KR960001587B1 - Method and arrangement for resistance welding - Google Patents

Abstract

No content.

Description

Resistance welding method and apparatus for implementing this method

1 shows a circuit diagram of a resistance joint welder having a device for regulating welding current according to the present invention.

FIG. 2 shows a detailed view of a part of the device according to the invention shown above line II-II in FIG. 1.

FIG. 3 shows a detailed view of the regulator represented by one block in FIG.

4 shows a first example of the pulse width modulated primary alternating voltage of the welding transformer and the resulting sinusoidal welding current.

FIG. 5 shows a second example of the first alternating voltage pulsed and modulated in a manner different from that in FIG.

6 shows a third example of the pulse width modulated primary alternating voltage in which a trapezoidal welding current is generated.

7a-c show various forms of preselectable trapezoidal welding currents with inclined impulse tops.

8A-8C show several examples of preselectable trapezoidal welding currents in which the impulse top has one or more protrusions.

9A-9C show several examples of preselectable trapezoidal welding currents in which the impulse top has one or more lower portions.

10 shows an example of a triangular welding current.

11 to 37 show preferred current forms.

* Explanation of symbols for main parts of the drawings

10, 12: welding electrode 14: frequency converter

16: welding transformer 19: welding machine control system

20: Current transformer 27: A / D converter

54: multiplier

The present invention is carried out using a welding current generated from a primary alternating voltage and regulated by its pulse width modulation in a periodic half wave, and the primary alternating voltage is applied during pulse width modulation. It is related to a resistance welding method in which the welding current is pumped in each half wave by a chipping frequency, and the welding current is controlled in the half wave based on the rated value-actual value comparison by affecting the impact coefficient due to the pulse width modulation.

In addition, a static frequency converter having a DC intermediate circuit of the present invention, which has a output as a output terminal, which has a output circuit which generates a first alternating voltage and transmits the secondary circuit to a welding transformer connected to a welding electrode of a resistance welding machine. And a control device capable of controlling the chopper for multiple chopping of each half-wave of the first alternating voltage, wherein the control device controls the welding current by pulse width modulation of the first alternating voltage. An apparatus for carrying out the method as claimed in claim 1 comprising such a regulator connected to a bumper of said frequency converter and having a welding current reference element.

The method and the apparatus are known from EP-A2-0 078 252, which will be discussed in more detail below. In addition, a method and apparatus for providing pulse width modulation and pulse width modulation are known from EP-A2-0 260 963.

In a well-known joint welding device (EP-A1-0 261 328) for longitudinal resistance joint welding of superposed edges of bodies such as cans for canning, the three-phase power AC voltage is converted into DC voltage. Converted to an impulse voltage with smoothed and alternating polarity. This impulse voltage is applied to the welding electrode of the joint welding apparatus. The frequency of the impulse voltage is chosen such that the resulting welding current is continuous and for this reason the individual welding nuggets or weld points, each generated by one rectangular half wave of the impulse voltage, overlap each other. Since each welding current half wave is generated by the half wave of the impulse voltage, the shape of the welding current depends on the duration of the impulse voltage half wave. If the impulse duration varies during one half wave when adjusting the welding current, this results in a significant variation in the welding current shape and this variation should be regarded as disadvantageous. It would be considerably more advantageous if the welding current type does not depend on the parameters of the machine and thus is not determined by it, but can be preselected to optimize the welding result.

Moreover, in known joint welding apparatuses, a welding current regulator which regulates the current at each welding spot operates in each case with the value measured at the preceding welding spot. The response time of the regulator, which is also determined by the calibration device controlled by the regulator, is therefore relatively long (with a welding cycle of 500 Hz and the response time reaches 1 ms). As a result, the regulator cannot correct rapid changes in welding parameters (eg on contaminated sheet metal surfaces). Therefore, in order to improve the controllability of the known joint welding apparatus, the response time of the regulator is required to be shortened. For this purpose, the switching frequency can be significantly increased by a predetermined factor.

However, the frequency of the welding current will also increase as a result of the above factors. As a result of the strong inductive load of the joint welding device, the impedance will increase in proportion to frequency, so the welding current is reduced by a factor equal to the inverse of the factor that increases the switching frequency. To compensate for this, the voltage and power of the frequency converter and the welding transformer of known joint welding devices shall be increased by the same factor as the factor of increasing the switching frequency. In addition, the requirement that the welding frequency maintain a constant proportion to the welding speed will no longer be fulfilled. In the known joint welding apparatus, welding parameters such as contact resistance at the welding point (surface quality of the welding material), material properties of the welding material, etc., due to the long regulator response time cannot be considered quickly enough and also It is not possible to adapt the welding current type to different welding conditions, for example the requirements for different materials to be machined.

The aforementioned EP-A2-0 260 963 proposed the use of a high frequency current source to make it possible to use smaller welding transformers. This inevitably leads to problems associated with the phase control of the thyristor, so that the feedforward of the welding current can be achieved by using a precomputed precalculated value at half wave during phase control based on the preceding half wave measurement. Or forward control is activated. In addition, since the welding current is connected-shorted only once per welding point, the device known in the above publication also does not operate satisfactorily in all working conditions. The response time of the regulator is also relatively long here as the regulator also operates with the value measured at the preceding welding point in each case. If the pulse length is varied during the pulse width modulation, the welding current pattern is also changed, and for this reason cannot be adapted to the latter special material or special operating conditions.

Moreover, a common fact for both devices already known is that only the secondary average of the welding current is measured as its actual value and therefore only the average of the welding current can be controlled. For the above reason, the predetermined average value of the welding current is set in advance as the rated value.

CH-A5-668 842 is known a device for stepless control of the amplitude of sinusoidal alternating current. In each half-wave of the alternating current, controllable circuit elements can change from a blocking state to a transmitting state over its variable portion. Certainly this gives a kind of adjustable transformer which can be controlled electronically with virtually no delay, but here too the possibility of affecting the welding current is limited to one switching procedure per half wave. For this reason a faster adjustment time cannot be obtained in this case either.

DE-C2-30 05 083 describes a method for the production of longitudinally welded curved bodies, in which one half-wave duration of a nearly rectangular welding current is obtained to obtain a continuous, unbroken weld seam. The time is adapted to the time for transporting one body between the welding electrode rollers and the energy required during the welding operation can be controlled directly by adding a high frequency current component to the welding current thereby. The possibility of adjustment by adding a high frequency current component is naturally limited not only with respect to the adjustment range but also with respect to the adjustment time.

Finally, experts are well known to vary the welding current strength by phase-shifting control, for example from Soudronic's publication "Electric Resistance-welding", MDI 00188 D, pages 9 and 10. Unfortunately, the welding current pattern also varies in each case. In both cases, the angle of the phase shift must be varied, so that the welding current is the same as the above case if the welding current is kept constant when the load conditions vary. Moreover, phase shift control of the primary alternating voltage of the welding transformer likewise produces an adverse interrupted welding current.

From EP-A-0 078 252 it is known to generate a welding current and modulate the pulse width with a switched power supply for argon-arc welding or TIG welding. . However, an analog controller is provided for generating only one rated value and continuous welding current for the welding current.

It is an object of the present invention to provide an apparatus for the implementation of one such method and the kind of method described above, in which the welding current form can be readily adapted to the requirements of the materials described above to be processed. Furthermore, this method and apparatus are intended to be suitable for the rapid regulation of welding current.

The above object is solved by the features of claim 1 devised in accordance with the present invention from methods of the foregoing kind.

The above problem is also solved by the features of claim 39 devised from devices of the above kind.

In the prior art described above, the primary alternating voltage is only choked once within each half wave during pulse width modulation, whereas according to the invention it is n times n times, where n> 1. With this feature a short adjustment time can be obtained, which allows several rated-actual comparisons of the welding current to be carried out at each half wave, thus repeatedly affecting the duty ratio during pulse width modulation. Because it can have. Accordingly, the chipping frequency is a predetermined multiple of the welding current frequency, so a quick adjustment process is obtained during the process at each welding point. This allows the regulator to correct for quick changes in welding parameters (such as for example on contaminated sheet metal surfaces). The shape of the impulse formed by the first alternating voltage being doped in each half wave is close to a rectangle. The duty ratio, ie the impulse width / impulse spacing, can vary within wide limits.

This can directly affect the average value of the primary alternating voltage, and the current form can be given in advance in the required form and thus is variable, which is not possible in the prior art. In the prior art, as described above, the current shape is influenced or fixed by the system (eg during phase shift control). Whereas in the prior art described above, the hopping frequency is fixed (e.g. during phase shift control) or equal to the welding frequency even in the largest case, it is a predetermined multiple of the welding frequency in the method and apparatus according to the invention. Therefore, the above method and apparatus provide the following advantages through the preferred specific embodiment;

The welding frequency may vary, the chipping frequency being a selectable multiple of the welding frequency;

The type of current can be preselected, and therefore can vary, and not significantly change by modification of the impact coefficient during the adjustment of the welding current;

If the preselected welding current shape is not maintained during the process, the current shape is correspondingly corrected by the adjustment process, ie by affecting the impact factor;

The welding current form stored in the memory can be selected as a trapezoid, such as a triangle, sinusoidal or trapezoid with an inclined impulse top or trapezoid with a protrusion or down, depending on the requirements (one welding Depending on the required heat-energy balance within the point, the more controlled the heating and cooling steps within a single welding point, such as, for example, chrome-plated parts, were not considered to be weldable before The welding process can be better controlled so that the materials can now be welded according to the invention). ;

Since the welding is done n times in one half wave and the current is readjusted in each case, the response time of the regulator is significantly shorter than in the prior art.

Advantageous developments of the invention are formed by the subject matter of the dependent claims.

In the configuration of the invention according to claim 5, each welding current type is defined by one selectable rating table. Thus, for example, one rating table is stored for sinusoidal current forms, the other for triangular current forms, the other for trapezoidal current forms, and so on.

Since the transistors used as circuit elements in the arrangement of the invention according to claim 40 have particularly short switching times, the adjustment process according to the invention can be most easily realized.

In the configuration of the invention according to claims 41 or 42 a special sub-table or rating table can be used for each welding current type or frequency or both. As a result, more ratings per half wave can be used for lower welding frequencies than for higher welding frequencies.

In the configuration of the invention according to claim 43, the short response times of the regulator are particularly well used since the current ratings and the respective variations between adjacent current ratings can conveniently be precomputed and stored in a table. Can be done. The first derivative of the welding current curve is preferably stored as a variation between adjacent current ratings. This has the advantage that the adjustment process can be performed in anticipation, that is, the overshoot during the adjustment process can be substantially avoided from the start, which pre-dates the position of the next current rating by the stored variations. Because you can see.

In the arrangement of the invention according to claims 44 or 45, the required amplitude of the welding current can be obtained in a simple way by multiplying the stored ratings in the table by a corresponding factor which can be entered as needed. .

The present invention relates to resistance welding with a pulsating current in periodic half waves, more particularly with a welding current of alternating current. Until now, such welding has been done with sinusoidal currents. In tin plate welding, a problem arises in the case of very thin metal plate or very thin tin plated metal plate welding. Especially in the welding of candles (tin containers), the metal plates cause a problem that it is difficult to control the production technology. The same problem arises in the welding of black plates and in particular plated metal plates, more particularly chrome plated metal plates. So far, attempts have been made to overcome this problem with different welding current amplitudes and welding current frequencies of sinusoidal welding currents, but the results are often unsatisfactory.

The task is therefore to make welding of thin or thin tinned metal sheets and other metal sheets possible. Within a very narrow tolerance band width, in order to prevent the formation of spatters (if the energy supply is excessive) or the formation of gaps in the overlaps (if the energy supply is insufficient), the energy supply is particularly important during the welding operation. To make it possible.

Preferably the welding current is made in the manner described below by deviating from the sinusoidal form.

The individual half-waves of the welding current can take any form so that it is possible to supply exactly the energy required at the welding position for optimum welding operation.

Through the course of the current, the required heating and cooling at the welding position can be made very precisely controlled to provide the required electrical resistance at the welding position, which has never been possible.

Embodiments of the present invention are described in more detail below with reference to the drawings. 1 is a simplified circuit diagram of a resistance joint welder for longitudinal joint welding of an unshown curved can body between the roller-shaped welding electrodes 10 and 12. This resistance joint welder has an energized frequency converter 14 which is powered from the main power source shown by conductors L1-L3 and is connected to an output stage 14b designed as a chopper via a conventional DC intermediate circuit 14c. (14a). This output stage 14b is connected to the primary circuit of the welding current transformer 16, to which the output stage supplies the primary alternating voltage U _p . The secondary circuit of the welding transformer 16 is connected to the welding electrodes 10 and 12.

According to the enlarged view in FIG. 2, the input stage 14a of the static frequency converter 14 has one three-phase rectifier, which is generally known at the same time and is not important for the understanding of the present invention. It forms the input of the DC intermediate circuit 14c, which does not need to be described in more detail here. As shown in FIG. 2, the winder of the output stage 14b of the frequency converter 14 (FIG. 1) is a switching element, a transistor T ₁ -T ₄ and a freewheel diode in parallel with these transistors. and a bridge circuit having freewheel diodes (F ₁ -F ₄ ). Four gate drivers are connected to the transistor and the freewheel diode in the manner shown in FIG. 2 and controlled by the regulator 18 (FIG. 1) via leads 15. A current transformer 20 is provided in the primary circuit of the welding transformer 16, which detects the actual value of the current flowing in the primary circuit of the welding transformer 16. As shown in FIG.

The present invention will be described with reference to the adjustment process in the following because it allows for the first time rapid adjustment.

According to the representation of FIG. 1, the current actual value from the current transformer 20 is supplied to the input of the regulator 18 designed as a process computer via the A / D converter 22. The regulator 18 may be set via the potentiometers 24 and 26 with a rating I _soll for the welding current or a rating f _s for the welding frequency. The analog voltages set in potentiometers 24 and 26 are applied to the process computer via A / D transformer 25 or 27. Additionally the welding current supply size I _F can be supplied to the regulator 18 via an input marked MANUAL or via a welder control system 19. This magnitude is connected, for example, to the rated current I _soll to take into account that the current through the can body is not constant. Thus, the welder control system 19, which accurately recognizes where the welded can body is at each point in time, also corresponds to the rated value I _soll set such that the welding is performed at a suitable welding current amplitude at each point of the can body. It can be changed. The regulator 18 determines the set point by comparing the rated-actual value with respect to the welding current, which is set via the A / D converter 28 and the lead 15 to the output stage of the frequency converter 14 (FIG. 1). Supplied to the gate driver (FIG. 2) in 14b). The set point affects the impact factor of the rectangular impulse chopped from the direct current voltage smoothed in each half wave from the DC intermediate circuit 14c by the bumper at the output stage 14b, which is described later in more detail with respect to FIG. It is to control the welding current by the pulse width modulation to the first alternating voltage with the impact coefficient affected by the set value as will be.

Various methods for generating a first alternating voltage by mapping smoothed DC voltage into rectangular impulses are shown in FIGS. In the example in FIG. 4, the smoothed DC voltage is pumped into a rectangular impulse with a polarity that changes every half wave, resulting in a first-order alternating voltage U ^j in the form of a sinusoid on average, followed by a welding current of substantially sinusoidal I. Occurs in the form.

In the example of FIG. 5, the same as above, in this example, the smoothed direct current voltage is mapped to rectangular impulses of the same height, and the height of the primary alternating voltage, which is the sinusoidal shape on average in each case, is the peak value of U _p . Equal to 2 times

In the example according to FIG. 6, the smoothing of the DC voltage is driven according to the same principle as in FIG. 4, but in this case, a trapezoidal welding current I is generated.

The regulator 18 is represented in more detail in FIG. As already mentioned above, the regulator 18 is designed as a process computer, of which only those parts important to the invention are represented in FIG. 3 and described later. This includes a welding current reference element 52 in the form of a PID control circuit 50 and a storage device, each of which is for comparison with each current actual value detected at each of the sampling intervals. For the current rating corresponding to the type of welding current. For each type of welding current (sinusoidal, triangular, trapezoidal, etc.), the storage device 52 stores a rating table, which can be selected via the input W _tab . One output of the storage device 52 is connected to the input of a multiplier 54. The output of this multiplier 54 is connected to a summing point 56. The input signal received from the multiplier 54 of the summing point 56 is connected with the current actual value. The output signal of the summation point 56 formed by the rated value-actual value comparison is applied to the input portion of the PID control circuit 50.

The PID control circuit 50 supplies a set signal at its output to one input of the summing point 58. Another output of the storage device 52 is connected to another input of the summation via a feed-forward or forward drive loop 60. Through this feedforward loop, the memory device changes from the actual current rating supplied to the proportioner 54 to the next rated value, i.e., the weld current curve at the actual current rating in the direction of the first derivative dI / dt or the next current rating. The increase amount is supplied to the summing point 58. The output signal of the data PID control circuit 50 in the above direction is interconnected. As a result, the output signal of the summing point 50 constitutes a setting signal. With this setting signal, the welding current is determined in the correct direction and ratio. It can be installed as a result, so that no overshoot occurs in the process of adjusting the current. Inside the rating table corresponding to each type of welding current, a sub-table can also be selected specifically for each welding frequency f _s , which is described in more detail below. The ratings of the current curve selected by the input signal W _Tab and also its first derivative are stored in each rating table. For each measurement and chipping interval the corresponding ratings from the table above are multiplied by the required current amplitude in the proportioner 54 and then supplied to the summing point 56 as a rating. The required current amplitude is supplied to the multiplier 54 via the A / D converter 25 as an I _soll signal, and multiplied therein by the current rating from the storage device 52. The required current amplifier I _soll may be, for example, further inclined to the upper part of the impulse as shown in FIGS. 7A to 7C or as shown in FIGS. 8A to 8C or 9A to 9C. Selectively or additionally, manual input may be applied to the welding current I in order to have a predetermined process applied to the welding current I within a range of half-waves of the first alternating voltage at one welding point, such as giving the upper part of the impulse a protrusion or a lower part. It may be affected via or through the welder control system 19 (FIG. 1).

As mentioned above, the memory device 52 stores one rating table for each current type, and in the embodiment shown, four rating tables. The welding current type required in each table is stored by several preset current ratings. In the present example, 256 ratings are stored per cycle of welding current. With a welding frequency of 500 Hz and a chipping frequency of 10 kHz, 10 100 kW or switching intervals are allowed for each half-wave. Thus, the welding current can be pumped ten times per half wave. That is, it can be switched on and off ten times. Therefore, 20 welding current ratings per cycle, ie 10 ratings per half wave, are selected from the 256 available welding current ratings, and used in the regulator 18 for rated-actual value comparison. If the welding frequency is only 50 Hz, 200 ratings per cycle of welding current, and therefore 100 ratings per half wave can be selected. According to the selected welding frequency f _s , an appropriate subtable corresponding to the welding current type among the rated values is selected via the A / D converter 27.

The rating table also stores the amount of change from one welding current rating to the next, i.e., the dI / dt values of the 256 predetermined welding current ratings. When working with welding frequencies between 35 and 40 Hz, all 256 points will be used in the nominal-actual comparison. Typically, however, it works with a welding frequency of 500 Hz, with the result that only 20 points per cycle of welding current are used in the nominal-actual comparison. Therefore, if a sub-table for a welding frequency higher than f _s is selected instead of a rating table with 256 ratings, the computer automatically adapts to this change, so that the change corresponds to the spacing between the selected weld current electric shock values. Will be done. Another possible method is to pre-calculate the subcategories with 256 points per weld current cycle from the beginning and then select subcategories with fewer weld current ratings, but precompute these subcategories and the amount of change per rating. Together with them, it is possible to select them as a rating table in the storage device 52.

The rated current value supplied by the storage device 52 corresponds exactly to the desired welding current type but does not correspond to the required amplitude yet. As already explained, the latter is determined by a separate factor which can be supplied to the multiplier 54 via another three inputs as described above.

The adjustment process is performed as follows: Referring to the above example, it is assumed that the operation should be performed with a welding frequency fs of 500 Hz and a chipping frequency of 10 kHz. The welding current I has a sine wave shape and is obtained by pulse width modulation of the primary alternating voltage U by the method shown in FIG. The rating table includes 10 ratings per half wave for the welding current (I). The smoothed DC voltage given by the DC intermediate circuit 14c is pumped at 10 kHz so that a welding current curve corresponding to the current rating is generated. The measurement frequency for determining the actual value of the welding current from the current transformer 20 is equal to the chipping frequency. Therefore, the actual welding current value is measured for each welding current rating. In each rating-actual value comparison, it is determined whether the measured actual value is equal to the rated value of the welding current shown in the rating table. If this is not the case, the summation point 56 and the PID control circuit 50 supply an error signal, from which the setting signal for the shock coefficient is formed using the feedforward signal in the manner described above. do. The set signal above affects the impact coefficient in the following manner: during the pulse width modulation of the primary alternating voltage, the ratio between the impulse width and the impulse spacing is such that the difference between the weld current actual value and the weld current rating value is eliminated. Is modified in a way.

Thus, the welding current can be readjusted within one half wave of the welding current, i.e., within one welding point, within an extremely short adjustment time. Another particular advantage of the above adjustment method is that each required current form is additionally memorized as a rating table and can be selected if necessary. The welding current type can be freely selected within certain limits set by the machine only in practice (eg when there is a maximum possible increase in the welding current curve which cannot be exceeded due to existing physical factors). .

In the so-called full sine welding of the can bodies between the upper and lower welding rolls, as in the case of the welding electrodes 10 and 12 shown here, the overall contact length between the welding rolls and the metal plate The heating distance over is pumped in six stages, where these stages generate three half-waves resulting from a welding speed of 60 m / min and a welding frequency of 500 Hz and also from a total contact length of 3 mm, these three cold and It is mapped to three hot periods (see “Soudronic” Company Journal, 1st year of Publication, No. 1, June 1985, page 3). As a result, the generation of each welding spot between the welding rolls consists of three alternating actions between heating and cooling. The control method according to the invention allows for optimal control of heating and cooling steps within one welding point. 7 to 9 show suitable welding current forms for this. Thus, adaptation of various materials to the welding behavior is possible with the present invention. Metal plates that have never been welded without generating spatter can now be well welded with flat welding current impulses that do not have current peaks.

11 shows a current process in which the welding current initially increases in each half-wave in sinusoidal form, decreases and then increases again before reaching the sinusoidal peak and then decreases toward zero crossing. With the above specific embodiment of the present invention, the welding spot formation heat (not liquid phase) can have a very good and purposeful effect. In roll seam welding, it is also very good for sheet metal materials, which have typically been difficult to control by working with a welding frequency of 500 Hz, a welding current of 3700 A and a welding speed of 60 m / min, for example. You can get the result.

12 and 13 show another preferred current form having a repeated reduction of the welding current in the center of the half wave; FIG. 12 with an increase in sinusoidal shape initially from zero intersection; FIG. 13 with a linear increase to the first amplitude peak lying higher than two successive peaks. With the above rectification forms, maximum welding speeds can be obtained with a low welding frequency, which prevents excessive heating of the welding device and generates less energy loss. For example, a frequency of 250 Hz, a current of 3780 A, and a speed of 60 m / min can be set for the roll joint.

Figure 14 shows the preferred current form with a gently decreasing current process in each case in the center of each half wave. With the above process, a wide welding range (between the attachment and the limit of the spatter) can be realized depending on the sheet metal quality.

15 shows a process in which the welding current is triangular. Advantages here can be obtained, in particular in the welding of a plated (not tinned) metal sheet, unlike in the prior art.

Figure 16 shows a similar form of current that provides lower energy to the material to be welded.

17 to 25 show in each case the current forms in which the welding current remains constant for a period of time within a half wave. In the case of special welding this results in a particularly good energy supply to the weld zone.

26, 27 and 28 degrees show a slightly slower decrease in welding current during half wave.

Figures 29-36 show such current forms in which the energy supply is severely reduced during half-waves by reducing the current to zero, or the current reverses for a short period of time during each half-wave.

37 shows a current form having constant portions, where the first constant interval has a higher amplitude than the following intervals.

Claims (30)

In a periodic half-wave, specifically, it pulsates with alternating current and is generated using a welding current generated from the primary alternating voltage and controlled by its pulse width modulation, and the primary alternating voltage is applied to the chipping frequency during the pulse width variation. In the resistance welding method in which the welding current is pumped in each half wave, and the welding current is controlled in the half wave based on the rated value-actual value comparison by affecting the impact coefficient due to the pulse width modulation. Any number greater than 1 is greater than n times the frequency of the welding current, a series of welding current ratings are provided for each half wave, and n nominal-actual values comparisons for the welding current are made within each half wave. And the impact coefficient is affected n times within each half wave. method. The resistance welding method according to claim 1, wherein the type of welding current can be selected in advance by a rating table. The resistance welding method according to claim 1 or 2, wherein the welding current deviates from the sinusoidal wave shape. 4. The welding current of claim 3 wherein the welding current initially increases primarily in sinusoidal form after zero crossing, decreases and then increases again before reaching sinusoidal peak, and thereafter decelerates mainly in sinusoidal wave towards zero crossing. Resistance welding method. 4. A process as claimed in claim 3, characterized in that the welding current is initially increased primarily in sinusoidal form after zero crossing and then repeatedly decreased-increasing thereafter, and thereafter decreasing in mainly sinusoidal form towards zero crossing. Resistance welding method. 6. The resistance welding method according to claim 5, wherein the welding current follows a single process of two decrease-increase between sections which are mainly sinusoidal waveforms. 4. The method of claim 3, wherein the welding current initially increases steeply, then slowly decreases and then steeply decreases. 4. The method of claim 3, wherein the welding current increases substantially linearly from zero crossings, repeatedly decreasing-increasing in the peak value interval of half wave and then decreasing substantially linearly toward zero crossings. 9. The method of claim 8, wherein the welding current is reduced-increased twice between generally straight sections. The resistance welding method according to claim 3, wherein the welding current follows a triangular process. The resistance welding method according to claim 3, wherein the welding current follows a trapezoidal process. 12. The method of claim 11, wherein additional trapezoidal current variation occurs between shoulders of a longer trapezoidal process. 13. The method of claim 12, wherein a plurality of similar trapezoidal current variations occur between shoulders of a longer trapezoidal process. 4. The method of claim 3, wherein the welding current initially increases substantially linearly, then decreases and increases substantially linearly thereafter, and then decreases substantially linearly again. The resistance welding method according to claim 14, wherein the reduction of the current in the half wave occurs mainly at a zero value of the current. 15. The method of claim 14, wherein the reduction in current in the half wave occurs below the zero value of the current. 5. The resistance welding method according to claim 4, wherein the decrease in current in the half wave occurs to zero. 5. The method of claim 4, wherein the reduction in current in the half wave occurs below the zero value of the current. The resistance welding method according to claim 12, wherein in the trapezoidal current variation, one horizontal section is higher than subsequent horizontal sections in the plurality of current variations. The resistance welding method according to claim 3, wherein the welding frequency is 500 Hz or 250 Hz. The method of claim 1, wherein the joints are roller welded. The resistance welding method according to claim 1, wherein the metal sheets are welded. A bumper 14b having a DC intermediate circuit 14c, generating a primary alternating voltage Up, which is transmitted to a welding transformer 16 connected to the welding electrodes 10 and 12 of the resistance welding machine. And a control device capable of controlling the bumper for multiple chopping of each half-wave of the first alternating voltage. Claims having such a regulator 18 connected to a bumper 14b of the frequency converter 14 for controlling the welding current I by means of pulse width modulation of the primary alternating voltage and having a welding current reference element. In the apparatus for carrying out the method as claimed in claim 1, the welding current reference element of the regulator 18 is the type of welding current for each pumping interval in order to compare with each current actual value measured at each pumping interval. Resistance welding apparatus, characterized in that the storage device (52) for storing a set of the current rated value corresponding to. 24. The chipper 14b of the frequency converter 14 has transistors T ₁ -T ₄ as circuit elements and free-wheel diodes F ₁ -F ₄ connected in parallel thereto. Resistance welding device comprising a bridge circuit. 25. The apparatus of claim 23 or 24, wherein a plurality of current rating sequences can be stored, and the memory device 52 has a selection input section W _Tab for selecting one of the stored rated rating sequences. Resistance welding device. 26. The memory device 52 according to claim 25, wherein the memory device 52 has one welding frequency input f _s and a current rated value sequence corresponding to the required shape of the welding current is inputted to the memory device 52 for each welding frequency. Resistance welding device characterized in that stored. In the resistance welding device as claimed in claim 25, wherein the regulator 18 comprises a PID control circuit 50 having a feed forward loop 60, a subsequent current rating for each current rating in the storage device. Resistance is stored and supplied as a feedforward value to the output of the PID control circuit (50). 24. The memory device 52 according to claim 23, wherein the storage device 52 has a multiplier 54 connected thereto, the multiplier being selectable which can be input to the rated value via an adjustable current rating input I _soll of the regulator 18. Resistance welding device, characterized in that multiplying the factors. 29. A resistance welding device according to claim 28, wherein the multiplier (54) has additional inputs that can be input from a welding machine control system (19) with additional factors being manually or superimposed. A resistance welder comprising the device as claimed in claim 23.