The welding of aluminium and its alloys

Process principles

The MIG welding process, illustrated in Figs. 7.1 and 7.2, as a rule uses direct current with the electrode connected to the positive pole of the power source, DC positive, or reverse polarity in the USA. As explained in Chapter 3 this results in very good oxide film removal. Recent power source developments have been successful in enabling the MIG process to be also used with AC. Most of the heat developed in the arc is generated at the positive pole, in the case of MIG welding the electrode, resulting in high wire burn-off rates and an efficient transfer of this heat into the weld pool by means of the filler wire. When welding at low welding currents the tip of the continuously fed wire may not melt sufficiently fast to maintain the arc but may dip into the weld pool and short circuit. This short circuit causes the wire to melt somewhat like an electrical fuse and the molten metal is drawn into the weld pool by surface tension effects. The arc re-establishes itself and the cycle is repeated. This is known as the dip transfer mode of metal transfer. Excessive spatter will be produced if the welding parame­ters are not correctly adjusted and the low heat input may give rise to lack-

of-fusion defects. At higher currents the filler metal is melted from the wire tip and transferred across the arc as a spray of molten droplets, spray trans­fer. This condition gives far lower spatter levels and deeper penetration into the parent metal than dip transfer. When MIG welding aluminium the low melting point of the aluminium results in spray transfer down to relatively low welding currents, giving a spatter-free joint.

The low-current, low-heat input dip transfer process is useful for the welding of thin plate or when welding in positions other than the flat (PA)

Table 7.1 Metal transfer modes and wire diameter

Metal transfer mode Wire diameter

TOC o "1-5" h z Dip 0.8mm

Pulsed 1.2 and 1.6mm

Conventional spray 1.2 and 1.6mm

High-current spray 1.6mm

High-current mixed 2.4mm

spray/globular

GLOBULAR/SPRAY

2.4

E

cc 1.6

PULSED SPRAY

DIP

0.8

0 50 100 150 200 250 300 350 400 450 500 550 600 650 700

WELDING CURRENT (A)

7.2 Typical welding current ranges for wire diameter and welding current.

position (see Fig. 10.3 for a definition of welding positions). It has, however, been supplanted in many applications by a pulsed current process, where a high current pulse is superimposed on a low background current at regular intervals. The background current is insufficient to melt the filler wire but the pulse of high current melts the filler metal and projects this as a spray of droplets of a controlled size across the arc, giving excellent metal trans­fer at low average welding currents.

Table 7.1 lists the likely and/or commonest methods of metal transfer with respect to wire diameter. Figure 7.3 illustrates the typical current ranges for a range of wire diameters.

7.2.1 Power sources

The MIG arc requires a power source that will provide direct current and with a suitable relationship established between welding current and voltage, this relationship being known as the power source dynamic char­acteristic. As mentioned above the MIG process uses a continuous wire feed and for the majority of welding operations it is important that the rate at which the wire burns off in the arc is matched by the wire feed speed. Failure to do this can result in an unstable arc and variable weld quality. To achieve this control many MIG/MAG welding power sources are designed with a flat or constant voltage characteristic. The importance of this characteristic becomes apparent when we consider what happens during manual welding. The manual welder cannot maintain a fixed invariable arc length while welding - an unsteady hand or repositioning himself during welding means that the arc length varies and this in its turn causes variations in arc voltage. When this happens with a flat characteristic power source a small increase in the arc length results in an increase in arc voltage, giving a large drop in arc current, as illustrated in Fig. 7.4. Since the wire burn-off rate is deter­mined by the current this also decreases, the tip of the wire moves closer to the weld pool, decreasing the voltage and raising the current as it does so. The burn-off rate therefore rises, the arc length increases and we have what is termed a self-adjusting arc where a constant arc length and filler metal deposition rate are maintained almost irrespective of the torch movement.

During both dip and spray transfer the speed at which the power source responds to the changes in the arc length is determined by the inductance

in the welding circuit. This controls the rate of current rise or fall and can have a significant effect on weld quality. Insufficient inductance permits the welding current to rise extremely rapidly, giving rise to excessive spatter and burning back of the wire to the contact tip. Too high an inductance means that the wire does not melt sufficiently rapidly and the wire tip may stub into the weld pool or be pushed through the root pass to protrude from the root. It is essential therefore that the power source is adjusted for the correct amount of inductance when, for example, the wire diameter or wire feed speed is changed.

The converse of the flat characteristic power source is the drooping char­acteristic or constant current power source, illustrated in Fig. 7.5. This design of power source is generally used in MMA and TIG welding but it also has some advantages when MIG welding aluminium. With a drooping charac­teristic a large change in arc voltage results in only a small change in arc current. Heat input is therefore reasonably constant, unlike that from a flat characteristic power source arc, giving more consistent penetration.

The problem with the drooping characteristic power source when used for MIG welding is that it requires more skill on the part of the welder. With push wire feeders the soft aluminium wire can buckle within the wire feed conduit, particularly with long and flexible conduits. This results in the wire feed speed at the contact tip fluctuating and, if no action is taken, vari­ations in the heat input to the weld. When using a flat characteristic power source these fluctuations are compensated for by the power source and the welder may not appreciate that this is occurring - with the drooping char­acteristic the arc length changes and the welder may experience what are perceived as arc stability problems. If the welder is sufficiently skilled, cor­rective action can be taken before this results in welding defects, whereas with the flat characteristic power source the welder can produce lack of fusion or excess penetration defects unknowingly. An advantage of the drooping characteristic power source is that as the welding current and the wire feed speed are fixed the welder can employ these features to enable the wire tip to be pushed into the joint, a useful feature when making the root pass.

The drooping characteristic unit is also useful in deep weld preparations. In such joints the constant voltage power source may measure the arc voltage from the side wall, rather than from the bottom of the weld prepa­ration, resulting in an unstable arc condition, poor bead shape and variable penetration. The same restrictions apply when welding the root pass in fillet welds where a drooping or constant current unit may give better results than the constant voltage power source. Weaving of the torch may also cause problems where the torch is moved simply by pivoting the wrist. This gives a regular increase and decrease in arc length, causing a loss of pene­tration at the limits of the weave with the flat characteristic power source. However, despite the apparent advantages of a drooping characteristic power source the bulk of MIG welding units in production today use a flat characteristic with consistent and acceptable results.

7.2.1.1 Pulsed MIG welding

Pulsed MIG welding was developed in the early 1960s but it was not until the late 1970s that the process began to be widely adopted on the shop floor. Prior to this date the equipment had been expensive, complicated and difficult to set up for optimum welding parameters, making it welder - unfriendly and impeding its acceptance by the most important individual in the welding workshop. Solid state electronics started to be used in welding power sources in the 1970s and ‘single knob’ control became pos­sible with the advent of synergic logic circuits. The synergic capability enabled all of the welding parameters to be controlled from a single dial control which optimised the current peak pulse and background values, the voltage and the wire feed speed. It has also became possible to repro­gramme the power source instantly when wire size, shield gas, filler metal composition, etc. are changed, simply by dialling in a programme number (Fig. 7.6). These programmes have been established by the equipment man­ufacturer with the optimum parameters for the application. Initially these units were expensive but the price has been steadily reduced such that they

are now only marginally more costly than a conventional power source, leading to a far wider usage. The modern inverter-based units (Fig. 7.7), are also far lighter, far more energy efficient and more robust than the older units that they are replacing.

The pulsed MIG process uses a low ‘background’ current, sufficient to maintain the arc but not high enough to cause the wire to melt off. On this background current a high-current, ‘peak’ pulse is superimposed. Under optimum conditions this causes a single droplet of molten filler wire to be projected across the arc into the weld pool by spray transfer. It is thus pos­sible to achieve spray transfer and a stable arc at low average welding cur­rents. This enables very thin metals to be welded with large diameter wires where previously very thin wires, difficult to feed in soft aluminium, needed to be used. The lower currents also reduce penetration, useful when welding thin materials and also enable slower welding speeds to be used, making it easier for the welder to manipulate the torch in difficult access conditions or when welding positionally.

The use of electronic control circuitry enables arc starting to be achieved without spatter or lack of fusion defects. Some units now available will slowly advance the wire until the tip touches the workpiece, sense the short circuit, retract the wire to the correct arc length and initiate the full welding current (Fig. 7.8). Similarly, in most of these modern units a crater filling facility is built in, which automatically fades out the current when the trigger on the gun is released.

7.7 Modern 500 amp inventor-based programmable synergic pulsed MIG power source. Courtesy of TPS-Fronius Ltd.

If you are contemplating purchasing new or replacement MIG equip­ment it is recommended that pulsed MIG power sources are purchased, even though they are more expensive than conventional equipment. This will give the fabrication shop a more flexible facility with a wider range of options than with the straight DC units.

7.2.1.2 Fine wire MIG

As the name suggests the fine wire MIG process uses a fine, small diameter wire, less than 1.2 mm and as small as 0.4 mm in diameter, although wires of 0.4 and 0.6 mm in diameter need to be specially ordered from the wire drawer. Small diameter wires are notoriously difficult to feed and to eliminate feeding problems a small wire reel and a set of drive rolls are mounted directly on the welding torch. Welding parameters are in the ranges 50-140 A and 17-22 V, resulting in a short-circuiting mode of metal transfer. Travel speeds are generally around 320mm/min, giving low heat input and enabling thin sheets, around 1 mm in thickness, to be welded without burn-through, excessive penetration or excessive cap height. The fine wire process, although successful, has now largely been replaced by pulsed MIG welding.

7.2.1.3 Twin wire MIG

A relatively recent development has been the twin-wire process. The current that can be used is limited in the single wire process by the forma­tion of a strong plasma jet at high welding currents. This jet may cause an irregular bead shape, porosity or excess penetration. The twin wire process overcomes these difficulties with two independent arcs operating in the same weld pool, enabling major improvements in productivity to be achieved. The basis of this is the use of two inverter-based pulsed MIG power sources coupled in series, each complete with its own microproces­sor control unit and wire feeder (Fig. 7.9). The two units are linked by a controller that synchronises the pulses from each unit such that when one unit is welding on the peak of a current pulse the other unit is on back­ground current. By this means a stable welding condition is created with the two arcs operating independently of each other. The wires are fed to a single torch carrying two contact tips insulated from each other. The wires may be positioned in tandem, side by side or at any angle in between enabling the bead width and joint filling to be precisely controlled.

The limitation of twin wire MIG is that the process can only be used in a mechanised or robotic application. With suitable manipulators, however, it is capable of very high welding speeds, a 3 mm leg length fillet weld, for

instance, being capable of being made at travel speeds of over 2 metres per minute. The welding torch is large, making access a problem, and the capital cost of the equipment is high.

7.2.2 Wire feeders and welding torches

7.2.2.1 Welding torches

The MIG process requires the filler wire to be delivered to the welding torch (Fig. 7.10) at a fixed speed and for the welding current to be trans­ferred to the wire via a contact tip within the torch. The torch must also be equipped with a means of providing the shield gas and of enabling the welder to commence and end the welding sequence. This is generally achieved by means of a trigger on the handle of the torch. Operating the trigger initiates the shielding gas flow and the welding current when the wire tip is scratched on the workpiece surface. This, in its turn, starts the wire feed. Releasing the trigger stops the wire feed and shuts off the current and shielding gas. The heat generated in the torch during welding may also require the torch to be water-cooled. All of these services must be deliv­ered to the torch via an umbilical cable containing a wire feed conduit,

7.10 Exploded view of a typical MIG torch: A ergonomically shaped handle; B contact tip, C gas shroud, D gas diffuser, E power cable connector, F umbilical containing gas hose, power cable and control cable, G power switch, H replaceable liner, I adjustable nozzle. Courtesy of Bernard Welding Equipment Company.

welding current cable, shield gas hose, cooling water delivery and return hoses and the electrical control cables. At the same time the torch must not be made so heavy and cumbersome that the welder cannot easily manipulate the torch with a minimum of effort. A well-designed torch therefore needs to be lightweight, robust and easily maintained and the umbilical cable needs to be light and flexible. It is most important if con­sistent quality is to be achieved that the welder is provided with the best torch available.

7.2.2.2 Wire feed systems

There are three basic forms of wire feeders: the ‘push’ system, the ‘pull’ system and the ‘push-pull’ system. As the name suggests, in the push system, the wire is pushed by the wire feed drive rolls along the conduit to the welding torch. The flexibility of aluminium wire means that the wire can buckle and jam inside the conduit, resulting in irregular wire feeding at the welding torch and, in extreme cases, a ‘bird’s nest’ of tangled wire at the wire feed unit. Such wire feeders are generally restricted to a minimum wire diameter of 1.6 mm and the wire feed conduit to a length of 3.5 m.

The pull system utilises a set of wire rolls in the torch handle which pull the wire from the wire reel (Fig. 7.11).This arrangement increases the weight of the torch and does not increase the distance over which the wire can be fed, this still being limited to around 3.5m, although the consistency of the wire feed is improved and wire diameters down to 0.8 mm can be used.

The push-pull system is a combination of the above two systems with a set of drive rolls at both the wire reel feeder and in the torches illustrated in Fig. 7.11. This enables small diameter wires to be fed up to 15 m from the wire reel. The final variation on this theme is the spool on gun torch which utilises a small 100 mm diameter wire reel mounted on the welding torch and a set of drive rolls in the torch body. These rolls push the wire the short distance from the reel to the contact tip, enabling wires as small as 0.4 mm in diameter to be used. The length of the umbilical cable is limited only by the voltage drop in the power delivery and return leads and perhaps the need to provide water cooling to the torch.

All of these systems require that the wire is driven at a constant, con­trolled rate unaffected by continuous operation, variations in supply voltage or fluctuations in temperature. They must also be able to reach the desired wire feed speed as rapidly as possible in order to give good and stable arc starting. The control for feed speed may be mounted on the torch or on the wire feeder.

While manual welding may use any of the systems mentioned, push-pull systems are becoming the standard method of wire feeding in robotic appli­cations because of the need for highly consistent feed speeds and defect - free arc starting.

7.2.2.3 Wire drive rolls

Aluminium wire is very much softer than steel and this can result in feeding difficulties, the wire being easy to deform by excessive roll pressure, causing the wire to jam in the feed conduit or in the contact tip. With push wire feeders any impediment to the wire feed, such as metal shavings or wire drawing soap compacted in the contact tip, kinks in the wire feed liner or spatter on the contact tip, may cause the wire to buckle within the wire feed conduit. Wire feed rolls must not be knurled but should be smooth, grooved rolls or, better still, one flat roll and one with a 60° V-groove. Wire feeding systems for aluminium also employ four drive rolls (Fig. 7.12) rather than the two rolls that conventionally are used to feed steel wires. It is impor­tant that the roll pressure is adjusted such that the wire is not grooved or flattened by the rolls since this will also lead to wire feeding problems. The wire should be kept as clean as possible. Covers to protect the reel from dust and heated cabinets are available and it is recommended that these are used where the highest quality is required. Also available are wire clean­ing devices comprising a cloth or felt pad clamped around the wire and soaked in a cleaning fluid such as alcohol or acetone. This can be used to remove grease, drawing soap and loose particles of swarf or oxide at the point at which the wire enters the conduit.

A relatively recent innovation in wire drive rolls is finding increasing use. This is the orbital welding system in which the wire passes through the hollow centre of the drive motor and is driven by a set of rolls set at an angle to and orbiting around the wire. This method of driving the wire has the advantages of both straightening and vibrating the wire, aiding in feeding the wire through the conduit.

7.2.2.4 Contact tip (tube)

The contact tip is a small but vital component in the welding power circuit. The tip is formed from a tube made to be a sliding fit for the wire. It is screwed into the torch head, ‘B’ in Fig. 7.10, and is the point at which the welding current is picked up by the filler wire. The contact tip is made from copper or brass and wears in use. It is therefore made to be replaceable. The tip for aluminium welding may vary in length from 25 mm to 100 mm. The longer contact tips provide the best current transfer conditions and therefore the most stable welding conditions. Tips have been designed that carry either a spring-loaded shoe to maintain a constant pressure on the wire or with the hole offset in order to force the wire against one wall, thereby improving and maintaining contact.

A worn contact tip may cause the wire to jam, resulting in a tangle at the wire drive rolls. A perhaps more serious weld quality problem may also arise from arc instability caused by the point at which the wire picks up the current moving up and down the contact tip. This effectively changes the wire stick-out length which in its turn affects the voltage, leading to arc instability and lack of penetration defects. Poor contact between the tip and the wire may cause arcing within the tip, giving rise to arc instability and perhaps wire feed problems. Damage to the tip from spatter, accidental touch-down or mechanical damage may cause similar problems.

The tip should be recessed in the gas shroud by at least 5 mm when welding in spray transfer. If the tip is too close to the end of the gas shroud there is an increased risk of spatter damaging the tip. If the tip protrudes from the shroud then there is a risk of the tip touching and melting into the weld pool. This will cause weld pool cracking, may give rise to ‘bird’s nesting’ and will require the tip to be replaced.