DE2000143A1 - Device and method for controlling spark machining - Google Patents

Description

The invention relates to material processing by spark erosion and relates in particular to a method for controlling a spark machining process and an improved embodiment of an electrical discharge machine which is used in the method.

Spark processing machines usually consist of an apparatus body or a support structure, as well as a tool carrier for holding a tool electrode, a work surface for fastening a workpiece and a dielectric liquid, such as paraffin oil, which is in a tank or a container in which the workpiece is in operation the machine, and an electrical supply circuit for providing a series of successive voltage pulses at a controlled repetition rate to a circuit containing the tool electrode and the workpiece to create a series of spark discharges therebetween to enable material to be removed from the workpiece.

The terms "spark" and "sparking" as used below refer not only to short-term discharges (ie microseconds), but also to those that are relatively long-lasting (ie milliseconds), often referred to as arcing will.

The tool electrode is normally made of carbon in the form of graphite and, during operation of the spark machining device, small pieces of carbon are detached from the tool electrode and are deposited on the machining piece. If this occurs, successive current pulses can be detected between the graphite electrode and the carbon particles deposited on the workpiece.

Such a current flow is very stable and can cause more serious damage to both the electrode and the workpiece, even if it can only flow for a very short period of time.

This state, i.e. a steady flow of a pulse current between a graphite electrode and the carbon particles attached to the workpiece, is known as a burning state.

The main object of the present invention is to propose a method for controlling a spark machining process and an apparatus for carrying out this method, in which a burning state is quickly determined and remedied.

According to one embodiment of the present invention, a method for controlling a spark machining process is proposed, which consists in monitoring the voltage and / or the current inflow at the spark gap in order to be able to determine the presence of a high-frequency component in the system voltage while a current is flowing, and to ensure that the successive voltage pulses in the spark gap stop if no high-frequency component is found in the gap voltage during spark formation.

The tool electrode can then be automatically withdrawn from the workpiece and the dielectric liquid in which the process takes place can circulate around the workpiece so that the surface thereof is cleaned.

According to a further embodiment of the present invention, a spark machining device of the type referred to above is proposed, which has a control circuit with devices for determining the presence of a spark and devices for interrupting the delivery of successive voltage pulses to the spark gap in the event that, while a spark exists, no high frequency component is detected in the spark gap voltage.

Devices can also be provided for automatically withdrawing the tool electrode from the workpiece if the delivery of the successive voltage pulses to the spark gap is consequently interrupted.

The voltage present at the spark gap can be fed via a high-pass filter to a comparison circuit in order to feed this a first input signal which indicates the presence or absence of a high-frequency component in the spark gap voltage, while a second input signal fed to the comparison circuit is a current signal from an arrangement which is based on a Current responds, can be to indicate whether or whether no current is flowing to the spark gap. In this case, the comparison circuit is designed such that the supply of the voltage pulses to the spark gap can be interrupted when a pulse current is flowing, while no frequency component is detected in the spark gap voltage.

For a more detailed explanation, the invention is described in more detail with reference to the figure. Here show:

Figs. 1A to 1C show the respective waveforms of the spark gap voltage in the case of open circuit, normal sparking and burning;

2 shows a schematic circuit diagram for controlling a spark machining device;

FIGS. 3A to 3E correspond to circuits which represent parts of the schematic circuit of FIG. 2;

FIGS. 4A to 4C show the corresponding output voltage of a high-pass filter, which is contained in FIG. 3A, during a state in which the spark gap is an open circle, normal sparking takes place and burning occurs, and

FIG. 5 shows the output signal from the first stage of the pulse-generating circuit which is contained in FIG. 3A, but only for the state of normal spark formation.

Reference is first made to FIG. 2, from which it can be seen that the spark gap 10 between a tool electrode and a workpiece is supplied with successive voltage pulses which are formed in a spark generator circuit 11. The actuation of the spark generator circuit 11 is controlled by the various control circuits, which are summarized diagrammatically at 12, so that it is ensured that the voltage pulse can only be generated when the correct conditions are present in the device. The voltage at the spark gap 10 is measured and the information derived therefrom is fed to a servo control circuit 13, which is designed such that it can control the movement of the tool electrode relative to the workpiece, so that it is ensured that the spark gap is kept at any predetermined value, as the spark machining process proceeds. The current flowing to the spark gap is also monitored by a short-circuit display circuit 14 which, in the event that an exceptionally high output current flows from the spark generator circuit 11, interrupts the delivery of successive voltage pulses within a few microseconds.

The circuits described so far are generally known and are therefore not explained further in detail.

The typical waveforms of the spark gap voltage for the various operating states shown in FIGS. 1A to 1C make it clear that the voltage at the spark gap when a normal spark is present (Vs) is lower than that which is present when there is no spark or an open circuit state (Vg) is present but is greater than the voltage that occurs during a burning state (Vb).

However, the difference between Vs and Vb is too small to have a reliable indication of whether there is proper sparking over a wide range of operating conditions, which is required for each individual piece.

However, high frequency ripple is observed on the spark gap voltage during normal sparking, but not when there is a burning condition or when the gap is open. It is this high frequency ripple that is used in the present invention to determine whether there is proper sparking; therefore, such a high frequency ripple is monitored with the aid of a fire detection circuit 15.

However, it is apparent that no such high frequency ripple is present on the waveform either in a burning condition or in an open condition. It is therefore necessary to determine at the same time whether there is a spark. This is simply done with a current-responsive arrangement, shown as a current detector circuit, which indicates the presence of a current greater than a certain minimum value flowing to the tool electrode. A current and a voltage signal are therefore compared in a comparison circuit 17, which is shown schematically in FIG.

The output from the comparison circuit 17 is fed to a logic circuit 18 which is designed such that it can cause the spark generator circuit 11 to stop generating successive voltage pulses for a period of a few seconds and then to resume the generation of such pulses. Under these conditions, the servo control circuit 13 responds and causes the tool electrode to be withdrawn from the workpiece and the dielectric liquid circulating through the tank containing the workpiece to flow freely over the machined surface, so that the graphite pieces that have settled on the workpiece and cause the burning condition can be washed away

When voltage pulses are again applied to the tool electrode, the servo control circuit 13 responds and causes the electrode to be advanced and the spark machining operation to be resumed.

If the graphite pieces have not been removed during the previous rinsing process, the firing process starts again and the process cycle described above is repeated. In order to prevent this sequence from being repeated indefinitely in the presence of a persistent error condition, the output of the logic circuit 18 is also fed to a storage circuit 19, which is designed such that it can supply the control circuit 12 with an output signal if more than a predetermined number Burning states are counted within a certain period of time, whereupon the control circuits 12 cause the spark generator circuit 11 to be completely switched off. In this case, the spark machining process can only be resumed after the fault has been removed by hand by an operator.

The only other condition that can be confused with the burning condition is the short circuit at the spark gap. In this case, too, a current would flow and there would be no high-frequency component in the spark gap voltage. The aforementioned spark machining devices, however, are usually provided with devices (referred to as short-circuit indicator circuit 14 in FIG. 2) in order to interrupt the voltage supply to the tool electrode if there is a short circuit in the spark gap. This can be determined by the increase in the flowing current above a predetermined maximum value, whereby under these circumstances the voltage supply is automatically interrupted and the tool electrode is automatically withdrawn.

Under certain circumstances, the burn detector circuit can be used to display both the burn status and the short circuit, so that no special short circuit detector circuit which only serves to display the short circuit status needs to be provided.

The operation of the various circuits 15 to 19 will now be described in detail with reference to the corresponding FIGS. 3A to 3E.

A signal is picked up from the spark gap 10 via a coaxial cable and fed to the input 15a of the fire detector circuit 15, which is shown in FIG. 3A. The signal is fed to a high-pass filter, consisting of a resistor R [low] 1 and capacitors C [low] 2 and C [low] 3, which is used to block all frequencies below 15 MHz. The output voltage from the high-pass filter is tapped by diodes MR5 and MR6 and at this point the output voltage has the forms shown in FIGS. 4a to 4c. This circuit is designed to provide almost the same amplitude output over a wide range of AC signal input amplitudes. The signal is fed to a two-stage amplifier consisting of transistors VT1 and VT2. Two levels of amplification are necessary because AC coupling of the emitter series resistors cannot be used because this would produce pulses that would have a duration of tenths of a microsecond in the open circuit state; therefore, if there is no decoupling capacitor, the amplification ratio is reduced, so that two amplifier stages are required.

The amplified AC signal from transistor VT2 is fed to a frequency-voltage inversion circuit composed of consensators C6 and C7, diode MR7, transistor VT3 and resistor R11. The waveform of the output at resistor R11 is shown in FIG. When the immediate input voltage at the collector of transistor VT2 is high, capacitors C6 and C7 charge and transistor VT3 becomes reverse biased and therefore non-conductive. On the other hand, when the input voltage is low, transistor VT3 is switched forward as the charge on capacitor C7 discharges capacitor C6, which is charged, which is 0-24 volts relative to the return line. When the input voltage rises again, the capacitor C7 charges even more and therefore the voltage across it reaches the line voltage within very few cycles of high frequency ripple.

This height is then maintained while this waviness is present. Between the successive pulses of the high frequency ripple, i.e. between these successive voltage pulses which are fed to the spark gap 10, the capacitor C7 discharges through the resistor R11, resulting in the waveform shown in Figure 5.

This waveform is converted into a true square wave with the help of a Schmitt trigger that contains the transistors VT4 and VT5 and their associated components. The square wave output from the Schmitt trigger is reversed by the transistor VT7 so that a square wave with an amplitude of 24 volts is present at the resistor R20 when there is proper sparking. The signal from the output 15b of the fire detector circuit 15 therefore consists of a sequence of 24 volt pulses which are coincident with the successive voltage pulses which are supplied to the spark gap 10, but only during normal spark formation. The output signal from the burning detector circuit 15 disappears when the open circuit is present, a short circuit on the track or when there is a fire on the track.

In order to be able to determine which of these conditions is present in the absence of an output signal from the fire detector circuit 15, the current detector circuit 16, which is shown in FIG. 3b, is provided.

The input signal for the current detector circuit 16 is tapped from a potential divider, which consists of the resistors 1R1 and 1R2, which are connected to a resistor RX, which is in the collector load circuit of the output transistor TRX of the spark generator circuit 11, which is not described in detail here. The resistor RX is in series with a tool electrode, so that a voltage drop occurs across this resistor when a current flows to the tool electrode. The voltage divider consists of the resistors 1R1 and 1R2 and is arranged in such a way that if, when a few milliamperes are exceeded, a current flows through the resistor RX, which generates a sufficiently high voltage across the resistor 1R2 to bring the transistor 1VT1 into the conductive state.

When this transistor is conductive, the voltage at its collector drops and consequently the voltage at the base of transistor 1VT2 drops enough to enable this transistor to go into the conductive state. The voltage at the collector of this transistor therefore rises approximately to the voltage of the supply line 24, i.e. up to 24 volts. Therefore, at this point in the circuit there is a train of voltage pulses corresponding to the pulses applied to the tool electrode.

The signal from the collector of the transistor 1VT2 is fed to the base of the transistor VT8, so that this also becomes conductive when the current flows through the resistor RX, which is in series with the tool electrode. As a result, the voltage at the collector of transistor VT8 drops enough to cause transistor VT9 to conduct, with the result that the voltage applied to the base of transistor VT10 increases to a value sufficient to cause this transistor to fail -to bring it into the conductive state, so that the voltage at its collector drops to zero.

The output signal from the output 16b of the current detector circuit 16 therefore contains the voltage which is applied to the resistor R25, this voltage being zero when a current flows through the resistor RX which is in series with the tool electrode and which is 24 volts when no current flows through this resistor.

The output signals from the fire detector circuit 15 and the current detector circuit 16 are fed to the corresponding inputs 17a1 and 17a2 of the comparison circuit 17, which is shown in detail in FIG. 3c. These two inputs are connected to the base of the transistor VT11, which is arranged in such a way that it can go into the conductive state when a voltage is supplied to one of the inputs 17a1 or 17a2.

The various possible states that can arise are summarized in the following table.

From this table it can be seen that the transistor VT11 is continuously conductive when there is an open circuit and normal spark formation and is conductive with an interruption when there is a short circuit or a burning condition. Under short-circuit conditions, however, the short-circuit detector circuit 14 responds within a few microseconds and interrupts the delivery to the spark gap 10 of the successive voltage pulses, so that the open circuit conditions then prevail.

The transistor VT11 can therefore only be switched off for periods longer than a few microseconds if a burning state occurs at the spark gap.

The signal from the collector of the transistor VT11 is passed through a diode
<nonreadable>
and a low-pass filter consisting of the resistors R29, R30, R31 and the capacitor C9, fed to the base of the transistor VT12. The time constant of the low pass filter is such that the transistor VT11 remains in the non-conductive state for more than 10 microseconds or, alternatively, repeatedly at a duty ratio greater than 1: 5 or at some other reasonable ratio in accordance with those in the Device is switched off under the prevailing operating conditions. It will be understood that since the current flowing across spark gap 10 takes the form of a series of pulses, a burning condition normally causes transistor VT11 to turn on and off repeatedly.

Since this transistor becomes non-conductive while a current flows through the spark gap when there is a burning state, it follows that the ratio of the non-conductive to the conductive periods of this transistor is equal to the ratio of the duration of the spark to the duration of the intervals between the consecutive sparks.

It is therefore understandable that the transistor VT12 only becomes conductive when a burning state appears at the spark gap.

The signal from the collector of the transistor VT12 is fed to a further transistor VT13 which acts as a phase inverter. Therefore, in all normal states, transistor VT13 is in a non-conductive state and the voltage across resistor R34 increases to approximately 24 volts. The voltage at the resistor R 34 represents the output signal which is taken from the output 17b of the comparison circuit 17 and is fed to the input 18a of the logic circuit 18 which is shown in FIG. 3d.

For purposes of describing the logic circuit, the use of a 24 volt signal applied to input 18a will be referred to as "1", while the use of a zero voltage will be referred to as "0" in accordance with common logic nomenclature.

The input 18a of the logic circuit 18 is fed to the logic units via a switch S1 which is designed such that it can select either the input 18a or a “0” derived from the return conductor 0. When the logic circuit 18 is switched on for the first time, the switch S1 is brought into the position shown in FIG. 3d, so that the input signal for the logic elements is a "0".

The logic circuit 18 consists mainly of or-not gates NOR1 to NOR7.

When logic circuit 18 is therefore switched on for the first time, a "0" signal is applied to contact 1a of OR gate NOR1 and a second "0" signal is applied to contact 1b of this gate in a manner which will be described below so that the output signal at contact 1
<nonreadable>
is a "1". This "1" is fed to the contact 2b of the OR-not gate NOR2, which creates a "0" at the contact 2c, which is fed back to the contact 1b of the OR-not gate NOR1 via the diode MR9. The "0" from contact 2c is also fed to contact 3a of the or-not gate NOR3, which creates a "1" at contact 3b. This "1" is fed directly to the input contact 4b of the OR-NOT gate NOR4, so that a "0" is produced on its output contact 4c. This output signal is fed via a diode MR13 and a resistor R35 to the input contact 7a of the OR-NOT gate NOR7, which causes a "1" to occur at the output at the contact 7b. The contact 7b is connected to the spark generator circuit 11, but the output "1" is blocked by the diode MR14, so that the spark generator circuit is ineffective.

The “0” at contact 2c of the or-not gate NOR2 is also fed to the contact 5a of the or-not gate NOR5, so that a “1” is produced at contact 5b. This output is fed to capacitor C10 through diode MR11 and resistor R36. Since this "1" represents a 24 volt signal, the capacitor C10 charges up quickly. The voltage on the capacitor C10 is fed as an input to the contact 6a of the OR-not gate NOR 6, so that a "1" enters the OR-not gate NOR 6 such that a "0" appears at the contact 6b. This zero is fed to the input contact 4a of the OR-not gate NOR4 and is also applied to the input contact 2a of the OR-not gate NOR2 via the diode MR10, so that the logic units are kept in the state described above.

The signal from contact 4c of OR gate NOR4 is also brought to output 18b of the logic circuit and fed to memory circuit 19 at input 19a. This storage circuit is shown in FIG. 3E, from which it can be seen that the input signal at 19a is fed to the base of the transistor VT14 and a capacitor C12 via a diode Mr 15 and resistor R39. Since the output signal from the OR gate NOR4 at contact 4c is a "0", the transistor VT14 is non-conductive and the capacitor C12 is not charged.

The voltage at the emitter of the transistor VT14 is fed to the base of a transistor VT15 via a Zener diode MR16 and resistor R41 and consequently the transistor VT15 is also non-conductive with the result that the voltage at its collector will be equal to the supply voltage on line 24 , ie 24 volts. The collector voltage is fed to the input contact 8a of the OR-NOT gate NOR6, so that in this case a "1" is entered. This creates a "0" at output contact 8b, which is fed to input contact Aa of an inverting amplifier A, so that a "1" is created at its output contact Ab. This output signal is applied to one side of a relay R2A, the other side of which is connected to the 24 volt supply line 24. As a result, no current flows through this relay and its contact A1 is held in its normally closed position as shown. These contacts are connected to the control circuit 12 and, when closed, allow the spark generator circuit 11 to be energized.

Therefore, when the logic circuit 18 is energized for the first time, the switch S1 is set to the position shown and the logic elements are set to this initial state without in any way effecting a spark machining operation which is currently in the process.

When the switch S1 is actuated in order to take the logic circuit 18 into use, the output signal from 17b of the comparison circuit 17 is fed to the first input contact 1a of the first OR gate NOR1. Assuming normal sparking occurs, this results in a "0" which remains applied so that the state of the logic elements remains unchanged.

If, however, there is a burning state, a 24 volt signal appears at output 17b, which is fed as a logic "1" signal to input contact 1a of OR-not gate NOR1. Since the input signal at its other input 2a is also a "0", a "1" appears at the output contact 2c. This is fed to the second input contact 1b of the first or-not gate NOR1 via the diode MR9 in order to effect a holding circuit.

The "1" from the output contact 2c of the OR-NOT gate NOR2 also becomes the input contact
<nonreadable>
of the or-not gate NOR3, so that at the output contact
<nonreadable>
and therefore a "1" at the output contact 4e of the OR-NOT gate NOR4.

This "1" is fed to the input contact 7a of the OR-not gate NOR7 via the diode MR13 and resistor R35, so that a "0" is produced at its output contact 7b. This “0” contains a zero voltage state with the result that the current can now flow from the spark generator circuit 11 through the diode MR14. Such a current is established to bring the output transistor TRX into a non-conductive state, so that the supply of successive voltage pulses to the spark gap immediately disappears and does not reappear until the "0" state of the output contact 7b of the OR gate NOR7 is eliminated.

When the supply of the successive voltage pulses to the spark gap disappears, the servo control circuit 13 responds in a known manner and causes the tool electrode to be withdrawn from the workpiece, so that the circulation of the dielectric fluid around the latter increases significantly, so that any pieces of graphite are washed away that are on the workpiece and that caused the burning condition.

In order to remove the "0" state from contact 7b of the or-not gate NOR7, the "1" at the output contact 2c of the or-not gate NOR2 is fed to the contact 5a of the or-not gate NOR5. This creates a "0" at the output contact 5b of this or-not gate, so that the capacitor C10 begins to discharge via the input contact 6a of the or-not gate NOR6. This discharge takes place in a controlled manner and, when the capacitor C10 has discharged sufficiently, a "1" appears at the output contact 6b of the OR-not gate BOR6. This "1" is fed to the input contact 4a of the OR-NOT gate NOR4, so that the "0" at its output contact 4
<nonreadable>
immediately becomes a "1" and the output transistor TRX of the spark generator circuit 11 is again enabled to become conductive, so that the supply of the successive voltage pulses to the spark gap takes place again. This has the effect that the servo control circuit 13 responds in a known manner, so that the tool electrode comes into its working position and the spark machining process begins again.

At the same time, a "1" is fed via the diode MR10 from the contact 6b of the or-not gate NOR6 to the contact 2a of the or-not gate NOR2. This results in a "0" immediately at the output contact 2c, so that the hold state is interrupted, which has existed via the diode MR9 at the contact 1b of the OR-not gate NOR1.

When the piece of graphite on the workpiece has been removed and normal sparking begins again, the 24 volt signal at output 17b of comparison circuit 17 will have disappeared, so that the input at contact 1b of OR-not gate NOR1 will again be a "0" . Under these conditions, the logic circuit 18 immediately reverses to its initial state, since the capacitor C10 is rapidly discharging.

This operating cycle can then be repeated whenever another burn condition arises. However, the state of the memory circuit 19 immediately after a burning state is removed is different from that which existed before the burning state arose. The time in which the output transistor TRX of the spark generator circuit 11 is held in the non-conductive state is determined by the value of the capacitor C10 in the logic circuit 18, since this controls the time required for the input to the capacitor 6a of the OR -Not required gate NOR6 to convert from a "1" to a "0". The value of capacitor C10 is normally chosen so that this time is about 4 seconds. During this period of 4 seconds, a "1" appears at the output contact 4c of the OR-NOT gate NOR4 in the manner described above. This signal is applied to input 19a of storage circuit 19, thereby causing capacitor C12 to charge at a rate determined by the value of resistor R39.

This is selected so that the voltage on capacitor C10 increases from zero to about 5 volts. As a result, the voltage applied to the resistor R40 in the emitter circuit of the transistor VT14 also rises to approximately 5 volts because of the emitter follow-up effect of this transistor. The Zener diode MR16, however, ensures that the transistor VT 15 remains in the non-conductive state.

If the "1" is removed from the output contact 40 of the OR gate NOR4 and the successive voltage pulses are delivered again to the spark gap, the capacitor C12 slowly begins to discharge via the emitter circuit of the transistor VT14, which has a time constant of approximately 30 seconds owns. Therefore, if no burning condition occurs again within about one minute, the capacitor C12 is almost completely discharged and the storage circuit 19 is brought to its initial condition.

If another burn condition is detected within a few seconds, taking into account the fact that there is a time delay of a few seconds before sparking is resumed due to the time it takes for the servo system to return the tool electrode to the working position, it disappears the "1" state at contact 4c of OR gate NOR4 again, so that the 24 volt supply at input 19a of memory circuit 19 occurs again. Therefore, the capacitor C12 continues to charge at a rate determined by the value of the resistor R39 until the "1" is removed from the output contact 4c of the OR gate NOR4. Under these conditions the voltage on capacitor C12 will rise to just below 10 volts.

Assuming that the operating voltage of the Zener diode MR 16 is 10 volts, the transistor VT15 will not yet be conductive because of the continuing blocking effect of the Zener diode.

However, if a third burn condition occurs in rapid succession, capacitor C12 will continue to charge and its voltage will rise above 10 volts, typically to approximately 13 volts.

Under these conditions, the Zener diode MR16 becomes conductive and the signal applied to the base of the transistor VT15 is effected so that this also becomes conductive. As a result, the voltage at the collector of this transistor drops to almost zero, so that a logic zero signal is fed to the input contact 8a of the OR-not gate NOR8. This causes a "1" output signal from the output contact 8b, which is fed to the inverting amplifier A, as a result of which a "0" signal is produced at the output contact AB of the same. Under these conditions, a "24 volt potential difference occurs across relay RLA, which is therefore energized so that its normally closed contacts A1 are opened. This is done to cause control circuit 12 to energize spark generator circuit 11 so that the spark machining operation terminates and is not automatically restarted.

Under these conditions, it is necessary for an operator to check the work piece and ensure that the burning condition is corrected before restarting the spark machining operation.

This arrangement ensures that, in the event of a persistent burning state, the spark machining process is interrupted before any damage to the electrode or the workpiece can occur, whereas without such a memory circuit the spark machining process would be repeatedly interrupted and resumed as long as the burning state exists or until an operator recognizes this condition and takes the appropriate remedial measures by hand.

The number of burning states which must be determined in quick succession in order to complete the spark machining operation can of course be varied by suitable settings of the values of the switching components so that the response of the memory circuit can be varied, if necessary, in accordance with the local conditions .

Instead of using the output signal from the memory circuit positively in order to end the spark machining process, it is alternatively also possible to use this output merely to generate a signal indicating the existence of a persistent fault condition so that the operator's attention is drawn to the need for manual control to clear the fault condition.

In the device described, the voltage applied to the spark gap and the current flowing to the spark gap are monitored by observing various parts of the circuit. However, in some cases it may be possible to obtain both pieces of information by observing a single part of the circuit.

Claims (11)

1. A method for controlling a spark machining process, dg that the voltage at the spark gap and / or the flowing current are monitored in order to determine the presence of a component of the high-frequency ripple in the system voltage during the current flow, and that the delivery of the successive voltage pulses to the spark gap is interrupted if no high-frequency component is found in the line voltage during the sparking. 2. The method according to claim 1, dg that when the delivery of the successive voltage pulses to the spark gap is set, the tool electrode is automatically withdrawn from the workpiece and that the dielectric liquid in which the process takes place circulates around the workpiece in order to be machined Clean surface. 3. The method according to claim 2, dg that the spark machining process is automatically terminated when the delivery of the successive voltage pulses to the spark gap is interrupted due to the absence of any high frequency component detected in the gap voltage when the spark formation
<nonreadable>
Number occurs within a certain period of time. 4. Spark machining device for performing the method according to claim 1 to 3, dg that a support structure with a tool head for attaching a tool electrode made of graphite and a work surface for supporting a workpiece are present and that the workpiece is immersed in a container with a dielectric liquid, when the machine is in operation and that an electrical supply circuit for supplying a series of successive voltage pulses at a controlled repetition rate to a circuit containing the tool electrode and the workpiece to produce a series of spark discharges therebetween so as to be released from the workpiece To remove material, wherein a control circuit is provided, the devices (1 5) for detecting any high frequency component in the spark gap voltage as well as devices for determining the existence of a spark and devices (17, 18) for interrupting the delivery of successive voltage pulses to the spark gap (10) in the event that no frequency component is found in the spark gap voltage while a spark is present. Apparatus according to claim 4, d.g. that the control circuit comprises means (13) for automatically withdrawing the tool electrode from the workpiece when the supply of successive voltage pulses to the spark gap (10) is so interrupted. 6. Apparatus according to claim 4 or 5, dg that the control circuit comprises a comparison circuit to which the voltage at the spark gap (10) is fed via a high-pass filter in order to provide the comparison circuit (17) with a first input signal which indicates the on or Indicates absence of a high frequency component in the spark gap voltage, the control circuit further comprising a current responsive arrangement (16) for generating a current signal indicating whether or not current is flowing to the spark gap, the current signal of the comparison circuit as a second input signal is fed. 7. Apparatus according to claim 6, d.g. that the output from the comparison circuit is arranged so that the delivery of the successive voltage pulses to the spark gap can be interrupted in the event that a current pulse flows while no high frequency component is detected in the spark gap voltage. 8. Apparatus according to claim 7, dg, that the control circuit further comprises means (14) for determining the existence of an extraordinarily large current flowing to the tool electrode so as to determine a short-circuit condition on the route, and which respond to the delivery taking place one after the other To terminate voltage pulses to the spark gap almost immediately, the comparison circuit being set up in such a way that it responds more slowly in order to be able to differentiate between a burning state and a short circuit on the path. 9. Apparatus according to claim 7 or 8, dg that the output from the comparison circuit (17) is fed to a memory circuit (19) which is designed such that it can end the spark machining process if more than a certain number of burning states are present of a current flowing to the tool electrode can be detected in the absence of a high frequency component in the spark gap voltage within a certain period of time. 10. Apparatus according to claim 9, dg that by determining a burning state the supply of successive voltage pulses to the spark gap (10) are interrupted for a certain time and that the memory circuit (19) has a capacitor (C12) which is designed in such a way that its state of charge is changed at a certain speed during this certain period of time and that a change in the state of charge of the capacitor by more than a certain amount can terminate the spark machining process. 11. Apparatus according to claim 10, dg that the specific period of time in which the delivery of successive voltage pulses to the spark gap is interrupted in the event that a burning state is detected, is controlled by a further capacitor (10), the charge state of which changes at a specific speed .