GB2051647A - An improvement in or relating to electric arc welding - Google Patents

Abstract

A method of electric arc welding using a non-consumable electrode (14) and including the step of imparting a modulated movement to the electrode (14) so that is moves in regular increments and produces a continuous weld constituted by a series of spot welds which overlap by a predetermined and regular extent, includes the further step of feeding a filler wire 24 to the weld pool, the movement of this feeding preferably being in step with the modulated movement of the electrode (14). The welding may be carried out in an inert atmosphere and electrode movement modulation may be accompanied by current modulation. <IMAGE>

Description

SPECIFICATION An improvement in or relating to electric arc welding.

The invention relates to electric arc welding and has for its object to provide an improvement therein.

In particular, the invention relates to the carrying out of programmed welding whereby welds of repeatable quality and profile can be made so that product quality is consistently high. The invention is particularly applicable to circumferential welding, for example the welding of tubes to tube plates in the manufacture of high pressure heat exchangers.

It is known to perform electric arc welding, using a so-called non-consumable electrode, in which the parts to be united are locally melted, forming a so-called weld "pool" which subsequently solidifies to produce the finished weld. However, such an autogenous technique can only be used where the materi als to be joined are metallurgically compatible.

In other cases a so-called "filler wire" of weld metal is used to supplement the weld "pool" and in this case the chemical composition of the weld metal, that is to say the so-called weld "pool", can be modified by the composition and amount of the filler metal which is added. It is to this latter technique that the invention relates.

According to one aspect of the invention, there is provided a mehod of electric arc welding, using a non-consumable electrode, which includes the step of imparting a modulated movement (as herein defined) to the electrode so that is moves in regular increments and produces a continuous weld constituted by a series of spot welds (as herein defined) which overlap by a predetermined and regular extent, and the further step of imparting a controlled feed movement to a filler wire so that a controlled quantity of filler metal is added to the continuous weld constituted by said series of spot welds.Preferably, a modulated feed movement (as herein defined) will be imparted to the filler wire, in step with the modulated movement of the electrode, so that filler wire is fed to the individual spot welds in turn. (By modulated movement of the electrode is herein meant a movement which is not constant but which is varied at regular intervals of time, either by the rate of movement being reduced to zero at these time intervals or by the rate of movement being varied at these time intervals. The term spot weld is to be taken to mean a weld which is concentrated at or in the region of a single point.By modulated feed movement of the filler wire is herein meant a feed movement which is not constant but which is varied at regular intervals of time, either by the rate of feed movement being reduced to zero at thes time intervals or by the rate of movement being varied at these time intervals). The degree of overlap expressed as a percentage of the size of each spot weld may be in excess of 10% and less than 90% and may be determined according to the composition and thickness of the material being welded and the electric power being used.

Other ranges of overlap may be employed, for example, ranges of, say, 25-75%, 40-60% and 45-55%. Depending upon the size of each spot weld, and upon the degree of overlap, and upon the size of the workpiece, the extent of angular movement of the electrode in successive increments may be variable from, say 0.25 to 20 . as an example of this, when welding a tube of say 31' diameter in a tube plate successive angular movements of the electrode may be of one degree, in linear measurement amounting to approximately six thousandths of an inch.Similarly, the extent of feed movement of the filler wire in successive increments, that is to say, to each spot weld if the filler wire is fed in incremental movements, may be variable, and it will be understood that the average filler wire feed rate will be chosen to maintain the correct volume of filler wire in the weld "pool" and this will depend on type of material being welded and on the joint geometry.

The time cycle, that is to say the periods of time between successive increments of movement of the electrode, will be determined according to the amount of incremental movement of the electrode (and the amount of feed of the filler wire if this is fed in incremental movements in step with the modulated movement of the electrode), the composition and thickness of the material being welded, and the electric power being used, and will generally lie within the range one tenth of a second to two seconds, although time cycles outside this range may be used, for example lying within the range one fiftieth of a second to four seconds.It may be found that other ranges of time cycle are appropriate to cover the time cycles which are useful in practice or for welding a particular type or thickness of material, for example cycle times within the ranges one twenty fifth of a second to three seconds, one fifth of a second to one second, one third of a second to three quarters of a second and two fifths of a second to three fifths of a second. The welding will generally be carried out in an inert atmosphere. The method may also include the modulation of electric current in combination with the modulated movements of the electrode, the current peaks taking place during the periods in which the electrode is moving at a reduced speed or is stationary.

According to a further aspect of the invention, there is provided a method of electric arc welding tubes of a heat exchanger to a tube plate (sometimes called a tube sheet) the welding being carried out according to the method described above.

According to a still further aspect of the invention, there is provided apparatus for performing the methods referred to, the apparatus including a holder for an electrode, means for conveying electric current to the electrode, means for imparting a modulated movement (as hereinbefore defined) to the holder so that it can impart movement to the electrode in regular increments to produce a continuous weld constituted by a series of spot welds (as hereinbefore defined) which overlap by a predetermined and regular extent, and means for guiding a filler wire and for imparting a controlled feed movement to said filler wire so that a controlled quantity of filler metal is added to the continuous weld constituted by said series of spot welds.Preferably, the means for imparting a controlled feed movement to the filler wire will be such as to impart thereto a modulated movement (as hereinbefore defined) in step with the modulated movement of the holder for the electrode so that filler wire is fed to the individual spot welds in turn.

In order that the invention may be fully understood and readily carried into effect, the same will now be described, by way of example only, with reference to the accompanying drawings, of which: Figure 1 is a diagrammatic illustration of apparatus embodying the invention for welding the tubes of a heat exchange to their tube plates, Figure 2 is a graph which illustrates the effects of various controls present in the apparatus of Fig. 1; Figure 3 is a diagrammatic illustration of further apparatus embodying the invention, again for welding the tubes of a heat exchanger to their tube plates; Figures 4, 5 & 6 are circuit diagrams of parts of the apparatus which will presently be referred to; Figure 7shows programme switches and display; and Figure 8 is a diagram illustrating a typical welding programme which can be employed using this further apparatus.

Referring now to Fig. 1 of the drawings, the apparatus there illustrated includes a drive shaft 10 carrying at one end a rotatable weld head 12 having an adjustable mounting for a non-consumable electrode 14. A welding power source with an associated ammeter A and voltmeter Vis indicated diagrammatically at 16 and connected to the electrode by means of a cable 18. An electric motor 20 is provided for rotating the drive shaft by way of a gearbox 22. As shown in chain-dotted lines, the electrode in the drawing is adjusted so that at one end it is located in close proximity to the periphery of a heat exchanger tube where the latter projects through a tube plate.

The arrangement is such that as the drive shaft rotates, moving the electrode around the periphery of the tube and in close proximity to the tube plate, the arc which is produced between the electrode and the work generates intense heat and locally melts the wall of the tube and the adjacent area of the tube plate so that the two are fused together.

The apparatus also includes a filler wire 24 which extends through the hollow interior of the drive shaft, a wire feed device indicated 26 and a rotatable spool 28 from which the filler wire is drawn off at a controlled rate by said feed device. Guide means (not shown) are provided on the weld head whereby the end of the filler wire remote from the spool 28 is guided towards the weld "pool", that is to say the molten metal of the heat exchanger tube and of the adjacent area of the tube plate, to augment said weld "pool". The extent by which the chemical composition of the weld metal, that is to say of the weld "pool", is modified by the addition of the filler metal is of course determined by the rate at which the feed device 26 is set to operate in relation to the rate of rotation of the electrode.Means (not shown) are provided for adjusting the rate at which the wire feed device operates.

Means are provided for controlling the electric motor 20 in a predetermined manner to drive the drive shaft with a modulated rotational movement, said means comprising a motor control translator 30 under the control of control logic generally indicated 32. The control logic includes three adjustable controls, that is to say an electrode speed control 34, a modulation cycle time control 36 and a modulation bias control 38. Programme logic 40 is associated with the motor control translator and with the control logic. A tachometer 42 is provided to show the speed of the drive shaft in degrees per second and a further dial indicator 44 is provided to show the angular position of the electrode at any particular instant (this being of use when the apparatus is used for bore welding when of course the electrode is hidden from view).

In operation, the electrode is brought into position adjacent a workpiece, for example adjacent a tube plate of a heat exchanger with the drive shaft 10 located coaxially of a tube which is to be welded in position in said tube plate, and the drive shaft is rotated so that the free end of the electrode is moved around in close proximity to the workpiece. When the electric current is switched on, an arc is produced between the electrode and the workpiece and is of sufficient intensity to locally melt the metal of the workpiece, that is to say, in the example illustrated in chain-dotted lines in Fig. 1 a portion of the tube and a portion of the tube plate through which the tube extends so that the tube and tube plate are fused together. Throughout the welding operation, the local melting of the workpiece produces a weld "pool" which moves around as the electrode moves around.Also throughout the welding operation, a controlled feed movement is imparted to the filler wire 24 and this supplements the weld pool.

During the welding operation, the movement of the electrode is modulated, that is to say is not constant but is varied at regular intervals of time, either by the rate of movement being reduced to zero at these time intervals or by the rate of movement being different from the general rate of movement at these time intervals. This modulated movement can be brought about by the controls 34, 36 and 38.

Fig. 2 is a graph which plots rate of movement in degrees per second against time. In a first part of the graph with no modulation the electrode is shown to be moving at a constant speed. At a time t1, however, modulation is shown to have been brought into effect so that initially, that is from t1 to t2, the rate of movement of the electrode has been doubled and then from t2 to t3 its rate of movement has been reduced. The modulation in this case is such that the amount of increase is equal to the amount of decrease, and consequently, at such intervals of time and in periods of "dwell" the parts of the work in the region of the free end of the electrode are heated to a greater degree than intermediate parts of the work.It will be understood therefor that the peaks of the graph represent periods of relatively fast movement of the electrode and that the lowermost portions of the graph represent the periods of "dwell" (which may or may not be periods of zero movement represented by the base line of the graph).

In a further part of the graph, the modulation of th electrode movement has "forward bias", that is to say, the amount of increase is greater than the amount of decrease and between the points t8 and t9 the electrode is not quite stationary although it is moving at a very much reduced rate so that the effect is substantially the same as before. In other works, the electrode is again moving at a rate which is not constant so that at regular intervals of time and in periods of "dwell" the parts of the work in the region of the free end of the electrode at those intervals of time are heated to a greater degree than intermediate parts of the work.

In a still further part of the graph, the modulation of the electrode movement has reverse bias" that is to say, the amount of increase is less than the amount of decrease and, for example, between the points t,2 and t,3 the electrode is actually moving in reverse, although at a very slow rate. Consequently, as before, at such intervals of time, the parts of the work in the region of the free end of the electrode are heated to a greater degree than intermediate parts of the work.

In each case the effect of this modulation is in many ways the same as if the electrode were moved at a constant rate with the electric power being increased and decreased at regular intervals of time; and weld penetration is increased locally in the regions where "dwell" occurs. In the result there is produced a continuous weld constituted by a series of spot welds which overlap by a predetermined and regular extent. In one important respect, however, the modulation of electrode movement is preferably to modulation of electric power even though the resulting weld may be similar. This is that changes in current affect the arc characteristics, and current induced magnetic fields are known to cause arc deflection.Consequently, modulation of the electric power can cause a physical shift in the position of the weld pool and the point of penetration, and in cases where the point of penetration is critical, for example, in the welding of tubes to tube plates of heat exchangers, it has been found advantageous to modulate the rate of electrode movement and to maintain the arc current constant.

It will be understood that when using the apparatus illustrated in Fig. 1 the general rate of movement of the electrode cn be adjusted by means of the control 34, the forward or rearward modulation bias can be adjusted by means of the control 38 and the cycle time between successive modulations can be adjusted by means of the control 36. The motor is in this case a so-called stepping motor and the arrangment is such that when the electrode has initially been moving at a constant rate, i.e. without modulation, the controls can be adjusted so that for example at a first modulated setting the electrode will move at twice the initial rate of movement for, say, one second and will then be stationary for an equal period of time. It will, however, be undestood that other types of electric motors may be used if desired.

However, referring now to Figs. 3 to 7 of the drawings, there is illustrated further apparatus by means of which welding can be carried out using both modulated electrode movement and modulated current control simultaneously. Referring in particular to Fig. 3, the further apparatus includes a conventional transducr controlled welding powr source 46 having an associated ammeter A and voltmeter Vand having a drooping output characteristic and also incorporating a high frequency arc starter 48 for arc initiation. A welding gun generally indicated 50 (and including a hollow drive shaft carrying a rotatable weld head; a non-consumable electrode; a filler wire and means for advancing the filler wire, all as in the previously described embodiment) is driven by a precision stepping motor 52.The motor 52 is energized from a power amplifier 54 which in turn is fed from a motor control translater 56. The translator 56 is controlled from command logic indicated by the chain line block 58. The command logic 58 includes programme logic 58a control logic 58b and operator controls 58c. The control logic has one output controlling a switch 60. An electrode speed control unit indicated by the chain line block 62 includes a voltage/frequency pulse generator 64 and a control potentiometer 66. The generator 64 is supplied from an output of the control logic 58b. An electrode movement modulation control is indicated by the chain line block 68 and includes a control potentiometer 70 and a modulation control unit 72. The control unit 72 has an input connected to the output of the electrode speed control unit 62.The output of the electrode speed control 62 is also connected to a tachometer 74 via an integrating circuit 76, the tachometer giving an indication of electrode speed throughout the operation of the apparatus. The output of the control 62 is also connected to a decimal counter 78 via a dividing circuit 80. A numerical read out 82 is connected to the decimal counter 78 and indicates the actual position of the electrode at any particular time (this again being of use if the apparatus is to be used for bore welding). The output of the control 62 is also connected to a gate 84 and a dividing circuit 86 both of which are connected to the change-over switch 60. The modulation control unit 72 is also connected to an input of the gate 84. A current control amplifier 88 is connected between the control logic and the welding power source 46.Feedback is provided from the welding power source to the control logic 58 and to the circuit control amplifier 88. Electrode adjustment towards or away from the work (as distinct from the modulated movement of the electrode) is provided by means of a cable C connected to the control logic 58b. The programme logic is connected to a programme socket 90 to which either a programme plug 92 or programme simulator 94 can be connected.

The d.c. power source has a current range of 25 to 180 amps with an open circuit voltage of 70 volts, the transductor control provides a low level input voltage control with linear input-output characteristics.

A wide electrode speed range and control flexibility is made possible by the use of a precision stepping motor; the shaft of the motor moves in discrete angular steps and has an inherent locking action and so allows a direct digital drive method to be employed. By including a digital counter operating from the motor control pulses it is possible to have a remote memory and display of the angular position of the motor shaft or welding electrode. The stepping motor has a permanent magnet rotor and a stator with two sets of centre tapped windings, the windings are energized in a particular sequensc the order of this sequence decides the direction of rotation and at each step in the sequence the rotor or shaft moves through 1.8 or 200 steps per revolution.Furthermore with steady state current in the windings the rotor has a substantial holding torque.

The motor translator module converts the incoming control pulses into the correct sequence to energise the stator windings of the motor. Forward and reverse input lines select the correct sequence for directional control and the frequency of the signal pulses to the translator module determine the actual speed of the motor. The speed ranges of the electrode traverse can be 0.2-12 rpm. Pulses of constant frequency to the motor translator will result in precise electrode movements of constant frequency.

When the electrode is stationary, maximum penetration occurs and when using the technique of modulated electrode movement it is found that the weld takes the form of a series of precisely positioned over-lapping spot welds. This technique has been found to have advantages when welding difficult materials such as Nickel alloys.

Referring to Fig. 4 there is shown one of four identical sections making up the motor power amplifer 54 in Fig. 3. The motor 52 has four stator windings 52a, 52b, 52c and 52d, one section of the amplifier 54 being connected with each winding so as to energize each winding individually.

The section shown is connected to the winding 52a. One terminal of the winding 52a is connected to the negative terminal of a power supply via a resistor R1. The other terminal of the winding 52a is connected to the collector of a PMP silicon power transistor TR1 whose emitter is connected to the positive terminal of the supply. The base of transistor TR1 is biased by means of resistor R2 connected between the base and positive terminal of the supply. The base control signal for the transistor TRI is supplied by an NPN drive transistor TR2 whose collector is connected to the base of transistor TR 1 via a resistor R3. The emitter of transistor TR2 is connected to the negative terminal of the supply. Transient voltages across the winding 52a are suppressed by a reverse biased diode D1 and resistor R4 in series between the terminals of the supply. The signal from the motor control translator 56 correponding to the required state of energization of the winding 52a is fed to the base of the transistor TR2 via resistor R5.

The winding 52a is either energized or deenergized depending on whether transistor TR2 does not conduct or conducts respectively. If the transistor TR2 conducts, transisor TOR 1 conducts and if transistor TR2 does not conduct then transistor TR1 does not conduct.

Referring to Fig. 5 there is shown that part of the command logic which regulates the arc current. The arc current is determined in relation to the electrode traverse by means of a current "slope" control. Two gates G1 and G2 are provided for increasing arc current and decreasing arc current respectively. The gates G1 and G2 are connected to, so as to control, switches S1 and S2 respectively. The switches S1 and S2 are connected in series with constant current supply circuit 96 and 98 respectively, these circuits 96 and 98 being connected to a stabilized power supply and being set by means of potentiometers R6 and R7 respectively. A capacitor 100 is connected between one terminal of the stabilized supply and a point between the two switches S1 and S2.The capacitor 100 can thus be charged through the switch S1 and discharged through the switch S2 at a rate determined by the value of the constant current through the circuits 96 and 98 respectively, these values being set by means of the potentiometers R6 and R7 respectively. Also the voltage across the capacitor 100 depends on the frequency and time duration of the switching of the switches S1 and S2. A buffer amplifier 102 and potentiometer 104 are connected in series, this series arrangement being connected in parallel with the capacitor 100.

The output of the potentiometer 104 is connected to the current control amplifier 88.

The gates G1 and G2 each have two inputs. One input to each gate comprises pulses derived from the generator 64 via the counter 78 and programme logic 58a. These pulses are fed to a multivibrator unit 106 which produces pulses of a substantially constant time duration which occur in synchronism with the incremental movements of the stepping motor. An "increase" signal is applied to the other input of the gate G1 and a "decrease" signal is supplied to the other input of the gate G2 when the programme requires the voltage across the capacitor 100 to be increased or decreased respectively. When the voltage is to be increased a steady "increase" signal is applied to the gate G1 and the capacitor 100 is charged in steps due to the multivibrator pulse being applied to the gate G1 which causes the switch to conduct a steady current for the duration of each pulse applied to the gate G1.This causes the voltage across the capacitor 100 to increase in regular steps, this increase being transmitted to the welding power source 46 so as to give a steady linear increase in arc current. For steady arc current no "increase" or "decrease" pulse is supplied to either of the gates G1 or G2.

Referring to Fig. 6 the current control amplifier 88 which is shown in detail is supplied with power via a transformer T1. The control signal output from the command logic 58 is connected to a terminal X of a phase shift control module 108 which is also connected to a winding 110 of the transformer T1.

Another winding 11 2 of the transformer is connected across a rectifier bridge 114 which provides a D.C. supply across the control winding of a transductor 116. The positive terminal of this D.C. supply is connected to earth. A silicon controlled rectifier, or thyristor 118 is connected in series with the control winding of the transducer 11 6 and the thyristor is controlled by the module 108. A reverse biased diode D2 is connected across the output of the rectifier bridge 114 to protect the thyristor 118 from reverse surge voltages. A pre-set variable resistor R8 is also connected in series with the control winding of the transductor. Feed-back connection is made from the terminal of the control winding connected to the resistor R8 to the module 108.

Connection across the D.C. output of the bridge 114 is made to the module 108. The output of the transductor 116 is connected to the welding power source 46.

Referring now to Fig. 7 the counter 78 is of conventional design having a counting capacity of 119 pulses. The pulses are supplied via the divider 80 and are 10 apart so that the readout 82 of the counter 78 has a fixed zero in the units readout. The counter thus counts over an electrode traverse range of 1 990' before resetting. Connections are made from the counter 78 to the programme logic which incorporates position switches two of which PS1 and PS2 are shown. Each switch such as PUS 1 comprises an AND gate 120 connected to and controlling a bistable 122. Each bistable 122 gives an output only when the number registered in the counter 78 corresponds to the connections made from the counter to the AND gate controlling that particular bistable.The bistables are triggered using the pulses from the divider 80 as clock pulses.

The operation of the apparatus shown in Fig. 3 will now be outlined. The programme plug 92 is plugged into the socket 90. The apparatus is then switched on at the operator controls 58c. The command logic then starts the welding procedure by initiating the arc via the high frequency arc starter in the welding power source 46. At the same time the command logic starts controlling the output of the generator 64. The pulse output from this generator drives the motor 52 via the dividing circuit 86 if the modulation is OFF or via the gate 84 if the modulation is ON, the modultion being ON or OFF depending on the state of the switch 60 which is controlled by the control logic 58b. As the welding progresses the counter 78 supplies the command logic 58 with information on electrode position.

The command logic 58 controls the generator 64, the switch 60, the arc current and the electrode adjustment according to the programme, the different steps in the programme being effectively triggered by the counter 78.

The modulation control depends not only on the potentiometer 70 but also on the output pulse frequency from the generator 64. This output pulse frequency is integrated with respect to time by the circuit 76 and the resulting analogue signal representing motor speed is displayed on the tachometer 74. The arc current feedback to the amplifier 88 and the command logic 58 gives a check on arc current and allows automatic correction of the arc current as the welding proceeds.

The average speed of the motor is controlled by the pulse generator 64 with a linear voltage/frequency characteristic; the nominal speed can be set by the potentiometer 66 and further speed adjustments can be made by low level signal inputs from the command logic 58 thus allowing weld penetration adjustments related to the position of the electrode around the periphery of the weld. The modulating action is produced by a pulse dividing technique in which the "gate" 84 allows a certain number of pulses to pass through and then inhibits a number of pulses to produce the pattern shown at the left hand side of Fig. 2 between t1 and t3.Typical values of "Weld Pitch" that can be preselected by means of the modulation control 68 and/or the electrode speed control unit 62 are 2.5 , 5 , 10 , 20 , 40". these values produce circumferential distances of approximately 0.016", 0.033", 0.066", 0.132", and 0.264" respectively for a diameter of i".

The modulation in effect reduces the average motor speed and in order to preserve the same average speed when switched to unmodulated operation, the control pulses are divided by means of the circuit 86 by a factor depending on the modulation ON/modulation OFF ratio of motor speed.

Referring now to Fig. 8, this illustrates a typical welding programme which can be employed using the apparatus just described, electrode speed in r.p.m. and arc current in amps being plotted against angular position of the electrode.

The programme includes: a) A slow initial speed to provide the necessary heat build-up.

b) An increased speed for the remaining section, with a suitable overlap at 360 .

c) The current is gradually reduced to the manimum level with the electrode speed unmodulated.

d) The electrode is accelerated and when at maximum speed, the welding current is switched off.

The zones at (c) and (d) are important in order to produce the ideal weld termination.

The current down slope gradually reduces weld penetration, this is further reduced by accelerating the electrode and the tendency to produce a crater at the point where welding current is switched off is therefore eliminated.

A suitable programme can be established in a test procedure by the use of a "programme simulator" which plugs into the programme socket. This unit is essentially a pluggable patch board which permits the selection of parameter changes and the angular position at which these changes in the zone (c) occur.

The programme positions can be selected from zero to the full capacity of the digital counter (1 990') in increments of 10 , changes in electrode traverse speed; modulation; arc current and when necessary electrode position are possible. When the programme giving optimum results has been established, a programme plug can be wired with suitable links and so provide an instant and repeatable weld procedure for future use.

It will be understood that whatever the mode of modulated operation of the electrode, the filler wire may be fed to the weld "pool" at a controlled but constant rate. However, it is thought preferable that the filler wire should be fed to the weld "pool" with a modulated movement in step with the modulated movement of the electrode, that is to say so that the filler wire feeds into the weld "pool" during the intervals of "dwell". This will require suitable interconnection of the wire feed device 26 with the means provided for controlling the electic motor 20 for driving the drive shaft 10 with a modulated rotational movement.

Various other modifications may be made without departing from the scope of the invention, particularly to the control apparatus for varying the rate of movement of the electrode.

For example, the electrode could be driven by a constant speed motor through variable speed gearing and Geneva mechanism. The control apparatus for feeding the filler wire to the weld "pool" may also be arranged to operate in a variety of ways, being either independent of the means for driving the electrode or coupled therewith so that the filler wire feed, or a substantial part of the filler wire feed, is brought about during the periods of "dwell" of the electrode. For example, in this latter case pulsing movements of the filler wire may be achieved by a stepping motor drive arrangement, and this could conveniently be energised by the same circuitry as the electrode drive motor to provide the necessary synchronisation.

Claims (12)

1. A method of electric arc welding, using a non-consumable electrode, the method including the step of imparting a modulated movement, as hereinbefore defined, to the electrode so that it moves in regular increments and produces a continuous weld constituted by a series of spot welds, as herein be fore defined, which overlap by a predetermined and regular extent, and the further step of imparting a controlled feed movement to a filler wire so that a controlled quantity of filler metal is added to the continuous weld constituted by said series of spot welds.

2. The method accoring to claim 1, wherein a modulated feed movement, as here inbefore defined, is imparted to the filler wire, in step with the modulated movement of the electrode, so that filler wire is fed to the individual spot welds in turn.

3. The method according to either one of the preceding claims, in which the degree of overlap expressed as a percentage of the size of each spot weld is in excess of 10% and less than 90%, determined according to the composition and thickness of the material being welded and the electric power being used.

4. The method according to any one of the preceding claims, in which the extent of angular movement of the electrode in successive increments is variable from 0.25 to 20 depending upon the size of each spot weld and upon the degree of overlap and upon the size of the workpiece.

5. The method according to any one of the preceding claims, in which the extent of feed movement of the filler wire in successive increments, that is to say, to each spot weld if the filler wire is fed in incremental movements, is variable the average filler wire feed rate being chosen to maintain the correct volume of filler wire in the weld "pool" depending on the type of material being welded and on the joint geometry.

6. The method according to any one of the preceding claims, the welding being carried out in an inert atmosphere.

7. The method according to any one of the preceding claims, including the step of modulating the electric current in combination with the modulated movements of the electrode, the current peaks taking place during the periods in which the electrode is moving at a reduced speed or is stationary.

8. A method of electric arc welding tubes of a heat exchanger to a tube plate (sometimes called a tube sheet) the welding being carried out using the method claimed in any one of the preceding claims.

9. Apparatus for electric arc welding using a non-consumable electrode, the apparatus including a holder for the electrode, means for conveying electric current to the electrode, means for imparting a modulated movement, as hereinbefore defined, to the holder so that it can impart movement to the electrode in regular increments to produce a continuous weld constituted by a series of spot welds, as hereinbefore defined, which overlap by a predetermined and regular extent, and means for guiding a filler wire and for imparting a controlled feed movement to said filler wire so that a controlled quantity of filler metal is added to the continuous weld constituted by said series of spot welds.

10. Apparatus according to claim 9, in which the means for imparting a controlled feed movement to the filler wire are such as to impart thereto a modulated movement, as hereinbefore defined, in step with the modulated movement of the holder for the electrode so that filler wire is fed to the individual spot welds in turn.

11. A method of electric arc welding using a non-consumable electrode, the method being substantially as hereinbefore described with reference to the accompanying drawings.

12. Apparatus for electric arc welding using a non-consumable electrode, substantially as hereinbefore described with reference to and as illustrated by the accompanying drawings.