JPH0127826B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a control device for a pulsed arc welding power source that performs welding while feeding a consumable electrode. The pulsed arc welding method, in which a consumable electrode (hereinafter referred to as a wire) is fed to a welding area for welding, generally transforms the molten droplets at the tip of the wire into fine particles for smooth transfer (hereinafter referred to as spray transfer). )
By keeping the arc length constant, a stable arc is generated, a uniform welding result is obtained, and a welded product with less spatter is obtained. This wire is supplied with a base current that is low enough to prevent spray transfer, that is, below a critical current value, and a pulsed current that periodically becomes a high current that exceeds the critical current value, thereby causing spray transfer.

The effect of pulsed current in such pulsed arc welding on welding results will be discussed. In pulsed arc welding, the pulse current elements to maintain smooth spray transition are the pulse current peak value Ip and pulse duration (hereinafter referred to as
This is called pulse width. )t _p . Among these, the pulse current peak value Ip is Ip ² from the tip of the molten wire.
This is the most important element that has the effect of detaching droplets by a pinch force proportional to the pulse current value.
As Ip increases, fine droplets move regularly,
Conversely, when the pulse current value Ip becomes small, large droplets begin to move irregularly. Therefore, unless the pulse current value Ip is kept constant, smooth spray transfer cannot be achieved. Furthermore, a predetermined pulse width t _p is also required as a condition for separating the melt from the tip of the molten wire, but since the separation of the droplet is due to the pinch force, the pulse width t _p is determined by the above-mentioned pulse current value. This is an important requirement that affects droplet detachment in relation to Ip. By selecting appropriate values for these, the cycle of the pulse current and the cycle of droplet transfer can be matched, and smooth spray transfer can be achieved.

In arc welding using pulsed current as described above, the peak value Ip of the pulsed current and the pulse width _tp , which is its duration, are important requirements. Among these, setting the pulse peak value Ip to a predetermined value is
This can be achieved relatively easily by using methods such as using a pulse current power supply with approximately constant current characteristics, or using feedback control to suppress the pulse peak value to a constant current when extracting pulse current from a DC power supply with constant voltage characteristics. . However, the pulse width _tp is usually determined by controlling the time from its rise to its fall. However, the actual pulse current waveform is strongly influenced by the power supply voltage, the impedance of the output circuit cable of the welding machine, the arc length, the wire protrusion length, etc. The effective width changes, and stable spray transfer in which pulse current and droplet transfer are synchronized cannot be expected.

The cause of this will be explained using a diagram. FIG. 1 is a connection diagram showing a conventional example of a control device for this type of pulse arc welding power source. In the same figure, 1 is a DC power supply for supplying pulse current, and 2 is a DC power supply 1.
3 is a DC reactor, and 4 is a DC power supply for supplying a base current sufficient to maintain an arc to the welding current. Reference numeral 5 denotes a pulse waveform control circuit that outputs a signal for determining the conduction amount and conduction period of the current control element 2, and a pulse oscillation circuit that generates pulses with a constant amplitude and frequency is usually used. Reference numeral 6 denotes a drive circuit for converting the output of the pulse waveform control circuit 5 into a signal suitable for controlling the current control element 2. Also 31,
32 is an output terminal, and 10 is a flywheel diode.

In the device shown in FIG. 1, the current control element 2 becomes conductive in response to the output signal from the pulse waveform control circuit 5. As a result, a pulse current corresponding to the output signal from the pulse waveform control circuit 5 is superimposed on the output current of the base current DC power source 4 in the wire 7 and the workpiece 8 connected by the output cable between the output terminals 31 and 32. The current is output and pulse welding arc 9
will occur. In this way, in the conventional device, the amount of conduction of the current control element is controlled so that the pulse width is determined by the pulse waveform control circuit 5. will be examined in detail. Figure 2a shows the equivalent circuit on the right side of the load from the output terminals 31 and 32 of the device in Figure 1, where r _l and L _l are the resistance components of the output circuit from the output terminal to the arc generation point, which is the welding load. and the inductance component, and Ea and r _a are the welding voltage of the arc load and the resistance at the protruding length portion of the consumable electrode wire beyond the power supply part. Further, the inductance of the reactor 3 connected in series to the pulse current circuit is assumed to be L ₀ .

Consider a case where, in such a circuit configuration, a rectangular waveform drive signal having a width _tp as shown in FIG. 2c is supplied from the pulse waveform control circuit 5 to the current control element 2. At this time, if the output voltage of the pulse DC power supply 1 is E ₁ , and the resistance and inductance of the output circuit are R = r _l + r _a and L = L _l + L ₀ , the pulse output current Ip is shown in Fig. 2b. As shown, the current rises along the curve E ₁ -E _a /R(1-ε ^-R/L t) toward the steady state current (E ₁ -Ea)/R with a rising time constant L/R. This current stops increasing at time _t1 when it reaches the peak current Ip determined by the peak value of the output signal of the pulse waveform control circuit 5, and is maintained at a substantially constant value thereafter. When the output of the pulse waveform control circuit 5 becomes zero at time t ₃ after time t _p , _the current control element 2 is cut off, and the electromagnetic energy (1/2L ₀ Ip ² ) is emitted through the flywheel diode 10. Here, consider a case where the power supply voltage fluctuates and the output voltage of the DC power supply 1 drops from _E1 to _E2 . At this time, the pulse current is
As shown by B in Figure b, since the steady current (E ₂ - Ea)/R rises with the same time constant L/R, which is lower than the previous value, its rise speed becomes slower and the set current Ip becomes lower. The time t ₂ reached is later than t ₁ in the case of A where the power supply voltage is high. On the other hand, the output of the pulse waveform control circuit 5 to the current control element 2 becomes zero at time t ₃ regardless of fluctuations in the power supply voltage, and the current decreasing speed is 1/2L ₀ Ip of the electromagnetic energy stored in the reactor 3. Since it is the same value as ² , it will not change. As a result, the time that the set current Ip is held becomes shorter even though the command signal is the same, and the true pulse time width that is effective for spray transfer becomes narrower and becomes insufficient. . As can be seen from the above equation, the rate of rise of the pulse current is related to the cable length as well as the power supply voltage.
Welding voltage determined by r _l , L _l , arc length
It is similarly affected by each component of Ea and r _a related to the wire protrusion length. Therefore, the pulse current waveform changes its appearance from moment to moment, and no matter how accurately the waveform of the command signal from the pulse waveform control circuit 5 is constant, it is only the bottom of the output pulse current waveform, that is, the rising edge of the pulse current. Since it only controls the pulse current from the time to the moment of falling, it cannot regulate the width of the pulse current above the critical current that is effective for spray transfer, and the droplets of the consumable electrode synchronize with the period of the pulse current. It is impossible to perform welding with smooth spray transfer.

Furthermore, as a result of experiments using pulsed current with a waveform close to a rectangular wave, the present inventors found that in order to spray droplets, the pulse frequency is irrelevant, and the pulse current value and pulse time are We found that the condition is that a specific relationship holds true. Figure 3 shows the relationship between pulse frequency, time, and current on spraying in aluminum and stainless steel welding, with the vertical axis representing the pulse current Ip(A) and the horizontal axis representing the pulse time width. t _p (ms), with material, wire diameter, and pulse frequency as parameters. In the figure, 1 to 4 are stainless steel, A and B are aluminum, 1 and A are 1.6 mm, and 2 and B are 1.2 mm.
mm, 3 indicates when using a wire with a diameter of 0.8 mm, and 4 indicates when using a wire with a diameter of 0.6 mm. In addition, the 〇 and □ marks indicate when the pulse frequency is 25 Hz, and the + mark indicates when the pulse frequency is 50 Hz. The area above the curve in the figure shows the area where the spray transfer occurs, and the area below the curve shows the area where the drop transfer occurs. As can be seen from the figure, the pulse frequency is almost irrelevant as a condition for spray transfer, and the pulse time t _p and pulse current Ip are (Ip - I _c ) ⁿ・t _p = A... (1) It becomes a hyperbolic shape showing the relationship. Here, I _c is a critical current for forming droplets into a spray, and n and A are constants determined by the material and diameter of the wire used. Therefore, in order to obtain stable spray transfer, instead of controlling the pulse peak value and pulse time width to maintain them at predetermined values as in the conventional method, it is necessary to control (Ip−I _c ) ⁿ・t _p to be equal to or greater than a predetermined value. It is conceivable that it may be controlled so that

However, it is difficult to make the pulse current waveform into a perfect rectangular waveform as in the above experiment, and it usually becomes a trapezoidal waveform that increases with a certain time constant as explained in FIG. Therefore, the experimental formula (1) above should hold for each instantaneous value of the pulse current that is greater than the critical current, so (Ip−I _c ) ⁿ t _p in formula (1) is the instantaneous value of the pulse current. When i is (i-I _c ), it can be replaced by the time integral of ⁿ . Therefore, ∫(i-I _c ) ⁿ dtA (2) becomes the condition for turning the droplets into a spray. In the above equation (2), the critical current I _c is a value determined mainly by the diameter of the wire used, but instead of this I _c , it is smaller than the peak value Ip of the pulse current, and the arc is It is considered that the objective can be generally achieved by determining a value Ix that can be clearly distinguished from the base current that is passed for maintenance.

Therefore, in the present invention, the pulse current starts flowing at a predetermined frequency, and the pulse current becomes the reference current.
This system enables reliable spray transfer of droplets by stopping the flow of pulse current when the time integral value ∫ (i - Ix) ⁿ dt of the period exceeding Ix exceeds the reference value k. be.

FIG. 4 is a connection diagram showing an example of a device implementing the present invention. In the figure, parts having the same functions as those of the apparatus shown in FIG. 1 are given the same reference numerals. 11 is a current detector for detecting the output current i, 12 is a reference current setter, and 13 is a difference signal i- when the output i of the current detector 11 and the output Ix of the reference current setter 12 are compared and iIx is reached. 14 is a pulse frequency control circuit consisting of an oscillator for determining the frequency of the pulse current; 15 is an integrating circuit which receives the output of the comparator 13 and calculates s=∫(i-Ix) ⁿ dt
Perform the calculation. Here, n is a positive number, and as mentioned above, the value is in the range of 1 to 3 depending on the material of the wire used for welding. 16 is a reference value setting device for setting the reference value k of the integral value of the integrating circuit 15;
A comparator 7 compares the output s of the integrating circuit 15 with the output k of the reference value setter 16, and generates an output when sk is reached. 18 is a comparator 17 from the time when the output signal of the pulse frequency control circuit 14 is received.
This is a pulse signal generation circuit that outputs a continuous signal until the pulse frequency control circuit 14 generates an output. It will be done. The pulse DC power supply 1' is a power supply whose output current is always set to a constant pulse peak value by a pulse peak current setting device 19, and the pulse DC power supply 1' is connected to the current control element 2'. A switching element is used to open and close the output. The operation of the apparatus shown in FIG. 4 will be explained with reference to the explanatory diagram shown in FIG. When the output of the pulse frequency control circuit 14 outputs the signal p at a predetermined period as shown in FIG _.
will occur. The drive circuit 6 makes the switching element 2' conductive by this output _tp , and the output of the pulse DC power source 1' is supplied to the welding part. The output current is detected by a current detector 11 and compared with a set value Ix of a reference current setter 12. The pulse current rising from time 0 increases according to the time constant of its output circuit, and when it reaches iIx, the subtracter 13 outputs the difference (i-
Ix) to start the integrating operation of the integrating circuit 15. The integral value s of the integrating circuit 15 is determined by the reference value setting device 16.
When the set value k or more is exceeded, the comparator 17 issues a signal, and the pulse signal generating circuit 18 stops its output based on the output of the comparator 17. Also, comparator 17
The integration circuit 15 is reset by the output of . The switching element 2' is cut off and remains cut off until the next output from the pulse frequency control circuit 14 is supplied.

In the device shown in Figure 4, fluctuations in the power supply voltage, cable length, welding voltage, wire protrusion length, etc., cause fluctuations in the impedance of the output circuit, and the rise speed of the pulse current changes as shown by the solid line a in Figure 5. Consider the case where the speed decreases from the speed shown by to the speed shown by the broken line. First, if the rising speed of the pulse current starting from time 0 is fast as shown by the solid line, then the current i
reaches the set value Ix of the reference current setter 12 at a relatively early time _t1 . The integrating circuit 15 starts integrating (i-Ix) ⁿ from time _t1 , and the comparator 17 starts integrating (i-Ix)n at time _t3 when the integral value s reaches the set reference value, as shown in FIG. 5c. Since the output is generated, the output of the pulse signal generating circuit 18 continues for a period _tp1 from time 0 to _t3 , as shown in FIG. 5d. When the output signal of the pulse signal generation circuit 18 disappears at time _t3 , the switching element 2' is cut off, so that the output current i is generated by first releasing the electromagnetic energy stored in the reactor 3 through the flywheel diode 10, and at a certain time. Decrease by a constant. On the other hand, when the rising speed is slow as shown by the broken line, the current i reaches the set value Ix at time _t2 later than in the case of the solid line, but the integrating circuit 15
Integration is started from t ₂ and at time t ₄ when the integral value s reaches the reference value k, the comparator 17 generates an output. As a result, the switching element 2' continues to conduct from time 0 to _t4 . Therefore, if the reference current Ix is set to a value greater than or equal to the value effective for spray transfer of droplets, a pulse current with a time product necessary for stable spray transfer will always be obtained. It is desirable to determine this reference current Ix within a range that is higher than the critical current at which the droplets spray and does not exceed the peak value of the output current, but even if the value is smaller than the critical current, there is a clear difference between the base current and the base current. If there is a difference, a considerable effect can be expected.

In this way, by controlling the peak current Ip and the time integral of the pulse current exceeding the reference value to be constant, stable welding is possible. However, in arc welding, the arc length may change due to unevenness of the workpiece, fluctuations in the position of the welding torch, or slight discrepancies between the wire feeding speed and the amount of wire melting. On the other hand, if the peak current Ip and time integral value are controlled constant, the average value of the current used to heat and melt the wire will also be kept constant against these changes in arc length, so it depends on the power supply characteristics. There is no self-control of arc length. As a result, once the arc length changes due to any of the causes mentioned above, it cannot be recovered. Therefore, for example, if there is a slight slowdown in the wire feeding speed relative to the wire melting speed, this will immediately grow and lead to wire burnback or protrusion. In order to prevent such a phenomenon, the welding voltage Ea is detected as shown by the dashed line in FIG. It should be controlled accordingly. That is, if the pulse frequency is increased while keeping the time integral of the pulse current constant, the average value of the welding current will increase, and the amount of wire melting will also increase accordingly.
Therefore, as the arc length becomes shorter and the welding voltage becomes lower, increasing the pulse frequency inversely proportional to this will increase the average value of the pulse current, which increases the amount of wire melting and reduces the arc. Recovery based on length. Conversely, when the arc length becomes longer and the welding voltage becomes higher, if the frequency is reduced in inverse proportion to this, the amount of wire melting will decrease and the arc length will return to its original value. In addition, the average value of the welding current corresponds to the welding voltage while keeping the pulse frequency constant by detecting the welding voltage and using it as a command signal for the reference value setting value 16 of the comparator 17, as shown by the broken line in Fig. 4. Since the arc length can be adjusted by changing the reference value of the integral value of the pulse current, automatic control of the arc length is also possible with this method.

Further, the wire feeding speed is generally determined depending on the diameter of the wire and the value of the welding current used. Therefore, it is necessary to adjust the wire feeding speed and the average value of the pulse current in conjunction with each other. In this case, it is also possible to uniformly adjust the welding conditions by changing the time integral or pulse frequency of the pulse current approximately in proportion to the wire feeding speed. At this time, self-control of the arc length due to power supply characteristics cannot be expected in response to changes in the arc length, so the welding voltage is detected and the pulse frequency or time integral of the pulse current is controlled in approximately inverse proportion to this. Therefore, it becomes possible to automatically control the arc length.

Fig. 4 shows an example of a method in which a power supply whose peak value is controlled to a constant value is used as a pulse supply DC power supply, and the output of this DC power supply is opened and closed by a switching element connected in series to the output circuit. However, the pulse supply DC power supply is not limited to this, and a DC power supply with substantially constant voltage characteristics may be used, and in this case, both the pulse period and the peak value may be controlled by a current control element. FIG. 6 is a connection diagram showing an embodiment of a system in which both the pulse period and the peak value are analog-controlled by a current control element. In this figure, parts having the same functions as those in FIGS. 1 and 4 are given the same reference numerals. Further, the current detector 11' is different from the case shown in FIG. 4 and is a current detector that detects only pulse current. 20 is current detector 1
This is a first comparator that compares the output i of 1' and the output Ip of the peak current setting device 19, and for example, a subtracter that obtains the difference (Ip-i) between both signal values is used. 21 is the output i of the current detector 11' and the reference current setter 1
A subtracter that compares the output Ix of 2' and generates a difference output (i-Ix) only when iIx, 22 is a subtracter 21
An integration circuit 23 integrates the output of and obtains s=∫(i-Ix) ⁿ dt, and 23 compares the output s of the integration circuit 22 with the output k of the reference value setter 16 and generates an output when sk. This is the second comparator. 24 is a signal synthesis circuit which inputs the outputs of the first and second comparators 20 and 23 and the output of the pulse frequency control circuit 14; Using the subtracter obtained, the peak value is determined by this difference signal, and a pulse command signal is output that starts from the rising edge of the output signal of the pulse frequency control circuit 14 and ends when the output signal of the second comparator 23 is generated. The current is supplied to the drive circuit 6 to control the conduction of the current control element 2. This signal synthesis circuit may be of any type as long as it has the above-mentioned functions, and an example thereof is shown in FIG. 7th
The figure is a connection diagram showing the signal synthesis circuit 24 of the apparatus shown in FIG. 6 together with its input/output sections. In the figure, 25 is the pulse frequency control circuit 1
26 is a flip-flop circuit that is set by the rising edge of the output of 4 and reset by the output of the second comparator 23, and 26 is a flip-flop circuit that connects the output signal (Ip-i) of the first comparator to the high level of the output of the flip-flop circuit 25. This is an analog switch circuit that allows the signal to pass through and transmits the signal to the drive circuit 6.

FIG. 8 is a connection diagram showing an embodiment of a signal synthesis circuit configured to perform control using a switching method, in which the first comparator 20' compares the peak current setter Ip and the detected current value i. An example will be shown in which a circuit that generates a signal that makes the current control element conductive only when i<Ip is used. In the same figure, 27
is an AND circuit that receives the outputs of the first comparator 20' and the flip-flop circuit 25, and makes the switching element 2 conductive only while the flip-flop circuit 25 is generating an output and the pulse current is smaller than the reference current Ip. Output a signal. The operation of the embodiment shown in FIG. 8 will be explained. Generally, a comparator always has a finite dead zone. Therefore, when the output current i increases from a value lower than the reference value Ip, the first comparator 20' outputs a value Ip ₁ that is slightly higher than the reference value Ip.
When the value becomes , an output is generated, and conversely, when decreasing from a value higher than the reference value, an output is generated where i>Ip until the value Ip ₂ , which is slightly lower than the reference value Ip, is reached. and the difference between these operating currents Ip ₁ and Ip ₂ ΔIp = Ip ₁
−Ip ₂ is always a finite value. As a result, the current control element 2 changes the output current i from a small value to i=Ip ₁
Continuity continues until . The current control element 2 is i=
When the circuit is shut off at Ip ₁ , the electromagnetic energy stored in the reactor 3 until then is released, and the output current decreases from the value at Ip ₁ . Output current decreases to Ip ₂
When the current control element 2 becomes conductive again, i=
Increases up to Ip ₁ . In this way, the output current i
increases and decreases repeatedly between Ip ₁ and Ip ₂ , and is maintained at the reference value Ip as an average value.

When using the circuit shown in Figure 8 as a signal synthesis circuit, the current control element 2 is used not as an analog element but as a switching element, which not only enables more stable control but also reduces loss during switching. Since it is only necessary to bear the load, the capacitance of the element can be made small.

Furthermore, in the signal synthesis circuit shown in FIG. It may also be input together with the signal to perform chopper control on the output signal.

In the embodiment shown in Fig. 6, as in Fig. 4, the pulse current exceeding the reference current Ix is integrated regardless of the rise speed of the pulse current, so that spray transfer can be controlled regardless of fluctuations in various external conditions. Since the total amount of pulse current after it becomes effective is always constant, stable welding can always be performed. Of course, in the embodiment shown in FIG. 6, either a signal corresponding to the welding voltage is obtained and this is used as the frequency command for the pulse frequency control circuit 14, or it is supplied to the reference value setting circuit 16 and the second comparator By setting the reference value of 23, it is possible to automatically recover from fluctuations in welding voltage due to fluctuations in arc length. Furthermore, if a reference value k, which is the pulse frequency or the time integral value of the pulse current, is determined in accordance with the wire feeding speed, it becomes possible to centrally adjust the wire feeding amount and the average value of the pulse current. . In this case, in order to automatically control changes in the arc length, it is sufficient to detect the welding voltage and control the reference value k or the frequency in substantially inverse proportion to the welding voltage.

Further, in each embodiment, the drive circuit 6 is not particularly required if the output of the pulse signal generation circuit or signal synthesis circuit in the previous stage is sufficient. In addition, the current detector may be one that detects the entire output current including the base current, or one that detects only the pulse current, and in each case, when setting the reference current Ix, this base current is Just consider it. Further, when a small current signal for causing a base current to flow through the drive circuit 6 is superimposed on the output signals of the pulse signal generation circuit and the signal synthesis circuit, the DC power supply 4 for supplying the base current can be omitted. In this case, in Fig. 6, the current detector 11'
If the detector is configured to detect the entire output current including the base current as in the example shown in FIG. 4, it is possible to control the base current to a constant value by feedback control even when the base current is present.

As described above, according to the present invention, the integrated value of the pulse current, which is the most important requirement in pulsed arc welding, is controlled as an integrated value at a value that is effective for actual spray transfer of droplets. Even if the impedance of the output circuit changes due to fluctuations in voltage, changes in output circuit impedance due to replacement of connected output cables, or changes in arc voltage, the required value can always be maintained, and droplets can be accurately synchronized with the pulse current. Spray transfer is possible and extremely stable welding can be performed.

[Brief explanation of drawings]

Figure 1 is a connection diagram showing the control method in conventional pulsed arc welding, Figures 2 a to c are explanatory diagrams to explain the operation of the device in Figure 1, and Figure 3 is the effect on spraying of droplets. Fig. 4 is a connection diagram showing an embodiment of the present invention, and Figs. 5 a to f show the operation of the embodiment of Fig. 4. An explanatory diagram to explain the
FIG. 6 is a connection diagram showing another embodiment of the present invention, and FIG.
8 and 8 are connection diagrams showing examples of the internal structure of the signal synthesis circuit in the embodiment of FIG. 6, respectively. 1...DC power supply, 2...Current control element, 2'...
...Switching element, 11, 11'... Current detector, 14... Pulse frequency control circuit, 15, 22
...Integrator circuit, 16...Reference value setter, 17...
Comparator, 18... Pulse signal generation circuit, 19...
Peak current setter, 20, 20'...first comparator, 21...subtractor, 23...second comparator, 2
4... Signal synthesis circuit, 25... Flip-flop circuit, 26... Analog switch circuit, 27...
AND circuit.

Claims (1)

[Claims] 1. In pulsed arc welding, which is performed while feeding a consumable electrode to a welding part at a predetermined speed, a DC power source that outputs a DC current of a predetermined peak value, and an output switch for switching the output of the DC power source are provided. A switching element that generates a pulsed current, a reference current setter that is set to a value Ix that does not exceed the peak value of the output current, a current detector that detects the output current i, and a switching element that repeatedly applies the pulsed current at a predetermined period. a pulse frequency control circuit for
A subtracter that compares the output signal Ix of the reference current setter and the output signal i of the current detector and outputs a difference signal i-Ix when the output signal iIx is equal to iIx, and integrates the output of the subtracter ∫( i-Ix) ⁿ dt (n is a positive number); a reference value setter that sets a predetermined reference value k for the output of the integration circuit; and an output of the integration circuit and the reference value. A comparator that compares the output of the setting device and generates an output when ∫(i-Ix) ⁿ dtk, and a comparator that generates an output when the output signal of the pulse frequency control circuit is received until the comparator generates an output. A control device for a power source for pulsed arc welding, comprising: a pulse signal generation circuit that outputs a continuous signal; and a drive circuit that conducts the switching element in accordance with an output signal of the pulse signal generation circuit. 2. The control device according to claim 1, wherein the pulse frequency control circuit is a circuit that generates a pulse with a frequency corresponding to a welding voltage. 3. The control device according to claim 1, wherein the pulse frequency control circuit is a circuit that generates pulses at a frequency corresponding to the feeding speed of the consumable electrode. 4. The control device according to any one of claims 1 to 3, wherein the reference value setting device is a setting device that outputs a signal having a value corresponding to a welding voltage. 5. The control device according to any one of claims 1 to 3, wherein the reference value setting device is a setting device that outputs a signal having a value corresponding to a feeding rate of the consumable electrode. 6 In pulsed arc welding performed while feeding a consumable electrode to a welding part at a predetermined speed, a DC power source, a current control element that controls the output of the DC power source in a pulsed manner, and a peak value Ip of the output current are set. a reference current setter that is set to a value Ix that does not exceed the peak value set by the peak current setter, a current detector that detects the output current i, and a pulse current a pulse frequency control circuit for repeating the circulation at a predetermined period;
a first comparator that compares the output signal Ip of the peak current setter and the output signal i of the current detector;
a subtracter that compares the output signal Ix of the reference current setter and the output signal i of the pulse current detector and outputs a difference signal (i-Ix) when the output signal Ix reaches iIx; and a subtracter that integrates the output of the subtracter. an integrating circuit that obtains ∫(i-Ix) ⁿ dt (n is a positive number), a reference value setting device that sets a predetermined reference value in advance, and an output of the integrating circuit and an output of the reference value setting device. A second comparator that outputs an output when the output of the comparison and integration circuit exceeds the output of the reference value setting device, and the outputs of the first and second comparators and the output of the pulse frequency control circuit are inputted. Pulse arc welding, comprising: a signal synthesis circuit that outputs a signal that determines a pulse current value and a pulse width; and a drive circuit that conducts the current control element in accordance with the output signal of the signal synthesis circuit. Control device for power supply. 7. The first comparator is a subtracter that obtains a difference signal (Ip-i) between the output signal Ip of the peak current setter and the output signal i of the current detector, and the signal synthesis circuit a pulse time control circuit that outputs a continuous signal from the time it receives the output signal of the pulse frequency control circuit until it receives the output signal of the second comparator; and the pulse time control circuit generates an output. 7. The control device according to claim 6, further comprising an analog switch circuit that transmits the output of the first comparator to the drive circuit during a period in which the first comparator is in the driving circuit. 8 The first comparator is a comparator that compares the output Ip of the peak current setter and the output signal i of the current detector and generates a signal that makes the current control element conductive only when i<Ip. a pulse time control circuit that outputs a signal that continues from the time when the signal synthesis circuit receives the output signal of the pulse frequency control circuit until the time when the signal synthesis circuit receives the output signal of the second comparator; 7. The control device according to claim 6, comprising an AND circuit whose inputs are the output of the time control circuit and the output of the first comparator. 9. The control device according to any one of claims 6 to 8, wherein the pulse frequency control circuit is a circuit that generates a pulse with a frequency corresponding to a welding voltage. 10. The control device according to any one of claims 6 to 8, wherein the pulse frequency control circuit is a circuit that generates pulses at a frequency corresponding to the feeding speed of the consumable electrode. 11. The control device according to any one of claims 6 to 10, wherein the reference value setting device is a setting device that outputs a signal having a value corresponding to a welding voltage. 12. The control device according to any one of claims 6 to 10, wherein the reference value setting device is a setting device that outputs a signal corresponding to the feeding speed of the consumable electrode.