The serviceability of a product or structure utilizing the type of information presented herein is, and must be, the sole responsi-bility of the builder/user. Many variables beyond the control of The James F. Lincoln Arc Welding Foundation or The LincolnElectric Company affect the results obtained in applying this type of information. These variables include, but are not limited to,welding procedure, plate chemistry and temperature, weldment design, fabrication methods, and service requirements.

This guide makes extensive reference to the AWS D1.1 Structural Welding Code-Steel, but it is not intended to be a comprehen-sive review of all code requirements, nor is it intended to be a substitution for the D1.1 code. Users of this guide are encouragedto obtain a copy of the latest edition of the D1.1 code from the American Welding Society, 550 N.W. LeJeune Road, Miami,Florida 33126, (800) 443-9353.

Fabricators’ and Erectors’ Guide toWelded Steel Construction

By Omer W. Blodgett, P.E., Sc.D.R. Scott FunderburkDuane K. Miller, P.E., Sc.D.Marie Quintana, P.E.

This information has been provided byThe James F. Lincoln Arc Welding Foundation

to assist the general welding industry.

Copyright © 1999

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1

2 Welding Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12.1 SMAW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12.2 FCAW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32.3 SAW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .62.4 GMAW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .82.5 ESW/EGW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10

3 Welding Process Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .113.1 Joint Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .113.2 Process Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .123.3 Special Situations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12

4 Welding Cost Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14

5 Welding Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .155.1 Effects of Welding Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .155.2 Purpose of Welding Procedure Specifications (WPSs) . . . . . . . . . . .175.3 Prequalified Welding Procedure Specifications . . . . . . . . . . . . . . . .185.4 Guidelines for Preparing Prequalified WPSs . . . . . . . . . . . . . . . . . .205.5 Qualifying Welding Procedures By Test . . . . . . . . . . . . . . . . . . . . . .205.6 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .225.7 Approval of WPSs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22

6 Fabrication and Erection Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . .236.1 Fit-Up and Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .236.2 Backing and Weld Tabs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .236.3 Weld Access Holes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .246.4 Cutting and Gouging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .256.5 Joint and Weld Cleaning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .256.6 Preheat and Interpass Temperature . . . . . . . . . . . . . . . . . . . . . . . . . .256.7 Welding Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .266.8 Special Welding Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .296.9 Weld Metal Mechanical Properties . . . . . . . . . . . . . . . . . . . . . . . . . .296.10 Intermixing of Weld Deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33

Table of Contents

Fabricators’ and Erectors’ Guide toWelded Steel Construction

7 Welding Techniques and Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . .357.1 SMAW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .357.2 FCAW-ss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .367.3 FCAW-g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .377.4 SAW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .387.5 GMAW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .397.6 ESW/EGW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40

8 Welder Qualification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40

9 Weld Cracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .409.1 Centerline Cracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .419.2 Heat Affected Zone Cracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .429.3 Transverse Cracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .44

10 Weld Quality and Inspection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4410.1 Weld Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4410.2 Weld Quality and Process-Specific Influences . . . . . . . . . . . . . . . . .4610.3 Weld Inspection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .46

11 Arc Welding Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49

1

Fabricators’ and Erectors’ Guide toWelded Steel Construction

1 Introduction/Background

This Fabricators’ and Erectors’ Guide to Welded SteelConstruction has been produced by The Lincoln ElectricCompany in order to help promote high quality and cost-effective welding. This guide is not to be used as a sub-stitute for the AWS D1.1 Structural Welding Code, or anyother applicable welding code or specification, and theuser bears the responsibility for knowing applicablecodes and job requirements. Rather, this documentincorporates references to the D1.1-96 code, and addsexplanation, clarification, and guidelines to facilitatecompliance with the code. At the time of writing, thisguide reflects the current industry views with respect tosteel fabrication, with specific emphasis on the new pro-visions that have been recently imposed for fabrication ofstructures designed to resist seismic loads. These provi-sions are largely drawn from the Federal EmergencyManagement Administration (FEMA) Document No.267, produced by the SAC Consortium, whose membersinclude the Structural Engineers Association ofCalifornia, Applied Technology Council, and CaliforniaUniversities for Research and Earthquake Engineering.Another cited document is the AWS D1 StructuralWelding Committee’s Position Statement on theNorthridge earthquake. Research is still underway, andadditional provisions may be found that will furtherincrease the safety of welded steel structures. The userof this document must be aware of changes that mayoccur to codes published after this guide, specific jobrequirements, and various interim recommendations thatmay affect the recommendations contained herein.

The January 1994 Northridge earthquake revealed anumber of examples of lack of conformance to D1.1code mandated provisions. Lack of conformance to codeprovisions, and the poor workmanship revealed in manysituations, highlight the need for education. This docu-ment is one attempt to assist in that area.

The information contained herein is believed to be cur-rent and accurate. It is based upon the current technolo-gy, codes, specifications and principles of weldingengineering. Any recommendations will be subject tochange pending the results of ongoing research. Asalways, it is the responsibility of the Engineer of Record,

and not The Lincoln Electric Company, to specify therequirements for a particular project. The prerogative tospecify alternate requirements is always within theauthority of the Engineer of Record and, when morerestrictive requirements are specified in contract docu-ments, compliance with such requirements would super-sede the preceding recommendations. Acceptance ofcriteria by the Engineer of Record that are less rigorousthan the preceding does not change the recommendationsof The Lincoln Electric Company.

2 Welding Processes

A variety of welding processes can be used to fabricateand erect buildings. However, it is important that all par-ties involved understand these processes in order toensure high quality and economical fabrication. A briefdescription of the major processes is provided below.

2.1 SMAW

Shielded metal arc welding (SMAW), commonly knownas stick electrode welding or manual welding, is the old-est of the arc welding processes. It is characterized byversatility, simplicity and flexibility. The SMAWprocess commonly is used for tack welding, fabricationof miscellaneous components, and repair welding. Thereis a practical limit to the amount of current that may beused. The covered electrodes are typically 9 to 18 inch-es long, and if the current is raised too high, electricalresistance heating within the unused length of electrodewill become so great that the coating ingredients mayoverheat and “break down,” potentially resulting in weldquality degradation. SMAW also is used in the field forerection, maintenance and repairs. SMAW has earned areputation for depositing high quality welds dependably.It is, however, slower and more costly than other meth-ods of welding, and is more dependent on operator skillfor high quality welds.

The American Welding Society (AWS) publishes a vari-ety of filler metal specifications under the jurisdiction ofthe A5 Committee; A5.1 addresses the particular require-ments for mild steel covered electrodes used with theshielded metal arc welding process. The specificationA5.5 similarly covers the low alloy electrodes.

2

For welding on steels with minimum specified yieldstrengths exceeding 50 ksi, all electrodes should be of thelow hydrogen type with specific coatings that aredesigned to be extremely low in moisture. Water, orH2O, will break down into its components hydrogen andoxygen under the intensity of the arc. This hydrogen canthen enter into the weld deposit and may lead to unac-ceptable weld heat affected zone cracking under certainconditions. Low hydrogen electrodes have coatingscomprised of materials that are very low in hydrogen.

The low hydrogen electrodes that fit into the A5.1 classi-fication include E7015, E7016, E7018, and E7028. TheE7015 electrodes operate on DC only. E7016 electrodesoperate on either AC or DC. The E7018 electrodes oper-ate on AC or DC and include approximately 25% ironpowder in their coatings; this increases the rate at whichmetal may be deposited. An E7028 electrode containsapproximately 50% iron powder in the coating, enablingit to deposit metal at even higher rates. However, thiselectrode is suitable for flat and horizontal welding only.

Under the low alloy specification, A5.5, a similar format isused to identify the various electrodes. The most signifi-cant difference, however, is the inclusion of a suffix letterand number indicating the alloy content. An examplewould be an “E8018-C3” electrode, with the suffix “-C3” indicating the electrode nominally contains 1% nick-el. A “-C1” electrode nominally contains 2.5% nickel.

In AWS A5.1, the electrodes listed include both lowhydrogen and non-low hydrogen electrodes. In AWSD1.1-96, Table 3.1, Group I steels may be welded withnon-low hydrogen electrodes. This would include A36steel. For Group II steels and higher, low hydrogen elec-trodes are required. These steels would include A572grade 50. For most structural steel fabrication today, lowhydrogen electrodes are prescribed to offer additionalassurance against hydrogen induced cracking. When lowhydrogen electrodes are used, the required levels of pre-heat (as identified in Table 3.2 of D1.1-96) are actuallylower, offering additional economic advantages to thecontractor.

All the low hydrogen electrodes listed in AWS A5.1 haveminimum specified notch toughnesses of at least 20 ft.lb. at 0°F. There are electrode classifications that have no

notch toughness requirements (such as E6012, E6013,E6014, E7024) but these are not low hydrogen elec-trodes. Although there is no direct correlation betweenthe low hydrogen nature of various electrodes and notchtoughness requirements, in the case of SMAW electrodesin A5.1, the low hydrogen electrodes all have minimumnotch toughness requirements.

Care and storage of low hydrogen electrodes— Lowhydrogen electrodes must be dry if they are to performproperly. Manufacturers in the United States typicallysupply low hydrogen electrodes in hermetically sealedcans. When electrodes are so supplied, they may beused without any preconditioning; that is, they need notbe heated before use. Electrodes in unopened, hermeti-cally sealed containers should remain dry for extendedperiods of time under good storage conditions. Onceelectrodes are removed from the hermetically sealedcontainer, they should be placed in a holding oven tominimize or preclude the pick-up of moisture from theatmosphere. These holding ovens generally are electri-cally heated devices that can accommodate several hun-dred pounds of electrodes. They hold the electrodes at atemperature of approximately 250-300°F. Electrodes tobe used in fabrication are taken from these ovens.Fabricators and erectors should establish a practice oflimiting the amount of electrodes discharged at anygiven time. Supplying welders with electrodes twice ashift — at the start of the shift and at lunch, for example— minimizes the risk of moisture pickup. However, theoptional designator “R” indicates a low hydrogen elec-trode which has been tested to determine the moisturecontent of the covering after exposure to a moist envi-ronment for 9 hours and has met the maximum level per-mitted in ANSI/AWS A5.1-91. Higher strengthelectrodes will require even more rigorous control.Electrodes must be returned to the heated cabinet forovernight storage.

Once the electrode is exposed to the atmosphere, itbegins to pick up moisture. The D1.1 code limits thetotal exposure time as a function of the electrode type(D1.1-96, paragraph 5.3.2.2, Table 5.1). Electrodes usedto join high strength steels (which are particularly sus-ceptible to hydrogen cracking) must be carefully caredfor, and their exposure to the atmosphere strictly limited.

3

Some electrodes are supplied in cardboard containers.This is not commonly done for structural fabrication,although the practice can be acceptable if specific andappropriate guidelines are followed. The electrodes mustbe preconditioned before welding. Typically, this meansbaking them at temperatures in the 700 to 900°F range toreduce moisture. In all cases, the electrode manufactur-er’s guidelines should be followed to ensure a bakingprocedure that effectively reduces moisture without dam-age to the covering. Electrodes removed from damagedhermetically sealed cans should be similarly baked athigh temperature. The manufacturer’s guidelines shouldbe consulted and followed to ensure that the electrodesare properly conditioned. Lincoln Electric’s recommen-dations are outlined in Literature # C2.300.

Redrying low hydrogen electrodes— When containersare punctured or opened so that the electrode is exposedto the air, or when containers are stored under unusuallywet conditions, low hydrogen electrodes pick up mois-ture. The moisture, depending upon the amountabsorbed, impairs weld quality in the following ways:

1. If the base metal has high hardenability, even a smallamount of moisture can contribute to underbead cracking.

2. A small amount of moisture may cause internal poros-ity. Detection of this porosity requires X-ray inspec-tion or destructive testing.

3. A high amount of moisture causes visible externalporosity in addition to internal porosity. Proper redry-ing restores the ability to deposit quality welds. Theproper redrying temperature depends upon the type ofelectrode and its condition (D1.1-96, paragraph5.3.2.4, Table 5.1).

2.2 FCAW

Flux cored arc welding (FCAW) uses an arc between acontinuous filler metal electrode and the weld pool. Theelectrode is always tubular. Inside the metal sheath is acombination of materials that may include metallic pow-der and flux. FCAW may be applied automatically orsemiautomatically.

The flux cored arc welding process has become the mostpopular semiautomatic process for structural steel fabri-cation and erection. Production welds that are short, thatchange direction, that are difficult to access, that must bedone out-of-position (e.g., vertical or overhead), or thatare part of a short production run, generally will be madewith semiautomatic FCAW.

The flux cored arc welding process offers two distinctadvantages over shielded metal arc welding. First, theelectrode is continuous. This eliminates the built-instarts and stops that are inevitable with shielded metal arcwelding. Not only does this have an economic advantagebecause the operating factor is raised, but the number ofarc starts and stops, a potential source of weld disconti-nuities, is reduced.

Another major advantage is that increased amperagescan be used with flux cored arc welding, with a corre-sponding increase in deposition rate and productivity.With the continuous flux cored electrodes, the tubularelectrode is passed through a contact tip, where electricalenergy is transferred to the electrode. The short distancefrom the contact tip to the end of the electrode, known aselectrode extension or “stickout,” limits the build up ofheat due to electrical resistance. This electrode extensiondistance is typically 3/4 in. to 1 in. for flux cored elec-trodes, although it may be as high as two or three inches.

Within the category of flux cored arc welding, there aretwo specific subsets: self shielded flux core (FCAW-ss)and gas shielded flux core (FCAW-g). Self shielded fluxcored electrodes require no external shielding gas. Theentire shielding system results from the flux ingredientscontained within the core of the tubular electrode. Thegas shielded versions of flux cored electrodes utilize anexternally supplied shielding gas. In many cases, CO2 isused, although other gas mixtures may be used, e.g.,argon/CO2 mixtures. Both types of flux cored arc weld-ing are capable of delivering weld deposits that meet thequality and mechanical property requirements for moststructure applications. In general, the fabricator will uti-lize the process that offers the greatest advantages for theparticular environment. Self shielded flux cored elec-trodes are better for field welding situations. Since no

4

externally supplied shielding gas is required, the processmay be used in high winds without adversely affectingthe quality of the deposit. With any of the gas shieldedprocesses, wind shields must be erected to preclude inter-ference with the gas shield in windy weather. Many fab-ricators have found self shielded flux core offersadvantages for shop welding as well, since it permits theuse of better ventilation.

Individual gas shielded flux cored electrodes tend to bemore versatile than self shielded flux cored electrodes,and in general, provide better arc action. Operator appealis usually higher. While the gas shield must be protectedfrom winds and drafts, this is not particularly difficult inshop fabrication situations. Weld appearance and qualityare very good. Higher strength gas shielded FCAW elec-trodes are available, while current technology limits selfshielded FCAW deposits to 90 ksi tensile strength or less.

Filler metals for flux cored arc welding are specified inAWS A5.20 and A5.29. A5.20 covers mild steel elec-trodes, while A5.29 addresses low alloy materials.Positive polarity is always used for FCAW-g, althoughthe self shielded electrodes may be used on either polar-ity, depending on their classification. Under A5.29 foralloy electrodes, a suffix letter followed by a numberappears at the end. Common designations include “Ni1”indicating a nominal nickel content in the depositedmetal of 1%. The letter “M” could appear at the end ofthe electrode classification. If this is done, the electrodehas been designed for operation with mixed shieldinggas, that is an argon-CO2 blend that consists of 75 - 80%argon. Other suffix designators may be used that indicateincreased notch toughness capabilities, and/or diffusiblehydrogen limits.

Table 2.1 describes various FCAW electrodes listed inAWS A5.20 and A5.29. Some of the electrodes haveminimum specified notch toughness values although oth-ers do not. Some are gas shielded, while others are selfshielded. Some are restricted to single pass applications,and others have restrictions on the thickness for theirapplication. The electrical polarity used for the variouselectrodes is also shown. For critical applications inbuildings that are designed to resist seismic loading asdetermined by the Engineer of Record, only electrodesthat are listed in Table 2.1 as having the required mini-

mum specified notch toughness levels should be used.The corresponding Lincoln Electric products are alsoshown.

Shielding gases for FCAW-g — Most of the gas shield-ed flux cored electrodes utilize carbon dioxide for theshielding media. However, electrodes may also beshielded with an argon-CO2 mixture. All gases should beof welding grade with a dew point of -40°F or less. Thecarbon dioxide content is typically 10% to 25%, with thebalance composed of argon. This is done to enhancewelding characteristics. In order to utilize the argonbased shielding gases, arc voltages are typically reducedby two volts from the level used with carbon dioxideshielding.

The selection of shielding gas may affect mechanicalproperties, including yield and tensile strength, elonga-tion, and notch toughness. This is largely due to the dif-ference in alloy recovery—that is, the amount of alloytransferred from the filler material to the weld deposit.Carbon dioxide is a reactive gas that may cause some ofthe alloys contained in the electrode (Mn, Si and others)to be oxidized, so that less alloy ends up in the deposit.When a portion of this active carbon dioxide is replacedwith an inert gas such as argon, recovery typicallyincreases, resulting in more alloy in the weld deposit.Generally, this will result in higher yield and tensilestrengths, accompanied by a reduction in elongation.The notch toughness of the weld deposit may go up ordown, depending on the particular alloy whose recoveryis increased.

Storing FCAW electrodes — In general, FCAW elec-trodes will produce weld deposits which achievehydrogen levels below 16 ml per 100 grams of deposit-ed metal. These electrodes, like other products whichproduce deposits low in hydrogen, must be protectedfrom exposure to the atmosphere in order to maintainhydrogen levels as low as possible, prevent rusting ofthe product and prevent porosity during welding. Therecommended storage conditions are such that theymaintain the condition of 90 grains of moisture perpound of dry air. Accordingly, the following storageconditions are recommended for FCAW electrodes intheir original, unopened boxes and plastic bags.

5

Table 2.1 FCAW Electrode Classification

6

For best results, electrodes should be consumed as soonas practicable. However, they may be stored up to threeyears from the date of manufacture. The Lincoln distrib-utor or sales representative should be consulted if there isa question as to when the electrodes were made.

Once the electrode packaging is opened, Innershield andOutershield electrodes can be subject to contaminationfrom atmospheric moisture. Care has been taken in thedesign of these products to select core ingredients thatare essentially resistant to moisture pick-up; however,condensation of the moisture from the atmosphere ontothe surface of the electrode can be sufficient to degradethe product.

The following minimum precautions should be taken tosafeguard product after opening the original package.Electrode should be used within approximately 1 weekafter opening the original package. Opened electrodeshould not be exposed to damp, moist conditions orextremes in temperature and/or humidity where surfacecondensation can occur. Electrodes mounted on wirefeeders should be protected against condensation. It isrecommended that electrode removed from its originalpackaging be placed in poly bags (4 mil minimum thick-ness) when not in use.

In the case of FCAW-s, excessively damp electrodes canresult in higher levels of spatter, poorer slag cover andporosity. FCAW-g electrodes will display high moisturelevels in the form of gas tracks, higher spatter and poros-ity. Any rusty electrode should be discarded.

Products used for applications requiring morerestrictive hydrogen control — The AWS specificationfor flux cored electrodes, ANSI/AWS A5.20, states that“Flux cored arc welding is generally considered to be alow hydrogen welding process.” To further clarify theissue, this specification makes available optional supple-mental designators for maximum diffusible hydrogenlevels of 4, 8 and 16 ml per 100 grams of deposited weldmetal.

Some Innershield and Outershield products have beendesigned and manufactured to produce weld depositsmeeting more stringent diffusible hydrogen require-ments. These electrodes, usually distinguished by an “H”added to the product name, will remain relatively dryunder recommended storage conditions in their original,unopened package or container.

For critical applications in which the weld metal hydro-gen must be controlled (usually H8 or lower), or whereshipping and storage conditions are not controlled orknown, only hermetically sealed packaging is recom-mended. Innershield and Outershield electrodes areavailable in hermetically sealed packages on a specialorder basis.

Once the package has been opened, the electrode shouldnot be exposed to conditions exceeding 80% relativehumidity for a period greater than 16 hours, or any lesshumid condition for more than 24 hours. Conditions thatexceed 80% RH will decrease the maximum 16 hourexposure period.

After exposure, hydrogen levels can be reduced by con-ditioning the electrode. Electrodes may be conditioned ata temperature of 230ºF ± 25ºF for a period of 6 to 12hours, cooled and then stored in sealed poly bags (4 milminimum thickness) or equivalent. Electrodes on plasticspools should not be heated at temperatures in excess of150ºF. Rusty electrodes should be discarded.

2.3 SAW

Submerged arc welding (SAW) differs from other arcwelding processes in that a layer of fusible granularmaterial called flux is used for shielding the arc and themolten metal. The arc is struck between the workpieceand a bare wire electrode, the tip of which is submergedin the flux. Since the arc is completely covered by theflux, it is not visible and the weld is made without theflash, spatter, and sparks that characterize the open-arcprocesses. The nature of the flux is such that very littlesmoke or visible fumes are released to the air.

Typically, the process is fully mechanized, although semi-automatic operation is often utilized. The electrode is fedmechanically to the welding gun, head, or heads. In semi-automatic welding, the welder moves the gun, usuallyequipped with a flux-feeding device, along the joint.

Maximum %Ambient Temperature Relative Humidity

Degrees F Degrees C60 - 70 16 - 21 8070 - 80 21 - 27 6080 - 90 27 - 32 4590 - 100 32 - 38 30

7

High currents can be used in submerged arc welding andextremely high heat input levels can be developed.Because the current is applied to the electrode a short dis-tance above its arc, relatively high amperages can beused on small diameter electrodes, resulting in extreme-ly high current densities. This allows for high depositionrates and deep penetration.

Welds made under the protective layer of flux are excel-lent in appearance and spatter free. Since the processdevelops a minimum amount of smoke, the surroundingplate surfaces remain clear of smoke deposits. The highquality of submerged arc welds, the high depositionrates, the deep penetration characteristics, and the easyadaptability of the process to full mechanization make itpopular for the manufacture of plate girders and fabricat-ed columns.

One of the greatest benefits of the SAW process is free-dom from the open arc. This allows multiple arcs to beoperated in a tight, confined area without the need forextensive shields to guard the operators from arc flash.Yet this advantage also proves to be one of the chief draw-backs of the process; it does not allow the operator toobserve the weld puddle. When SAW is applied semiau-tomatically, the operator must learn to propel the guncarefully in a fashion that will ensure uniform bead con-tour. The experienced operator relies on the uniform for-mation of a slag blanket to indicate the nature of thedeposit. For single pass welds, this is mastered fairlyreadily; however, for multiple pass welding, the degree ofskill required is significant. Therefore, most submergedarc applications are mechanized. The nature of the jointmust then lend itself to automation if the process is toprove viable. Long, uninterrupted straight seams are idealapplications for submerged arc. Short, intermittent weldsare better made with one of the open arc processes.

Two electrodes may be fed through a single electricalcontact tip, resulting in higher deposition rates.Generally known as parallel electrode welding, theLincoln trade name for this is Tiny Twin® or Twin Arc®.The equipment is essentially the same as that used forsingle electrode welding, and parallel electrode weldingprocedures may be prequalified under AWS D1.1-96.

Multiple electrode SAW refers to a variation of sub-merged arc which utilizes at least two separate powersupplies, two separate wire drives, and feeds two elec-trodes independently. Some applications such as themanufacture of line pipe may use up to five independent

electrodes in a multiple electrode configuration. ACwelding currently is typically used for multi-electrodewelding. If DC current is used, it usually is limited to thelead electrode to minimize the potentially negative inter-action of magnetic fields between the two electrodes.

Submerged arc filler materials are classified underAWS A5.17 for mild steel and AWS A5.23 for low alloyfiller materials. Both fluxes and electrodes are coveredunder these specifications. Since submerged arc is a two-component process, that is, flux and electrode, the classi-fication system is slightly different than for other fillermaterials.

Electrodesare classified based on the composition of theelectrode. Under A5.17, the electrode will carry a classi-fication that consists of two letters, one or two numericaldigits and, in some cases, a final letter. The first letter isan E, which stands for electrode. The second letter willbe L, M, or H, referring to a low, medium, or high levelof manganese in the electrode. The next one or two dig-its refer to the nominal carbon content in hundredths of apercent. A “12” in this location, for example, wouldindicate a nominal carbon content of 0.12%. It should beemphasized that this is the nominal value; it is possible tohave higher and lower carbon contents in a specific elec-trode. In some cases, the electrode will be made of killedsteel. When this is the case, silicon normally is addedand the electrode will have a “K” at the end of the clas-sification (e.g., EM13K).

Electrodes classified under A5.23, the low alloy variety,have a more complex nomenclature, because of the vari-ety of alloys that may be involved. The most importantalloys for structural welding are the “Ni,” or nickelalloys, and “W,” or weathering alloys (e.g., ENi1K).

Fluxesare always classified in conjunction with an elec-trode. The flux-electrode combination must meet specif-ic mechanical property requirements. After a flux isselected and a classification test plate welded, a flux-electrode classification may be established. Specimensare extracted from the weld deposit to obtain the mechan-ical properties of the flux-electrode combination. Theclassification will follow the format of an “F” followedby a single or two digit number, an “A” or “P,” a singledigit and a hyphen which separates the electrode classifi-cation. Thus, a typical flux-electrode may be classifiedas an F7A2-EM12K. The “F” stands for flux, and the“7” indicates all of the following: a 70-95 ksi tensilestrength deposit, a 58 ksi minimum yield strength, and a

8

minimum of 22% elongation. The “A” indicates thedeposit is tested in the as-welded condition. The “2”indicates 20 ft. lbf. at -20°F, and the balance of the clas-sification identifies the electrode used.

Because of the popularity of the submerged arc processfor pressure vessel fabrication where assemblies are rou-tinely stress relieved, submerged arc products may beclassified in the post weld heat treated, or stress relieved,condition. When this is done, a “P” replaces the “A.” Forstructural work, which is seldom stress relieved, the “A”classification is more common.

For products classified under A5.23, a format similar tothat of A5.17 is used, with this major exception: at theend of the flux-electrode classification, a weld depositcomposition is specified. For example, an F7A2-ENi1-Ni1 would indicate that the electrode, an ENi1, deliversan F7A2 deposit when used with a specific flux. In addi-tion, the deposit has a composition that meets therequirements of an Ni1. In this case, a nickel bearingelectrode deposits a weld that contains nickel. Theexample is straightforward. However, it is also possibleto use alloy fluxes which, with mild steel electrodes, arecapable of delivering alloy weld metal. In this case, atypical classification may be an F7A2-EL12-Nil. In thisexample, an EL12 electrode (a non-alloy electrode thatcontains a low level of manganese) is used with an alloyflux. The result is an alloyed deposit. This is common-ly done when nickel bearing deposits are desired onweathering steel that will not be painted.

Only part of the flux deposited from a hopper or a gun isfused in welding. The unfused, granular flux may berecovered for future use and is known as reclaimed flux.The unmelted flux does not undergo chemical changesand may therefore be capable of delivering quality weldswhen used the next time. However, this flux can be con-taminated in the act of recovery. If it comes in contactwith oil, moisture, dirt, scale or other contaminants, theproperties of the weld deposit made with reclaimed fluxmay be adversely affected. Care should be exercised toensure that flux is not thus contaminated. Another prob-lem with reclaimed flux is the potential for the break-down of particles and the modification of the particle sizedistribution. This can affect the quality and/or proper-ties. The method of flux recovery can range from sweep-ing up the flux with broom and pans, to vacuum recoverysystems; the method chosen should take into account theneed to avoid contamination.

Larger pieces of fused slag should be separated from therecovered flux in order to avoid flux feeding problems.The automated systems typically have screening to han-dle this. The fused slag may be chemically different thanthe unfused flux. For less critical applications, this slagmay be crushed and thoroughly intermixed with newflux. This is sometimes called “recycled flux,” but sincereclaimed flux is sometimes referred to by the same term,a better description for this product is “crushed slag.”Performance and mechanical properties of crushed slagmay differ from those of virgin flux. AWS D1.1-96requires that crushed slag must be classified in much thesame way as new flux. (See AWS D1.1-96, paragraph5.3.3.4)

Flux must be stored so that it remains dry. The manu-facturer’s guidelines regarding storage and usage of theflux must be followed. In use, granules of flux must notcome in direct contact with water since weld crackingcan result. Fluxes can be contaminated with moisturefrom the atmosphere, so exposure should be limited.When not in use, flux hoppers should be covered or oth-erwise protected from the atmosphere. Lincoln Electric’srecommendations for storage and handling of flux areoutlined in Literature # C5.660.

2.4 GMAW

Gas metal arc welding (GMAW) utilizes equipmentmuch like that used in flux cored arc welding. Indeed,the two processes are very similar. The major differencesare: gas metal arc uses a solid or metal cored electrode,and leaves no appreciable amount of residual slag. Gasmetal arc has not been a popular method of welding inthe typical structural steel fabrication shop because of itssensitivity to mill scale, rust, limited puddle control, andsensitivity to shielding loss. Newer GMAW metal coredelectrodes, however, are beginning to be used in the shopfabrication of structural elements with good success.

A variety of shielding gases or gas mixtures may be usedfor GMAW. Carbon dioxide (CO2) is the lowest cost gas,and while acceptable for welding carbon steel, the gas isnot inert but active at elevated temperatures. This hasgiven rise to the term MAG (metal active gas) for theprocess when (CO2) is used, and MIG (metal inert gas)when predominantly argon-based mixtures are used.

While shielding gas is used to displace atmospheric oxy-gen, it is possible to add smaller quantities of oxygen into

9

mixtures of argon — generally at levels of 2 - 8%. Thishelps stabilize the arc and decreases puddle surface ten-sion, resulting in improved wetting. Tri- and quad-mixes of argon, oxygen, carbon dioxide and helium arepossible, offering advantages that positively affect arcaction, deposition appearance and fume generation rates.

Short arc transfer is ideal for welding on thin gaugematerials. It is generally not suitable for structural steelfabrication purposes. In this mode of transfer, the smalldiameter electrode, typically 0.035 in. or 0.045 in., is fedat a moderate wire feed speed at relatively low voltages.The electrode will touch the workpiece, resulting in ashort in the electrical circuit. The arc will actually go outat this point, and very high currents will flow through theelectrode, causing it to heat and melt. Just as excessivecurrent flowing through a fuse causes it to blow, so theshorted electrode will separate from the work, initiatinga momentary arc. A small amount of metal will be trans-ferred to the work at this time.

The cycle will repeat itself again once the electrodeshorts to the work. This occurs somewhere between 60and 200 times per second, creating a characteristic buzzto the arc. This mode of transfer is ideal for sheet metal,but results in significant fusion problems if applied toheavy materials. A phenomenon known as cold lap orcold casting may result where the metal does not fuse tothe base material. This is unacceptable since the weldedconnections will have virtually no strength. Great cau-tion must be exercised in the application of the short arcmode to heavy plates. The use of short arc on heavyplates is not totally prohibited however, since it is theonly mode of transfer that can be used out-of-positionwith gas metal arc welding, unless specialized equipmentis used. Weld joint details must be carefully designedwhen short arc transfer is used. Welders must pass specific qualification tests before using this mode oftransfer. The mode of transfer is often abbreviated asGMAW-s, and is not prequalified by the D1.1 code.

Globular transfer is a mode of gas metal arc weldingthat results when high concentrations of carbon dioxideare used, resulting in an arc that is rough with larger globsof metal ejected from the end of the electrode. This modeof transfer, while resulting in deep penetration, generatesrelatively high levels of spatter. Weld appearance can bepoor and it is restricted to the flat and horizontal position.Globular transfer may be preferred over spray transferbecause of the low cost of CO2 shielding gas and thelower level of heat experienced by the operator.

Spray arc transfer is characterized by high wire feedspeeds at relatively high voltages. A fine spray of moltendrops, all smaller in diameter than the electrode diame-ter, is ejected from the electrode toward the work. Unlikeshort arc transfer, the arc in spray transfer is continuous-ly maintained. High quality welds with particularly goodappearance are the result. The shielding used for sprayarc transfer is composed of at least 80% argon, with thebalance made up of either carbon dioxide or oxygen.Typical mixtures would include 90-10 argon-CO2, and95-5 argon-oxygen. Other proprietary mixtures areavailable from gas suppliers. Relatively high arc volt-ages are used with the spray mode of transfer. However,due to the intensity of the arc, spray arc is restricted toapplications in the flat and horizontal position, becauseof the puddle fluidity, and lack of a slag to hold themolten metal in place.

Pulsed arc transfer utilizes a background current that iscontinuously applied to the electrode. A pulsing peakcurrent is optimally applied as a function of the wire feedspeed. With this mode of transfer, the power supplydelivers a pulse of current which, ideally, ejects a singledroplet of metal from the electrode. The power supplyreturns to a lower background current which maintainsthe arc. This occurs between 100 and 400 times per sec-ond. One advantage of pulsed arc transfer is that it canbe used out-of-position. For flat and horizontal work, itmay not be as fast as spray transfer. However, used out-of- position, it is free of the problems associated with gasmetal arc short circuiting mode. Weld appearance isgood and quality can be excellent. The disadvantage ofpulsed arc transfer is that the equipment is slightly morecomplex and more costly. The joints are still required tobe relatively clean, and out-of-position welding is stillmore difficult than with processes that generate a slagthat can support the molten puddle.

Metal cored electrodesare a relatively new develop-ment in gas metal arc welding. This is similar to fluxcored arc welding in that the electrode is tubular, but thecore material does not contain slag forming ingredients.Rather, a variety of metallic powders is contained in thecore. The resulting weld is virtually slag-free, just aswith other forms of GMAW. The use of metal coredelectrodes offers many fabrication advantages. Theyhave increased ability to handle mill scale and other sur-face contaminants. Finally, metal cored electrodes per-mit the use of high amperages that may not be practicalwith solid electrodes, resulting in potentially higherdeposition rates. The properties obtained from metal

10

cored deposits can be excellent. Appearance is verygood. Because of the ability of the filler metal manufac-turer to control the composition of the core ingredients,mechanical properties obtained from metal coreddeposits may be more consistent than those obtainedwith solid electrodes. However, metal cored electrodesare in general more expensive.

2.5 ESW/EGW

Electroslag and electrogas welding (ESW/EGW) areclosely related processes that offer high deposition weld-ing in the vertical plane. Properly applied, these process-es offer significant savings over alternativeout-of-position methods and in many cases, a savingsover flat position welding. Although the two processeshave similar applications and mechanical set up, thereare fundamental differences in the arc characteristics.

Electroslag and electrogas are mechanically similar inthat both utilize copper dams or shoes that are applied toeither side of a square edged butt joint. An electrode ormultiple electrodes are fed into the joint. A starting sumpis typically applied for the beginning of the weld. As theelectrode is fed into the joint, a puddle is established thatprogresses vertically. The copper dams, which are com-monly water cooled, chill the weld metal and prevent itfrom escaping from the joint. The weld is completed inone pass.

These processes may be used for groove welds in butt,corner and tee joints. Typical applications involve heav-ier plate, usually 1” or thicker. Multiple electrodes maybe used in a single joint, allowing very heavy plate up toseveral inches thick to be joined in a single pass.Because of the sensitivity of the process to the variety ofvariables involved, specific operator training is required,and the D1.1-96 code requires welding procedures to bequalified by test.

In building construction, applications for ESW/EGWwith traditional connection designs are somewhat limit-ed. However, they can be highly efficient in the manu-facture of tree columns. In the shop, the beamflange-to-column welds can be made with the column inthe horizontal plane. With the proper equipment andtooling, all four flange welds can be made simultaneous-ly. In addition, continuity plate welds can be made withESW/EGW. Future connection designs may utilize con-figurations that are more conducive to these processes.

Another common application is for the welding of conti-nuity plates inside box columns. It is possible to weldthree sides of the continuity plate to the interior of thebox prior to closing the box with the fourth side.However, once this closure is made, access to the finalside of the continuity plate is restricted. It is possible touse these processes to make this final closure weld byoperating through a hole in the outside of the box col-umn. This approach is very popular in Asia, where boxcolumns are widely used.

In electroslag welding, a granular flux is metered into thejoint during the welding operation. At the beginning, anarc, similar to that of submerged arc welding, is estab-lished between the electrode and the sump.

After the initial flux is melted into a molten slag, thereaction changes. The slag, which is carefully designedto be electrically conductive, will conduct the weldingcurrent from the electrode through the slag into thepieces of steel to be joined. As high currents are passedthrough the slag, it becomes very hot. The electrode isfed through the hot slag and melts. Technically, elec-troslag welding is not an arc welding process, but a resis-tance welding process. Once the arc is extinguished andthe resistance melting process is stabilized, the weld con-tinues vertically to completion. A small amount of slagis consumed as it chills against the water cooled coppershoes. In some cases, steel dams instead of copper damsare used to retain the puddle. After completion of theweld, the steel dams stay in place, and become part of thefinal product. Slag must be replenished, and additionalflux is continuously added to compensate for the loss.

One aspect of electroslag welding that must be consid-ered is the very high heat input associated with theprocess. This causes a large heat affected zone (HAZ)that may have a lower notch toughness. Electrogas weld-ing is different from electroslag, inasmuch as no flux isused. Electrogas welding is a true arc welding processand is conceptually more like gas metal arc or flux coredarc welding. A solid or tubular electrode is fed into thejoint, which is flooded with an inert gas shield. The arcprogresses vertically while the puddle is retained by thewater cooled dams.

The Lincoln Vertishield® system uses a self shieldedflux cored electrode, and while no gas is required, it isclassified as EGW since it is an open arc process.

11

The HAZ performance is dependent not only on the heatinput, but also on the nature of the steel. While allprocesses develop a heat affected zone, the large size ofthe electroslag heat affected zone justifies additionalscrutiny. Advances in steel technology have resulted inimproved steels, featuring higher cleanliness and tough-ness, that better retain the HAZ properties in ESW/EGWwelds.

3 Welding Process Selection

Any of the common arc welding processes can be used toachieve the quality required for structural steel applica-tions. While each may have a particular area of strengthand/or weakness, the primary consideration as to whichprocess will be used is largely driven by cost. The avail-ability of specialized equipment in one fabrication shop,compared to the capabilities of a second shop, may dic-tate significantly different approaches, both of whichmay prove to be cost effective. A history of successfulusage offers a strong incentive for the fabricator to con-tinue using a given process. The reasons for this go wellbeyond familiarity and comfort with a specific approach.When welders and procedures are established with agiven process, significant costs will be incurred with anychange to a new approach.

3.1 Joint Requirements

Each individual weld joint configuration and preparationhas certain requirements of the welding process in orderto achieve low cost welding. Four characteristics mustbe considered: deposition rate, penetration ability, out-of-position capability, and high travel speed capacity.Each process exhibits different capabilities in theserealms. Once the joint and its associated requirementsare analyzed, they should be compared to the variousprocess options and the ability of the process to achievethose requirements. A proper match of weld jointrequirements and process capabilities will lead todependable and economical fabrication.

Some welds, such as large fillet welds and groove weldsrequire thathigh deposition ratewelding be used (Fig.3-1) for the most economical fabrication. The cost ofmaking these welds will be determined largely by thedeposition rate of the process. The amount of weld mate-rial required may be measured in pounds per foot ofjoint. Once the deposition rate of a process in pounds perhour is known, it is possible to determine the number of

feet of weld that can be made in a given hour assuming100% arc time. This, of course, translates directly to pro-ductivity rates.

The second criterion imposed by weld joints is therequirement for penetration. Examples are listed underFig. 3-2 and would include any complete joint penetra-tion groove weld that has a root face dimension. Thesejoints will be made by welding from one side and backgouging from the second to ensure complete fusion.With deeper penetration afforded by the welding process,a smaller amount of base metal will be required to beremoved by back gouging. Subsequent welding will thenbe proportionately reduced as well.

While all welding requires fusion, not all joints requiredeep penetration. For example, simple fillet welds arerequired by AWS D1.1-96 to have fusion to the root of thejoint, but are not required to have penetration beyond theroot. This has a practical basis: verification of penetrationbeyond the root is impossible with visual inspection.Fusion to the root, and not necessarily beyond, ensuresthat sufficient strength is generated, provided the weld isproperly sized. While penetration can be verified withultrasonic inspection, fillet welds routinely receive onlyvisual or magnetic particle inspection. Thus, no penetra-

Figure 3-1 Joints requiring substantial fill

Figure 3-2 Joints requiring substantial penetration

12

tion beyond the root is required, nor is design credit givento deeper penetration in fillet welds if it happens to be pre-sent. Figure 3-3 illustrates this requirement.

The out-of-position capability of a given welding processrefers to the ability to deposit weld metal in the verticalor overhead positions. It is generally more economical toposition the work in the flat and horizontal positions.However, this is usually impossible for field erection,and may be impractical under other conditions.

The ability to obtain high travel speeds is important forsmall welds. It may not be possible for a high depositionwelding process to be used at high travel speeds. Thesize of the droplet transferred, puddle fluidity, surfacetension, and other factors combine to make someprocesses more capable of high travel speeds than others.

3.2 Process Capabilities

After the joint is analyzed and specific requirementsdetermined, these are compared to the capabilities of var-ious processes. The process with capabilities most close-ly matching the requirements typically will be the bestand most economical option.

Submerged arc welding and electroslag/electrogas weld-ing have the greatest potential to deliver high depositionrates. Multiple electrode applications of submerged arcextend this capability even further. For joints requiringhigh deposition rates, submerged arc andelectroslag/electrogas welding are ideal processes to con-tribute to low cost welding. When the specific conditionsare not conducive to SAW but high deposition rates arestill required, flux cored arc welding may be used. Thelarger diameter electrodes, which run at higher electricalcurrents, are preferred.

Deep penetration is offered by the submerged arc weld-ing process. While electroslag/electrogas also offersdeep penetration, the joints on which electroslag are usedtypically do not require this capability. Where open arcprocesses are preferred, gas shielded flux cored weldingmay offer deep penetration.

Out-of-position capability is strongest for the flux coredand shielded metal arc welding processes. The slag coat-ings that are generated by these processes can be instru-mental in retaining molten weld metal in the vertical andoverhead positions. Submerged arc is not applicable forthese joints.

The requirement for high travel speed capability is fairlylimited in terms of welding structural steel members.This typically consists of the travel speed associated withmaking a 1/4 in. fillet weld. All of the popular process-es, with the exception of electroslag/electrogas, are capa-ble of making 1/4 in. fillet welds under the properconditions. Among the variables that need to be consid-ered are electrode size and procedure variables. A com-mon mistake of fabricators is to utilize a process andprocedure capable of extremely high deposition rates, butlimited travel speeds. Oversized welds can result fromthe inability to achieve high travel speeds. A more eco-nomical approach would be to optimize the procedureaccording to the desired travel speed. This may result ina lower deposition rate but a lower overall cost becauseoverwelding has been eliminated.

3.3 Special Situations

Self shielded flux cored welding is ideal for outdoor con-ditions. Quality deposits may be obtained without theerection of special wind shields and protection fromdrafts. Shielded metal arc welding is also suitable forthese conditions, but is considerably slower.

Figure 3-3 Fillet weld requirements

13

The welding process of choice for field erectors for thelast 25 years has been FCAW-ss. It has been the com-monly used process for fabrication of steel structuresthroughout the United States. Its advantages arereviewed in order to provide an understanding of why ithas been the preferred process. In addition, its limita-tions are outlined to highlight areas of potential concern.

The chief advantage of the FCAW-ss process is its abili-ty to deposit quality weld metal under field conditions,which usually involve wind. The graph shown in Figure

3-4 illustrates the effect of shielding gas loss on welddeposits, as well as the resistance to this problem withthe FCAW-ss process. The code specifically limits windvelocity in the vicinity of a weld to a maximum of 5miles per hour (2.3 m/s) (D1.1-96, paragraph 5.12.1). Inorder to utilize gas shielded processes under these condi-tions, it is necessary to erect windshields to precludemovement of the shielding gas with respect to the moltenweld puddle. While tents and other housings can be cre-ated to minimize this problem, such activities can becostly and are often a fire hazard. In addition, adequateventilation must be provided for the welder. The mostefficient windshields may preclude adequate ventilation.Under conditions of severe shielding loss, weld porositywill be exhibited. At much lower levels of shielding loss,the mechanical properties (e.g., notch toughness andductility) may be negatively affected, although there willbe no obvious evidence that this is taking place.

A variety of other gas-related issues are also eliminated,including ensuring availability of gas, handling of highpressure cylinders (always a safety concern), theft ofcylinders, protection of gas distribution hosing underfield conditions, and the cost of shielding gas. Leaks inthe delivery system obviously waste shielding gas, but aleak can also allow entry of air into the delivery system.Weld quality can be affected in the same way as shield-ing loss. Most field erectors have found it advantageousto utilize the self-shielded process and circumvent allsuch potential problems.

Some projects permit multiple welding heads to besimultaneously operated in the same general vicinity. Forsuch applications, submerged arc is an ideal choice.Because of the lack of arc flash, one operator can controlmultiple arcs that are nearly impossible to control in a sit-uation where the arc intensity from one arc would make itdifficult to carefully control another. A typical examplewould be the use of welding systems that simultaneouslymake fillet welds on opposing sides of stiffeners.

The easiest way tocontrol smoke and fumesin thewelding environment is to limit their initial generation.Here, submerged arc is ideal. Smoke exhaust guns areavailable for the flux cored arc welding processes. Themost effective process for use with these smoke exhaustguns is FCAW-ss. Because the process is self shielded,there is no concern about the disruption of the gas shield-ing. See 11 on arc welding safety.

Figure 3-4a Comparison of the effect of side wind ontensile elongation of: (a) CO2-shielded and (b) selfshielded ferritic steel weld metals.(source: Self-Shielded Arc Welding. T. Boniszewski, 1992.)

Figure 3-4b Comparison of the effect of side wind speedon Charpy V-notch impact toughness of: (a) CO2 -shield-ed at room temperature and (b) self shielded ferritic steelweld metals at 0°C.(source: Self-Shielded Arc Welding. T. Boniszewski, 1992.)

14

4 Welding Cost Analysis

Welding is a labor intensive technology. Electricity,equipment depreciation, electrodes, gases, and fluxesconstitute a very small portion of the total welding cost.Therefore, the prime focus of cost control will be reduc-ing the amount of time required to make a weld.

The following example is given to illustrate the relativecosts of material and labor, as well as to assess the effectsof proper process selection. The example to be consid-ered is the groove weld of beam flange to column con-nections. Since this is a multiple pass weld, the mostappropriate analysis method is to consider the weldingcost per weight of weld metal deposited, such as $/lb.Other analysis methods include cost per piece, ideal formanufacturers associated with the production of identicalparts on a repetitive basis. Another method is cost perlength, appropriate for single pass welds with substantiallength. The two welding processes to be considered areshielded metal arc welding and flux cored arc welding.Either would generate high quality welds when properlyused.

To calculate the cost per weight of weld metal deposited,an equation taking the following format is used:

The cost of the electrode is simply the purchase cost ofthe welding consumable used. Not all of this filler metalis converted directly to deposited weld metal. There arelosses associated with slag, spatter, and in the case ofSMAW, the stub loss (the end portion of the electrodethat is discarded). To account for these differences, anefficiency factor is applied. The following efficiency fac-tors are typically used for the various welding processes:

The cost to deposit the weld metal is determined bydividing the applicable labor and overhead rate by thedeposition rate, that is, the amount of weld metal deposit-ed in a theoretical, continuous one hour of production.This cannot be maintained under actual conditions sincewelding will be interrupted by many factors, includingslag removal, replacement of electrode, repositioning of

the work or the welder with respect to the work, etc. Toaccount for this time, an “operating factor” is used whichis defined as the “arc-on” time divided by the total timeassociated with welding activities. For SMAW, replace-ment of electrodes takes place approximately everyminute because of the finite length of the electrodes used.The following operating factors are typically used for thevarious processes and method of application:Operating factors for any given process can vary widely,

depending on what a welder is required to do. In shopsituations, a welder may receive tacked assemblies andbe required only to weld and clean them. For field erec-tion, the welder may “hang iron,” fit, tack, bolt, clean thejoint, reposition scaffolding and other activities in addi-tion to welding. Obviously operating factors will be sig-nificantly reduced under these conditions.

The following examples are the actual procedures usedby a field erector. The labor and overhead cost does notnecessarily represent actual practice. The operating fac-tors are unrealistically high for a field erection site, buthave been used to enable comparison of the relative costof filler metals vs. the labor required to deposit the weldmetal, as well as the difference in cost for differentprocesses. Once the cost per deposited pound is known,it is relatively simple to determine the quantity of weldmetal required for a given project, and multiply it by thecost per weight to determine the cost of welding on theproject.

Method Operating FactorManual SMAW 30%Semiautomatic 40%Mechanized 50%

Process EfficiencySMAW 60%FCAW 80%GMAW 90% (CO2 shielding)

98% (Mixed gas)SAW 100% (Flux not included)

Process SMAW FCAWElectrode Classification E7018 E70TG-K2Electrode Diameter 3/16” 7/64”Amperage 225 430Voltage N.A. 27Electrode Efficiency 60% 80%Electrode Cost $1.23/lb. $2.27/lb.Operating Factor 30% 40%Deposition Rate 5.5 lb./hr. 14.5 lb./hr.Labor and Overhead Rate $50/hr $50/hr.

Cost perweight

= +Electrode Cost

EfficiencyLabor + Overhead Rate

(Deposition Rate) (Operating Factor)

Cost perweight

= $2.05 + $30.30 = $32.35/lb.= +$1.2360%

$50.00(5.5) (30%)

Cost perweight

= +$2.2780%

$50.00(14.5) (40%)

= $2.84 + $8.62 = $11.46/lb.

For SMAW:

For FCAW:

15

In the SMAW example, the electrode cost is approxi-mately 6% of the total cost. For the FCAW example, pri-marily due to a decrease in the labor content, theelectrode cost is 25% of the total. By using FCAW, thetotal cost of welding was decreased approximately 65%.While the FCAW electrode costs 85% more than theSMAW electrode, the higher electrode efficiency reducesthe increase in electrode cost to only 39%.

The first priority that must be maintained when selectingwelding processes and procedures is the achievement ofthe required weld quality. For different welding methodswhich deliver the required quality, it is generally advan-tageous to utilize the method that results in higher depo-sition rates and higher operating factors. This will resultin reduced welding time with a corresponding decreasein the total building erection cycle, which will generallytranslate to a direct savings for the final owner, not onlylowering the cost of direct labor, but also reducing con-struction loan costs.

5 Welding Procedures

Within the welding industry, the term “WeldingProcedure Specification” (or WPS) is used to signify thecombination of variables that are to be used to make acertain weld. The terms “Welding Procedure,” or simply“Procedure,” may be used. At a minimum, the WPS con-sists of the following:

WPS Variables

Process (SMAW, FCAW, etc.)Electrode specification (AWS A5.1, A5.20, etc.)Electrode classification (E7018, E71T-1, etc.)Electrode diameter (1/8 in., 5/32 in., etc.)Electrical characteristics (AC, DC+, DC-)Base metal specification (A36, A572 Gr50, etc.)Minimum preheat and

interpass temperatureWelding current (amperage)/wire feed speedArc voltageTravel speedPosition of weldingPost weld heat treatmentShielding gas type and flow rateJoint design details

The welding procedure is somewhat analogous to acook’s recipe. It outlines the steps required to make aweld of the required quality under specific conditions.

5.1 Effects of Welding Variables

The effects of the variables are somewhat dependent onthe welding process being employed, but general trendsapply to all the processes. It is important to distinguishthe difference between constant current (CC) and con-stant voltage (CV) electrical welding systems. Shieldedmetal arc welding is always done with a CC system.Flux cored welding and gas metal arc welding generallyare performed with CV systems. Submerged arc mayutilize either.

Amperage is a measure of the amount of current flowingthrough the electrode and the work. It is a primary vari-able in determining heat input. Generally, an increase inamperage means higher deposition rates, deeper penetra-tion, and more admixture. The amperage flowingthrough an electrical circuit is the same, regardless ofwhere it is measured. It may be measured with a tongmeter or with the use of an electrical shunt. The role ofamperage is best understood in the context of heat inputand current density considerations. For CV welding, anincrease in wire feed speed will directly increase amper-age. For SMAW on CC systems, the machine settingdetermines the basic amperage, although changes in thearc length (controlled by the welder) will further changeamperage. Longer arc lengths reduce amperage.

Arc voltage is directly related to arc length. As the volt-age increases, the arc length increases, as does thedemand for arc shielding. For CV welding, the voltageis determined primarily by the machine setting, so the arclength is relatively fixed in CV welding. For SMAW onCC systems, the arc voltage is determined by the arclength, which is manipulated by the welder. As arclengths are increased with SMAW, the arc voltage willincrease, and the amperage will decrease. Arc voltagealso controls the width of the weld bead, with highervoltages generating wider beads. Arc voltage has a directeffect on the heat input computation.

The voltage in a welding circuit is not constant, but iscomposed of a series of voltage drops. Consider the fol-lowing example: assume the power source delivers a totalsystem voltage of 40 volts. Between the power sourceand the welding head or gun, there is a voltage drop ofperhaps 3 volts associated with the input cable resistance.From the point of attachment of the work lead to thepower source work terminal, there is an additional volt-age drop of, say, 7 volts. Subtracting the 3 volts and the

16

7 volts from the original 40, this leaves 30 volts for thearc. This example illustrates how important it is toensure that the voltages used for monitoring welding pro-cedures properly recognize any losses in the welding cir-cuit. The most accurate way to determine arc voltage isto measure the voltage drop between the contact tip andthe work piece. This may not be practical for semiauto-matic welding, so voltage is typically read from a pointon the wire feeder (where the gun and cable connectionis made), to the workpiece. For SMAW welding, voltageis not usually monitored, since it is constantly changingand cannot be controlled except by the welder. Skilledwelders hold short arc lengths to deliver the best weldquality.

Travel speed, measured in inches per minute, is the rateat which the electrode is moved relative to the joint. Allother variables being equal, travel speed has an inverseeffect on the size of the weld beads. As the travel speedincreases, the weld size will decrease. Extremely lowtravel speeds may result in reduced penetration, as the arcimpinges on a thick layer of molten metal and the weldpuddle rolls ahead of the arc. Travel speed is a key vari-able used in computing heat input; reducing travel speedincreases heat input.

Wire feed speedis a measure of the rate at which theelectrode is passed through the welding gun and deliv-ered to the arc. Typically measured in inches per minute(ipm) the wire feed speed is directly proportional todeposition rate, and directly related to amperage. Whenall other welding conditions are maintained constant(e.g., the same electrode type, diameter, electrode exten-sion, arc voltage, and electrode extension), an increase inwire feed speed will directly lead to an increase inamperage. For slower wire feed speeds, the ratio of wirefeed speed to amperage is relatively constant and linear.

For higher levels of wire feed speed, it is possible toincrease the wire feed speed at a disproportionately highrate compared to the increase in amperage. When theseconditions exist, the deposition rate per amp increases,but at the expense of penetration.

Wire feed speed is the preferred method of maintainingwelding procedures for constant voltage wire feedprocesses. The wire feed speed can be independentlyadjusted, and measured directly, regardless of the otherwelding conditions. It is possible to utilize amperage asan alternative to wire feed speed although the resultantamperage for a given wire feed speed may vary, depend-

ing on the polarity, electrode diameter, electrode type,and electrode extension. Although equipment has beenavailable for twenty years that monitors wire feed speed,many codes such as AWS D1.1 continue to acknowledgeamperage as the primary method for procedure docu-mentation. D1.1 does permit the use of wire feed speedcontrol instead of amperage, providing a wire feed speedamperage relationship chart is available for comparison.Specification sheets for various Lincoln electrodes pro-vide data that report these relationships.

Electrode extension, also known as “stickout,” or ESO,is the distance from the contact tip to the end of the elec-trode. It applies only to the wire fed processes. As theelectrode extension is increased in a constant voltage sys-tem, the electrical resistance of the electrode increases,causing the electrode to be heated. This is known asresistance heating or “I2R heating.” As the amount ofheating increases, the arc energy required to melt theelectrode decreases. Longer electrode extensions may beemployed to gain higher deposition rates at a givenamperage. When the electrode extension is increasedwithout any change in wire feed speed, the amperage willdecrease. This results in less penetration and less admix-ture. With the increase in electrode stickout, it is com-mon to increase the machine voltage setting tocompensate for the greater voltage drop across the elec-trode.

In constant voltage systems, it is possible to simultane-ously increase the electrode stickout and wire feed speedin a balanced manner so that the current remains con-stant. When this is done, higher deposition rates areattained. Other welding variables such as voltage andtravel speed must be adjusted to maintain a stable arc andto ensure quality welding. The ESO variable shouldalways be within the range recommended by the manu-facturer.

Electrode diameter— Larger electrodes can carry high-er welding currents. For a fixed amperage, however,smaller electrodes result in higher deposition rates. Thisis because of the effect on current density discussedbelow.

Polarity is a definition of the direction of current flow.Positive polarity (reverse) is achieved when the electrodelead is connected to the positive terminal of the directcurrent (DC) power supply. The work lead is connectedto the negative terminal. Negative polarity (straight)occurs when the electrode is connected to the negative

17

terminal and the work lead to the positive terminal.Alternating current (AC) is not a polarity, but a currenttype. With AC, the electrode is alternately positive andnegative. Submerged arc is the only process that com-monly uses either electrode positive or electrode negativepolarity for the same type of electrode. AC may also beused. For a fixed wire feed speed, a submerged arc elec-trode will require more amperage on positive polaritythan on negative. For a fixed amperage, it is possible toutilize higher wire feed speeds and deposition rates withnegative polarity than with positive. AC exhibits a mixof both positive and negative polarity characteristics.The magnetic field that surrounds any DC conductor cancause a phenomenon known as arc blow, where the arc isphysically deflected by the field. The strength of themagnetic field is proportional to the square of the currentvalue, so this is a more significant potential problem withhigher currents. AC is less prone to arc blow, and cansometimes be used to overcome this phenomenon.

Heat input is proportional to the welding amperage,times the arc voltage, divided by the travel speed. Higherheat inputs relate to larger weld cross sectional areas, andlarger heat affected zones, which may negatively affectmechanical properties in that region. Higher heat inputusually results in slightly decreased yield and tensilestrength in the weld metal, and generally lower notchtoughness because of the interaction of bead size andheat input.

Current density is determined by dividing the weldingamperage by the cross sectional area of the electrode. Forsolid electrodes, the current density is therefore proportion-al to I/d2. For tubular electrodes where current is conduct-ed by the sheath, the current density is related to the area ofthe metallic cross section. As the current density increases,there will be an increase in deposition rates, as well as pen-etration. The latter will increase the amount of admixturefor a given joint. Notice that this may be accomplished byeither increasing the amperage or decreasing the electrodesize. Because the electrode diameter is a squared function,a small decrease in diameter may have a significant effecton deposition rates and plate penetration.

Preheat and interpass temperature are used to controlcracking tendencies, typically in the base materials.Regarding weld metal properties, for most carbon-man-ganese-silicon systems, a moderate interpass temperaturepromotes good notch toughness. Preheat and interpasstemperatures greater than 550°F may negatively affect

notch toughness (AWS Position Statement, p. 7). Whenthe base metal receives little or no preheat, the resultantrapid cooling may also lead to a deterioration of notchtoughness. Therefore, careful control of preheat andinterpass temperatures is critical.

5.2 Purpose of Welding Procedure Specifications

The particular values for the variables discussed in 5.1have a significant effect on weld soundness, mechanicalproperties, and productivity. It is therefore critical thatthose procedural values used in the actual fabrication anderection be appropriate for the specific requirements ofthe applicable code and job specifications. Welds thatwill be architecturally exposed, for example, should bemade with procedures that minimize spatter, encourageexceptional surface finish, and have limited or no under-cut. Welds that will be covered with fireproofing, in con-trast, would naturally have less restrictive cosmeticrequirements.

Many issues must be considered when selecting weldingprocedure values. While all welds must have fusion toensure their strength, the required level of penetration is afunction of the joint design and the weld type. All weldsare required to deliver a certain yield and/or tensilestrength, although the exact level required is a function ofthe connection design. Not all welds are required todeliver minimum specified levels of notch toughness.Acceptable levels of undercut and porosity are a functionof the type of loading applied to the weld. Determinationof the most efficient means by which these conditions canbe met cannot be left to the welders, but is determined byknowledgeable welding technicians and engineers whocreate written Welding Procedure Specifications (WPSs)and communicate those requirements to welders by themeans of these documents. The WPS is the primary toolthat is used to communicate to the welder, supervisor, andthe inspector how a specific weld is to be made. The suit-ability of a weld made by a skilled welder in conformancewith the requirements of a WPS can only be as good asthe WPS itself. The proper selection of procedural vari-able values must be achieved in order to have a WPSappropriate for the application. This is the job of thewelding expert who generates or writes the WPS. Thewelder is generally expected to be able to follow theWPS, although the welder may not know how or whyeach particular variable was selected. Welders are expect-ed to ensure welding is performed in accordance with theWPS. Inspectors do not develop WPSs, but should ensurethat they are available and are followed.

18

The D1.1-96 Structural Welding Code - Steel requireswritten welding procedures for all fabrication performed(D1.1-96, paragraph 5.5). The inspector is obligated toreview the WPSs and to make certain that productionwelding parameters conform to the requirements of thecode (D1.1-96, paragraph 6.3.1). These WPSs arerequired to be written, regardless of whether they are pre-qualified or qualified by test (sections 5.3and 5.5). Eachfabricator or erector is responsible for the developmentof WPSs (D1.1-96, paragraph 4.1.1.1, 4.6). Confusionon this issue apparently still exists since there continue tobe reports of fabrication being performed in the absenceof written welding procedure specifications. One preva-lent misconception is that if the actual parameters underwhich welding will be performed meet all the conditionsfor “prequalified” status, written WPSs are not required.This is not true. As has been shown in the cited code ref-erences, the requirement is clear.

The WPS is a communication tool, and it is the primarymeans of communication to all the parties involvedregarding how the welding is to be performed. It musttherefore be readily available to foremen, inspectors andthe welders.

The code is not prescriptive in its requirements regardingthe availability and distribution of WPSs. Some shopfabricators have issued each welder employed in theirorganization with a set of welding procedures that aretypically retained in the welder’s locker or tool box.Others have listed WPS parameters on shop drawings.Some company bulletin boards have listings of typicalWPSs used in the organization. The AWS D1. PositionStatement suggest that WPSs should be posted near thepoint where welding is being performed. Regardless ofthe method used, WPSs must be available to those autho-rized to use them.

It is in the contractor’s best interest to ensure efficientcommunication with all parties involved. Not only canquality be compromised when WPSs are not available,but productivity can suffer as well. Regarding quality,the limits of suitable operation of the particular weldingprocess and electrode for the steel, joint design and posi-tion of welding must be understood. It is obvious that theparticular electrode employed must be operated on theproper polarity, that proper shielding gases are used, thatamperage levels be appropriate for the diameter of elec-trode, and appropriate for the thickness of material onwhich welding is performed. Other issues may not be soobviously apparent. The required preheat for a particular

application is a function of the grade(s) of steel involved,the thickness(es) of material, and the type of electrodeemployed (whether low hydrogen or non-low hydrogen).The required preheat level can all be communicated bymeans of the written WPS.

Lack of conformance with the parameters outlined in theWPS may result in the deposition of a weld that does notmeet the quality requirements imposed by the code or thejob specifications. When an unacceptable weld is made,the corrective measures to be taken may necessitate weldremoval and replacement, an activity that routinelyincreases the cost of that particular weld tenfold.Avoiding these types of unnecessary activities by clearcommunication has obvious quality and economic rami-fications.

Equipment such as Lincoln Electric’s LN-9 wire feeder,which has the ability to have preset welding parameters,coupled with a digital LED display that indicates opera-tional parameters, can assist in maintaining and monitor-ing WPS parameters. Analog meters can be used onother wire feeders.

The code imposes minimum requirements for a givenproject. Additional requirements may be imposed bycontract specifications. The same would hold trueregarding WPS values. Compliance with the minimumrequirements of the code may not be adequate under allcircumstances. Additional requirements can be commu-nicated through the WPS. For example, the D1.1-96code permits the use of an E71T-11 FCAW electrode formultiple pass welding without any restriction on platethickness. The Lincoln Electric product, InnershieldNR211MP, has a maximum thickness restrictionimposed by the manufacturer of 1/2 in. This additionalrequirement can be incorporated into the applicableWPS. Other recommendations that may be imposed bythe steel producer, electrode manufacturer, or others canand should be documented in the WPS.

5.3 Prequalified Welding Procedure Specifications

The AWS D1.1 code provides for the use of prequalifiedWPSs. Prequalified WPSs are those that the AWS D1Committee has determined to have a history of acceptableperformance, and so they are not subject to the qualifica-tion testing imposed on all other welding procedures. Theuse of prequalified WPSs does not preclude their need tobe in a written format. The use of prequalified WPSs stillrequires that the welders be appropriately qualified. All

19

the workmanship provisions imposed in the fabricationsection of the code apply to prequalified WPSs. The onlycode requirement exempted by prequalification is thenondestructive testing and mechanical testing required forqualification of welding procedures.

A host of restrictions and limitations imposed on pre-qualified welding procedures do not apply to weldingprocedures that are qualified by test. Prequalified weld-ing procedures must conform with all the prequalifiedrequirements in the code. Failure to comply with a sin-gle prequalified condition eliminates the opportunity forthe welding procedure to be prequalified (D1.1-96, para-graph 3.1).

The use of a prequalified welding procedure does notexempt the engineer from exercising engineering judg-ment to determine the suitability of the particular proce-dure for the specific application (D1.1-96, paragraph 3.1).

In order for a WPS to be prequalified, the following con-ditions must be met:

• The welding process must be prequalified. OnlySMAW, SAW, GMAW (except GMAW-s), and FCAWmay be prequalified (D1.1-96, paragraph 3.2.1).

• The base metal/filler metal combination must be pre-qualified. Prequalified base metals, filler metals, andcombinations are shown in D1.1-96, paragraph 3.3,Table 3.1.

• The minimum preheat and interpass temperatures pre-scribed in D1.1-96, paragraph 3.3, Table 3.2 must beemployed (D1.1-96, paragraph 3.5).

• Specific requirements for the various weld types mustbe maintained. Fillet welds must be in accordancewith D1.1-96, paragraph 3.9, plug and slot welds inaccordance with D1.1-96, paragraph 3.10, and groovewelds in accordance with D1.1-96, paragraphs 3.11,3.12, and 3.13, as applicable. For the groove welds,whether partial joint penetration or complete joint pen-etration, the required groove preparation dimensionsare shown in D1.1-96, Figures 3.3 and 3.4.

Even if prequalified joint details are employed, the weld-ing procedure must be qualified by test if other prequali-fied conditions are not met. For example, if aprequalified detail is used on an unlisted steel, the weld-ing procedures must be qualified by test.

Prequalified status requires conformance to a variety ofprocedural parameters. These are largely contained inD1.1-96, Table 3.7, and include maximum electrodediameters, maximum welding current, maximum rootpass thickness, maximum fill pass thicknesses, maxi-mum single-pass fillet weld sizes, and maximum singlepass weld layers (D1.1-96, Table 3.3).

In addition to all the preceding requirements, weldingperformed with a prequalified WPS must be in confor-mance with the other code provisions contained in thefabrication section of AWS D1.1-96 Structural WeldingCode.

The code does not imply that a prequalified WPS willautomatically achieve the quality conditions required bythe code. It is the contractor’s responsibility to ensurethat the particular parameters selected within the require-ments of the prequalified WPS are suitable for the spe-cific application. An extreme example will serve as anillustration. Consider a (hypothetical) proposed WPS formaking a 1/4 in. fillet weld on 3/8 in. A36 steel in the flatposition. The weld type and steel are prequalified. SAW,a prequalified process, is selected. The filler metalselected is F7A2-EM12K, meeting the requirements ofD1.1-96, Table 3.1. No preheat is specified since itwould not be required according to D1.1-96, Table 3.2.The electrode diameter selected is 3/32 in., less than the1/4 in. maximum specified in D1.1- 96, Table 3.7. Themaximum single pass fillet weld size in the flat position,according to D1.1-96, Table 3.7, is unlimited, so the 1/4in. fillet size can be prequalified. The current levelselected for making this particular fillet weld is 800amps, less than the 1000 amp maximum specified inD1.1-96, Table 3.7.

However, the amperage level imposed on the electrodediameter for the thickness of steel on which the weld isbeing made is inappropriate. It would not meet therequirements of D1.1-96, paragraph 5.3.1.2, whichrequires that the size of electrode and amperage be suit-able for the thickness of material being welded. Thisillustration demonstrates the fact that compliance with allprequalified conditions does not guarantee that the com-bination of selected variables will always generate anacceptable weld.

Most contractors will determine preliminary values for aprequalified WPS based upon their experience, recom-mendations from publications such as Lincoln Electric’sProcedure Handbook of Arc Welding, the AWS Welding

20

Handbooks, AWS Welding Procedures Specifications(AWS B2.1), or other sources. It is the responsibility ofthe contractor to verify the suitability of the suggestedparameters prior to applying the actual procedure on aproject, although the verification test need not be subjectto the full range of procedure qualification tests imposedby the code. Typical tests will be made to determinesoundness of the weld deposit (e.g., fusion, tie-in of weldbeads, freedom from slag inclusions, etc.). The platecould be nondestructively tested or, as is more common-ly done, cut, polished, and etched. The latter operationsallow for examination of penetration patterns, beadshapes, and tie-in. Welds that are made with prequalifiedWPSs meeting the physical dimensional requirements(fillet weld size, maximum reinforcement levels, and sur-face profile requirements), and that are sound (that is,adequate fusion, tie-in and freedom from excessive slagcharacterized by inclusions and porosity) should meetthe strength and ductility requirements imposed by thecode for welding procedures qualified by test. Weldsoundness, however, cannot be automatically assumedjust because the WPS is prequalified.

5.4 Guidelines For Preparing Prequalified WPSs

When developing prequalified WPSs, the starting point isa set of welding parameters appropriate for the generalapplication being considered. Parameters for overheadwelding will naturally vary from those required fordown-hand welding. The thickness of the materialinvolved will dictate electrode sizes and correspondingcurrent levels. The specific filler metals selected willreflect the strength requirements of the connection.Many other issues must be considered.

Depending on the level of familiarity and comfort thecontractor has with the particular values selected, weld-ing a mock-up may be appropriate. Once the parametersthat are desired for use in production are established, it isessential to check each of the applicable parameters forcompliance with the D1.1-96 code.

To assist in this effort, Annex H has been provided in theD1.1-96 code. This contains a check list that identifiesprequalified requirements. If any single parameter devi-ates from these requirements, the contractor is left withtwo options: (1) the preliminary procedure can be adjust-ed to conform with the prequalified constraints; or, (2)the WPS can be qualified by test. If the preliminary pro-cedure is adjusted, it may be appropriate to reexamine itsviability by another mock-up.

The next step is to document, in writing, the prequalifiedWPS values. A sample form is included in Annex E ofthe code. The fabricator may utilize any convenient for-mat (D1.1-96, paragraph 3.6). Also contained in AnnexE are a series of examples of completed WPSs that maybe used as a pattern.

5.5 Qualifying Welding Procedures By Test

Conducting qualification tests— There are two prima-ry reasons why welding procedures may be qualified bytest. First, it may be a contractual requirement.Secondly, one or more of the specific conditions to beused in production may deviate from the prequalifiedrequirements. In either case, a test weld must be madeprior to the establishment of the final WPS. The first stepin qualifying a welding procedure by test is to establishthe precise parameters to be qualified. The same sourcescited for the prequalified WPS starting points could beused for WPSs qualified by test. These will typically bethe parameters used for fabrication of the test plate,although this is not always the case, as will be discussedlater. In the simplest case, the exact conditions that willbe encountered in production will be replicated in theprocedure qualification test. These would include thewelding process, filler metal, grade of steel, joint details,thicknesses of material, minimum preheat temperature,interpass temperature, and the various welding parame-ters of amperage, voltage, and travel speed. The initialparameters used to make the procedure qualification testplate beg for a name to define them, although there is nostandard industry term. It has been suggested that“TWPS” be used where the “T” could alternately be usedfor temporary, test, or trial. In any case, it would definethe parameters to be used for making the test plate sincethe validity of the particular parameters cannot be veri-fied until successfully passing the required test. Theparameters for the test weld are recorded on a ProcedureQualification Record (PQR). The actual values usedshould be recorded on this document. The target voltage,for example, may be 30 volts whereas, in actual fact, only29 volts were used for making the test plate. The 29 voltswould be recorded.

After the test plate has been welded, it is allowed tocool and the plate is subjected to the visual and nonde-structive testing prescribed by the code. The specifictests required are a function of the type of weld beingmade and the particular welding consumables. Thetypes of qualification tests are described in D1.1-96,paragraph 4.4.

21

In order to be acceptable, the test plates must first passvisual inspection followed by nondestructive testing(NDT) (D1.1-96, paragraphs 4.8.1, 4.8.2). At the con-tractor’s option, either RT or UT can be used for NDT.The mechanical tests required involve bend tests (forsoundness), macro etch tests (for soundness), andreduced section tensile tests (for strength). For qualifi-cation of procedures on steels with significantly differentmechanical properties, a longitudinal bend specimen ispossible (D1.1-96, paragraph 4.8.3.2). All weld metaltensile tests are required for unlisted filler metals. Thenature of the bend specimens, whether side, face, or root,is a function of the thickness of the steel involved. Thenumber and type of tests required are defined in D1.1-96,Table 4.2 for complete joint penetration groove welds,D1.1-96, Table 4.3 for partial joint penetration groovewelds, and D1.1-96, Table 4.4 for fillet welds.

Once the number of tests has been determined, the testplate is sectioned and the specimens machined for test-ing. The results of the tests are recorded on the PQR.According to D1.1-96, if the test results meet all the pre-scribed requirements, the testing is successful and weld-ing procedures can be established based upon thesuccessful PQR. If the test results are unsuccessful, thePQR cannot be used to establish the WPS. If any onespecimen of those tested fails to meet the test require-ments, two retests of that particular type of test may beperformed with specimens extracted from the same testplate. If both of the supplemental specimens meet therequirements, the D1.1-96 allows the tests to be deemedsuccessful. If the test plate is over 1-1/2 in. thick, failureof a specimen necessitates retesting of all the specimensat the same time from two additional locations in the testmaterial (D1.1-96, paragraph 4.8.5).

It is wise to retain the PQR’s from unsuccessful tests, asthey may be valuable in the future when another similarwelding procedure is contemplated for testing.

The acceptance criteria for the various tests are pre-scribed in the code. The reduced section tensile tests arerequired to exceed the minimum specified tensilestrength of the steel being joined (D1.1-96, paragraph4.8.3.5). Specific limits on the size, location, distribu-tion, and type of indication on bend specimens is pre-scribed in D1.1-96, paragraph 4.8.3.3.

Writing WPSs from successful PQR’s— When a PQRrecords the successful completion of the required tests,welding procedures may be written from that PQR. At aminimum, the values used for the test weld will consti-tute a valid WPS. The values recorded on the PQR aresimply transcribed to a separate form, now known as aWPS rather than a PQR.

It is possible to write more than one WPS from a suc-cessful PQR. Welding procedures that are sufficientlysimilar to those tested can be supported by the samePQR. Significant deviations from those conditions, how-ever, necessitate additional qualification testing.Changes that are considered significant enough to war-rant additional testing are considered essential variables,and these are listed in D1.1-96, Tables 4.5, 4.6, and 4.7.For example, consider an SMAW welding procedure thatis qualified by test using an E8018-C3 electrode. Fromthat test, it would be possible to write a WPS that utilizesE7018 (since this is a decrease in electrode strength) butit would not be permissible to write a WPS that utilizesE9018-G electrode (because Table 4.5 lists an increase infiller metal classification strength as an essential vari-able). It is important to carefully review the essentialvariables in order to determine whether a previously con-ducted test may be used to substantiate the new proce-dure being contemplated.

D1.1-96, Table 4.1 defines the range of weld types andpositions qualified by various tests. This table is bestused not as an “after-the-fact” evaluation of the extent ofapplicability of the test already conducted, but rather forplanning qualification tests. For example, a test plateconducted in the 2G position qualifies the WPS for use ineither the 1G or 2G position. Even though the first antic-ipated use of the WPS may be for the 1G position, it maybe advisable to qualify in the 2G position so that addi-tional usage can be obtained from this test plate.

In a similar way, D1.1-96, Table 4.7 defines whatchanges can be made in the base metals used in produc-tion vs. qualification testing. An alternate steel may beselected for the qualification testing simply because itaffords additional flexibility for future applications.

If WPS qualification is performed on a non-prequalifiedjoint geometry, and acceptable test results are obtained,WPSs may be written from that PQR utilizing any of theprequalified joint geometries (D1.1-96, Table 4.5, Item 32).

22

5.6 Examples

To provide some insight into the thought process that awelding engineer may follow to develop a WPS, twoexamples will be given. In both cases, the weld is thesame, namely, a 5/16 in. fillet weld. The specific appli-cation conditions, however, will necessitate that a sepa-rate WPS be developed for each situation. A sampleWPS is included for each situation.

Situation One:The weld to be made is a 5/16 in. filletweld that connects the shear tab to the column. Thisweld will be made in the fabrication shop with a columnin the horizontal position. The fillet weld is applied toeither side of a 1/2 in. shear tab. It is welded to a W14 X311 column with a flange thickness of 2-1/4 inches. Theshear tab is made of A36 steel, while the column is ofA572 Gr 50.

The welding engineer recognizes that for the grades ofsteel involved, and for the type of weld specified, a pre-qualified WPS could be written. The process of choicefor this particular shop fabricator is gas shielded fluxcored arc welding, a prequalified welding process. FromTable 3.1 of the D1.1-96 code, a list of prequalified fillermetals is given. Outershield 70, an E70T-1 electrode, isselected because, for semiautomatic welding, it is likelyto be the most economical welding process consideringdeposition rate and cleanup time. The electrode operateson DC+ polarity. From experience, the engineer knowsthat a 3/32 in. diameter is appropriate for the application,and specifies that the shielding gas should be CO2 basedupon the electrode manufacturer’s recommendation andits low cost characteristics. From Table 3.2 of the D1.1-96 code, the preheat is selected. It is controlled by thethicker steel, that is, the column flange, and required tobe a minimum of 150°F since the column flange thick-ness is 2-1/4 inches. From recommendations supplied bythe electrode manufacturer, the welding engineer selectsa welding current of 460 amps, 31 volts, and specifiesthat the welding speed should be 15-17 inches perminute. The final variable is determined based uponexperience. If any doubts still exist, a simple fillet weldtest could be made to verify the travel speed for the givenamperage.

As a quick check, the engineer reviews Annex H toensure that all the prequalified conditions have beenachieved. Finally, these are tabulated on the WPS (seeFigure 5-1 for an example of a WPS). The particularform used was a copy from the D1.1-96 code, although

any convenient format could have been used, provided allthe required information was given.

Situation Two: The second weld to be made is also a 5/16in. fillet weld, but in this case, the weld will be made inthe field. The weld will be made between the shear tabdescribed above, and the beam web. In this situation, thebeam is a W36 X 150, specified to be of A36 steel.Under field conditions, the weld must be made in the ver-tical position.

The welding engineer again recognizes that the WPS forthis application could be prequalified if all the applicableconditions are met. Self shielded flux cored arc weldingis selected in order to ensure high quality welds underwindy conditions. This is a prequalified process. InD1.1-96, Table 3.1, the engineer locates suitable fillermetals and selects Innershield NR232, an E71T-8 self-shielded flux cored electrode which operates on DC neg-ative polarity. Because the welding will be made in thevertical position, a 0.068 in. diameter electrode is speci-fied. From technical literature supplied by LincolnElectric, a middle-of-the-range procedure suitable forvertical position welding is selected. The engineer spec-ifies the current to be 250 amps, 19-21 volts, with a trav-el speed of 5.5-6.5 inches per minute. The controllingvariable is the thickness of the beam web, which is 5/8 in.In this situation, Table 3.2 of the D1.1-96 code does notrequire any minimum preheat. These parameters arerecorded on the WPS form shown in Figure 5-2.

The two welds to be made are remarkably similar, andyet the WPS values specified are significantly different.In order to ensure that quality welds are delivered at eco-nomical rates, it is imperative that a knowledgeable indi-vidual establish WPS values. These values must beadhered to during fabrication and erection in order toensure quality welds in the final structure.

5.7 Approval of WPSs

After a WPS is developed by the fabricator or erector,D1.1 requires that it be reviewed. For prequalifiedWPSs, the inspector is required to review the WPSs toensure that they meet all the prequalified requirements(AWS D1.1-96, paragraph 6.3.1). The code requiresWPSs qualified by test to be submitted to the engineerfor review (D1.1-96, paragraph 4.1.1).

The apparent logic behind the differences in approvalprocedures is that while prequalified WPSs are based

23

upon well established, time proven, and documentedwelding practices, WPSs that have been qualified by testare not automatically subject to such restrictions. Eventhough the required qualification tests have demonstratedthe adequacy of the particular procedure under test con-ditions, further scrutiny by the engineer is justified toensure that it is applicable for the particular situation thatwill be encountered in production.

In practice, it is common for the engineer to delegate theapproval activity of all WPSs to the inspector. There is apractical justification for such activity: the engineer mayhave a more limited understanding of welding engineer-ing, and the inspector may be more qualified for thisfunction. While this practice may be acceptable for typ-ical projects that utilize common materials, more scruti-ny is justified for unusual applications that utilizematerials in ways that deviate significantly from normalpractice. In such situations, it is advisable for the engi-neer to retain the services of a welding expert to evaluatethe suitability of the WPSs for the specific application.

6 Fabrication and Erection Guidelines

6.1 Fit-Up and Assembly

It is important that proper dimensional tolerances bemaintained in fabrication. The code prescribes “asdetailed” tolerances that apply to variations that can betaken from the prequalified dimensions (D1.1-96, Figures3.3, 3.4). These dimensions must be shown on shopdrawings, but allow the contractor to deviate from the pre-scribed dimensions within the allowable limits. If fromexperience, for example, a contractor realizes that a larg-er root opening is favorable for obtaining the required rootpenetration, an increase can be made within the limitshown and the geometry is still considered prequalified.“As fit” tolerances apply to the shop drawing dimensions,which may have been increased by the “as detailed” tol-erance added to the geometries prescribed for the pre-qualified joint. For joints that are qualified by test,D1.1-96, Figure 5.3 contains tolerance limitations.

It is particularly important that adequate access to theroot of groove welds is provided. Narrow root openingsand tight included angles make it more difficult to obtainuniform fusion in the root. Excessively tight root geome-tries encourage improper width-to-depth ratios that maypromote centerline cracking in the weld bead. A width-to-depth ratio of 1.2:1 is generally adequate to preventcenterline cracking.

Excessively wide included angles and root openingswaste weld metal, and increase the degree of shrinkagethat will occur in the connection, resulting in an increasein distortion and/or residual stress. When the root open-ings are greater than those permitted, but not greater thantwice the thickness of the thinner part or 3/4 in.,whichever is less, correction may be made by “buttering”the surface of the short member prior to the membersbeing joined by welding. If welding is performed ononly one member at a time, the weld beads are free toshrink on the surface without significantly adding toresidual stresses or distortion (D1.1-96, paragraph5.22.4.3). Greater corrections than these require theapproval of the engineer (D1.1-96, paragraph 5.22.4.4).

For joints receiving fillet welds, the parts are to bebrought as closely into contact as practicable. If the rootopening is greater than 1/16 in., the leg of the fillet weldis required to be increased by the amount of the rootopening. The root opening is not allowed to exceed 3/16in. except in cases involving shapes or plates greater than3 inches in thickness, in which case a maximum allow-able gap of 5/16 in. is permitted, providing backing isused in the joint. Greater dimensions require build-up ofthe surfaces to reduce the gap (D1.1-96, paragraph5.22.1).

6.2 Backing and Weld Tabs

Steel backingthat will become part of the final structure(i.e., left in place), is required to be continuous for thelength of the joint (D1.1-96, paragraph 5.10.2).Thorough fusion between the weld and the backing isrequired (D1.1-96, paragraph 5.10.1). Suggested back-ing thicknesses to prevent burn through are shown inD1.1-96, paragraph 5.10.3. For statically loaded struc-tures, D1.1-96 allows the backing to be left in place, andattaching welds need not be full length, unless otherwisespecified by the engineer (paragraph 5.10.5). For cycli-cally loaded structures, steel backing in welds that aretransverse to the direction of computed stress arerequired to be removed, while backing on welds that areparallel to the direction of stress is not required to beremoved (D1.1-96, paragraph 5.10.4). For structuresdesigned to resist seismic loading, post-Northridge spec-ifications have often called for the removal of backingfrom the bottom beam-to-column connection in SpecialMoment Resisting Frames (SMRF). (AWS PositionStatement, p. 5; SAC References p. 6-5, 18, 9-7, 11-6)Contract documents should be carefully reviewed toensure conformance with any special requirements

24

regarding backing removal from other locations of thestructure such as the top flange connection, continuityplates, etc.

It is generally advisable to tack weld the backing to thejoint inside the joint, where the tack welds can be incor-porated into the final weld, although this is not a coderequirement. D1.1-96, paragraph 5.18.2.2, requires theremoval of any tack welds that are not incorporated intothe final weld, except for statically loaded structures,unless required by the engineer. Definitive requirementsfor structures subject to seismic loading have not beenestablished at this point. The code does require, howev-er, that tack welds meet the same quality requirements asfinal welds with two exceptions: First, preheat is notrequired for single pass tack welds which are remeltedand incorporated into continuous subsequent submergedarc welds, and secondly, discontinuities such as undercut,unfilled craters, and porosity need not be removed beforethe final submerged arc welding. Notice that both ofthese deviations are permitted for only applicationsinvolving subsequent welding with submerged arc,which is not the common process for many structuralapplications (D1.1-96, paragraph 5.18.2).

Because of the newer requirements that specify steelbacking removal for certain connections in seismic appli-cations, there is renewed interest in ceramic backing.This is certainly an option worthy of investigation,although the historic limitations of steel backing stillapply. It is essential that the metallurgical compatibilityof the ceramic backing and the filler metal being used beestablished. The self shielded flux cored products are notcompatible with all ceramic backing systems. Evenwhen ceramic backing is applied, it is still advisable toback gouge or grind the root of the joint from the backside in order to ensure complete fusion has been gainedto the root, and then follow this operation with the appli-cation of a reinforcing fillet weld. The D1.1-96 codedoes not allow for the use of ceramic backing for pre-qualified WPSs, so such approaches are subject to quali-fication testing (paragraph 5.10, Fig. 3.4).

Welds are required to be terminated in a manner that willensure sound welds for the full length of the joint. Tofacilitate this,weld tabsmay be necessary. They are tobe aligned in a manner which will provide for an exten-sion of the joint preparation, i.e., a continuation of thebasic joint geometry (D1.1-96, paragraph 3.31.1). Theuse of plates that are perpendicular to the axis of theweld, commonly known as “end dams” do not constitute

weld tabs (AWS Position Statement p. 5). For staticallyloaded structures, weld tabs could be left in place whilefor cyclically loaded structures, they are removed (D1.1-96, paragraphs 5.31.2, 5.31.3). In structures designed toresist seismic loads, weld tabs should be removed fromcritical beam-to-column connections (AWS PositionStatement p. 6; SAC References p. 6-4, 18, 20, 9-7).

The code does not define the length of weld tabs, but theymust be sufficiently long to ensure that the weld is fullsize for the length of the joint. As a rule of thumb, weldtabs should be at least as long as the thickness of thegroove weld. It is advisable that the backing (when used)extend beyond the width of the joint by at least the lengthof the weld tabs.

Acceptable steelsfor weld tabs and backing are definedin paragraph 5.2.2 of the D1.1-96 code. Unacceptablematerials would include rebar (sometimes improperlyused for weld backing), and the twist-off ends of bolts(sometimes improperly used for weld tabs/end dams).

6.3 Weld Access Holes

Weld access holes serve two important functions in weld-ed structures. First, as their name implies, they have thepractical function of providing access to the weld joint.Secondly, the weld access hole prevents the interaction ofthe residual stress fields of the flange and web. As shownin Figure 6-1, the longitudinal shrinkage of the web andflange weld, as well as the transverse shrinkage of theflange weld would induce 3-dimensional triaxial residualstresses at some point if it were not for the weld accesshole. By reducing triaxial stresses, cracking tendenciescan be minimized.

Figure 6-1 Minimization of triaxial stress at the weldaccess hole

25

It is important that weld access holes be properly sizedand of appropriate quality. Minimum dimensions forweld access holes are prescribed in D1.1-96. The mini-mum height of the access hole is not to be less than thethickness of the material in which the hole is made, butmust be adequate in size for the deposition of sound weldmetal. The minimum length of the access hole from thetoe of the weld preparation to the end of the radius is 1-1/2 times the thickness of the material in which it is cut.Examples are provided in the code in Figure 5.2.

For heavy Group 4 and 5 rolled shapes, and for built-upsections where the web material thickness is greater than1-1/2 in., the thermally cut surfaces of weld access holesare to be ground to bright metal and inspected by mag-netic particle or dye-penetrant methods (D1.1-96, para-graph 5.17.2). In addition, AISC LRFD M2.2, J1.6would require that the steel be positively preheated to150°F prior to the application of the thermal cuttingoperations. Gouges and nicks in weld access holes canact as stress concentrations and, particularly when sub-ject to seismic loading, can be the point of fracture initi-ation. Drilling the radius of weld access holes,particularly on heavy members, can assist in the devel-opment of high quality surfaces, while eliminating thegrinding and subsequent nondestructive testing require-ments of the code. It also proves to be a highly cost-effective method, particularly for fabricators withautomated drill lines.

6.4 Cutting and Gouging

Back gouging is required for all Complete JointPenetration (CJP) groove welds made with a prequalifiedWPS, unless steel backing is used. The resultant cavitycreated by the back gouging operation is expected to bein substantial conformance with the groove profiles spec-ified for prequalified groove details (D1.1-96, paragraph5.22.5).

Irregularities in the surface of thermally-cut edgesmaybe an indication of inclusions in the steel. These irregu-larities deserve investigation to preclude the possibilityof inducing weld defects that will be determined onlyafter weld completion. Specific acceptance criteria forcut surfaces are contained in paragraph 5.15 of the D1.1-96 code.

6.5 Joint and Weld Cleaning

The surfaces on which weld metal is to be deposited mustbe conducive to good fusion. Excessive scale, rust andsurface contaminants that would inhibit good fusionmust be removed. The code provides a practical test todetermine excessive levels of scale: mill scale that canwithstand rigorous wire brushing may remain (D1.1-96,paragraph 5.15).

The code requires that the slag be removed from all weldpasses before subsequent welding. The weld and theadjacent metal are required to be brushed clean. This notonly applies to subsequent layers, but also to the craterarea of any weld that is interrupted (D1.1-96, paragraph5.30.1).

Completed welds are required to have the slag removed,and the weld and surrounding base metal cleaned bybrushing or other suitable means (D1.1-96, paragraph5.30.2).

6.6 Preheat and Interpass Temperature

Preheating the steel to be welded will slow the rate atwhich the heat affected zone and weld metal cool. Thismay be necessary to avoid cracking of the weld metal orheat affected zone. The need for preheat increases as thesteel becomes thicker, the weldment is more highlyrestrained, the carbon and/or alloy content of the steelincreases, or the diffusible hydrogen level of the weldmetal increases.

The code requires that the preheat used in actual weldingbe in accordance with the WPS requirements. Preheat isto be measured a minimum of 3 inches away from thejoint or through the thickness of the part, whichever isgreater (in all directions including the thickness of thepart). These are minimum levels of preheat, and thesevalues can be exceeded (D1.1-96, paragraph 5.6). Thecode does not dictate how preheat is applied, and in mostcases, the contractor will use fuel gases as a source ofthermal energy, although resistant strip heaters can beused. It is important during the application of preheat thatthe steel not be heated too rapidly so as to cause hot spots,local distress that would be induced by shrinkage, or inthe extreme case, localized melting of the steel surface.

26

Although preheat can be measured in a variety of ways,the most accurate and practical method is to employ tem-perature indicating crayons. For some thicknesses andgrades of steel, and for select electrodes, there is no needto preheat. This is because, for the specific conditionsencountered, the hardenability of the steel is sufficientlylow that there is no expected danger of weld cracking.The secondary benefit of preheating, however, is that itdries the joint of moisture that may be on the surface ofthe steel. The code requires the surfaces to be free ofmoisture (D1.1-96, paragraph 5.15). Welding is not to beperformed when the surfaces are wet or exposed to rain,or snow, or when welding personnel are exposed toinclement conditions (D1.1-96, paragraph 5.12.2). Ifmoisture from condensation, for example, is on the sur-face of the steel, even though preheat is not required fortraditional reasons, it is necessary to dry the steel toensure quality welding. This could be done with a gastorch, but condensation of water from the flame’s prod-ucts of combustion can actually result in more surfacemoisture. The use of heated air guns or infrared heaterscan eliminate this.

When making beam-to-column connections wherejumbo sections (Group 4 and 5 shapes) are used for thecolumn member, it is advisable to increase the preheatlevel beyond the minimum prequalified level to thatrequired by AISC for making butt splices in jumbo sec-tions, namely 350°F (AISC LRFD J2.8). This conserva-tive recommendation is made acknowledging that theminimum preheat requirements prescribed by the D1.1-96 code for prequalified WPSs may not be adequate forthese highly restrained connections.

Interpass temperature refers to the temperature of thesteel just prior to the deposition of an additional weldpass. Its effect is identical to preheat temperature, exceptthat preheating is performed prior to any welding. Aswelding is performed, the temperature of the steel willnaturally increase momentarily in the vicinity of the welddue to the introduction of thermal energy from the arc.Immediately thereafter, the surrounding steel begins todraw this heat away, reducing the temperature. If anoth-er weld pass is made fairly quickly, the temperature ofthe steel may be higher than that of the minimum preheattemperature. If a longer period of time exists betweenweld passes, the steel may have cooled to a temperaturebelow the minimum interpass temperature whichshould be the same as the minimum preheat temperature.Many factors influence this condition in addition to time,including the thickness of the steel, and the amount of

energy added by the welding process. For weldmentswhere the joint has a smaller cross-sectional area, there isa natural tendency for the interpass temperature toincrease beyond that of the preheat level. On joints witha large cross-sectional area, just the opposite occurs. Forbutt splices in flanges, as well as beam flange-to-columnconnections, it is natural for the interpass temperature toincrease to higher levels. Care must be taken not toexceed specified interpass temperatures because exces-sively high interpass temperatures can result in a deteri-oration of weld metal and HAZ properties.

The D1.1-96 code does not impose specific maximuminterpass temperatures on welding operations, either as ageneral fabrication requirement or a prequalified con-straint. The AWS Position Statement (p. 7), however,has noted this as an area in which the code could beimproved, and has suggested that, for joints where notchtoughness is required, the maximum interpass tempera-ture for prequalified WPSs be limited to 550°F. Higherinterpass temperatures generally result in decreasedtoughness and lower strength of the weld metal. TheAWS Position Statement does not preclude the qualifica-tion of a higher maximum and interpass temperature bytest. Monitoring a maximum interpass temperature ismuch like maintaining minimum preheat and interpasstemperatures, but obviously a second, high-temperaturecrayon is required. The typical means by which this iscontrolled will be simply to not weld on that particularjoint until it has cooled to an acceptable temperature.This may necessitate movement to another weld joint.

6.7 Welding Techniques

Regardless of whether a welding procedure is prequali-fied or is qualified by test, it is essential that these valuesbe maintained for production welding. It is the inspec-tor’s responsibility to make certain that the actual weld-ing parameters used conform to the WPS (D1.1-96,paragraph 6.5.2). The code also requires the inspector toperiodically observe the welding techniques and the per-formance of each welder (D1.1-96, paragraph 6.5.4).This is an opportune time to ensure that the actual para-meters being used are appropriate. Under Section 5.11,the D1.1-96 code requires that the welding and cuttingequipment be designed so as to enable designated per-sonnel to follow the procedures and obtain the resultsrequired by the code. This requires that the equipment bein appropriate working condition, and implies that prop-er metering devices will be available to demonstrate thatthe correct welding conditions are being maintained.

27

These meters could be built into the equipment, or alter-nately could consist of hand held metering devices.

Arc welding processes are inherently dynamic. As afunction of time, amperage and voltage are not absolute-ly constant. A nominal, or average value, is what thecode intends to have maintained. A sensitive ammeter,for example, may register swings of several hundredamps from the nominal value. It is not possible, or evendesirable, for the electrical values to be absolutely con-stant. For control of welding parameters, only the aver-age or nominal value is important. Voltage is measuredbetween two points in an electrical circuit. It is impor-tant that voltage be read as near the arc as possible.Amperage is the same through an electrical circuit, andcan be read at any convenient point.

The most accurate method to control welding parametersfor wire feed processes is to use wire feed speeds. In aconstant voltage circuit, the amperage is not set by theoperator. The power source will deliver the requiredamperage to maintain the voltage selected by the opera-tor. The required amperage is a function of the wire feedspeed, electrode diameter, electrical stickout and elec-trode polarity. When wire feed speeds are used as need-ed as the primary variable of control, the resultantamperage can be used to verify that all the other variablesare correct. However, when amperage is used as the pri-mary controlling variable, an improper electrical stickoutor polarity can go undetected.

Travel speed is one of the more difficult elements to con-trol, particularly for manual and semiautomatic welding.The most direct and practical way to ensure that travelspeeds are appropriate is to monitor the relative size ofthe weld passes being deposited. Since the weld size isinversely proportional to the travel speed, acceptablyaccurate control of travel speed can be maintained in thismanner.

The root pass is generally the most critical pass in aweld, and is arguably the most difficult to make.Adequate access to the root must be achieved by properfit-up of properly prepared joints. As with all welding, itis essential that the operator keep the electrode on theleading edge of the weld pool to ensure that the intenseenergy of the arc is available to melt the base metal, thusassuring adequate fusion. When the weld pool rollsahead of the arc, the arc energy is concentrated on the liq-uid metal of the weld pool instead of the base metal.Lack of fusion and slag inclusions may result.

Excessively slow travel speeds encourage this type ofbehavior, and welds made with slow travel speeds haveinherently larger beads. For this reason, the AWSPosition Statement (p. 6) has recommended a maximumroot pass thickness of 1/4 in. to encourage root fusion.For all processes except SMAW, the code mandates amaximum 1/4 in. layer thickness for all intermediateweld layers (that is, except the root and the cap pass), butthis is more easily achieved because the width of the jointfor subsequent layers is larger than in the root pass(D1.1- 96, Table 3.7). For SMAW, the maximum layerthickness is restricted to 3/16 in.

The bottom beam flange-to-column connectionforbeam-to-column assemblies is one of the most difficultwelds to make because of the geometric constraintsimposed by the presence of the beam web. Weld beadsare, of necessity, interrupted in the middle of the lengthof the weld. Welding must be performed through theweld access hole, adding additional difficulty to the oper-ation. One of two sequence procedures can be employed,as follows:

The following techniques are recommended (adoptedfrom FEMA 267, p. 6-20):

1. The root pass thickness should not exceed 1/4 in.

2. The first half-length root pass should be made withone of the following techniques, at the option of thecontractor:

a) The root pass may be initiated near the center of thejoint. If this approach is used, the welder shouldextend the electrode through the weld access hole,approximately 1 in. beyond the opposite side of thegirder web. This is to allow adequate access for clean-ing and inspection of the initiation point of the weldbefore the second half-length of the root pass isapplied. It is not desirable to initiate the arc in theexact center of the girder width, since this will limitaccess to the start of the weld during post-weld opera-tions. After the arc is initiated, travel should progresstowards the end of the joint (that is, toward the beamflange edge), and the weld should be terminated on aweld tab.

b) The weld may be initiated on the weld tab, with travelprogressing toward the center of the girder flangewidth. When this approach is used, the welder shouldstop the weld approximately 1 in. before the beam

28

web. It is not advisable to leave the weld crater in thecenter of the beam flange width since this will hinderpost-weld operations.

3. The half-length root pass should be thoroughlydeslagged and cleaned.

4. The end of the half-length root pass that is near thecenter of the beam flange should be visually inspectedto ensure fusion, soundness, freedom from slag inclu-sions and excessive porosity. The resulting bead pro-file should be suitable for obtaining fusion by thesubsequent pass to be initiated. If the profile is notconducive to good fusion, the start of the first root passshould be ground, gouged, chipped, or otherwise pre-pared to ensure adequate fusion.

5. The second half of the weld joint should have the rootpass applied before any other weld passes are per-formed. The arc should be initiated at the end of thehalf-length root pass that is near the center of the beamflange, and travel should progress to the outer end ofthe joint, terminating on the weld tab.

6. Each weld layer should be completed on both sides ofthe joint before a new layer is deposited.

Regardless of the approach used, it is important that thestart-and-stop point be carefully determined to providefor adequate cleaning of the previously made half-lengthweld pass, and to facilitate adequate tie-in of the remain-ing half-length weld pass.

It is recommended that no additional layer of weld metalbe deposited before the previous layer is completed forthe full length of the weld joint (AWS PositionStatement, p. 8). Notice this does not dictate that theindividual beads be completed full length, but rather thatthe layers be completed full length. The layer-comple-tion approach has been suggested to prohibit the tech-nique that would allow for the completion of an entirehalf-length of the weld prior to welding on the oppositeside of the web. Completing an entire half-length firstwould commonly result in fusion problems in the centerof the length of the weld.

In Annex B of D1.1-96 code, stringer beads are definedas a type of weld bead made without appreciable weav-ing motion. Weave beads are defined as a type of weldbead made with transverse oscillation. The D1.1 codedoes not mandate nor restrict stringer passes or weaving.

Final bead widths are not directly prescribed, althoughthe maximum root width is controlled. For prequalifiedWPSs, the maximum full width weld beads are restrictedto joints with a root face of 5/8 in. for FCAW performedin other than the vertical position. For vertical welding,this dimension is increased to 1 in. (D1.1-96, Figure 3.7,footnote 5). It would be possible to increase the maxi-mum width of a weld bead by qualification testing.

A balance must be established between many smallstringer passes, and a few larger weave passes. As thenumber of passes increases, the amount of distortion andthe corresponding residual stresses increase. In mostcases, however, increased notch toughness is obtained inthe weld beads when smaller stringer passes areemployed. Either extreme can cause problems. Anexcessive number of small stringer passes will result inextremely high cooling rates of the weld beads, and acorresponding increase in yield and tensile strength, adecrease in elongation, and perhaps a decrease in notchtoughness in the weld deposit. Excessively large weldbeads result in a decrease in strength, and a decrease innotch toughness, although the measured elongation mayactually increase. Excessively large weld beads areinvariably associated with high welding heat input whichmay have negative effects on the heat affected zone.

Bead size and deposition rate are not directly relatedalthough bead size (cross section) and heat input are. Itis possible to maintain smaller bead sizes with highdeposition rate procedures, providing the travel speed isappropriately adjusted (i.e., increased). For most groovewelding procedures, weld passes that deposit the equiva-lent volume of weld metal as would be required for 1/4in. to 3/8 in. fillet welds is optimum. This equates to aheat input of approximately 30 to 70 kJ/in. By maintain-ing the prequalified bead sizes specified in the code(width and thickness), these conditions are generallyachieved.

Aside from the cited effect on mechanical properties, thegreatest concern associated with excessively wide weldpasses is the probability of inducing fusion-related prob-lems and slag inclusions.

It is inappropriate to fully restrict oscillation of the elec-trode. Skilled welders are aware of the approximatewidth that the weld bead will assume with a straight, lin-ear progression. When this bead width is inadequate tocompletely fuse between the sides of groove weld, orbetween a previous pass in the side of the groove, a slight

29

oscillation of the electrode not only should be permitted,but is desirable. The AWS D1 Position Statement (p. 6)permits such oscillations, suggests that the oscillationshould be within certain limitations, and advises inspec-tors to monitor final bead widths, not oscillation dimen-sions.

The D1.1-96 code neither requires nor prohibits peening,except for root and cap passes, where peening is prohib-ited (paragraph 5.27). Peening is not justified for routinefabrication (AWS Position Statement, p. 8). For highlyrestrained members, and particularly for critical repairs,peening of intermediate weld beads may be helpful inreducing the level of residual stress that is imposed bythe shrinkage of the weld beads.

Whenever possible, welders should initiate the arc in theweld joint or on weld tabs, eliminating the possibility ofarc strikes outside of the weld joint. The D1.1-96 codedoes not expressly prohibit arc strikes, but advises thatthey should be avoided. If they occur, they are to beground to a smooth contour and checked to ensuresoundness (D1.1-96, paragraph 5.29). It is important thatwelding cables be properly insulated to ensure that inad-vertent arc strikes are not created in other areas of thestructure.

6.8 Special Welding Conditions

Weld repairs — Welds that do not meet the acceptancecriteria can be repaired. Repairs fit into three categories:(1) removal of excess metal; (2) deposition of additionalmetal; and, (3) removal and reapplication of metal.Overlap, excessive convexity, or excessive reinforcementcan be addressed by removal of excess metal (D1.1-96,paragraph 5.26.1.1). Excessive concavity, undersizedwelds, and undercut can be repaired by the application ofadditional metal (D1.1-96, paragraph 5.26.1.2). Weldmetal porosity, incomplete fusion, slag inclusions, andcracks must be repaired by removing the deficient weldmetal, and applying sound metal. In the case of cracks,not only must the crack be removed, but sound metal 2in. beyond the end of the crack must also be removed.The extent of cracking is required to be determined bydye penetrant or magnetic particle inspection, or othermeans (D1.1-96, paragraph 5.26.1.4). The ends of deepcavities should be cascaded to facilitate quality welds atthe ends of the cavity.

Protection for gas shielded processes— For GMAW,FCAW-g, and other gas shielded processes such asGTAW and EGW, it is important that the gas shield notbe disturbed by air movement. Tenting of one type oranother is generally supplied for field welding, and thecode requires that the velocity of any wind be reduced toa maximum of 5 miles per hour (D1.1-96, paragraph5.12.1). It is important that tenting be of materials thatare not flammable, or that welding take place in a loca-tion where it does not constitute a fire hazard. Highlyeffective tenting may not provide adequate ventilation forthe welder, another consideration. For these reasons,SMAW and FCAW-ss are the preferred processes formost field welding conditions.

Cold temperature welding — Completely aside frompreheat requirements, welding should not be done whenwelding personnel are exposed to conditions underwhich quality will suffer. The code prescribes the lowerbound for “ambient” temperature to be 0°F. The word“ambient” is emphasized because it is possible to changeambient conditions by tenting, or other enclosures. Theoutside environmental temperature could drop below0°F, but if within the shelter the temperature is above thislower limit, welding may continue (D1.1-96, paragraph5.12.2).

When welding under cold conditions, it is also importantto reexamine preheat requirements. For prequalifiedWPSs, the D1.1 code mandates that, even for situationswhere no preheat application is required, the temperatureof the steel be raised to a minimum of 70°F when theambient temperature drops below 32°F (D1.1-96, seeFootnote 1 to Table 3.2). Although this is restricted to pre-qualified conditions, it is a good practice for all welding.If extremely low temperature conditions are anticipatedand no preheat is going to be applied, the WPS should bequalified by test on plate held at the cold temperature.

6.9 Weld Metal Mechanical Properties

The first important point in any discussion aboutmechanical properties is to recognize that a weld, likeany other material, is not uniform. When a weld bead isdeposited, three distinct zones are created — the fusionzone, which is the product of melting welding electrodewith some base material, the heat affected zone (HAZ) inthe base material which was heated to a temperature high

30

enough to alter the microstructure and properties withoutmelting, and the zone comprised of the surrounding unaf-fected base material (see Figure 6-2).

Figure 6-2 Three zones created by welding

This is a somewhat simplified example, but it will illus-trate the point. It is reasonable to expect the mechanicalproperties and the chemical composition to vary acrossthese zones. Chemical composition in the weld beadmay be different than that of the electrode compositiondue to the mixing with the base material. Themicrostructure variation from the HAZ to the weld metalis always significant, even if the base metal and fillermetal chemical compositions are very similar.

Now, consider what happens in a multiple pass weldwhere several overlapping weld beads are deposited.Each successively deposited bead reheats and refines anarrow zone around it, creating a series of overlappingHAZs both in the base plate and in the weld metal (seeFigure 6-3).

Figure 6-3 Grain refinement in multiple pass weld

The areas of weld metal that are not affected by the heat-ing of subsequent passes are usually referred to as pri-mary weld metal. Consequently, many small zones areclosely created in close proximity, which will not neces-sarily exhibit consistent chemical composition, metallur-gical structure or mechanical properties.

Additionally, regions of the weld metal close to the basematerial may be somewhat different in chemical compo-sition than regions in the middle of the weld near the sur-face, simply because of dilution. The differences will

depend on how closely the compositions of the weldelectrode and base material match. Further, the weldingprocedures used can significantly change the number andarrangement of these various zones in the weld. If anelectrode diameter and/or welding parameters wereselected that produced smaller and flatter weld beads, theamount of primary weld metal relative to reheated weldmetal would be significantly different (see Figure 6-4).Rarely, if ever, are the mechanical properties of therefined HAZs the same as those of the primary weldmetal, even when the chemical composition does notvary significantly. It is, therefore, important to under-stand the purpose of any test and what is actually beingmeasured, since the results can vary according to the testspecimen location and fabrication procedures.

AWS classification testing —Classification of weldingconsumables for welding steels often requires mechani-cal testing, which may include tests for mechanical prop-erties such as ultimate tensile strength, yield strength andCharpy V-notch impact toughness. It is important to notethat not all of these tests are required in every case. Forexample, AWS A5.20-95 specifies the classificationrequirements for approximately 37 types of FCAW elec-trodes. All of them have an ultimate tensile strengthrequirement, but less than half of them have yieldstrength and/or Charpy V-notch requirements.

In the context of the preceding discussion about inherentvariability, the classification test represents a single weld-ing condition for a consumable. As such, the mechanicalproperties determined for a welding consumable duringclassification might not be representative of the propertiesachievable under actual production welding conditions.Consequently, fabricators are cautioned that the AWSclassification test results which form the basis for mostpublished literature on welding consumables should beused as a relative guide for consumable selection and nota guarantee that the specified minimum properties will beachieved under actual production welding conditions.

Figure 6-4 Grain refinement for flat weld beads

31

Such differences between classification tests and actualproduction welding conditions, represented in a proce-dure qualification test, for example, arise because themechanical properties achieved in any weld depend asmuch on the welding and testing procedures as on thewelding consumables selected. AWS cautions that thefollowing can influence weld metal mechanical proper-ties: electrode size, current type and polarity, voltage,type and amount of shielding gas, welding position, elec-trode extension, plate thickness, joint geometry, beadplacement, preheat and interpass temperatures, travelspeed, base material composition and extent of dilution.It is almost impossible to change one variable withoutalso affecting several others.

Strength — For classification, weld metal strength istypically measured by means of a 0.500 in. diameterspecimen located in the weld cross-section as illustratedin Figure 6-5.

Figure 6-5 Typical tensile specimen location

Strength measured in this way is influenced to a largeextent by both the chemical composition of the consum-able and the welding procedure. A standard base mater-ial type is used and the root opening is wide enough tominimize the effects of dilution at the weld center.Further, the specimen is large enough that it samples bothprimary and reheated weld metals, making variationsbetween these two zones relatively insignificant.Strength usually fluctuates directly with alloy level.Certain procedure variables can actually influence howmuch of the alloy in the electrode is transferred to theweld deposit. Some alloy elements (e.g., Mn) are easilyoxidized in the arc. Higher arc voltages (i.e., longer arclength) tend to result in lower weld metal alloy levels.With a longer arc length there is more time for oxidationto occur across the arc. This corresponds to a greater lossof alloying elements and lower strength. Conversely,with shorter arc lengths there is less time available foroxidation in the arc atmosphere, resulting in a potential-

ly higher alloy level and higher strength. Therefore,selection of welding voltage and electrode extension willcontrol the arc length and influence the alloy level andstrength.

Oxidation which occurs as a result of slag/metal reac-tions also influences alloy levels. A higher heat inputusually results in slower cooling rates. Slower coolingrates lead to longer reaction times and possibly greaterloss of alloy due to slag/metal reactions. Conversely, alower heat input, resulting in a faster cooling rate, maypromote higher alloy content.

The cooling rate also influences the weld metalmicrostructure, which influences strength. Faster cool-ing rates can promote harder structures, which elevatestrength measurements. Slower cooling rates promotesofter structures and also contribute to grain coarseningin the reheated regions, both factors that can lowerstrength measurements. The effect of cooling rate isillustrated graphically in Figure 6-6. The weld metalstrength may also differ from an AWS test plate due tothe joint geometry.

Figure 6-6 Strength related to cooling rate

In a very narrow groove weld, dilution effects from thebase material become more of a factor. Also, when thejoint volume becomes so small that a smaller diametertensile specimen is necessary, variations in measurementcan occur simply as a result of specimen size.

Tests that vary from the standard AWS classification con-dition provide useful information about the mechanicalperformance of the weld metal. However, it is importantto remember what is being tested and for what purpose.

32

Charpy V-notch impact toughness —Charpy V-notch(CVN) impact test results also vary significantly withcomposition and welding procedure variables, and aredependent on test temperature. A schematic illustrationof CVN response with temperature is illustrated in Figure6-7. At lower test temperatures, a lower shelf existswhere the fractures are brittle in appearance and furtherreductions in test temperature achieve no further substan-tial reduction in energy. By contrast, at higher test tem-peratures, an upper shelf is observed where the fracturesare ductile in appearance and further increases in test

temperature achieve no substantial improvement in CVNenergy. Between these two shelves is the transition regionwhere the fractures have some ductile and some brittlecharacter. All steel weld metals have these characteris-tics, although the specific shape and position of the curverelative to temperature depend on the alloy type, fabrica-tion details and the location of the CVN notch in the weld.Variations in chemical composition and welding proce-dure that influence weld thermal cycles can alter theupper and lower shelf energies as well as the temperaturerange over which the transition occurs.

For this reason, AWS classification tests are conductedwith the CVN located at the mid-thickness on the weldcenterline, as shown in Figure 6-8.

These tests are conducted at a single test temperaturewith the results evaluated based on a minimum energyrequired. In most cases, where CVN specimens arerequired for classification, the test temperature is in thetransition region where test scatter is typically the high-est and variations in composition and welding processvariables can have a large effect on the impact energymeasurements. The specimen is oriented in the grooveweld so that variation due to location is minimized. Notethat the notch is located on the weld centerline in thebead overlap region where the greatest amount of refine-ment takes place. While this provides for greater testingconsistency, it also samples the region of the weld withpotentially the highest impact toughness.

The finer grained material in the reheated regions usual-ly has higher CVN values than columnar structuresencountered in the primary weld metal. This is not truein every case. There are a few welding consumablesspecifically designed to achieve high levels of toughnessin the primary weld metal. However, AWS classificationtests, which form the basis for most product literature,will not usually make this distinction. Rather, the moreusual differences in performance between as-welded andreheated weld metals in the same weld joint can give riseto large differences in apparent CVN performance. Forexample, a qualification test that samples CVN from theroot region of a double V-groove is not likely to achievethe same level of CVN energy reported for AWS classifi-cation tests. Such tests are intended to simulate produc-tion welding conditions, not differences from AWSclassification testing conditions.

CVN performance can be influenced by chemical com-position variations in ways that can either increase ordecrease CVN energy at a given test temperature. Theresulting impact on CVN energy will depend on how thecomposition variation influences the weld microstruc-ture. If grain refinement can be achieved without otherundesirable changes in the structure, CVN energy shouldincrease. However, in most cases, a shift in compositionaway from the optimum design for the welding consum-able has the effect of reducing CVN energy to somedegree.

Welding process variables can also influence the CVN testresults. For example, higher welding heat inputs and inter-pass temperatures have the effect of slowing the cooling rateas seen in the discussion on strength. This allows moretime at elevated temperature, which widens the bands ofreheated metallurgical structure, but also promotes a coarser

Figure 6-8 Typical CVN specimen location

Figure 6-7 Schematic CVN transition curve

33

grain structure in these regions. Consequently, it is impossi-ble to determine which variable(s) will govern performancewithout specific knowledge of the welding consumable andalloy system. Chemical composition variations that resultfrom dilution can be minimized if welding conditions whichminimize penetration are selected.

Variation is an inherent part of a weld, but the extent ofvariation is influenced both by the materials selected for aparticular weldment and the procedure variables chosenfor the fabrication. Both affect the mechanical test resultsobtained with a given welding consumable. Wheneverthe mechanical properties of a weld are determined bytesting, it is important to recognize that such variationsexist and can influence the results. Consequently, litera-ture based entirely on AWS classification testing shouldbe used only as a relative guide to weld metal perfor-mance. Fabricators and designers with more stringentperformance requirements should always consult theelectrode manufacturer for additional guidance.

6.10 Intermixing of Weld Deposits

For most arc welding processes, the molten metal is pro-tected from the atmosphere by a combination of shield-ing and/or a slag system. In this respect, FCAW-ss isunique. FCAW-ss consumables produce very littleshielding gas, relying heavily on slag systems. Ratherthan protecting the molten metal from atmospheric cont-amination, the slag relies on the addition of largeamounts of deoxidizers to react with oxygen and nitro-gen. While aluminum is the primary deoxidizer, smallerquantities of titanium and zirconium may also be present.Consequently, aluminum levels in excess of 1% byweight are common, significantly higher than the alu-minum contents typically found in steel base materials.Low levels of aluminum in both base metal and weldmetal have been known to cause a reduction in tough-ness, so the higher levels of aluminum found in the welddeposits of FCAW-ss have generated curiosity. Becauseof its function as a primary deoxidizer/denitrider, alu-minum is essential to the metallurgy of the typicalFCAW-ss weld deposit. For example, significant quanti-ties of nitrogen may be contained in the weld metal ofFCAW-ss, but this nitrogen is in the compound of alu-minum nitride (AlN). While the nitrogen content may be500 ppm, the available 10,000 ppm level of aluminumensures that there is always excess aluminum to ensureformation of AlN (which requires an Al:N ratio ofapproximately 2:1). The remaining aluminum thereforeacts as an alloying agent in the weld metal.

The balance between aluminum and nitrogen, as well ascarbon and other alloying elements, must be properlymaintained to ensure the specified mechanical propertiesare obtained from the weld deposit. As a result of theseunique characteristics, many FCAW-ss weld deposits cancontain substantially higher carbon (up to 0.45 wt. pct.),lower manganese (as low as 0.5 wt. pct.), lower oxygen(as low as 30 ppm), and significantly higher nitrogen (upto 700 ppm) than would be found in weld metals pro-duced by other arc welding processes. When weldingconsumables which derive their properties from differentmetallurgical mechanisms and alloy balances are mixedin a single joint there is the potential for negative inter-action. For example, if a root pass is made with FCAW-ss, and the subsequent fill passes are made with SMAWor FCAW-g, the properties of the fill passes may be neg-atively affected in terms of ductility and notch toughness.

This phenomenon is, at least in part, the result of excessaluminum picked up through dilution from the underly-ing FCAW-ss weld metal. The SMAW or FCAW-g weldmetals, typically manganese-silicon deoxidized metallur-gical systems, are not designed to accommodate excessaluminum. The presence of even a small amount ofexcess aluminum alters the normal deoxidationsequence, which influences the formation of non-metal-lic inclusions and disrupts the normal alloy balance.Thus, the notch toughness of the manganese-silicondeoxidized weld metal may be significantly reduced bythe introduction of small levels of aluminum. The use ofmultiple weld processes and electrode or consumabletypes in a single weld joint may occur for several rea-sons. Tack welding might be performed with oneprocess/electrode, and the completion of the joint madewith another. The more demanding welding conditionsof root pass welding, and particularly open root joints(i.e., no backing), may dictate a different weldprocess/electrode for the root pass.

Multiple processes also may result when repair weldingis being performed with a different process/electrodethan the original method used to fabricate the joint.When intermixing of weld process/electrode typesoccurs in the same weld joint, the potential for negativeeffects must be investigated. The use of non-FCAW-ssweld process/electrodes on top of welds made by FCAW-ss is of particular concern. The resulting mechanicalproperties of subsequent welds will be dependent onmany variables, including: the original composition ofthe FCAW-ss deposit; the degree of admixture (related topenetration) that will occur in the subsequent welding;

34

and the actual level of mechanical properties delivered bythe subsequent weld process when they are unaffected byFCAW-ss interactions. For example, when deep-pene-trating SAW is used upon high aluminum contentFCAW-ss welds, notch toughness may decrease from amoderate level of 40 ft·lbf. at -20°F to less than 15 ft·lbfat the same temperature. In contrast, a shallow penetrat-ing SMAW weld deposit made with E7018 may find thenotch toughness decreases from 150 ft·lbf to 80 ft·lbf atthe same temperature, but the resultant notch toughnessmay be more than adequate for the application.

It has been believed in the past that this concern existedonly for non-FCAW-ss weld deposits on top of FCAW-ss. However, tests indicate that reductions in toughnessare possible with combinations of other welding process-es and electrode types. Two approaches can be takenwith respect to this issue. First, the same process can beused throughout, eliminating potential concerns.Secondly, the potential interaction of the two processescan be evaluated by testing. The latter approach is rec-ommended by the SAC Interim Guidelines. A test pro-gram was undertaken in order to provide guidanceregarding the effects of some admixture combinations onmechanical properties, specifically toughness, and toassess the probable mechanisms at work. Initial interestfocused on a case in which a SMAW, FCAW-g or SAWdeposit was made over FCAW-ss. The program wasexpanded to include deposition of FCAW-ss weld metalover SMAW, FCAW-g and SAW. The emphasis was con-sistently on combinations of welding consumables whichderive their properties from significantly different metal-lurgical mechanisms and alloy balances. The objectivein each case was to determine the magnitude of tough-ness reduction in the fill material resulting from dilutionfrom the root material.

To this end, the joint geometry and bead sequence wereselected to achieve a reasonably high level of dilution. Atypical joint is illustrated in Figure 6-9. Each combina-tion of fill and root materials was evaluated based onCharpy V-notch impact test specimens located at themaximum dilution location. That is, the bottom of thetest specimen was located as closely as possible to thefusion boundary between the fill and root materials, thusmaximizing the effect of dilution on the test. Specimenlocation is illustrated in Figure 6-9.

All notches were approximately centered between theside walls in the bead overlap regions to be as consistentas possible with standard AWS certification testing. Theextent of toughness reduction was determined by com-paring the results from each admixture weld with theresults from similar locations in test welds fabricatedentirely with the fill electrode. So, for example, SMAWdiluted with FCAW-ss is compared with the sameSMAW from the same general area in the groove withoutdilution from FCAW-ss.

It is important to note that actual field conditions may bemore or less severe than those in the tests representedhere. Any reduction in notch toughness will be highlydependent upon the extent of dilution from underlyingweld metal. Consequently, variations in welding proce-dure, joint geometry and configuration will influenceresults. Further, it should be noted that some reductionin notch toughness may be tolerable for an application,depending upon the specific requirements prevailing.Consequently, the user must be the ultimate judge as tothe applicability of specific electrode combinations for aparticular field application. While the recommendationsprovided below are based on conservative tests achievingrelatively high levels of dilution, the best test is onewhich simulates the specific application and weldingprocedures to be employed. Users are encouraged toconduct their own tests to verify that desired perfor-mance is obtained.

Table 6.1 summarizes the recommendations based on testresults for some combinations. Some recommendationsare based on results expected simply on the basis ofchemical composition differences between the consum-ables. However, most recommendations result from thecomparison of a single test weld for each intermixedcombination against a corresponding baseline test weldas previously described. The recommendations are

Figure 6-9 Locations of CVN test specimens

35

based on satisfying a 20 ft·lbf requirement at either -20°For 0°F. The combinations have been categorized basedon the performance of the fill passes only.

A No change in notch toughness is expected.

B Some decrease in notch toughness is anticipated.However, Charpy V-notch exceeding 20 ft·lbf at both-20°F and 0°F is expected.

C Some decrease in notch toughness is anticipated.However, Charpy V-notch results are expected toexceed 20 ft·lbf at 0°F but not necessarily at -20°F.

D Significant reductions in notch toughness are antici-pated. Charpy V-notch results exceeding 20 ft·lbf arenot expected consistently at either -20°F or 0°F.

E Significant reductions in notch toughness are expect-ed. Charpy V-notch results approaching 20 ft·lbf arenot expected at either -20°F or 0°F.

Table 6-1

7 Welding Techniques and Variables

7.1 SMAW

Amperage: Covered electrodes of a specific size andclassification will operate satisfactorily at various amper-ages within a certain range. This range will vary some-what with the thickness and formulation of the covering.Manufacturer’s guidelines should be followed becauseflux coverings are of different compositions.

Voltage:Voltage on a CC machine will be determined byarc length, which is determined by the operator. It is nota presettable parameter. Operators should hold short arclengths to minimize the generation of spatter, and to opti-mize mechanical properties.

Travel speed:Travel speed is determined by the rate atwhich the electrode travels along the joint. The propertravel speed is the one which produces a weld bead ofproper contour and appearance. Training and experiencewill help in determining proper appearance. Coderestrictions on bead sizes will directly affect the requiredtravel speeds for a given deposition rate.

Electrode diameter: The correct electrode diameter isone that, when used with the proper amperage and travelspeed, produces a weld of the required quality and size inthe least amount of time. Also, the correct electrode isdetermined by the thickness of the material to be welded.As a general rule, 1/8 and 5/32 in. diameter electrodes areused for out-of-position welding. For flat and horizontalwelding, typical electrode diameters are 3/16 and 7/32 in.Larger electrodes may be used, but suitability is highlydependent on the joint and weld type.

Drag angle: The drag angle is the angle between theelectrode centerline and the seam centerline in the direc-tion of travel.

Polarity: Use DC+ whenever possible if the electrodesize is 5/32 in. or less. For larger electrodes, use AC forbest operating characteristics (but DC can also be used).

Flat position: Use low current within the acceptablerange of the electrode on the first pass, or whenever it isdesirable to reduce admixture with a base metal of poorweldability. On succeeding passes, use currents that pro-vide the best operating characteristics. Drag the elec-trode lightly or hold an arc of 1/8 in. or less. Do not usea long arc at any time, since E7018 electrodes rely prin-cipally on molten slag for shielding. Stringer beads orsmall weave passes are preferable to wide weave passes.When starting a new electrode, strike the arc ahead of thecrater, move back into the crater, and then proceed in thenormal direction. On AC, use currents about 10% high-er than those used with DC. Govern travel speed by thedesired bead size.

Vertical: Weld vertical-up with electrode sizes of 5/32in. or less. For E7018, use a triangular weave for heavysingle pass welds. For multipass welds, first deposit astringer bead by using a slight weave. Deposit addi-tional layers with a side-to-side weave, hesitating at thesides long enough to fuse out any small slag pocketsand to minimize undercut. Do not use a whip techniqueor take the electrode out of the molten pool. Travelslowly enough to maintain the shelf without causing

36

metal to spill. Use currents in the lower portion of theacceptable range.

Overhead: Use electrodes of 5/32 in. or smaller. ForE7018, deposit stringer beads by using a slight circularmotion in the crater. Maintain a short arc. Motionsshould be slow and deliberate, but fast enough to avoidspilling weld metal; do not be alarmed if some slag spills.Use currents in the lower portion of the acceptable range.

7.2 FCAW-ss

Semiautomatic operating techniques:Before welding,control settings should be carefully checked. The controlsettings should be within the range specified by the pro-cedures, and adjusted according to past experience withthe specific joint. Drive rolls and wire guide tubesshould be correct for the wire size, and drive-roll pres-sure should be adjusted according to the manufacturer’sinstructions. The wire feeder and power source shouldbe set for constant-voltage output. The gun, cable, andnozzle contact tip should be correct for the wire size andfor the stickout.

Amperage: Manufacturer’s guidelines should be fol-lowed because flux core contents are of different compo-sitions.

Wire feed speed:The value selected should be withinthe manufacturer’s recommended range of operation, andappropriate for the specific joint and skill of the operator.

Voltage: The arc voltage will be specified by the elec-trode manufacturer. System voltages may be higher dueto voltage drops. Excessive arc voltage leads to porosi-ty. Voltage should be measured from the wire feeder tothe work.

Travel speed:Bead sizes are dependent on travel speeds.Excessively large or excessively small weld beads canreduce weld quality. Excessively slow travel speeds canlead to fusion problems and encourage slag entrapment.Very high travel speeds can result in reduced penetration,irregular wetting, lack of slag coverage and rough-appearing beads.

Electrode diameter: For out-of-position welding rec-ommendations, electrode diameters are typically 1/16 to5/64 in. For flat and horizontal welding, electrode diam-eters range from 5/64 to 0.120 in.

Electrical stickout: The distance between the point ofelectrical contact and the electrode tip is referred to as“stickout” or “electrical stickout.” The entire length ofelectrode — not just the visible portion protruding fromthe nozzle — is subject to resistance heating as the cur-rent passes through it. The longer the projection of theelectrode from the point of electrical contact, the greaterthe heat build-up within it. This heat can be used to goodadvantage to increase the melting rate and reduce pene-tration. Electrode extensions vary from 1/2 to 3-3/4 in.,so the electrode manufacturer’s recommendations shouldbe consulted. The selection of an electrode extensionmust also be appropriate for the joint being welded. Rootpasses, and joints requiring penetration, are best madewith shorter electrode extension dimensions.

Starting the arc: To start the arc, the electrode is“inched” out beyond the nozzle to the visible stickoutrecommended for the electrode size and guide tip. Thetip of the electrode is positioned just off, or lightly touch-ing, the joint, and the trigger is pressed to start the arc.The electrode should not be pushed into the joint as itmelts away, as in stick electrode welding, since themechanical feed will take care of advancing the elec-trode. Welding is stopped by releasing the trigger orquickly pulling the gun from the work. The instructionmanual for a specific wire feeder usually gives recom-mendations on setting feed speed and open-circuit volt-age to facilitate starting.

Accommodating poor fit-up: One of the advantages ofFCAW is the ability to handle poor fit-up. Pulling thegun away from the work to increase the visible stickoutreduces the current, and thus the penetration, and helps toavoid melt-through. After a poor fit-up area has been tra-versed, normal stickout should be used for the remainderof the joint.

With electrodes designed for out-of-position welding,poor fit-up can be handled by reducing the welding cur-rent to the minimum value specified in the procedures.Increasing the stickout to 1-1/2 in. also helps to reducepenetration and melt-through.

Removing slag:Slag removal is easy in most self shield-ed flux cored electrode welding. In heavy fast-fill work,the slag often curls up and peels off behind the weldinggun. Otherwise, a light scrape with a chipping hammeror wire brush is usually all that is needed to dislodge theslag. Slag is occasionally trapped on fillet welds made in

37

the vertical position, or on groove welds in a flat positionthat have a convex bead. Entrapment can be avoided byproper bead location and drag angle and by using asmooth, even travel speed to insure good bead shape.

Drag angle: The drag angle is the angle between theelectrode centerline and the seam centerline in the direc-tion of travel. The desired drag angle is approximatelythe same as in SMAW. If slag tends to run ahead of thearc, the drag angle should be decreased.

Joint angle: The joint angle is the angle between theelectrode centerline and the joint on a plane perpendicu-lar to the weld axis.

For the best bead shape on most 5/16 in. and larger hori-zontal fillets, the electrode should point at the bottomplate and the angle between the electrode and bottomplate should be less than 45°. With this arrangement, themolten metal washes up onto the vertical plate. Pointingthe electrode directly into the joint and using a 45 to 55°angle will decrease root-porosity problems, if they occur,but may produce spatter and a convex bead. For 1/4 in.and smaller fillets, the electrode should be pointed direct-ly into the joint and the electrode angle held at about 40°.

For vertical and overhead welding with E71T-8 typeelectrodes, similar techniques are used. Vertical up, mul-tiple pass welding should generally be with a split weavepass sequence, with each bead approximately 3/8 in.wide. Some E71T-8 type electrodes are capable of verti-cal down operation; for those products, a straight pro-gression technique is normally recommended. Verticaldown WPSs must be qualified by test for work done toD1.1-96.

7.3 FCAW-g

Generally the same operating techniques and precautionsas those recommended for self shielded flux cored arcwelding should be observed for gas shielded flux coredarc welding. The use of a shielding gas allows a widervariation in the metal transfer mode than is attainablewith the self shielded process.

Wire feed speed:The value selected should be withinthe manufacturer’s recommended range of operation, andappropriate for the specific joint or operator skill.

Amperage: Welding current is directly proportional toelectrode feed rate for a specific electrode diameter, com-position, and electrode extension. Manufacturer’s guide-lines should be followed because flux core contents areof different compositions.

Voltage: If the arc voltage is too high, the bead tends towiden in an irregular manner, with excessive spatter. Toolow an arc voltage results in a narrow, high bead withexcessive spatter and reduced penetration. Differenttypes of gases will affect voltage (higher or lower).

Travel speed:As with other mechanized processes, trav-el rate affects the build-up of molten metal and penetra-tion into the base material. Slow travel increasespenetration, but excessively slow travel can lead toexcessive build-up of molten metal, overheating of theweld area, and a rough-appearing bead. Too high travelrates may result in inadequate penetration and a ropy,irregular bead. Travel rates between 12 and 30 in. perminute usually give satisfactory results.

Electrical stickout: Varying the electrode stickout — aswith self shielded flux cored welding and submerged arcwelding — offers a method of controlling deposition rateand penetration. At a given rate of wire feed, a shorterstickout results in deeper penetration than a longer stick-out. With gas shielded flux cored electrodes, stickout of3/4 to 1-1/4 in. is usually recommended, depending onthe type of nozzle. If the stickout is excessive, spatteroccurs, and arc shielding is lost. The gas nozzle must beadjusted as stickout is changed to ensure adequate shield-ing at the arc.

Electrode diameter: Proper electrode diameter for agiven weld will produce a weld in the least amount oftime for a given thickness. For out-of-position welding,electrode diameters of 0.045 to 1/16 in. are typicallyused. For flat and horizontal welding, electrode diame-ters of 1/16 to 3/32 in. are typical.

Polarity: DC+ is always used for FCAW-g.

Drag angle: In a butt joint, the electrode should be per-pendicular to the joint and slanted from 2° to 15° in thedirection of travel. The leading angle results in a “lag-ging” gas shield, with much of the gas flowing back overthe newly deposited weld metal. With a fillet weld, theelectrode is dropped off center of the joint approximate-ly half the diameter of the electrode, and a leading angleof 2° to 15° is used.

38

Distance to nozzle:The distance of the nozzle to thework, as well as the electrical stickout, influences theperformance. The recommended nozzle-to-work dis-tance is 3/4 to 1 in., which, with concentric-type nozzles,will give an electrical stickout of about 1 to 1-1/4 in. Ifthe nozzle-to-work distance is too short, spatter mayrapidly build up on the nozzle and contact tube. Withside-shielded nozzles, electrical stickout is normally setat 3/4 to 1-1/4 in.

Shielding gas: Shielding gas is used to exclude theatmosphere from contact with the molten weld metal.Proper shielding gas and flow will depend on materialtype and location. CO2 is generally used; however, mix-tures of CO2, argon and oxygen can be used.

7.4 SAW

Amperage:Welding current determines the rate at whichthe electrode is melted, the depth of penetration of theweld pool into the base metal, and the amount of basemetal fused. An increase in current increases penetrationand melt-off rate, but an excessively high current pro-duces a high, narrow bead, an erratic arc, and undercut.Excessively low current produces an unstable arc.Welding amperages range from: 200-600 amps for 5/64in., 230-700 amps for 3/32 in., 300-900 amps for 1/8 in.,420-1000 amps for 5/32 in., and 600-1200 amps for 7/32in. electrodes.

Wire feed speed:SAW welding procedures can be con-trolled with wire feed speed as well as amperage. Wirefeed speed control would be the preferred method forSAW welding on CV systems. For CC welding, amper-age is usually used as the reference control.

Voltage: Welding voltage influences the shape of theweld cross section and the external appearance of theweld. Increasing voltage produces a flatter wider bead,increases flux consumption, and increases resistance toporosity caused by rust or scale. Excessively high volt-age produces a “hat-shaped” bead that is subject tocracking, makes slag removal difficult, and produces aconcave fillet weld that is subject to cracking.

Travel speed: Travel speed is set primarily to controlbead size and penetration. In single pass welds, the cur-rent and travel speed should be set to get the desired pen-etration without melt-through. For multiple-pass welds,the current and travel speed should be set to get thedesired bead size.

Electrical stickout: Typically, electrical stickout forSAW ranges from 1 to 2 in. Deposition rates with long-stickout welding (2 to 4 in.) are typically increased some25 to 50% with no increase in welding current. With sin-gle electrode, fully automatic submerged arc welding,the deposition rate may approach that of two-wire weld-ing with a multiple power source. Limitations exist andliterature should be evaluated for each application.

Electrode diameter:Generally, only the 1/16, 5/64, and3/32 in. electrodes are used for semiautomatic welding.The 1/16 in. wire is used for making high-speed filletwelds on steel ranging from 14-gauge to 1/4 in. thick.The 5/64 in. electrode is used for fillet and groove weldswhen the welding gun is hand-held. The 3/32 in. wire isused primarily when the gun is carried mechanically.Electrodes of this diameter can be used for hand-heldoperation, but the stiffness of the wire tends to make thecable rigid and thereby decrease the maneuverability ofthe gun. Fully automatic submerged arc welding gener-ally employs electrodes from 5/64 to 7/32 in. diameter.

Polarity: DC+ is recommended for most submerged arcwelding where fast-follow or deep penetration are impor-tant. Negative polarity gives a melt-off rate about one-third greater than that of positive polarity, but negativepolarity also produces less penetration for a given wirefeed speed. AC is recommended for two specific sub-merged arc applications: for the trailing electrode whentandem-arc welding, or for occasional single-arc applica-tions where arc blow is limiting DC current and slowingtravel speed.

Flux Type: Fluxes can be defined by a variety of meth-ods, depending on the particular operating characteristicof interest. Fluxes may generally be classified as fused,bonded, or agglomerated. Fused fluxes require that theingredients used be melted, and then crushed into smallparticles. A general disadvantage of fused fluxes is thatit is impossible to incorporate into these materials ingre-dients that perform their functions at temperatures lowerthan the melting point of the ingredients. In general, thislimits the applicability of fused fluxes to applicationswith less demanding mechanical properties. Bondedfluxes utilize water as a binder, and the resulting productis generally quite hydroscopic, absorbing moisture out ofthe atmosphere. Agglomerated fluxes utilize a high tem-perature silicate binder. This system permits the use of awider range of chemical ingredients for optimized weld-ing conditions and mechanical properties. The use of thehigh temperature binder minimizes moisture pickup ten-

39

dencies, although fluxes still must be protected frommoisture absorption where moisture condenses on thesurface of the flux.

Fluxes may also be classified based on their relative neu-trality. Neutrality is defined as the sensitivity of a flux tosilicon and/or manganese changes as a function of arcvoltage. Fluxes that are relatively immune from changesin the manganese and silicon content as the voltage ischanged are known as neutral fluxes, while those thatexperience a significant change in composition withrespect to these two elements over a range of voltage arecalled ‘active’. The Wall Neutrality Number is defined asthe absolute change in composition (that is, an increase ordecrease) with respect to manganese and silicon betweenweld deposits made at 28 volts and 36 volts. Since theresultant number is small, the sum of the absolute valuesof the percentage change to these elements is multipliedby 100. When the Wall Neutrality Number is 40 or less,the product is considered neutral, and when the value isgreater than 40, it is considered active.

Active fluxes are ideal for single or limited pass welding,particularly on surfaces that contain contaminants suchas oil, scale, or rust. The added level of oxidizers asso-ciated with active fluxes encourages high resistance toporosity and generates good bead shapes with excellentslag removal. The higher manganese content can also behelpful in minimizing centerline cracking problems dueto high sulphur plate. Neutral fluxes are ideal for multi-ple pass welds on heavier material, particularly whenhigh notch toughness levels are desired. While the sili-con and manganese content may increase to unaccept-able levels on multiple pass welding when active fluxesare used, the use of neutral fluxes minimizes this prob-lem. In general, neutral fluxes are best for multiple passwelding on 1 in. plate and greater, and active fluxes arebest when applied for single or limited pass welding. Formultiple pass welding with active fluxes on plate lessthan 1 in. thick, it is recommended that the electrode bea low silicon, low manganese electrode, and the arc volt-age should be controlled to minimize the potential ofalloy (manganese and silicon) accumulation.

Flux depth: The width and depth of the granular layer offlux influence the bead appearance and soundness of thefinished weld as well as the welding action. If the gran-ular layer is too deep, the arc is too confined and a roughweld with a rope-like appearance will result. The gasesgenerated during welding cannot escape. If the granularlayer is too shallow, the arc will not be entirely sub-

merged in flux. Flashing and spattering will occur. Theweld will have a poor appearance, and it may be porous.

Electrode type: The primary selection criteria for solidelectrode are the required mechanical properties and thechemistry of the deposited weld metal. Specific condi-tions may further affect electrode choice, such as thelevel of rust or mill scale that will be encountered.

Electrode alignment: The electrode should be kept inposition using the alignment guide. The plane of align-ment should be perpendicular to the surface of the work.Improper alignment produces undesirable bead shapes.

Work leads: The location of work connections is impor-tant, especially on short welds. Both AC and DC workconnections should be at the start of the weld, unlessback blow is desired to keep weld metal from runningahead of the arc. Where back blow is desired, the weldshould be made toward the work connection.

Nozzles:The cone tip establishes proper flux coveragewhen using the drag technique that is recommended formost joints. In general, use the smallest cone tip that willprovide sufficient flux coverage to avoid flash-throughand permit making the desired size weld. The smallerthe cone tip, the more positive the alignment of the wirewith the seam. This is particularly important when mak-ing small welds at high speed.

7.5 GMAW

Amperage:The current is adjusted by changing the wirefeed speed. Welding current will largely depend uponthe type of metal transfer that is required for a givendesign. Details of these modes are in section 2.

Voltage: Arc voltage is considered to be the electricalpotential between electrode and the workpiece.Increasing or decreasing the output voltage of the powersource will increase or decrease the arc length. Differenttypes of gases will affect voltage (higher or lower).

Travel speed: Desired bead width and penetration aredirectly affected by travel speed. Travel speed should befast enough to acquire the desired penetration. Travelthat is too slow will allow the welding arc to concentrateon the hot weld pool, resulting in a flat wide bead. Travelthat is too fast will place the arc on cold metal, giving anarrow mounded bead.

40

Electrical stickout: Electrode extension will dependupon the current and voltage used with a specified wire.GMAW electrode extensions range from 1/4 to 1 in.

Electrode diameter: The electrode size influences theweld bead configuration. Proper electrode diameter isdependent upon the application being used, materialthickness, and weld size desired. For flat and horizontalwelding, electrode diameters of 0.045, 0.052, and 1/16in. are used. Out-of-position welding is typically donewith 0.035 and 0.045 in. diameter electrodes.

Polarity: A DC+ constant voltage power source is rec-ommended for GMAW. More sophisticated powersources have been developed especially for gas metal arcwelding. Pulsed arc equipment should be considered forout-of-position GMAW.

Shielding gas: Shielding gas is used to exclude theatmosphere from contact with the molten weld metal.Proper shielding gas and flow will depend on materialtype and location. Two inert gases are used: argon andhelium. CO2 is also used either as a sole shielding gas oras a mixture with argon or helium. Small amounts ofoxygen may be added to argon. Tri- or quad-mixes arealso available for use.

7.6 ESW/EGW

Due to the complexity of these processes and the fact thatthey are not prequalified processes, ESW/EGW variablesare not listed here.

8 Welder Qualification

Qualification tests are specifically designed to deter-mine the ability of a welder to produce sound welds byfollowing a WPS. The code does not imply that anyonewho satisfactorily completes qualification tests can dothe welding for which he or she is qualified under allconditions that might be encountered during productionwelding. It is essential that welders receive some degreeof training for these differences (D1.1-96, CommentaryC4.1.2).

The most efficient route to qualify for a particularmethod is to perform tests in the 3G and 4G positionsusing 1 in. plate. Successful completion of these testswould qualify a welder in all groove and fillet positionsfor any plate thickness (D1.1-96, paragraph 4.18.1.2-2.1,Tables 4.8, 4.9).

All WPS qualification test plates are required to be visual-ly inspected as well as radiographically or ultrasonicallytested to demonstrate soundness, before being mechani-cally tested. The type of tests required are found in D1.1-96, paragraph 4.19, whereas the exact testing requirementsare listed in D1.1-96, paragraph 4.30. Also, all CJP, PJP,and fillet welds for nontubular connections will be inaccordance with D1.1-96, paragraphs 4.23, 4.24, 4.25.

A welder’s qualification will remain in effect for sixmonths beyond the date that the welder last used thewelding process, or until there is a specific reason toquestion the welder’s ability. The requalification testneed only be made using a 3/8 in. thick plate. If thewelder fails the requalification test, then a retest shall notbe permitted until further training and practice havetaken place. If there is a specific reason to question thewelder’s ability, then the type of test shall be mutuallyagreed upon between the contractor and the Engineer,and shall be within the requirements of Section 4, Part C(D1.1-96, paragraph 4.32).

Welders must be trained to understand the proper weld-ing techniques and approaches necessary to make quali-ty welds under specific conditions. Qualification tests donot realistically duplicate many field conditions.Training on mock-up assemblies can help to develop theskills required for specific situations.

9 Weld Cracking

Several types of discontinuities may occur in welds orheat affected zones. Welds may contain porosity, slaginclusions or cracks. Of the three, cracks are by far themost detrimental. Whereas there are acceptable limitsfor slag inclusions and porosity in welds, cracks arenever acceptable. Cracks in a weld, or in the vicinity ofa weld, indicate that one or more problems exist thatmust be addressed. A careful analysis of crack charac-teristics will make it possible to determine the cause andtake appropriate corrective measures.

For the purposes of this section, “cracking” will be dis-tinguished from weld failure. Welds may fail due toover-load, underdesign, or fatigue. The cracking dis-cussed here is the result of solidification, cooling, and thestresses that develop due to weld shrinkage. Weld crack-ing occurs close to the time of fabrication. Hot cracksare those that occur at elevated temperatures and are usu-ally solidification related. Cold cracks are those thatoccur after the weld metal has cooled to room tempera-

41

ture and may be hydrogen related. Neither is the resultof service loads.

Most forms of cracking result from the shrinkage strainsthat occur as the weld metal cools. If the contraction isrestricted, the strains will induce residual stresses thatcause cracking. There are two opposing forces: thestresses induced by the shrinkage of the metal, and thesurrounding rigidity of the base material. The shrinkagestresses increase as the volume of shrinking metalincreases. Large weld sizes and deep penetrating weld-ing procedures increase the shrinkage strains. The stress-es induced by these strains will increase when higherstrength filler metals and base materials are involved.With a higher yield strength, higher residual stresses willbe present.

Under conditions of high restraint, extra precautionsmust be utilized to overcome the cracking tendencieswhich are described in the following sections. It isessential to pay careful attention to welding sequence,preheat and interpass temperature, postweld heat treat-ment, joint design, welding procedures, and filler mater-ial. The judicious use of peening as an in-process stressrelief treatment may be necessary when fabricating high-ly restrained members.

9.1 Centerline Cracking

Centerline cracking is characterized as a separation in thecenter of a given weld bead. If the weld bead happens tobe in the center of the joint, as is always the case on a sin-gle pass weld, centerline cracks will be in the center ofthe joint. In the case of multiple pass welds, where sev-eral beads per layer may be applied, a centerline crackmay not be in the geometric center of the joint, althoughit will always be in the center of the bead (Figure 9-1).

Centerline cracking is the result of one of the followingphenomena:segregation induced cracking, bead shapeinduced cracking, or surface profile induced cracking.Unfortunately, all three phenomena reveal themselves inthe same type of crack, and it is often difficult to identi-

fy the cause. Moreover, experience has shown that oftentwo or even all three of the phenomena will interact andcontribute to the cracking problem. Understanding thefundamental mechanism of each of these types of center-line cracks will help in determining the corrective solu-tions.

Segregation induced crackingoccurs when low meltingpoint constituents such as phosphorous, zinc, copper andsulfur compounds in the admixture separate during theweld solidification process. Low melting point compo-nents in the molten metal will be forced to the center ofthe joint during solidification, since they are the last tosolidify and the weld tends to separate as the solidifiedmetal contracts away from the center region containingthe low melting point constituents.

When centerline cracking induced by segregation isexperienced, several solutions may be implemented.Since the contaminant usually comes from the basematerial, the first consideration is to limit the amount ofcontaminant pick-up from the base material. This maybe done by limiting the penetration of the weldingprocess. In some cases, a joint redesign may be desir-able. The extra penetration afforded by some of theprocesses is not necessary and can be reduced. This canbe accomplished by using lower welding currents.

A buttering layer of weld material (Figure 9-2), deposit-ed by a low energy process such as shielded metal arcwelding, may effectively reduce the amount of pick-up ofcontaminant into the weld admixture.

In the case of sulfur, it is possible to overcome the harm-ful effects of iron sulfides by preferentially forming man-

Figure 9-1 Centerline cracking

Figure 9-2 Buttering layers

42

ganese sulfide. Manganese sulfide (MnS) is createdwhen manganese is present in sufficient quantities tocounteract the sulfur. Manganese sulfide has a meltingpoint of 2,900°F. In this situation, before the weld metalbegins to solidify, manganese sulfides are formed whichdo not segregate. Steel producers utilize this conceptwhen higher levels of sulfur are encountered in the ironore. In welding, it is possible to use filler materials withhigher levels of manganese to overcome the formation oflow melting point iron sulfide. Unfortunately, this con-cept cannot be applied to contaminants other than sulfur.

The second type of centerline cracking is known asbead shape induced cracking. This is illustrated inFigure 9-3 and is associated with deep penetratingprocesses such as SAW and CO2 shielded FCAW.When a weld bead is of a shape where there is more

depth than width to the weld cross section, the solidify-ing grains growing perpendicular to the steel surfaceintersect in the middle, but do not gain fusion across thejoint. To correct for this condition, the individual weldbeads must have at least as much width as depth.Recommendations vary from a 1:1 to a 1.4:1 width-to-depth ratio to remedy this condition. The total weldconfiguration, which may have many individual weldbeads, can have an overall profile that constitutes moredepth than width. If multiple passes are used in this sit-uation, and each bead is wider than it is deep, a crack-free weld can be made.

When centerline cracking due to bead shape is experi-enced, the obvious solution is to change the width-to-depth relationship. This may involve a change in jointdesign. Since the depth is a function of penetration, it isadvisable to reduce the amount of penetration. This canbe accomplished by utilizing lower welding amperagesand larger diameter electrodes. All of these approacheswill reduce the current density and limit the amount ofpenetration.

The final mechanism that generates centerline cracks issurface profile conditions. When concave weld sur-faces are created, internal shrinkage stresses will placethe weld metal on the surface into tension. Conversely,when convex weld surfaces are created, the internalshrinkage forces will pull the surface into compression.These situations are illustrated in Figure 9-4. Concaveweld surfaces frequently are the result of high arc volt-ages. A slight decrease in arc voltage will cause the weldbead to return to a slightly convex profile and eliminatethe cracking tendency. High travel speeds may alsoresult in this configuration. A reduction in travel speedwill increase the amount of fill and return the surface toa convex profile. Vertical-down welding also has a ten-dency to generate these crack-sensitive, concave sur-faces. Vertical-up welding can remedy this situation byproviding a more convex bead.

9.2 Heat Affected Zone Cracking

Heat affected zone (HAZ) cracking (Figure 9-5) is char-acterized by separation that occurs immediately adjacentto the weld bead. Although it is related to the weldingprocess, the crack occurs in the base material, not in theweld material. This type of cracking is also known as“underbead cracking,” “toe cracking,” or “delayed crack-ing.” Because this cracking occurs after the steel hascooled below approximately 400°F, it can be called “coldcracking”, and because it is associated with hydrogen, itis also called “hydrogen assisted cracking.”

Figure 9-3 Bead shape induced cracking

Figure 9-4 Surface profile induced cracking

Figure 9-5 Heat affected zone cracking

43

In order for heat affected zone cracking to occur, threeconditions must be present simultaneously: there must bea sufficient level of hydrogen; there must be a sufficient-ly sensitive material involved; and, there must be a suffi-ciently high level of residual or applied stress. Adequatereduction or elimination of one of the three variables willgenerally eliminate heat affected zone cracking. In weld-ing applications, the typical approach is to limit two ofthe three variables, namely the level of hydrogen and thesensitivity of the material.

Hydrogen can enter into a weld pool from a variety ofsources. Moisture and organic compounds are the pri-mary sources of hydrogen. It may be present on the steel,the electrode, in the shielding materials, and is present inthe atmosphere. Flux ingredients, whether on the outsideof electrodes, inside the core of electrodes, or in the formof submerged arc or electroslag fluxes, can absorb mois-ture, depending on storage conditions and handling prac-tices. To limit hydrogen content in deposited welds,welding consumables must be properly maintained, andwelding must be performed on surfaces that are clean anddry.

The second necessary condition for heat affected zonecracking is a sensitive microstructure. The area of inter-est is the heat affected zone that results from the thermalcycle experienced by the region immediately surround-ing the weld nugget. As this area is heated by the weld-ing arc during the creation of the weld pool, it istransformed from its room temperature structure of fer-rite to the elevated temperature structure of austenite.The subsequent cooling rate will determine the resultantHAZ properties. Conditions that encourage the develop-ment of crack sensitive microstructures include highcooling rates and higher hardenability levels in the steel.High cooling rates are encouraged by lower heat inputwelding procedures, greater base metal thicknesses, andcolder base metal temperatures. Higher hardenabilitylevels result from greater carbon contents and/or alloylevels. For a given steel, the most effective way to reducethe cooling rate is by raising the temperature of the sur-rounding steel through preheat. This reduces the tem-perature gradient, slowing cooling rates, and limiting theformation of sensitive microstructures. Effective preheatis the primary means by which acceptable heat affectedzone properties are created, although heat input also hasa significant effect on cooling rates in this zone.

The residual stresses of welding can be reduced throughthermal stress relief, although for most structural appli-cations, this is economically impractical. For complexstructural applications, temporary shoring and other con-ditions must be considered, as the steel will have a great-ly reduced strength capacity at stress relievingtemperatures. For practical applications, heat affectedzone cracking will be controlled by effective low hydro-gen practices, and appropriate preheats.

For HAZ hydrogen cracking to occur, it is necessary forthe hydrogen to migrate into the heat affected zone, whichtakes time. For this reason, the D1.1 Code (D1.1-96, para-graph 6.11) requires a delay of 48 hours after completionof welds for the inspection of welds made on A514, A517and A709 Gr. 100 and 100W steels, known to be sensitiveto hydrogen assisted heat affected zone cracking.

With time, hydrogen diffuses from weld deposits.Sufficient diffusion to avoid cracking normally takesplace in a few weeks, although it may take many monthsdepending on the specific application. The concentra-tions of hydrogen near the time of welding are always thegreatest, and if hydrogen induced cracking is to occur, itwill generally occur within a few days of fabrication.However, it may take longer for the cracks to grow to suf-ficient size to be detected.

Although a function of many variables, general diffusionrates can be approximated. At 450°F, hydrogen diffusesat the rate of approximately 1 in. per hour. At 220°F,hydrogen diffuses the same 1 in. in approximately 48hours. At room temperature, typical diffusible hydrogenrates are 1 in. per 2 weeks. If there is a question regard-ing the level of hydrogen in a weldment, it is possible toapply a postweld heat treatment commonly called “postheat.” This generally involves the heating of the weld toa temperature of 400 - 450°F, holding the steel at thattemperature for approximately one hour for each inch ofthickness of material involved. At that temperature, thehydrogen is likely to be redistributed through diffusion topreclude further risk of cracking. Some materials, how-ever, will require significantly longer than 1 hour perinch. This operation may not be necessary where hydro-gen has been properly controlled, and it is not as power-ful as preheat in terms of its ability to prevent underbeadcracking. In order for post heat operations to be effec-tive, they must be applied before the weldment is allowedto cool to room temperature. Failure to do so could resultin heat affected zone cracking prior to the application ofthe post heat treatment.

44

9.3 Transverse Cracking

Transverse cracking, also called cross cracking, is char-acterized as a crack within the weld metal perpendicularto the direction of travel (Figure 9-6). This is the leastfrequently encountered type of cracking, and is general-ly associated with weld metal that is higher in strength,significantly overmatching the base material. This typeof cracking can also be hydrogen assisted, and like theheat affected zone cracking described in 9.1, transversecracking is also a factor of excessive, hydrogen, residualstresses, and a sensitive microstructure. The primary dif-ference is that transverse cracking occurs in the weldmetal as a result of the longitudinal residual stress.

As the weld bead shrinks longitudinally, the surroundingbase material resists this force by going into compres-sion. The high strength of the surrounding steel in com-pression restricts the required shrinkage of the weldmaterial. Due to the restraint of the surrounding basematerial, the weld metal develops longitudinal stresseswhich may facilitate cracking in the transverse direction.

When transverse cracking is encountered, a review of thelow hydrogen practice is warranted. Electrode storageconditions should be carefully reviewed. If this is a prob-lem, a reduction in the strength of the weld metal willusually solve transverse cracking problems. Of course,design requirements must still be met, although mosttransverse cracking results from weld metal over match-ing conditions.

Emphasis is placed upon the weld metal because thefiller metal may deposit lower strength, highly ductilemetal under normal conditions. However, with the influ-ence of alloy pick-up, it is possible for the weld metal toexhibit extremely high strengths with reduced ductility.Using lower strength weld metal is an effective solution,but caution should be taken to ensure that the requiredjoint strength is attained.

Preheat may have to be applied to alleviate transversecracking. The preheat will assist in diffusing hydrogen.

As preheat is applied, it will additionally expand thelength of the weld joint, allowing the weld metal and thejoint to contract simultaneously, and reducing the appliedstress to the shrinking weld. This is particularly impor-tant when making circumferential welds. When the cir-cumference of the materials being welded is expanded,the weld metal is free to contract along with the sur-rounding base material, reducing the longitudinal shrink-age stress. Finally, post weld hydrogen releasetreatments that involve holding the steel at 250-450°F forextended periods of time (generally 1 hour per in. ofthickness) will assist in diffusing any residual hydrogen.

10 Weld Quality and Inspection

10.1 Weld Quality

A weld must be of an appropriate quality to ensure that itwill satisfactorily perform its function over its intendedlifetime. Weld “quality” is therefore directly related tothe purpose the weld must perform. Codes or contractdocuments define the required quality level for a specif-ic project, meaning that a quality weld is one that meetsthe applicable requirements. Ensuring that the require-ments have properly addressed the demands upon theweld is ultimately the responsibility of the Engineer.

All welds contain discontinuities, which are defined asan interruption in the typical structure of the material,such as a lack of homogeneity in its mechanical, metal-lurgical, or physical characteristics (AWS A3.0). Suchirregularities are not necessarily defects. A defect isdefined as a discontinuity that is unacceptable withrespect to the applicable standard or specification.Defects are not acceptable; discontinuities may, or maynot, be acceptable.

Welds are not required to be “perfect,” and most weldswill contain some discontinuities. It is imperative that theapplicable standards establish the level of acceptability ofthese discontinuities in order to ensure both dependableand economical structures. AWS D1.1 is the primarystandard used to establish workmanship requirements. Ingeneral, these are based upon the quality level achievableby a qualified welder, which does not necessarily consti-tute a boundary of suitability for service. If the weld qual-ity for each type of weld and loading condition werespecified, widely varying criteria of acceptable workman-ship would be required. Moreover, acceptable weld qual-ity (in some cases) would be less rigorous than whatwould be normally produced by a qualified welder.

Figure 9-6 Transverse cracking

45

(D1.1-96, page 404, Commentary C6.8). This suggeststhat, in some instances, the D1.1 requirements exceed theactual requirements for acceptable performance. TheEngineer of Record can use a “fitness for purpose” evalu-ation to determine alternate acceptance criteria in such sit-uations. Some specific loading conditions require morestringent acceptance criteria than others. For example,undercut associated with fillet welds would constitute astress riser when the fillet weld is loaded in tension per-pendicular to its longitudinal axis. However, when thesame fillet weld is loaded in horizontal shear, this wouldnot be a stress riser, and more liberal allowances are per-mitted for the level of undercut.

A variety of types of discontinuities can exist in welds.Characteristics, causes, and cures of common examplesmay be summarized as follows:

Undercut is a small cavity that is melted into the basemetal adjacent to the toe of a weld that is not subse-quently filled by weld metal. Improper electrode place-ment, extremely high arc voltages, and the use ofimproper welding consumables may result in undercut.Changes to the welding consumable and welding proce-dures may alleviate undercut.

Excess concavityor excess convexityare weld surfaceprofile irregularities. These may be operator- and/orprocedure-related.

Overlap (or cold-lap) is the protrusion of weld metalbeyond the toe of the weld where the weld metal is notbonded to the base material. Overlap usually is associat-ed with slow travel speeds.

Incomplete penetration is associated with weld jointdetails that rely on melting of base metal to obtain therequired weld strength. A typical example would be asquare-edged butt joint. Incomplete penetration occurswhen the degree of penetration is inadequate, and is gen-erally attributable to insufficient current density, improp-er electrode placement, or excessively slow travel speeds.

Lack of fusion, or incomplete fusion, is the result of thefailure of the weld metal and the base metal to form themetallurgical bonds necessary for fusion. Lack of fusioncan range from small, isolated planes, or, in extremecases, may consist of a complete plane between the weldmetal and the base metal where fusion does not exist.Improper filler metal selection, improper welding proce-

dures, and poor surface preparation are common causesof this condition. Improper use of GMAW short-circuit-ing transfer is a common cause of lack of fusion.

Arc strikes consist of small, localized regions of metalthat have been melted by the inadvertent arcing betweenelectrically charged elements of the welding circuit andthe base metal. Welding arcs that are not initiated in thejoint leave behind these arc strikes. Arcing of workclamps to the base metal can cause arc strikes, as canwelding cables with improper insulation. SMAW is par-ticularly susceptible to creating arc strikes since the elec-trode holder is electrically ‘hot’ when not welding. Theuse of properly insulated welding equipment and properwelding practices minimize arc strikes. Grinding awaythe affected (melted) metal is an effective way of elimi-nating any potential harm from arc strikes.

Slag inclusionsdescribe non-metallic material entrappedin the weld metal, or between the weld metal and basemetal. Slag inclusions are generally attributed to slagfrom previous weld passes that was not completelyremoved before subsequent passes were applied. Slagmay be trapped in small cavities or notches, makingremoval by even conscientious welders difficult. Properjoint designs, welding procedures, and welder techniquecan minimize slag inclusions.

Spatter is the term used to describe the roughly spheri-cal particles of molten weld metal that solidify on thebase metal outside the weld joint. Spatter is generallynot considered to be harmful to the performance of weld-ed connections, although excessive spatter may inhibitproper ultrasonic inspection, and may be aestheticallyunacceptable for exposed steel applications. Excessivespatter is indicative of less than optimum welding condi-tions, and suggests that the welding consumables and/orwelding procedure may need to be adjusted.

Porosity consists of spherical or cylindrical cavities thatare formed as gases entrapped in the liquid weld metalescape while the metal solidifies. The D1.1-96, Table 6.1code defines acceptable limits for porosity as a functionof its type, size, and distribution. Porosity occurs as theresult of inadequate shielding of the weld metal, orexcessive contamination of the weld joint, or both. Theproducts used for shielding weld deposits (gases, slags)must be of appropriate quality, properly stored, anddelivered at a rate to provide adequate shielding.Excessive surface contamination such as oil, moisture,

46

rust, or scale increases the demand for shielding.Porosity can be minimized by providing proper shield-ing, and ensuring joint cleanliness.

Cracking is the most serious type of weld discontinuity.Weld cracking is extensively discussed in section 9.

10.2 Weld Quality and Process-Specific Influences

Some welding processes are more sensitive to the gener-ation of certain types of weld discontinuities, and someweld discontinuities are associated with only a few typesof welding processes. Conversely, some welding process-es are nearly immune from certain types of weld discon-tinuities. Contained below are the popular weldingprocesses and their variations, along with a description oftheir associated sensitivity relative to weld quality.

SMAW — The unique limitations of shielded metal arcwelding fall into three categories: arc length related dis-continuities, start-stop related discontinuities, and coat-ing moisture related problems. In SMAW, the operatorcontrols arc length. Excessively short arc lengths canlead to arc outages, where the electrode becomes stuck tothe work. When the electrode is mechanically broken offthe joint, the area where the short has occurred needs tobe carefully cleaned, usually ground, to ensure condi-tions that will be conducive to good fusion by subsequentwelding. The electrode is usually discarded since a por-tion of the coating typically breaks off of the electrodewhen it is removed from the work. Excessively long arclengths will generate porosity, undercut, and excessivespatter. Because of the finite length of the SMAW elec-trodes, an increased number of starts and stops is neces-sitated. During arc initiation with SMAW, startingporosity may result during the short time after the arc isinitiated and before adequate shielding is established.Where the arc is terminated, under-filled weld craters canlead to crater cracking. The coatings of SMAW elec-trodes are sensitive to moisture pick-up. While newerdevelopments in electrodes have extended the period forwhich electrodes may be exposed to the atmosphere, it isstill necessary to ensure that the electrodes remain dry inorder to be assured of low hydrogen welding conditions.Improper care of low hydrogen SMAW electrodes canlead to hydrogen assisted cracking, i.e., underbead crack-ing or transverse cracking. See 2.1 on care and storageof low hydrogen electrodes.

FCAW-g — In FCAW-g, as with all gas shieldedprocesses, it is important to protect the gas shielding

around the weld deposit. If FCAW-g gas shields are dis-turbed by winds, fans, or smoke exhaust equipment,porosity can result. The deep penetrating characteristicsof FCAW-g are generally advantageous, but excessivepenetration can lead to centerline cracking because of apoor width-to-depth ratio in the weld bead cross section.

FCAW-ss — Excessively high arc voltages, or inappro-priately short electrode extension dimensions can lead toporosity with FCAW-ss. When excessive voltages areused, the demand for shielding increases, but since theamount of shielding available is relatively fixed, porositycan result. When the electrical stickout distance is tooshort, there may be inadequate time for the various ingre-dients contained within the electrode core to chemicallyperform their function before they are introduced into thearc. This too can lead to porosity. Because of theextremely high deposition rate capability of some of theFCAW-ss electrodes, it is possible to deposit quantitiesof weld metal that may result in excessively large weldbeads, leading to a decrease in fusion, if not balancedwith a corresponding increase in travel speed.

SAW — Submerged arc welding is sensitive to align-ment of the electrode with respect to the joint. Misplacedbeads can result from improper bead placement. Thedeep penetration of the SAW process can lead to center-line cracking due to improper width-to-depth ratios in thebead cross section.

GMAW — When solid electrodes are used, and particu-larly when welding out-of-position, the short arc transfermode is frequently used. This can directly lead to cold-lap, a condition where complete fusion is not obtainedbetween the weld metal and base material. This is amajor shortcoming of the GMAW process and is one ofthe reasons its application is restricted by the D1.1 codewith respect to its prequalified status. As with all gas-shielded processes, GMAW is sensitive to the loss of gasshielding.

10.3 Weld Inspection

Weld quality is directly tied to the code or specificationunder which the work is being performed. Welds areacceptable when they conform to all the requirements ina given specification or code.

Five major non-destructive methods are used to evaluateweld metal integrity in steel structures. Each has uniqueadvantages and limitations. Some discontinuities are

47

revealed more readily with one method as compared toanother. It is important for the fabricator to understandthe capacities and limitations of these inspection meth-ods, particularly in situations where interpretation of theresults may be questionable.

Visual inspection (VT) is by far the most powerfulinspection method available. Because of its relative sim-plicity and lack of sophisticated equipment, some peoplediscount its power. However, it is the only inspectionmethod that can actually increase the quality of fabrica-tion and reduce the generation of welding defects.

Most codes require that all welds be visually inspected.Visual inspection begins long before an arc is struck.Materials that are to be welded must be examined forquality, type, size, cleanliness, and freedom from defects.The pieces to be joined should be checked for straight-ness, flatness, and dimensions. Alignment and fit-up ofparts should be examined. Joint preparation should beverified. Procedural data should be reviewed, and pro-duction compliance assured. All of these activitiesshould precede any welding that will be performed.

During welding, visual inspection includes verificationthat the procedures used are in compliance with theWelding Procedure Specification (WPS). Upon comple-tion of the weld bead, the individual weld passes areinspected for signs of porosity, slag inclusion, and anyweld cracks. Bead size, shape, and sequences can beobserved.

Interpass temperatures can be verified before subsequentpasses are applied. Visual inspection can ensure compli-ance with procedural requirements. Upon completion ofthe weld, the size, appearance, bead profile and surfacequality can be inspected.

Visual inspection may be performed by the weld inspec-tor, as well as by the welder. Good lighting is imperative.In most fabrication shops, some type of auxiliary lightingis required for effective visual inspection. Magnifyingglasses, gauges, and workmanship samples all aid invisual inspection.

Liquid penetrant testing (PT) involves the applicationof a liquid which by a capillary action is drawn into a sur-face breaking discontinuity, such as a crack or porosity.When the excess residual dye is carefully removed fromthe surface, a developer is applied, which will absorb the

penetrant that is contained within the discontinuity. Thisresults in a stain in the developer showing that a discon-tinuity is present.

Dye penetrant testing is limited to surface discontinu-ities. It has no ability to read subsurface discontinuities,but it is highly effective in identifying the surface dis-continuities that may be overlooked or be too small todetect with visual inspection. However, because it is lim-ited to surface discontinuities, and because these discon-tinuities also will be observed with magnetic particleinspection, this method is not specified by most structur-al steel welding codes.

Magnetic particle inspection (MT) utilizes the changein magnetic flux that occurs when a magnetic field is pre-sent in the vicinity of a discontinuity. This change inmagnetic flux density will show up as a different patternwhen magnetic powders are applied to the surface of apart. The process is effective in locating discontinuitiesthat are on the surface and slightly subsurface. For steelstructures, magnetic particle inspection is more effectivethan dye penetrant inspection, and hence, is preferred formost applications. Magnetic particle inspection canreveal cracks very near the surface, slag inclusions, andporosity.

The magnetic field is created in the material to beinspected in one of two ways. Current is either directlypassed through the material, or a magnetic field isinduced through a coil on a yoke. With the first method,electrical current is passed through two prods that areplaced in contact with the surface. When the prods areinitially placed on the material, no current is applied.After intimate contact is assured, current is passedthrough. Small arcs may occur between the prods andthe base material, resulting in an arc strike, which maycreate a localized brittle zone. It is important that prodsbe kept in good shape and that intimate contact with thework is maintained before the current is passed throughthe prods.

The second method of magnetic field generation isthrough induction. In what is known as the yoke method,an electrical coil is wrapped around a core, often witharticulated ends. Electrical current is passed through thecoil, creating a magnetic field in the core. When the endsof the yoke are placed in contact with the part beinginspected, the magnetic field is induced into the part.Since current is not passed into the part, the potential for

48

arc strikes is eliminated. Along with this significantadvantage, comes a disadvantage: the yoke method is notas sensitive to subsurface discontinuities as the prodmethod.

Cracks are most easily detected when they lie perpendic-ular to the magnetic field. With the prod method themagnetic field is generated perpendicular to the directionof current flow. For the yoke method, just the opposite istrue. Magnetic particle inspection is most effective whenthe region is inspected twice: once with the field locatedparallel to, and once with the field perpendicular to, theweld axis.

While magnetic particle inspection can reveal some sub-surface discontinuities, it is best used to enhance visualinspection. Fillet welds can be inspected with thismethod. Another common use of MT is for the inspec-tion of intermediate passes on large groove welds, partic-ularly in crack sensitive situations.

Radiographic inspection (RT) uses X-rays or gammarays that are passed through the weld and expose a pho-tographic film on the opposite side of the joint. X-raysare produced by high voltage generators, while gammarays are produced by atomic disintegration of radioactiveisotopes.

Whenever radiography is used, precautions must betaken to protect workers from exposure to excessive radi-ation. Safety measures dictated by the OccupationalSafety and Health Administration (OSHA), the NationalElectrical Manufacturer’s Association (NEMA), theNuclear Regulatory Commission (NRC), the AmericanSociety of Nondestructive Testing (ASNT) and otheragencies should be carefully followed when radiograph-ic inspection is conducted.

Radiographic testing relies on the ability of the materialto pass some of the radiation through, while absorbingpart of this energy within the material. Different materi-als have different absorption rates. Thin materials willabsorb less radiation than thick materials. The higher thedensity of the material, the greater the absorption rate.As different levels of radiation are passed through thematerials, portions of the film are exposed to a greater orlesser degree than the rest. When this film is developed,the resulting radiograph will bear the image of the planviews of the part, including its internal structure. A radi-ograph is actually a negative. The darkest regions arethose that were most exposed when the material being

inspected absorbed the least amount of radiation. Thinparts will be darkest on the radiograph. Porosity will berevealed as small, dark, round spots. Slag is also gener-ally dark, and will look similar to porosity, but will beirregular in its shape. Cracks appear as dark lines. Lackof fusion or underfill will show up as dark spots.Excessive reinforcement on the weld will result in a lightregion.

Radiographic testing is most effective for detecting volu-metric discontinuities: slag and porosity. When cracksare oriented perpendicular to the direction of the radia-tion source, they may be missed with the RT method.Tight cracks that are parallel to the radiation path havealso been overlooked with RT.

Radiographic testing has the advantage of generating apermanent record for future reference. With a “picture”to look at, many people are more confident that the inter-pretation of weld quality is meaningful. However, read-ing a radiograph and interpreting the results requiresstringent training, so the effectiveness of radiographicinspection depends to a great degree upon the skill of thetechnician.

Radiographic testing is best suited for inspection of com-plete joint penetration (CJP) groove welds in butt joints.It is not particularly suitable for inspection of partial jointpenetration (PJP) groove welds or fillet welds. Whenapplied to tee and corner joints, the geometric constraintsof the applications make RT inspection difficult, andinterpretation of the results is highly debatable.

Ultrasonic inspection (UT)relies on the transmission ofhigh frequency sound waves through materials. Solid,discontinuity-free materials will transmit the soundthroughout a part in an uninterrupted fashion. A receiv-er “hears” the sound reflected off of the back surface ofthe part being inspected. If a discontinuity is containedbetween the transmitter and the back side of the part, anintermediate signal will be sent to the receiver indicatingthe presence of this discontinuity. The pulses are dis-played on a screen. The magnitude of the signal receivedfrom the discontinuity is proportional to the amount ofreflected sound. This is indirectly related to the size,type, and orientation of the reflecting surface. The rela-tionship of the signal with respect to the back wall willindicate its location. Ultrasonic inspection is sensitiveenough to read discontinuities that are not relevant to theperformance of the weld. It is a sophisticated device thatis very effective in spotting even small discontinuities.

49

UT is most sensitive to planar discontinuities, such ascracks, laminations, and non-fusion perpendicular to thedirection of sound transmission. Under some conditions,uniformly cylindrical or spherical discontinuities can beoverlooked with UT.

Ultrasonic inspection is very effective for examination ofCJP groove welds. While UT inspection of PJP groovewelds is possible, interpretation of the results can be dif-ficult. UT inspection can be applied to butt, corner, andT-joints, and offers a significant advantage over RT.

A common situation in UT inspection is worth notingbecause of the problems encountered. In tee and cornerjoints, with CJP groove welds made from one side andwith steel backing attached, the interpretation of resultsis difficult at best. It is difficult to clearly distinguishbetween the naturally occurring regions where the back-ing contacts the adjacent vertical tee or corner joint mem-ber and an unacceptable lack of fusion. There is alwaysa signal generated in this area. This of course is the sit-uation that is encountered when steel backing is left inplace on a beam-to-column moment connection. To min-imize this problem, the steel backing can be removed.This offers two advantages: First, the influence of thebacking is obviously eliminated; and secondly, in theprocess of backing removal, the joint can be backgougedand the root inspected prior to the application of the backweld and the reinforcing fillet weld (FEMA 267).

It is also important to note that when a bottom beam-to-column connection is inspected from the top side of theflange, it is impossible for the operator to scan across theentire width of the beam flange because of the presenceof the beam web. This leaves a region in the center of theweld that cannot be UT inspected. Unfortunately, this isalso the region that is most difficult for the welder todeposit sound weld metal in, and has been identified asthe source of many weld defects. When the beam isjoined to a wide flange column, this is also the mostseverely loaded portion of the weld. Backing removaland subsequent backgouging operations help overcomethis UT limitation since it affords the opportunity of visu-al verification of weld soundness.

11 Arc Welding Safety

Arc welding is a safe occupation when sufficient mea-sures are taken to protect the welder from potential haz-ards. When these measures are overlooked or ignored,welders can encounter such dangers as electric shock,over-exposure to radiation, fumes and gases, and fire andexplosion; any of these can result in fatal injuries.Everyone associated with the welding operation shouldbe aware of the potential hazards and ensure that safepractices are employed. Infractions should be reported tothe appropriate responsible authority.

Supplement One is a guide for the proper selection of anappropriate filter or shade for eye protection when direct-ly observing the arc. Supplement Two lists published stan-dards and guidelines regarding safety. Supplement Threeconsists of a series of precautions covering the major areaof potential hazards associated with welding. SupplementFour is a checklist which gives specific instructions to thewelder to ensure safe operating conditions.

50

SUPPLEMENT 1Guide for Shade Numbers

51

SUPPLEMENT 2Arc Welding Safety Precautions

52

53

SUPPLEMENT 3Welding Safety Checklist

54

SUPPLEMENT 4References

ANSI/AWS D1.1-98 Structural Welding Code - Steel. The American Welding Society, 1998.

AWS A3.0-94 Standard Welding Terms and Definitions.The American Welding Society, 1994.

AWS A5.17-97 Specification for Carbon Steel Electrodesand Fluxes for Submerged Arc Welding. The American Welding Society, 1997.

AWS A5.1-91 Specification for Carbon Steel Electrodesfor Shielded Metal Arc Welding. The American Welding Society, 1991.

AWS A5.20-95 Specification for Carbon Steel Electrodesfor Flux Cored Arc Welding. The American Welding Society, 1995.

AWS A5.23-97 Specification for Low Alloy SteelElectrodes and Fluxes for Submerged Arc Welding. The American Welding Society, 1997.

AWS A5.29-98 Specification for Low Alloy Electrodesfor Flux Cored Arc Welding.The American Welding Society, 1998.

AWS A5.5-96 Specification for Low Alloy Electrodes for Shielded Metal Arc Welding. The American Welding Society, 1996.

AWS B2.1 Welding Procedure Specifications. The American Welding Society, 1999.

AWS Structural Welding Committee Position Statementon Northridge Earthquake Welding Issues. The American Welding Society, November 10, 1995.

Boniszewski, T. Self-Shielded Arc Welding.Cambridge: Abington, 1992.

Interim Guidelines: Evaluation, Repair, Modificationand Design of Welded Steel Moment Frame Structures(FEMA 267). Federal Emergency Management Agency,August 1995.

Load & Resistance Factor Design. The American Institute of Steel Construction, 1999.

The Procedure Handbook of Arc Welding. 13th Edition.The James F. Lincoln Arc Welding Foundation, 1995.

Welding Handbook, Vol. 1 - Welding Technology. The American Welding Society, 1987.