

yield. This raises the tensile, yield strength,

reason sulfur is detrimental is that it solidi-

and hardness of the steel considerably.

fies at a low temperature. The liquid sulfur is

carried to the grain boundaries and reduces

Silicon
Silicon is added mainly as a deoxidizer. It
combines with oxygen to form SiO2. This
silicon dioxide, (also known as glass), floats
to the surface of the weld puddle in combination with manganese oxide to form the
brown slag islands seen in the weld surface.
Silicon can also be added as an alloying
element; this is very beneficial in electrical
steels used in transformers.

the strength of the weld. Any manganese

remaining after formation of MnO and MnS
will form manganese carbide (Mn3C) which
strengthens and toughens the matrix. A
special category of steel containing greater
than 10% manganese, is called Hadfield
steels after the metallurgist that discovered
them. Hadfield steels harden very quickly
to high hardness levels. They are also very
abrasion resistant. The most common appli-

Manganese

cation of these materials is in railroad cross-

Manganese is used in small amounts in most

ings or frogs. Hadfield steels are also used

steels to deoxidize, desulfurize, and improve

in rock crushers, dredging pumps, and

material properties.

dippers for power shovels.

Copper

Increasing the chromium level above

Copper is added in very small amounts to

12-13% causes the material to resist cor-

increase the hardness of the steel. Copper

rosion; these materials are known as stain-

can also be added to improve the corrosion

less steels. In chrome plating, a very thin,

resistance of weathering steels such as

tight layer of chromium oxide is formed that

Cor-Ten. These materials can be seen in

resists further oxidation. As the layer grows

the unpainted condition on bridges, guard

in thickness, the color begins to change to a

rails, and light poles. The copper aids in the

straw color, and then to blue. This bluing

formation of a very tight oxide that is not as

can be seen on motorcycle exhaust pipes.

prone to flaking as carbon/manganese steel;

this slows the corrosion rate of the material.

Nickel
While nickel does not form any carbide in

Molybdenum

a steel matrix, hardenability, ductility and

Molybdenum is another carbide forming

toughness are all improved when nickel is

element, and is added to most alloy steels

added. Nickel is added to austenitic stain-

from .5% to 1.5%. Molybdenum improves

less steels (300 series) in concentrations of

yield strength and resistance to high tem-

7-35%. Nickel is also an important element

perature creep (deformation due to high

found in the materials used for storage of

temperature and stress). It also helps to

cryogenic fluids. An alloy containing 9%

maintain the strength of the steel after it

nickel is used to fabricate the inner vessel

undergoes stress relief. Molybdenum is

in liquid nitrogen, argon, and oxygen tanks.

added to stainless steels to reduce pitting

(highly localized corrosion) in corrosive
environments.

Boron
Boron is added in very small amounts
to increase the hardenability of steels.

Chromium

The amount added is typically below .01%.

Chromium is added to increase the strength,

King pins, used on semi-trailers, are made

wear resistance, heat resistance, corrosion

from boron-containing steels.

resistance, and hardness of steel. Chromium

is also a carbide former, and forms chromium carbide (Cr3C). In steels, chromium can
also form complex carbides, Fe2CrC and
Cr2FeC.

Vanadium is a strong carbide former, and

increases the hardenability of the steel.
Vanadium is expensive, so it is usually used
in percentages of less than .2%. This ele-

Bearing steels typically contain about

ment reduces the grain size and increases the

1% carbon and 1.5% chromium. When

toughness of the material. Vanadium steels

chromium levels rise to about 4%, and

are used to make axles, connecting rods,

tungsten and molybdenum are added, tool

hand tools, and engine crankshafts.

steels are the result. These materials are

made into the high speed cutting tools
used in machining.

Vanadium

Titanium

Sulfur

Titanium is a strong carbide former that will

Sulfur, also considered an impurity like

also form oxides and nitrides. A large use of

phosphorus, is usually specified as a maxi-

titanium is to stabilize certain grades of the

mum allowable concentration. Sulfur in the

stainless steels. Titanium combines with any

puddle moves to the grain boundaries of the

carbon in the matrix to form carbides before

solidifying weld metal because of its low

chromium carbide precipitation can occur.

melting temperature. This segregation in the

When chromium forms carbides, the corro-

grain boundaries reduces the strength of the

sion resistance of the material will deterio-

material.

rate. Titanium also helps to reduce grain

growth in high strength steels, improving
strength and toughness.

Manganese is added to prevent this as it

combines with the sulfur (Mn + S = MnS)
before it can react with iron. Certain steels

Phosphorus

called free machining steels, contain up to

Phosphorus is generally considered an im-

.3% sulfur; these alloys are difficult to weld

purity in steels; a maximum percentage is

and have poor strength characteristics.

generally listed. Phosphorus tends to segregate forcing carbon into the surrounding
matrix. This can lead to brittle materials.

With this greater understanding of the alloys

that are added to steel, lets look at carbon
and alloy steels.


B.
Carbon
Steels

Carbon steels are categorized as low,

medium, and high carbon. The main
alloying elements are carbon and manganese. Typical products made of these
three categories of materials are:
Low Carbon Steel

1. Auto frames and bodies

(.005 - .3 C)

2. Auto and truck wheels

3. Structural shapes
(I-beams, channel, angle)

Medium Carbon Steel

1. Machine parts, pins

(.3 - .6 C)

2. Tools

High Carbon Steel

1. Railroad rail

(.6 - 1.0 C)

2. Dies
3. Springs

Carbon is a very powerful alloying element

high carbon groups. Notice that the only

because it forms iron carbide (Fe3C).

real change in most of the alloys is in carbon

Manganese also forms carbides, but has

content. At the right of the chart notice that

much less of an effect on strength and

the tensile strength rises rapidly as the car-

hardness. The following chart shows two

bon level is increased with little or no

carbon steels from the low, medium, and

change in the manganese level.

Table 1

SAE/AISI

Carbon

Manganese

P (max)

S (max)

Chemical

1008

.08 max

.25 - .40

.04

.05

43,000

1018

.14 - .21

.6 - .9

.04

.05

58,000

1040

.36 - .45

.6 - .9

.04

.05

76,000

1050

.47 - .55

.6 - .9

.04

.05

90,000

1070

.64 - .76

.6 - .9

.04

.05

102,000

1090

.89 - 1.4

.6 - .9

.04

.05

122,000

Compositions of

Tensile

Carbon Steels and

Tensile Strengths

C.
Alloy
Steels

Alloy steels are generally classified by

there are. The carbon steels, discussed

alloying additions. There are currently

earlier, are in the first classification. There

38 different classifications of steels that

are 15 to 20 different carbon steels available,

are recognized by both organizations that

and there are 38 different classifications

categorize steels (AISI and SAE). The

listed. Table 2 gives specification numbers

following table is included to give a better

and alloy classification.

understanding of how many steel alloys

Table 2

Specification

Classification

Alloy Steel

Number

Classification and

10XX

Carbon Steels

Specification

11XX

Carbon Steels, Resulfurized

Numbers

12XX

Carbon Steels, Resulfurized and Rephosphorized

13XX

Manganese Steels

2XXX

Nickel Steels

31XX

Nickel-Chromium Steels

33XX

High Nickel-Chromium Steels

40XX

Carbon-Molybdenum Steels

41XX

Chromium-Molybdenum Steels

43XX

Chromium-Nickel-Molybdenum Steels

46XX

Nickel-Molybdenum Steels

48XX

High Nickel-Molybdenum Steels

50XX

Low Chromium Steels

51XX

Chromium Steels

52XXX

Carbon-Chromium Steels

61XX

Chromium-Vanadium Steels

86XX

Low Nickel-Chromium-Molybdenum Steels

92XX

Silicon-Manganese Spring Steels

92XX

Silicon-Manganese-Chromium Spring Steels

93XX

Nickel-Chromium-Molybdenum Steels

98XX

Nickel-Chromium-Molybdenum Steels

XXBXX

Boron Containing Steels

XXBVXX

Boron-Vanadium Containing Steels

WX

Water-Hardening Steels

SX

Shock-Resisting Steels

OX

Oil-Hardening Steels

AX

Air-Hardening Steels

DX

High Carbon-High Chromium Tool Steels

HXX

Hot Work Tool Steels

TX

High Speed Tungsten Based Tool Steels

MX

High Speed Molybdenum Based Tool Steels

LX

Special Purpose Tool Steels

FX

Carbon-Tungsten Tool Steels

PX

Mold Steels

2XX

Chromium-Nickel-Manganese Stainless Steels

3XX

Chromium-Nickel Stainless Steels

4XX

Chromium-Stainless Steels

5XX

Low Chromium Heat Resisting Stainless Steels

SECTION 3

3
Electrical Characteristics

A.
Power Supply
Basics

The power supply (or welding machine)

Congress is made of:

connected to the torch is basically a big

transformer/rectifier. Its purpose is to take

SENators And REPresentatives

high voltage (440v or 220v) and low current

(20-50 amps/leg) AC power and transform

Straight Electrode Negative

it to low voltage (16-40v), high current

(80-500 amp) DC power. To change AC

Reverse Electrode Positive

to DC, a device called a rectifier is used.

Direct current provides a much more stable
arc. Most GMAW power supplies are set-up

Almost all GMAW power supplies are

using a reverse polarity connection. Reverse

constant voltage machines, where Stick

polarity is designated as DCEP, which means

Electrode (SMAW) machines are

Direct Current, Electrode Positive. An easy

constant current.

way to remember this is:

DC
Low Voltage
High Current

AC
High Voltage
Low Current

Work

Reverse Polarity - DCEP (DC - Electrode Positive)

Straight Polarity - DCEN (DC - Electrode Negative)

Figure 3
Typical GMAW
Power Supply


B.
Constant Voltage
Power Supply
Controls

All constant voltage power supplies have

supply or on the remote control. Increasing

at least two operator-adjustable settings:

wire feed increases current proportionally so

current and voltage. Current is set by

that enough current is available to melt the

adjusting the wire feed rate; voltage is set

wire and deposit it in the weld pool. Voltage

with a voltage adjustment on the power

adjusts the length of the arc.

Some power supplies also provide the options of adjustable slope and inductance. The

Figure 4

purpose of these controls will be discussed

Power Supply

in the section on power supply characteris-

Adjustments

tics. Figure 4 shows a power supply with the

standard voltage (arc length) and wire feed
speed (current) adjustments. This power
supply also allows the operator to change

the inductance and slope.

Flat

I
+

Steep

C.
Electrical
Stick Out

Electrical stick out (ESO) is the distance

1. Electrode preheat

measured from the contact tip in the torch to

2. Burning off of drawing lubricants

the workpiece as figure 5 shows. ESO is very

3. Determination of current level

important, and can affect the following:

ESO

Figure 5
Measuring Electrical
Stick Out
9

The wire in the GMAW process is called

rates. This is a big help in controlling pene-

an electrode because it conducts electricity.

tration, because as current increases, so does

The current is transferred to the wire in

the depth of penetration. By using a slightly

the contact tip. The energy resulting from

longer stickout, more weld metal can be

the welding current is distributed to two

deposited without burning through thinner

different places in the welding circuit;

parts. Increasing ESO makes the arc harder

(1) resistance heating of the electrode, and

to start because less current is available at

(2) penetration into the base metal as figure

the arc due to resistance heating. As more

6 shows. The electrode acts like the elements

resistance is put into the welding circuit

in a home toaster. As current passes through

(increased ESO), the effective slope of the

it, resistance heating occurs and its tempera-

system is also increased. This also tends to

ture rises. The increased temperature burns

reduce the short-circuit current. ESO also

off drawing lubricants used in the manufac-

affects shielding gas coverage. As the

turing of the wire. The temperature rise also

distance increases from the contact tip to

helps make it easier to melt the electrode.

the work (also called TWD tip to work

This is the reason that deposition rate incr-

distance), you reach a point where the

eases as ESO increases. As ESO increases,

shielding gas cannot effectively blanket the

current is decreased. This also helps to keep

molten weld puddle. This will be covered

the contact tip cooler at higher deposition

in more detail in the shielding gas section.

Current from the power supply

is distributed to:
1. Resistance heating of the
electrode (ESO dependent)
2. Penetration into base metal

Figure 6
Current
Distribution

10


D.
Constant Voltage
Power Supply
Characteristics

1. Slope

This means that for each increase of

100 amps, the power supply will produce

The characteristics of a power supply are

determined by the components used in its
design. The performance of a typical machine is described by a graph such as figure 7.
Most constant voltage power supplies without slope adjustment are factory preset at
about 2 volts/100 amps (flat slope).

2 volts less at the same voltage setting.

The lower slope line is 6 volts/100 amps,
and is about the maximum slope seen in
a constant voltage power supply (steep
slope). A few machines are still available
with continuously adjustable slope; others
have external or internal taps to switch
between slopes. Increasing the slope of a
power supply to control short-arc welding at

Figure 7 -

low currents is necessary because the short-

Power Supply

circuit current is limited. This reduces the

Characteristics

tendency to burn-through on thinner materi-

Curves

als and decreases spatter on arc starts. This

will be explained further in the section on

32

lat S

30
28
Volts

metal transfer.

CV-F

CV
-S
te

26

-D

ro
o

24

pi

lope

ep

Sl

Figure 8 shows a typical 2 volt/100 amp

100A

slope power supply characteristic curve.

A review of this graph helps to explain

op

ng

22

(2V/

(6

why the arc changes during welding. As

V/

10

20

0A

an example, select a welding condition of

)

27 volts and 250 amps. As welding continues, if the stickout (ESO) is reduced the

18
100

150
200
250
Current (Amps)

300

350

welding conditions change. As that change

is made, the spatter level begins to increase.
As ESO is decreased, less current goes into
preheating the wire and more goes into the

Figure 8

arc. Suppose the current increases 50 amps,

The Effect of Slope on

which is easily done with a small torch

Current and Voltage

movement of about 1/4". This moves the

operating point to the second point in figure
8; here the voltage decreases to 26v while
Flat Slope
2V/100A

22

the current increases to 300 amps. This

voltage is at the minimum for spray transfer;

Volts

21

this would account for the slight increase

in spatter that is observed. This will be

20

explained in more detail in Section 6.

19
18
17
225

250
275
Current (Amps)

300

11

Measuring Actual Welding Voltage

2. Inductance

Measuring actual welding voltage is a useful way to be certain that the condition is
within the range specified by the welding
procedure. Hard starts can result from bad
connections in the welding circuit. The voltage drop due to bad connections increases
the slope of the system, and reduces the
available short-circuit current. Comparing
the voltage at the power supply terminals
and between the feeder and the work (figure
9) will give the voltage drop due to resistance. It can measure as high as 5 volts
which makes starting the arc difficult.

Inductance is an adjustment that is provided

more frequently than slope on CV power
supplies. Inductance is another method for
controlling the arc. This is done by controlling the rate at which the welding current
reaches the setting selected. Figure 10 shows
a plot of inductance vs. time. The top curve
shows what happens with no additional inductance and the current rises as quickly as
the power supply will allow it to rise. This
can result in very hot starts or even in the
wire exploding at these very high current
levels. Inductance should be kept low for
spray transfer. This produces better arc start-

Figure 9

Volt Meter

Measuring Actual Welding Voltage

27.5

ing and more stable arc at high currents.

High inductance settings can make it hard
to initiate the arc because it limits the maxi-

+
A

Connect to
+ terminal on
feeder and work
piece and read
while welding

mum short circuit current available for this

purpose.
Referring again to figure 7, as the electrode
first touches the work to strike an arc, the
voltage falls from open circuit voltage
(40-50v) to 0 arc volts. At 0 volts (no arc
length or a dead short), the power supply
produces the maximum short-circuit current.

V
Flat

I
Steep

Work

For a machine rated at 450 amps (at 100%

duty cycle) this might be 550-600 amps. A
600 amp machine might produce up to 800
amps during this initial start. This is more
than enough current to explode the wire and
make the arc difficult to start. If the arc start

Figure 10
The Effect of Increasing Inductance

is too hot, either the slope can be increased

or inductance added to reduce the current at
arc initiation. Inductance can be beneficial

Current (amps)

Least Inductance

it makes the puddle more fluid and allows it

600

to better wet the base material.

500
400
More Inductance
300
Most Inductance

12

when welding with low current short-arc, as

Time (milliseconds)

3. Heat Input

The heat input formula is also used for

when welding high alloy materials, but it

A useful formula often used by welding

engineers is the heat input equation. If

also helps in understanding the uses of the

power supply characteristic curve.

you have been welding high strength or

corrosion resistant alloys, you may already

The heat input formula can assist in pre-

be familiar with this formula. It allows you

paring welding procedures where similar

to calculate the amount of heat delivered

material thicknesses and joint configurations

to the workpiece.

are welded. After determining the range of

The formula is:

heat inputs that produce an acceptable weld,

the range of heat inputs can be calculated.

Amps X Volts X 60
Heat Input =
Travel Speed (in/min)

This is really helpful in increasing speeds

for robotic welding. The required travel
speed can be calculated as wire feed speed
(current) and voltage are increased. The

Heat input is measured in units of energy/

unit of length (joules/inch). A joule is a unit
of energy equal to 1 watt of energy into the
workpiece per second. While a joule is not a

formula is also helpful in controlling distortion. Conditions that put less energy into
the metal can be calculated, which reduces
distortion.

familiar unit of measurement, there are a lot

of uses for the heat input formula itself.
For example, suppose that welding is done
at the second condition used as an example
in figure 8. That condition was 300 amps
and 26 volts. For this example, say the travel
speed is 10 in/min. This works out to a heat
input of 46,800 joules/in. If we know that
the welding involves bridging a gap, ESO
can be increased to reduce current and
increase voltage while reaching the first
condition indicated. That was 250 amps and
27 volts, still at 10 in/min. The heat input is
now 40,500 joules/in. By increasing ESO
(stickout), heat input has been reduced by
almost 15%.

13

SECTION 4

4
Shielding Gases

At high temperature, all metals commonly

has very high moisture content. The moisture

used for fabrication will oxidize in the pre-

produces oxygen and hydrogen in the arc

sence of the atmosphere. Every welding

environment.

process provides shielding from the atmosphere by some method. When welding
steels, we want to exclude oxygen, nitrogen, and moisture from the area above the
molten puddle.

Submerged-arc welding shields the puddle

by a different method. As the puddle progresses, the intense heat melts the flux in the
joint area; this forms a slag that covers the
weld and excludes the atmosphere.

In the Oxy-fuel process, the weld pool

is shielded from the atmosphere by the
combustion by-products of carbon monoxide (CO) and carbon dioxide (CO2). In
stick welding (SMAW), CO and CO2 are
also the shielding gases. The 60XX type of

GMAW (MIG) and GTAW (TIG) are both

gas shielded processes in which the shielding
gas is provided from an outside source. No
fluxing agents are included in the filler metal
of solid wires.

electrode uses a cellulosic coating, which

A.
Shielding Gas
Functions

For the purpose of this discussion the

As was mentioned earlier, the atmosphere

GMAW process will be emphasized because

must be displaced while the puddle is cool-

it constitutes the greatest portion of welding

ing or oxidation will occur rapidly. This

done in industry. A good portion of this

appears as a gray surface on the weld bead.

information is applicable to GTAW too.

One cause of porosity is the result of poor

shielding when atmospheric oxygen com-

The major functions of a shielding

gas are to:
1. Protect the puddle from the

atmosphere

metal cools, porosity occurs as this carbon

monoxide escapes from the center of the
bead. If air is aspirated into the shielding gas

2. Provide arc plasma

line through a leak, nitrogen and moisture

3. Provide oxygen for wetting

will also contaminate the shielding gas.

(ferrous alloys)

Nitrogen, while very soluble in the puddle at

4. Control type of metal transfer

high temperatures, will cause porosity as it

5. Affect arc stability

escapes during cooling of the weld bead.

6. Control welding costs.

14

bines with carbon in the puddle. As the weld

The shielding gas also provides a portion of

This is accomplished by ionizing the gas,

the arc plasma, which transfers the welding

which frees electrons to transfer the current

current across the gap between the electrode

from the work to the electrode. Metallic and

and the work as figure 11 shows.

argon ions (atoms stripped of an electron)

transfer the positive charge across the arc.
This explains in part why the arc becomes
very unstable when a torch is hooked up

Figure 11

using straight polarity (DCEN) rather than

Transfer of Current

reverse polarity (DCEP). In DCEN, the posi-

Across the Arc Plasma

tive current is trying to remove iron atoms

from the plate, which are much harder to
melt than a small diameter electrode.
In steels (carbon and stainless), oxygen sta-

bilizes the arc and reduces the surface tension of the weld metal. Oxygen is obtained
DCEP

from direct additions of oxygen or carbon

dioxide to the shielding gas. Surface tension,

Fe+

the force that causes water to bead up on

a waxed car surface, is not desirable when
depositing a weld bead. If pure argon is used

Ar+

instead of a mixed gas, the bead does not

wet out and appears as though it is sitting
on top of the part surface (convex bead).
Figure 12 shows the basic gases used in
shielding GMAW. The ionization potential
is the amount of energy (in electron volts)
Gas

Ar

Characteristic

Totally Inert (Cool)

Ionization

required to establish an arc. The use of a

Potential (eV)

helium rich gas such as Praxairs HeliStar

15.759

A-1025 blend (for short-arc welding of

stainless steel), requires 3-5 volts more than

He

Totally Inert (Hot)

24.587

O2

Highly Oxidizing

13.618

CO2

Oxidizing (Dissociates)

13.769

H2

Highly Reducing

13.598

an Ar/CO2 mixture at the same current. The

shielding gas used also has a pronounced
effect on the type of metal transfer obtained.

Figure 12
Gases Used for
GMAW Shielding

15

There is a best gas for almost every app-

Argon/oxygen mixtures are also very stable,

lication, but there may be 2 or 3 gases that

and are used in steel welding applications.

will do a very reasonable job. Gases are

Pure carbon dioxide provides a less stable

selected on the basis of performance, avai-

arc plasma, but its additions to argon can be

lability, cost, and many other variables.

very beneficial where depth and width of

Further discussion of gases and metal

penetration need to be controlled. Some

transfer is found in section 6.

shielding gases also use additions of oxygen

and carbon dioxide in one mixture (Praxair's

Stability of the arc plasma is another factor

Stargon gas blend).

influenced by the shielding gas. Pure argon

provides a stable arc, and is used when
welding reactive metals such as aluminum.

B.
Flow Rates

Once the shielding gas is selected, it is cri-

Regulator-flowmeters can vary the inlet

tical to make sure that the flow rate is with-

pressure to the flowmeter. As inlet pressure

in certain limits. For low current short-arc

falls when a cylinder gets low, the flow rate

applications, 25-35 scfh (standard cubic feet

is actually skewed to a higher reading. For

per hour) is adequate if ESO is held from

example, a regulator-flowmeter, installed in

3/8" to 1/2". For high current short-arc and

a low pressure line (20 psig), showed a flow

the spray transfer mode, flow rates need to

reading of 70+ scfh, but the actual flow rate

be increased to the 35-45 scfh range. Figure

was 15 scfh. The regulator had reduced the

13 shows the best way to measure the flow

pressure in the flowmeter to about 5 psig

rate using a torch flowmeter.

instead of the correct design pressure of 50

psig. Flow rates must be kept in a controlled
range so that the shielding gas column does

70
60
50
40
30
20
10
0

not become unstable and mixes with air at

turbulent (non-axial flow).

both low and high flow rates or when forced

to flow past an obstruction in the nozzle
such as spatter. This type of flow is called

30 - 70 scfh
measured
at the nozzle

Figure 13
Measuring Actual
Flow Rate

16

Testing has shown that, the shielding gas

The erratic quality of the shielding can pro-

column stays in laminar flow from 30 scfh

vide a weld that looks satisfactory, but can

up to about 70 scfh with a 400 amp gun

contain subsurface (honeycomb) porosity.

(using a 5/8" diameter nozzle). Above

As deposition rates increase, a 35-45 scfh

70 scfh, the flow becomes turbulent and

flow rate is still satisfactory unless there are

mixes with air. A cigarette in an ashtray can

breezes or drafts. In high deposition MIG

illustrate the difference between laminar and

welding, a .045 wire, can be run at 1300 ipm

turbulent flow. The smoke initially leaves

(35 lb/hr) at 50-60 scfh with no problems.

the tip, in a tight orderly laminar flow. A few

Fans and drafts will displace a shielding gas,

inches above, the flow becomes turbulent

and may require increasing flow rates to

and the smoke mixes with the air rapidly as

50-70 scfh. Reducing cup-to-work distance

shown in figure 14. This same thing happens

can also improve shielding.

with a column of shielding gas.

Figure 14
Laminar and
Turbulent Flow

Air
Turbulent
Flow
Turbulent
Flow

Laminar
Flow
Laminar
Flow

Air

C.
Gas Losses

Loss of shielding can cause problems in

GMAW. Inadequate gas coverage can result
in an oxidized surface or porosity. The first
step in troubleshooting what you think may
be a gas loss is the use of a torch flowmeter.
This flowmeter fits over the nozzle on a
torch and measures the actual flow rate of
your shielding gas (figure 13). Compare the
actual reading with that of the station flowmeter (if used). The two readings should be
very close to each other. If not, there may be
some potential problem areas.
17

Threaded Connections are notorious for

If the leak is large enough, no amount of

leaking if not properly sealed. A pipeline

flow can give a good quality weld. Use a

with shielding gas at 50 psi can leak a lot

liquid leak detector to find any leaks.

of expensive gas. It will also allow air to

diffuse in, as figure 15 shows.

O2 Concen

Hose Wall

N2, O2, H2O

Process
Gas

Air

Figure 15

Velocity

N2, O2, H2O

Hose Wall

Back Diffusion Allows

Pressurized Line

Distance

Wall

Wall

Air to Leak into a

C
L

A few shops use quick disconnects for

by disconnecting the supply hose at the wire

their shielding gasses. There are also prime

feeder. Use the torch flow meter to check

suspects when investigating leaks. Use a

the flow out of the hose. If the flow meter is

liquid leak detector solution to check them.

correct, check if a leaking fitting or if the

If the flow measurement at the torch indi-

o-rings in the back of the liner have been

cates a much lower flow than the station

damaged (figure 16). These can be easily

flowmeter, find the source of the problem

replaced, and should be lightly coated with

silicone grease to avoid damaging them

Lube O-Ring with silicone

grease before installing

during reinstallation. Another place where

shielding gas flow can be disrupted is in
the diffuser. The gas diffuser is found at the
point where the contact tip is mounted. Its
purpose is to distribute the gas evenly to
produce laminar flow out of the gas nozzle.
If spatter builds up on the diffuser, it can
clog it and reduce the gas flow enough to
provide poor shielding. If the diffuser is only
partially blocked, the entire gas flow may try
to exit the holes still open and create unbal-

18

Figure 16

anced turbulent flow. This in itself will as-

O-Rings Seal the Gas

pirate air into the shielding gas column and

Ports at the Feeder

may once again, cause porosity.

Figure 17

Holding the torch at too small an angle

Air Aspirated by the

can also create a venturi effect between the

Venturi Effect can

plate and the nozzle. This will also contami-

Contaminate the

nate the shielding gas stream with air and

Shielding Gas

cause porosity (figure 17).

Some welders clean spatter from nozzles by
lightly tapping the gas cup against something
which knocks the spatter out. This can create
Venturi effect pulls air
into the shielding gas

a problem by eventually causing the cable to

gooseneck connection to loosen. When this
happens, it is possible to lose your gas coverage in the torch handle and aspirate air into

Figure 18

the shielding gas stream.

Pressure Correction
Formula for Flowmeters

One final thing to check is incorrect inlet

pressure to the flowmeter. All flowmeters

Example: A flowmeter calibrated at 20 psig

on a 50 psig line indicating 40 scfh

are calibrated for one specific inlet pressure,

and the actual flow reading will be incorrect
if the inlet pressure does not match the cali-

Actual
Flow Rate

Indicated
Flow Rate

Actual Pressure (psia)

bration pressure. Figure 18 shows what hap-

Calibration Pressure (psia)

pens when a 20 psig calibrated flowmeter is

attached to a 50 psig line. The actual flow is

Actual
Flow Rate

= 40 scfh

50 + 14.7

26% higher than indicated on the flowmeter.

20 + 14.7

The easiest way to check this is by using a

torch flowmeter, because it is calibrated for

= 40 x 1.37 = 54.6 scfh

atmospheric pressure at the outlet.

Figure 19

Penetration Patterns

Weld Terminology

The penetration pattern of a weld can be ex-

and Penetration

amined by cutting and etching the weld as

Measurement

shown in figure 19. To examine a weld, cut

through it at 90 to the face. The sample is
easiest to prepare when a saw is used. The
cut face is then sanded with progressively

Effective
Throat

finer sandpaper until a 320 grit paper is

Face

reached. An etchant reveals the penetration

pattern. For steels, there are many different
etchants available. For macro-etching, a
mixture of 10% nitric acid in methanol
works well. This etchant is called nital-10%.

Toe
Root

19

When the etchant is swabbed on the weld

The cooling is caused by the quenching

area, the different microstructures react at

effect of the colder base metal. By measur-

different rates. The darkest area will be the

ing from the face of the weld to the root, we

heat-affected zone, or HAZ. This is base

find the effective throat of the fillet weld as

metal that is adjacent to the weld metal. The

shown in figure 19.

HAZ has been heated to near the melting

temperature and then cooled very rapidly.

Different gases will change the shape and

depth of the penetration pattern. Oxygen
additions will reduce the diameter of the

Figure 20

plasma, and lead to a deep and narrow pene-

Different Gases

tration pattern as shown in figure 20. This is

Provide Different

sometimes called finger-like penetration.

Penetration Patterns

Additions of carbon dioxide and helium

increase the diameter of the plasma because
C-25

O-5

C-15/C-8

of their increased thermal conductivity. This

tends to decrease the effective throat and
broaden the penetration pattern.
Figure 21 shows the penetration patterns of
tubular (flux-cored or metal cored) and solid
wires at the same power (current and voltage) level. The tubular wires are made with a
solid sheath and filled with a powder; most
of the current is carried by the sheath. This
increases the diameter of the plasma, and
makes for a less penetrating or softer arc.
Because of the axial metal transfer character-

Figure 21

istics in spray transfer with solid wire, the

Penetration Patterns

penetration pattern is narrower and deeper

of Solid and Tubular

than that of a tubular wire. Because the pow-

Wires in Spray Transfer

er of the two arcs is the same in the example,

the areas melted in the base metal are equal,
but different in shape. In some welding applications where a solid wire produces burn-

Solid

Tubular

through on thinner parts, a tubular wire may

be a good choice because the larger diameter
plasma spreads the heat over a larger area
and reduces the tendency to burn-through.

Equal Areas
20

Table 3
Shielding Gas Selection Guide for GMAW

Material

Thickness

Transfer
Mode

Recommended
Shielding Gas

Description

Carbon Steel

Up to 14 gauge

Short Circuiting

StarGold C-8,
C-15, C-25
Stargon CS

Good penetration and distortion control to reduce

potential burnthrough; good gap-bridging ability

14 gauge 1/8"

Short Circuiting

StarGold C-8,
C-15, C-25
Stargon CS

Higher deposition rates without burnthrough;

minimum distortion and spatter; good puddle
control for out-of-position welding

Over 1/8"

Short Circuiting

StarGold C-15,
C-25
Stargon CS
CO2

High welding speeds; good penetration and

puddle control; applicable for out-of-position
welds

Globular

StarGold C-8,
C-25
CO2

Suitable for high current and high speed welding

Short Circuiting

StarGold C-50

Deep penetration; low spatter, high travel speeds;

good out-of-position welding

Short Circuiting
and Globular
(Buried Arc)

CO2

Deep penetration and fast travel speeds but with

higher burnthrough potential; high current
mechanized welding

Spray Arc

StarGold O-1,
O-2, O-5
Mig Mix Gold
StarGold C-10, C-15
Stargon CS
RoboStar CS

Good arc stability; produces a more fluid puddle

as O2 increases; good coalescence and bead
contour; good weld appearance and puddle control

Short Circuiting
and Spray
Transfer

StarGold C-5,
C-8, C-10
Stargon CS
Mig Mix Gold

Applicable to both short circuiting and spray

transfer modes; has wide welding current range
and good arc performance; weld puddle has good
control which results in improved weld contour

High Density
Rotational
Transfer

Stargon CS
StarGold C-8
HeliStar CS
RoboStar CS

Used for high deposition rate welding where

15 to 30 lbs/hr is typical; special welding
equipment and techniques are sometimes required
to achieve these deposition levels

Pulsed Spray

StarGold C-5
Stargon CS
Robostar CS
HeliStar CS

Used for both light gauge and out-of-position

weldments; achieves good pulsed spray stability
over a wide range of arc characteristics and
deposition ranges

Short Circuiting

Stargon CS
StarGold C-5
HeliStar CS

Good coalescence and bead contour with

excellent mechanical properties

Gauge

21

Table 3
Shielding Gas Selection Guide for GMAW (continued)

Material

Thickness

Transfer
Mode

Recommended
Shielding Gas

Description

Alloy Steel

Up to 3/32"

Short Circuiting

StarGold C-8,
C-15
Stargon CS

High welding speeds; good penetration and

puddle control; applicable for out-of-position
welds; suitable for high current and high speed
welding

Spray Arc
(High Current
Density and
Rotational)

StarGold O-5
Stargon CS
HeliStar CS

Reduces undercutting; higher deposition rates

and improved bead wetting; deep penetration
and good mechanical properties

Pulsed Spray

StarGold O-2,
C-5, C-8
Stargon CS

Used for both light gauge and heavy out-ofposition weldments; achieves good pulsed spray
stability over a wide range of arc characteristics
and deposition ranges

Over 3/32"

22

SECTION 5

5
Electrodes

A.
Alloying
Additions

Although a welder doesnt often get the

The list shown in figure 22 outlines the

chance to select a filler material, this section

elements added to steel and the reasons for

is included for information. Knowing how

their addition. These are the same alloying

wires are alloyed and why, can sometimes

elements that are added to the base metals,

be helpful when a problem arises. We will

proportions differ slightly for filler metals.

concentrate on a few carbon steel wires, because they are used in the biggest portion
of GMAW welding.

1. Carbon The addition of carbon to iron

has a very strong influence on its properties.
Mild steels, 1010 and 1020 for instance,
have low carbon contents (0.1 and 0.2%

Figure 22

respectively.) Carbon is a very potent

Steel Alloying

strengthener, and when added above about

Additions

0.3%, requires special welding procedures

to keep the material from cracking (preheat,
Steel is iron that is alloyed with

interpass and post heat, etc.) Most common

wires are low in carbon content.

Carbon

Strength

Manganese

Deoxidation and Strength

Silicon

Deoxidation

Aluminum

Deoxidation

before the carbon does so there is little or

Zirconium

Deoxidation

no oxidation of carbon in the weld puddle

Titanium

Deoxidation

to produce carbon monoxide (and cause

2. Manganese This element is added for

three reasons: (1) Deoxidation. Manganese
combines with oxygen in the weld metal

porosity. (2) Desulfurization. Manganese

combines with sulfur to form manganese
sulfides before the sulfur can segregate to
Figure 23
Deoxidation

the grain boundaries and form low melting

Liquids

point iron sulfides. Iron sulfides can cause

Reactions

hot cracking in steels. (3) Strengthening.

Si + 2O -> Si O2
Mn + O -> MnO

Manganese remaining after these other

reactions combine with carbon to form
manganese carbides, which strengthen the

Gases

weld deposit.

C + O -> CO
C + 2O -> CO2

23

3. Silicon This element is mainly added as

4. Aluminum The main function of this

a deoxidizer. Silicon combines vigorously

element is also deoxidation. It is a very

with oxygen in the weld puddle and forms a

strong deoxidizer and forms aluminum oxide

silicon dioxide (see figure 23) slag. Beach

(Al2O3). A secondary function is that of a

sand is silicon dioxide. When the silicon

grain refinement, which produces a stronger,

combines with the oxygen, heat is generated

tougher deposit.

because of this oxidation reaction. This is

one reason why a wire higher in silicon will
provide a more fluid puddle. The brown

5. Zirconium This is also a deoxidizer,

and is used in only a few wires.

glassy solid that forms on the weld deposit

6. Titanium This element is also a

is a combination of silicon dioxide and

deoxidizer in low carbon steels.

manganese oxide.

B.
1. Solid Wire
Designations
and Chemistry

With some understanding of why alloy

additions are made to low carbon wires, the
compositions of some commonly used wires
Rod

are more meaningful. Figure 24 shows the

Solid

American Welding Society (AWS) nomenclature used for solid wires. Table 4 shows

ER70S-X
Figure 24

Electrode

AWS Solid Wire

70,000 UTS
70 - 120 ksi

Designation

24

Table 4

AWS

ER70S - Wire

Electrode

Chemistries

Class.

Carbon

the chemistries of the available ER70Selectrodes.

Chemistry
2, 3, 4, 5, 6,
7, 8, 9, 10, G

Manganese

Silicon

Phos.

Sulfur

(max)

(max)

ER70S-2

.07 Max

.9 - 1.4

.4 - .7

.025

.035

ER70S-3

.06 - .15

.9 - 1.4

.45 - .7

.025

.035

ER70S-4

.07 - .15

1.0 - 1.5

.65 - .85

.025

.035

ER70S-5

.07 - .19

.9 - 1.4

.3 - .6

.025

.035

ER70S-6

.07 - .15

1.4 - 1.8

.8 - 1.15

.025

.035

ER70S-7

.07 - .15

1.5 - 2.0

.5 - .8

.025

.035

Other

TI,ZR,AL

AL

These wire designations are set by the AWS,

(TIG). The third and fourth characters in-

to standardize welding electrodes and filler

dicate the minimum tensile strength of the

metals. The classification system is based

weld metal in thousand psis. An ER70S-X

on chemical composition and strength of the

wire would have a tensile strength of 70,000

deposited weld metal. A typical solid wire

psi. The fifth character, S, indicates that this

designation would be ER70S-3. The E de-

is a solid wire. The number after the dash

signates a wire can be used as an electrode,

indicates the composition classification of

meaning it can carry current. The second

the alloy. These numbers run from 2 to 7 and

character, R, indicates that this alloy is

G. The G classification (stands for general)

available as rod. Rods are usually the 36"

indicates a chemistry agreed upon by the

straight lengths and are used for GTAW

supplier and purchaser.

Since metal-cored wires perform like solid

Where the C Indicates a cored wire

wires during welding, they have recently

and the X indicates the type of shielding

been included for classification purposes in

gas used in qualification of the wire

the AWS specification for solid GMAW

(C=100% Carbon Dioxide, M=Mixed Gas

wires, but they follow a classification pro-

of 75-80% Ar/balance CO2).

B.
2. Metal-Cored
Wire Designations
and Chemistry

cess more like a flux-cored wire than a solid

wire. The two basic wire types are:

E70C-3X
E70C-6X

C.
Flux-Cored
Tubular Wire
Designations

A second classification document covers

Notice there is no R, because straight lengths

flux cored and wires. An example of the

of these wires are not usually practical.

designation for these wires is E70T-X

The second character is the tensile strength

(see figure 25). As with solid wires, E

multiplied by 10,000 psi. The third character

designates an electrode that carries current.

is either a 0 or a 1. A 0 indicated that

the electrode is for use only in the flat and

Electrode

Tubular

horizontal welding positions. A 1 indicated that the electrode is suitable for all
position work. The T in the fourth position
designates this as a tubular electrode. The

E70T-X

digit after the dash indicates the type of

shielding required with this wire.
For Example, an

Figure 25
AWS Tubular Wire
Classification

70,000 UTS
70 - 120 ksi
70 = flat only
71 = all position

Shielding Type
1,4,5,6,7,8,11,G
1,2,5 - with gas
3,4,6,7,8, - w/o gas

E70T1 = 70,000 tensile, flat position

E71T1 = 70,000 tensile, all position

25

Deoxidation of the weld pool is very im-

an S-3 to an S-7, S-6 or even an S-2 solid

portant in metal joining. There are at least

electrode is made. To determine the deoxi-

5 elements added for deoxidation, each

dation potential of a flux-cored wire, the

doing a slightly different job. As a general

manufacturers literature must be consulted.

rule, the rustier or more mill-scaled a plate

Flux-cored wires contain increasing amounts

is, the more deoxidation required from the

of deoxidizers to remove the oxygen being

electrode. The shielding gas is also a source

deposited in the weld puddle by the shield-

of oxidation. If a rusty plate is welded with

ing gas and by any mill scale or rust. Both

a gas of high oxidation potential, a cleaner

mill scale and rust are iron oxides (FeO-

deposit would be obtained if a change from

mill scale, and Fe2O3) .

Silicon and manganese are oxidized in the

The SiO2 and the MnO are liquids that float

molten weld puddle. Silicon forms silicon

to the surface of the puddle. Upon cooling,

dioxide and manganese forms manganese

they are commonly referred to as slag is-

oxide. These reactions occur to remove the

lands. Higher oxygen contents can remove

oxygen and keep the carbon and iron from

much of the silicon and manganese so that

being oxidized. Higher oxidizing potential

the oxygen will then begin to combine with

gases, such as pure carbon dioxide will oxi-

carbon. Carbon monoxide and carbon diox-

dize more of the alloying additions and pro-

ide result, and most of the gas formed is

duce a deposit of slightly lower strength and

carbon monoxide. The gases will evolve

toughness. The oxygen in the weld puddle

from the molten puddle, but if cooling is

will also combine with the carbon to form

rapid and the gas concentration is high gas

carbon monoxide if all of the manganese

bubbles will be trapped and create porosity.

and silicon have been oxidized. Porosity in

Nitrogen is also soluble in the puddle and

a weld is usually caused by the evolution of

will cause porosity if shielding is inadequate.

D.
Slag and Gas
Formation

carbon monoxide. The reactions that occur

in the molten puddle are:

2Mn + O2
2C + O2
C + O2

26

Si + O2

SiO2
2MnO
2CO
CO2


E.
Solidification
of the Weld
Puddle

Figure 26 shows how a weld puddle cools

(and also possibly nitrogen), these gases

and solidifies. At the weld puddle to base

will be pushed to the centerline as the puddle

metal interface, crystals begin to grow into

freezes. This causes porosity along the soli-

the molten weld pool. This is very similar to

dification line and is known as centerline

the growth of ice crystals on a window seen

porosity. Larger amounts of contamination

in time-lapse photography. The crystals are

can cause gross porosity in the weld and

called grains, and where they meet and stop

lead to the condition shown in the bottom

growing is called the grain boundary. As

illustration in figure 26.

the metal solidifies, the solubility of gases

decreases greatly. If there is just slightly
more oxygen in the puddle than manganese
and silicon available for deoxidization

As the weld is finished, the cooling rate at

the crater increases because the weld is
losing heat in all directions. This rapid
cooling rate leaves less time for the gases
to leave the puddle, and a hollow gas cavity

Figure 26

can form at the crater.

Solidification of
a weld

Calculation of Deposition Rates

A very useful piece of information needed
when calculating the cost of welding is the
deposition rate. The deposition rate is usually stated in pounds of wire per hour of
weld time. Figure 27 shows the multipliers
that can be used to determine deposition
rate for different diameters of solid wires.
The multiplier is a factor that takes into
account the cubic inches of wire per hour
consumed and the density of steel, to arrive
at the rate in pounds per hour. To calculate
the deposition rate of an .045" diameter
wire at 500 ipm wire feed speed, multiply

Wire Diameter

Multiplier

the 500 ipm times the .027 multiplier, and

.030

.012

determine a rate of 13.5 lbs/hr of arc-on

.035

.0163

.045

.027

.052

.0361

actually on) and the deposition efficiency

.0625

.0521

of the process.

time. To determine the actual amount of

metal deposited, multiply this weight by
the duty cycle (% of an hour that the arc is

Example: an .045 wire at 500 ipm

500 ipm x .027 = 13.5 lbs/hour

Figure 27
Calculating Deposition
Rates for Solid Electrodes

27

SECTION 6

6
Metal Transfer

An understanding of metal transfer is very

obtained. In this section, electrical charac-

helpful when trying to solve a welding

teristics, wires and shielding gases all come

problem such as why is there so much

together. There are four major types of metal

spatter?, or how can more penetration be

transfer that will be discussed. They are:

1. Short-Arc

Figure 28

2. Globular Transfer

The Pinch Effect

3. Spray Transfer

Tries to Pinch Off

4. Pulsed Spray Transfer

the End of the

Electrode

Figure 28 shows the pinch effect. The pinch

effect is a function of current, and tries to
pinch off the molten tip of the electrode.
P

P
Short-Arc

Globular

P
Spray

The pinch effect is a function of current

Higher currents and smaller areas increase

the pinch effect and give cleaner metal
transfer with less spatter.

A.
Short Arc

Short-arc, or short-circuit transfer, is basi-

This process is a good choice on thin

cally a low heat input, low penetration

material and sheet metal and has been used

process. Currents range from 40-50 amps

extensively for out-of-position MIG weld-

(.023" wire) up to 250-275 amps (0.052"

ing. Short-arc is also a good choice where

diameter wire). Voltage ranges from 14-21v.

bridging gaps is a problem.

This form of metal transfer is called short-

Figure 29

arc because the wire does electrically short

In Short-Arc Transfer,

to the workpiece. When the wire touches

the Electrode Shorts

the base material, the arc goes out, and the

60-120 Times Per

current flowing through the wire begins to

Second

rapidly raise the temperature of the wire.

As seen in the power supply characteristic
curve, at 0 volts the power supply tries to
Current Voltage

produce a maximum current output. When

the wire reaches its melting point, it flows
into the puddle and the arc reignites. This
shorting takes place very rapidly, from 60
to 120 times per second. Figure 29 shows
Time (milliseconds)

the droplet of molten weld metal pinching

off just before the arc reignites in the third

28

illustration.

After adjusting the wire feed to do the job,

A power supply that has an inductance

the voltage can be fine-tuned to where the

control is easier to use for short-arc transfer.

sound from the arc becomes smooth and

As shown in figure 10 adding inductance

very regular. Electrical stick-out must be

slows the rate of current rise. With this

closely controlled as it has a great impact

available, after wire feed and voltage are

on current levels. At the low-end condition

adjusted, inductance should be increased to

of 40-50 amps and 14-15 volts with a .023"

the point where the metal begins to transfer

electrode, stick-out should be about 1/4".

smoothly. By increasing the inductance,

With a .035" electrode at 80-90 amps and

the current rises at a slower rate, and is at a

15-16 volts, ESO should be about 3/8". At

lower level when the cycle begins again than

the high end of short-arc transfer with a

it would be without the inductance in the

.045" electrode, current will be at 225-235

circuit. Limiting short current reduces ex-

amps and 20-22 volts. Because of the high

plosive, harsh metal transfer. Most short arc

current levels, we increase the ESO to pre-

welding is done with .023", .035" or .045"

heat the electrode and reduce the current.

electrodes. Larger wires require too much

The ESO with these conditions would be in

current for most applications.

the 5/8"-3/4" range. This method gives a

very controllable arc. If the arc is unstable,
the usual cause is that the ESO is too long
or voltage is too high. There should be very
little spatter with this process, regardless
of shielding gas. Argon mixtures, however,
provide smaller droplets with better gap
bridging and arc stability as a result.

Gases that work well in short-arc span

from C-8 through C-25 mixtures (Praxair's
StarGold blends), straight CO2 and
Praxair's Stargon CS gas blend. For low
current applications carbon dioxide is sometimes a good choice. The arc is hotter than
with an argon mixture, and at low current
levels, there is not much spatter. For in-

To adjust the power supply for short-arc

creased deposition rates and travel speeds,

transfer, two variables can help, if they are

C-25, C-15, or C-8 mixtures will usually

available. The slope on some machines is

provide better results with decreased spatter

adjustable, either externally, or with internal

levels. Reducing the amount of carbon

taps. Figure 7 shows that the steeper curve

dioxide makes the puddle less fluid and

will limit the maximum current that the

easier to control. Burn-through is also

power supply can deliver, which is a real

reduced. On thinner materials, gases lower

benefit when short-arc welding at low cur-

in carbon dioxide content work best by

rents on thin materials. If the majority of

minimizing burn-through and permitting

your work is in the short-arc range, it is

higher currents and travel speeds. For

probably worthwhile to change an internal

welding higher alloys, like stainless steel,

tap if the machine has one. An internal tap

helium is sometimes added to the shielding

isnt something you change for each job.

gas to increase heat input at lower

deposition rates.

29


B.
Globular
Transfer

Globular transfer is usually not the recom-

The arc is continuously moving to the place

mended way to deposit weld metal because

where the glob of metal is closest to the

of the inefficiency of the process. This

work, where the minimum voltage is re-

type of transfer produces the most spatter.

quired to sustain the arc. This creates the

Depending on the current range, shielding

instability that you see and hear in the arc.

gas, and power supply settings, globular

When the surface tension and the force of

transfer can waste 10-15% of the weld metal

the arc are finally overcome by gravity the

as spatter. Because of the inefficiency of the

glob transfers. As the glob of metal hits the

process, slower travel speeds or smaller bead

work, it tends to splash, throwing spatter

sizes result at wire feeds comparable to

out of the puddle onto the work.

spray or short-arc transfer.

Globular transfer occurs when voltages and
When the tip of the wire begins to melt in

currents exceed that of the short-arc range.

globular transfer as shown in figure 30, it

Other than short-arc, this is the only type of

only shorts to the workpiece occasionally

transfer you get with carbon dioxide in the

due to higher voltages. The inconsistent

current range used in industry. If you are

cracks and pops you hear are the breaking

using gas blends like Praxair's Stargon CS

of the short circuits. Unlike short-circuit

and StarGold (Ar/O2 or Ar/CO2), globular

transfer, an arc is present most of the time,

transfer is what you get when voltage or

and the metal begins to form a ball on the

current falls below the spray transfer range.

end of the wire. This ball is held by the

This is where spatter develops when a spray

surface tension and the force of the arc.

arc mixture is used improperly.

Unstable Arc
High Spatter Levels

Figure 30
Globular Transfer
Produces High
Levels of Spatter

30


C.
Spray
Transfer

Spray transfer is a very clean, high efficiency

Spray transfer can be used on materials as

process. All wire diameters can be used. For

thin as 14 and 16 gauge metals with the

most applications in the 175 amp to 500 amp

right wire diameter (.023"). Thicker section

range, .035" to 1/16" wires work well. When

welding is where spray really gains an ad-

the welding equipment is set up properly,

vantage, especially in the flat and horizontal

there is almost no spatter and 97-98% of the

positions. This type of metal transfer can

filler weld is deposited in the weld puddle

be used out of position but wire diameter

(deposition efficiency).

should be smaller and the operating conditions less than in the flat position. All steels

In spray transfer, the tip of the electrode becomes pointed as figure 31 shows. Because

(carbon and stainless), and most other materials, can be GMAW welded in spray transfer.

the tip is so small, the current density (amps/

square inch) and the pinch force are very

The gases used for spray are lower in active

high. This pinches off metal droplets that are

gases (CO2 and O2) than gases for short-arc

smaller than the diameter of the wire. The

and globular transfer. Most contain from

droplets are accelerated by the magnetic

85-90% argon, and some blends contain

field, around the arc instead of transferring

both carbon dioxide and oxygen. Some of

by gravity as in globular transfer. The small

the newer gases also contain small additions

droplets are absorbed into the weld pool

of helium (Praxair's HeliStar gas blends)

rather than splashing.

to increase the energy in the arc.

Droplets smaller
than diameter of
electrode
Very low spatter
Minimum voltage
and current
required
Figure 31
Spray Transfer
is a Very High
Efficiency Process

31

Transition Currents
Figure 32

(Steel and Stainless Steel)

Transition Currents

To set a welding system for spray, there are

for 95% Ar/5% O2

minimum voltages and currents required.

Shielding Gas

Voltages range from 24 v (small diameter

with Ar/O2 ) to 30 v (hi-deposition with He).
A good place to start is around 26-27 volts.
To estimate what the minimum transition

.035 x 10,000 = 350/2 = 175 amps

current for spray transfer would be, multiply

.045 x 10,000 = 450/2 = 225 amps

the wire diameter (in thousandths of an inch)

by 10,000 and divide by two as figure 32

.052 x 10,000 = 520/2 = 260 amps

shows.
.0625 x 10,000 = 625/2 = 312amps

This approximation is accurate for a 95%

Ar/5% O2 shielding gas. If you are using a
gas with 10% CO 2 , a closer approximation
is made by adding the % CO2 to the transition current calculated above and tabulated
Figure 33

in figure 33. For example, a .045" wire

Transition Currents

with C-10 would produce spray transfer at

with Various

approximately 225 + 10 = 235 amps and

Shielding Gases

about 27 arc volts. Figures 34 and 35 show

the spray transfer ranges for 95% Ar/5% O2
(figure 34) and the short-arc and spray

32

Wire

O-5

C-5

C-10

C-15

.035

175

180

185

190

.045

225

230

235

240

.052

260

265

270

275

.0625

320

320

325

330

transfer regions for Ar/8% CO2 (figure 35).

Figure 34
Spray Transfer
Ranges for
95% Ar/5% O2
.035" and .045"

36

.035" and .045"

wire with
O-5 gas

Electrodes

Hiss

.045

Spatte

Voltage

32

28

.035

ter

Spat
24

22

18
100

200

300

400

Current

Figure 35
Short-Arc and Spray
Transfer Ranges for

.045" wire
with C-8 gas

Ar/8% CO2 (C-8)

with 0.35" and 0.45"

40

Electrodes

Hiss

.045

Spatte

Voltage

35

30

.035

ter

Spat
25

20

Short-arc
15
100

200

300

400

Current

33


D.
Pulsed Spray
Transfer

Pulsed spray transfer is a process that com-

Due to its low heat input, pulsed spray is

bines the lower heat inputs associated with

beneficial for out of position work and for

short-arc with the clean metal transfer and

filling gaps. Since it can produce high peak

good penetration associated with spray trans-

currents, a larger wire can usually be used at

fer. A graph of current vs. time (figure 36)

lower deposition rates. A larger wire (.045

shows the shape to be a square wave. The

instead of .035) will usually reduce wire

current at the top of the square wave is call-

costs and reduce wire feeding problems,

ed the peak current, and the current at the

especially for materials such as aluminum.

bottom of the square wave is called the background current. The background current
keeps the arc lit, but at very low currents
typically 20-40 amp. When the current rises
to the peak current, one droplet is transferred
in spray transfer. Because of the small size
of the droplet, spatter is minimized and
penetration is maximized due to the spray
transfer.

Recent research has shown that inverter pulsed power supplies with very rapid current
rise can reduce the fume associated with
higher current GMAW welding. The fuming
is caused by superheating the molten tip of
the wire and causing the metal to boil. The
very rapid current rise reduces the superheating, leading to the reduced fume generation rates.

Peak Current

Current

275

140

Average Current

20
Background Current
Time (milliseconds)

Figure 36
Pulsed Spray
Transfer Produces
Low Heat Inputs
With Very Clean
Transfer

34

SECTION 7

7
Welding of High Strength Steels

Higher tensile and yield strengths can be

The ASME (American Society of Mech-

achieved by increasing carbon content, add-

anical Engineers) grades for these alloy

ing alloys, or a combination of both. In the

steels are ASME SA517, grades B and F.

section on materials, it was seen that there

are hundreds of different steels available
today. US Steels T1 construction alloy
will be used as an example of an alloy steel.
Most of the major steel producers now make
similar High Strength Low Alloy (HSLA)
steels, which are designated A514 and A517
grades B, Q, H, and F by the ASTM.

Table 5
Mechanical Properties
Comparison of 1080

Tensile
Yield

A comparison will be made between T1

steel and a carbon steel of comparable
strength (SAE 1080) in table 5 to look at
the procedures required to weld the two
different materials. It is usually easier to
weld an alloy steel than a carbon steel of
equivalent strength.

1080

T1

T1A

T1B

T1C

112,000

110,000

110,000

110,000

110,000

61,000

100,000

100,000

100,000

100,000

High Carbon Steel and

4 T1 Construction
Alloy Steels

Elongation

10%

18%

18%

18%

18%

% RA*

25%

40%

40%

40%

40%

* Reduction of Area

The tensile strengths of the four materials

(stretch), and the structural members would

are all in the 110,000 to 112,000 psi range.

take a permanent set instead of returning to

Most products are designed using the mater-

their original shape. Elongation and reduc-

ials yield strength; in this case there is a

tion in area are measures of the ductility of

dramatic difference in the yield strengths

the material. A ductile material will deform

of the materials. Yield strength is where

instead of fracture under severe loading.

the material begins to plastically deform

35

T1

T1 A

T1 B

T1 C

1080

Carbon

.1 - .2

.12 - .21

.12 - .21

.14 - .21

.78 - .89

Manganese

.6 - 1.0

.7 - 1.0

.95 - 1.3

.95 - 1.3

.6 - .9

Phos (max)

.035

.035

.035

.035

.04

of Equal Tensile

Sulfur (max)

.04

.04

.04

.04

.05

Strengths

Silicon

.15 - .35

.2 - .35

.2 - .35

.15 - .35

Nickel

.7 - 1.0

Copper

.15 - .5

Chrome

.4 - .65

.4 -.65

.4 - .65

1.0 - 1.5

Molybdenum

.4 -.6

.15 - .25

.2 - .3

.4 - .6

Vanadium

.03 -.08

.03 - .08

.03 - .08

.03 - .08

Boron

.0005 - .006

.0005 - .005

.0005 - .005

Table 6
Comparison
of Chemical
Compositions of
1080 and T1 Alloys

Titanium

1.2 - 1.5

.01 - .03

Welding of high strength steels is different

than welding low carbon steels. The welder
must pay a lot more attention to detail when
welding high strength steels.

A.
Select the
Proper
Filler Metal

If standard T1 steel were to be welded the

tensile strength of the material should be
considered. From table 5 it is seen that the
tensile strength is 110,000 psi. For the
GMAW process, an ER110S-1 electrode
could be used. The 110 indicates a tensile
strength of 110,000 psi. Some manufacturers
also develop electrodes that do not exactly fit
in the AWS classifications. These electrodes
can also be used, but a welding engineer
should look at the mechanical properties to
ensure compatibility. If the FCAW process
were to be used, an E110T5-K3 (Mn-Ni-Cr)
or an E110T5-K4 (Mn-Ni-Cr-Mo) would
be selected according to the mechanical
properties required.

36


B.
Minimize
Hydrogen
Contamination

Molten weld metal is capable of absorbing

3. Moisture

considerable amounts of hydrogen. Hydro-

Moisture can come from surface contami-

gen in weld metal causes two problems,

nants such as rust and mill scale, and it can

porosity and cracking. Hydrogen can come

be adsorbed on the surface of a clean sheet

from a variety of sources, and all of them

of steel. Bringing a piece of high strength

can cause problems. Hydrogen is normally

steel inside in the winter will sometimes

found in nature as a diatomic molecule, H2.

cause moisture to condense on the surface.

For T1, it is recommended that the material

In the arc, the hydrogen immediately dissociates to monatomic hydrogen (H), which is

temperature be at least 70 oF before welding

begins.

the smallest atom known. A single hydrogen

atom is about 1/100,000 the diameter of an

Moisture can also come from the shielding

iron atom. The hydrogen atoms are very sol-

gas if there is a leak in the supply line as

uble in the weld metal. Some of the hydro-

discussed in the shielding gas section. Air

gen sources are:

contains oxygen, nitrogen, and moisture, all

1. Grease, Oil, Oxidation or Paint on the

of which are detrimental to the integrity of

the weld metal.

Part or Electrode
Grease and oil are both hydrocarbons, and

As the puddle freezes, the solubility of

in the heat of the arc (about 10,000 oF), will

hydrogen in the puddle decreases. The

rapidly dissociate to produce hydrogen, car-

hydrogen tries to escape from the weld

bon, and other contaminants in the puddle.

metal by two different mechanisms.

The carbon can over-harden the weld metal

with possible carbon increases of .1 to .25%.
Hydrogen goes into solution in the weld
metal. Some paints are also hydrocarbon
based, leading to the same problems as oil
and grease. Mill scale and rust can also contain moisture. When the arc heats these materials, the water molecules are dissociated
into hydrogen and oxygen. Both of these
gases go into solution in the weld metal and

Porosity Formation
If the concentration of hydrogen in
the matrix exceeds its solubility in the
solidified weld metal, the excess hydrogen can form bubbles of hydrogen gas.
These bubbles can show up as porosity
in an x-ray. At very high concentrations
of hydrogen they can cause visible porosity at the face of the weld.

can cause weld problems such as porosity

and cracking.
2. Excess Drawing Lubricants

on the Electrode
Drawing lubricants on the wire electrode can
also contain hydrocarbons. When they are
exposed to the arc heat, they dissociate and
contaminate the puddle.

37

 Slowly Diffuses Out of the Weld Metal

hydrogen content is high enough, there

After welding is completed, hydrogen

are visible streams of bubbles rising from

continues to rapidly diffuse out of the

the coupon. The rest of the diffusible hy-

weld metal. In lab tests for measuring

drogen will escape within 20 to 30 days.

diffusible hydrogen, the welded coupon

The remaining hydrogen in the weld metal

is quickly inserted in a bath of liquid

is called residual hydrogen, and it can

nitrogen. The sample is then placed in

cause cracking problems after welding.

mercury, and the quantity of hydrogen

released is measured over time. If the

During the welding process, the base metal

in contact with the molten puddle was subjected to temperatures very close to the

Figure 37

melting temperature of the material. In that

Hydrogen is Driven Into the

period, hydrogen atoms also diffuse into the

HAZ by Stress Gradients

base metal in what is referred to as the heat

affected zone (HAZ - see figure 37). The
solubility of hydrogen increases with temperature even when the metal remains a
solid.
The diffusion of the hydrogen into the base
metal in conjunction with the very rapid
cooling rate can lead to underbead cracking
or delayed brittle fracture.

HAZ

C.
Control
Heat Input

38

To understand the welding of the higher

metal and the HAZ is very important in

strength steels, it is necessary to know a little

determining the properties of the material.

about what happens when steels solidify. As

A familiar microstructure found in metal

molten metal cools, it undergoes a transfor-

files is called martensite. Martensite is very

mation from a phase called austenite to a

strong, with a tensile strength of 200,000 psi

number of different structures. A low carbon

or more, but the microstructure is very brit-

steel like a 1008 grade changes from aus-

tle. Martensite can be formed by very rapid

tenite to ferrite, which is a very soft and

cooling of the material. If a file is heated to

ductile microstructure. Adding slightly more

1300 oF - 1400 oF and then cooled very slow-

carbon and cooling more slowly forms a

ly, a hole can be drilled in it. If it is then re-

microstructure called pearlite, which is fer-

heated and quenched in oil, martensite will

rite and cementite (Fe3C). When carbon and

reform and the file can be used again. There

other alloying elements are added, the trans-

is also an intermediate microstructure called

formation from austenite is modified to form

bainite, between pearlite and martensite,

harder and sometimes more brittle micro-

which has properties that fall between the

structures. The cooling rate of the weld

two.

Metallurgists use charts called continuous

base metal right next to the weld metal that

cooling transformation (CCT) diagrams to

almost reached its melting temperature. The

determine the cooling rates required to ob-

cooling rate of the HAZ is much faster than

tain certain microstructures. A CCT diagram

the weld metal because the base metal is

tells what microstructure to expect when

acting like a quenching medium. This is the

cooling occurs at different rates. At the bot-

reason that the heat input must be controlled

tom of the chart are three different scales

during the welding of the high strength

for air, oil, and water quenching. Air cool-

steels. For T1, higher heat inputs lead to

ing produces the slowest cooling rate with

grain growth in the HAZ, and the strength

a water quench being the fastest. Welding

of the material is reduced.

on a piece of base metal that has just been

To control the cooling rate, minimum

brought in the shop from the yard on a very

preheat and interpass temperatures are

cold day can produce some very rapid

specified to avoid a brittle crack sensitive

cooling rates.

HAZ and weld metal.

The reason why a CCT diagram is mentionTable 7 below shows the maximum heat

ed is that weld metal and the heat affected

inputs for T1 for different thicknesses and

zone (HAZ) also go through the same cool-

preheat and interpass temperatures.

ing process after welding. The HAZ is the

Preheat and Interpass Temperature (oF)

Table 7
Maximum Allowable
Heat Input for T1

70

150

200

300

400

27

23

21

17

13

1/4"

36

32

29

24

19

1/2"

70

62

56

47

40

3/4"

121

107

99

82

65

1"

Any

188

173

126

93

1 1/4"

Any

Any

Any

175

127

1 1/2"

Any

Any

Any

Any

165

2"

Any

Any

Any

Any

Any

Thickness 3/16"

Steel in Kilojoules
Per Inch of Weld

39

The thicker the material is the higher the

Any hydrogen in the weld metal will diffuse

allowable heat input. Some of the thicker

into the highly strained microstructure and

materials have no maximum heat input be-

could lead to underbead cracking. Hydrogen

cause there is sufficient material to ensure

in a strained microstructure can also lead to

a rapid cooling rate with currently available

a problem called delayed brittle fracture.

welding processes. To check interpass tem-

Delayed brittle fracture is a weld defect that

perature, use a temperature indicating crayon

does not show up immediately. The product

1/2" to 1" away from the joint. It is possible

can be in service for a period of time, and

to contaminate the weld with the temperature

then fail at loads well below the yield str-

stick, and it is the temperature of the base

ength of the material. Since the hydrogen

metal, not the previous pass, that is

atom is about 1/100,000 the size of an iron

important.

atom, it can diffuse through the metal easily.

Stress gradients in the material cause the hy-

If the preheat and interpass temperatures for

T1 steel are compared with those for 1080, a
real difference is seen. The 1080 steel has a
carbon equivalent of up to 1.03%. The steel
with 1% carbon would require preheat and
interpass temperatures of between 600 oF
and 800 oF depending on thickness. A post
weld heat treatment at high temperature
would also probably be required to prevent
cracking. Compared to the preheat requirements (70 oF) in table 7, it is easy to see
that T1 steel is much easier to weld than the
equivalent strength carbon steel. The alloying additions in T1 steel undergo transformation to carbides at higher temperatures than
the carbon steels, which allows the greatly
reduced preheat and interpass temperatures.
For materials that require a slow cooling rate
to avoid cracking, it is also very important to
control preheat and interpass temperature. A
highly strained microstructure, such as that
created by rapid cooling, is very sensitive to
hydrogen.

40

drogen atom to migrate and concentrate in

areas of high triaxial (3 dimensional) stress.
In the matrix, these areas would be at the tip
of a microcrack, a grain boundary imperfection, or at the base of a surface imperfection,
such as undercut. When the hydrogen concentration reaches a critical level, the crack
grows slightly to relieve the stress. The diffusion mechanism begins again at the tip of
the new crack, and the process repeats itself.
As the crack continues to grow, the material
resisting the load decreases. At some point,
rapid crack growth occurs, and there is
catastrophic failure.


D.
Use the
Correct
Technique

The correct technique incorporates all of the

When beginning to weld, it is very good

topics just mentioned, such as controlling

practice to use the back step technique. The

hydrogen and watching heat input. With high

back step technique involves starting to weld

strength materials, it is also very important

1/4"-3/8" ahead of where the beginning of

to watch torch angle. The ideal fillet has a

the weld is required, and then backing up

flat face and the toes blend into the side wall

to the beginning. The beginning of a weld

smoothly. No undercut can be tolerated in

usually cools too rapidly because the base

high strength materials, because undercut

metal has not been preheated from the heat

acts just like a notch to greatly reduce the

of the arc. Back stepping begins to preheat

stress at which the component fails. A bolt

the material, and greatly reduces the possi-

that fails in the first thread illustrates how

bility of lack of fusion because the material

undercut can affect weld properties. The

where you strike the arc will be re-melted.

thread, just like undercut, (called a stress

riser) is basically the beginning of a crack.
Undercut will also increase the strain in the
material directly below it and make it more
sensitive to any residual hydrogen that may
be in the material.

Weaving is also not recommended on higher

strength materials. A stringer bead technique
gives better results, and helps control heat
input. If heat input were calculated using a
weaving technique, it would probably be
difficult to control the maximum heat input
due to the slow travel speeds associated with
weaving.

41

SECTION 8

8
Technique and Equipment Set-Up

Technique is very important when welding

puddle position, vertical down welds,

any type of material, and gets more impor-

gaps, crater filling, and arc starting will

tant as the material strength increases. Torch

be discussed here.

angle, feed roll tension, burnback, arc and

A.
Torch
Angle

There is a specific amount of energy avai-

elevated temperature of the base metal, the

lable from the arc to heat and melt the base

bead will cool more slowly. This allows the

metal. Torch angle plays a very important

face of the weld to come to equilibrium and

role in the shape of the bead and the depth of

will give a relatively flat face. If a lagging

penetration into the base metal as figure 38

(or drag) angle is used, very little of the arc

shows. A leading (or push) angle will use

energy goes into preheating the base metal,

some of the arc energy to preheat the base

and deeper penetration is the result. Because

metal before it is welded. Because of the

of the lack of preheat, the bead will tend to

be convex (humped) because the weld will
cool more quickly. If a line were drawn bet-

Travel

ween the toes of the weld, the weld metal

above that line is wasted. If a fracture were
to occur, it would start at the toe of the weld,
not through the thick section. In fatigue service, a humped bead actually reduces the
service life of the component. The decreased
Lead Angle (Push)

Lag Angle (Pull)

reentry angle (angle that the bead face makes

with the base metal) tends to raise the stress
level at the toe of the weld. A flat bead face
distributes the stress more evenly across the
joint. A slight drag angle does work well in
a deep groove, and also in the first pass of a
multi-pass weld. As a first pass, the benefit
of increased penetration is obtained, but it is
critical to be sure to melt out the toes of the
first pass when the cap passes are put in. A

Figure 38

lack of fusion at the toe of the first pass is a

Torch Angle Affects

defect that will reduce the service life of the

Penetration and

component.

Bead Shape

42


B.
Feed Roll
Tension

Setting feed roll tension is important to

the wire to deform it between the feed rolls.

improving consumable life and reducing

Figure 39 shows the shape of the wire as

downtime due to feeding problems. How

it leaves a two U-grooved roller setup with

many times have you seen a welder have a

excessive pressure. The fin will be scrap-

feeding problem corrected by increasing

ed off as it feeds through the liner, and will

feed roll pressure? There are two main types

make it more difficult to feed the wire as

of feed roll designs used for solid wires, one

the excess material begins to clog the liner.

grooved and one flat roll, or two grooved

Excessive pressure on the feed roll will wear

rolls. Increasing feed roll tension on either

out the feed rolls, plug up the liners, and

design can actually put enough pressure on

usually lead to a burnbacks at the tip.

The best method of adjusting feed roll ten-

Figure 39

sion involves running about a foot of wire

Feed Roll Tension Adjustment

out of the gun. Bend the wire 180o to form

Too much feed roll pressure deforms the wire

and causes slivers to form

a curved end and run the wire into a gloved

hand. Pull the trigger and slowly adjust the
feed roll pressure until the wire will make

Slivers clog the liner and make the wire

harder to feed

the turn and feed smoothly. At this feed roll

pressure, the wire will feed without deform-

Apply just enough tension that the wire feeds

into a gloved hand and exits at 180

ing and the rolls will slip if you get a burnback instead of bird nesting.

C.
Burnback

The burnback control is designed to reduce

the stick-out at the end of a weld and keep
the wire from sticking in the puddle. When
welding begins, the trigger on the gun clos-

Wire feed speed

Burnback is a timer that controls the

ESO for the start of the next weld

es the contacts in the welder, opens the gas

Contactor
off

solenoid, and starts the wire feed motor as

shown in figure 40. When you release the
trigger at the end of the weld, the gas solenoid and the contactor can react very quickly. The motor has the inertia of the armature,
the gearbox, and the feed rolls to overcome,

Time

so it does not stop instantly. What a burnGas on

contactor
closed feed
motor on

Motor up
to speed

Motor
off

Gas off

back control does is put a timer in the circuit

that delays the opening of the contactor and
the closing of the gas solenoid. This allows
the motor to spin down and the wire to con-

Figure 40

tinue to burn off so you dont have to try to

The Burnback

start the next weld with too much stick-out.

Control Is Essentially

A short stick-out greatly improves the start-

a Timer

ing of the arc because more current is

available.
43


D.
Arc and
Puddle Position

It is very important to watch the position of

your puddle in relation to the arc. In order
to get good fusion into both pieces of base
metal, the heat from the arc must be directed onto the base metal (figure 41). If a very

If you get on top

of the puddle, lack
of fushion and coldlap can result

large fillet is attempted in one pass, the tendency is to try to hold the puddle back with
the arc. The puddle will continue to advance,
and trying to hold the puddle back with the
arc will lead to incomplete fusion into the
side wall. When the puddle rolls under the
arc, the heat of the arc is no longer being
put into the base metal, it is going into the
puddle. This increases the temperature of the
puddle and reduces the fusion into the walls.
This is sometimes called slugging the

Figure 41

joint. This is a very poor practice and leads

Stay in Front of the

to welds that will fail at well below their

Puddle with the Arc

design strength.

E.
Vertical Down
Welding

Vertical down welding (figure 42) can be

done properly, but the weld parameters must
be set very carefully. The problem with vertical down welding is that it is very easy to
Use a drag angle

get lack of fusion into the side wall. If you

Stay in front of the

puddle

have watched a welder putting in a vertical

Use short-arc pulsed spray

or low current spray transfer
Use a smaller
diameter electrode

down weld and seen the molten metal hitting

the floor, this is very similar to slugging a
joint mentioned before in that the only path
to the floor is between the arc and the base
metal. If the weld metal is taking this path,
the arc energy cannot be going into the base
metal. A high quality vertical down weld can
be made in either short-arc or spray transfer,
but the arc must stay in front of the puddle.

Very little
metal should
fall past the
puddle

Here is one more place where a lagging angle is beneficial, because the arc force can be
used to hold the puddle in place as it begins
to cool.

Figure 42
Vertical Down Welding Can
Produce a High Quality Joint

44


F.
Gaps

Gaps can cause a lot of problems because

Reducing the electrode diameter reduces

they vary in consistency and also change the

the current due to the higher resistance of a

maximum heat input that the part can handle.

smaller wire. A .035" (.9mm) wire will give

A real problem arises on critical parts where

the same deposition as a .045" (1.2 mm) at

depth of penetration is important, and, there-

about 20 amps less because of the higher

fore, the heat input is near the upper end of

resistance of the smaller electrode. Shielding

the range that the part can handle. As soon

gas mixtures with lower carbon dioxide con-

as a gap appears, the heat input must be re-

tents can also reduce the energy in the arc

duced, or burn-through will occur.

and make gaps easier to bridge. If a C-25

mixture is used, try C-8 for more gap tol-

The best solution is to fix the parts so the

gaps dont exist, but, this is usually not possible. In a manual welding operation, the
solutions are a lot easier than in a robotic
operation. A welder is a lot smarter than a
robot, and can quickly adjust conditions to
reduce heat input. Increasing the electrical
stick-out can reduce the heat input enough

erance. Consider also, changing to short-arc

transfer if spray is currently used. Short-arc
is a lot better at filling gaps due to the reduced heat input. Another very good solution is
to reduce heat input with the use of pulsed
spray transfer. Pulsed spray allows low heat
inputs with almost no spatter, and makes it
easier to fill gaps.

in some cases to eliminate the burn-through.

Remember that increasing ESO reduces current and increases voltage slightly, reducing
heat input.

Use short-arc pulsed

spray or low current
spray transfer
Use smaller electrodes
Reduce the CO2 in the
shielding gas
Increase ESO
(stick-out)

Figure 43
Welding Gaps

45


G.
Crater
Filling

The crater is the last bit of weld metal to

The solution to this is to spend a little more

freeze at the end of a weld. Due to the

time filling in the crater before ending the

shrinkage that occurs as the weld metal

weld. This will add a little more weld metal

cools, the crater will sometimes appear

to increase the reinforcement and also put

concave. This can cause problems because

some additional energy into the weld to slow

it is highly strained weld metal with very

the cooling rate. Slowing the cooling rate

little reinforcement because of the conca-

will allow more time for the gases in solu-

vity. Highly strained weld metal with very

tion to escape through the face of the weld.

little reinforcement can develop cracks very

easily. Porosity in the crater is also possible.

H.
Arc Starting

Initiating the arc is mainly a function of the

The current rise begins when the arc starts,

current available. The biggest cause of poor

is slowed by the additional resistance in the

arc starting is too much stick-out. The extra

welding circuit. Replacement of the poor

resistance heats the wire and reduces the

connections will restore arc starting. Sealing

current available for arc initiation. When the

the exposed copper with a liquid electrical

wire has heated enough to soften it buckles

tape will prevent oxidation from occurring.

and the arc goes out. Then the whole process

begins again, leaving little stubs of wire at
the start of the weld.

At higher current levels, also consider the

slope and inductance setting on the power
supply. A steep slope or high inductance set-

Another problem seen on a regular basis is

ting is designed to limit short-circuit current.

poor connections on the work and the hot

If the material to be welded is not too thin,

leads. When copper is new, it is bright and

a flat slope and/or minimum inductance will

highly conductive. As it ages, the copper

improve the arc starting by increasing the

surface oxidizes and acts as an insulator.

rate of current rise.

This resistance causes the cables to get hot

during welding.

46

SECTION 9

9
Weld Discontinuities and Defects

The AWS defines discontinuities as interruptions in the typical structure of a weldment.

A discontinuity only becomes a defect when
the component is rendered unserviceable.

A.
Lack of
Fusion

Lack of fusion, is little or no penetration and

torch moves back into the puddle, the molten

tie-in to the base metals, as shown in figure

metal doesnt stop and wait. It continues to

44. This can be caused by many different

advance, but without the benefit of the in-

reasons. Some of the common ones are:

tense heat of the arc to melt the base metal.

1. Torch Not Centered: This concentrates

the heat of the arc on only one of the pieces.

The puddle just lies on top of the work instead of fusing to it. Exaggerated oscillation
produces the same problem.

If both pieces to be joined are not melted

by the arc, the molten puddle will have a

3. Excessive Travel Speed: If the travel

tendency to lie on top of the second piece

speed is too high, it is possible to not spend

without fusing to it.

enough time to allow fusion with the base

2. Poor Torch Oscillation: If a welder

makes little circles with the torch, it is

easier to make a pretty bead, but there is

metal. There is a minimum amount of heat

required for every welding joint, because the
base metal can cool the puddle very rapidly
after it is deposited.

a good chance for lack of fusion. As the

4. Vertical Down Welding is notorious for

lack of fusion at the toes (where the weld

metal intersects the base metal) and very
little penetration. This lack of fusion can also
occur in flat and horizontal positions where
it is called overlap, or cold lap. Lack of fusion in vertical down welds is usually caused
by getting on top of the weld puddle. This
prevents the arc from adequately heating the
base material to produce proper fusion. An
old slang term for this defect is fingernailing
because if it is bad enough, you can insert a
fingernail beneath it.

Figure 44
Lack of Fusion

47

5. Travel Speed Too Slow: It is similar to

7. Incorrect Welding Parameters is the

improper torch oscillation. The heat from the

last major cause for lack of fusion. This

arc is used to further heat the already molten

category would include all the variables

puddle and raise its temperature instead

covered plus incorrect ESO (electrical stick

melting the base metal.

out), torch angle wrong, incorrect shielding

6. Current Too Low can also cause lack

gas, etc.

of fusion. This is usually only a problem on

thicker sections, where the base metal pulls
heat out of the puddle very rapidly.

B.
Porosity

Porosity (figure 45) can be a significant

Another form of contamination can come

problem, and not easily solved. The biggest

from the base metals. It is against better

causes are probably contamination of the

welding practice to weld over paint, marker

shielding gas, followed by filler metal and

lines, water, rust, oil, and heavy mill scale.

base metal contamination. A leak anywhere

Paint and markers are typically made of hy-

in the distribution system from a leaking fit-

drocarbons. The heat from the arc breaks

ting in the ceiling to a loose hose fitting at

down these compounds to release hydrogen,

the feeder will allow gas to leak out and air

carbon and other contaminants into the weld

to diffuse into the shielding gas. Molten

pool. Mill scale and rust can also contain a

weld metal holds a lot more nitrogen, oxy-

lot of moisture that is dissociated in the arc

gen, and hydrogen (from moisture in the air)

to form oxygen and hydrogen. Although

than solid metal. As the weld puddle freezes,

this is not such a big problem on low carbon

the gases come out of solution and form

steels, on higher alloys it can cause cracking

porosity.

that sometimes doesnt show up until the

component is in service.
Additions of higher thermal conductivity
gases (He & CO2) and higher alloy content
wires (ER70S-6, S-7) can help to reduce
porosity. Helium and carbon dioxide increase the energy being put into the puddle,
so the cooling rate is slower. The slower
cooling rate gives more time for the gases to
leave the puddle instead of causing porosity.
The increased deoxidizer levels in some
wires will also reduce porosity if the cause
of the porosity is associated with oxygen
levels (not high nitrogen levels due to gas
leaks). These deoxidizers will tie up the
oxygen and float it to the surface as liquid

Figure 45
Porosity in a Weld

48

oxides that freeze to form slag islands.

Porosity can also be caused by excessive

Sometimes just moving a fan a few degrees

tip to work distance. This can create turbu-

is enough to solve a problem. Barriers must

lence in the shielding gas column, aspirating

sometimes be put between the welding area

oxygen and nitrogen from the atmosphere

and an open door. This can usually be some-

that then react with the high temperature

thing as simple as a welding screen.

weld metal.
As mentioned under gas losses, a dirty torch
This also occurs when the torch angle from

can cause porosity by blocking and disturb-

vertical is too great. When the torch angle

ing the laminar flow of the shielding gas.

is too severe, a venturi is set up between the

Spatter adherence to the nozzle is an obvious

gas nozzle and workpiece. This pulls in a

culprit, as is spatter on the gas diffuser.

great deal of atmospheric air contaminating

the shielding gas.

If the operating voltage is too high problems can occur because the arc will have a

Additional causes of porosity are shielding

tendency to wander, especially on fillets. As

gas flow too low and shielding gas flow too

the arc seeks the closest point, it can disturb

high. At low rates, the gas cannot exclude

the gas column enough to cause turbulence

the atmosphere. At high flows, turbulence

and entrain air.

in the gas column causes mixing with the

atmosphere.

A further cause of porosity is excessive

torch oscillation. Too much torch oscilla-

Drafts, winds and fans can also cause poro-

tion can cause porosity because it is possible

sity. All of these sources disturb the shield-

to either induce turbulence or even run out

ing gas column and must be controlled.

from under the gas column.

C.
Burn-Through

Burn-through is usually caused by excessive current for the application, as shown

in figure 46. Sometimes an easy solution is
to increase your ESO (electrical stick out.)
This reduces heat input into the part.
Another problem can be that the travel
speed is too low. Traveling too slowly increases heat input greatly. When possible,
increase the travel speeds and make multiple passes, if necessary.

Figure 46
Burn-Through

49

Incorrect wire diameter is also a cause of

Excessive gaps can also be a cause of burn-

burn-through. Reducing the wire diameter

through, as every manual welder knows.

allows you to weld at lower currents but at

This is easy to forget on robots and mecha-

the same deposition rate with reduced heat

nized welds, and sometimes, mysterious

input. One more variable that can cause ad-

burn-through occurs because the robot cant

ditional heat is using the incorrect shielding

compensate by changing conditions when

gas. The shielding gas can be tailored to the

faced with gaps like a welder can. If the

application to help control the heat input.

parts cant be made to fit better, gaps can be

pre-welded or stringered with a short-bead
prior to final welding.

D.
Undercut

Undercut (figure 47) has three main causes.

The first is excessive voltage, which can
cause the arc to wander. The melted area of
the basemetal is too large for the puddle to
flow into and fill the undercut area. High
travel speeds can lead to under-cut for the
same reason. The heat input is too low to
allow the weld metal to flow and the bead
is generally convex. On horizontal fillets,
undercut can be caused by incorrect torch
position. If the torch is positioned on the
vertical leg, there is no way for the weld metal to flow uphill against the force of gravity.
Repositioning the torch 1-2 wire diameters
out from the root on the horizontal plate will
eliminate the undercut. Undercut is a very
serious defect on high strength materials

50

Figure 47

because it greatly increases the strain in the

Undercut on the

material just below the notch and increases

Upper Leg of a Fillet Weld

its sensitivity to hydrogen cracking.


E.
Spatter

Spatter is usually caused by operating out-

In spray transfer, too low a voltage will also

side acceptable parameters for the desired

cause spatter. For the same .045" diameter

metal transfer mode. For example, operating

electrode and C-8, the minimum conditions

in the short arc mode with an .045" diameter

are about 235 amps and 27 arc volts. The

electrode and a C-8 gas, the hottest condition

key here is arc volts. The power supply may

would be about 223-235 amps and 20 to 21

say that it is putting out 32 volts, and there

volts. This is a sliding scale, so at 150 amps

is still spatter. Use a volt meter and measure

it might be 17 or 18 volts. As the voltage

the voltage from the positive connection

climbs into the 23-25 volt range the spatter

at the wire feeder to the work piece while

level increases rapidly because the transfer

welding. It is possible that there is 100 feet

mode is now globular transfer.

of work leap with two frayed, oxidized connections. The 32 volts at the machine may be
only 25 volts at the arc due to the resistance
in the circuit.

F.
Cracking

Cracking (figure 48) in low carbon steels

is not usually a problem because the
materials are so ductile. In high strength
materials, cracking can be caused by:
High Restraint A highly restrained part
will not allow the material to move to relieve
the stresses of the weld metal during cooling. This also leads to local stresses beyond
the yield strength of the material. A highly
restrained joint using high strength material
can require preheat and interpass temperatures be increased to slow the cooling rate.
Post-weld heat treating may also be
necessary.
Poor Bead Shape A convex bead will
greatly increase the stress at the toes of the
weld. The increased thickness of the center

Figure 48

of the bead forces the stress into the toes,

Centerline and

and can rapidly exceed the yield strength of

Underbead

the material. A concave bead may not have

Cracking

the required effective throat to withstand the

design forces, and crack through the throat.
The optimum fillet shape is a flat face with
the toes blending into the side walls very
smoothly.

51

S E C T I O N 10

10
Conclusion

Hopefully this course has given you a little

better understanding of the Gas Metal Arc
Welding Process.
With the knowledge you have, and this
booklet to use as a reference, it is hoped
that youre now a more informed welder
and your job has been made easier. Please
call your local Praxair engineering representative with any questions that you have
in the future.

52

Inside
back cover
is blank

 Copyright 1996-2005 Praxair Technology, Inc.

All rights reserved
Praxair, Inc.
39 Old Ridgebury Road
Danbury, CT 06810-5113
USA
www.praxair.com
info@praxair.com
Telephone:
1-800-PRAXAIR
(1-800-772-9247)
(716) 879-4077
Fax:
1-800-772-9985
(716) 879-2040
Praxair the Flowing Airstream design, Helistar,
Stargold and Stargon are trademarks or registered
trademarks of Praxair Technology, Inc. in the
United States and/or other countries.
The information contained herein is offered
for use by technically qualified personnel at
their discretion and risk without warranty of
any kind.

P-8107-C

Revised 11/2005