The main problem is that the cause and effect of certain faults are still unclear. More research is required to find out which sensors are more adequate for moni- toring specific faults so that redundant information can be minimized. Further- more, more work has to be done to reduce the size and cost of the sensors so that a ‘complete sensing’ approach can be seriously taken into account. 277 4.5.5 References 1 Barthel, K., Trunzer, W., Laser-Praxis (1994) 10. 2 Anon., Opto Laser Eur. (1997) 3. 3 Schwede, H., Kramer, R., in: Proceedings of ICALEO 1998, Orlando, FL; www.primes.de. 4 Eg, Prometec UFC60; www.prometec.com. 5 Www.precitec.com. 6 Friedel, R:, EuroLaser (2000) 1. 7 Tönshoff, H. K., Schumacher, J., in: Proceedings of ICALEO 1996, Detroit, MI. 8 Ikeda, T., Kojima, T., Tu, J. F., Ohmura, E., Miyamoto, I., Nagashima, T., Tsubo- ta, S., Ishide, T., in: Proceedings of ICA- LEO 1999, San Diego, CA. 9 Www.precitec.som/PPS130E.htm. 10 Rinke, M., Dissertation; Düsseldorf: VDI- Verlag, 2000. 11 Graumann, C., Dissertation; Düsseldorf: VDI-Verlag, 1998. 12 Tönshoff, H.K., Alvensleben. v. F., Os- tendorf, A., Hillers, O., Stallmach, M., in: Proceedings of Photonics East, 1999, Boston, MA. 13 Seto, N., Katayama, AS., Matsunawa, A., in: Proceedings of ICALEO 1999, San Diego, CA. 4.6 Electrical Discharge Machining T. Masuzawa, University of Tokyo, Tokyo, Japan 4.6.1 Introduction Electrical discharge machining (EDM) has been known as one of the non-tradi- tional machining processes since its invention in early the 1940s. However, cur- rently, EDM is one of the most commonly used processes in various kinds of fac- tories. Most EDM machines are used in the field of die and mold making. The material removal mechanism in the case of EDM is completely different from that in the case of highly conventional machining processes such as cutting and grinding. It is based on melting and vaporization of workpiece materials by heating introduced by a spark or, more precisely, by a transient arc discharge. Thus the machining force is a minor or, rather, a negligible parameter in the control of this process. On the other hand, precise, dynamic control of the spark gap is essential for stable repetition of the discharge. These features make the control system unique and, consequently, the sensing strategy must be different from that in the case of conventional machining processes. Sensors in Manufacturing. Edited by H.K. Tönshoff, I. Inasaki Copyright © 2001 Wiley-VCH Verlag GmbH ISBNs: 3-527-29558-5 (Hardcover); 3-527-60002-7 (Electronic) 4.6.2 Principle of EDM From the viewpoint of shape specification of the workpiece, EDM is included in the same group as cutting where the shape of the tool is copied on to the work- piece or further modified according to the trajectory of the tool movement (rela- tive to the workpiece). The tool is referred to as the electrode. The main difference between EDM and conventional machining processes such as cutting is in the mechanism of material removal. The material removal process in EDM is illustrated in Figure 4.6-1. A voltage pulse from a pulse generator initiates and maintains a short arc dis- charge through the dielectric fluid in the gap between the electrode and the work- piece. The discharge heats the surface layer of the workpiece locally and melts (and also partially vaporizes) a small part of the workpiece. The simultaneous va- porization of the fluid produces a high pressure and the molten material disinte- grates into microsize particles, forming a shallow crater on the workpiece surface. The process leading to this crater formation is repeated over the entire work- piece surface facing the electrode, producing a thin cavity layer on the workpiece. By continuing this process with the feed of the electrode facing downward, the en- tire electrode shape is replicated on to the workpiece. 4 Sensors for Process Monitoring278 Fig. 4.6-1 Principle of EDM The discharge also forms a crater on the electrode. However, specifically chosen discharge conditions can minimize the size of the crater to, for example, less than 0.1% of the crater size on the workpiece. 4.6.3 Process Control Discharge occurs only when the gap distance is within a certain range. If the dis- tance is too large, the discharge cannot be initiated. If the distance is too small, a short-circuit occurs after the discharge, because a bridge is formed by the molten material. This short-circuit terminates any further discharge. Because of the above characteristics, gap control is the main issue in the case of EDM control A constant electrode feed rate is not usually adopted, because the occurrence of the discharge is a more or less random process and the speed of removal in the electrode feed direction is not constant. Therefore, the control strategy involves maintaining a constant gap distance throughout the operation. However, direct in-process measurement of the gap dis- tance is almost impossible. The main reasons for this are that (a) the gap is very narrow, 0.5–50 lm., and (b) the gap distance varies throughout the machining area and the nearest point determines the occurrence of the discharge and short- circuit. For these reasons, indirect detection of the gap distance is adopted in all EDM machines. The basic control algorithm is shown in Figure 4.6-2. 4.6.4 Sensing Technology In EDM, sensing technology is mainly required for obtaining information on the gap condition. As described previously, the information finally required is on the gap distance, which cannot be obtained directly. In order to overcome this diffi- culty, relationships between the gap distance and other detectable parameters have been analyzed. The parameters that represent the state of the gap condition are used as control parameters. Fortunately, as the removal mechanism of EDM is based on an electrical phe- nomenon, useful information is obtainable from the electrical parameters of the circuit for discharge. The most frequently used parameters are the gap voltage and the current through the gap, as shown in Figure 4.6-3. Some other parameters can also provide information on the machining status. The more promising ones are the electromagnetic radiation from the arc and the acoustic radiation caused by discharges. In the detection of these parameter values, sensors may play a useful role. To date, sensors have played a minor role in EDM control, because the more useful parameters, such as the voltage and the current, can be detected without special sensors. However, in some cases, sensors offer merits from the viewpoints of im- proved technology and economics. 4.6 Electrical Discharge Machining 279 The following is an overview of the major technologies used in obtaining gap information. 4.6.4.1 Gap Voltage In EDM, the gap voltage contains numerous information. The status of the gap is indicated by the gap voltage as shown in Figure 4.6-4. The figure shows only typical examples but there also occur various transient and combined types of voltage waveforms. These voltage pulses are input into the control system and statistical calculation is performed which provides the neces- sary values that can be compared with the set values for proceeding according to the flow chart shown in Figure 4.6-3. In this detection of gap voltage, no special sensor is usually required. A conven- tional high-impedance probe commonly used for measuring equipment such as an oscilloscope can be used without any problems. The voltage range is usually from 50 to 200 V. 4 Sensors for Process Monitoring280 Fig. 4.6-2 Basic control algorithm for EDM When very short pulses are applied for micromachining or very fine finishing, a special pulse circuit called a ‘relaxation-type generator’ is often used. In such a case, obtaining the input from a probe is difficult because the impedance of the gap is too high when it is open-circuited. Moreover, the input capacitance of the probe may change the energy of the discharge. Special voltage sensors may pro- vide a solution to this problem, but in most cases the detection of the gap voltage is abandoned and replaced by the detection of the current through the gap. This method is discussed in the next section. 4.6.4.2 Current Through Gap Since discharges for EDM are the current flow from the electrode to the work- piece (or in the reverse direction), the dynamic value of the current through the gap contains numerous information. The typical relationship between the current waveform and the status of the gap is shown in Figure 4.6-5. As is apparent when Figures 4.6-4 and 4.6-5 are compared, the current waveform can be as useful as the gap voltage described earlier. Since the difference between the cases of normal discharge and abnormal arc is only the high-frequency component in the wave- form, control using current information requires more careful data processing than that using gap voltage. For detecting the current waveform, two types of devices are currently used. (A) Shunt resistor. The most conventional and reliable method involves using a low-resistance resistor inserted in the loop of the gap current flow, as shown in 4.6 Electrical Discharge Machining 281 Fig. 4.6-3 Parameters used for controlling EDM Fig. 4.6-4 Voltage change according to the status of gap Figure 4.6-6. In a simple control system, a general-purpose solid resistor can be used as a detector. However, when more detailed information is required, a spe- cially designed shunt is necessary because the current signal contains a very high- frequency component. A suitable design for a shunt is illustrated in Figure 4.6-7, where the residual inductance can be sufficiently suppressed. (B) Current probe with Hall sensor. Another method involves using a commercial current probe which outputs current waveforms only by clipping one of the con- necting wires to the electrode and the workpiece. In this application, careful atten- 4 Sensors for Process Monitoring282 Fig. 4.6-5 Current change according to the status in gap Fig. 4.6-6 Circuit for current detection Fig. 4.6-7 A shunt for current detection tion is required because the direct current (DC) level is important in EDM control. The probes customized alternative current (AC) cannot be used, or at the very least requires complicated data processing. 4.6.4.3 Electromagnetic Radiation Since transient arc discharges radiate a wide spectrum of electromagnetic waves, medium-frequency (MF), high-frequency (HF), or very-high-frequency (VHF) re- ceivers can offer useful information concerning discharges. A small dipole (centi- meters long) fixed near the machine head can detect the radiation. The signal is processed through bandpass filters and analyzed. This method is particularly ef- fective in distinguishing normal discharges from other types of current such as abnormal arcs and short-circuits, because the radiation mainly occurs during nor- mal discharge pulses except during the rise and fall of pulses. Arc discharge also radiates high intensity light. The spectrum and intensity of this light contains information on the discharge status. However, this information has not yet been practically applied. 4.6.4.4 Acoustic Radiation The rapid rise and fall of temperature at the point of discharge produces shock waves and pressure pulses. This causes acoustic radiation. The spectrum contains the audio-frequency range, and one can audibly determine whether discharges are taking place in the gap. This signal can be detected using conventional micro- phones. However, the dynamic range of a microphone is usually not sufficient to cover the rough machining to very fine finishing range, because the pulse energy ranges from 1 J to 0.1 lJ. Detecting acoustic emission by attaching sensors on the workpiece or on the electrode (or its holder) is another approach for obtaining acoustic information. Since the generated acoustic spectrum in fine finishing tends to be more intense in the higher frequency range, a bandpass filter designed to transfer HF or VHF may be effective in obtaining information covering a wider range of discharge en- ergy distribution. In this type of application, commercially available acoustic emis- sion (AE) sensors can be used. This type of detection technique is, however, also still under development. 4.6.5 Evaluation of Machining Accuracy In micromachining, evaluation of machining performance is sometimes difficult, particularly in the case of machining precision. Drilling microholes is the simplest typical example of micromachining. However, suitable means for dimensional mea- surement are not provided by the manufacturers of measurement equipment, if the object is a hole with a diameter of less than 0.5 mm and a depth of more than 1 mm. Since recent applications often require holes of diameter 100 or 50 lm, the de- velopment of technology for measuring the internal dimensions of such holes is 4.6 Electrical Discharge Machining 283 an urgent requirement in production engineering. However, in the measurement of such small-sized objects, errors in fixing the object often strongly influence the result of the measurement. This leads to the requirement for on-the-machine measurement (OMM). OMM means the measurement is performed without removing the object from its clamped position on the machine tool which machines the microhole on the object. In OMM, no extra error is introduced after machining until measurement is performed. On the other hand, the tools for measurement must be installed on the ma- chine, because the common use of coordinates between the machining system and measuring system is essential in OMM. This type of set-up is usually difficult because of the wide gap in the design concept of the equipment between machine tools and measuring machines. A recently developed measuring technique called the vibroscanning (VS) meth- od can provide a solution to this problem in some cases of application. 4.6.5.1 VS Method The VS method utilizes a probing system to obtain the surface profiles, using the principle of electrical contact between the probe and the object. It is well known that the detection of electrical contact is not precisely reproducible. Therefore, this type of detection has not been applied for dimensional measurement. Several ex- amples of application have been found only in the positioning of workpieces in machine tools with medium accuracy. The difference between the VS method and the conventional method using elec- trical contact is that the VS method applies an intermittent contact derived from the vibration of the probe for contact. Figure 4.6-8 illustrates the basic principle of the VS method. The cantilever for surface detection is vibrated at a low frequency so as to avoid mechanical reso- nance. The amplitude is adjusted to approximately 2 lm. When the cantilever is 4 Sensors for Process Monitoring284 Fig. 4.6-8 Principle of VS method close to the surface, its tip contacts the surface intermittently. This contact pro- duces a series of rectangular pulses as output. The relationship between the probe movement and the output signal is illustrated in Figure 4.6-9. The pulse duration changes according to the distance between the center of vibration and the surface, or the x-coordinate of the probe position. If the probe position is controlled so as to maintain a constant duty cycle of the output, eg, 50%, the vibration center fol- lows the surface curve when the probe is driven in the longitudinal direction, or y-direction. Storing the x and y data of the probe position is equivalent to storing the data for the surface curve. Owing to the dynamic contact, the output reproducibility of this method can be in the submicron range. Since the deformation of the cantilever at its tip is only around 1 lm, it is possible to measure the internal dimensions of a microhole provided that the cantilever can be inserted in the hole. 4.6.5.2 Application of Micro-EDM The simple mechanism of the VS method realizes OMM. A practical example of the application of the VS method for measuring EDMed microholes is outlined below. The set-up of the measuring attachment (VS head) on a micro-EDM machine is shown schematically in Figure 4.6-10. The center lines of the machining head (z'- axis) and VS head (z-axis) are fixed in parallel at known positions on the ma- chine. After a microhole has been machined in the workpiece, the workpiece is moved horizontally to the VS head position and measurement can be performed immediately. This simple and rapid transition from machining to measurement on the same machine introduces two advantages. First, the internal surface of the microhole is fresh after machining and provides a good electrical contact for carry- ing out VS method. Second, the result of measurement can be fed back to the 4.6 Electrical Discharge Machining 285 Fig. 4.6-9 Relationship between probe movement and output signal machining system by moving the workpiece back to the machining position and performing additional machining. In other words, the dimensional error of ma- chining can be corrected before demounting the workpiece from the machine. Concerning the VS method, applications for the measurement of practical parts and improvements in reproducibility and accuracy are currently being investigated. 4.7 Welding H. D. Haferkamp, Universität Hannover, Hannover, Germany, F. v. Alvensleben, Laser Zentrum Hannover, M. Niemeyer, Universität Hannover, W. Specker, Laser Zentrum Hannover, M. Zelt, Universität Hannover 4.7.1 Introduction Quality in welding depends not only on the various process parameters such as welding speed, current, and voltage, eg, for arc welding. In most cases, the weld- ing process is affected by: · tolerances and mismatch of the workpiece geometry; · tolerances in machines and clamping devices; · variations in groove shape; · tack beads; and · welding deformations. In manual welding, trained welders are able to compensate for all these influ- ences, because their senses, especially the eyes and ears, give them the informa- tion they need to produce a high-quality weld. For mechanical and automatic welding, all this information must be detected by sensors. Such a sensor for weld- ing can be defined as follows [1]: 4 Sensors for Process Monitoring286 Fig. 4.6-10 Set-up for on-the-machine measurement . measurement of such small-sized objects, errors in fixing the object often strongly influence the result of the measurement. This leads to the requirement