Final Report Summary - µECM (Micro ECM for SMEs)
Current ECM process technology is limited in machining accuracy and process stability. The primary reason for this limitation is that the power supply units, which are at the heart of ECM are traditionally designed for the millisecond pulsed current range. Developing a power supply unit that has the capability of pulsing in the microsecond / nanosecond range, will result in more control of the process and better product, leading to improved uptimes on the shop floor and improved product quality. The SME partners in this project consortium have identified a commercial opportunity for a new generation of ECM that can offer:
- Improved process stability and control - Increased accuracy (at a micron level)
- Better process uptime and increased efficiency / outputs
- Consistently better performing and better quality products
- Shorter product development lead time This project will undertake the research and development work necessary to ensure that the SMEs in the consortium can realise this commercial opportunity through the development of a power supply unit (SMPS) that can support the next generation of ECM. The project aims to develop an SMPS that will deliver quicker electrical current pulsing rates in the microsecond / nanosecond range, resulting in an ECM with a machining accuracy better than 1 micron. The project will also develop a demonstrator / prototype that will be used for validating and demonstrating the next generation ECM. Based on research undertaken by CDAMC (in conjunction with SMPS manufacturers and suppliers), it is clear that power units in this performance range are currently not commercially available. The availability of a new generation of SMPS could result in a major step change in ECM capability and help sustain Europe’s leading position in ECM technology.
The project's main achievements are summerised as:
The development of a novel Power Supply Unit (PSU) that utilizes recently emerged industrial technology and components for developing nano second-pulses at MHz frequency ranges. This PSU has potential applications not just limited to micro ECM processes.
An ECM demonstrator that can machine copper using nano second pulses. The project has also provided a general advancement of knowledge in understanding the cause of an oxide layer that occurs when machining certain commercial materials within the nano-second range. Investigation trials using the demonstrator have provided preliminary results for setting process parameters that can discharge the oxide layer ensuring maximum efficiency and utlisation of faradic current for machining within nano second pulse range. Further process trial investigation and development work on the Demonstrator has been identified and is being continued post completion of this project lead by Sonplas with agreed support from all other partners.
Simulation technology providing an in-depth understanding of the electrical behaviour of the micro-ECM process. Based on this understanding a mathematical model was proposed and implemented to quantify the effect of the pulsing parameters on the behaviour of the micro-ECM process. Simulations and modelling showed that the capacitive behaviour of the oxide layer is a dominant effect during the micro-ECM process making the control of the process based on the total current inaccurate. One of the key simulation aspects identified during this project is the importance of the mesh quality/properties on simulation results. For 3D simulations, the meshing becomes even more important as it has a huge effect on the calculation time. This project helped devised a new non-uniform and non-isotropic surface meshing approach to generate and remesh high quality surface and volume meshes, even on poor quality CAD.
Project Context and Objectives:
The SME partners in this project consortium have identified a commercial opportunity for an affordable new generation micro-ECM sinking machine. The ECM process is traditionally viewed as a dirty process that is difficult to control, and there are concerns over the stability and repeatability that can be achieved. This stems from traditional ECM sinking machines using an electrolyte which was basically salt water and poor maintenance, and resulted in machines that were covered in salt and not very clean or nice to
look at. Many end users adopt the process initially as they cannot find another process that can machine the feature that they require. Once the process is adopted by a company they then tend to realise the potential benefits of ECM and it is then one of their key process and very often gives them a competitive advantage. The process of implementing an ECM process typically involves much iteration. Products are designed, the individual component drawings are then passed to the manufacturing department for their feedback. Quite often, in the case of ECM, trials need to be carried out to determine what level of accuracy, shape and surface finish that can be achieved. This information is then fed back to the design team with design change recommendations. This is a costly time consuming process, through the advancements proposed in this project the aim is to
• Improve the accuracy and capability of ECM to the micron level,
• develop a power supply unit will be able to control the material removal rate to a finer level.
• Develop an accurate simulation package that will shorten the time from component design to sign off by the manufacturing team.
• Develop an integrated machining head that has a power supply unit and piezo actuation capabilities.
• Bring this together in a demonstrator machine that will prove that this micro-ECM sinking machine can achieve micron accuracy and improved surface finish quality.
The work proposed by this group will result in a new approach to representing the ECM process to potential customers. The simulation package will allow designers to see very
quickly if their concept can be manufactured and from a manufacturing point of view it will give a more accurate estimate of cycle time and quality and machine capacity. The advances in the power supply unit will allow more control over material removal rate and the quality of surface finish which will allow more control of the process and therefore a more stable process for the end-user. The consortium has recognised an opportunity in fuel injection system components through enquiries received from customers of an ECM company. ECM is currently widely used in the manufacture of fuel injection systems. The main benefit of using ECM over other manufacturing processes for these products is that it allows the products to operate at higher pressure without having to make them significantly larger to withstand the final products working pressures which can be above 2000bar. Fuel systems operating at a higher pressure allows the engine to burn the fuel more efficiently and therefore reduce exhaust emissions. Automotive companies see ECM as one of the key processes that allow them to achieve the emissions legislation. In the case of medical device companies the ECM process allows intricate shapes and surface textures to be created in hard materials that could not otherwise be processed. Repeatability and accuracy of the process is key in this sector. Micro-ECM will bring together three SME’s with technology that is key to advancing the ECM process. Through their collaboration, and with the research support of the RTD providers, the
• Develop their own existing technology further by understanding the needs of the other partners.
• Understand the influence of certain feature of the process that they may not have considered previously and which might be key to stabilising the process
• Target markets and sectors that may not have been involved in before through the wider network that this collaboration will result in
• Streamline their products by having a common objective and a common goal
• Utilise the expertise available from the research providers and use this to carry out research that would not have been possible or that would have been done on an adhoc basis, improve their own know-how in certain areas and utilise the extra resource capacity.
As stated there are three key technological areas that the consortium will focus on:
1. The power supply unit (PSU). As already mentioned the power supply unit is one of the key parts of an ECM process. The problem to be overcome here is to get a PSU that allows more precise control of its pulses. The goal is to have a switch mode power supply that can work in the nanosecond range. PSU’s that are used in ECM today are limited to working in the millisecond to microsecond pulse range due to the technology that is used. As with all commercially available power supplies, certain protection features are provided in the standard PSUs used in current ECM machines. Current, voltage and temperature are all monitored for excess conditions and appropriate protection measures taken in the event of a fault. Although, some of the protection features such as over temperature are required for safety, others such as over current and over voltage are critical to the ECM process. At present there are no commercially available PSUs that have optimised these parameters for the ECM process. Additionally, currently used ECM PSUs do not provide adequate monitoring of PSU parameters such as voltage and current which are crucial to the accuracy and repeatability of the ECM process.
2. ECM process simulation. An effective ECM process simulation package will need to be developed to model what the PSU is capable off. The issues to be overcome in this caseinclude the solving of very large non-linear systems. Also advanced time stepping is of paramount importance. When using classical time stepping the number of time steps is dictated by the typically very short pulse period, burdening the simulation with a huge number of time steps to be calculated. Hybrid time stepping solves this problem by partially averaging. A small amount of accuracy is sacrificed to render the simulation independent of the small pulse period.
3. Micro-ECM machine. A working demonstrator micro-sinking ECM machine is the expected outcome of this proposed work. Through the integration of the PSU and piezo actuation for efficient machining and the design of the actual tool taken from the simulation work, a final micro sinking ECM machine tool will be produced. The key is to focus on the physical integration of the power supply unit on the machining head and minimising the inductance loop. Developing a machining head that will have a vibration frequency up to 600 Hz to aid the machining process, this should be achieved through the use of the piezo actuator. The final machine will have to be capable of maintaining a +/-1 micron accuracy, be able to drill holes of 100 micron diameter with aspect ratios of up to 10 and to machine that type of hole or feature in 25 seconds to an acceptable level of quality.
To enable the project to achieve its objectives of a commercially viable micro-sinking ECM machine with micron accuracy and improved surface finish capability and with the unique selling point of an accurate simulation package, the following objectives are key:
• Develop of a power supply unit that can deliver current pulses of 10ns duration with a maximum voltage of 30V and a maximum current of 20A.
• Develop a software simulation package that can replicate how the power supply functions and that can accurately predict material removal for given process parameters and tool shape.
• Integrate the power supply unit and piezo actuator in a new micro-ECM sinking machine with input taken from the simulation work to influence the optimum machine design.
• The final demonstrator machine must be accurate to within+/-1 μm, must be capable of drilling
100 micron holes with 1:10 aspect ratio, and a surface finish accuracy of Rt0.1micron
• To develop a new ECM product that is priced competitively, that is commercially viable and that can achieve the specifications set out.
• To secure more end-user’s for this product. DELPHI one of the worlds largest automotive part suppliers has also expressed a strong interest in the product. They will be involved in the project from an end-user point of view as will Creganna-Tactx who have expressed an interest in the technology being developed
• To ensure that the SME’s involved acquire know-how that they currently do not have, that they benefit from the new network that will be established.
• To create a niche market within the ECM sector and to diversify and compete in the micro EDM market.
• To have a product that will disrupt the EDM micro-machining market.
• To ensure that the aims of the consortium are met and for the best chance of success the consortium have worked together to define the following scientific and quantifiable technical objectives.
S&T Results for WP2 - PSU Development. WP2 Leader Vox-Power, RTD Provider IT Sligo
A prototype power supply has been produced that matches the specification that evolved from the initial specification (TA2.2) with feedback from testing and analysis performed as part of part of this work package and others. As a result of which the remaining tasks laid out in the document of work have been brought to conclusion and, with the submission of this report, all deliverables and milestones have been met. The power supply produced uses novel technology, such as delay line pulse generators, for the generation of nano-second range pulses and gallium nitrate field effect transistors, for the use of those pulses in gating a suitable supply of machining power. Use of these novel technologies required that high specification supporting circuitry be developed to ensure correct delivery and feedback measurement of the high speed signals within a timeframe that allowed relevant responses by the control system.Integration with the demonstrator required the design and manufacture of a new method of connection to the inter-electrode gap, which resulted in a PCB with a circular arrangement of sprung probes and Brunel and Sonplas’ introduction of multiple spindle electrical interface concepts.
Trials of the current sense circuit on-board were carried out using a dummy load of 1 Ω. Each test varied parameters and checked the level output by the current sense circuit for correlation to the calculated value, for that set of parameters.
Figure 1: Output comparison with varied TON at 1 V bias (L) and 2 V bias (R) (Note: All figures for WP2 contained in Appendix A)
Figure 2: Output comparison with varied TON at 5 V bias Figure 3: Output comparison with varied bias
Figure 4: Current sense input and output with 1 V bias, 100 kHz and 1000 nS TON (LEFT) and Current sense input and output with 2 V bias, 100 kHz and 14 nS TON (RIGHT)
It was seen from the results recorded that the system works within expected parameters, except for ON-times below 35 nS with a frequency of 100 kHz, where there is a significant linear fall-off of the level that is output. There was a small amount of undershoot and overshoot observed on all measurements, additionally a small amount of ringing has also been observed at the load resistor output.
Trials of the circuit on-board were carried out using a dummy load of 1Ω. Each test varied parameters and checked the level output by the peak detect circuit for correlation to the calculated value, for that set of parameters.
Figure 5: Output comparison with varied TON at 1 V bias (L) and with varied frequency at 2 V bias (R)
Rise time measured with pulse settings of 100 kHz frequency and 200 nS ON-time and a bias voltage of 5 volts, was shown to be 51 nS, including overshoot of 0.3 volts.
Fall time measured with the same pulse settings and bias voltage, was shown to be 509 µS, with no apparent undershoot.
It was seen from the results reorded that the system works within the expected parameters except for ON-times below 200 nS where there is fall-off, due to a correlating fall-off of the level supplied by the current sense circuit.
The variance in output level over the range of frequencies tested is minimal and within allowable operating tolerances.
Peak Detect Filtering
Trials of the circuit on-board were carried out using a dummy load of 1Ω. Each test varied parameters and checked the level output by the peak detect filter circuit for correlation to the calculated value, for that set of parameters.
Figure 6: Output comparison with varied TON at 1 V bias (L) and with varied frequency at 2 V bias (R)
It was seen from the results recorded that the system works within the expected parameters except for ON-times below 200 nS where there is a fall-off, due to a correlating fall-off of the level supplied by the current sense circuit to the peak detector circuit. The amount of variance in the output level over the range of frequencies tested varies and is due to overshoot on the level supplied by the current sense circuit to the peak detector circuit. Other than this variance, the filter output matches the theoretically expected operation.
With a dummy load acting as the IEG, specific voltages were induced across the IEG while the output of the voltage sense circuit was monitored by oscilloscope. The amplitude of the resulting waveform plots were measured and compared to the expected output as determined by theoretical calculation of the circuit’s gain.
It was seen from the graph created that the measured output is linear and matches the theoretically calculated output very closely, with error in the order of < ±1%.
It was seen from the results presented that the system works within the expected parameters. The variance in output level over the range of voltages tested is minimal and within a tolerance of ±1%.
OCP Level Setting:
OCP limit values were requested and the resulting DAC output level was measured on an oscilloscope.
The OCP inputs, outputs and the output of the pulse generator were monitored with an oscilloscope while the current sense value was induced to rise above the OCP limit value by increasing the bias power voltage. From the resulting plots an OCP response time could be obtained, that is, the time between the current sense value rising above the OCP limit and the time at which pulses cease to be produced by the pulse generator.
The OCP limit was set to a specific value, and the current sense value was again induced to rise above this value. The state of the OCP output was monitored as the current sense value crossed the OCP limit value. With the OCP output in the latched state the state of the OCP output was again monitored as the DAC attempted to reset the OCP latch by rising above the maximum allowable value.
OCP Alert LED:
Again the current sense was induced to rise above the OCP limit value, this time the OCP alert LED was monitored visually to observe any change in state. A second visual test was performed to ensure the LED reverts to its original state when the OCP latch is reset.
OCP Level Setting
The requested values, measured values and the difference error were recorded. A current scaling factor of 125 mV/A was used. Data points appear linear. Error is low at a maximum of ±4% at very small voltages and falling to less than ± 1% at higher voltages.
When the current sense output rises above the OCP level it causes the OCP output to switch from low to high state, preventing the pulse generator from producing any further pulses. Also noted was that the current sense line appears to continue to rise sharply as the OCP output changes. This is a result of the OCP feedback loop and allows the OCP to remain in the high state irrespective of any further changes in the current sense levels and so produces the latching functionality. This test was performed with a pulse frequency of 80 kHz, a pulse width of 400 nS, a current scaling factor of 125 mV/A and an OCP limit of 413 mA. It was seen from the measurements taken, that with the first marker positioned where the current sense value appears to first rise above the limit, the response time is 280 ns. This point is very hard to judge however. Taking that into consideration there are two other points it may be feasible to measure from, one where the OCP output begins to rise, in which case the response time is 180 ns, or from where the last pulse begins to rise after the OCP output has begun to rise, in which case the response time is 30 ns.
With the current sense output below the OCP limit, the output of the LM7171 reads -1.5 volts. When the output of the current sense is raised above the OCP, the output of the LM7171 reads 3.5 volts, a potential difference of 5 volts. The LM7171 output remains at the high state until the OCP is reset by raising the OCP limit level to be the maximum possible value
OCP Alert LED:
Initially the LED was observed to be in the inactive state. When the current sense value rose above the OCP limit the red LED turned on, indicating an OCP alert condition. When the latch was instructed to reset the LED turned off again, thus reverting to its original, non-Alert, state.
The level output by the ADC to set the OCP limit is linear in nature and correlates correctly to the requested value with a minor tolerance.
The OCP circuit latches correctly when the current sense level rises above the OCP limit level with, no discernable margin or error. The circuit can be correctly reset by setting and OCP limit above the user maximum allowable range. The worst case system response time is just under 200 nS. Whether another pulse will occur during this time is dependent on the frequency being used. The OCP Alert indicator LED lights correctly when the OCP is latched and turns off again when the latch is reset.
The possible I2C communications to and from the temperature sensor were requested from a PC controlled test program and monitored with an oscilloscope, allowing plots to be taken of the each of the waveforms. The expected change in function of the sensor after communication was also monitored.
Results showed waveforms as expected, i.e. tones that matched theoretical logic diagrams; they are not included here for brevity’s sake. Subsequent readings at different induced temperatures also showed correct plots and returned correct readings. When the induced temperature rose above the set limit, the state of the ALRT line changed as required.
As the results show, the sensor is operating and communicating as required.
DC PSU Control
Module A & B Inhibit Signal
The PSU software was started and the state of the inhibit line and the Nevo output was monitored to check its default state was HIGH, enabling the module’s inhibits.
Using PC test software, a request for a certain period of machining was made and the state of the inhibit line and the Nevo output was monitored with an oscilloscope.
DC PSU Current Control
Using the PC test software, a request for certain levels of current limit at the Nevo were made. For each level requested, plots of the related PWM output and voltage drop across resistor R36 were taken on the oscilloscope and also monitored with a voltmeter.
DC PSU Voltage Control
Using the PC test software, a request for certain levels of voltage at the Nevo were made. For each level requested, plots of the related PWM output and voltage drop across resistor R26 were taken on an oscilloscope and also monitored with a voltmeter. The voltage output by the Nevo was also monitored with a voltmeter.
Module A & B Inhibit Signal
During the initial idle state the inhibit line was set to a HIGH state (3.3 V) thus enabling the modules inhibits and the Nevo output was observed to be off using a voltmeter.
When a period of machining was requested a LOW (0v) pulse was observed on the inhibit line on the plot displayed on the oscilloscope. The output of the Nevo was monitored with a voltmeter and was observed to turn ON and OFF correctly in accordance with the inhibit signal.
DC PSU Current Control
The smallest effective ΔICtrl setting observed is equal to 375 mA, which allows for 100 different ICtrl Levels, meaning there is an effective resolution of approximately 6.5 bits.
DC PSU Voltage Control
The smallest effective ΔVCtrl setting observed is equal to 112 mV, which allows for close to 89 different VCtrl Levels, meaning there is an effective resolution of approximately 6.5 bits.
The Inhibit signal is working as required and the current and voltage control signals are working within allowable tolerances. For operation above 1.1 volts the Nevo voltage output matches the VCtrl setting closely and appears linear.
Pulse Width Modulated Frequency Generator
The resulting generated pulse is able to be varied between 10 kHz and 45 MHz with a minimum pulse width or TON of 14ns and a maximum of about 9000 nS.
At lower frequencies some trigger skipping has been seen with pulses that approach 100 percent duty cycle, this has been traced to pulse generator pulse stability tolerances and can be averaged to occur only from the 98th percentile upwards.
The circuit works as intended. However it does not quite reach the required specification maximum frequency of 50 MHz due to the pulse generator’s inherent delays. The achievable limit of 45 MHz was put to the consortium members and after discussion was agreed to be acceptable. Another point to note is the increased noise at the higher frequencies; this is disappointing but not unexpected with frequencies this high. This noise and its effect may possibly be reduced in further circuits by filtering and better use of low impedance power planes and short signal traces.The error in actual versus requested pulse width can be reduced such that it becomes less than ±1%.
Bias Power Supply Distribution
The inputs and outputs of the bias supply distribution and regulator circuits were monitored, with an oscilloscope, under a range of different conditions, with a dummy load attached in place of an actual IEG. The resulting waveforms were then compared with the previously calculated theoretical values.
Control and Pulser Bias Supply
The input and output voltages of the control and pulser bias power supply circuit matche the expected operation of a 5 volt DC voltage being regulated from a 7 volt DC supply voltage.
Isolated Bias Supply Distribution
The pulse train at the gate and source of the bias supply distribution level shifter matches the expected operation of 3.3 volt pulse train switching a supply of greater than 5 volts.
The voltage across the A and D windings of the transformer matches the expected operation of a primary winding (A) pulse of greater than 5 volts inducing a similar pulse in the sense circuitry winding (D) with double the voltage level.
Power Stage Bias Supply
The input and output voltages of the high side power stage bias power supply circuit match the expected operation of a 7 volt DC supply being regulated to a 5 volt DC voltage.
The distribution circuit and the four bias supplies’ operation all match the expected operation with a dummy load and a range of operating parameters.
Power Stage Gate Drive
Oscilloscope plots were recorded at the outputs of both the low and high side gate drives, for various drive signals and compared against the expected output.
Across the test range it was seen that as higher frequencies (> 5 MHz) and narrower pulse widths (< 100 nS) were approached, the pulse shape began to deteriorate. However, the actual effective pulse width matched that of the requested nominal width across all tests.
While the pulse shape at higher frequencies and narrower pulse widths is not as well formed as may be desired, it is sufficient to correctly drive the eGaN FETs and thus it can be allowed that the power stage is operating as required.
ECM Machine Trials
The PSU was mounted on the micro-ECM demonstrator machine and test settings initiated to allow the output of the PSU to be observed under normal conditions and under alert conditions.
The PSU was determined to be correctly integrated with the demonstrator machine and communications with the machine’s control system were observed to be operating as intended.The pulses observed at the gap were of the expected shape, size and quality, with little deterioration. The induced process faults or alert conditions were detected and reported as expected with the appropriate changes to the process observed.
The PSU is operating as required.
• S&T Results for WP3 – ECM Simulation Software. WP3 Leader Elsyca, RTD Provider VUB
The main outcome from the project regarding the simulation technology is an in-depth understanding of the electrical behaviour of the micro-ECM process. Based on this understanding a mathematical model is proposed and implemented (D3.5) to quantify the effect of the pulsing parameters on the behaviour of the micro-ECM process. Furthermore simulations and modelling show that the capacitive behaviour of the oxide film is a dominant effect during the micro-ECM process making the control of the process based on the total current inaccurate.
Additional measurements and quantification of the polarization data is required to increase the accuracy of the simulations. Furthermore the proposed algorithm for modelling the effect of the pulses using a quasi-DC approach needs to be refined to allow for shape change simulations.
One of the key aspects identified by Elsyca during this project is the importance of the mesh quality/properties on simulation results. As we focus on 3D simulations, the meshing becomes even more important as it has a huge effect on the calculation time. Elsyca devised a new non-uniform and non-isotropic surface meshing approach to generate and remesh high quality surface and volume meshes, even on poor quality CAD.
The main S&T results for WP3 are summarised below:
1. A software that enables to model real time µ-ECM considering pulses in the nanosecond range. The current density distribution is calculated while considering the electrolyte resistance, the polarisation the influence of the charging and un-charging the double layer capacitance. Two-dimensional and axi-symmetrical configurations can be simulated. Electrode shape change is possible but not performed. Simulations would simply take too much time. The aim of this software is to study the parameters that influence the confinement of metal removal.
2. A software that makes use of time averaging techniques in order to make electrode shape change simulations possible without loss of accuracy. Two-dimensional and axi-symmetrical configurations can be simulated. The averaging is making use of the different time scales that are playing a role.
3. Based on a one-dimensional approximation, that is valid along a current line, a simplified analytical model was developed that provides already qualitative estimations on the possible confinement that can be obtained locally.
4. The software mentioned in 1 was adapted to make it possible to take into account the capacitive effect of oxide layers that are present on work pieces that have oxide layers. With this software it was possible to elucidate the unexpected issue that was encountered during testing. We showed that forward polarisation could overcome the problem.
5. The outcome of 2 revealed that also time averaging would not sufficiently reduce the simulation time for real applications. Therefore, a theory was developed that makes it possible to calculate (µ-)ECM pulsing at millisecond up to nanosecond range. This theory is based on a two-step calculation, one that defines the electrolyte resistance seen at each point of the electrodes and one that makes use of this resistance to calculate a local pulse averaged current density or removal rate.
6. A measuring device was suggested (constructed by Sonplas) that would make it possible to characterise the electrochemical behaviour of electrode material. This is needed as input for all type of simulations and machine settings.
S&T Results for WP4 – Design Test and Validate Micro ECM. WP4 Leader Sonplas, RTD Providers Brunel and CDAMC
1. Micro ECM process was interesting as new approach in micro manufacturing 3 years ago when the project was initiated and today it is still very attractive area of constantly growing research initiatives.
2. Design of unique equipment on which to be used the micro electrochemical process poses an interest in other companies not only the partner companies.
3. Designed machine tool is on modular principal and can and will be used to accommodate also different processes like micro milling, micro EDM.
In micromanufacturing, the ultraprecision machine tool axes must be friction free, without stick-slip, no backlash, smooth, easy maintenance and high acceleration capabilities to provide high accuracy motion and positioning at high or low machining speeds.
In all latest ultraprecision machine tools, direct drive systems are implemented. The usage of drive system which converts rotational movement into linear movement is not popular now. With direct drive systems, there is no more mechanical transmission which results in wear. Faster and accurate motion will be achieved using direct drive systems.
The bearings of the drives are equally important. Before, contact bearings such as roller bearings were widely implemented in machine tools. Roller bearings are no longer seen as suitable at high cutting speeds. Even oil lubricated bearings could not cope in high speed machining as there would be too much heat generated. Axes with air bearings are the best solution and becoming the norm in the latest commercial ultraprecision machine tools.
A new technique in air-bearings design is called the groove technique. This design merges aerostatic and aerodynamic design principles for ultra-high speed performance optimization. By feeding pressurized air through the orifice restrictors, aerostatic lift is generated.
These linear axes or slideways manufactured by Loadpoint Bearings Limited are driven by three phase DC brushless linear motors. Optical linear encoders by Renishaw were used as feedback for the all linear axes with a resolution of 20 μm before digital interpolation and 2 nm after digital interpolation to provide motion accuracy of less than 1 μm over the total length of travel.
Slides were designed to reflect the need of the accuracy required. The slides are novel design of air-bearing slides with an attempt of increasing the rigidity of the slide to withstand greater forces and stability.
• The structure of the machine tool is from natural granite in order to reduce severely the thermal deformation in the system and therefore to preserve the high accuracy achieved by the slides.
• The fixture is furnished with anti vibration pads in order to avoid influence of low frequency vibration on the machine accuracy.
• The weight of the Z slide is supported by two frictionless air cylinders to compensate for the weight of the slide and accompanied with two air brakes.
Figure 11: Housing of the machine tool was designed to fit the normal look of the machines produced by Sonplas
4. Development of the control algorithm:
Proposed control logic
➢ Fuzzy logic controlled (FLC) feedrate of the tool: according to the measured current the FLC determines a step size that is sent to the motion controller
➢ Integration of the overcurrent protection to retract the tool in case of short circuits
➢ Current is the main sensing parameter: it helps evaluate the gap size
The behaviour of the system has been simulated using Matlab/Simulink.
The simulated fuzzy logic controlled system has been implemented on a test rig to verify the viability of this control approach.
A control program has been developed using LabVIEW running on a desktop PC.
The PC was interfaced with:
• The motion controller (controlling the position of the tool-electrode)
• The power supply (applying short voltage pulses at high frequency)
• A data acquisition module
The process can be decomposed in different stages which require different control algorithms :
a) First stage of the process: the interelectrode gap (IEG) initialization
The interelectrode gap size is a very important parameter in electrochemical micromachining and must therefore be defined precisely.
When the IEG initialization programme is started, the PSU is switched on, the over current protection is set to a low threshold and the tool is moved down step by step towards the workpiece. When contact is made, the over current protection is triggered and the power supply stops pulsing. Signals are sent to the PC to get the state of the over current protection (OCP) (i.e. triggered (ON) or not triggered (OFF)) and the status of the power supply (i.e. pulsing (ON) or not pulsing (OFF)). When the PC receives the ON signal from the OCP, it stops moving the tool forward and retracts it step by step - with a 0.1µm steps length - until the electrical contact between the tool and the surface of the workpiece disappears.
The programme flow chart can be described with the following diagram:
b) Second stage of the process: gap measurement and fuzzy logic controller configuration.
Once the gap is initialized, electrolyte starts flowing and pulses are applied to measure the value of the current flowing through the gap.
The user can then choose if this value is acceptable or not. If it is acceptable, the measured current is set as a target for the fuzzy logic controller. The Minimum machining current is set to 70% of the value of the measured current. The maximum machining current is set to 130% of the value of the measured current.
The over current protection is also set at this stage: its value will be equal to 150% of the value of the measured current value.
c) Third stage of the process: inter-electrode gap control throughout machining time via a fuzzy logic algorithm embedded in a state machine
After the fuzzy logic controller configuration, the actual control programme can be started and machining can occur. The control programme consists of a fuzzy logic controller embedded into a state machine to handle the over current protection triggering events.
The PC communicates with the motion controller to move the stage up/down according to the measured value of the current (through the current sensor).
Control system summary :
• The control algorithm consists of a fuzzy logic algorithm embedded in a state-machine allowing it to react to unexpected events such as short circuits.
• This fuzzy logic algorithm is based on monitoring and maintaining constant current through the inter-electrode gap.
• The position of the tool electrode is controlled with respect to the variation between the desired value of the machining current and its actual value.
• The LabVIEW programme has then be translated in C language and adapted to the Delta Tau Power PMAC motion controller. The control system used is the latest model from Delta Tau (Power PMAC). This control system allowed the implementation of the complex control algorithms and knowledge gained will be also used for the control of other processes and machines.
The Universal Motion and Automation Controller (UMAC) by Delta Tau Data Systems Inc. is an open architecture controller (OAC) and is used as the machine control unit (MCU) of the micromilling machine. Currently, the controller has employed the most basic main processor which is a Power PC.
The UMAC controller does all the interpretation, interpolation, servo loop computation, position control, etc. The controller could also be called Numerical Control unit. The UMAC has interfaces for amplifiers (drives) and feedback from encoders and would make the system a closed loop system. There are input and output (I/O) interfaces on the controller which are used to drive or obtain signal feedback for auxiliary equipment or functions which are controlled by PLC programs. In the controller, lie the servo algorithms (PID plus feedforward).
The µECM machine employs pulse width modulation (PWM) amplifiers for all the linear axes. The PWM amplifiers are products from Delta Tau Data Systems Inc. Each amplifier has two channels, which makes it possible to drive two motors with one amplifier. PWM amplifiers are used instead of linear amplifier due to the fact that linear amplifiers consume more power and PWM amplifiers are able to meet the motors performance requirements.
Brunel has developed the control system of the machine, which involved the following achievements:
• Supervision of the assembly of the electrical cabinet of an electrochemical micromachining machine.
• Setup and gain tuning of the closed loop control for ultra precision DC brushless linear motors (on air bearings) using a Delta Tau Power PMAC motion controller to achieve optimum performances on the X,Y and Z axis.
• Conception and development of a Human-Machine Interface (HMI) in C programming language using NI LabWindows/CVI. This HMI connects directly to the motion controller using the Telnet communication protocol and allows the user to setup and monitor the machining current, tool position and other essential machining parameters.
• Set up of the communication between the High frequency Power Supply unit - delivering the voltage pulses necessary to machine the workpiece - and the HMI.
• Integration of the HMI into the proprietary standard Sonplas software using LabWindows/CVI.
• Programming of the Delta Tau Power PMAC controller to automate the setup of motors at power up.
• Development of fuzzy logic control and finite state machine algorithms (in C) on the Delta Tau Power PMAC to control the tool position during the initialization of the interelectrode gap and the machining process.
• Realization of machining trials to validate the control strategy and refine the control algorithms for better machining efficiency and surface finish.
• Training of the Sonplas employees in order to use the developed software.
Figure 20: Screen shot of the developed HMI of the machine for the company Sonplas GmbH
Figure 21: Picture of the µECM machine in Sonplas
Machining test results:
A series of tests have been undertaken to study the efficiency of the control strategy. The machined parts were made of the following materials: copper, 18NiCr6, stainless steel, iron, chromium based alloys.
Holes were drilled using NaNO3 electrolyte and 10V pulses with a duty cycle of 1:4 and a frequency of 120 kHz.
Figure 22: Picture of a part being machined
Figure 23: Picture of holes drilled in fuel injectors (chromium- based alloy)
Figure 24: Picture and profile of a hole machined in a fuel injector with the sidewalls of the tool non-insulated.
For the first time, fuzzy logic has been successfully applied to control a Micro-ECM process. The above results show that fuzzy logic is a viable control method for maintaining the inter-electrode gap throughout machining time. The present lack of localization is mainly due to the fact the frequency applied during machining was quite low (120kHz). Moreover, to improve localization, the sidewalls of the cathode should be coated with insulating material.
A spindle was designed and built with the aim to conduct high frequency pulses and to minimize the RF emitted
Figure 25: Spindle
WP2 - PSU Development. WP2 Leader Vox-Power, RTD Provider IT Sligo
This project work package has facilitated knowledge transfer between IT Sligo and Vox Power allowing Vox Power the information for a novel power supply technology applicable to micro-electrochemical machining and possibly other pulsed power supply applications. This knowledge transfer has also assisted in the advancement of knowledge, and qualification, of some of those involved in IT Sligo in new areas. One of the IT Sligo researchers employed for this project has now taken up employment with Vox-Power.
WP3 – ECM Simulation Software, WP3 Leader Elsyca, RTD Provider VUB
Based on the scientific results obtained in WP3, Elsyca is able to extend its 3D modelling capabilities to simulate micro-ECM processes. Of course the potential impact of the modelling software is directly linked to the success of the equipment/hardware.
A similar approach as developed here for incorporating the effect of the pulsing on the macro-scale distribution can be used to improve the Elsyca modelling capabilities for high frequency electroplating processes. This would allow Elsyca to increase significantly the range of applicability of the simulation technology, as many new processes now are based on high frequency pulsing. Typical examples are plating of copper in through holes for printed circuit boards and micro/nano structures on semiconductor wafers.
For VUB the main impact of the results is situated on a scientific level. The obtained results increase the knowledge on electrochemical modelling on several aspects: time dependent electrode processes, time averaging considering different time scales, electrode shape change technology. These results can be used in the framework of other electrochemical processes (see below).
The results are integrated in advanced courses on numerical modelling.
• 219th ECS Meeting, Montreal, Canada, May 1-6, 2011.
• MMLive, Birmingham, UK, September 27-29, 2011.
• INSECT, Vienna, Austria, November 3-4, 2011.
• INSECT, Krakow, Poland, October 18-19, 2012.
• ISEM XVII, Leuven, Belgium, April 9-12, 2013.
• Simulation of nano-second pulsed phenomena in electrochemical micromachining processes – Effects of the signal and double layer properties, Electrochimica Acta, Volume 93, 30 March 2013, Pages 8–16.
• VUB intends to publish its results in an additional number of journal papers. This will be done after the end of the project.
• Simulation of electrochemical micromachining with nanosecond pulses, in proceedings 219th ECS Meeting, Montreal, Canada.
• Simulation of electrochemical micromachining with nanosecond pulses, MMLive, Birmingham, UK.
• Simulation of electrochemical micromachining with nanosecond pulses, in proceedings, INSECT 2011, Vienna, Austria.
• Time accurate simulations of electrochemical micro-machining with nanosecond pulses, in proceedings, INSECT 2012, Krakow, Poland.
• Time-efficient simulations of nano-pulsed electrochemical machining, in proceedings ISEM XVII, Leuven, Belgium.
• A time-efficient numerical approach to electrochemical processes - Application to nano-second pulsed electrochemical micro-machining, June 2013.
Application of results
VUB will apply the developed software to the simulation of corrosion processes. In fact ECM is a controlled corrosion process. Software for time averaging and electrode shape change are (re)usable and/or adaptable. There is a growing interest in corrosion simulation tools and VUB SURF is participating actively in that field. The final aim is to come to reasonable (life time) predictions of corroding components or structures (in cars, tooling, building, etc.) in different circumstances. The theoretical developments mentioned in result 5 can be implemented in the industrial ECM simulation software of the Elsyca.
WP4 – Design Test and Validate Micro ECM. WP4 Leader Sonplas, RTD Providers Brunel and CDAMC
This can be considered in few groups:
- Communications and engagements
1. The project put together people from different background and different specialities, to work and develop a new product.
2. The web site helped to spread the word and we had number of enquiries regarding the development
3. During the annual workshops at the university we always mentioned and demonstrated the result of the project to number of companies from the UK and from all over the world.
4. Each of the partners gained knowledge from the pre-existing knowledge and expertise of the other partners and also gaining knowledge together during the working meetings when hundreds of tests were done.
- Collaboration and co-production
5. This project helped us at Brunel to have established name in the UK for development of non-traditional manufacturing technologies and related equipment.
6. Also we had the chance to gain our place in the European developers of precision equipment.
7. Advances in knowledge achieved from this project will help contribute to a new FP7 project we are taking part in devoted to minimizing the defects in micro manufacturing technologies (grant agreement 285614) where there 17 partners from across Europe.
- Exploitation and application
8. The results achieved (good and bad) will be used in our further development of the equipment and the technology
9. Use of different workpiece materials and different electrolytes are already in progress.
10. New application areas like drilling holes in the turbine blades for Rolls Royce are under tests
11. Developed by NPL and IBS micro CMM machines require manufacturing of highly accurate styli for the micro CMM probing systems. The developed technology is probably the best suited for this purpose and first attempts to machine micro styli is commencing.
Potential impacts of the project at Sonplas
The project has not provided any socio-economic impact so far. The required activities have been covered with existing work forces. Since dissemination activities are still on going, a future impact will not be excluded in case that interesting customers will provide samples to Sonplas for tests on the demonstrator.
Main dissemination activities
Several dissemination activities have been developed during the project run time. Sonplas has mainly focussed on presenting this new technology to customers, having actual and future need to employ the µECM process on their products. The µECM-Projects was presented to the following companies :
• Magneti Marelli
Due to the project delay (the demonstrator has been delivered incomplete and late from the RTD-partner) it was not possible so far to officially process customer sample parts, since the basic functionality of the demonstrator has still to be improved.
Exploitation of results
Sonplas is still working on the further optimization of the demonstrator, to improve the efficiency and the understanding of this complex process. Therefore Sonplas will start an activity that will last beyond the official project to improve the demonstrator function closer to industrial applications.
If it can be successfully demonstrated on customer parts that µECM is a competitive alternative to existing micro mechanical drilling processes, an exploitation is still planned. But it must be committed, that during the past 3 years the competitive laser drilling technology has improved significantly and seems to be favoured from the automotive industry.
A further exploitation is considered for the aircraft engine industry like the drilling of cooling holes into turbine blades.
Also of interest are high pressure exposed components of Diesel pumps like the so called “hydraulic head” where traditional ECM is often applied, but is not longer fulfilling the requirements to surface roughness.
List of Websites: