METHOD FOR ELECTRICAL DISCHARGE MACHINING

Abstract

The invention relates to a method and device for wire electrical discharge machining (WEDM), in which a first voltage UCH1 is measured between a first current feeder and the workpiece, and a second voltage U.sub.CH2 is measured between a second current feeder the workpiece, whereas said first- and second voltage U.sub.CH1, U.sub.CH2 are measured when current of the machining discharge pulse is zero, for instance at the end of the of the machining discharge current pulse, and a voltage difference of said measured voltages ΔU.sub.CH=U.sub.CH1−U.sub.CH2 is determined, and a discharge position of said machining discharge pulse is determined in real time, as a function of said voltage difference ΔU.sub.CH.

Claims

1. A wire electrical discharge machining (WEDM) method, in which a preliminary voltage pulse is applied to a gap between a wire electrode and a workpiece, and, at the breakdown of said preliminary voltage pulse, a machining discharge pulse is applied to the wire electrode, comprising wherein, a first voltage U.sub.CH1 is measured, between a first current feeder and the workpiece, and that a second voltage U.sub.CH2 is measured, between a second current feeder and the workpiece, wherein: said first voltage U.sub.CH1 and said second voltage U.sub.CH2 are measured when current of the machining discharge pulse is zero, a voltage difference of said first- and said second measured voltage ΔU.sub.CH=U.sub.CH1−U.sub.CH2 is determined, and a discharge position of said machining discharge pulse is determined in real time, as a function of said voltage difference ΔU.sub.CH.

2. A WEDM method according to claim 1, wherein the voltage is measured, at the beginning of a machining discharge pulse, and/or at the end of a machining discharge pulse.

3. A WEDM method according to claim 1, wherein the determined voltage difference is a function of the wire segment inductance L.sub.1 of the partial wire segment length s.sub.U between the upper current feeder and the discharge position, and the wire segment inductance L.sub.2 of the partial wire segment length s.sub.L between the lower current feeder and the discharge position.

4. A WEDM method according to claim 1, wherein a discharge position is discretized by dividing a height H.sub.WP of the workpiece into a number N.sub.S of vertical sections S, and, with each discharge D.sub.i, determining the voltage difference ΔU.sub.CH, based on said voltage difference ΔU.sub.CH, determining a discharge position Z.sub.Di of each discharge Di along an engagement line of a wire electrode and the workpiece, and assigning each discharge Di to a matching vertical section S.sub.j of the workpiece based on the determined discharge position Z.sub.Di of each discharge.

5. A WEDM method according to claim 1, wherein said first voltage U.sub.ch1 and said second voltage U.sub.ch2 are measured by a voltage measurement means V.sub.1 and V.sub.2, which is connected to the respectively at the upper- and lower current feeder by means of a measuring cable.

6. A WEDM method according to claim 5, wherein said measurement cables are the gap voltage measurement cables.

7. A WEDM method according to claim 1, wherein the machining discharge pulse is a capacitive discharge pulse.

8. A WEDM method according to claim 1, wherein the determined discharge position Z.sub.Di is combined or compared with one or more other discharge position detection methods, including: partial currents detection method, a discharge frequency based workpiece height detection method, geometrical data about the workpiece height stored in the control unit.

9. A WEDM method according to claim 1, wherein a lookup table is provided for indicating the vertical section S in which the discharge occurs for a voltage difference ΔU.sub.CH, in particular, the lookup table comprises data for each wire diameter used in the WEDM process.

10. A WEDM method according to claim 1, wherein a calibration step is conducted by determining the voltage difference ΔU.sub.CH measured with discharges occurring at the top- and/or bottom of the workpiece surface.

11. A voltage measurement means for measuring the first voltage U.sub.CH1 and the second voltage U.sub.CH2 to determine the discharge position according to claim 1.

12. The voltage measurement means of claim 11 which includes an operation amplifier (52) for subtracting the measured voltages U.sub.CH1 and U.sub.CH2 and generating the difference voltage ΔU.sub.CH; a sample and hold circuit (53) for receiving the different voltages and sampling the voltages when the current zero crossing signal is active.

13. A wire electrical discharge machine for machining a workpiece by a wire electrode comprising: a table for mounting the workpiece thereon; an upper current feeder and an lower current feeder for carrying machining current pulses to the wire electrode for provoking an electrical discharge between the wire electrode and the workpiece; and a machining process monitoring and control board including the voltage measurement means according to claim 11.

14. The wire electrical discharge machine according to claim 13, wherein the machining process monitoring and control board is mounted in a cabinet (8) for accommodating generator boards and/or power supplies and/or a control unit and/or motor drives.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Preferred embodiments of the invention will now be detailed with reference to the attached drawings, in which

[0029] FIG. 1 is a sketch of the machining area of a WEDM, with a prior art discharge position detection circuit;

[0030] FIG. 2 is a known discharge position discretization scheme;

[0031] FIG. 3a is a sketch of the machining area of a WEDM, with a discharge position detection circuit, according to the preferred embodiment of the invention;

[0032] FIG. 3b is a sketch of the machining area of a WEDM, with a discharge position detection circuit, according to the preferred embodiment of the invention;

[0033] FIG. 4a is a circuit model of a discharge position detection circuit according to the preferred embodiment of the invention;

[0034] FIG. 4b is a circuit model of a discharge position detection circuit according to the preferred embodiment of the invention;

[0035] FIG. 5a is a signal for the voltage measurement;

[0036] FIG. 5b is a signal for the voltage measurement;

[0037] FIG. 5c is a plot of machining currents i1 and i2;

[0038] FIG. 5d is a plot of a plurality of measured voltage signals at different discharge positions;

[0039] FIG. 6a is a digital oscilloscope acquisition of the current of a machining discharge pulse, and the voltage according to the discharge position;

[0040] FIG. 6b is a digital oscilloscope acquisition of the current of a machining discharge pulse, and the voltage according to the discharge position;

[0041] FIG. 7a is a comb-shaped workpiece generating the waveforms of 7b which shows a digital oscilloscope measurement of the voltage in persistence mode;

[0042] FIG. 7b is a comb-shaped workpiece generating the waveforms of 7a which shows a digital oscilloscope measurement of the voltage in persistence mode;

[0043] FIG. 8 is a plot of an upper- and a lower machining discharge pulse current, i1 and i2;

[0044] FIG. 9 is a functional diagram showing the principle of signal acquisition and processing;

[0045] FIG. 10a is an exemplary calibration method for the discharge position detection system;

[0046] FIG. 10b is an exemplary calibration method for the discharge position detection system; and

[0047] FIG. 11 is a graph illustrating the relation between the wire self-inductance and the wire segment length.

DESCRIPTION OF THE PREFERRED EMBODIMENTS OF THE INVENTION

[0048] FIG. 1 shows the machining area of a wire electrical discharge machine of the prior art. A workpiece 2 is mounted on a table 6 in a worktank 5 filled with a dielectric machining fluid (not shown). The workpiece 2 is cut by means of a wire electrode 1, which is precisely guided by means of wire guides 22, 23. The wire guides are located in the upper- and the lower wire guiding head 20, 30. Wire guiding heads 20, 30 further include an upper- and a lower current feeder 21, 22. An upper- and a lower electrode cable 25, 35 connects the generator to the upper- and lower current feeder 21, 31. In operation, the wire electrode 1 slides on the current feeders 21, 22 while the machining current pulses are carried to the wire. The discharges occur in the gap 3, along the engagement line of wire 1 and workpiece 2.

[0049] According to the prior art, as discussed e.g. in EP3446820A1 the partial discharge currents I20, I30 flowing to the wire through the upper- and lower feeding path are measured separately using toroidal current transformers 29, 39.

[0050] The discharge position referred to a position at half distance Z.sub.FM between the current feeders 21, 31 is proportional to the difference of the partial discharge currents. More specifically, the difference of partial discharge current signals (I20−I30), divided by the total discharge current (I20+I30), multiplied by the half distance between the current feeders (Z.sub.FU+H.sub.WP+Z.sub.FL)/2 determines the discharge position Z.sub.Di of each discharge Di occurring along an engagement line of the wire and the workpiece, from the midpoint between the current feeders. The location of discharges is discretized for further processing. FIG. 2 shows a discretization scheme in which the workpiece height is divided into a discrete number of sections, and the detected discharge positions are assigned to the respective section, as known in the art.

[0051] As mentioned previously, the method of partial currents flowing in an upper- and a lower current feeding branch has a number of disadvantages. Therefore, the aspect of the invention is to provide an improved discharge position detection method.

[0052] In the wire electrical discharge machining method according to the invention, a preliminary voltage pulse is applied to a gap between a wire electrode and a workpiece, and, at the breakdown of said preliminary voltage pulse, a machining discharge pulse is applied to the wire electrode, wherein, a first voltage U.sub.CH1 is measured between a first current feeder and the workpiece, and a second voltage U.sub.CH2 is measured between a second current feeder and the workpiece, whereas said first voltage U.sub.CH1 and said second voltage U.sub.CH2 are measured when current of the machining discharge pulse is zero and the voltage difference ΔU.sub.CH=U.sub.CH1−U.sub.CH2 of said first- and said second measured voltage is determined, and a discharge position of said machining discharge pulse along the engagement line of the wire electrode and the workpiece is determined in real time, as a function of said voltage difference ΔU.sub.CH.

[0053] For instance, the first current feeder is upper current feeder, and the second current feeder is the lower current feeder, or vice versa. In this way, the voltage between the wire electrode and the workpiece is measured, in both, the upper- and the lower contact position.

[0054] According to an embodiment, the voltage is measured at the beginning of a machining discharge pulse, and/or at the end of a machining discharge pulse, when current of the machining discharge pulse is zero.

[0055] The method according to the invention will now be presented in detail. At the beginning, a preliminary voltage pulse is applied to a gap between a wire electrode and a workpiece, whereby the gap is ionized, and a breakdown is generated. A machining discharge pulse is applied immediately after the breakdown.

[0056] The first voltage U.sub.CH1, between the upper current feeder and the workpiece, and the second voltage U.sub.CH2, between the lower current feeder and the workpiece are measured when the discharge current is zero, and their difference ΔU.sub.CH=U.sub.CH1−U.sub.CH2 is determined. The discharge position of said machining discharge pulse is computed based on said difference ΔU.sub.CH. FIG. 3a,b illustrates the machining area of a WEDM, with a voltage measurement means for the measurement of the voltages U.sub.CH1 and U.sub.CH2, used by the discharge position detection circuit according to the invention.

[0057] Referring now to FIG. 4a, there is shown a circuit model of a first embodiment, illustrating the current path between the generator of the WEDM and the wire electrode. In this embodiment, the machining discharge pulses are provided to the wire by both, the upper electrode cable (25) and lower electrode cable (35). In general, one or more upper- and lower electrode cables connect the generator 7 to the upper current feeder C.sub.PU and to the lower current feeder C.sub.PL. The arrows of currents i.sub.0, i.sub.1 and i.sub.2 point towards the generator since usually the workpiece is polarized positively and the wire electrode negatively. For simplicity, when illustrating the formulas for U.sub.CH1 and U.sub.CH2 a reverse polarity is assumed (wire positive and workpiece negative).

[0058] L.sub.U and L.sub.L are respectively the upper- and lower electrode cable inductances; R.sub.U and R.sub.L represent the upper- and lower electrode cable resistances respectively including the resistance of the current feeder C.sub.PU, C.sub.PL.

[0059] As explained above with reference to the prior art, the influence of inductances and resistances L.sub.U, R.sub.U, L.sub.L, R.sub.L must be taken into account when measuring the partial currents i1 and i2. Therefore, any change in these values affects the measurement, making it unreliable, e.g. when machining at different positions of the worktable.

[0060] As shown in FIG. 3b, s.sub.U is the partial wire segment length between the discharge location and the upper current feeder C.sub.PU, and s.sub.L is the partial wire segment between the discharge location and the lower current feeder C.sub.PL. R.sub.WU and R.sub.WL are the partial wire segment resistances, L.sub.1 and L.sub.2 are the partial wire segment inductances above and underneath the discharge location. U.sub.CH1 and U.sub.CH2 are respectively the voltages measured at the upper and lower current feeder with respect to the ground.

[0061] During the discharge, the gap voltage U.sub.Gap between the wire and the workpiece is typically 25-30 V. This voltage is the discharge voltage or burning voltage. The gap voltage is independent from the discharge location; the gap voltage must be eliminated in order to get the relevant voltage ΔU.sub.CH for the determination of the discharge position.

[0062] The following formulas yield both voltages:

U.sub.CH1=i1*R.sub.WU+L.sub.1*di.sub.1/dt+U.sub.Gap  {circle around (1)}

U.sub.CH2=i2*R.sub.WL+L.sub.2*di.sub.2/dt+U.sub.Gap  {circle around (2)}

[0063] According to the invention disclosed herein, the voltage is sampled exactly, with typically 50 ns accuracy at the time where the discharge pulse current is zero, as shown e.g. in FIGS. 6a, 6b. In other words, the voltage must sampled pretty much exactly at the detection of the zero pulse current. For the avoidance of doubts, the wording “the first voltage U.sub.CH1 and the second voltage U.sub.CH2 are measured when current of the machining discharge pulse is zero” includes a voltage measurement in which the current is nearly zero, i.e. within the required accuracy.

[0064] In this way, the resistive voltage drop caused by the current on the wire segment is zero. Above formulas {circle around (1)} and {circle around (2)} are thus simplified:

U.sub.CH1=L.sub.1*di.sub.1/dt+U.sub.Gap  {circle around (3)}

U.sub.CH2=L.sub.2*di.sub.2/dt+U.sub.Gap  {circle around (4)}

[0065] Moreover, their difference is:

ΔU.sub.CH=L.sub.1*di.sub.1/dt−L.sub.2*di.sub.2/dt  {circle around (5)}

[0066] One sees that both voltages U.sub.CH1, U.sub.CH2 depend on the derivative of the current di.sub.1/dt, di.sub.2/dt and on the partial wire inductances L.sub.1, L.sub.2.

[0067] Moreover, the voltages U.sub.CH1 and U.sub.CH2 are measured at the current feeders, so that the impedance of the cabling from the generator to the current feeders is now irrelevant.

[0068] These aspects are now illustrated again with reference to the circuit model shown in FIG. 4a. By measuring the voltages U.sub.CH1 and U.sub.CH2 at the current feeders, the electrode cables 25, 35 are by-passed; only the relevant part of the circuit is measured. Moreover, by measuring at zero discharge pulse current, the partial wire segment resistances R.sub.WU and R.sub.WL have no effect to the voltages U.sub.CH1 and U.sub.CH2. Lastly, the measured channel voltages, respectively the first voltage and the second voltage U.sub.CH1 and U.sub.CH2 are subtracted from each other, so that the voltage difference ΔU.sub.CH is not affected by the gap voltage U.sub.Gap.

[0069] In summary, the voltage difference ΔU.sub.CH is only determined by the partial wire segment inductances L.sub.1 and L.sub.2, and the derivatives of the discharge pulse currents di.sub.1/dt, di.sub.2/dt. Moreover, in practice the current difference I.sub.1, I.sub.2 between the upper and lower paths only minimally affects the derivative of the current. FIG. 8 shows exemplary capacitive discharge pulse waveforms of different upper- and lower currents, i1 and i2. The derivative di/dt is only slightly affected by the difference of the discharge currents i1 and i2. This means, that the voltage difference ΔU.sub.CH is mainly determined by the partial wire segment inductances L.sub.1 and L.sub.2.

[0070] ΔU.sub.CH contains the information regarding the exact discharge location: The inductance L.sub.1 of the partial wire segment between the upper current feeder and the discharge position, and the inductance L.sub.2 of the partial wire segment between the lower current feeder and the discharge position are a function of the respective partial wire segment lengths s.sub.U and s.sub.L, from the upper-, respectively the lower current feeder to the discharge location.

[0071] According to Rosa, E. B. (1908). “The self and mutual inductances of linear conductors”. Bulletin of the Bureau of Standards. U.S. Bureau of Standards. 4 (2): 301 ff. doi:10.6028/bulletin.088, the relation between self-inductance L of a straight wire and the wire segment length is approximately:

L=k*s*[ln(2*s/r)−1]  {circle around (6)}

[0072] Where k=200 nH/m; r=wire radius in meters; s=relevant wire segment length in meters, e.g. the upper wire segment length s.sub.U; L=wire inductance in nH.

[0073] The formula {circle around (6)} shown above is transcendental and cannot be solved easily in real time. However, lookup tables can be used to link the wire segment length to the self-inductance L. Such lookup tables can be calculated in advance for the relevant of wire electrode diameters and with the required step size. The lookup tables can be stored e.g. in a fast memory, for real time access. Lookup tables are practical to discretize the detected discharge position, in view of a further use, e.g. for discharge position monitoring purposes or for discharge position control.

[0074] Moreover, the above formula {circle around (6)} can be linearized to find the approximate wire segment length s, as exemplarily shown in FIG. 11 for a wire having diameter 0.2 mm:

L=1.3138*s−3.5483

s=(L+3.5483)/1.3138 [0075] (e.g. using a simple linear interpolation)

[0076] Where s=wire segment length in meters, L=wire inductance in nH.

[0077] This is meant to show that the voltage, and hence the voltage difference, contains the information about the wire segment lengths. However, the present invention proposes a simpler method to directly find the wire segment length starting from the voltage difference. This will become clearer later.

[0078] Even for smaller currents, the derivatives di.sub.1/dt and di.sub.2/dt are quite large, since in order to achieve high quality results the WEDM machining pulses must be as narrow and steep as possible. Therefore, the measure is reliable not only for the main cut, but also for the trim cuts.

[0079] The relevant theory, i.e. self-induction is now illustrated with reference to FIGS. 6a and 6b, which show digital storage oscilloscope (DSO) acquisitions of the current of a machining discharge pulse, and the voltage difference ΔU.sub.CH according to the discharge position, for a particular discharge occurring respectively at the top of the workpiece and at the bottom thereof.

[0080] In detail, FIG. 6a shows the course of the channel voltage difference ΔU.sub.CH, when the discharge occurs at the bottom of the engagement line of wire and workpiece. Further, FIG. 6b shows the course of the channel voltage difference ΔU.sub.CH, when the discharge occurs at the top of the engagement line of wire and workpiece.

[0081] The invention is based on the fact, that a voltage proportional to the derivative of the current is induced by the current pulse in the relevant wire portion s, between the upper- and the lower current feeder. Knowingly, a change in current through a straight wire will produce a magnetic field around the wire. Self-induction voltage is generated by magnetic field variations and is proportional to the derivative of the current pulse. Thus, as illustrated in the figures for the case of capacitive discharge pulse, the derivative of the current is steepest at the beginning and at the end of the pulse. Means, the induced voltage is at the maximum, exactly when the discharge current is at zero.

[0082] As said, FIGS. 6a and 6b illustrate the two extreme cases of a top respectively a bottom discharge position. In case of a discharge occurring exactly at the midpoint between the upper- and the lower current (FIG. 3b: Midpoint), the two partial wire segment lengths s.sub.U and s.sub.L, from the upper-, respectively the lower current feeder to the discharge location have the same length. The induced voltages U.sub.CH1 and U.sub.CH2 have the same value, and the voltage difference ΔU.sub.CH of the measured channel voltages U.sub.CH1 and U.sub.CH2 is thus zero.

[0083] Notably, the current gradient di/dt of a wire electrical machining discharge pulse is large at the beginning and at the end of the pulse. Therefore, the voltage signal is very suitable for detection, also in the case of trim cutting, where conventional discharge position detection fails.

[0084] Now, coming back to FIG. 3a, 3b there is illustrated a preferred embodiment in which the first voltage U.sub.ch1 and the second voltage U.sub.ch2 are measured using measuring cables which are connected respectively at the upper- and at the lower current feeder. Preferably, the measuring cable 26 is directly connected to the upper current feeder 21 for the measurement of the channel voltage U.sub.CH1, and a measuring cable 36 is directly connected to the lower current feeder 21 for the measurement of the channel voltage U.sub.CH2.

[0085] For instance, channel voltages U.sub.CH1 and U.sub.CH2 are measured by means of voltage measurement means V.sub.1 and V.sub.2, outside of the machining area, e.g. in a cabinet 8. Thus, the current feeders are branched to the voltage measurement means V.sub.1 and V.sub.2 by means of measurement cables 26, 36. These cables are not the electrode cables 25, 35, that carry the high current machining discharge pulse.

[0086] A particular advantage of this method is that no particular changes to the electrode cables are required. These electrode cables can be used unchanged. During the measurement, only the voltage between the current feeder and the workpiece is measured. The gap voltage and the small voltage drop through the workpiece is eliminated by subtracting channel voltage 2 from channel voltage 1.

[0087] Moreover, in WEDM, the gap voltage is measured routinely for real time process monitoring and control. Preferably, the measuring cables 26, 36 for the measurement of the channel voltages U.sub.CH1 and U.sub.CH2 are the same used to measure the gap voltage, a.k.a. “reference cables”, which carry a negligible current and have virtually no voltage drop, according to the principle of the Kelvin bridge. In this way, no additional cables are installed for the measurement of the channel voltages U.sub.CH1 and U.sub.CH2; complexity and costs are not needlessly increased. The workpiece side is typically at ground potential, as usual.

[0088] In summary, the invention poses no burden to the machining area of the WEDM; it is possible to implement the invention with no additional cables. However, dedicated measuring cables for the measurement of channel voltages U.sub.CH1 and U.sub.CH2 is also possible.

[0089] According to an embodiment, a machining process monitoring and control board includes the voltage measurement means V.sub.1 and V.sub.2, for the acquisition of the channel voltages U.sub.CH1 and U.sub.CH2. Such machining process monitoring and control board is routinely present in electrical discharge machines for real time process monitoring and control. Said e.g. printed circuit board may be mounted in a cabinet 8, further accommodating e.g. generator boards, power supplies, control unit, motor drives, etc.

[0090] The use of appropriate measuring cables and their proper connection are very important. Preferably, the measuring cables are coaxial cables or twisted pair cables, having a low resistance, inductance and capacitance. As an example, a specific inductance of 600 nH/m and capacitance of 28 pF/m will provide an accurate voltage signal. In this way, the measuring cables allow an accurate acquisition of the gap voltage.

[0091] FIG. 5c is a plot of the machining currents i1 and i2, which in this exemplary embodiment have the typical waveform of capacitive discharge pulses. However, the invention can be used with square, trapezoidal, and any other current waveform. In this example, the current i1 through the upper branch is higher than the current i2 through the lower branch, which means that the discharge position is closer to the upper current feeding point.

[0092] FIG. 5d is a plot of the course of the difference ΔU.sub.CH of measured voltages over time, for a plurality of discharges. The voltage curves are proportional to the difference of the pulse current differentials di/dt. The thick line illustrates the progression of the voltage difference ΔU.sub.CH of a discharge occurring at the top of the workpiece. The thin lines are examples to show the course of the voltage difference ΔU.sub.CH of discharges occurring at other positions along the engagement line of wire electrode and workpiece.

[0093] FIGS. 5a and 5b are current zero crossing signals, by which the voltage measurement is executed at the ideal moment. As discussed above, the channel voltages U.sub.CH1 and U.sub.CH2 are preferably measured at the beginning, or at the end of the discharge current pulse, when the current is zero and the current differentials di/dt are at the maximum. Particularly preferably, the channel voltages U.sub.CH1 and U.sub.CH2 are measured at the end of the discharge pulse, because here the noise is low compared with the noise at the beginning of the discharge pulse.

[0094] FIG. 7a is a comb-shaped workpiece 2′, which is processed over its entire height by means of the wire electrode 1. FIG. 7b is a digital oscilloscope measurement of the voltage difference ΔU.sub.CH over pulse time, which is captured in the course of the WEDM machining of the comb-shaped workpiece of FIG. 7a, by using the oscilloscope persistence display mode. The generated waveforms reflects the comb-shape of workpiece, and illustrates the relation between the discharge position and the voltage difference ΔU.sub.CH.

[0095] Referring to FIG. 9, the principle of signal acquisition and processing is now illustrated exemplarily:

[0096] The voltages U.sub.CH1 and U.sub.CH2 are acquired on the upper- and lower current feeders. Then a voltage divider and limiter circuit 51 scales down the measured voltages. Next, a fast operational amplifier 52 subtracts these voltages. This difference ΔU.sub.CH is fed to a sample and hold circuit 53, that samples the voltages when the current zero crossing signal is active. The subtractor- and sample and hold circuit are not detailed further, since they are well known in the art.

[0097] Then the voltage signal is input to a fast A/D converter 54 and is discretized. Lastly, the output of the A/D converter is input to a FPGA (field programmable gate array) component 55. The FPGA is used for the further processing of the discharge positions, for simple monitoring purposes, such as e.g. the detection of concentrated subsequent discharges, or to actively control the position of a machining discharge pulse, for the application of an arbitrary discharge pattern to the workpiece cutting surface (see patent application EP21213367.2), or other introductorily mentioned “uses of the discharge position detection”.

[0098] For practical use and further processing of the detected analogue discharge position, it is of advantage to discretize the discharge positions. To this purpose, in analogy to the method disclosed in EP3446820A, discretization is made by dividing the workpiece height H.sub.WP into a discrete number of N.sub.S distinct vertical sections S, e.g. starting from the lower edge of the workpiece (Z=0), and assigning the discharge position of each discharge to the respective vertical section. In the present example it is assumed, that the wire guide heads have symmetrical structure, i.e. the distance from the lower current feeder to the nozzle is the same for the upper wire guiding head 20 as for the lower wire guiding head 30.

[0099] Preferably, a lookup table is provided, indicating the vertical section S in which the discharge occurs for a voltage difference ΔU.sub.CH. The lookup table comprises data for each wire diameter used in the WEDM process.

[0100] According to a preferred embodiment, a discharge position is discretized by dividing a workpiece height H.sub.WP into a number N.sub.S of vertical sections S, and, with each discharge Di, determining the voltage difference ΔU.sub.CH, and then, based on said voltage difference ΔU.sub.CH, determining a discharge position Z.sub.Di of each discharge D.sub.i along an engagement line of a wire electrode and the workpiece, and lastly, assigning each discharge D.sub.i to a matching vertical section S.sub.j of the workpiece based on the determined discharge position Z.sub.Di of each discharge.

[0101] The present invention now discloses how the exact location of the discharge is easily found by knowing the voltage difference ΔU.sub.CH. Referring to the symbols of FIG. 2:

Z.sub.Li=H.sub.WP/2+k*ΔU.sub.CH

k=−H.sub.WP/(2*U.sub.max)

[0102] Where ΔU.sub.CH=U.sub.max is found, as illustrated by FIG. 10a, by slightly inclining the wire at an angle α (which wire length becomes now S.sub.adj) and translating with the inclined wire against the workpiece to provoke a short circuit discharge at the top of the workpiece. The wire can also be inclined as illustrated in FIG. 10b to provoke a short circuit discharge at the bottom of the workpiece. In this way, it is possible to adjust the circuit symmetry.

[0103] In theory, the absolute values of U.sub.max found by inclining as per FIGS. 10a and 10b should be identical. In this case ΔU.sub.CH=0 would correspond to the exact middle of the workpiece height. However, a difference ε=Umax.sub.upper−Umax.sub.lower can be used to slightly adjust the midpoint H.sub.WP/2.

[0104] The wire inclination α for the detection of the top and bottom edges of the workpiece can be small, e.g. <2°, so that H.sub.WP and S.sub.adj are nearly same. However, very small wire inclination α produces less distinct discharge positions. To prevent this, larger wire inclination α may be adopted, and compensated by computing the length of the inclined wire, as follows:

H.sub.WP=S.sub.adj*cos α

[0105] As discussed above, the voltage difference ΔU.sub.CH is proportional to the discharge position. The voltage difference ΔU.sub.CH of a discharge occurring at the midpoint between the electrode feeders is zero. With a symmetrical constitution of the wire guide heads (Z.sub.FU=Z.sub.FL) and with flushing nozzles proximal to the workpiece surface, this midpoint corresponds to the midpoint of the workpiece height. The discharge position is referred e.g. to said midpoint, at half distance Z.sub.FM between the current feeders 21, 31. As said, the distance of the actual discharge from the midpoint between the current feeders is determined by the voltage difference ΔU.sub.CH. For instance, a positive value of ΔU.sub.CH means that the discharge occurs between the midpoint and the top of the workpiece, a negative value of ΔU.sub.CH means that the discharge occurs between the midpoint and the bottom of the workpiece.

[0106] The following is a simplified example (e.g. the nonlinearity of the wire electrode self-inductance is neglected) for the determination of the discharge position Z.sub.Li based on the difference of the channel voltages U.sub.CH1 and U.sub.CH2:

H.sub.WP=100 mm

[0107] (Let the zero coordinate of Z.sub.Li be at the workpiece bottom) [0108] Wire diameter d=0.2 mm [0109] U.sub.max=100V; U.sub.CH1=33.3 V; U.sub.CH2=66.6 V

[0110] Discharge position Z.sub.Li=?

ΔU.sub.CH=U.sub.CH1−U.sub.CH2=−33.3 V

k=−H.sub.WP/(2*U.sub.max)=−0.5 mm/V

Z.sub.Li=H.sub.WP/2+k*dU=50+(−0.5)*(−33.3)=66.6 mm

[0111] As already mentioned above, the position detection system is calibrated, for instance by determining the voltage difference ΔU.sub.CH measured with discharges occurring at the top- and/or bottom of the workpiece surface, whereby the such discharges can unequivocally be assigned to a specific vertical sections S.

[0112] Discharges occurring at the top- and/or bottom of the workpiece surface can be generated deliberately, e.g. by tilting the wire in cutting direction, as shown in FIGS. 10a and 10b. Calibration can also be made taking advantage of the workpiece height information stored in the control unit.

[0113] Another calibration includes the processing of a comb-shaped workpiece, similar to the one shown in FIG. 7a, e.g. having a number of protrusion equivalent to the number of vertical sections S. The scale can be easily adjusted in process, by looking for the top- and bottom discharge positions of a distribution of a significant number of discharges.

[0114] If needed, the measure can be linearized, as illustrated by FIG. 11.

[0115] Referring now to FIG. 4b, there is shown a circuit model of a second embodiment, illustrating the current path between the generator of the WEDM and the wire electrode. FIG. 4b is similar to FIG. 4a, but the lower electrode cable is not present or simply disconnected. In fact, the method according to the preferred embodiment of the invention works also with a one sided supply.

[0116] The method applied to measure the first voltage U.sub.CH1 and the second voltage U.sub.CH2 of this second embodiment is similar to the one illustrated above with regard to the first embodiment, but now the formulas for U.sub.CH1 and U.sub.CH2 are as follows:

U.sub.CH1=L.sub.1*di.sub.1/dt+U.sub.Gap [0117] U.sub.CH2=U.sub.Gap (since there flows no current past the partial wire segment s.sub.L, between the discharge point and the lower current feeder)

[0118] The voltage difference is:

ΔU.sub.CH=U.sub.CH1−U.sub.CH2=L.sub.1*di.sub.1/dt

[0119] The measure of U.sub.CH2 serves only to determine and subtract the gap voltage U.sub.Gap from the formula.

[0120] The various applications of the discharge position detection known in the art, including the applications listed introductorily may use the improved method disclosed by the present invention, taking advantage of an improved accuracy, robustness, and its inexpensiveness.

[0121] Moreover, the inventive discharge position detection method may also be used in combination, i.e. “fusion” or plausibility check with known discharge position detection methods, e.g. the mentioned partial currents detection method, by workpiece height detection methods or, geometrical data about the workpiece and/or the machining program for instance taken from CAD/CAM data.